EPA-660/2-73-019
DECEMBER 1973
Environmental Protection Technology Series
Color Removal from
Kraft Mill Effluents
by Ultrafiltration
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series. These five broad
categories were established to facilitate further
development and application of environmental
technology. Elimination of traditional grouping
was consciously planned to foster technology
transfer and a maximum interface in related
fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. This series
describes research performed to develop and
demonstrate instrumentation, equipment and
methodology to repair or prevent environmental
degradation from point and non-point sources of
pollution. This work provides the new or improved
technology required for the control and treatment
of pollution sources to meet environmental quality
standards.
-------
EPA-660/2-73-019
December 1973
COLOR REMOVAL FROM KRAFT MILL
EFFLUENTS BY ULTRAFILTRATION
by
H. A. Fremont
D. C. Tate
R. L. Goldsmith
Project S800261
Program Element 1B2037
Project Officer
Mr. Edmond P. Lomasney
Environmental Protection Agency
Atlanta, Georgia 30309
prepared for the
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price $i*i
-------
EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents necessarily
reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or
commercial products constitute endorsement or recommenda-
tion for use.
-------
ABSTRACT
Reduction of color in pulp mill effluents by ultrafiltration
has been examined with a 10,000 gallon per day (gpd) pilot
plant. Treated streams included Decker effluents and pine
bleachery caustic extraction filtrate, which together com-
prise about 80% of the color from a bleached kraft mill.
High color removal (90-97%) was demonstrated when operating
at water recovery ratios of 98.5-9970. Pilot plant capacity
(membrane flux) was 15-20 gal./day-ft^ when operation pro-
ceeded smoothly. However, plugging of the membrane car-
tridges by residual particulates (even after precoat fil-
tration) was troublesome.
Several prefiltration, concentrate disposal, and water
reuse alternatives were evaluated.
Full-scale plant designs, and approximate capital and
operating costs were estimated for systems of 1 and 2 MM
gpd capacity. Capital costs are about $700,000 for a 1 MM
gpd plant, and $1,200,000 for a 2 MM gpd plant. Corresponding
operating costs are about 45c/Mgal. (1 MM gpd) and 38c/Mgal.
(2 MM gpd).
Additional pilot plant tests are recommended to demonstrate
long-term solutions to the particulate problem and long-
term membrane cartridge life.
This report was submitted in fulfillment of Project Number
S800261, by Champion International Corp. under the partial
sponsorship of the Environmental Protection Agency. Work
was completed as of May, 1973.
111
-------
TABLE OF CONTENTS
Page
I. SUMMARY AND CONCLUSIONS 1
A. NATURE OF PROBLEM 1
B. NATURE OF STREAMS TREATED 1
C. PROCESS CONCEPT 2
D. PROGRAM RESULTS 3
E. FULL-SCALE PLANT DESIGN AND PROCESS COSTS 5
II. RECOMMENDATIONS 7
IIII. BACKGROUND 9
A. NATURE OF PROBLEM AND PROJECT GOALS 9
B. PROCESS DESCRIPTION 12
IV. PILOT PLANT DESCRIPTION 25
A. GENERAL 25
B. PRETREATMENT SEQUENCES 28
C. OTHER SYSTEM MODIFICATIONS 30
D. DESCRIPTION OF MEMBRANE CARTRIDGES 33
E. SAMPLING AND ANALYTICAL PROCEDURES 50
V. RESULTS AND DISCUSSION 53
A. FEED PRETREATMENT AND CHARACTERISTICS 53
B. REJECTION DATA AND EVALUATION 80
C. ULTRAFILTRATION RATE DATA AND
EVALUATION 88
D. PRESSURE DROP DATA 115
-------
TABLE OF CONTENTS
(continued)
Page
E. IMPORTANT FACTORS CONTROLLING MEMBRANE
FLUX 121
F. CLEANING PROCEDURES AND EFFICIENCY 125
G. MODULE MECHANICAL FAILURES 129
H. INCINERATOR STUDIES 134
VI. FULL SCALE PLANT DESIGN AND COSTS 139
A. FIRST STAGE PINE CAUSTIC EXTRACTION
FILTRATE 139
B. DECKER EFFLUENTS 173
VII. WATER REUSE 189
A. PINE CAUSTIC EXTRACTION FILTRATE
PERMEATE 189
B. DECKER EFFLUENT PERMEATES 189
VIII. ACKNOWLEDGEMENTS 191
IX. REFERENCES 193
APPENDIX A - DETAILED PILOT PLANT PROCESS
DESCRIPTION 197
APPENDIX B - ULTRAFILTRATION PILOT PLANT
SPIRAL MEMBRANE CARTRIDGE
IDENTIFICATION 217
APPENDIX C - DETAILS OF FILTERS USED IN
PILOT PLANT PROGRAM 223
APPENDIX D - MECHANICAL PROBLEMS OF DIFFERENT
SPIRAL MEMBRANE CARTRIDGES 231
APPENDIX E - SLIME ANALYSIS 237
VI
-------
FIGURES
Pa^e
1. SIMPLIFIED ULTRAFILTRATION FLOW SCHEMATIC 13
2. CAUSTIC EXTRACTION FILTRATE—FLOW SCHEMATIC 18
3. TYPICAL MATERIAL BALANCE FOR TREATMENT OF
CAUSTIC EXTRACTION FILTRATE BY ULTRAFILTRATION 19
4. DECKER EFFLUENTS—FLOW SCHEMATIC 21
5. TYPICAL MATERIAL BALANCE FOR TREATMENT OF
HARDWOOD DECKER EFFLUENT BY ULTRAFILTRATION 22
6. SIMPLIFIED PILOT PLANT FLOW SCHEMATIC 26
7. FEED PRETREATMENT FLOW SCHEMATIC 29
8. BAUER HYDRAS IEVE 34
9. TOP OF 500 GAL. FEED TANK 35
10. ACID PUMP AND ACID DRUM 36
11. DETERGENT MIXING TANK AND PRECOAT MIXING
TANK WITH MIXER 37
12. PRETREATMENT SEQUENCE 38
13. DETAILS OF SPARKLER FILTER 39
14. SHRIVER FILTER PRESS 40
15. 100 GAL. FILTRATE SURGE TANK 41
16. SINGLE CARTRIDGE AND FILTERS 42
17. SPARKLER VELMAC DISC FILTER 43
18. ULTRAFILTRATION UNIT 44
19. ULTRAFILTRATION UNIT (TOP VIEW) 45
20. DETAILS OF CONTROL PANEL RIGHT SIDE 46
21. DETAILS OF CONTROL PANEL LEFT SIDE 47
VII
-------
FIGURES
(continued)
Page
22. SPIRAL WOUND MODULE DESIGN 48
23. SULFURIC ACID REQUIREMENT TO NEUTRALIZE
CAUSTIC EXTRACTION FILTRATE 54
24. SULFURIC ACID REQUIREMENT TO NEUTRALIZE
PINE DECKER EFFLUENT 55
25. SULFURIC ACID REQUIREMENT TO NEUTRALIZE
HARDWOOD DECKER EFFLUENT 56
26. EFFECT OF pH ON COLOR OF PINE CAUSTIC
EXTRACTION FILTRATE 58
27. SUSPENDED SOLIDS OF CAUSTIC EXTRACTION
FILTRATE 70
28. SUSPENDED SOLIDS OF PINE DECKER EFFLUENT 71
29. SUSPENDED SOLIDS OF HARDWOOD DECKER EFFLUENT 72
30. TOTAL SOLIDS OF PINE CAUSTIC EXTRACTION
FILTRATE AND PINE DECKER EFFLUENTS 73
31. TOTAL SOLIDS OF HARDWOOD DECKER EFFLUENT 74
32. COLOR OF PINE CAUSTIC EXTRACTION FILTRATE
AND PINE DECKER EFFLUENT 75
33. COLOR OF HARDWOOD DECKER EFFLUENT 76
34. EFFECT OF FEED PRETREATMENT ON COLOR OF
PINE CAUSTIC EXTRACTION FILTRATE 78
35. REJECTION AND CONVERSION DATA 81
36. MEMBRANE FLUX FOR STAGE la 94
37. MEMBRANE FLUX FOR STAGE Ib 95
38. MEMBRANE FLUX FOR STAGE Ic 96
39. MEMBRANE FLUX FOR STAGE 2 97
Vlll
-------
FIGURES
(continued)
Page
40. MEMBRANE FLUX FOR STAGE 3 98
41. MEMBRANE FLUX FOR STAGE 4 99
42. MEMBRANE FLUX FOR STAGE 5 100
43. EFFECT OF PREFILTRATION EFFICIENCY ON
ULTRAFILTRATION RATE 102
44. ULTRAFILTRATION RATE DATA: STAGE Ib 108
45. ULTRAFILTRATION RATE DATA: STAGE 2 110
46. FLUX VS. TIME FOR STAGES la AND Ib 114
47. STAGE 1 PRESSURE DROP 116
48. STAGE 2 PRESSURE DROP 117
49. STAGE 3 PRESSURE DROP 118
50. STAGE 4 PRESSURE DROP 119
51. STAGE 5 PRESSURE DROP 120
52. CLEANING EFFICIENCY OF STAGE Ib MEMBRANES 130
53. CLEANING EFFICIENCY OF STAGE 3 MEMBRANES 131
54. STAGE 3 COMPACTION CURVE 132
55. SIMPLIFIED FLOW SCHEMATIC: TREATMENT OF
PINE CAUSTIC EXTRACTION FILTRATE 140
56. FLOW SCHEMATIC FOR TREATMENT OF PINE CAUSTIC
EXTRACTION FILTRATE: LOW FLOW CASE 141
57. FLOW SCHEMATIC FOR TREATMENT OF PINE CAUSTIC
EXTRACTION FILTRATE: HIGH FLOW CASE 142
58. SIMPLIFIED FLOW SCHEMATIC: TREATMENT OF
DEKCER EFFLUENTS 174
ix
-------
TABLES
1. WASTE CHARACTERISTICS AT THE NORTH CAROLINA
MILL 11
2. FILTERS USED 31
3. CHARACTERISTICS OF MEMBRANE CARTRIDGES 49
4. ANALYTICAL PROCEDURES 51
5. PARTICLE SIZE DISTRIBUTION OF SUSPENDED
SOLIDS IN PINE CAUSTIC EXTRACTION FILTRATE 61
6. PERFORMANCE - MAIN FILTERS 63
7. TESTS WITH DIFFERENT FILTER AIDS ON SPARKLER
15.3 SQ FT LEAF FILTERS 67
8. SUMMARY OF FEED CHARACTERISTICS 69
9. WASTE CHARACTERISTICS AT THE NORTH CAROLINA
MILL - DETAILED ANALYSES 79
10. REJECTION 82
11. COMPOSITION DATA: PINE DECKER EFFLUENT
SAMPLES 85
12. COMPOSITION DATA: HARDWOOD DECKER EFFLUENT
SAMPLES 86
13. COMPOSITION DATA: PINE CAUSTIC EXTRACTION
FILTRATE SAMPLES 87
14. MEMBRANE FLUX BY STAGE DURING PILOT PLANT
PROGRAM 89
15. EFFECT OF FEED PREFILTRATION ON CLEANING 127
16. CASES FOR CAPITAL COST ESTIMATES, PINE
CAUSTIC EXTRACTION FILTRATE 145
7. CHARACTERISTICS OF MEMBRANE CARTRIDGES 147
-------
TABLES
(continued)
Page
18. CHARACTERISTICS AND COSTS OF MEMBRANE MODULES 148
19. ULTRAFILTRATION SECTION DESIGN DETAILS AND
COSTS—CASE 1 153
20. ULTRAFILTRATION SECTION DESIGN DETAILS AND
COSTS—CASE 2 158
21. SUMMARY OF INSTALLED CAPITAL COST ESTIMATES
PLANT TO TREAT PINE CAUSTIC EXTRACTION
FILTRATE 162
22. OPERATING COSTS FOR TREATMENT OF PINE CAUSTIC
EXTRACTION FILTRATE, CASE 1 163
23. OPERATING COSTS FOR TREATMENT OF PINE CAUSTIC
EXTRACTION FILTRATE, CASE 2 164
24. OPERATING COSTS FOR TREATMENT OF PINE CAUSTIC
EXTRACTION FILTRATE, CASE 3 165
25. OPERATING COSTS FOR TREATMENT OF PINE CAUSTIC
EXTRACTION FILTRATE, CASE 4 166
26. OPERATING COSTS FOR TREATMENT OF PINE CAUSTIC
EXTRACTION FILTRATE, CASE 5 167
27. DAILY INCREMENTAL OPERATING COSTS, TREATMENT
OF PINE CAUSTIC EXTRACTION FILTRATE 168
28. CASES FOR CAPITAL COST ESTIMATES, DECKER
EFFLUENTS 175
29. INSTALLED CAPITAL ESTIMATES—PLANT TO TREAT
DECKER EFFLUENTS 176
30. OPERATING COSTS FOR TREATMENT OF DECKER
EFFLUENT, CASE 6 178
31. OPERATING COSTS FOR TREATMENT OF DECKER
EFFLUENT, CASE 7 179
XI
-------
TABLES
(continued)
Page
32. OPERATING COSTS FOR TREATMENT OF DECKER
EFFLUENT, CASE 8 180
33. INCREMENTAL OPERATING COSTS FOR TREATMENT
OF DECKER EFFLUENT, CASE 9 181
34. INCREMENTAL OPERATING COSTS FOR TREATMENT
OF DECKER EFFLUENT, CASE 10 182
35, DAILY INCREMENTAL OPERATING COSTS, TREATMENT
AND REUSE OF DECKER EFFLUENTS 183
xi i
-------
SECTION I
SUMMARY AND CONCLUSIONS
A. NATURE OF PROBLEM
At present, in kraft paper mills color pollutants are
not substantially removed by conventional biological
waste treatment methods. The available means, such as
lime precipitation and carbon or resin adsorption are
highly expensive.
In the Champion Papers' North Carolina mill, approx-
imately 607o of the mill color discharge is contained in
the bleach plant caustic extraction filtrate, and 20%
in the Decker effluents. These numbers are fairly typical
for a bleached kraft mill. Thus, reduction of color from
these two streams can greatly reduce the overall color
discharge from a kraft mill.
The purpose of this program has been to examine ultra-
filtration as a means of reducing color in kraft mill
effluents more efficiently and/or more economically than
the presently available methods. The scope of the
program included the six month operation of a 10,000 gpd
pilot plant at the Champion Papers' North Carolina mill.
B. NATURE OF STREAMS TREATED
The major experimental effort dealt with treatment of
pine caustic extraction filtrate, with lesser emphasis
placed on pine and hardwood Decker effluents. All
three streams are highly alkaline (pH 10-12) and hot
(120°F-135°F), and require neutralization (to pH7)
and cooling (to 100°F-105°F) before treatment by
ultrafiltration. These limits are imposed by the
characteristics of current cellulose acetate
membranes, which will not exhibit long (economical)
life if exposed to high pH and high temperature.
-------
These streams also contain substantial quantities of
particulates (e.g. 100-300 ppm), of which 50% are
smaller than 10/j. To obtain acceptable membrane
equipment performance it is essential to employ mem-
brane equipment (configuration) which is not susceptible
to plugging by particulates. In addition, flocculation
occurs in a freshly filtered feed on "aging".
C. PROCESS CONCEPT
Ultrafiltration, can selectively concentrate color bodies
and organics contained in the streams. A simplified
flow schematic is shown below:
Organic
O Concentrate
5-20% organics
0.5-2% ash
,:^!
Feed to ^
Ultrafilter \f^^T' ^ f Semipermeable
0.1-0.2% organics " TMembrane
0.4-0.6% ash
Decolorized
Permeate
—0.01% organics
0.3-0.5% ash
Desirable features of the process (compared to reverse
osmosis) are:
o Since only 5-30% of the dissolved solids are
returned in the concentrate, very high water
recovery (e.g. 99%) can be achieved;
o A low-volume, high-organic-solids concentrate
is obtained, which has substantial heating value
if ultimate disposal is by burning;
o Low-pressure systems (e.g. 100 psig) can be
employed; and
o High capacity can be achieved since higher-flux
membranes can be used.
-------
Limitations of the process are:
o The effluent is not demineralized, and the
residual salt content (especially chlorine
in caustic extraction filtrate) limits reuse
potential; and
o Color removal is typically 90-97%, somewhat
lower than can be achieved by reverse osmosis.
For all three streams examined, pretreatment (neutra-
lization, temperature reduction, and filtration) is
required.
D. PROGRAM RESULTS
A pilot plant was operated from August, 1972 through
February, 1973. Operation was normally 24 hrs/day,
7 days/week. During this period four experimental
aspects of the process were evaluated:
o Feed pretreatment (especially filtration);
o Ultrafiltration (separation efficiency and capacity);
o Concentrate disposal (incineration); and
o Water reuse potential.
Feed pretreatment work focused primarily on means to
remove particulates, by surface, packed-bed, and pre-
coat filtration, to obtain satisfactory ultrafiltration
performance. Using filtration which removed particles
of about 2p, and larger, 50-80% particulate removal was
achieved.
The ultrafiltration system membranes were in a spiral
wound configuration, and were obtained from three
vendors (T. J. Engineering, Gulf Environmental Systems,
and Eastman Chemical Products). Most membrane cartridges
had the standard "mesh" flow-channel spacer; a few had
"corrugated" spacers. Operation was usually at about
100°F, 100 psig, and pH 6-7.
Color removal efficiency was satisfactory; typical
results are:
-------
7. Water % Solids In % Color
Influent Recovery Concentrate Remova1
Pine caustic
extraction filtrate 98.5-99 15-20 90-92
Pine Decker
effluent 98.5-99 5-8 95-97
Hardwood Decker
effluent 98.5-99 5-8 95-97
Capacity (membrane flux) was variable. When operation
proceeded smoothly, fluxes of 15-20 gallons/day-ft *
of membrane were achieved. At other times, fluxes were
substantially lower. Important factors which reduced
flux were:
o Reversible membrane surface fouling by colloidal
and macromolecular feed constituents. This was
reversed by water flushing and detergent cleaning;
o Irreversible membrane "compaction". This was
minor; and
o Reversible and irreversible particulate collection
within the membrane cartridges.
The latter limitation is the most serious problem encoun-
tered in the test program. Due to incomplete particulate
removal in the pretreatment operation, particulates
collected in the membrane cartridges. This was especially
severe for cartridges with "mesh" spacers, but not for
the "hydrodynamically-clean" corrugated spacers.
Manifestations of cartridge plugging were:
o Occlusion and inactivation of membrane surface,
with reduced capacity;
o High pressure drop across the cartridges which
resulted in: (a) cartridge deformation and
(b) seal failure;
o Flow maldistribution within cartridges and
between cartridges in parallel, which aggravated
cartridge plugging by particulates; and
o Various modes of module mechanical failures.
Particulate plugging can be minimized substantially by:
-------
o Operation at high flow, which prevents particulate
collection;
o Regular and efficient cartridge cleaning; and
o Use of cartridge flow-channel spacers which are
not susceptible to particulate collection (e.g.
corrugated spacers).
Concentrate disposal for pine caustic extraction filtrate
by incineration or evaporation and admixture with primary
sludge was demonstrated. Return of the concentrates
from Decker effluents to the weak black liquor system
was chosen as an optimum means of disposal.
Water reuse potential was examined. Analyses of treated
effluents and comparison to mill water standards led
to the following conclusions; treated effluent from
pine caustic extraction filtrate has limited or no
reuse potential due to a high chloride content and
some residual color; treated water from the Decker
effluents may be beneficially reused in pulp washing.
E. FULL-SCALE PLANT DESIGN AND PROCESS COSTS
A series of design cases were examined. Design para-
meters included:
o Plant capacity (1 MM gpd and 2 MM gpd);
o Membrane type (mesh and corrugated spacers); and
o Different prefiltration alternatives.
The key cost factor was plant capacity, as expected,
since treatment costs are nearly proportional to the
volume treated and not the contaminant loading.
The installed capital costs are estimated to be:
Pine Caustic Decker
Extraction Filtrate Effluents
$770,000 $690,000
$1,250,000 $1,100,000
-------
The difference is due to concentrate disposal equipment
required for the pine caustic extraction filtrate.
The total operating costs (including amortization)
are estimated to be:
Pine Caustic Decker
Flow Extraction Filtrate Effluents
1 MM gpd 46C/M gal. = 44(?/M gal. =
$0.58/ton bleached $0.33/ton total
pulp
2 MM gpd 35C/M gal. - 37
-------
SECTION II
RECOMMENDATIONS
The results of the project studies using the nominal 10,000 gpd
ultrafiltration pilot plant demonstrate that the processing
system is technically sound as a means for reducing color
from the pine caustic extraction filtrate of a pulp bleachery
and from the Decker effluents from pulp washing. Mechanical
difficulties, however, were encountered which should be
resolved. These problems and unknown variables are highly
critical in developing reliable cost estimates for full scale
installations.
It is recommended that pilot plant studies be continued
for the purpose of confirming the present results over a
longer time span, and further examining the cost sensitive
areas of the system. More specifically, these additional
studies could accomplish the following tasks;
1. Examine membrane cartridges from several manufacturers
for long-term reliability.
2. Verify optimum operating conditions such as feed flow
rates, pH, temperatures and operating pressures.
3. Further specify prefiltration conditions for each
influent.
4. Test any new membranes that may become commercially
available and which could afford the elimination of
the neutralization and cooling treatments of the
influent.
5. Obtain more extensive operating experience using Decker
effluents.
6. The information from items 1 through 5 above could
result in redesigning a full scale plant and assessing
more reliably capital and operating cost estimates for
such a plant.
-------
SECTION III
BACKGROUND
A. NATURE OF PROBLEM AND PROJECT GOALS
The 121 kraft pulp mills in the United States produce
about 85% of the chemical wood pulp consumed. In these
pulping operations a substantial volume of wastewater
is discharged, typically about 25,000 gallons per ton
of pulp. Of concern are the pH, temperature, BOD, and
color loading of this effluent. Conventional and
generally inexpensive techniques are adequate for waste
treatment except for color removal. It may also be noted
that conventional waste treatment does not provide for
water reuse or chemical recovery and reuse, and, as such,
is not conducive to eventual close- loop operations.
Color bodies found in pulp mill wastes unfotunately are
resistant to biological degradation. The effluent color is
due primarily to lignin and its degraded products which are
chemically stable and are intractible to separation by
presently proven commercial processes. Consequently, new
treatment techniques for color removal are undergoing
active development and actual plant scale demonstration.
Promising processes developed include chemical precipitation
(1-10) , including lime precipitation (stoichiometric and
massive lime), adsorption (11-14), and reverse osmosis and
ultraf iltration (15-25) . For controlling color from
bleaching effluents, it is also possible to. modify the
bleaching sequence to CHE .... from CEH . . . ; where C=
chlorination, H =hypochlorite bleaching, and E = caustic
extraction.
Segregation of mill wastes is often practiced and it is
likely that segregation of waste streams by color will
eventually be required for adequate waste treatment.
Tertiary treatment systems which could not be considered
cost-wise for treatment of the total effluent, might be
applicable if the bulk of the color is contained in a
relatively small fraction of the mill effluent.
For example, in Champion Papers' North Carolina m
about 60% of the mill color is present in 2 x 10b gpd
-------
of bleachery first-stage pine caustic extraction fil-
trate. This flow amounts to only about 4% of the
total wastewater. Thus, if color could be removed
from this stream at total operating costs of $450 to
$800 per day, over 60% of the mill color could be re-
moved at a cost of about 0.9$ to 1.6C/1000 gal. of
total effluent or about 55$ to $1.00 per ton of
bleached pine pulp.
The second most important controllable source of color
in a kraft mill is the decker effluent. This waste is
present in all kraft mills, while the pine caustic
extraction filtrate is found only in mills producing
bleached pulp. At the North Carolina mill, approxi-
mately 2 x ID** gpd of mixed pine and hardwood decker
effluents are currently discharged. This waste con-
tributes about 20% of the mill color. Thus, decker
effluents are another case of interest for segregated
waste treatment.
Ranges of flow and composition for these streams at the
North Carolina mill are given in Table 1.
The program described in this report encompasses the
initial phase of a two-phase project for the develop-
ment and demonstration of ultrafiltration as a means of
color removal from kraft mill effluents. The site of the
project is Champion Papers' North Carolina mill which
is a representative bleached kraft mill. Phase I of
the project (the subject of this report) included on-
site operation of a nominal 10,000 gpd membrane demon-
stration plant to supplement data from previous pilot
tests and to specify more accurately design bases for a
full-scale demonstration plant. Additional studies
determined requirements for feed pretreatment prior to
ultrafiltration, disposal of the concentrated wastes
produced by ultrafiltration, and the potential for water
reuse.
Briefly the project objectives have been threefold.
1. To demonstrate with commercially-available
equipment the effectiveness of ultrafiltration to re-
duce color in first-stage pine caustic extraction
filtrate and decker effluents to low levels;
10
-------
TABLE 1
WASTE CHARACTERISTICS AT THE NORTH CAROLINA MILL
Flowrate* ,
Million
Waste
Pine Caustic
Extraction
Filtrate
Pine Decker
Effluent
Hardwood
Decker
Effluent
Gal/Day
1.3 to
2.1
0.5 to
2.7;
Avg. =
0.5
1.5 to
4.1;
Avg . =
1.5
pH
11.5
to
11.8
10.7
to
11.0
10.0
to
10.6
Color, ppm
12,000 to
50,000;
Avg. ^
28,000
5,000 to
10,000;
Avg. "*>
6,000
4,000 to
24,000;
Avg. ^
11,000
Total
Solids,
ppm
4,000 to
14,000;
Avg. ^
7,000
1,500 to
6,000;
Avg. i>
2,400
1,200 to
5,000;
Avg. ^
3,000
Suspended
Solids,
ppm
50 to
Avg. ^
50 to
Avg. 'v
100 to
Avg. ^
500;
80
250;
80
500;
200
* maximum flow is future projection
11
-------
2. To examine the potential for reuse of purified
effluents and means of disposal of the concentrated
wastes produced by the membrane process; and
3. To demonstrate that process economics will
be sufficiently attractive to lead to widespread adop-
tion of this pollution abatement process in the
industry.
B. PROCESS DESCRIPTION
1. Principles of Ultrafiltration
Ultrafiltration is a membrane process for concentration
of dissolved materials in aqueous solution. A semi-
permeable membrane is used as the separating agent and
pressure as the driving force. In an Ultrafiltration
process (Figure 1), a feed solution is fed into the mem-
brane unit, where water and certain solutes pass through
the membrane under an applied hydrostatic pressure.
The solutes whose sizes are larger than the pore size
of the membrane are retained and concentrated. The
pore structure of the membrane thus acts as a molecular
filter, passing some of the smaller solutes and re-
taining the larger solutes. The pore structure of this
molecular filter is such that it does not become plugged
because the solutes are rejected at the surface and do
not penetrate the membrane. Furthermore, there is no
continuous buildup of a filter cake which has to be
removed periodically to restore flux through the mem-
brane since concentrated solutes are removed in solu-
tion. Many Ultrafiltration applications involve the
retention of relatively high molecular weight solutes
accompanied by the removal through the membrane of lower
molecular weight impurities. Thus concentration of
specific solution components can be achieved.
Considerations important for determining the technical
and economic feasibility of Ultrafiltration as applied
to a specific process are the rate of solution trans-
port through the membrane (flux) and the separation
efficiency (rejection). Factors which control flux
and rejection have been described elsewhere (.?_§.r22).
Membrane processes for treating pulp mill wastes have
been under development for several years, focusing
primarily on reverse osmosis. In reverse osmosis all
12
-------
PERMEATE
' ' . .
' ' ' ' * MEMBEANE
• • • • - _ X
FEED Q O . * O *. ^ . o 9 • o' • O' O • Op O. O CONCENTRATE
- c . -o --
' '
' • -
.o
'
. • • , ' ' . * I. * • • • I- . ^MEMBRANE
• *
- Dissolved Salts
O Color Bodies
FIGURE 1: SIMPLIFIED ULTRAFILTRATION FLOW SCHEMATIC
-------
dissolved solutes are concentrated, and a demineralized
aqueous effluent is produced. Ultrafiltration is, in
fact, a variation of reverse osmosis. The fundamental
difference relates to the retention properties of the
membranes: ultrafiltration membranes do not retain
salts and other low molecular weight solutes.
There are several potential advantages of ultrafiltra-
tion when compared to reverse osmosis, which are:
a. The use of ultrafiltration membranes leads
to operation at lower pressures, typically 50 to
300 psig. The strength requirements of the mem-
brane system are much less stringent than those for
reverse osmosis syterns, which typically operate
above 400 psig. At the lower pressures used in
ultrafiltration, power costs are less; membranes
can last longer since the rate of "compaction" may
be reduced; and lower capital costs can be
achieved due to less demanding requirements for
pressure vessels, pumps, etc. Reliability is also
an important consideration, and system failure ob-
served in reverse osmosis tests to date can be tied
to the high operating pressures used. Important
factors have been membrane compaction (resulting
in reduced capacity), membrane catastrophic
failure (membrane support rupture), and pump
failure.
b. With reverse osmosis the feed pine caustic
extraction filtrate or decker effluent can be con-
centrated only to 8 to 10% total solids, or about
20-fold. This is due to limitations of both mem-
brane fouling and a buildup in feed solution
osmotic pressure. Both problems are of substan-
tially less importance for ultrafiltration, in which
very high volumetric concentration ratios are
obtainable, up to or exceeding 200-fold. This is
due primarily to a relatively slow increase in feed
solids content with concentration by ultrafiltration,
That is, only a small portion of the feed solids,
specifically the higher molecular-weight organics,
is retained by the ultrafiltration membrane. Note
that about 80% of the dissolved solids in the wastes
of interest are low molecular-weight salts. Since
the retained solutes have relatively high molecular
weights, osmotic pressure limitations are of minor
14
-------
importance, and solids levels up to 20% can be
achieved in low-pressure ultrafiltration.
c. Disposal cost of an ultrafiltration con-
centrate by incineration or other means would be
substantially less than that for a reverse osmosis
concentrate, since a substantially smaller volume
must be treated. For example, in treating
2 x 106 gpd, about 100,000 gpd of reverse osmosis
concentrate would be generated, but only 10,000
gpd of ultrafiltration concentrate. Furthermore
the high organic content of the latter provides
substantial heating value, almost sufficient in
itself to sustain combustion. A major fuel cost
would be required to burn a reverse osmosis con-
centrate which contains primarily inorganics. In
addition a high chlorine content in a reverse
osmosis concentrate of pine caustic extraction fil-
trate could cause severe corrosion problems during
incineration, and would require extensive off-gas
scrubbing to remove volatile chloride particulates.
d. Membrane life data in the field in other
applications has shown that ultrafiltration mem-
branes are less susceptible than reverse osmosis
membranes to deterioration in flux and rejection
due to alkaline hydrolysis and/or compaction under
pressure.
The major disadvantage of ultrafiltration vis-a-vis re-
verse osmosis lies in the quality of the aqueous efflu-
ent produced by the membrane system. Two factors are of
importance. First, since ultrafiltration does not
demineralize that waste treated, the effluent does con-
tain residual salts. As discussed below, this should
not present any problem in the treatment of decker efflu-
ents. However, for pine caustic extraction filtrate these
residual salts can limit the reuse potential of the
water. Second, color rejection in ultrafiltration.is
somewhat lower than that in reverse osmosis. The greater
residual color in the effluent from an ultrafiltration
system can be a limiting factor.
A review of the costs of commercially-available membrane
equipment at the beginning of the project showed that
spiral wound membranes would be the lowest in cost.
Although some process limitations exist for membranes
15
-------
in this configuration, it was felt that this system
would be adequate for processing both pine caustic
extraction filtrate and decker effluents. The advan-
tages and disadvantages of other membrane configura-
tions have been discussed elsewhere (^6). The two
alternatives, hollow fine fiber and tubular configura-
tions, are not thought to be attractive. The former is
extremely susceptible to plugging by suspended solids,
and the latter is a relatively high cost system. For
these reasons the present study has focused solely on
the use of membranes in the spiral wound configuration.
2- Need for Pilot Plant Studies
In a previous program sponsored by Champion Papers
(unpublished) the necessity of conducting in-field pilot
studies was apparent. In the preliminary program,
several hundred gallon samples were shipped to pilot
plant facilities for ultrafiltration tests. Data for
membrane flux, membrane rejection efficiency, and
material balances were obtained. However, factors which
could not be evaluated were:
a) differences in treating "fresh" (minutes-
old) and "aged" (days-old) feed materials;
b) pretreatment requirements for particulate
removal to obtain stable long-term operation;
c) operability of a staged membrane system
as a means of continuously producing a high-solids
concentrate; that is, with feed "conversion" to
permeate of 99+%;
d) effective techniques for membrane cleaning
as a means to sustain high-flux in long-term
operation;
e) membrane flux and life in intermediate-term
tests (six months). Note that long-term life tests
were not a part of this program;
f) requirements for pH and temperature control,
and performance at different pH, temperature, pres-
sure and feed flow conditions;
g) examination of a statistically meaningful
number of membrane modules to determine failure
modes; and
h) system operability in the field.
Furthermore, feed pretreatment (filtration) data could
only be obtained with fresh samples since particulate
16
-------
level and size were known to change with aging for
pulp mill effluents.
Finally, variability in flow and composition of the
wastes of interest, extremely important parameters in
process design, had to be determined at the mill.
For these reasons, a pilot plant program was clearly
warranted to realistically determine process technical
and economic feasibilities.
3. Pine Caustic Extraction Filtrate; Flow Schematic
and Typical Material jjalance
Figure 2 shows a flow schematic for the treatment of
pine caustic extraction filtrate by ultrafiltration. In
the bleaching of pine pulp, the pulp flows through a
series of stages. First is chlorination, with the
aqueous waste discharged to drain. In the second stage,
the pulp is extracted with caustic, and it is the efflu-
ent from this operation which contains the bulk of the
color discharged from the bleaching system. The pulp is
subsequently treated with calcium hypochlorite, and in.
other bleaching and washing stages.
One means of reducing the color discharge from a bleach
plant is to reverse the hypochlorite and caustic extrac-
tion stages. Through this sequence a substantial frac-
tion of the color is oxidized by hypochlorite, but at a
cost of increased chemical usage.
In the proposed process, the traditional CEH ... bleach-
ing sequence is used and the pine caustic extraction
filtrate is processed by an ultrafiltration system
(Figure 2). As described above, a membrane is selected
which specifically concentrates the dissolved organics
and color bodies. A purified, but not demineralized,
water is obtained for reuse or disposal. Reuse would be
limited to non-pulping operations because of the high
chloride content. The organic concentrate is disposed
of either by evaporation to approximately 50% solids,
with subsequent ultimate disposal on land, or by inciner-
ation. A typical material balance for the ultrafiltra-
tion operation is shown in Figure 3.
By obtaining 90-96% color removal from the pine caustic
extraction filtrate, approximately 50-60% of the total
17
-------
Chlorine
Water
Caustic
Calcium
Hypochlorite
Pine
Pulp
Chlorination
Stage
Drain
Caustic
Extraction
Stage
Other
Bleaching
Stages
00
Calcium
Hypochlorite
Stage
I
Drain
Organic Concentrate
Disposal by Evaporation
or Incineration
Ultrafiltration
Sy a tern
Purified Water for
Reuse or Disposal
FIGURE 2: CAUSTIC EXTRACTION FILTRATE—FLOW SCHEMATIC
-------
CONCENTRATE
1.67 Ibs
1,300,000 ppm color
14.5% total solids
4,020 ppm total chloride
1,590 ppm ionic chloride
FEED -
100 Ibs
VO
36,700 ppm color (unneutralized)
23,600 ppm color (neutralized)
0.91% total solids (neutralized)
785 ppm total chloride
705 ppm ionic chloride
MEMBRANE
PERMEATE
98.33 Ibs
1,920 ppm color
0.68% total solids
730 ppm total chloride
690 ppm ionic chloride
FIGURE 3: TYPICAL MATERIAL BALANCE FOR TREATMENT OF
CAUSTIC EXTRACTION FILTRATE BY ULTRAFILTRATION
(Neutralization with H2SO.)
-------
mill color will be removed. In addition, about 2 x 10
gpd of water will be available for potential reuse
within the mill. If the water cannot be reused and is
instead sewered to the waste treatment system, the
organic loading (as B.O.D.) of the waste treatment
system will be reduced by approximately 207o.
4. Decker Effluents: Flow Schematic and Typical
Material Balance
The treatment of decker effluents is the other case of
interest. While bleach plant effluents are found only
in bleached kraft mills, decker effluent is found in
all kraft mills. Thus, the treatment of decker efflu-
ent by ultrafiltration is an application of broader
interest to the industry. Referring to Figure 4, wood
chips are digested and the resulting pulp is transferred
to a blow tank. The black liquor removed from the blow
tank is processed in evaporators for chemical recovery.
The pulp from the blow tank goes to a series of washers.
At present, treated water is used in the washers, with
the spent wash water processed in the black liquor
recovery system. The pulp is then washed in a final
decker operation, again with the addition of treated
water. The decker effluent is too dilute to be processed
with the black liquor, and is discharged to the waste
treatment plant. At the North Carolina mill, approxi-
mately 2 x 106 gpd of treated water is added to the
decker.
In the proposed process, the decker effluent is concen-
trated in an ultrafiltration system. The purified water
is recycled to the final washer, or other pulp system uses,
eliminating the need for adding treated water. This
purified decker effluent could also be used at other
points in the mill. The organic concentrate from the
ultrafiltration system is processed in the black liquor
recovery system. A typical material balance for the
ultrafiltration operation is shown in Figure 5. Note
that the color removal efficiency is better than for
the pine caustic extraction filtrate case. This is
because the color bodies are lower molecular weight for
the latter, since substantial lignin fragmentation occurs
in the bleach plant chlorination stage.
This treatment of decker effluent by ultrafiltration has
several desirable features. These are
20
-------
Wood
Chips
f
Digester
ro
H
To
Black
Liquor ""^
Evaporators
Chemicals
(other pulp
mill usage)
Blow
Tank
To
Black <-s£
Liquor *^
Evaporators
Treated
Water
Pulp
Water
1
Washers
Decker
(hardwood
and pine)
Organic
Concentrate
To Black
Liquor
Evaporators
Ultrafilte:
Purified
—Watrex
FIGURE 4: DECKER EFFLUENTS—FLOW SCHEMATIC
-------
NEUTRALIZED
FEED
100 Ibs
0.26% total solids
11,000 ppm color
MEMBRANE
CONCENTRATE
2 Ibs
6.8% total solids
467,000 ppm color
- PERMEATE
48 Ibs
0.13% total solids
500 ppm color
FIGURE 5: TYPICAL MATERIAL BALANCE FOR TREATMENT OF
HARDWOOD DECKER EFFLUENT BY ULTRAFILTRATION
(Neutralization with H-SO.)
-------
--the water usage, and subsequent treatment require-
ment in the waste treatment system, is reduced
by 2 x 106 gpd;
--approximately 20% of the color in the total mill
effluent is removed;
--salts ordinarily lost in the decker effluent are
recovered and recycled to the black liquor
evaporation system;
--organics ordinarily lost in the decker effluent
are returned to the black liquor system and have
a corresponding heating value; and
--the organic loading (as B.O.D.) to the existing
waste treatment system is reduced by approximately
20%.
As described in detail below, the technical feasibility
of treating both pine caustic extraction filtrate and
decker effluents has been demonstrated. Projected costs
for both cases are thought to be attractive. For the
treatment of pine caustic extraction filtrate, ultra-
filtration costs will be less than those for lime preci-
pitation, activated carbon adsorption, or utilization of
the alternative bleaching sequence. For the treatment
of decker effluents, a positive return on investment
might be realized.
23
-------
SECTION IV
PILOT PLANT DESCRIPTION
A. GENERAL
A generalized flow schematic is shown in Figure 6. A
more detailed flow schematic and process description is
contained in Appendix A. Referring to Figure 6, feed
material, either pine caustic extraction filtrate, pine
decker effluent, or hardwood decker effluent was piped
from the mill to a 500-gallon Fiberglas feed tank.
Temperature, pH, and level were controlled at this point.
Neutralized feed from the feed tank was pumped through
a filter(s) for removal of suspended solids, and then to
the ultrafiltration unit. Details of the various filtra-
tion sequences employed are given below.
In the ultrafiltration system color bodies and organics
were concentrated and removed as a low-volume concentrate,
while water and salts were removed as permeate. During
operation, feed, concentrate, and permeate flows were
measured, sampled, and analyzed, as well as other inter-
mediate process flows.
In the ultrafiltration part of the pilot plant, five
recirculated "stages-in-series" were used. This design
was selected so that high conversion with high color
removal efficiency could be achieved (see Appendix A
for details). Spiral wound membrane cartridges were
installed in seven housings ("shells"); three shells were
used in Stage 1; and one in each of Stages 2 through 5. A
single shell held either three 24-inch long cartridges
(T.J. Engineering, Downey, California; or Eastman
Chemical Products, Kingsport, Tennessee) or two 36-inch
long cartridges (Gulf Environmental Systems Co., San
Diego, California). Cartridges from all three companies
were used in the pilot plant operation. A detailed
listing identifying cartridge use by stage and time is
given in Appendix B.
2
Membrane areas were approximately 300 ft for Stage 1,
and 100 ft2 for each of Stages 2 through 5. Total mem-
brane area was about 700 ft^. At a nominal membrane flux
of 15 gal./day-ft (gfd), the pilot plant capacity was
10,500 gpd.
25
-------
Recirculation Flow
(3 si ells)
Booster
Pump
Stage 1 Filter
Pump
Stage 1
permeate
Sampling points:
1) raw feed
2) feed and concentrate for each
stage (10 points)
3) permeates from each shell
(7 points).
4) mixed permeate
Recirculation Flow
Stage 2
Pump
(1 sh
all each)
Stage 2,3/4
permeate
Three stages connected
in series.
Concentrate
Flow
Stage 5
Pump
'Stage 5
permeate
Mixed
Permeate
FIGURE 6. SIMPLIFIED PILOT PLANT FLOW SCHEMATIC
-------
After pilot plant start-up in mid-August, 1972, it became
apparent that the pre-filtration system provided with
the pilot plant was inadequate for the removal of sus-
pended solids. The specific manifestations of inadequate
filtration were a rapid increase in pressure drop across
the membrane cartridges, and a drastically reduced ultra-
filtration rate with time. In addition, it was suspected
that membrane fouling by colloidal and dissolved macro-
molecular materials was also a problem, although definitely
of lesser importance than the flow-channel plugging created
by particulates. Several changes were made to the pilot
plant, including:
1. the installation of a single test cartridge
which was used for tests to determine means of
controlling the plugging and fouling problems;
2. the installation of more effective filters;
3. piping modifications to allow high-capacity,
once-through water flushing of all stages; and
4. the installation of piping and tanks to allow
detergent cleaning of membrane cartridges.
The sequence of changes can be summarized as follows:
Mid-August startup period; The originally installed
Broughton lOy basket filters were used.
Early September: A single cartridge was installed
for special tests to examine the plugging/fouling
problem. Also installed were a Bauer Hydrasieve
filter for the removal of coarse fibers, and ly
Cuno cartridge filters for more complete removal
of suspended solids.
Mid-September; A Hydromation depth filter was put
into use upstream of the Cuno polishing filters to
prolong the polishing filter element life.
Mid October: A detergent cleaning system was inte-
grated into the pilot plant. The Hydromation filter
was replaced with a Shriver filter press to obtain
increased capacity (prolonged filtration cycles).
End of November; Flush valves and the necessary
piping for once-through flushing of all membrane
shells was installed.
December: An automatic filter aid feeder was in-
stalled to permit unattended operation of the pre-
27
-------
coat filter over 24-hr periods.
Early February: The Shriver filter press was replaced
by a Sparkler leaf filter to obtain more "representa-
tive" filtration data. A Sparkler Velmac disc filter
was substituted for the Cuno polishing filters to
examine a "backwashable" polishing filter.
Changes in the flow diagram that resulted from the equip-
ment modifications are shown in detail in Appendix A
(Figure A. 2).
B. PRETREATMENT SEQUENCES
A more detailed flow schematic for the pretreatment opera-
tion is shown in Figure 7. Raw feed was piped from the
mill to the 500-gallon feed tank. During part of the test
period a screen (Bauer Hydrasieve) was used to remove
coarse fibers and particulates from the raw feed prior
to introduction into the feed tank.
As indicated, acid from a 55-gallon drum was pumped by
an acid pump into the feed tank for neutralization. For
the bulk of the program, sulfuric acid was used, but for
a two-week period, hydrochloric acid was added.
When pre-coat filtration was employed, a continuous filter
aid addition unit (BIF Industries screw-type adder) intro-
duced filter aid directly into the feed tank, where
vigorous mixing was achieved with a Lightnin mixer.
Neutralized/ cooled feed was pumped by a filter pump
through one of three filters, a Hydromation depth filter,
a Shriver filter press/ or a Sparkler leaf filter. The
filtrate was held in a 100-gallon surge tank (two 55-
gallon drums). In general/ flow into this surge was
through a float valve which kept the surge tank full.
Thus, the filtrate flow was controlled by the subsequent
rate of consumption in the pilot plant. At time,s ex-
cess filtrate was drawn off and either sewered or re-
turned to the 500 gal. feed tank. When a precoat fil-
ter was being used, a precoat suspension was mixed in
a 55 gal. drum connected to the suction of the filter
pump. Filtrate was returned to the precoat mixing
tank until an adequate precoat was developed.
Filtrate from the filtrate surge was pumped into the
suction of the Goulds multistage Stage 1 pump. The
28
-------
Raw
Feed
Feed Sampling Point
excess filtrate
Drain
Broughton
lOy
Basket
Filter
55 Gal.
Stage 1
Membranes
Polishing
Filters
Pretreated
Feed Sample
FI6UBE 7. FEED PRETKEATMENT FLOW SCHEMATIC
-------
outlet from the pump passed through two Broughton nominal
lOu basket filters. Flow was then through polishing fil-
ters (Cuno cartridges or Sparkler Velmac pot filter) to
remove residual particulates. To avoid the buildup of
an excessive pressure drop across the polishing filters,
a differential pressure switch was installed. Excess
pressure drop resulted in system shutdown. The final
polished filtrate was introduced into the three parallel
membrane shells of Stage 1.
A summary of the location in the process flow and the
characteristics of the filters is given in Table 2. Manu-
facturer's information on the filters is contained in
Appendix C.
C. OTHER SYSTEM MODIFICATIONS
1. Single Cartridge Tests
The single spiral wound cartridge was used for tests to
characterize prefiltration efficiency. Raw feed was
filtered by various means and processed in the single
cartridge. The buildup in pressure drop across the
cartridge and decrease in ultrafiltration rate were con-
sidered to reflect prefilter performance.
The single cartridge was connected in parallel to the
Stage 1 shells, and fed by the Stage 1 pump (Appendix A,
Figure 199). With this arrangement it was possible to
operate the single cartridge simultaneously, or inde-
pendently of, the pilot plant. The permeate and con-
centrate from the single cartridge were either sewered
or returned to the 100 gal. filtrate surge tank.
2. Detergent Cleaning System
A 55-gal. barrel served as a mixing tank for cleaning
solutions. The tank outlet was connected to the suction
of the booster pump (Figure 7). When cleaning, the
concentrate(s) and permeate(s) from the pilot plant
were returned by hoses to the detergent solution tank,
allowing the indefinite recirculation of detergent solu-
tion through the system. This eliminated the need to
prepare large volumes of cleaning solutions.
3. Once-through Flushing
The original piping of the pilot plant did not permit
30
-------
TABLE'2
FILTERS USED (IN ORDER OF FLOW SEQUENCE)
Filter
Location
Description
Function
1. Bauer Hydrasieve
in raw feed line
to 500 gal. feed tank
Curved wire screen/ 6 in.
wide, 2 ft long/ screen
openings approx. 0.01 in.
(see Fig. 8 )•
to remove coarse
fibers and solids
2. Hydromation
Depth Filter
after filter pump/
before filtrate
surge tanks
Granular PVC depth filter,
with automatic backwashing.
1 ft2 cross-sectional area.
fine filtration to
remove particles of
about 1v and larger
3. Shriver
Filter
Press
24 plate filter press;
16 in. x 16 in. plates;
total area of about 75 ft2.
Cloth used initially to hold
filter aid; subsequently
switched to paper sheets
(see Fig. 14) .
4. Sparkler
Leaf
Filter
Vertical leaf filter with
15.3 ft area. Leaves with
316 ss wire mesh, covered
with nylon bags, in carbon
steel vessel. Contained
high-pressure spray nozzles
for cake removal (see Fig. 13).
-------
Filter
TABLE 2 (continued)
FILTERS USED (IN ORDER OF FLOW SEQUENCE)
Location Description
Function
5. Broughton
Basket
Filters
after Stage 1 pump/
before Stage 1
membranes
Two model 3000 basket fil-
ters with nominal lOy
stainless steel mesh
screens. Total area
5.6 ft . Included an
automated differential-
pressure-controlled back-
washing sequence. A third
basket filter (200 mesh
screen, ) was used to filter
water for backwashing (see Fig. 16).
to remove suspended
solids of 10 y and
larger -(this was
the initial filter
provided with the
pilot plant)
6. Cuno
K> Cartridge
Filters
after Broughton
filter, before
Stage 1 membranes
Two parallel filter
housings, each containing
two cartridges. Disposible
1-5v cartridges in stain-
less steel housings. Area
per filter of about 1.6 ft2
for Micro-klean II elements
and 1 ft2 for Micro-Wynd II
elements. Parallel filters
permitted element change
without system shutdown.
final polishing
filtration before
membrane elements
7. Sparkler Velmac
Disc Filter
in place of
Cuno filters
Contained backwashable
polypropylene circular
discs* 5 v retention,
15 ft2 area.
-------
once-through flushing of the membrane shells, i.e. the
piping connected each stage to the next. It was decided
that cleanup of plugged membrane cartridges could be
facilitated by once-through flushing with water. Conse-
quently, a water line was connected to the suction of the
pump for each stage/ and flush valves were piped into the
concentrate lines from each stage. This arrangement per-
mitted once-through flushing of each shell in a "forward-
flow" direction. A subsequent piping modification allowed
once-through "reverse-flow" flushing.
Photographs of various system components are contained
in Figures 8-21.
D. DESCRIPTION OF MEMBRANE CARTRIDGES
The spiral wound cartridges consist of a membrane envelope
and mesh spacer that have been rolled around a PVC tube
(see Figure 22). One or more cartridges in series can
be installed in a housing, which serves as a pressure
vessel. The process fluid passes through the open space
provided by the mesh spacer. The permeate produced spirals
through the porous backing inside the membrane envelope
into the PVC center tube. A seal between the OD of the
cartridge and the ID of the shell prevents the process
liquid from bypassing the cartridge.
For the pilot plant program, 7-ft long, 4-in diameter
epoxy-coated steel shells were used. For the special
tests with a single cartridge, a 30-in long, 4-in diameter
PVC shell was employed.
Four different types of cartridges were purchased from
three different suppliers. All membranes were cellulose
acetate. More specific information on the cartridges
can be found in Table 3.
The Gulf cartridges had the highest salt rejection and
lowest flux. These were therefore used primarily in
Stages 4 and 5, where the greatest loss of color into
the permeate could occur and where the lowest flux was
expected due to high feed concentrations. One set of
three cartridges with corrugated flow channel spacers
was purchased for the purpose of comparing this type of
spacer with the standard mesh spacer. The former is
less susceptible to plugging by suspended solids.
33
-------
Fig. 8 - Bauer Hydrasleve
Bauer
Hydrasieve
-------
Feed
Line
Solids
Adder
Cooling
Coil
Connection
Mixer
Acid
Line
Fig. 9 -
Left:
of 500 Gal
Feed Tank
mixer and body feeder
Center: acid and raw feed lines
Right: connections to pli probe
and heating/cooling coil
-------
Fig. 10 - Acid Pump and Acid Drum
-------
Detergent Mixing Tank (fore
ground) and Precoat Mixing
Tank with Mixer (rear).
Precoat
Mixing
Tank
Detergent
Solution
Tank
-------
Temperature
Controller
Sparkler
Filter
12 -Pre treatment Sequence (in
art) .
From left to right: Sparkler
filter; filter pump; bottom of
500-gal. feed tank; temperature
controller and indicator; booste
-------
Flush
Connections
Fig. 13 - Details of Sparkler Filter
Pressure gauges, filtrate flow
meter, flush connections.
-------
Fig. 14 - £i Press
-------
Level
k.Switch
Float
Valve
^^•^•^•^••^
Fig- 15 - 100 Gal. Filtrate Surge Tanl
Two 55-gal. drums connected at the
bottom to provide 100-gal. surge
volume; level switch (left) and
level control valve (right; volume
ahead of the ultrafiltration unit.
-------
Broughton
Basket
Filters
Single
Cartridge
- Single Cartridge and Filters
Single cartridge in PVC shell
deft); Broughton basket filter
(center); Cuno filter (right).
-------
7 - Sparkler Velmac Disc Filter
Sparkler
Velmac
-------
Fig. 18 - Ultrafiltration Unit:
Center tier
Left: control panel
Top tier: circulation pumps,
Stages 2 through 5.
shells with membrane
cartridges, Stages 2
through 5
Bottom tier: Stage 1 pump and
membrane cartridges
(flush valves in-
stalled)
irculation
Pumps
Stages 2-5
Membrane
Shells
Stages 2-5
Membrane
Shells
Stage 1
-------
. 19 -.UltraflUr'ablon Unit: Crop
view")
Stages 2 through 5 pumps and shells
with membrane cartridges in fore-
ground
-------
Main switch, pump switches, alarms
outlet pressure gauges and control
valves, Stage 1-5
-------
ontroller
Recirculation
Flow Meters
pH, Temp.
Recorders
Inlet
Pressure
Gauges
Sampling
Valves
Permeate
Flow
Meters
concentrate
Flowmeter
Mixea
Permeate
Flowmeter
Control Panel
47
Left Side
Stages 1-5 recirculation flow
meters, inlet pressure gauges
(Stages 1-5), and control valves
(Stages 2-5); permeate flow meter,
one for each of the shells (7);
pH controller with indicator; pH
and temperature recorders.
-------
Brine Seal
ROLL TO
ASSEMBLE
FEED SIDE
SPACER
PERMEATE OUT
PERMEATE SIDE BACKING
MATERIAL WITH MEMBRANE ON
EACH SIDE AND GLUED AROUND
EDGES AND TO CENTER TUBE
FEED FLOW
PERMEATE FLOW
(AFTER PASSAGE
THROUGH MEMBRANE)
SEE DETAIL A
MESH SPACER
MEMBRANE
FIGURE 22. SPIRAL
WOUND MODULE DESIGN
BACKING MATERIAL
PERMEATE TUBE
GLUE LINE
DETAIL A
-------
TABLE 3
CHARACTERISTICS OF MEMBRANE CARTRIDGES
SUPPLIER
model no.
membrane type
spacer material
length of
cartridge,
inches
number per shell
membrane area
per cartridge,
nominal salt
rejection, %
nominal flux,
gfd at
100 psig, 70° F.
number purchased
number of brine seals
brine sealing point
during normal
operation
Gulf Environ-
mental Systems
4000
UF
mesh
36
2
50
approx .
30"
7-10
4
1
upstream
TJ
Engineering
UF-H-32
Eastman
HT 00
mesh
24
3
32-35
0-10.
35-50
45
2
downstream
TJ
Engineering
UF
Eastman
HT 00
corrugated
24
3
12-15
0-10
35-50
3
2
downstream
Eastman
Chemical Products
MK 00 A
HT 00
mesh
24
3
32-35
0-10
35-50
3
1
upstream
-------
E. SAMPLING AND ANALYTICAL PROCEDURES
Referring to Figures 6 and 7, the following samples were
collected:
—raw (untreated) feed;
—neutralized and filtered feed (feed to Stage 1);
—final concentrate (Stage 5 concentrate);
—mixed permeate (from all 5 stages);
—feed samples for all stages (5); and
—permeate samples for all stages (5).
Analytical procedures are detailed in Table 4. The analyses
for pH, total solids, suspended solids and color of the
raw and pretreated feed, final concentrate and mixed per-
meate were performed on a regular basis. Occasionally,
analyses of feed and permeate samples from individual
membrane stages were also carried out. A few samples
were analyzed for other constituents (ash, various ions,
and suspended solids composition and particle size dis-
tribution) to better characterize performance of the sys-
tem.
Suspended solids analyses were performed by the standard
gravimetric method. However, it was found that only
relatively small samples, 50-100 ml, could be filtered
through the 0.45y Millipore membrane filters before
blinding occurred. Thus, the suspended solids analyses
were only approximate, and poorly reproducible, since
the total amounts of suspended solids present in the
samples were small, and difficult to measure accurately.
A turbidimeter and a nepholometer were also tried to
assay for suspended solids, but both instruments were
found to lack sensitivity because of strong light ab-
sorption by the highly colored samples.
50
-------
Analysis For
PH
Color, Cobalt units
TABLE 4
ANALYTICAL PROCEDURES
Procedure *
Standard pH meter
Absorbance at 465 mja
compared to absorbance of
standard Pt/Co. solution,
No. 118 at pH 7.6, "Field"
color measured at as-is pH.
Gravimetric, No. 148A
Gravimetric, No. 148B
Gravimetric, No. 148C
Gravimetric, No. 148D
Calibrated chloride ion
electrode, No. 203C
Flame Photometric Method,
No. 153A
EDTA Titrimetric, No. 67C
Gravimetric, No. 238A
Colorimetric, No. 144B
Particle Size Distribution Sizing on filters fllowed
of Suspended Solids by analysis of fraction
passing through 44u filter
with Coulter counter.
Total Solids
Total Dissolved Solids
Suspended Solids
Ash
Ionic Chloride
Na+
Ca-H-
So =
Fe
^Referenced number is the method in Standard Methods
for Examination of Water and Wastewater, APHA, 13th
Edition, 1971.
51
-------
SECTION V
RESULTS AND DISCUSSION
A. FEED PRETREATMENT AND CHARACTERISTICS
The pilot plant was operated on pretreated pine caustic
extraction filtrate for most of the test program, and
on hardwood and pine decker effluents for a limited
period during January , 1973.
Before introduction into the ultrafiltration section of
the pilot plant all three effluents were neutralized
to a pH of 6 to 7 , cooled to a temperature of 95 to
105 °F, and filtered to remove suspended solids. A
description of the pretreatment section of the pilot
plant was given in Section III.B. Results pertinent
to the evaluation of the pretreatment operations are
discussed below.
1. Temperature
Cellulose acetate ultrafiltration membranes show
accelerated hydrolysis and compaction {irreversible
loss of flux due to collapse of the porous membrane
structure) with increasing temperature/ and especially
when exposed to temperatures in excess of 110 °F. All
three effluents treated in the pilot plant are generally
discharged at a temperature of 120-135 °F/ depending on
mill operating conditions and the season. Therefore,
it was necessary to cool the feed, which was done by
passing cold water through the coil heat exchanger in
the feed tank. In the course of the experimental pro-
gram, the operating temperature was not intentionally
varied, and most data are for operation at a feed tempera-
ture between 95 and 105 °F.
2.
Cellulose acetate membranes show satisfactory life when
operated within a pH range of 3 to 7. All three efflu-
ents, however, are highly alkaline, with pH's ranging
from pH 10 to pH 12. Therefore, the effluents were
neutralized to a pH of 6 to 7, by addition of sulfuric
acid. Figures 23 to 25 contain pH curves for the neutral-
ization of pine caustic extraction filtrate, pine decker
53
-------
12J
10
BASIS: 100 ml of Caustic Extraction Filtrate
•\
\
u.
H
2 .
mm
V
V
\
234
Amount of 0.5N I^SC^ added, ml
FIGURE 23: SULFURIC ACID REQUIREMENT TO NEUTRALIZE
CAUSTIC EXTRACTION FILTRATE
-------
11
BASIS: 200 ml of Pine Decker Effluent
PH
10
3/22/73
18
Amount of 0.1N H2SO4 added/ ml
FIGURE 24. SULFURIC ACID REQUIREMENT TO NEUTRALIZE PINE DECKER EFFLUENT
-------
11.
BASIS: 200 ml of Hardwood Decker Effluent
10
PH
-rir
-rt-
-te
Amount of 0.1N H-SO, added, ral
FIGURE 25. SULFURIC ACID REQUIREMENT TO NEUTRALIZE HARDWOOD DECKER EFFLUENT
-------
effluent, and hardwood decker effluent respectively. About
4 Ibs of 100% sulfuric acid are required to neutralize
1000 gals, of any one of the three effluents.
The major effects on feed characteristics of pH adjust-
ment were:
—a reduction in feed color;
—an increase in the total solids of the feed by
about 400-500 ppm, as a result of the addition
of sulfate ion (from H2S04); and
—some change in the colloidal nature of the feed
as evidenced by a change in suspended solids
content.
Figure 26 shows the effect of pH adjustment on color for
a sample of pine caustic extraction filtrate. The color,
measured by sample absorbance at 625 my, gradually dropped
with decreasing pH until at pH 2 the sample became turbid
due to heavy flocculation of organics.
It was observed during several experiments with pine
caustic extraction filtrate that the suspended solids
level increased with reduced pH. Typical data are given
below.
SUSPENDED SOLIDS OF PINE CAUSTIC EXTRACTION
FILTRATE AS A FUNCTION OF pH
pH Adjusted With H^SO, Suspended Solids, ppm
8 43
7 39
6 58
5 113
2 turbid (not measured)
This behavior of pine caustic extraction filtrate is fur-
ther discussed in Section V.E : Important Factors
Controlling Membrane Flux.
3. Prefiltration
In a previous laboratory test program, the necessity of
feed pretreatment for particulate removal to obtain
satisfactory ultrafiltration rates was apparent. The
57
-------
Ul
00
p.
6
in
CM
vo
-P
(0
M
O
en
1.0
.8
.6
.4
0
TT
pH
FIGURE 26. EFFECT OF pH ON COLOR OF PINE CAUSTIC EXTRACTION FILTRATE
-------
exact type of filtration and filter equipment could not
be determined under laboratory conditions with "aged"
feeds. Hence, a substantial amount of time was devoted
during this pilot plant program to investigate filters
and their operating conditions in order to obtain satis-
factory removal of suspended solids. The following
filters and filter combinations were examined in the pilot
plant program:
8/14/72-9/27/72
9/27/72-10/5/72
10/17/72-2/1/73
Broughton nominal 10 y mesh filter;
1) Bauer Hydrasieve (removal of
fibers and gross suspended solids;
2) Hydromation depth filter (l-5y
solids removal); and
3) Cuno polishing filters (ly
cartridges;
1) Shriver filter press with wood
flour as precoat material and
filter aid (body feed, between
150 and 300 ppm); and
2) Cuno polishing filters (ly
cartridges;
1) Sparkler pressure leaf filter
with wood flour as the precoat
material and filter aid (body
feed, between 150 and 300 ppm); and
2) Sparkler Velmac 5y disc filter
Four different main filters were used: the Broughton mesh
filter, the Hydromation depth filter, the Shriver filter
press, and the Sparkler leaf filter. The installation
of each was carried out to improve pilot plant perform-
ance. The Hydromation depth filter used a packed bed of
granular PVC as the filtration medium. The Shriver and
Sparkler filters were precoat filters and operated with
filter aid added to body the feed. For most of the experi-
mental program wood flour was used as the body feed; how-
ever, some tests were performed with other filter aids,
and the results are discussed below.
2/2/73-3/1/73
59
-------
a. Operational Experience
Immediately after pilot plant startup, it became apparent
that the Broughton filter was not providing adequate sus-
pended solids removal. Specifically, the membrane car-
tridges were plugging rapidly with suspended solids, as
evidenced by a rapid increase in pressure drop across
the cartridges and a rapid decrease in ultrafiltration
rate. At that time a sample of the suspended solids in
the pine caustic extraction filtrate was analyzed for
particle size distribution. The results of this analysis
are contained in Table 5. It is apparent that about 50%
of the suspended solids are below lOy, and is not sur-
prising that the nominal 10y Broughton filters proved in-
effective for particulate removal. No actual data were
obtained on the removal efficiency of suspended solids
by the Broughton filter, but it was probably 50% or less.
Samples of the pine caustic extraction filtrate were
submitted to filter vendors to obtain recommendations for
prefiltration. Vendor tests showed that surface filtra-
tion resulted in rapid filter blinding, and that depth
filtration was required. As an initial approach, a
Hydromation depth filter was installed. Excellent re-
moval of suspended solids was obtained, as determined
both by the reduction in suspended solids level and more
stable ultrafiltration rates in the pilot plant (data
discussed in Section V.C.). Unfortunately, the filter
size was inadequate for operation of the entire pilot
plant on a continuous basis. The vendor was unable to
deliver a larger filtration unit within the schedule and
budget constraints of the program, and it was decided to
switch to a precoat filter.
2
An existing Shriver filter press (75 ft area) was avail-
able within the mill, and was installed in the pilot
plant. Based on a recommendation by the Sparkler Manu-
facturing Co. , Wilner wood flour # 139 was chosen as the
precoat and body feed material. Substantially improved
performance of the pilot plant was obtained with the
Shriver filter press compared to the Broughton basket
filters. Pilot plant operation continued for about 3-1/2
months using the Shriver filter. However, it was real-
ized in December that the filter was operating at~an
unrealistically low filtration rate (^ 0.1 gpm/ft ) and
it was decided to install another filter which could
treat the feed at more typical filtration rates
60
-------
TABLE 5
PARTICLE SIZE DISTRIBUTION OF SUSPENDED SOLIDS
IN PINE CAUSTIC EXTRACTION FILTRATE (UNNEUTRALIZED)
Particle Size (microns) Weight/ % Cumulative Weight, %
>149
88-149
44-88
40-44
32-40
20-32
10-20
5-10
2.5-5
<2.5
9.9
2.8
3.9
1.03
1.88
13.76
17.68
19.84
20.94
8.26
99.99
90.09
87.29
83.39
82.36
80.48
66.72
49.04
29.2
8.26
61
-------
2
(^ 1 gpm/ft ). Furthermore, it was suspected that the
Shriver press did not provide the quality of filtration
normally obtained with a precoat filter. This conclusion
was based on often erratic performance of the ultrafiltra-
tion unit. The following considerations may explain the
less than adequate performance of the Shriver filter:
—it was an old filter and possibly contained
defects;
—the flow distribution between the different plates
and frames (24 plates) was poor and the precoat
thickness was probably not uniform. Furthermore
it was not possible to form the precoat under high
flow conditions which would have produced a more
uniform precoat;
—the filter press was operated at high pressure
drop, between 30 to 60 psi, and this could have
resulted in the "extrusion" of solids into the
filtrate;
—operational problems may have allowed the precoat
to fall off when switching valves to change from
precoat application to treating neutralized feed.
For these reasons a Sparkler leaf filter was ordered
and installed at the beginning of February, 1973. The
Sparkler filter provided excellent suspended solids re-
moval and filtration rates. It was possible to main-
tain filtration rates of approximately 1 gpm/ft2 for
periods of up to 24 hrs, depending on the feed type and
characteristics. (Note, filtration characteristics of
the caustic extraction filtrate and decker effluents
differed, as discussed below).
Table 6 contains a summary of the performance data of
the main filters in processing pine caustic extraction
filtrate. The data for removal of suspended solids are
based on a suspended solids analysis by filtration on
0.45y Millipore filters. As described before, this
method provided only approximate values for suspended
solids levels. However, it is possible to recognize
major trends in filter performance. Based on the suspend-
ed solids removal data in Table 6, it was judged that
the Sparkler and Hydromation filters gave the best per-
formance; and performance of the Shriver filter press was
good at times, but erratic. The filtration rates of
both the Sparkler and Hydromation filters are judged to
be in the range suitable for commercial application.
62
-------
TABLE 6
PERFORMANCE - MAIN FILTERS
Filter
Area Filter
Filter sq ft Medium
Max/Min
Flows
During
Operation/
gpm
Max/Min
Filtration
Rates During
Operation,
gpm/sq ft
Min/Max
Duration
Filtration
Cycle, hrs
Filtration
Efficiency as
indicated by
Subsequent Ultra-
filtration Rates
Removal of
Suspended
Solids
% Removed
Broughton 5.6 10u 8-3
mesh
1.5-0.5
inadequate
Hydromation 1 granular PVC 8-3
8-3
0.1-1.0
acceptable
50-80
Shriver 75 wood flour 10-3
u>
0.13-0.04
4-20
sometimes
acceptable
20-80
Sparkler 15.3 wood flour 25-5
1.6-0.3
4-16
acceptable
50-80
-------
Perhaps the most relevant information on suspended
solids removal by the different filters is obtain-
able from the ultrafiltration rates of the pilot plant.
In general, the ultrafiltration rate data show that the
Sparkler pressure leaf filter and the Hydromation depth
filter were equally effective in removal of suspended
solids. At times the Shriver filter press performed
adequately.
Performance of Other Filters
The performance characteristics of the Bauer Hydrasieve,
the Cuno cartridges, and the Sparkler Velmac disc
filter were not closely monitored because of their rela-
tively limited influence on the performance of the
ultrafiltration unit.
The Bauer Hydrasieve adequately removed fibers and gross
suspended solids from the raw feed, which prevented
clogging of valves and pumps used before the main fil-
ter. However, the performance of the Bauer Hydrasieve
was not a critical part of the filtration operation from
the point of view of ultrafiltration performance.
The Cuno cartridge filters were installed just before
the ultrafiltration unit for final polishing of the
feed. These filters also served to protect the ultra-
filtration unit in case of gross failure of the main
filter. The Cuno filters were primarily used in combina-
tion with the Shriver filter press. Some additional
suspended solids removal was achieved by the Cuno fil-
ters and it was necessary to change the disposable
cartridge elements every 5-10 hrs of operation (1000-
3000 gals.). It cannot be ascertained whether the
solids removed by the Cuno filters are particulates
which would pass through any precoat filter, or are
suspended solids which passed through the Shriver filter
due to below-par performance.
In the final phase of program, the Cuno filters were re-
placed by a Sparkler Velmac disc filter because of the
larger filter area and the capability of cleaning the
filter discs. In combination with the Sparkler leaf fil-
ter, the Velmac filter lasted for about 60 hrs (20,000
gal.) when it was used for the first time. Most of the
solids removed by the Velmac filter could be removed by
64
-------
washing. However, some reduction in filter capacity
was observed in subsequent runs due to an inability
to completely regenerate the filter medium. The
ultimate useful life of the Velmac filter has not yet
been established, but it is anticipated that in opera-
tion of a full-scale plant, life will be economically
acceptable.
Tests with Different Filter Aids
Two sets of experiments were performed. One with the
Shriver filter press, and another with the Sparkler leaf
filter.
The data from the Shriver filter tests were obtained 2
under conditions of low filtration rates (^0.1 gpm/ft ),
and can only be used as a guide in evaluating perfor-
mance characteristics of different filter aids. In these
tests, five different filter aid materials were used:
# 139 wood flour, (Wilner Wood Products, Norway, Maine),
Dicalite 436 (Grefco diatomaceous earth), Celite 545
(Johns-Manville diatomaceous earth), Hyflo Super Cel
(Johns-Manville), and Perlite 400F (Chemrock Corp.).
In the tests a precoat of about 6 to 10 Ibs was applied
to the filter, and the filter aid was added as body
feed at a level of approximately 150 ppm.. In these tests,
no observable difference in removal efficiency of
suspended solids was noted. However, a substantially
longer filtration cycle was obtained with the wood flour
than with the other filter aids. This suggests that
the filter cake formed with wood flour is less suscepti-
ble to plugging or blinding by the suspended solids-
filter aid suspension.
In the tests with the Sparkler filter, three different
types of filter aid were examined: wood flour, dia-
tomaceous earth (two grades), and a Perlite. The tests
were performed with two different effluents: pine
caustic extraction filtrate and hardwood decker efflu-
ent. In these tests, a 3-lb precoat of each of the fil-
ter aid materials was applied to the 15.3 ft2 Sparkler
filter. After the precoat was applied, wood flour was
ised as the bodying material, and added at a level of
Approximately 150 ppm.
65
-------
The effluent filtered was always fresh, and was with-
drawn directly from the 500 gal. feed tank. Operation
was at a standard filter pressure drop of 20 psi
(except for the one run with Hyflo Super Cel since this
is a looser type of filter aid and could compress at a
20 psi pressure differential.
A summary of the test results is presented in Table 7.
The data in Table 7 show that initial filtration rates
for all of the precoat materials were excellent, in the
range of 1.4 to 1.8 gpm/ft . In the tests with the pine
caustic extraction filtrate, wood flour and the Dicalites
maintained their original filtration rates over the full
21 minute test period. In the tests with the hardwood
decker effluent, which contained a much higher level of
suspended solids, an appreciable falloff in filtration
rate was observed, as expected.
The Hyflo Super Cel is a much coarser filter aid, with
a greater porosity and higher initial filtration rate.
However, the filtration rate dropped off sharply with
both pine caustic extraction filtrate and hardwood
decker effluent, indicating that a fairly dense filter
aid is to be preferred as a precoat medium if long fil-
tration cycles are to be achieved.
Removal of suspended solids in all tests was about the
same, approximately 50 to 75%.
On the basis of the tests with both the Shriver and the
Sparkler filters, it was concluded that the Wilner # 139
wood flour was the preferred filter aid. In addition,
wood flour has the advantages that it is inexpensive
and can be easily disposed of by incineration with the
concentrate from the ultrafiltration unit. For these
reasons, wood flour was used for the experimental
program.
4. Feed Characteristics
The detailed experimental data on feed characteristics
for caustic extraction filtrate, pine wood decker
effluent and hardwood decker effluent, respectively,
are presented in summarized form in Table 8.
66
-------
TABLE 7
TESTS WITH DIFFERENT FILTER AIDS
ON SPARKLER 15.3 sq ft LEAF FILTER
Filter Aid
Wilner |139
Dicalite 436
c>
"^ Dicalite 436
Hi call te 476
Hyflo Super
Cel
Wilner §139
Dicalite 476
Hyflo Super
Cel
Type of Filter Aid
wood flour
diatomaceous earth
"
H
Per lite
wood flour
diatomaceous earth
Perlite
Effluent
Pine caustic
extract
n
n
n
n
Hardwood deck-
er effluent
"
n
Pressure
Drop
psi
20
20
20
20
10
20
20
20
Precoat
Ibs
3
3
3
3
3
3
3
3
Filtration rate
gpm/ft2
initial
1.4
1.7
1.5
1.6
1.6
1.4
1.8
1.5
final
(21 min)
1.4
1.5
1.6
1.6
0.85
0.9
1.2
0.2
Suspended
Solids
in/out
ppm
132/47
20/10
75/18
44/20
50/19
300/160
260/180
280/100
-------
The range of compositions shown in Table 8 are based on
analyses of all samples collected throughout the pilot
plant program. The detailed data presented in Figures
27 to 33 show that the feed composition varied widely
over relatively short periods of time, often within
the same day. Sharp variations in feed composition
were noticed particularly when start up or shutdown
operations took place in the mill. This alone suggests
that improved effluent flow control could reduce the
mill washwater volume. Presumably operation at high
solids-high color loadings did not impair pulp or paper
quality. It should be possible to maintain these
loadings near the highest tolerable level, thus re-
ducing water consumption.
a Suspended Solids
Table 8 and Figures 27 to 29 give the ranges of suspended
solids that were determined for the three effluents,
both untreated and pretreated. As discussed in the
section on Analytical Methods, the suspended solids data
are only approximate, and this accounts, in part, for
the scatter in the data.
The suspended solids level of the hardwood decker efflu-
ents , during the limited time when measurements were
made, was found to be higher than those for the other
two effluents.
, Total Solids
D.
Figures 30 and 31 show the total dissolved solids contents
of pine caustic extraction filtrate, pine decker effluent,
and hardwood decker effluent, respectively. Pine caustic
extraction filtrate had a much greater amount of total
solids (avg. about 7000 ppm) compared to the decker ef-
fluents which averaged about 2500-3000 ppm solids.
Feed neutralization and prefiltration introduced about
400-500 ppm of solids into the three effluents, due to
the addition of sulfate ion.
68
-------
TABLE 8
SUMMARY OF FEED CHARACTERISTICS
VO
Pine Caustic
Extraction Filtrate
Pine Decker Effluent
Hardwood
Decker Effluent
1.
2.
3.
4.
pH
Untreated Feed
Treated Feed*
Color, ppm, Cobalt units
Untreated Feed
Treated Feed*
Total solids, ppm
Untreated Feed
Treated Feed*
Suspended Feed, ppm
Untreated Feed
Treated Feed*
Range Average
11.5-12
6.0-6.5
12,000- 28,000
45,000
10,000- 19,000
27,000
4,400- 7,000
11,400
** **
50-200 80
10-90 30
Range Average
10.7-10.9
6.0-6.5
4,000- 6,000
9,300
3,000- 4,000
7,000
1,700- 2,400
8,200
** **
50-150 80
20-40 (limited data)
Range
10.0-10
6.0-6.
8,000-
22,000
6,000
13,000
1,200-
4,600
**
100-350
50-180
Average
.6
5
11,000
8,000
3,000
* *
200
(limited dat
* pH of feed adjusted to 6.5-6.9, followed by filtration through a depth or precoat filter
** total solids in neutralized, filtered feed were increased over untreated feed by about
400-500 ppm due to sulfate addition (H-SO.)
-------
40C-
£• unneutralized feed (untreated)
• neutralized and filtered feed (treated)
30C-
a
*
to
•H
H
0 A
•o
^20(*
0)
(X
CO
3
cn
&
average for average for * *& m a* A A
untreated .treated • *
irtrj vuA
J
/:
".A
A
®
1
. O • BJI
• •• *
August September October November December January February March
1972 1973
FIGURE 27. SUSPENDED SOLIDS OF CAUSTIC EXTRACTION FILTRATE
-------
400
300
o
w
•S
to 200
CO
unneutrailzed feed
(untreated)
neutralized and
filtered feed
(treated)
100
average for un-
neutralized feed
average for
neutralized
feed
L.
fSL
November December January February March
1972 1973
FIGURE 28. SUSPENDED SOLIDS OF PINE DECKER EFFLUENT
-------
400
300 h-
o
CO
•o
I
8, 200
3
to
100
average for unneutralized
X
a
A
&
unneutralized feed
(untreated)
neutralized and
filtered feed
(treated)
£>
February
March
November December January
1972 1973
FIGURE 29. SUSPENDED SOLIDS OF HARDWOOD DECKER EFFLUENT
-------
untreated Pine Caustic Extraction Filtrate
10,000
• neutralized and filtered Pine Caustic
Extraction Filtrate
o untreated Pine Decker Effluent
5,8000
•»
to
•H
rH
o
01
i 6000
«J
4J
O
4000
average for unneutralized
Pine Caustic Extraction Filtrate
&
&
&
average for unneutralized
Pine Decker Effluent
2000
a
G> a
a
a
August
September October
November December January February March
1972 1973
FIGURE 30. TOTAL SOLIDS OF PINE CAUSTIC EXTRACTION FILTRATE AND PINE DECKER EFFLUENTS
-------
10,000
8000
p<
6000
(0
4J
o
4000
2000
a
-fir
unneutralized feed
a
A
average for un-
neutralized feed
November December January February March
1972 1973
FIGURE 31. TOTAL SOLIDS OF HARDWOOD DECKER EFFLUENT
-------
48 —
40
32
01
X 24
8.
cu
8 16
G Pine Caustic Extraction Filtrate (untreated)
o
A Pine Caustic Extraction Filtrate
(neutralized and filtered)
a Fine Decker Effluent (untreated)
Pine Decker Effluent
(neutralized and filtered)
average for untreated Pine
Caustic Extraction Filtrate
00
average for treated Pine
Caustic Extraction
o
§
o
average for untreated Pine
Decker Effluent
o
o o
00
o o o
o
GO
O
0
o
o
o
average for_treated_ Pijie Decker Effluent
_L
i
J.
J_L
I I I I I
1
s
'' ' ' ' ' ' ' ' i
© O
___.,
"' ...... © o o
© o ©
o
00
O O O©
o o°©
QJ
• •
< ' I ' ' ' '
November December January February March
1972 1973
August September October
FIGURE 32. COLOR OF PINE CAUSTIC EXTRACTION FILTRATE AND PINE DECKER EFFLUENT
-------
48
O unneutralized feed
A neutralized and filtered feed
40
32
o
H
X
Os
24
8
16
average for unneutralized feed
O
O
o
o
o
o
o
0
&
average for neutralized and
filtered feed
©*~
i i t
I i i i i i I i i i i i I i i i i i
I
August September October November December January
1972 1973
February March
FIGURE 33. COLOR OF HARDW6)OD DECKER EFFLUENT
-------
c. Color
Figures 32 and 33 show the color contents of the three
effluents. It is seen from these figures that the
pine caustic extraction filtrate was much more highly
colored (avg 28,000 ppm), than the other two effluents.
Pine decker effluent had the least color, averaging
about 6000 ppm.
Figure 34 shows the effect of feed pretreatment (pH
adjustment to 6.5 to 6.9 and precoat filtration) on the
"field" color content of a series of samples of pine
caustic extraction filtrate. Feed pretreatment removed
about 30-35% of the color from the raw feed; the major
part was due to pH adjustment.
d. Other Solutes
Table 9 gives additional analytical data for a few feed
samples of the three effluents.
In summary, the following conclusions can be drawn about
the characteristics of the three effluents:
—Pine caustic extraction filtrate has higher
levels of total dissolved solids, ash and color
than either of the decker effluents;
—Although the two decker effluents have similar
levels of total dissolved solids, pinewood decker
effluent has less color;
—Hardwood decker effluent has the highest suspended
solids level; pine caustic extraction filtrate
and pine decker effluent levels are about the same;
—Pine caustic extraction filtrate contains much more
total and ionic chloride than the decker effluents;
--Decker effluents contain a substantial amount of
sulfate ion, originating in the pulp digester;
—The feed compositions of all three effluents vary
widely within a relatively short period of time;
--Pretreated effluents have about 30-35% less color
than the untreated (raw) effluents;
--pH adjustment of the effluents introduces about
400-500 ppm sulfate ion;
~-A major fraction of the suspended solids of the
three effluents is smaller than 10^1, but is
reasonably effectively removed by either depth
77
-------
-0
c
(0
•o
0)
N
•H
id
M
9
oo 0)
4-t
0
^1
o
r-l
O
U
JU
CO
1
o
I—I
X
1
•o 20
^1
^s^
*s^
0 .^^^ O O
.t^^
^
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
FIGURE 34,
20 30 40
Color of Unneutrailzed Feed, ppm x 10
EFFECT OF FEED PRETREATMENT ON COLOR OF PINE CAUSTIC EXTRACTION FILTRATE
-------
Components
TABLE 9
WASTE CHARACTERISTICS AT THE NORTH CAROLINA MILL - DETAILED ANALYSES
Untreated Pine Caustic
Extraction Filtrate
Untreated Pinewood
Decker Effluent
sampled on 1-3-73
Untreated Hardwood
Decker Effluent
sampled on 2-1-73
PH
10-13-69
11.3
Color, ppm
S . S . , ppm
T.S. t ppm
Ash / ppm
ft a , ppm
Ca , ppm
Cl~, ppm
1
S04~~' PP™
Fe , ppm
1,180
78
1,010
10-16-70
11.5
14,000
6,800
3,800
2,000
5
1,600
1.6
2-10-73
11,7
28,000
80
7,000
1,200
60
1,275
3 pm
10.7
7,700
230
2,244
1,010
480
20
14
260
1.4
5 pm
10.7
7,300
70
2,284
784
425
16
14
190
1.2
8 pm
10.7
8,300
140
2,468
908
516
4
14
240
0.8
1 pm
10.0
18,300
400
3,848
976
58
16
30
350
2
3 pm
10.5
8,300
114
2,520
848
480
4
6
200
1.2
5 pm
10.7
7,700
214
2,244
1,464
441
12
35
200
0.1
-------
filtration or precoat filtration.
B. REJECTION DATA AND EVALUATION
Comprehensive data were collected on detailed color and
total solids rejections data for pine caustic extraction
filtrate, pine decker effluent and hardwood decker efflu-
ent, respectively. Specifically, the following informa-
tion was included:
a. Color of unneutralized feeds; neutralized and
filtered feeds, final concentrates and mixed permeates
from the ultrafiltration unit;
b. Total solids of unneutralized feeds, final
concentrates and mixed permeates from the ultrafiltra-
tion unit;
c. Concentration ratios, defined as the ratio of
feed flow to the ultrafiltration unit to concentrate
flow from the ultrafiltration unit;
d. Percent color removal, defined as
color of unneutralized feed-color of mixed permeate x 100;
color of unneutralized feed
and
e. Color removal efficiency of the ultrafiltration
unit by stage for treatment of pine caustic extraction
filtrate.
For presentation here these data have been condensed and
displayed in Figure 35, which summarizes the separation
efficiency of the pilot plant during the operating period.
For all three effluents, the concentration ratio achieved,
the total solids level in the concentrate, and the percent
color removal are shown.
1. Color Rejection
As seen in Figure 35 excellent color removal was obtained.
For example, for about eighty-fold concentration, color
removal was approximately 90% for the pine caustic extrac-
tion filtrate and 95% for both decker effluents. The
higher color removal efficiency for decker effluents is
hypothesized to be related to the molecular weights of
80
-------
20
I
Sio-
H C
S
t, £>
7"
10 &
ft sa
»a ^
CEF ,£
(-X.80 fold
concentration)
(-v-120 fold
concentration)
a a
1
100
90
I
L-
«
n
60
40
A
& &
A
&
A
CEF
(i>80 fold
concentr ation)
^ Pine Caustic Extraction
Filtrate (CEF)
-=-Pine Decker Effluent (DE)
a Hardwood Decker Effluent (DE)
DE (^80 fold
concentration)
& * £ —''3
'X «.
e> &
ae>
a a
& e> &
i
1
August September October November December January
FIGURE 35. REJECTION AHD CONVERSION DATA
February March
-------
Table 10
REJECTION
Effluent
Pine
Caustic
Extraction
Filtrate
Pine
Decker
Effluent
Hardwood
Decker
Effluent
Period
11/5-
11/8/72
11/28-
12/1/72
2/22-
2/27/72
1/30-
1/31/73
1/18-
1/23/73
Color Concentration , ppm
Raw
Feed
14,000
18,000
20,000
30,000
30,000-
40/000
6,000-
8,000
12,000-
16,000
Neutralized
Filtered
Feed
Average
11,000
17,000
23,000
4,500
9,000
Permeate
1200-1500
2000-3000
4000-5600
430-470
600-900
Final
Concentrate
500,000-
600,000
1,000,000-
1,500,000
1,600,000-
2,500,000
180,000-
220,000
420,000
500,000
Total
Solids in
Concentrate
%
12-16
12-15
18-22
5
6-7
Concen-
tration
Ratio
80-90
80-100
100-130
70-90
80-90
Color
Removal , %
92
90
86
94
95
--
00
to
-------
the color bodies in the effluents. The presence of
lower molecular weight lignaceous materials in the pine
caustic extraction filtrate would be expected due to
lignin fragmentation in the bleach plant chlorination
stage.
The color removal was also dependent on the concentra-
tion ratio obtained: the greater the concentration
ratio, the higher the color content of the mixed per-
meate. This is to be expected since permeate color
increases with feed color, and at high conversion a
greater fraction of the membrane area is exposed to
highly colored feed.
This effect is further demonstrated by Table 10, which
contains rejection data for selected periods when
operation proceeded smoothly. Also shown is the higher
color removal observed when processing decker effluents.
Appendix I contains some limited color rejection data
for individual membrane stages when treating pine caustic
extraction filtrate. Color rejection increased with
feed concentration (progressing through the stages of
the pilot plant). This is due to the fact that with
increasing feed concentration a higher fraction of the
feed color bodies are high molecular weight species
since lower molecular weight solutes are preferentially
removed in the earlier stages.
In addition, color rejection was clearly dependent on
membrane type. The T.J. Engineering modules showed a
wide variation in color rejection, between 87 and 96%,
which is attributed in part to poor quality control in
both membrane and cartridge manufacture, as well as
occasional leaks. The color rejection of the four Gulf
Environmental Systems modules was between 98-99.9%, and
clearly superior to rejections for either the Eastman or
T.J. Engineering modules. This was expected since the
HT-00 membrane (used in the latter two modules) intrin-
sically had a higher "molecular weight cutoff" than the
Gulf membranes.
Color rejection of the pilot plant was low on a few
occasions due to mechanical failures of several T.J.
Engineering cartridges: Especially troublesome was o-
ring failures on the permeate collection tube (see
comments column in Appendix D, and discussion in section
V.G., Module Mechanical Failures).
83
-------
2. Total Solids Rejection
Figure 35 also shows the effect of concentration ratio
on total solids content of the final concentrate. At
eighty-fold concentration of pine caustic extraction
filtrate, the final concentrate contained about 1570
total solids. The greater part of the solids were high
molecular-weight organics since the ultrafiltration
membranes used did not appreciably retain salts and low-
molecular solutes. The detailed pilot plant data demon-
strate that the total solids rejections of the membranes
were between 5 and 30%.
3. Material Balances
Material balance data for color and total solids was
collected and evaluated at least every operating shift.
Both field color tests (at as-is pH) and the standard
color tests (at pH 7.6) methods were used. Because of
the pH sensitivity of the color values in the effluents,
the standard test data was used for color balance
calculations.
Tables 11 and 12 contain composition data of samples
from tests with pine and hardwood decker effluents.
These data also show that retention of non-complexed
salts and other low molecular solutes (sodium ion,
sulfate ion, ash and total solids) was real, but low.
This can be seen best by comparison of permeate and
concentrate assays. This was probably due to the use
of the Gulf modules in Stages 4 and 5. These membranes
had about 3070 NaCl rejection.
Rejection of Fe444 and Ca44" was very high, and this is
attributed to complex formation of these ions with re-
tained organics. The pH of concentrate samples was
consistently lower than the filtered feed; and that for
permeate samples was consistently higher. Evidently,
the retained organics were weakly acidic.
The assays in Table 13 for pine caustic extraction
filtrate samples also show similar behavior. Note
values for the contents of trivalent metal oxides
(mainly Fe444 ) , Ca44 , 804"„ ash and chloride. Un-
fortunately, the samples were collected on different
days and only qualitative conclusions can be drawn.
84
-------
COMPOSITION DATA:
TABLE 11
PINE DECKER EFFLUENT SAMPLES
Date Sampled 1-31-73 3:00
Assays
Total Solids
Ash
pH
Fe++
Na+
S04~
Ca++
Color
Assays
Total Solids
Ash
pH
Fe++
Na+
S04—
Ca++
Color
Assays
Total Solids
Ash
PH
Fe++
Na+
S04—
Ca++
Color
Permeate
1/376 ppm
948 ppm
7.0
0.8 ppm
425 ppm
468 ppm
0 ppm
430 ppm
Date Sampled
Permeate
1,296 ppm
868 ppm
6.6
0.6 ppm
325 ppm
470 ppm
0 ppm
430 ppm
Date Sampled
Permeate
1,392 ppm
716 ppm
6.5
0 . 4 ppm
350 ppm
420 ppm
0 ppm
470 ppm
Concentrate
( concentration
ratio = 50)
48,740 ppm
13,532 ppm
5.9
182 ppm
4,860 ppm
7,157 ppm
448 ppm
200,000 ppm
1-31-73 5:00 pm
Concentrate
( concentration
ratio = 50)
49,312 ppm
13,652 ppm
5.6
110 ppm
4,620 ppm
7,280 ppm
320 ppm
200,000 ppm
1-31-73 8:00 pm
Concentrate
(concentration
ratio = 50)
48,256 ppm
12,240 ppm
5.3
96 ppm
4,620 ppm
7,250 ppm
280 ppn
186,000 ppm
Neutralized &
Filtered Feed
3,500 ppm
1,700 ppm
6.5
13.97 ppm
600 ppm
839 ppm
16.05 ppm
Neutralized &
Filtered Feed
3,624 ppm
1,684 ppm
5.8
6.20 ppm
575 ppm
900 ppm
16.0 ppm
Neutralized &
Filtered Feed
3,688 ppm
1,648 ppm
5,8
5.2 ppm
600 ppm
970 ppm
8.0 ppm
Raw Feed
2,244 ppm
1,012 ppm
9.5
1.40 ppm
480 ppm
260 ppm
20 ppm
7,670 ppm
Raw Feed
2,284 ppm
784 ppm
8.5
1.2 ppm
425 ppm
190 ppm
16 . 0 ppm
7,330 ppm
Raw Feed
2,468 ppm
908 ppm
9.2
0 . 8 ppm
516 ppm
240 ppm
4.0 ppm
8,330 ppm
85
-------
TABLE 12
COMPOSITION DATA; HARDWOOD DECKER EFFLUENT SAMPLES
Date Sampled 2-1-73 1:00 pm
Assays
Total Solids
Ash
PH
Fe++
Na+
S04—
Ca++
Permeate
1,344
572
6.8
1.2
325
200
0
ppm
ppm
ppm
ppm
ppm
ppm
Concentrate
(concentration
ratio = 62)
36
8
2
4
,512
,072
5.8
134
,900
,000
96.0
ppm
ppm
ppm
ppm
ppm
ppm
Neutralized &
Filtered Feed
5,184
1,636
6.0
5.2
880
350
24
ppm
ppm
ppm
ppm
ppm
ppm
Raw __ Feed
3,848
976
10.5
2.0
580
350
16.0
ppm
ppm
ppm
ppm
ppm
ppm
Date Sample 2-1-73 3:00 pm
Assays
Total Solids
Ash
PH
Fe++
Na+
SO4 —
Ca++
Color
Assays
Total Solids
Ash
pH
Fe++
rfa+
804 —
Ca-H-
Permeate
2,316 ppm
1,192 ppm
6.8
1 . 2 ppm
558 ppm
591 ppm
0 ppm
730 ppm
Date Sampled
Permeate
2,432 ppm
1,460 ppm.
6.9
1.0 ppm
625 ppm
660 ppm
0 ppm
Concentrate
( concentration
ratio = 60)
59,652 ppm
12,896 ppm
5.7
148 ppm
4,500 ppm
5,100 ppm
1,281 ppm
400,000 ppm
2-1-73 5:00 pm
Concentrate
(concentration
ratio = 60)
62,948 ppm
13,704 ppm
5.7
125 ppm
5,300 ppm
5,950 ppm
481 ppm
Neutralized &
Filtered Feed
6,296 ppm
2,280 ppm
5.6
12 ppm
920 ppm
1,040 ppm
48 ppm
Neutralized &
Filtered Feed
6,608 ppm
2,348 ppm
5.8
7 . 0 ppm
972 ppm
1,600 ppm
80 ppm
Raw Feed
2,520 ppm
848 ppm
10.5
1.2 ppm
480 ppm
200 ppm
4.0 ppm
15,000 ppm
Raw Feed
2,244 ppm
1,464 ppm
10.7
0.1 ppm
441 ppm
220 ppm
12 ppm
86
-------
TABLE 13
COMPOSITION DATA: PINE CAUSTIC EXTRACTION FILTRATE SAMPLES
ASSAY
PERMEATE
sampled on
1-30-73
CONCENTRATE
sampled on
2-13-73
sampled on
2-14-73
pH
Total Solids
Ash
Trivalent
Metal Oxides
Na+
S04—
Ca++
Cl-
7.4
4,560 ppm
2,950 ppm
3.32 ppm
1,200 ppm
580 ppm
8 ppm
1,275 ppm
6.7
142,850 ppm
27,920 ppm
4,120 ppm
10,400 ppm
1,809 ppm
8,000 ppm
5,86 8 ppm
200,000 ppm
35,000 ppm
4,500 ppm
12,000 ppm
2,000 ppm
10,000 ppm
6,000 ppm
Specific
Gravity
0.975
1.0635
1.0766
87
-------
In summary, all the rejection data have confirmed the
basic assumptions made at the beginning of the program.
Color bodies and organics in the three effluents can be
selectively concentrated by ultrafiltration. Permeates
had low color contents, especially the decker effluents,
but still contained the greater part of the total dis-
solved solids (ash) contained in the original effluents.
High solids,levels in the concentrate were achieved,
at times over 20% in the pine caustic extraction filtrate
tests.
C. ULTRAFILTRATION RATE DATA AND EVALUATION
The detailed ultrafiltration rate (flux) data have been
condensed for presentation and discussion in this section.
The dependence of flux on time and operating conditions
was complex. Operating time was important for both
short-term operation, i.e. periods up to 48 hours, and
long-term, i.e. months. Time effects can be attributed
to four factors:
--reversible membrane fouling/flow channel plugging
(short term effect);
--irreversible membrane fouling/flow channel plugging
(long term effect);
—membrane compaction (long term effect); and
--membrane cartridge compression (long term effect).
During the pilot plant experiments, various changes were
made in operating conditions and pilot plant equipment,
including several changes of membrane cartridges. For
this reason it is logical to discuss the operation of
the pilot plant chronologically, and seven distinct
periods are discussed below.
Before detailing performance of the pilot plant during
these periods, however, flux data for the membrane
system during the entire program will be presented.
In Table 14 membrane flux by shell is presented, from
the beginning of pilot plant operation through the
end of the program. These fluxes were measured at
the start of experiments; that is, after membrane clean-
ing. The efficiency of membrane cleaning was highly-
variable, and this explains in part some of the varia-
tion in day-to-day operation. These data have been in-
cluded in Figures 36 to 42, which show the flux behavior
at both the beginning and end of experiments. An
evaluation of these data follows.
88
-------
TA8X.B X4
MEMBRANE FLUX BY STAGE DURING PILOT PLANT PROGRAM
Cumulative
FEED* Date Hours of
Operation
PCEF 8-19-72
00
SO
8-21
8-22
8-25
8-30
8-31
10-25
10-28
10-31
11-1-72
11-2
11-2
11-3
11-6
11-7
11-10
f 11-11
28
32
38
50
66
80
110
135
190
200
204
220
240
280
300
336
356
la
22.5
24
13.5
12
20.2
10.5
36
30
22.5
34.5
30
28.5
24
18.9
10.5
12.9
16.5
Ib
.75
16.5
20.2
12
22.5
8.2
13.5
7.5
34.5
30
28.5
24
18.9
10.5
12
15
Permeate
1C
.4
6
15
7.5
18
7.5
15
12
13.5
33
27
28.5
22.5
21
18.7
22
18.7
Flux by Stage/ gfd
2
18
18
15
7.5
37.5
9
15
15
34.5
30
28.5
22.5
22.5
15
18.7
19.5
3
21
24
14.2
15
13.5
10.5
14.3
12
6.8
30
22.5
24
18
21
15
20.2
20.2
4
6
7.5
6.9
7.5
6
9.3
15
15
15
30
30
28.5
21
15
12.7
22.5
22.5
5
1.5
3.0
5.3
4.4
4.5
2.7
6.9
9
6.3
15
15
15
13.5
4.5
3.7
6.6
6.6
Average
flux,
gfd
10.3
14.4
10.3
10.3
14.4
8.2
14.4
14.4
10.3
29
25.3
25
19.9
16.6
11.8
15.8
16.2
-------
TABLE 14
(continued)
MEMBRANE FLUX BY STAGE DURING PILOT PLANT PROGRAM
FEED* Date
PCEF 11-12-72
VO
o
'
11-16
11-17
11-18
11-28
11-29
12-1
12-2
12-3
12-5
12-6
12-7
12-8
12-9
12-10
r 12-11
Cumulative
Hours of
Operation
376
408
430
460
476
516
530
550
585
600
615
636
644
660
680
la
13.5
22.5
19.5
19.5
18
10.5
15
15
13.8
17.4
16.5
13.8
34.5
33.8
30
34.5
Ib
12.7
19.5
18
15
18
10.5
15
15
13
15
12.9
11.7
12
12
10.8
12
Permeate
Ic
17.2
20.2
18
18
18
9
14
21
12
15
14.4
11.9
21
18
18
19.5
Flux by Stage/ gfd
2
16.5
21
19.5
15
15
15
19.5
18
18.7
18
18
15
18
15
15
15
3
15
18.7
18
14
15
12.7
18
15
15
15
18
18
16.5
18
14
16.5
4
15
8.7
15
14
8.1
6.0
15
15
15
17
16.5
12.4
15
15
15
13.8
5
8.9
3.0
3.0
2.4
3.0
2.4
2.4
4.5
6.0
7.3
6
2
2
2
6.3
5.4
Average
flux,
gfd
13.6
15
13.4
13
8.8
13.6
14.2
12.9
14.3
13.9
11.5
16.3
15.6
15
16
-------
TABUS 14
(continued)
MEMBRANE FLUX BY STAGE DURING PILOT PLANT PROGRAM
FEED* Date
PCEF 12-12-72
o
.j
i
12-13
12-14
12-18
12-19
12-20
12-21
12-22
12-23
12-26
12-27
12-28
1-4-73
1-5
1-6
1-7
' 1-8
Cumulative
Hours of
Operation
700
714
734
760
781
797
813
824
846
869
897
919
943
967
989
1010
1033
Permeate Flux by Stage, gfd Average
la Ib
33 12
34.5 10.5
34.5 10.5
34.5 10.5
13.5 36
34.5
34.5
34.5
34.5
34.5
36
34.5
34.5
37.5
34.5
34.5
Ic 2
17.2 13.2
16.5 16.5
15 13.5
13.5 15
13.5 12
11.7 10.5
12 12
12.8 10.7
11.3
10.5
12
10.5
17.2
17.2
18.5
18.5
20.4
3
15
16.5
16.5
15
15
14.3
13.8
13.8
12.7
15
16.5
15
15
15
16.5
16.5
16.5
flux,
4 5 gfd
13.2 3.4
6.7
6.0
4.5
3.0
4.3
3.8 3.5
6 6
3.6 3
4.4 4.5
5.7 4.5
3.3 2.3
3.9 5
4.5 4.8'
5 2.7
3.3 1.5
4.5 3.9
-------
TABLE 14
(continued)
MEMBRANE FLUX BY STAGE DURING PILOT PLANT PROGRAM
Cumulative
FEED* Date Hours of
Operation
PCEF 1-9-73
i 1-10
HDE 1-18
> '
5
1-19
1-22
1-23
1-24
•J
PDE 1-26
>
1-29
1-30
' 1-31
HDE 2-1-73
MDE 2-2
PCEF 2-12
^
2-13
2-19
, 2-20
1056
1078
1092
1098
1110
1120
1127
1136
1146
1158
1172
1184
1196
1000
1016
1025
1035
la Ib
31.
33
27.
22
29.
26.
25.
29.
29.
34.
27.
26.
24
24.
25.
22.5 26.
21 25.
Permeate
Ic
5
6
4
4
7
1
1
2
3
9
8
5 22.5
3 22.5
5 21
Flux by Stage/
2 3
18 17.
18.5 17.
10.8 13.
9.6 13.
11.5 13
9.9 11.
9.6 10.
10.4 11.
12.
13
11.
12.
10.
13.
13.
15
12
gfd
2
2
5
5
4
7
6
9
1
4
5
8
5
4
4.
4.
5.
5.
5.
4.
3.
5.
4.
4.
4.
5.
4.
15
15
15
13.
5
5
3
3
3
2
8
3
5
5
5
3
5
5
Average
flux,
5 gfd
3
3.8
6
6
6
3.8
4.5
6
5.3
6
4.5
4.5
4.5
16
13.5 16.2
5.4 16.2
9.6 15.8
-------
TABLE 14
(continued)
MEMBRANE FLUX BY STAGE DURING PILOT PLANT PROGRAM
FEED* Date
PCEF 2-21
S
i
2-22
2-23
2-26
2-27
2-28
3-1
' 3-2
Cumulative
Hours of
Operation
1052
1072
1090
1106
1128
1140
1154
1171
la
19.
18.
18
17.
16.
16.
19.
17.
5
8
3
5
5
5
1
Ib
24
24.
24
24
22.
23.
19.
24.
8
5
3
5
6
Permeate Flux
Ic 2
21
21
19.5
18.7
16.5
18
19.5
17.1
by Stage/ gfd
3
12
13.5
12
13.8
12
12
14.9
12.2
4 5
13.5 9.6
15 10
15 8.3
15 4.5
14.2 10
13.5 9
11.8
13.2
Average
flux/
gfd
15.
15.
14.
14.
13.
13.
13
12.
6
3
8
1
7
7
6
*FEED Abbreviations Key:
PCEF - Pine Caustic Extraction Filtrate
HDE - Hardwood Decker Effluent
PDE - Pine Decker Effluent
MDE - Mixed Decker Effluents
-------
36
32
28
24
20
12
A
i_i at beginning of test
O at end of test
A
A
A
A
A
Eastman Cartridges
•si
I
A.
A
A
OO
,low circulation rate
A
circulation
rate
low circulation
rate
T.J. Cartridges
T.J. Cartridges (old set)
(1948,1970,3749)
T.J. Cartridges from
Stage Ib (low flux) T.J. Cartridges
(4713,4723,4724)
I
(4709,4711,4719)
I I
(1950,4714,4722)
I
August
September October
November December January
1972 1973
FIGURE 36 : MEMBRANE FLUX. FOR STAGE la
February
March
-------
32
28
24
•O
M-l
^20
vo
16
12
8
A at beginning of test
at end of test
A
A
A
\— A
A
T.J. Cartridges (old set)
A
T.J. Cartridges
0 574 71 i~'
T.J. Cartridges (old set)
(2026, 4283, 4395)
t I I
August
September October November December January
1972 1973
FIGURE 37: MEMBRANE FLUX FOR STAGE lb
February
March
-------
28
24
20
•o
VO CP
** .16
12
A at beginning of test
O at end of test
A
A
T.J. cartridges (old set)
A
1
A
A
A
T.J. Cartridges (4722,4725,3747)
' low circulatio
rate
\
T.J. Cartridges
(old set)
("3749, 3782, 3751BT'
August September October November December January February
1972 1973
March
FIGURE 38: MEMBRANE FLUX FOR STAGE Ic
-------
32
28
A at beginning of test
O at end of test
A
A
24
20
16
12
A
A
T.J. Cartridges (old set)
I I
I
T.J. Cartridges
(4714,4716,4717) "^
I I
T.J. Cartridges
(CorrugatedT^Spacer)
I 1
August September October November December January
1972 1973
FIGURE 39: MEMBRANE FLUX FOR STAGE 2
February
March
-------
TJ
00
28
24
20
16
12
T.J. Cartridges (old set)
at beginning of test
at end of test
A
A
A
A
T.J. Cartridges (4708,4712,4736)
A
I
I
i
I
August September October November December January February
1972 1973
March
FIGURE 40: MEMBRANE FLUX FOR STAGE 3
-------
Gulf
^ ..
Cartridges
_(#!, #2)
T.J. Cartridges
(old set)
A at beginning of test
at end of test
T.J. Cartridges Gulf Cartridges T.J. Cartridges
14737','4735,4731V ^ ":""(f 1, #2) ^ ""("4737", 4 731 f 4721JT
A
A
24
20
A
(n
12
A
o
°
A
August SeptemberOctober November December January February
1972 1973
March
FIGURE 41: MEMBRANE FLUX FOR STAGE 4
-------
Gulf Cartridges
(13, #4)
<—^ at beginning of
test
O at end of test
P.J. can
T.J. Cartridges
(4718,4720,4721)
Gulf Cartridges (#3, #4)
T.J. Cartridges
>- -33 —
(4718,4720,4735)
•O
§X
12
August
September
October November December January
1972 1973
February
March
FIGURE 42: MEMBRANE FLUX FOR STAGE 5
-------
1. Operating period: 8/19/72-8/31/72 (Start up).
Pilot plant operation was initiated with pine caustic
extraction filtrate on 8/19/72. At this time, the mem-
brane system was equipped with Eastman cartridges in
Stage la; T.J. Engineering cartridges in Stages Ib, lc,
2 and 3; and Gulf cartridges in Stages 4 and 5.
It was immediately apparent that two related problems
were present. First was that the ultrafiltration rate
declined very rapidly during the course of an experi-
ment due to membrane fouling. Second/ restoring the
flux by flushing or cleaning the membranes proved to be
difficult. Subsequent runs showed the identical pattern
of rapidly declining flux after start up, and great
difficulty in membrane cleaning.
This behavior was evaluated, and the following con-
clusions were drawn:
—the filtration system, consisting solely of the
Broughton lOia basket filters, was inadequate
(as discussed in Section V.A.); and
—the problem was possibly aggravated by the fact
that under certain conditions it was possible
for unfiltered feed to bypass the Broughton
filters and enter the membrane system. This
possibility was eliminated by installing a check
valve in the appropriate place (Stage 1 recir-
culation loop) on 8/31/72.
On the basis of these results it was decided to evaluate
other filtration alternatives to avoid or minimize the
membrane flux decline. In order not to expose the en-
tire membrane system to the possibility of inadequate
filtration by the filtration sequences to be examined,
it was decided to conduct these tests with a single
membrane cartridge.
2. Operating period: 9/1/72-10/24/72 (single cartridge
tests). A series of 50 to 100-hr tests were conducted
with various filtration sequences and a single membrane
cartridge. The first test was with the Broughton filters.
Data for short term operation are contained in Figure 43.
As is evident, there was a very rapid flux decline, which
confirmed the conclusion that the lOy Broughton filter
was inadequate.
A more retentive filtration system was then installed.
The new filtration sequence consisted of:
101
-------
o
10
60
Hydromation Filter
stopped for
pump repair
depressurization
-£>- —-
Filter
I
0
20
40
60 80 100 120
Operating Time, hours
140
160
180
FIGURE 43;
EFFECT OF PREFILTRATION EFFICIENCY ON ULTRAFILTRATION RATE
(T.J. Eng. Co. Cartridge, 110 psig; lOa'F, 5 gpm circulation rate)
-------
—a Bauer-hydrasieve filter (to remove fibers and
gross solids);
—a Hydromation depth filter (as the main filter);
and
—ly Cuno cartridge filters (for final polishing).
The suspended solids removal efficiency of this filtra-
tion sequence was substantially better than that of the
Broughton filters. This was reflected in a substantial
improvement in membrane flux stability, as seen in
Figure 43. It was possible to maintain a membrane flux
exceeding 20 gfd with this filtration sequence during
a 130-hr test. The short term flux decline (i.e. over
24 hrs) was about 25 to 35%, and it was possible to
largely recovery flux by intermittent shut down (system
depressurization). Complete flux recovery was achieved
by detergent cleaning.
The cleaning procedure consisted of a low-pressure (20-
40 psig) and high flow (5-6 gpm) once-through water
flushing, and detergent cleaning (described in Section
V. F.) .
Unfortunately the filtration capacity of the Hydromation
filter was less than that required to treat the feed
needed for operation of the entire pilot plant. The
filter vendor was unable to deliver a larger unit within
the budget and schedule constraints of the program.
Therefore it was decided to install a precoat filter,
which was expected to give particulate removal equivalent
to that of the Hydromation depth filter.
As described in the section on pretreatment, a Shriver
filter press was available within the mill. This was
installed in place of the Hydromation depth filter,
keeping both the Bauer hydrasieve and the Cuno polishing
filters in the system.
The single membrane cartridge was operated for about 50
hrs with the Shriver filter press, using wood flour as
a precoat material and body feed. This filtration se-
quence was also found to effectively remove suspended
solids as shown by the data in Figure 43. Its filtra-
tion efficiency was somewhat less than that for the
Hydromation filter, since the rate of membrane flux
decline within relatively short periods, e.g. 24 hrs,
was somewhat more rapid. However, it appeared that
103
-------
acceptable ultrafiltration rates could be maintained,
and membrane cleaning efficiency was greatly improved.
Moreover, the Shriver filter press had sufficient
capacity for operation of the entire pilot plant. Thus
at the end of these tests with the single cartridge it
was decided to resume operation of the entire pilot plant.
3. Operating period: 10/25/72-10/31/72 (renewed pilot
plant tests). The pilot plant operation was renewed
on 10/25/72 using the Shriver filter press. The first
step was to clean the pilot plant according to the
detergent cleaning method developed for the single car-
tridge. The result was a substantial improvement in
flux for all stages (see Table 14). However, during
operation on subsequent days, the flux decline was un-
satisfactory. This was particularly true for Stages Ib,
2 and 3. At the time it was concluded that the initial
set of membrane cartridges was irreversibly fouled or
plugged by inadequately filtered feed during August,
and was not suitable for obtaining meaningful results.
A new set of membrane cartridges was ordered, in order
to conduct the remainder of the experimental program
under representative conditions, that is, with mem-
brane cartridges which had not been subjected to ex-
treme upset conditions. It now appears likely that
the unsatisfactory flux decline observed at the end of
October was due to a combination of this hypothesis
and inadequate filtration with the Shriver filter press.
4. Operating period; 11/1/72-11/27/72. The new set
of T.J. Engineering cartridges was received and installed
in the pilot plant on 11/1/72. Operation was resumed
with pine caustic extraction filtrate. Initial fluxes
were high, in the range of 25-35 gfd. By reference
to Figures 36 to 42 and Table 14, it .can be seen that
performance was substantially improved over that of
any earlier period. Nevertheless, a continuous decline
in flux and increase in pressure drop across the mem-
brane stages was observed. Complete flux recovery by
the cleaning procedure developed for the single spiral
was not obtained. However, the system appeared to be
stabilizing at a flux rate level of approximately
15 gfd, but pressure drops in Stages 1, 2 and 3 were
high.
One membrane change was made during this period. When
a leak occurred in Stage 5 on 11/6/72, the cartridges
were replaced with two of the original Gulf membrane
104
-------
cartridges. These cartridges remained in the pilot
plant until 2/12/73.
During this start up period with the new membranes
several observations were made on the condition of the
cartridges. These were:
Date Observation
11/14/72 to Brine seals of almost all
11/17/72 cartridges, except the two
Gulf cartridges in Stage 5,
were found to have "flipped",
potentially allowing feed to
short-circuit the membranes.
Membrane cartridge compres-
sion effects were also noticed
(see Section V.G.: Module
Mechanical Problems). The
brine seals were reinforced
and the cartridges returned
to their respective shells.
12/7/72 The brine seals of the car-
tridges in shells la, Ib, and
Ic had flipped. The cartridges
were reinstalled with the
brine seals properly positioned.
In addition, other limitations of the system became
apparent during this period. Specifically, the pilot
plant piping did not permit once-through flushing of the
membrane cartridges at low pressure and high flow. Also,
it was suspected that the Shriver filter press was not
providing a high efficiency for suspended solids
removal.
Thus, the following system modifications were undertaken
and completed by 12/7/72.
—A continuous wood flour body feed addition
system was installed.
—A regular procedure was instituted to apply a
layer of fine paper on top of the filter cloth
of the Shriver filter press. This aided in
the formation of the precoat layer, as well as
facilitating filter cleanup.
105
-------
—Piping modifications were made such that each
membrane shell could be flushed at high flow and
low pressure on a once-through basis.
5. Operating period; 11/28/72-1/10/73. After the changes
to the filtration system and piping had been made, oper-
ation continued on a regular basis. Additional membrane
changes were made, as summarized below.
Change
Stage la cartridges leaked;
replaced with three TJ
Engineering cartridges from
the initial set.
12/19/72 Switched position of Stage
la and Ib cartridges.
12/21/72 Stage Ib(formerly la) car-
tridges were replaced by three
TJ cartridges from the initial
set.
12/21/72 Installed two Gulf cartridges
from the initial set in
Stage 4.
1/4/73 Installed three TJ Engineering
wide-channel corrugated-
spacer cartridges in Stage 2.
Observed that two of the three
cartridges removed had torn
brine seals.
Stage 1
On 12/8/72, Shell la was observed to have a leak. The
membrane cartridges were replaced with three of the
original set of TJ Engineering cartridges. Surprisingly,
these three older cartridges showed a very high flux,
about 30 gfd.
To determine if the position of the la Shell influenced
106
-------
performance, the three cartridges in Stage la were
shifted to the Ib position on 12/19/72. For the two
days during which these cartridges were tested in this
position, membrane flux continued to be high.
On 12/21/72, the three cartridges in the Ib position
were replaced by another three TJ Engineering cartridges
from the initial set, in order to see if they also had
high flux. As seen from the data in the Table 14,
these three cartridges did indeed exhibit high, stable
flux, and continued to do so for the rest of the test
program. Specifically, flux remained more or less
constant until 1/10/73, when operation with decker
effluents was initiated.
It was immediately suspected that unequal flow distribu-
tion between the parallel shells in Stage 1 caused this
behavior. Just before beginning the tests with the
decker effluent, flows through the shells in Stage 1
were measured. It was observed that approximately 65%
of the total flow was passing through Stage Ib, 30%
through Stage la, and only 5% through Stage Ic.
It was concluded from these results that it is neces-
sary to maintain a high flow through spiral-wound car-
tridges with standard mesh spacers to maintain a satis-
factory flux. High flow also results in facilitated
cleaning of the membrane cartridges and flux recovery.
Furthermore, in an actual plant installation, flow
distributors between parallel shells will be required.
Finally, it appears that a minimum circulation rate of
about 8 gpm will be required to maintain high flux.
Data showing performance of these membrane cartridges
operating under acceptable conditions are contained in
Figure 44. Shown is flux as a function of operating
time. During any period between cleaning, flux de-
creased due to reversible membrane fouling. However,
membrane flux was completely recovered during cleaning.
Stage 2
Stage 2 fluxes ranged between 15 and 22 gfd until mid-
December. After this time, flux began to decline slowly.
Pressure drop for Stage 2 increased to a level of about
70 psi (at 4.5 gpm circulation rate) on 11/29/72. After-
107
-------
•a
*.
x
rH
t,
40
30
20
10
1/4/73
Pine Caustic Extraction Filtrate
Decker Effluents
- -•>-1 -^ 1:- -3-
-f*-
_L
1000
?
1100
1/18/73
Cumulative Operating Time, hours
1200
FIGURE 44
ULTRAFILTRATION RATE DATA: STAGE Ib
-------
wards/ pressure drop decreased and this was due to brine
seal reversal. On 1/4/73, when the Stage 2 cartridges
were removed to replace them with three new TJ Engineer-
ing cartridges (wide-channel, corrugated spacers), it
was observed that two out of the three Stage 2 cartridges
had brine seal failures. It is likely that after brine
seals failed, feed partially bypassed the cartridges
which then slowly fouled due to low flow (see Table
14).
Three corrugated spacer cartridges were installed in
Stage 2 on 1/4/73 to determine if this "hydrodynamically
clean" flow channel spacer would permit operation with-.
out cartridge plugging. It is generally necessary to
operate this type of cartridge at a higher feed flow
rate than in mesh spacer cartridges. The open cross-
sectional area in corrugated space cartridges is greater,
and unless one operates at the same or higher linear
velocity than in a mesh-spacer cartridge, concentration
polarization can reduce flux below the water flux of the
cartridges. It was anticipated that the flow rate re-
quired to overcome this effect of concentration polari-
zation would exceed the 5-6 gpm which the Stage 2 circu-
lation pump could deliver. Thus, the purpose of the
test was not to collect meaningful ultrafiltration rate
data but to demonstrate that the cartridges could oper-
ate without flow channel blockage and the corresponding
irreversible buildup in pressure drop and flux decline.
An examination of the data in Figure 45 shows that no
long-term degradation in flux was obtained in the tests
with the pine caustic extraction filtrate. (The data
for the decker effluents are explained below). The
clean module rates, were observed to be about 50% of the
water flux of the cartridges, and this demonstrates that
concentration polarization was relatively severe. By
contrast, flux of pine caustic extraction filtrate in
all other stages was generally observed to be identical
to the water flux.
Since the initial flux of the corrugated spacer car-
tridges could be recovered repeatedly, it is concluded
that the use of this flow channel spacer can result in
complete membrane cartridge cleanup.
109
-------
30
g
8 20
tr
x
10
0
1/4/73
Pine Caustic Extraction Filtrate
Decker Effluents
1000
/ 1100
1/18/73
Cumulative Operating Time, hours
FIGURE 45. ULTRAFILTRATION RATE DATA: STAGE 2
-------
Stage 3
The membrane flux of the Stage 3 membrane cartridges
remained between 12 and 18 gfd througout this test peri-
od. It is concluded that neither brine seal failure
nor substantial irreversible flow channel plugging
occurred. The gradual decline in flux is attributed to
irreversible membrane compaction (see Section V.F. for
the effect of time on water flux of these membranes).
Stages 4 and 5
The flux of Stage 4 also remained reasonably high, until
12/12/72, at which point a leak occurred and flux was
no longer measured.
Membrane flux for the Gulf cartridges in Stages 4 and 5
generally ranged between 2 and 6 gfd, and this is
attributed to the lower flux characteristics of the Gulf
membranes. It is also possible that these membrane
cartridges may have partially dried out during the
period that they were not used.
6. Operating period; 1/18/73-2/2/73 (Operation with
Decker Effluents). During the end ot January, 1973, the
pilot plant was operated for periods of approximately
40 hrs with hardwood and pine decker effluents. Un-
fortunately, these tests were conducted when the pilot
plant system was not operating with acceptable flux in
most of the stages. Thus, the tests with the decker
effluents provide preliminary information on the separa-
tion efficiency of the membranes and, to a limited ex-
tent, values of membrane flux.
Referring to Table 14, meaningful flux can be obtained
from the data for Stages Ib, 2 (corrugated spacer car-
tridges) , and 3. The data for Stage Ib and 3 indicate
that the flux for the two decker effluents was some-
what less than flux for the pine caustic extraction
filtrate. This was more pronounced for Stage 2. This
behavior is explained by the following hypothesis. The
organics contained in the decker effluents have higher
molecular weights than those in the pine caustic
extraction filtrate, since fragmentation of the organics
in the latter waste occurs in the bleach plant chlorina-
tion stage. Thus, at a given retained organics loading,
111
-------
concentration polarization should be more severe for
the decker effluents since the diffusivity of the dis-
solved solutes is lower. For this reason, flux in
Stage 2 was reduced by almost a factor of two below
that for pine caustic extraction filtrate.
The lower flux in Stage 1 may also be due to a gradual
failure of the Stage 1 pump. It was discovered that
the capacity of the Stage 1 pump was low and this could
have resulted in a flux reduction due to concentration
polarization. Also, near the end of the tests with the
decker effluents, the Stage 2 pump began to fail, and
this may account in part for the reduced Stage 2 flux.
7. Operating period; 2/3/73-3/1/73. At the end of
the tests with the decker effluents it was apparent that
additional modifications to the pilot plant were desir-
able. The following changes were made:
—a Sparkler pressure leaf filter replaced the
Shriver-filter press;
—a Sparkler-Velmac 5y disc filter (backwashable)
replaced the Cuno cartridge filters;
—the Stage 1 and Stage 2 pumps were repaired,
and their flow capacity was restored.
During the first week in February, 1973, all the TJ
Engineering cartridges were tested in the single car-
tridge test shell. On the basis of the results of these
tests, the best membranes were selected and installed
in the pilot plant. Details are given in Appendix B.
Briefly, the cartridges in Stages Ib, 2, and 3 were not
changed; the four Gulf cartridges in Stages 4 and 5
were replaced by six TJ Engineering cartridges; two of
the three cartridges in each staqe had been previously
used in the same stage in November, 1972. Stages la and
Ic were filled with other good remaining TJ Engineering
cartridges.
Initially some persistent problems with O-ring leaks
were encountered, but eventually these were solved.
Operation began on 2/12/73 with pine caustic extraction
filtrate. In general, membrane fluxes were similar to
those observed in November and December.
112
-------
The membranes in Stages Ib and 3 had not been changed.
In treating pine caustic extraction filtrate, the
flux for Stage 3 returned almost to the original level.
But in the case of Ib there was a certain amount of
irreversible flux loss. The reason for this is not
clear.
In Table 14 flux levels are shown for both the begin-
ning of an experiment (approx. 1 hr) and the end
(usually 8 to 17 hrs). The flux decline for Shell Ib
usually did not exceed 20%, compared to declines of
40 to 50% for Shells la and Ic. This is presented
graphically in Figure 46, which contains data for
Shells la and Ib. It is seen that the initial water
flux for the two shells were identical. However, the
initial flux with pine caustic extraction filtrate was
higher for Shell Ib and did not fall off as rapidly as
that of Shell la. Furthermore, membrane cleaning
recovered the initial flux for Shell Ib, while for
Shell la the initial flux decreased with time. In
addition, the water fluxes of the two stages diverged,
with water flux for Shell la declining relative to that
for Shell Ib.
At this point, circulation flows in the different shells
were measured and it was found that the flow through
Shell la was approximately 1.3 gpm; Shell b, 9.5 gpm;
and Shell Ic, 4 gpm. So again it was apparent that the
shell with the highest circulation, flow exhibited
superior performance.
In general, flux for the pine caustic extraction filtrate
was equal to the water flux of the membranes. Notable
exceptions were the behavior in Stage 2, with corrugated
spacer spirals, and in Stage 5. As can be seen in Figure
42, the Stage 5 flux for the TJ Engineering cartridges
was more or less equal to the Stage 5 flux for the Gulf
cartridges, even though the TJ Engineering cartridges
had a substantially higher water flux. This demonstrates
that at least in Stage 5 the feed concentration at-
tained a level sufficiently high to reduce flux, that
is, flux in this stage was limited by concentration
polarization, and not by the membrane water flux.
113
-------
fr.
12
8
Stage la,~' pine Caustic.
' Extraction
G Stage Ib,
4 Stage la,^
I Stage Ib,
Filtrate Flux
Water
' 72°F, 90 psi
°J
2/20/73
10
20
30 40 50
Operating Time, hours
60
70
80
90
FIGURE 46: FLUX VS TIME FOR STAGES lA and IB
-------
Since the color rejection of the TJ Engineering
cartridges was lower than that of the Gulf cartridges
(more color was observed in the Stage 5 permeate)
it would be desirable in a full-scale installation
to use lower flux, more selective membranes in the
final stages).
D. PRESSURE DROP DATA
Pressure drop (and circulation rate) data for the five
stages are shown in Figure 47 to 51. Pressure drop was
strongly dependent on circulation rate, as can be seen
from inspection of the data. For example, when the
Stage 1 pump lost capacity (flow), pressure drop across
this stage decreased sharply. (Note: one would expect
turbulent-flow pressure drop to increase with the 1.8
power of circulation rate.)
During November, pressure drops for all stages became
quite high, exceeding 50 psig in the early stages. In
some stages (e.g. Stage 2), brine seal failure led to
a reduction in pressure drop since feed flow bypassed
the membrane cartridges.
The high pressure drop was due to inadequate prefiltra-
tion and an inability to flush and clean the cartridges
thoroughly.
The changes made in the pilot plant in early December,
piping modifications to permit once-through flushing and
the use of paper on the plates of the Shriver filter
press, resulted in improved membrane cleaning. This
was generally reflected in lower pressure drops in the
pilot plant after mid-December. The high pressure drop
in Stage 1 in February, 1973 was due to operation at
increased flow.
Pressure drop across Stage 2 when corrugated spacer car-
tridges were used was negligible, and no increase with
time was observed.
Thus, it is concluded that negligible flow-channel
plugging occurred with the corrugated-space cartridges.
115
-------
20
10
o;
• •
• tt
*•• •
M •
.V
60
50
V)
a 40
M
Q
0)
M
« 30
0)
0)
M
20
10
A A
4 44
4
4 4
4 4
4
4 44
**
4
4
4 4 4t
4 4 A
Auqust September October November December January February March
1972 1973
FIGURE 47: STAGE 1 PRESSURE DROP
-------
fa
•a
-------
& 6
8 «
1 2
0)
• t ••
V
•*•• •%*%• *—
60
§40
0)
n
9
CO
n
o*
20-
1C-
4
4
4 44
4 4
4
44
44* 44
4 4M
4
4 4
44
4
August September October November December January February March
y 1972 1973
FIGURE 49: STAGE 3 PRESSURE DROP
-------
•O 2
-------
i
• • • /• - % •
•» • »•»•»••• r • •«•«• *•
V
H 4
T3
0>
0)
0> 9
2
• w •
. • ft • * •
•
•
60
50
o
• •**
-H 4°'
to
30
0)
M
S 20
0)
M
10-
*
• .
* A *A A4 AM*«* / *
* * A
* A
* A* * *
•*•
August September October November December January February March
1972 1973
FIGURE 51: STAGE 5 PRESSURE DROP
-------
E. IMPORTANT FACTORS CONTROLLING MEMBRANE FLUX
1. Slime Formation on the Membrane Surfaces
The membrane flux .decline observed was controlled by
several interrelated phenomena. It is instructive first
to consider a "normal" case, in which a continuous and
uniform feed flow is maintained across the membrane
surface. In this case, a "gel" layer of slightly solu-
ble colloidal and macromolecular solutes can build up
on the membrane surface with time, reducing flux. This
behavior has been reported previously for a variety of
feed solutions (27)f and undoubtedly occurred in the
ultrafiltration of the three effluents processed in
this program.
During the program "slimes" removed from membrane
spirals were analyzed, and other tests were performed
in an attempt to characterize the troublesome "foul-
ants".
Analyses of the slimes indicated the presence of large
amounts of ash in the slime (60-70%). The ash con-
sisted primarily of finely divided clay, calcium com-
pounds and anti-foam agents such as silicones. These
ash materials were traced to paper machine white water
which was used for a time as make-up for the caustic
extraction stage. Beginning in January white water was
not used for this function.
Materials similar in composition to the slime are found
in the effluents of every pulp and paper mill process
using water which has been in contact with cellulose
pulp. In all cases, these effluents can be filtered
(or even ultrafiltered), and the filtrate when refiltered
24-72 hours later will have 10-100+ppm of a grayish floe
which appears similar to the material removed from the
membrane surfaces. This suggests that the minimal
membrane fouling and slime formation observed in earlier
laboratory experiments may have been a result of the
age of the material studied. The slime-forming material
may have flocculated during the time of sample shipping
and most of it may have been removed in the laboratory
prefiltration.
121
-------
In order to develop a better understanding of the
mechanism of slime formation, a series of tests and
analyses were conducted. The slimy solids were examined
bacteriologically in several laboratories. No evidence
was developed to indicate the presence of any micro-
organism.
An extensive experiment was conducted on fresh pine
caustic extraction filtrate. The fresh sample was
adjusted to a given pH. Suspended solids were deter-
mined by filtration of 500 cc samples through Whatman
No. 40 filter paper. The filtrates were allowed to
stand for 24 hrs at room temperature and at 120°F. The
suspended solids in these filtrates were then deter-
mined as before.
A typical set of data from these experiments is given
below:
Pine Caustic Extract Sample 12/7/72
Adjusted
PH
(original)
12.5
8
7
6
Suspended Solids in
Filtrate after 24 hrs,ppm
Suspended at Room
Solids, ppm Temperature at 120°F
53
43
39
58
113
4
29
41
1
28
43
From experiments of this type it appears that part of
the problem of prefiltration and membrane fouling may be
due to pH adjustment to pH 7 or below. The mechanism
which causes the slimy material to appear acts as
though it is a micellar system with an isoelectric point
between pH 6 and 7.
122
-------
A series of detailed chemical analyses were conducted
on the solids formed in the parent liquor, the retentate
and the concentrate as well as the solids flushed out
of the pilot unit during the washing operation.
One of the samples most fully investigated was Sample
#3, January 19, 1973, which was a water flushing from
the pilot plant. Details are given in Appendix E.
From such analyses, it appears that the basic slime
forming material is polysaccharidic in nature and there
is some component of long chain hydrocarbon materials.
Materials derived from the cellulosic fibers are
believed to be the basic contributors to this slime.
One other set of observations of merit in this discus-
sion is that during ultrafiltration of both caustic
extract filtrates and decker effluents with bulk pH of
about 7, the concentrates displayed pH's of about 5-6
and the permeates had pH's 1-2 points higher.
From considerations such as these, it is hypothesized
that the slime forming and membrane fouling constituents
are present in the effluents predominantly as small
miceliar structures. These constituents are derived
from the cellulose fibers and may be stabilized in the
liquid by small amounts of tall oil soaps.
It is hypothesized further that the concentration ef-
fects at the membrane surface, and the concommittant pH
depression, cause the micellar structure to agglomerate,
resulting in the slime formation on the membrane sur-
face.
The results obtained thus far in developing some under-
standing of this solid formation problem suggest that
the formation of these materials, and membrane flux
decline, can be minimized by:
—operating the system at as high a pH as
feasible consistent with the use of cellulose
acetate membranes;
—operating the membrane modules at high feed flow
to minimize concentration polarization.
123
-------
2. Plugging of Membrane Cartridge Flow Channels
The formation of slimes on the membrane surface can be
greatly aggravated by blockage of the cartridge flow
channels. Specifically, if the flow channels become
partly blocked, flow channeling through only part of the
cartridge results in a part of the membrane area being
exposed to stagnant, or nearly stagnant, feed. Concen-
tration polarization and slime formation become severe,
and part of the membrane area becomes ineffective. Also,
when the brine seals "flipped" flow bypassed the car-
tridges, at least in part, and the reduced net feed
flow led to greater concentration polarization and
membrane fouling.
The flow-channel plugging problem was encountered for
all membrane cartridges with standard mesh spacers.
The cartridges in shells la and Ib which operated at
very high feed flow (^7-10 gpm) showed substantially
less collection of solids - as concluded from flux
stability and nearly complete recovery of water flux on
cleaning.
Other cartridges which became plugged did so irreversibly,
Solids continued to accumulate with time, until the
brine seals flipped. Autopsies of a few cartridges with
high pressure drop showed extensive amounts of a grayish
gelatinous material which filled the mesh spacer, and
presumably, occluded membrane area.
For the TJ Engineering cartridges another problem arose.
The brine seal was on the downstream end of the cartridge
Thus, the entire cartridge exterior was at the cartridge
inlet pressure, while the interior pressure was at the
inlet pressure, less the pressure drop to that point in
the cartridge. If a 20 psig pressure differential
existed across a cartridge (inlet to outlet), the static
pressure difference between the exterior and interior,
at the outlet end, would be 20 psig. This resulted in
a net compressive force being exerted on the cartridges.
Since the TJ Engineering cartridges were loosely wound,
extensive deformation occurred. This further aggravated
the internal flow maldistribution problem.
124
-------
Finally, in addition to flow channeling within a car-
tridge, channeling occurred among the Stage 1 shells.
This appears to have been an unstable situation. When
a flow maldistribution occurred, the shell with high
flow remained clean, while the low flow shells continued
to collect solids and plug since the low flow was not
able to sweep the accumulated particulates out of the
modules.
3. Prefiltration Efficiency
The efficiency of suspended solids removal greatly in-
fluenced the degree of membrane cartridge plugging by
suspended solids, and hence, the membrane fouling
characteristics as discussed above. In addition,
depth and precoat filtration could have removed colloidal
material which contributed, in part, to the slime
formation.
4. Feed Circulation Rate
As discussed above, increasing the feed circulation rate
reduced the flux decline for two reasons. First, it
reduced the rate of particulate collection within the
modules. Second, it reduced concentration polarization
and, hence, the rate of slime formation.
F. GLEANING PROCEDURES AND EFFICIENCY
1. Procedures
The cleaning procedure that was found to be the most
effective consisted of the following three steps:
Step 1: Low-pressure and high-flow water flush of
the pilot plant in a once-through manner.
Step 2: Low-pressure and high-flow detergent
cleaning for about 30 minutes with recircu-
lation of the detergent solution.
Step 3: Repetition of Step 1 at the end of the
detergent cleaning.
The water flushing was the most important part of the
cleaning procedure and was done at 20-50 psig and 2-3
gpm per shell. Reverse-flow flushing was found to clean
the system faster than forward-flow flushing.
125
-------
Detergent cleaning was not always required and sometimes
was performed on alternate days. However, on most
occasions, detergent cleaning helped in removal of
the gel-type foulant at a faster rate. After 12/5/72,
the pilot plant was cleaned only at the end of every
15-22 hrs of continuous operation. The detergent
cleaning part consisted of circulating about 40 gallons
of cleaning solution, containing about 1% neutral or
slightly alkaline detergent, for about 30 minutes.
It was found that either one of the following sets of
operating conditions for detergent cleaning was
effective:
Procedure used during 10-40 psig pressure
December and January
2-3 gpm circulation rate
per shell
100°F
Procedure used during 40-70 psig pressure
October, February and
March 4-5 gpm circulation rate
per module
100°F
The following detergents were used for cleaning:
a) Ultraclean, Part A.; Abcor, Inc. (a phosphate
type low-alkaline detergent).
b) Tide (pH adjusted to 7).
c) 1% EDTA + 1% Triton X-100 (Rohm & Haas Co.).
d) 1% citric acid + 1% Tergitol 15-S-7 (Union
Carbide).
e) Ultraclean (enzyme detergent, Abcor, Inc.).
It was found that cleaning efficiency was not noticeably
dependent on detergent type. Ultraclean Part A, which
does not require pH adjustment, was used most frequently,
Tide was also used on several occasions.
2. Effect of Feed Prefiltration on Cleaning
Table 15 summarizes the effect of feed prefiltration and
different cleaning procedures on pilot plant performance
126
-------
TABLE 15
EFFECT OF FEED PREFILTRATION ON CLEANING
Operating
Period
Feed
Filtration
Systems
Operating
Time , hrs
Cleaning Procedures
Water
Flushing
Detergent
Cleaning
Cleaning
Efficiency
8-19-72 to
9-26-72
Broughton 10y
Filter
4-10
Inadequate;
needed modi-
fications in
the pipelines
es
Poor. Required about 4 hrs
(2-3 times repetition of
cleaning cycles) to recover
flux.
9-27-72 to
10-5-72
a. Hydromation,
depth filter
and
b. 1v Cuno
cartridge
filter
20
Yes
Improved. Required about
2 hrs to clean the system.
10-17-72 to
11-27-72
Plate & Frame
Press with
wood flour
precoat & body
feed; and
1v Cuno
cartridge
filter
20
Yes
Required about 2-2*$ hrs
to clean the system.
11-28-73 to
2-2-73
20
Modified pipina
improved flushing operation.
See details in text
for operating conditions.
Required about l-l>s hrs
to clean the system.
2-11-73 to
4-1-73
Sparkler leaf
filter with
wood flour
precoat & body
feed; and
1 v Cuno
cartridge
filters
20
Required about 1 hr to
clean the system.
-------
and cleaning efficiency. The pilot plant ultrafiltration
rate and cleaning efficiency were very poor at startup
of the plant (8/19/72 to 8/30/72), at which time the
filtration system was grossly inadequate. By contrast,
the pilot plant could be cleaned within an hour after
the installation of efficient filtration and flushing
systems.
3. Effect of Intermittent Shutdown on Ultrafiltration Rate
During a study with the single cartridge it was observed
that substantial flux recovery was possible with a
brief shutdown of the unit. Figure 43 shows the effect
of shutdown on this recovery. However, during the pilot
plant operation similar flux recovery was not obtainable.
The probable reason for this is that, in the single
cartridge case, the particulates and membrane foulants
were released during shutdown and on restarting were
flushed out of the system. There was negligible recircu-
lation of foulants in the single cartridge case. How-
ever, during the pilot plant operation, foulants released
during shutdown were picked up again either by the same
membrane stage or a subsequent one because of the high
degree of internal recirculation. This was the major
reason for poor cleaning efficiency of the pilot plant
during the period when it was not possible to flush on
a once-through basis.
Intermittent shutdown of the system during cleaning was
also found to improve the cleaning operation. The same
held for flow interruption during the water flush cycle.
4. Cleaning Efficiency in Different Membrane Stages
The cleaning efficiency of the membrane cartridges is,
for very obvious reasons, dependent on the flow distri-
bution within the cartridges. If the brine seal of a
cartridge fails and creates a short circuit, flow through
the cartridge will be reduced, and hence the effective-
ness of a flush or cleaning cycle. Also, if the flow
distribution within the different parts of the brine
channel is not uniform, those parts where there is lit-
tle flow will remain fouled. Many of the TJ Engineering
cartridges became deformed under the compressive force
discussed above, thereby affecting the flow distribu-
tion within their brine channels. These cartridges,
128
-------
compared in particular to the Gulf cartridges, showed
high pressure drop and inadequate cleaning efficiency
as measured by water and effluent fluxes.
Membrane cartridges with poor ultrafiltration performance
were not readily cleaned. Both characteristics can be
traced to unsatisfactory flow distribution within the
cartridges.
Stage 1 membrane modules were arranged in three parallel
shells. The cleaning efficiency and the ultrafiltration
rate of these shells were dependent on flow distri-
bution among the shells, in addition to the flow distri-
bution within each cartridge. The variation in flow
distribution in the three shells depended upon the type
and history of individual cartridges and their varying
resistance, as discussed before. Figure 52 shows the
effect of cleaning on Stage Ib. It is seen that clean-
ing was reasonably effective, and that the ultrafiltra-
tion rate remained relatively high. During the same
period, Stages la and Ic however showed lower flux.
These cartridges were inspected and found to have under-
gone severe deformation (compression). They had, as a
result, low flow during cleaning and ineffective
cleaning.
The three wide channel corrugated spacer cartridges
installed in Stage 2 were cleaned most easily.
Figure 53 shows the effect of cleaning on the Stage 3
membrane cartridges. These cartridges had the longest
operating life, from 11/1/72 to 4/1/73. Figure 53
shows that the water flux of these cartridges decreased
with time. The main reason for this decline appears to
be membrane compaction. The flux decline parameter,
measured by the slope of a log-log plot of water flux
vs. time was about -0.078 (Figure 54).
G. MODULE MECHANICAL FAILURES
During the operation of the pilot plant a variety of
failures with several of the 45 TJ Engineering car-
tridges used occurred. Mechanical problems with the
three Eastman and four Gulf cartridges were negligible.
Appendix D details the different problems experienced
during the pilot plant operation. By and large, these
problems appeared to be four fold:
129
-------
O Water flux before cleaning
A Water flux after cleaning
18
A
A
A
A
o
CM
16
14
(§>
O
O O
A
-H 12
(0
G
o
r-
10
8
0)
I
I
December January February March
1972 1973
FIGURE 52: CLEANING EFFICIENCY OF STAGE Ib MEMBRANES
130
-------
0 Water flux before cleaning {60 psig, 72°F)
3
rH
6-,
A Water flux after cleaning (60 psig, 72°F)
O Pine Caustic Extraction Filtrate Flux (100 psig,
100°F)
18
16
14
12
10
6
O
O O CO
A
A
A
I— A
o
i
O
O
A A
A A
A
O
O
O
O
COO
A
o
i
February
FIGURE 53: CLEANING EFFICIENCY OF STAGE 3 MEMBRANES
November December January
1972 1973
131
-------
iocr
so r-
Slope = -.078
M
0)
-P
it)
1
1
10 100
Cumulative Operating Time, hours
1000
FIGURE 54: STAGE 3 COMPACTION CURVE
132
-------
—O-ring failures;
~-membrane cartridge compression, resulting in
deformation of the cartridges;
—brine seal failures; and
—glue seam or membrane failures.
1. O-ring Failures. Membrane leaks in most cases were
due to inadequate O-ring seals on the permeate collec-
tion tubes, and poor color rejection was corrected
simply by replacing the O-ring seals. On at least five
different dates involving stages la, Ic, 4 and 5, the
poor color rejections were definitely identified as
O-ring failures.
2. Membrane Cartridge Compression. TJ Engineering
cartridges have brine seals on the downstream end, thus
the exterior pressure on the cartridges is higher than
the interior pressure due to pressure drop of feed
flowing through the cartridge. Several cartridges were
observed to "deform" and "shrink" from their round
shape. These cartridges were primarily in stages where
severe pressure drop across the cartridges was present.
It is suspected that module compression created dead-ends
which could not be cleaned, thus resulting in low ultra-
filtration rate and accumulation of fouling materials.
The latter phenomenon created additional pressure drop,
consequently more compression, and further aggravated
the problem.
Gulf cartridges, however, did not show this problem.
Gulf cartridges have brine seals on the upstream end,
and the pressure within the cartridge is higher than
the exterior pressure. The net result is that the Gulf
cartridges received an "expansive" force, which maintained
the brine channel width intact and prevented dead-ends.
3. Brine Seal Failures. As the membrane cartridges
collected solids and/or underwent compression, pressure
drop across the cartridges increased. This resulted in
brine seal reversal, permitting feed flow to bypass the
membrane cartridges. It is possible that this effect
accounted for gradually declining flux in several of the
membrane she11s.
133
-------
The four Gulf cartridges which were in the system for
a period of about seven months did not evidence brine
seal reversal. At the end of the test period, how-
ever, three of the four brine seals were observed to
have weak tapes which needed reinforcement. Several
TJ Engineering>cartridges required replacement of brine
seals (mainly foam gaskets and tapes).
The outer wraps of almost all the cartridges remained
in good shape despite the compression or expansion
forces (10 to 20 psi per cartridge).
The pressure drop data show that the brine seals of the
spiral cartridges probably failed to maintain a proper
seal when the pressure drop across the seals reached
about 20 psi.
4. Glue Seam or Membrane Failures. In some instances,
leaks can only be attributed to membrane cartridge
failures. There have been five failures of this kind
involving stages lbr lcr 4 and 5. All cartridges w.exe
from TJ Engineering Co.
Membrane cartridge failures may have happened because
of
—leakage at the glue seam;
—leakage at the material fold adjacent to the
permeate collection tube;
—membrane and/or backing material wrinkles
causing direct leakage;
—pinholes in membrane.
One membrane cartridge with poor color rejection was
autopsied and was found to have glue seam failures in
several places, especially at points where longitudinal
wrinkling was present. The Gulf cartridges had mem-
branes directly cast on membrane-support sheets. Gulf
has reported that spiral modules made in this manner
show more uniform brine channel thickness (28_) .
H. INCINERATOR STUDIES
Disposal of the concentrate from the ultrafiltration of
pulp mill effluents is a problem only if the concentrate
134
-------
contains substantial amounts of chlorides. In the
studies discussed here it was found that disposal of
the concentrate from the pine caustic extraction fil-
trate does require special processing, but that the
concentrate from the decker effluents can be bene-
ficially used in the weak black liquor system. The
following discussion applied, then, only to the concen-
trate from the pine caustic extraction filtrate.
The concentrate produced in ultrafiltration of the
caustic extraction filtrate would contain about 20%
total solids. These solids would be predominately
chlorinated lignins with some amount of sodium salts
such as sodium chloride.
The content of organic and inorganic chlorides in the
concentrate makes disposition a problem if it is to be
injected into any of the present pulp mill streams.
For example, if the concentrate were processed in the
black liquor recovery system, about 500 pounds per day
of hydrogen chloride would be liberated in the recovery
boiler from combustion of the chlorinated lignins and
also about 500 pounds per day of sodium chloride would
be introduced into the recovery boiler from the inor-
ganic solids in the concentrate. These chloride-con-
taining materials could lead to corrosion problems in the
boilers, dusting problems in the boiler stacks due to the
liberation of finely divided sodium chloride, and also to
raising the chloride level in the pulping system, as a
function of the material recycle system.
Injecting the concentrate into the lime mud kilns would
lead to similar problems.
As a result of the engineering evaluations it was con-
cluded that the disposition of the concentrate from the
pine caustic extraction filtrate would require either an
incinerator specifically designed to burn organic
chlorides, or, evaporation to high solids and subsequent
admixture with primary sludge for disposal as land fill.
Laboratory studies have been conducted on evaporation of
concentrate to various levels of solids. In all of the
tests the material dried with no apparent increase in
viscosity to about 80% solids. The material does not
scale nor foam when concentrated by evaporation. When
135
-------
a 507. solids concentrate was mixed with primary sludge,
the admixed sludge was judged to be satisfactory for
land fill. As a result of the laboratory tests it was
concluded that evaporation of the concentrate to about
50% solids could be done in commercially available equip-
ment using waste hydrogen available on the plant site
as the heat source. The 50% concentrate would be added
to the dewatered primary sludge and carried to land fill.
This method of disposal would add 10% to the present
daily primary sludge weight.
Concentrate from the ultrafiltration unit was incinerated
in February at the John Zink Company, Tulsa, Oklahoma
pilot facility. The purpose of the tests was to
establish parameters on which to base a budgetary
estimate for a full scale incineration system to dispose
of the concentrate.
The tests demonstrated that the Zink incinerator will
provide a method of disposing of the concentrate.
The operation of the incinerator requires the addition
of supplementary fuel. To minimize operating costs,
hydrogen from the electrochemical cells, which is
presently unused, would be burned.
The tests demonstrated the need for the use of a venturi
scrubber and a vent scrubber to dispose of the poten-
tially plume-forming, finely-divided inorganic salts
formed in the combustion of the concentrate. In a full
scale installation, the pine caustic extraction filtrate,
prior to neutralization, would be used as the scrubbing
fluid for the venturi scrubber and the stack scrubber.
The partially neutralized pine caustic extraction fil-
trate would then be treated in the ultrafiltration
system.
The economic feasibility of the use of an incinerator
for the disposal of the concentrate is heavily dependent
on the availability of an inexpensive fuel source. ' If
hydrogen is used as the fuel source, the fuel costs
will be those associated with the capital and operating
costs of the system required to deliver it to the
incinerator.
136
-------
The partial evaporation system for disposal of the
concentrate requires less energy input but does require
increased costs for land fill operations with the
primary sludge.
The capital and operating costs of both systems are
improved, the more concentrated the solids level pro-
duced by the ultrafiltration system. These costs are
also strongly influenced by the salt rejection charac-
teristics of the specific membrane system used; the
incinerator system is more cost sensitive than the
evaporation system. If low salt rejection membrane
systems are used, the ratio of the organic content to
inorganic content in the concentrate will be increased
and for a given solids level the heating value of the
concentrate will be increased. Increasing the organic
solids content of the concentrate will yield a material
of both higher heating value per unit concentrate
weight and also less water to be evaporated in the
combustor.
The choice of the system for disposal of the concentrate
depends on several design parameters of an ultrafiltra-
tion plant as well as the specific plant site for which
it is designed. At the present there are at least two
technically feasible methods for disposing of the
ultrafiltration concentrate from pine caustic extrac-
tion filtrate. In the capital estimates for these
systems (discussed in Section VI.A.) compromise cost
figures are used.
137
-------
SECTION VI
FULL SCALE PLANT DESIGN AND COSTS
A. FIRST STAGE PINE CAUSTIC EXTRACTION FILTRATE SYSTEM
1. Flow Schematic
A general flow schematic has been presented in the
Introduction as Figure 2. A second generalized pro-
cess flow schematic, more specific for the ultrafiltra-
tion system, is given in Figure 55. First-stage pine
caustic extraction filtrate flows from the pine pulp
bleachery through a pretreatment system. Pretreatment
consists of neutralization and filtration. The treated
feed is cooled to 100°F. Although future membranes
would process pine caustic extraction filtrate at its
normal temperature (120-130°F), current membranes prob-
ably cannot withstand a temperature in excess of 100°F
and exhibit long life. It is to be noted that no
long-term life data are available at temperatures
higher than 100°F for membrane systems treating pine
caustic extraction filtrate.
The pretreated and cooled feed is concentrated in the
ultrafiltration system. The permeate (treated effluent)
is sewered and flows to the mill's waste treatment sys-
tem. The concentrate from the ultrafiltration system
is used to sluice the filter cake, and this suspension
is pumped to an evaporator and evaporated to 50% solids.
Excess hydrogen currently available in the mill from
the electrochemical cells will be used as fuel. The
evaporator discharge will be mixed with primary sludge
and disposed of as land fill.
A more detailed process flow schematic is given in
Figures 56 and 57. These two figures differ in that
Figure 56 is for a low-flow membrane system (spiral
wound cartridges with standard mesh spacers), and
Figure 57 is for high-flow membranes (spiral wound
cartridges with corrugated spacers). The reasons for
examining these two cases are discussed below.
Sulfuric acid is mixed with pine caustic extraction
139
-------
Hydrogen
Evaporator
To
Caustic
Extraction
Filtrate
Pine Pulp
Bleachery
Water
Pretreatment
Filter
Cake
UF
Treatment
Permeate
to
Sewer
Available
2 MM gpd
at 100°F
Concentrate
FIGURE 55: SIMPLIFIED FLOW SCHEMATIC: TREATMENT OF PINE CAUSTIC EXTRACTION FILTRATE
-------
Feed
Solids
Filter Aid
Bodying Pump
Filter Pump
Dry Dumped
& Sluiced away in
UF Concentrate
To Evaporator *— -7-^
Back Washed
into Feed
Concentrate
used to Slurry •»—[Xj-
Dumped Filter Cake
Permeate
from Stage 1
Water
for
Back.
Flushing
Overflow
to Sump
HX—;£**
Connect
:o Points A
Flushing
Pump
Stages
& 3 Required
but not Shown
Stage
Circulation
i. £
Flush to Feed
Permeate Lines (.typical 1
Flush
to Feed
Stages 2, 3, 4
(Stage 1 Lines to Water Store)
Sump
Permeate Pump
(if reqd.)
Feed
Pump
FIGURE 56: FLOW SCHEMATIC FOR TREATMENT OF PINE CAUSTIC EXTRACTION FILTRATE:
LOW FLOV7 CASE
-------
reed
Dry Dumped and
Sluiced Away in
UP Concentrate
To Evaporator •*— -^_.
Concentrate Used
9 Slurry Dun1
Filter Cake
to Slurry Dump«d^__jsxL
Permeate
From Stage 1
Sta
*
Water
Store for
Back-
flushing
Overflow
to SumP
Flushing
Pump
-pad):
ranei
a-
Stage 4
Circulation
Stages 2 t 3
Required but
Not Shown
Flush to Feed
Permeate Lines (tyfical)
Membranes
Heat
Exchanger
Stage 1
Circulation
Feed
Pump
Flush to
Faed
I Sump \
Stages 2,3,4
CStage 1 lines to water store)
Permeate
Pump
(if required)
FIGURE 57t FLOW SCHEMATIC FOR TREATMENT OF PINE CAUSTIC EXTRACTION FILTRAT3
HIGH PLOW CASZ
-------
filtrate in an in-line mixing section. The rate of acid
addition is controlled by a downstream pH probe and con-
trol system.
A small portion of the neutralized feed flows to a bodying
tank. Filter aid (tentatively wood flour) is added by a
solids feeder. The rate of addition of filter aid is
such as to maintain a body feed level of 50-100 ppm. The
body feed slurry is pumped with the bodying pump back
into the feed line, where it mixes with the neutralized
pine caustic extract. This stream is pumped through leaf
filters to remove most of the suspended solids contained
in the feed. Instrumentation for the filter system will
control operating pressure, pressure drop, and mechanical
operation. Dry cake will be dumped, sluiced away with
the ultrafiltration concentrate, and sent to the evaporator,
A bypass line to the bodying tank on the downstream side
of the leaf filters is used to apply the precoat. Specif-
ically, after a cake has been dumped, pine caustic ex-
traction filtrate, with filter aid added, is circulated
through the leaf filters and returned to the bodying tank.
When an adequate precoat has built up, normal flow into
the membrane system is resumed.
The feed from the leaf filters,is further treated in back-
washable scavenging filters. These polishing filters re-
move any filter aid or other large particles which pass
through the leaf filters. The solids' discharge from the
scavenger filters (backwash) is mixed with the feed pine
caustic extraction filtrate for additional solids removal
in the leaf filters.
The filtered, neutralized feed flows to a surge tank
prior to treatment in the ultrafiltration system. This
surge capacity allows for short-term interruption in
the feed flow or filter operation, without causing an
immediate membrane system shutdown. Feed from the surge
tank is pumped through a heat exchanger and cooled
to 100°F. Flow is then through a series of four mem-
brane stages. For the low-flow membranes, Stage 1 can
be operated on a once-through basis; that is, it is
not necessary to recirculate concentrate within the
Stage 1 flow loop. The concentrate from Stage 4 is
used to sluice the filter cake, and flows to the evap-
orator for further concentration. The permeate from
Stage 1, used for backflushing the membranes, flows
to a water storage tank. The permeates from Stages 2,
3, and 4 are collected in a sump, and pumped to the sewer
143
-------
by a permeate pump. The membrane stages will be set up
for reverse-flow flushing, an operation which has been
found to be effective in membrane cleanup. The Stage 1
permeate will be used for this purpose. In addition,
if detergent cleaning of the membranes is required, the
Stage 1 permeate storage tank will be used to mix the
detergent solution.
The high-flow membrane system shown in Figure 57 is
identical to the low-flow system, except that recircula-
tion in Stage 1 is required to maintain high velocity
through the membranes.
Details of the equipment for these process flow schematics
are given below with cost estimates.
2. Equipment Description and Capital Cost Estimates
Five capital cost estimates for full-scale, battery-
limits plants are presented. These cases have been
prepared to provide some insight into the relative cost
sensitivity of the various parts of the process system.
The significant process parameters used for each case
are shown in Table 16.
The equipment cost estimates presented are based on
budgetary estimates obtained from potential vendors.
In the following. Case 1 will be described in detail;
the other cases will be presented as modifications of
Case 1.
Membrane Module Alternatives—Different Flow Channel
Spacers" Spiral wound membranes will be used. These
membrane cartridges are currently available from Gulf
Environmental Systems with either the standard mesh
spacers (60 ft2/cartridge @ $1.50/ft2) or corrugated
spacers (40 ft2/cartridge @ $2.25/ft2). In the pilot
plant operation, as described above, most of the mem-
brane cartridges had the standard mesh spacers. In
addition to the mechanical failures encountered with.
the T. J. Engineering cartridges, all cartridges with
mesh spacers were difficult to clean. In the early
part of the pilot plant operation, an inability to
thoroughly clean the membrane cartridges resulted in
increased pressure drop across the cartridges with time,
as well as reduced capacity. It was subsequently ob-
144
-------
TABLE 16
CASES FOR CAPITAL COST ESTIMATES
PINE CAUSTIC EXTRACTION FILTRATE
Pine Caustic
Extraction
Filtrate Flow
Case (x 106 gpd)
Membrane Modules circulation
Low Flow High Flow Membrane Rate
(corrugated Flux, gpm/membrane
(mesh spacer) spacer) aal/dav-£t2 cartridge .
Membrane Prefiltration operating
Cost Precoat Backwashable Screen Manpower ,
$/sq ft Filter Depth Filter Filter No. of Men
18
$1.50
18
15
$2.25
18
15
$2.25
18
$1.50
18
$1.50
-------
served that improved prefiltration and strict adherence
to prescribed cleaning procedures alleviated the prob-
lems associated with mesh-spacer cartridges. That is,
a "steady state" operation was achieved. Specifically,
over long periods of operation it was possible to re-
turn repeatedly to a base-line membrane flux and car-
tridge pressure drop after cleanup.
Experiments with membrane cartridges with corrugated
spacers suggest that it will be substantially easier to
clean these cartridges, possibly by water flushing alone,
In particular, reverse-flow flushing is anticipated to
be substantially more effective for corrugated-spacer
cartridges that for mesh-spacer cartridges. Thus, cor-
rugated-spacer cartridges offer several possible advan-
tages. These are:
—maintenance of a higher average flux (reduced
membrane area requirement);
—facilitated cleanup, reducing down-time for
cleanup, and possibly eliminating the need for
detergent cleaning;
—longer membrane cartridge life; and
--possible operation without precoat filtration.
These advantages can be of substantial importance in
terms of process costs. It is to be noted, however,
that these factors have not been demonstrated to date.
The cost figures for the two different membrane car-
tridges given above are identical, i.e. $90/cartridge.
These are the costs currently quoted by Gulf Environ-
mental Systems. It is not clear why costs should be
independent of the square footage of membrane in each
cartridge. On the contrary, for large-scale production
costs should be proportional more to membrane area, and
not to the number of cartridges. For present estimates,
however, the Gulf price structure will be used, al-
though it will tend to give a conservative picture for
corrugated-spacer membrane cartridges. The two cases
are summarized below.
2
1. $1.50/ft , for membrane cartridges with mesh
spacers. A circulation rate through these
cartridges of 8 gpm is assumed to be suffi-
cient to minimize membrane fouling and pre-
146
-------
vent extensive solids collection within the
cartridges. These membranes will be assumed
to have a three-year life. Precoat filtra-
tion is required.
2
2. $2.25/ft , for membrane cartridges with cor-
rugated spacers. A circulation rate of 15 gpm
through the membrane cartridges is assumed to
be sufficient to maintain cleanliness and
avoid cartridge plugging. Membrane cartridge
life is assumed to be three years. Precoat
filtration may not be required.
The relatively high flows through the membrane cartridges
have been selected on the basis of the pilot plant opera-
tion. For shells (with mesh-spacer cartridges) which
were operated at high feed flow, membrane fouling and
particulate collection within the cartridges was greatly
reduced. The three-year life assumed is an important,
but not critical, factor. For example, a two-year mem-
brane life would add about 4C/1000 gal. to the plant
operating costs (detailed below).
For design purposes, it has been assumed that corrugated-
spacer cartridges can be obtained with 60 ft2 of membrane
area. The desirability of having a high square footage
per cartridge relates to mechanical problems encountered
in the pilot plant operation. Specifically, it is
highly desirable to minimize the total number of seals,
gaskets, and glue seams in the system since these are
the points of potential operating problems.
The flow characteristics of the two different membrane
cartridges are given in Table 17.
TABLE 17
CHARACTERISTICS OF MEMBRANE CARTRIDGES
Spacer Mesh Corrugated
Feed Pressure 120 psi 120 psi
Outlet Pressure 80-90 psi 80-90 psi
Circulation Rate 8 gpm 15 gpm
Pressure Drop 8 psi/cartridge 8 psi/cartridge
Maximum Number of
Cartridges in Series 4 4
147
-------
The operating pressures chosen are 120 psig at the inlet
to the cartridges and 80-90 psig at the outlet from the
cartridges. At the circulation rates given, pressure
drop per cartridge will be about 8 psi. Correspondingly,
the maximum number of cartridges which can be assembled
in a series configuration is four. Thus, four cartridges
will be installed in a single housing, and this will be
denoted as a "membrane module".
In any single stage, capacity will be fixed by the total
membrane area. Since more than four membrane cartridges
will be required, it is necessary to pipe membrane
modules in parallel. The number of parallel modules is
determined by the total area requirement.
The costs of membrane modules for cartridges with mesh
and corrugated spacers are given in Table 18.
TABLE 18
CHARACTERISTICS AND COSTS OF MEMBRANE MODULES
Spacers Mesh Corrugated
Circulation Rate 8 gpm ,., 15 gpm
Area 240 ft 240 ft2
Bare Membrane Cost $360 $540
Housing and Installa-
tion $390 $390
TOTAL COST $750 $930
The modules with mesh spacers will cost $750 each; and
those with corrugated spacers, $930 each. The cost
difference is due solely to the greater bare membrane
area cost for corrugated-spacer spirals.
Installation includes all racks, internal piping,
shipping, and on-site direct labor. Excluded are costs
for pumps, the building and foundations, engineering
design, materials procurement, construction supervision,
and startup.
Membrane cartridges with corrugated spacers will prob-
ably be somewhat larger physically than membrane car-
148
-------
•bridges with mesh spacers, and the housings could cost
more. However, for present purposes, the housing and
installation costs have been assumed to be the same.
a. Details of Case__!_ Design
(i) Design Bases. The Case 1 design bases are as follows:
—2 x 10 gpd pine caustic extraction filtrate
processed;
—mesh-spacer membrane cartridges used;
—precoat filtration needed; and ^
—membrane flux of 18 gal./day-ft .
(ii) Feed Rate and Its Effect on Plant Size. It has been
assumed that the pine caustic extraction filtrate flow
is constant at 2 x 10° gpd (1,390 gal./min). However,
membrane flux depends on membrane cleanliness and
changes with time. In any case, the plant has been
sized for a flux of 18 gal./day-ft2, which is an average
flux over the operating period between cleaning cycles.
Flux varies from 25-30 gal./day-ft2 for clean membranes
to 10-15 gal./day-ft2 for fouled membranes.
The simplest way to operate a system is to process feed
at a rate greater than 2 x 106 gpd immediately after
membrane cleaning, and then to reduce the feed flow in-
to the membrane plant as the membranes become fouled.
During the latter part of this operation it would be
necessary to accumulate surplus feed for use during
the next run immediately after cleaning. However
this requires supplying a large surge capacity for the
feed, which would be prohibitively expensive. There-
fore, operation will be such that a small fraction
(one-quarter or less) of the membrane plant will be
flushed and cleaned at any single time. That is, the
flushing/cleaning sequence will be cycled through
isolatable sections of the ultrafiltration plant. This
will increase piping and operating costs somewhat, but
a net benefit will accrue due to the elimination of a
requirement for a large feed surge capacity.
It is further assumed that each segment of the plant
can be cleaned in one hour per day. Consequently, the
operating time available is 23 hrs/day.
149
-------
Excess membrane area will be required to produce the
permeate used for membrane flushing. Since the plant
volume is about 10,000 gals., conservatively not more
than 100,000 gpd will be required for once-a-day flush-
ing and cleaning. This will increase the membrane plant
area requirement (capacity) by 5% or less.
Finally, some surge capacity must be available, and a
one-hour surge (80,000 gals.) will be provided.
• (iii) Plant Size. The ultrafiltration system must process
2.1 x 106 gals, in 23 hrs, at an average flux of 18 gal./
day-ft2. The permeate rate is essentially the same,
since more than 98.5% of the feed must pass through
the membranes. Thus, the calculation of the membrane
area required is:
Membrane area =2.1 x 106 x 24/23 x i = 122,000 ft2.
J.C
The minimum size of the pretreatment section of the
process, including the filters, is 2.1 x 106 gals, within
23 hrs, or 1,500 gpm. A slightly larger filtration sys-
tem will be chosen to allow for cleaning the filtration
section and emergencies. Specifically, the filtration
section will be designed to handle 2.1 x 106 gals, with-
in 22 hrs, or 1,600 gpm.
(iv) Neutralization and Filtration Sections. The feed
will be neutralized with sulfuric acid before filtering.
Based on pilot plant experience, the feed will be
filtered first in a precoat filter, adding filter aid
to body the liquid at 50-100 ppm. This will be fol-
lowed by a scavenging filter. The filter aid tenta-
tively will be wood flour. Costs for equipment in the
neutralization and filtration sections are given below.
Cost
Acid store - 500 gals., assumed to be
available. n/c
Acid pump and piping - The probable acid
use rate is 8000 Ib/day, about 0.4 gpm.
On-off control through a low pressure
diaphragm metering pump (^1/2 hp) and
iron piping. $2,000
150
-------
Cost
pH measurement and control - In line pH
electrodes with a recorder, on-off control
and alarm switches, installed. $6,000
Filter aid solids handling - Filter aid
will be received in bags. Daily usage
about 1,000 Ib/day (about 70 ft3/day).
Hopper and controlled rate conveyor. $3,000
Bodying tank - 600 gal. carbon steel.
5 ft diameter x 4 ft high, with 1/4 hp
agitator. $1,600
Bodying pump - 1 gpm, 1/2 hp. $ 500
Precoat filter - Tests at Sparkler Manu-
facturing indicate that an achievable
filtration rate is 1.5 gpm/ft2. For con-
servative design purposes, 1 gpm/ft2 is
assumed and 1,600 ft-2 filter area will be
installed. Two units, Sparkler MCRO-
1000-3, each having 1000 ft2, vertical
leaf, horizontal tank filters for dry
case discharge cost $35,100. Scaled to
1,600 ft2. $30,000
Piping and valving - for the precoat
filters.$ 7,000
Scavenging filters - 10 units Velmac
backflushable felt-type filters, each
13" diameter and 62" high, rated 100-
200 gpm. $10,000
Piping and valving - for scavenging
filters. $10,000
In-line mixing section - $'1,500
Instruments and controls - other than
for pH, for measuring pressure and
differential pressure and automating
filter flushing. $16,000
Filter pump - Size 8 x 10-13 to pump
1,600 gpm into 94 feet (40 psi), 60 hp.
Operated at 62% efficiency. In iron.
Driven and started, not installed or
connected. $ 3,200
151
-------
Cost
Back washing and other filter cake re-
moval - The scavenging filters will be
back washed into the feed. The precoat
filters will be dumped (dry), probably
2 or 3 times each day. About 100 ftV
day of cake will be discharged. This cake
will be mixed with 20,000-30,000 gpd of
concentrate and conveyed either to the
evaporator (for pine caustic extraction
filtrate) or to the black liquor plant
(for decker effluent). $ 800
Slurry pump - to remove the mixture of
filter cake and concentrate, 14 gpm. $ 400
SUBTOTAL FOR EQUIPMENT $85,000
Transportation, based on non-instrument,
heavy pieces, totalling about $63,000
and coming from 800 to 1,000 miles. $ 2,500
Installation, direct labor only, in-
cluding rigging, pipe fitting and
electrical connections of (approx):
2 pieces filter hydraulic motors, 1/2 hp
acid pump, 1/2 hp
dry feeder, 1/2 hp
tank stirrer, 1/4 hp
filter pump, 60 hp
instruments and controls, 20 amp max.
filter cake mixer, 1/2 hp
slurry pump, 1 hp
$12,600
TOTAL PRETREATMENT SECTION $100,100
(Engineering design, equipment pro-
curement, supervision and startup
not included)
T
(v) Ultrafiltration Section. 122,000 ft"" of membrane
area will be used. This will be incorporated into four
stages (Figure 56); details are given in Table 19.
Each membrane module requires 8 gpm inlet feed flow.
Two-hundred parallel modules have been provided in
152
-------
Stage
Number
TABLE 19
ULTRAFILTRATION SECTION DESIGN
DETAILS AND COSTS—CASE 1
Number of
Modules
200
160
90
60
Circulation
Rates (gpm)
Pump Differ-
ential (psi)
Utilized
Horse Power
1600
(feed)
120
(feed)
175
1280
40
50
720
40
25
480
40
20
Efficiency for
Estimating h.p,
64%
60%
68%
56%
Approximate
Pump Size
6x8-16 6x8-13 6x8-11
4x6
Cost Estimate
(Driven, Started and
Installed
Electrical Connection
$5,200 $3,000 $2,600 $1,400
Flush Pump:
Addi tion a1 Equipment
1600 gpm, 40 psi, 60 HP
Permeate Pump: same as Flush Pump
Water Store:
20,000 gallons, delivered and
installed
Flush and drain piping, installed
$ 3,200,
$ 3,200
$10,000,
$ 9,000,
153
-------
Stage 1, requiring 1,600 gpm. The feed rate of 1,520 gpm
is adequate for once-through operation and circulation
in Stage 1 is not required. The feed pump delivers 120
psig, and a Stage 1 circulation pump is not provided.
The feed to Stage 2 is 920 gpm, which can feed 120 parallel
membrane modules on a once-through basis. Since a greater
number of modules is used, recirculation is necessary.
Perhaps of greater importance is the difficulty of op-
erating two, once-through stages in series. Specifically,
an imbalance would occur since the second pump would up-
set the flow and pressures in the first stage.
Stages 3 and 4 would require circulation in any event
since the net feed flow to each of the stage will be
relatively low.
(vi) Summary of Ultrafiltration Section Costs
Cost
Membranes (510 x $750) $382,500
(includes $183,600 for membrane
cartridges)
Pumps 18,600
Tanks and piping 19,000
Electrical connections for a total of
350 hp to 6 pumps (transformer ex- 9,000
eluded)
TOTAL $429,100
(Engineering design, equipment procure-
ment, supervision of construction, and
startup not included)
(vii) Other Equipment. Other equipment costs are given
belovr.
Cost
Clean feed surge tank - will have about
one hour's surge capacity (80,000 gals.),
installed. $ 30,000
Heat exchanger (counter current) - to
cool 1,600 gpm from 130°F to 100°F using
1,600 gpm of water which is heated from
70°F to 100°F. Installed cost.
154
-------
Cost
24 million BTU/hr
150 BTU/(hr)(ft2)(°F)
30°F AT ,,
5,350 ft $ 75,000
Control panel - $ 10,000
Building - about 40 ft x 50 ft con-
crete pad with sumps and pump pads of
cast concrete. A prefabricated type
of building 30 ft to cross beams.
Very limited heating as the system
releases substantial heat (electrical
substation, if required, is not in-
cluded) . $ 48,000
Evaporator - based on direct contact
between flame and solution to con-
centrate 20,000-30,000 gpd of 20%
solids to 50% solids. Assumed to
be fueled by burning hydrogen avail-
able from Electrochemical Plant. The
50% solids stream will be disposed of
with the primary sludge from the
waste treatment plant. Incineration
is preferable to evaporation and may
be possible. However, preliminary
capital costs provided by the John
Zink Company are high compared to
previous estimates. Furthermore,
poor burning efficiency will re-
quire additional fuel consumption.
Therefore, for present purposes,
incineration is questionable.
Evaporator, installed, including
hydrogen handling system. $125,000
Total Other Equipment - $288,000
Engineering charges - including all
design and drafting, equipment procure-
ment, construction supervision, and
shift engineers for startup and super-
vision for the first twelve months of
operation. $300,000
155
-------
(viii) Capital Cost Summary, Case 1.
Cost
Filtration and neutralization section $100,000
Ultrafiltration section (includes $429,000
membranes @$183,000)
Other equipment and building $288,000
Design, Administrative, and Supervision $300,000
$1,117,200
Contingency, @ 10% 111,700
TOTAL $1,228,900
Total utilized energy is 374 hp excluding the flushing
pump which is run only when a circulation pump is stopped.
b. Details of Case 2 Design
(i) Design Bases. The design bases for Case 2 are given
below:
—2 x 10 gpd pine caustic extraction filtrate
processed;
—corrugated-spacer membrane cartridges used;
—precoat filtration needed; and 2
—membrane flux of 18 gal./day-ft .
(ii) Ultrafiltration System Costs. For this case costs
increase due to (1) increased membrane costs, and {2} in-
creased feed circulation rate through the membrane
spirals. The costs for the neutralization and filtra-
tion section, other equipment and building, and design
and administrative costs remain unchanged. Only the
Ultrafiltration section costs must be changed.
As before, 510 membrane modules are required, but
each is more expensive. The net feed flow to the mem-
brane plant is 1,520 gpm. Since each module requires
15 gpm, the first stage can be operated only once-
through if it contains 102 modules. Instead of limiting
the number of modules in Stage 1 to this number, re-
circulation will be employed. This requires an addi-
tional circulation pump for Stage 1.
156
-------
Ultrafiltration design factors and modified costs are
given in Table 20. The appropriate flow schematic is
given in Figure 57.
Capital Cost Summary, Case 2.
Filtration and neutralization section
(from Case 1)
Ultrafiltration section (includes
membranes §$275,400)
Other equipment and building
(from Case 1)
Design, Administration, and Supervision
(from Case 1)
$1,217,100
Contingency, @ 10% 121,700
TOTAL $1,338,800
Total utilized energy is 509 hp, excluding the flushing
pump which is run only when a circulation pump is stopped.
c. Case 3 Design
Design Bases. The design bases for Case 3 are given
below:
—2 x 10 gpd of pine caustic extraction filtrate
processed;
—corrugated-spacer membrane cartridges used;
—no precoat filtration required;2and
—membrane flux of 18 gal./day-ft .
This case is similar to Case 2, but since corrugated-
spacer membrane cartridges are used, it has been assumed
that precoat filtration is not required. Instead, a
simple screen, such as a Bauer Hydrosieve, will be
employed in place of the precoat and scavenger filters.
Feed neutralization and mixing is still required.
Cost savings are achieved from (1) elimination of ex-
pensive filtration equipment; and (2) the building
cost is reduced by 15% ($7200). Thus, the capital costs
for Case 3 will be $80,000 less than Case 2. The total
157
-------
TABLE 20
ULTRAFILTRATION SECTION DESIGN
DETAILS AND COSTS—CASE 2
Stage
Number
4
Number of
Modules
200 120
120
70
Circulation
Rates (gpm)
3000 1800 1800 1050
Pump Differ-
ential (psi)
40
40
40
40
Utilized
Horse Power
100
70
70
40
Efficiency for
Estimating h.p,
70% 60%
60% 61%
Approximate
Pump Size
8x10-13 6x8-13 6x8-13 6x8-11
Pump Cost (Driven,
Started and Installed,
but without Electrical
Connections)
$4000 $3,500 $3,500 $2,800
158
-------
TABLE 20
(continued)
ULTRAFILTRATION SECTION DESIGN
DETAILS AND COSTS—CASE 2
Additional Equipment
Flush Pump: 1800 gpm, 40 psi, 60 HP. $ 3,500,
(note that Stage 1 must be
flushed in two halves to
obtain adequate flow)
Permeate Pump: 1600 gpm, 40 psi, 60 HP. $ 3,200,
Feed Pump: 1520 gpm, to 90 psi, 125 HP. $ 4,700,
(64% efficiency)
Water Store: 20,000 gallons, installed $10,000,
Flush and drain piping installed $ 9,000,
SUMMARY OF ULTRAFILTRATION SECTION COSTS
Membranes, (510 x $930)
(includes $275,400 for
membrane cartridges) $474,300
Pumps 25,200
Tanks and Piping 19,000
Electrical connections for a total of
485 HP to 7 pumps (transformer excluded) 10,500
TOTAL $529,000
159
-------
capital cost is then $1,258,000. This cost includes
membranes at $275,400. The total energy used is 447 hp.
d. Case 4 Design
Design Bases. The design bases for Case 4 are the same
as those presented for Case 1 with the exception that
a depth filter system is used in place of the precoat
filter system. No changes in capital requirements
have been included. The capital estimates from this
case are used to display the operating cost changes.
The capital cost for Case 4 is $1,228,900. Total
utilized energy is 374 hp.
e. Case 5 Design
Design Bases. Design bases for Case 5 are:
—1 x 10 gpd pine caustic extraction filtrate
processed;
—mesh-spacer membrane cartridges used;
—depth filtration system used; and
—membrane flux of 18 gal./day-ft2.
This case is similar to Case 1, but the system is designed
to handle 1 x 106 gpd rather than Case 1 design of
2 x 106 gpd. In addition, the system design uses a depth
filter in place of the precoat filter system used in
Case 1. The capital cost estimate for Case 5 has been
developed from the Case 1 capital estimate by scaling
as indicated below:
CAPITAL COST SUMMARY
Case 1 Scale Case 5
Capital Cost Factor Capital Cost
Filtration and
neutralization 1
section $ 100,000 $ 70,800
Ultrafiltration ,
section 429,000 - 225,800
Other Equipment 240,000 1^ 120,000
160
-------
CAPITAL COST SUMMARY (continued)
Case 1 Scale Case 5
Capital Cost Factor Capital Cost
Building 48,000 -^ 30,000
J.. b
Design, Administra-
tion and Supervision 300,000 250,000
Total $700,600
Contingency, @ 10% 70,000
TOTAL $770,600
Total utilized energy is 187 hp, exluding the flushing
pump which is run only when a circulation pump is stopped.
Summary of Capital Cost Estimates
A summary of the capital cost projections for the five
case studies for battery-limit plants to treat pine
caustic extraction filtrate is presented in Table 21.
Because a plant for this service would be the first of
its kind, conservative estimates have been used in pre-
paration of the projections. This is especially so in
the estimates for design, administration and supervision.
It is felt that the startup and first year supervision
should be amply provided for. As will be noted, in
Case 5 this accounts for about one-third of the capital
costs.
These capital values are used in the following sections
in the development of the projected plant operating
costs.
Operating Cost Summary
Projected operating costs for Cases 1 through 5 are
presented in Tables 22, 23, 24, 25, and 26 and are
summarized in Table 27.
Bases for Estimates. The estimates have been developed
on the basis of a 365-day operating year. The costs
are incremental operating costs for treating pine caustic
extraction filtrate in an existing pulp mill complex.
161
-------
TABLE 21
SUMMARY OF INSTALLED CAPITAL COST ESTIMATES
PLANT TO TREAT PINE CAUSTIC EXTRACTION FILTRATE
Case No. 11211
Filtration &
Neutralization $ 100,000 $ 100,000 $ 35,000 $ 100,000 $ 70,800
Section
429,000 529,000 529,000 429,000 225,800
Other Equipment 240,000 240,000 240,000 240,000 120,000
Building 48,000 48,000 40,000 48,000 30,000
Design,
Administration 300,000 300,000 300,000 300,000 250,000
& Supervision
Subtotal 1,117,200 1,217,100 1,114,000 1,117,200 700,600
Contingency 111,700 121,700 114,000 111,700 70,000
Total $1,228,900 $1,338,800 $1,258,800 $1,228,900 $770,600
Membrane Costs
Included in $ 183,600 $ 275,400 $ 275,400 $ 183,600 $ 91,800
Capital
Utilized HP 374 509 447 374 187
162
-------
OJ
TABLE 22
Operating Costs for Treatment of Pine Caustic Extraction Filtrate, Case :
Quantity Unit Cost $/Day $/Year
-------
TABLE 23
Operating Costs for Treatment of Pine Caustic Extraction Filtrate; Case 2
Quantity Unit Cost $/Day $/Year C/M-gal Total
Statistical
of
Material
Acid. . . 8/000 Ib/day IC/lb
Filter Aid 1/000 Ib/day 3£/lb
Total Material
Conversion Exoense
Labor (including benefits) 5 Man-yrs $ll,800/yr
Repair and Maintenance
Material. . . . .$ 763.400 i na ATV
Electric Power .... 509 HP 0. 746
-------
Ui
TABLE 24
Operating Costs for Treatment of Pine Caustic Extraction Filtrate; Case 3
Quantity Unit Cost $/Day $/Year C/M-gal Total
Statistical
of
Material
Acid 8,(
Filter Aid
Total Material . . ,
Conversion Expense
300
Labor (including benefits)
Repair and Maintenance
Material $
Labor
Electric Power
Insurance and Taxes 1/2
682
0.5
447
Ib/day
4 Man-yrs
,400
Man-yrs
HP
x Maintenance
l$/lb
$ll/800/yr
1.5%/yr
$20,000/yr
0.746
-------
TABLE 25
Operating Costs for Treatment of Pin<
Quantity
Material
Acid 8,0
Filter Aid
Conversion Expense
Labor (including benefits)
Repair and Maintenance
Insurance and Taxes I/2
Total, excluding Deprec
Depreciation
Other Facilities . . $1,
Total Incremental Cost
00. Ib/day
3 Caustic Ext:
Unit Cost
l«/lb
4 Man-yrs $ll,800/yr
744,300 1.5%/yr
0.5 Man-yrs $20,000/yr
374 HP 0.746$/HP-hr
x Maintenance Material
i at ion
183,600 3-year life
044,300 15-year life
ractio:
$/Da
$ 80.
$ 80.
$129.
30.
27.
66.
15.
$269.
$167.
190.
.$627.
$707.
n ri
£
00
00
32
58
40
96
29
55
67
74
9f>
96
.itrate: c<
$/Year
$ 29
$ 47
11
10
24
5
$ 98
61
69
$229
$258
,200
,200
,165
,000
,400
,580
,386
,200
,620
r2p5
,405
ase 4
<=/M-gal
4
6
1
1
3
0
13
8
3;
35
.0
.47
.52
.37
.35
.76
.47
.38
,39
.39
% of
Total
11.
18.
4.
3.
9.
2.
38.
23.
27.
8?,
100.
3
3
3
9
5
1
1
7
0
7
0
Statistical
Effluent Treated (x 106 gallons) 2 730
-------
TABLE 26
Operating Costs for Treatment of Pine Caustic Extraction Filtrate: Case 5 % -^
Quantity Unit Cost $/Day $Aear C/M-gal Total
Material
Acid. . . . ....... 4,000 Ib/day l£/lb $ 40.00
Filter Aid .
Total Material $ 40.00 $ 14,600 4.00 8.7
Conversion Expense
Labor (including benefits) 4 Man-yrs $ll,800/yr $129.32 $ 47,200 12.93 28.2
Repair and Maintenance
Material $ 364,800 1.5%/yr 15.00 54,720 1.5 3.3
Labor . 0.5 Man-yrs $20,000/yr 27.40 10,000 2.7 6.0
Electric Power 187 HP 0.746$/HP-hr 33.48 12,220 3.35 7.3
Insurance and Taxes 1/2 x Maintenance Material 7.50 27,360 0.75 1.6
Total, excluding Depreciation $212.70 $ 77,636 21.27 46.5
Depreciation
Membranes $ 91,880 3-year life 83.84 30,600 8.38 18.3
Other Facilities $ 664,800 15-year life 121.42 44,320 12.14 26.5
Total Conversion Expense $417.94 $152,548 41.79 91.3
Total Incremental Cost $457.94 $167,100 45.79 100-.0
Statistical
Effluent Treated (x 106 gallons) 1 365
-------
TABLE 27
DAILY INCREMENTAL OPERATING COSTS
TREATMENT OF PINE CAUSTIC EXTRACTION FILTRATE
Case No.
Pine Caustic
Extraction
Filtrate 22221
Treated
x 106 gpd
Capital
Investment $1,228/900 $1,338,800 $1,258,800 $1,228,900 $770,600
Daily Operating Costs - Dollars/day
Materials $110 $110 $ 80 $ 80 $ 40
Conversion
Expense less 301.87 327.23 278.81 269.55 212.70
Depreciation
Membrane 167.67 251.50 251.50 167.67 83.84
Depreciation
Other
Facilities 190.74 194.23 179.43 190.74 121.42
Depreciation
Total $770.28 $882.96 $789.74 $707.96 $457.94
Costs
$/1000 gal. 38.50$ 44.14* 39.48$ 35.39$ 45.79$
Treated
Costs
$/Ton $ 0.962 $ 1.103 $ 0.987 $ 0.085 $ 0.572
Pine Pulp
168
-------
These costs are presented as representing steady-state
operating costs for an ultrafiltration plant using
cellulose acetate membranes from the second year of
operation and on. As indicated previously, in prepara-
tion of the capital costs, since this plant would be
the first of its kind and scale, due allowance has been
made for special startup and supervisory costs for the
first year of operation.
The bases for estimating costs were as follows:
Materials; Acid costs and filter aid costs are
present plant or vendor estimates.
Labor; Labor costs have been estimated on the
basis of 1973 estimated base salaries plus 30%
added to cover benefits. It is assumed the
plant can be operated with 4 or 5 man-years per
year of operator time for the first year or so,
and that this manpower level could be substantially
reduced as operating experience is gained. For
these estimates, however, 4 or 5 operators are
used.
Electric power; Electric power costs are estimated
on the basis of power costs of IC/kw-hr.
Depreciation; The membrane life is taken as 3 years
and the costs are depreciated over this time period
on a straight-line basis. The remainder of the
plant is depreciated on a straight-line basis over
a 15-year period.
A summary of the daily operating costs in presented in
Table 27.
A comparison of Cases 1 and 2 shows that an operating
cost increase of approximately 6C/1000 gals, is in-
curred if corrugated-spacer membrane cartridges are
used instead of the mesh-spacer cartridges. Examination
of Cases 1 and 3 shows that there is only about a 1
-------
Case 4 considers the use of mesh-spacer cartridges but
installing backwashing filters with automatic cleaning
cycles. For this case, filter aid would not be re-
quired and four men would operate the system. The
capital cost of these filters has been assumed to be
equal to the capital cost of precoat filters, but
operating costs are reduced by elimination of filter aid
and a labor component. Some pilot plant experience
has been obtained with such a filter—the Hydromation
granuar PVC backwashing filter. The model tested fil-
tered adquately but was not used extensively in the
pilot plant since its capacity was too small. Case 4
shows operating costs for a system with a Hydromation-
type filter and with mesh-spacer cartridges. It is
seen that a savings of approximately 3C/1000 gals, over
Case 1 can be obtained.
Both the capital and operating costs for an ultrafiltra-
tion system are strongly dependent on the volume pro-
cessed. It has been observed in the pilot plant program
that the concentration of color bodies and organics in
the pine caustic extraction filtrate varies substantially
from day-to-day, and even hour-to-hour. In principle,
it should be possible to control the pine caustic
extraction filtrate flow rate such that the concentra-
tion of contaminants is at the maximum allowable level
in terms of obtaining bleached pulp of acceptable
quality. Through this means it is thought that the flow
of pine caustic extraction filtrate can be substantially
reduced from 2 x 106 gpd, possibly to as low a volume
as 1 x 10° gpd. By this means, the cost of waste treat-
ment per unit of pulp produced can be substantially
reduced, even though the treatment cost per unit of
effluent may increase. In Case 5, it has been assumed
that the total flow of pine caustic extraction filtrate
can be reduced to 1 x 106 gpd. Mesh spacers are used,
and a backwashing filter is employed. Capital costs
for this plant have been scaled from the 2 x 10^ gpd
plant using the scaling factors presented on page 148.
Case 5 is to be compared to Case 4. Although the cost
per unit effluent is higher in Case 5, the cost per
unit pulp produced is substantially lower.
Some Economic Evaluations
The capital cost estimatesgindicate that an ultrafiltra-
tion plant to treat 2 x 10 gpd of pine caustic extraction
filtrate would have an installed cost of $1.2-$1.4 million
170
-------
based on present equipment and labor costs. A similar
plant to treat 1 x 10° gpd would cost almost 750 to
800 thousand dollars.
The operating cost estimates indicate that the most
significant cost items are membrane depreciation, other
facilities depreciation, operating labor and materials —
which account for 70% to 90% of the daily costs in the
cases presented.
The most sensitive single cost factor displayed is the
total flow of caustic extraction filtrate to be treated.
If the flow of this material can be limited by judi-
cious bleachery process flow control to 1 x 10^ gpd,
the capital cost of the membranes would be reduced to
one-half, the total plant cost would be almost 60% and
the daily operating costs would decrease from a level
of $700 - $800 per day to about $450 per day.
Membrane flux is a second important parameter in process
costs. The cases examined above have not considered
variation in membrane flux, but have assumed a rate of
18 gal./day-ft2 . For a given capacity, plant cost will
vary almost inversely with membrane flux. Operating
costs will be greatly affected due to both membrane
replacement cost and capital depreciation. The value
chosen in the design calculations (18 gal./day-ft2) is
higher than that observed in most of the pilot plant
tests. However, since some of the membrane cartridges
exhibited this ultrafiltration rate, it is assumed to
be an achievable value. In addition, future advances
in membrane technology will undoubtedly provide
higher-flux membranes.
Membrane life is very important since membrane replace-
ment is a substantial operating cost factor. The 3-yr
life which has been chosen for design calculations is
longer than any life demonstrated to date with pine
caustic extraction filtrate. This life, however, is
not unreasonable for brackish water desalination
applications, and is considered realistic. It may be
noted that a 2-yr membrane life (mesh-spacer cartridges)
would increase operating costs by approximately
4C/1000 gals. Again, with new developments in mem-
brane technology, longer life should be obtainable.
171
-------
Other anticipated membrane improvements would permit
operation at alkaline pH's and at high temperatures.
Three such membranes are in an advanced stage of
development. These are the NS-1 membrane, developed
by the North Star Research and Development Institute,
Minneapolis, Minnesota (29) ,the polybenzimidazole
membrane, developed by the Celanese Research Company,
Summit, New Jersey (3P) ,and the dynamic membranes
developed by the Oak Ridge Natural Laboratory (16).Not
only can acid costs and cooling costs be eliminated,
but membrane flux should also increase. This is due to
two factors. First, membrane flux increases with
temperature; second, in the treatment of pulp mill
effluents membrane flux generally decreases when the
feed pH is changed. This is primarily a fouling
phenomenon, with non-neutralized feeds exhibiting a
substantially reduced fouling rate. Thus, new mem-
branes will not only reduce pre-treatment costs, but
can also reduce capital and operating cost since a
higher membrane flux can be realized.
When using cellulose acetate membranes, for which pre-
treatment is required, some additional cost savings may
be realized. First, the acid cost for neutralization
may be reduced or eliminated by neutralizing with waste
acid available within the mill. A second materials
cost which may be eliminated is that for filter aid.
If corrugated-spacer cartridges can be used without
precoat filtration, or if a backwashable depth filter
is adequate, filter aid would not be required.
Another major operating cost is labor. It has been
assumed that four or five man-years/year would be re-
quired to run the combined pre-treatment and ultra-
filtration systems. If the system is simplified by
modification or elimination of the pre-treatment system,
this operating labor load could be reduced substan-
tially. Furthermore, as operating experience is gained,
the labor requirement should decrease.
Thus, several potential savings in both capital and
operating costs can be obtained through new develop-
ments in membrane technology, as well as by obtaining
operating experience for a full-scale plant.
172
-------
B. DECKER EFFLUENTS
1. Flow Schematic
Figure 4 (page 21) has presented a generalized flow
schematic showing integration of an ultrafiltration
system in the decker and black liquor systems. A
generalized flow schematic, more specific to the ultra-
filtration process, is shown in Figure 58. The
effluent from the decker undergoes pre-treatment
(neutralization and filtration), is cooled by exchange
with incoming fresh water, and flows into the ultra-
filtration system. The permeate from the ultrafiltra-
tion system is recycled to the final stage of pulp
washing, or is used for other pulp mill water require-
ments. The concentrate from the ultrafiltration section
is used to sluice the filter cake, and this suspension
is processed in the black liquor recovery system.
More detailed flow schematics for the decker effluent
treatment system are given in Figures 56 and 57
(pages 129 and 130).These drawings, previously presented
for the pine caustic extraction filtrate system, are
applicable to the treatment system for decker effluents.
Only the means of disposal of the permeate and concen-
trate fractions from the ultrafiltration system are
different.
In preparation of the five cases of projected economics
presented, the process parameters have been varied as
shown in Table 28.
2. Capital Cost Projections
The five capital cost projections for installed battery-
limit plants are presented in Table 29. The vendor
equipment budgetary estimates used in preparation of the
pine caustic extraction filtrate capital projections
have been used as the basis for the estimates and are
not repeated here.
Cases 6 and 9 differ from the capital estimate for
Case 1 (pine caustic extraction filtrate) only by the
cost of the evaporator system used in Case 1. Details
for the costs are given under the Case 1 discussion.
173
-------
Other Pulp
Mill Water .
Requirements
Fresh Water
1
Pulp Washers
Black
Liquor
Plant
Decker
Dilute
Black
Liquor
Present
Effluent
T
Filter
Cake
Fresh Water
Concentrate
UF
Treatment
Permeate
FIGUP^ 58: SIMPLIFIED FLOW SCHEMATIC: TREATMENT OF DECKER EFFLUENTS
-------
TABLE 28
CASES FOR CAPITAL COST ESTIMATES
DECKER EFFLUEHTS
Decker Membrane Modules Circulation Prefiltration
Effluent Low Flow High Flow Membrane Rate Membrane Operating
Flow Rate (corrugated Flux gpm/roembrane Cost Precoat Backwashable Screen Manpower
Case (x 106 gpd) (mesh spacer) spacer) gal./day-ft^ cartridge S/sq ft Filter Depth Filter Filter No. of Men
6 2 x 18 8 $1.50 x 5
72 X 18 15 $2.25 x «
M
«J
Ui
8 1 x 18 8 $1.50 x 5
9 2 x 18 8 $1.50 x 4
10 1 X 18 8 $1.50 x 4
-------
TABLE 29
INSTALLED CAPITAL ESTIMATES - PLANT TO TREAT DECKER EFFLUENTS
Case No. 6a 7_b Bc £a 10_
Prefiltration &
Neutralization $100,000 $100,000 $100,000 $100,000
Section -65,000
=35,000 =70,800 $ 70,800
Ultrafiltration 429,100 529,000 429,100 429,100 225,800
Membrane 1.9
=225,800
Other Equipment 115,000 115,000 115,000/2 115,000 53,000
=5., 000
Building 48,000 40,800 48,000 48,000 30,000
1.6
Design, =30<000
Administration 300,000 300,000 250,000 300,000 250,000
& Supervision ' - - - - -
Subtotal 992,200 1,019,800 629,600 992,200 629,600
Contingency 98,200 101,000 60,000 98,200 60,000
Total $1,090,400 $1,120,800 $689,600 $1,090,400 $689,600
Membrane Costs
Included Tn183,600 275,400 183,600 183,600 91,800
Capital =91,800
Utilized HP 374 447 187 374 187
NOTES;
a - costs taken from Case 1 and adjusted
b - costs taken from Case 3 and adjusted
c - costs obtained by adjustment and scaling from Case 6
176
-------
Case 7 differs from Case 3 (pine caustic extraction
filtrate) only by the cost of the evaporator system.
Details are given under Case 3.
Case 8 and Case 10 are derived from Case 6 by the
scaling procedures used under the pine caustic extrac-
tion filtrate Case 5.
3. Estimated Operating Costs
Projected operating costs for each of the 5 cases are
presented in Tables 30, 31, 32, 33 and 34 . The
estimates are based on a 365-day operating year.
The projected operating costs as presented are incre-
mental operating costs for a plant for treating the
decker effluents and their reuse. These costs are
presented for steady state operations from the second
year of operation and on. In preparation of the
capital costs, since this plant would be the first of
its kind and scale, due allowance has been made for
special startup and supervisory costs for the first
year operations.
The same estimating bases have been used for the decker
effluent cases as were used in the pine caustic extrac-
tion filtrate cases. The life of the membranes is
taken as 3 years and the life of the other facilities
as 15 years for purposes of depreciation estimates.
Labor costs as presented include a 30% adder for
benefits.
The daily incremental operating costs for the five cases
are summarized in Table 35. These cases are evalua-
ted in greater detail below, which makes allowance for
process credits.
4. Potential Process Credits from Decker Effluent Recycle
The design of the ultrafiltration plant to treat decker
effluents is developed on the basis of splitting the
stream into a concentrate containing organic materials,
including color bodies, and a permeate stream of low
color containing the bulk of the dissolved materials,
primarily sodium sulfate. The low-volume concentrate
would contain 10-20% organic material, and would be
177
-------
-J
00
TABLE 30
Operating Costs for Treatment of Decker Effluent; Case 6 % of
Quantity Unit Cost $/Day $/Year C/M-gal Total
Material
Acid 8,000 Ib/day l«/lb $ 80.00
Filter Aid 1,000 Ib/day 3C/lb 30.00
Total Material $110.00 $ 40,150 5.5 14.9
Conversion Expense
Labor (including benefits) 5 Man-yrs $ll,800/yr $161.64 $ 59,000 8.08 21.9
Repair and Maintenance
Material $ 606,800 1.5%/vr
Electric Power 375 HP 0. 746£/HP-hr
Insurance and Taxes 1/2 Maintenance Material
Depreciation
Membranes $ 183,600 3-year life
Other Facilities $ 906,800 15-year life
Total Conversion Expense
Total Incremental Cost
24.94
27.40
66.96
12.26
.$293.20
$167.67
165.63
$626.50
$736.50
9,102
10,000
24,400
4,476
$107,017
61,200
60.453
$228,673
$268,823
1.25
1.37
3.35
.62
14.67
8.38
8.28
31.33
36.83
3.3
3.9
9.1
1.7
39.8
22.8
22.5
85.1
100.0
Statistical
Effluent Treated (x 106 gallons) 2 730.0
-------
TABLE 31
Operating Costs for Treatment of Decker Effluent: Case 7 % of
Quantity Unit Cost $/Day $/Year $/M-gal Total
Material
Acid ........... 8,000 Ib/day l$/lb $ 80.00
Filter Aid
Total Material $80.00 $ 29,200 4.0 10.6
Conversion Expense
Labor (including benefits) 4 Man-yrs $ll,800/yr $120.32 $ 47,200 6.47 17.1
Repair and Maintenance
Material $ 545,400 1.5%/yr 22.41 8,810 1.12 3.0
^> Labor 0.5 Man-yrs $20,000/yr 27.40 10,000 1.37 3.6
vo
Electric Power 447 HP 0.746C/HP-hr 80.03 29,210 4.00 10.6
Insurance and Taxes 1/2 x Maintenance Material 11.20 4,090 0.56 1.5
Total, excluding Depreciation $270.76 98,827 1.35 35.8
Depreciation
Membranes $ 275,000 3-year life $251.50 91,800 12.57 33.2
Other Facilities $ 845,400 ISryear life 154.41 56,360 7.72 20.4
Total Conversion Expense $676.67 $246,985 $33.81 89.4
Total Incremental Cost $756.67 $276,185 $37.81 100.0
Statistical
Effluent Treated (x 106 gallons) 2 730
-------
00
o
TABLE 32
Operating Costs for Treatment of Decker Effluent: Case 8
Quantity Unit Cost $/Day $/Year <=/M-gal
Material
Acid
Filter Aid
Total Material .
Conversion Expense
. .4.000 Ib/dav 16/lH
500 Ib/day 3£/lb
Labor (including benefits) 5 Man-yrs $ll,800/yr
Repair and Maintenance
Material s 2Q7."*nn i <;» An-
Labor . . . . . .
Electric Power
Insurance and Taxes
Total, excluding
Depreciation
Membranes ....
0* 5 Man— vrs $20 .000 A/r
187 HP 0.746C/HP-hr
1/2 x Maintenance Material
Depreciation
. . s qi . ann ^-v^ar i i f &
Other Facilities . .$ 597,300 15-year life
Total Conversion Expense
Total Incremental Cost
$ 40.00
15
$ 55 00
$161.64
12.22
27.40
33.48
6.11
$240.85
$ 83.84
109.10
.$433.38
.$488 38
$ 20,070
$ 59,000
4,450
10,000
12,220
2,230
$ 87,910
30,600
39,820
$158,333
$178,259
5.5
16.15
1.22
2.74
3.35
0.61
24.08
8.38
10.91
43.47
48.87
% of
Total
11.3
33.1
2.5
5.6
6.9
1.2
49.3
17.1
22.3
88.7
100.0
Statistical
Effluent Treated (x 10 6 gallons) 1 365
-------
TABLE 33
oo
jLiiwj.diidiv.aa. w^co. a. i, j.iivj \* wo 1.0 J.WJL j. j.cai,uicii u \JJ- L/cwrt
Quantity Unit Cost
Material
Acid 8,000 Ib/day l£/day
Filter Aid
Total Material
Conversion Expense
Labor (including benefits) 4 Man-yrs $ll,800/yr
Repair and Maintenance
Material & 606/800 1 5%/yr
Labor • ....... 0.5 Man— yrs $20,000/yr
Electric Power 374 HP 0. 746C/HP-hr
Insurance and Taxes 1/2 x Maintenance Material
Total, excluding Depreciation
Depreciation
Meirbranes $ 183,600 3-year life
Other Facilities . $ 906,800 15-year life,
C J. OJi. JL .LUC
$/Day
$ 80.00
.$ 80.00
$129.32
9494
27 40
66 96
12.26
.$260. 88
$167.67
165.63
.$594. 18
.$674. 18
a<- , v-aac y
$/Year
$ 29 .200
$ 47,200
9 102
10 000
24 400
4,476
$ 95,178
$ 61,200
60,453
$216 ,831
$246 ,031
C/M-gal
4 0
6.47
1 25
1 37
3 35
0.62
13.06
8 38
8.28
29. 72
33. 72
% of
Total
11 9
19.2
3 7
4 1
9 9
1.8
38.7
24 9
24.6
88. 1
100.0
Statistical
Effluent Treated (x 106 gallons) 2 730
-------
TABLE 34
Incremental Operating Costs for Treatment of Decker Effluent; Case 10 % of
Quantity Unit Cost $/Day $/Year C/M-gal Total
Material
Acid ........... 4,000 Ib/day l$/lb $ 40.00
Filter Aid
Total Material $ 40.00 $ 14,600 4.0 9.1
Conversion Expense
Labor (including benefits) 4 Man-yrs $ll,800/yr $129.32 $ 47,200 12.93 29.3
Repair and Maintenance
£ Material $ 297,,300 1.5%/yr 12.22 4,460 1.22 2.8
10 Labor. 0.5 Man-yrs $20,000/yr 27.40 10,000 2.74 6.2
Electric Power 187 HP 0.746«/HP-hr 33.48 12,220 3.35 7.6
Insurance and Taxes . . 1/2 x Maintenance Material 6.11 2 ,230 0.61 1. 4
Total, excluding Depreciation $208.53 $ 76,110 20.85 47.2
Depreciation ~~~~~
Membranes $ 91,800 3-year life $ 83.84 $ 30,600 8.38 19.0
Other Facilities . $ 597,300 15-year life $109.10 39,820 10.91 24.7
Total Conversion Expense $401.47 $146,530 40.14 90.9
Total Incremental Cost $441.45 $161,130 44.14 100.0
Statistical
Effluent Treated (x 106 gallons) 1 365
-------
TABLE 35
DAILY INCREMENTAL OPERATING COSTS
TREATMENT AND REUSE OF DECKER EFFLUENTS
Case No.
Decker Effluent
Feed Rate
FTo"6 gpd
Capital
Investment
2 2
$1,090,400 $1,120,800
121
$689,600 $1,090,400 $689,600
Daily Incremental Operating Costs
Materials
Conversion
Expense less
$110 $ 80
293.20 270.76
$ 55 $ 80 $ 40
240.85 260.88 208.53
Depreciation
Membrane
Depreciation
167.67
Other
Facilities 165.63
Depreciation
251.50
154.40
83.84
109.10
167.67
165.63
83.84
109.10
Total
$736.50
$756.67 $488.38 $674.18 $441.45
Costs
C/1000 gal.
Treated
36.83
37.81
48.87
33.72
44.14
183
-------
sent to the dilute black liquor system. The permeate
would be reused, either in the pulp washers or in other
fresh water pulping uses and hence would eventually be
returned to the dilute black liquor stream for chemical
recovery.
Operation with recycle of permeate and concentrate
should result in process credits which could offset the
costs associated with installation and operation of the
ultrafiltration plant. Potential credits could be in
the following:
—reduction of the fresh water requirements of
the pulp mill (^3-5%);
—reduction of the total mill effluent and waste
processing operations (^3-5%);
—retention of the salts presently being dis-
charged in the effluent (^16-24 tons/day);
—increasing the organic content of the dilute
black liquor without increase in liquor
volume (^8-10 tons/day);
—reduction of the mill effluent color by about
20%; and
—reduction of organic solids treated in the
secondary treatment plant by about 20%.
The potential credits to be derived are a function of
the specific pulp mill in which such an ultrafiltration
plant would be installed. The potential credits dis-
cussed below are those that might be applicable at the
pulp mill in which these pilot studies were conducted.
At other pulp mills the washing procedures/ water costs,
waste disposal costs and nature of the total mill
effluent may well differ. However, it is felt that the
conservative figures presented below are typical "ball-
park" credits which would be available at pulp mills
where the decker effluents are presently sewered.
a. Water Credits
At present, water at the Canton Mill costs c.bout
$110/106 gal. for fresh water feed and treatment of the
water in the waste treatment operation. The potential
water credit, then, is a function of the actual amount
of water used. In this study, two cases have been
184
-------
examined. At 2 x 10* gpd recycle flow, the credit
would be 2 x 110 = $220 per day. At 1 x 106 gpd
recycle flow the credit would be $110 per day. (It
should be noted those water costs are low compared to
many mills because of geographical location and well
planned waste disposal facilities.)
In the mill the water recycle would reduce the fresh
water requirements and also the effluent handling
requirements by 3-5%. No direct credit is taken for
this physical reduction of volume flow in this
presentation.
b. Retention of Salts
Recycling both the concentrate and the permeate frac-
tions of the decker effluents will retain all the con-
tained materials presently going to the waste treatment
plant. Depending on the operating conditions, the
decker effluents contain 1200-3300 ppm of sodium sulfate
For the calculation below, a value of 2500 ppm at
2 x 106 gpd flow is taken. A value of !$/# is assumed
for the sodium sulfate value.
Potential credit for retained salts =
(2500 x 10~6) (2 x 106) (8.33) (0.01) = $416
Recycle of these materials, of course, will reduce the
dissolved salt content of the total effluent. No credit
is taken for this result.
Organic Content
About 98% of the organic content of the decker effluent
is in the concentrate stream. This stream will contain
>10% organic material. The heating value of this
material is assumed to be 8000 BTU/f. Because no total
water increase in the weak black liquor stream is antici-
pated as a result of the recycle and reuse of the water,
the total heat value of the retained organics is taken
as a potential credit.
185
-------
The organic content is taken as 1000 ppm at a
2 x 10° gpd flow. The heat input is assumed to be
worth 50C/106 BTU. Then,
Potential heat value credit =
(1000 x 10"6)(2 x 106) (8.33) (8000) (0.50 x 10"6) =
$66.6
The decker effluents exit temperatures are 120-125°F.
Returning this stream to the process has two effects.
The thermal balance of the pulping process is improved
and the thermal load on the mill effluent is decreased.
No credit is taken for either effect in this calculation,
d. Total Potential Credits
The potential credits from installation of the process
are totaled below. In the calculations it has been
assumed that there is a basic load of material to be
removed in the washing operations and that the effect
of varying the washing volumes will be to change the
concentration of these materials, but that the total
amount removed will remain the same.
For 2 x 10^ gpd flow the total potential process credits
are $220 + 416 + 67 = $703 per day.
For 1 x 10 gpd flow the total potential process credits
are $110 + 416 + 67 = $593/day.
5. An Evaluation of Process Economics
The capital costs of the ultrafiltration plants for
treating the decker effluents are those for treating the
pine caustic extraction filtrate, modified by elimina-
tion of the concentrate disposal systems required for
the latter. Consequently the capital costs and the
daily operating costs displayed in Tables 29 and 35 are
lower than for similar cases in the pine caustic extrac-
tion filtrate treatment.
Again the most sensitive parameter in the study is the
flow of the decker effluent stream. For example,
186
-------
Case 9 (2 x 10 gpd) has a capital cost of $1,090,400
and a daily operating cost of $674.18; Case 10,
which presumably represents the same level of treat-
ment but on a more concentrated stream of 1 x 106 gpd,
has a projected capital cost of $689,600 and a daily
operating cost of $441.45.
As in the previous study the other parameters which
are highly sensitive are membrane flux and life, capi-
tal depreciation, operating manpower, and materials.
The decker effluent cases differ from the pine caustic
extraction filtrate cases in that for the former the
potential credits defray the operating costs of the
treatment plant. On the assumption that the projected
capital and operating costs and potential credits are
realistic, a decker effluent treatment plant, depending
on size and process configuration, would have either
a small net operating cost (comparison of $736.50/day,
operating cost, Case 6, and $703/day potential credit),
or a small net credit (comparison of $441.45/day, opera-
ting cost, Case 10, and $593/day potential credit).
Consistent with the precision of the cost estimates
and knowledge of the plant flow volumes, it is likely
that an ultrafiltration plant to treat decker effluents
could be a low-cost or net-credit operation. The
process could produce color and BOD reduction and plant
flow reduction values which have not been treated as
credits in this evaluation.
It must be emphasized that the attractive steady state
operating economics presented here are projections from
the pilot operation. As in the previous set of projec-
tions for the pine caustic extraction filtrate these
economics require additional experimental engineering
verification, especially in the area of:
—quantifying the minimum controllable flow of
decker effluents which can be used while pro-
ducing a pulp of acceptable quality for subse-
quent mill operations;
—demonstrating that reliable membranes can be ob-
tained commercially;
—evaluating alternative plant designs, including
additional experimental work, to assure that the
plant is capable of continuous 365-day operation
with minimum capital and operating costs / especial-
ly labor.
187
-------
The technical feasibility of treating the decker
effluents has been demonstrated using the feed treat-
ment and ultrafiltration processes described. It
would be premature, at present/ however, to proceed
with a full-scale plant design until the information
indicated above is obtained.
188
-------
SECTION VII
WATER REUSE
One potential value to be obtained from the use of an
ultrafiltration process to treat pulp mill process
effluents is to provide a permeate which can be re-
used within the mill. This would reduce the fresh
water requirements, as well as the total plant efflu-
ent volume.
During the course of the program, as permeate samples
were produced and analyzed, evaluations of water reuse
potential were performed.
A. PINE CAUSTIC EXTRACTION FILTRATE PERMEATE
The pilot plant produced permeates which had 90-96% of
the color removed and which represented 99-99.5% of
the flow stream fed to the system. The permeate con-
tained most of the non-organic dissolved materials (salts)
and a residual color of usually greater than 1000 ppm—
even with the high color removal effectiveness (see
Appendix for representative values).
Because of the stream color and high dissolved material
content (usually ^ 6000 ppm) it is felt that this per-
meate would have limited usefulness in the pulping and
bleaching process areas. At best, it is felt, this
permeate could be used to augment the volumes of other
"dirty" water presently used in operations such as wood
washing and cleaning of fly ash collectors. No high
value reuse capability has been developed for this per-
meate .
B. DECKER EFFLUENT PERMEATES
The pilot plant produced permeates from both the pine
and hardwood decker effluents which had about 98-99.5%
of the color removed, and which represented about 99%
of the stream feed to the system. As discussed in
other parts of this report, the concentrate from the
system would be returned to the weak black liquor stream.
The permeate with its low color (^ 100 ppm color) and
dissolved salts, primarily sodium and sulfate, is a
stream of high potential value for direct multiple re-
use in the washers of the pulping system, or as make up
water in the pulping operations. As such, the reuse of
189
-------
the decker effluents at a mill such as the Canton mill
would:
a) reduce the treated water demand by about 4%;
b) reduce the plant effluent by about 4%;
c) recycle 15-25 tons/day of chemicals normally
lost in the decker effluents; and
d) reduce the organic loading of the biological
waste treatment system by about 20%.
190
-------
SECTION VIII
ACKNOWLEDGEMENTS
The support of the project by the Environmental
Protection Agency and the help provided by Mr. George
Webster, Mr. Edmond P. Lomasney, the Federal Grant
Project Officer, Dr. Karl Suva and Ralph Scott is
acknowledged with sincere thanks.
The project was directed by Dr. Henry A. Fremont of
Champion International. The experimental program was
supervised by Mr. Dan C. Tate of Champion International.
The membrane equipment was provided by Abcor, Inc.
Drs. Robert L. Goldsmith, James R. Ryan, and David J.
Goldstein, and Messrs. Sohrab Hossain and A.C.F.
Ammerlaan, all of Abcor, Inc., contributed to the
technical support of the project and report preparation.
The progress of the project would not have been possible
without the cooperation and counsel of Champion Inter-
national personnel, including John 0. Parrott, Plant
manager, Champion International, Canton Mill, and his
very competent staff; Richard Wigger; Austin Moore;
Robert Townsend; C. E. Fredericks; and, Dr. A. E.
Vassiliades.
191
-------
SECTION IX
REFERENCES
1. Interstate Paper Corporation for the Environmental
Protection Agency, Program #12040 ENC, Grant #WPRD
183-01-68, "Color Removal from Kraft Pulping Effluent
by Lime Addition," December 1, 1971.
2. Spruill, Edgar L., "Paper Mill Waste: Treatment for
Color Removal," IW, March/April 1971, pp. 15, 21-23.
3. Gould, Matthew, "Lime-Based Process Helps Decolor
Kraft Wastewater," Chem. Eng. , January 25, 1971,
pp. 55-57.
4. Tejera, N.E. and Davis, M.W. , Jr., "Removal of Color
and Organic Matter from Kraft Mill Caustic Extraction
Waste by Coagulation," TAPPI, 53, No. 10, October
1970, pp. 1931-1934.
5. National Council for Air and Stream Improvement, Inc.,
"The Mechanisms of Color Removal in the Treatment of
Pulping and Bleaching Effluents with Lime. I. Treat-
ment cf Caustic Extraction Stage Bleaching Effluent,"
Technical Bulletin No. 239, July 1970.
6. National Council for Air and Stream Improvement, Inc.,
"The Mechanisms of Color Removal in the Treatment of
Pulping and Bleaching Effluents with Lime. II. Treat-
ment of Chlorination Stage Bleaching Effluents,"
Technical Bulletin No. 242, December 1970.
7. Davis, C.L., Jr., "Tertiary Treatment of Kraft Mill
Effluent Including Chemical Coagulation for Color
Removal," TAPPI, 52, No. 11, November 1969, pp. 2132-
2134.
8. Middlebrooks, E. J., Phillips, W.E., Jr. and Coogan,
F.J., "Chemical Coagulation of Kraft Mill Wastewater,"
IW Water and Sewage Works Supplement, March 1969,
pp. 7-9.
9. Smith, S.E. and Christman, R.F., "Coagulation of
Pulping Wastes for the Removal of Color," Journal
WPCF, 41, No. 2, Part 1, February 1969, pp. 222-231.
193
-------
REFERENCES
(continued)
10. "Projects of the Industrial Pollution Control Branch,
July 1971," Water Pollution Control Research Series
# 12000-07/71, pp. 5-22, 5-23, 5-24, 5-25.
11. Ibid., pp. 5-13, 5-26.
12. Proceedings of the TAPPI 8th Water and Air Conference,
Boston, Mass., 1971, paper entitled "Activated Carbon
System for Treatment of Paper Mill Washwaters."
13. Ibid., "Color Removal from Kraft Bleach Wastes by Ion
Exchangers."
14. "Carbon Treatment of Kraft Condensate Wastes," TAPPI,
5^, 241, 1968.
15. Lacey, R.E. and Loeb, S., eds., Industrial Processing
with Membranes, Wiley-Interscience, New York, 1972,
pp. 223+.
16. Moore, G.E. Minturn, R.E., et. al., "Hyperfiltration
and Cross-Flow Filtration of Kraft Pulp Mill and
Bleach Plant Wastes," ORNL-NSF-EP-14, May, 1972.
17. Wiley, A.J., Dubey, G.A. and Bansal, I.K., "Reverse
Osmosis Concentration of Dilute Pulp & Paper Effluents,"
for the Environmental Protection Agency, Project
#12040 EEL, February, 1972.
18. Morris, D.C., Nelson, W.R. and Walraven, G.O., "Recycle
of Papermill Waste Waters and Application of Reverse
Osmosis," for the Environmental Protection Agency,
Program #12040, January, 1972.
19. Bansal, I.K., Dubey, G.A. and Wiley, A.J., "Develop-
ment of Design Factors for Reverse Osmosis Concen-
tration of Pulping Process Effluents," presented at
Membrane Symposium, National Meeting of American
Chemical Society, Chicago, Illinois, September 14-
18, 1970.
20. Beder, H. and Gillespie, W.J., "Removal of Solutes
from Mill Effluents by Reverse Osmosis," TAPPI, 53,
No. 5, May 1970 , pp. 883-887.
194
-------
REFERENCES
(continued)
21. Bregman, Jacob I., "Membrane Processes Gain Favor
for Water Reuse/' Environmental Science £ Technology,
4_, No. 4, April 1970, pp. 296-302.
22. Wiley, A.J., Dubey, G.A., Holderby, J.M. and Ammerlaan,
A.C.F., "Concentration of Dilute Pulping Wastes by
Reverse Osmosis and Ultrafiltration," Journal WPCF,
42, No. 8, Part 2, August 1970, pp. R279-R289.
23. Ammerlaan, A.C.F. and Wiley, A.J., "Pulp Manufacturers
Research League Demonstrates Reverse Osmosis Process,"
TAPPI, 52, 1969.
24. Ammerlaan, A.C.F. and Wiley, A.J., "The Engineering
Evaluation of Reverse Osmosis as a Method of Pro-
cessing Spent Liquors of the Pulp and Paper Industry,"
prepared for the New Orleans Meeting of A.I.Ch.E.,
March 17-20, 1969.
25. Ammerlaan, A.C.F., Lueck, B.F. and Wiley, A.J.,
"Membrane Processing of Dilute Pulping Wastes by
Reverse Osmosis," TAPPI, 52, No. 1, January 1969,
pp. 118-122.
26. "Industrial Ultrafiltration," in Membrane Processes
in Industry and Biomedicine, Plenum Press, 1971.
27. Flinn, J.E., ed., Membrane Science and Technology,
Plenum Press, New York, 1970, pp. 33+, 47+.
28. Foreman, G.E., et al., "The Improvement of Spiral-
Wound Reverse Osmosis Membrane Modules," Research
and Development Progress Report, No. 675, Office of
Saline Water, April, 1971, pp.13.
29. Cadotte, J.E., and Rozelle, L.T. of North Star
Research and Development Institute, "In Situ-Formed
Condensation Polymers for Reverse Osmosis Membranes"
for Office of Saline Water, U.S. Department of the
Interior, Third Quarterly Report, January 9, 1972,
through April 8, 1972, Contract No. 14-30-2883.
195
-------
REFERENCES
(continued)
30. Hossain, S., Goldsmith, R.L., Tan, M., Wydeven, T.,
and Levan, M.I., "Evaluation of 165°F Reverse
Osmosis Modules for Washwater Purification", for
Intersociety Conference on Environmental Systems,
July 1973, San Diego.
196
-------
APPENDIX A
DETAILED PILOT PLANT PROCESS DESCRIPTION
(as initially installed)
197
-------
APPENDIX A
DETAILED PILOT PLANT PROCESS DESCRIPTION
(as initially installed)
A generalized description of the process was given in
Section iv.A. The following information describes the
pilot plant, as initially installed, in detail.
A detailed process flow is shown in Figure A-l. The
process description which follows is based on normal
operation. Components which were used at other times
are described in Table A-I, a complete listing of all
system components by intended function.
Referring to Figure A-l, the numbers contained in the
diamond symbols refer to the process streams listed at
the bottom of the Figure. This table gives average,
maximum, and minimum flow rates at different points in
the system.
Feed flow from the plant was to the 500 gal. fiberglas
feed tank (T-l). Flow was through a float valve (V-7),
which kept T-l filled so long as feed was available from
the mill. If feed flow were interrupted, a low level
shutdown switch in the tank (LS-F) shut down the pilot
plant and sounded an alarm. The unit was fitted with
a Lightnin mixer (M-l) which kept the tank contents well
mixed, enabling pH and temperature to be controlled ac-
curately.
Temperature was measured and controlled by a probe in-
stalled in the tank and control unit (TIS-1), which
controlled flow of cold water or steam through a heat
transfer coil installed in the tank (HE-1). Although
feed left the mill hot (approximately 120-135° F) some
cooling occurred before the feed flow reached the feed
tank. Consequently, a control system xvas provided to
allow either heating or cooling by manual selection.
An automatic shutdown switch and audible alarm was in-
cluded in the temperature controller, which would shut
down the system should a maximum preset temperature be
exceeded. A separate temperature probe in T-l was con-
nected to a recorder giving a continuous record of feed
temperature (TR-1).
198
-------
jtn*
n«s«
«r
fr*
Vi
«
M-7
^i..__.
1
c
i^
.,„ .,. , ••>— ' '
JMUMUIUUli Lc
MM MM>»tt« 1*7^
«• »»;iT '•':.!* '
*^^T^ "•« ^A»*« •J*4_s/K«k * ! O*«S* 4Ljd
_=)_._. _l
*r£
—T fruTB* lAfrWAjwl
ti ) <".*.*. |
J
-7~y«rw
ti««J
^°3
—ffitfj-
er
•m
A
;
NO
VO .
lUUt*
I
&1MC 4- 100 s«. PT. MMMAMi: AIC4
(12f&'-t)
in »w
±SM,^-©(2> »Mt~,
-tfT^^Hfe^EHI
fca ^/T^T""T^r'n
'Sf,**1^ * ft
r«i».-o ^
"• WAIVC
[Xj YAlVt OMN O««IN* I/' OH"»»TI<»H
^•4 IMIVt CLOSED PUHINC ur OCKKATItfM
»>A •»HHtrOKDl»
T»r Tfftt.*. ftKCOftOtM
f S • T(.M^. SWITCM
Ll • civil SWITCH
am
s
<
1>
k
t
t
OK
i
»,**
j«
*
!•
•TC i
IT
i
1
1
£!£
!•
!•*
••
«,•
*n i
4.1
L*
1
«T*
ns
'»
>
.»^
"? f
.>
i
p-i
UEl
11
it
^^^
ml
i.ii
i
a
11
*
nu
t
s
im
tB.
i
«•
1
41
>
t
i
a
i
i
H
n
f
wr
•
m
»?
n
1 1
I
n
•»i
•
JL
W
H
ff
aa
v~
MM
et
i*
™i
.i
f
V
Jw
t>
£
-
PB
1*
1
ID
• JKh
IV
PI
IB
i
nf
f t
I
r
«*
r.
M
_l
1
I
!
n
•* t-
K
"
nrc
|4
1
E
10
>0
•
tM
*
«
MM
1
I
k
M
L
u
•
i
•
r
i
i
at
»f
s
1
i
a
i
W
m
fT
M
to.
1
k
,•
• .
r?
ii
*
i
1
f
L
a
i
9,1
11
1
ff
4
1
•K
tti
s..
•DC
11
1
IK
Lft
»
,-
t , -
*T*
T
>
* v
BE
11
•"
•
1
tSS
M
H
I
• "now a i*rn MV
23S£.3r.1*
Mtinmtjn HM
m*jH
=aa
«* IVIt-lfnvMUlIT
CMMHDCt, MJW
«W
*r*irf
5136-5003
-------
TABLE A'I
COMPONENT LIST
Code
Name
Identification
CS-1
Composite
Sampler
o
o
F-1A \
F-1B J
P-2
FM-1
FI-1C
Process
Filters
Water Filter
Integral Feed
Flow Meter
First Stage
Concentrate
Flow Meter
Location
(assumes operator is
facing control panel)
Function
Sigmamotor Pump, Model
AJ>4, 4-channel finger
pump
Enclosed in dog house
on top of module tray
Two Broughton Corp.
Model 3000 'filters
with 10 y stainless
steel baskets
Broughton Corp.
Model 990 filter with
200 mesh stainless
steel basket
Water meter
Brooks Instrument
25 gpm rotameter
Model 1305
On module, left hand
side
On module, left hand
side
Mounted on top of
barrel of P-l; on
module
Control panel
Continuously pumps small
quantitites of four process
streams to composite
sample collector*. The
four boxes labelled CS-1A,
CS-1B, CS-1C, and CS-1D
in Figure 2 represent
channels for raw feed,
neutralized feed, final
concentrate, and mixed
permeate, respectively.
Filters neutralize the.feed
prior to introduction into
the membrane assemblies.
Filters water used, for
backwash of F-1A and
F-1B.
Records the total amount
of process fluid handled
by the membrane system.
Measures Stage 1- concen-
trate flow.
-------
TABLE A-I
Page 2
Code
Name
Identification
Location
Function
ts>
PI-C
FI-1PA
FI-1PB
FI-1PC
FI-2P
FI-3P
FI-4P
FI-5P .
FI-P
HE-1
LS-F
Concentrate
Flow Metera
Concentrate
Flow Regulator
and Flow Meter
Permeate Flow
Meters
Mixed Permeate
Flow Meter
Heat Transfer
coil
Feed Tank Level
Switch
Brooks Instrument
10 4pm rotameters,
Model 1305
Brooks Instrument
Self-contained flow
controller, Model
1350-8802-5-65C
Dwyer Instruments
polycarbonate rota-
meters, Models RMC
141,142,143
Brooks Instrument
25 gpm rotameter,
Model 1305
25 sq ft, 1/2"
diameter titanium
coil
Gems level switch,
Model LS-1900
Control panel
Control panel
Control panel
Measures concentrate
flows for Stages 2-5.
Regulates and measures
concentrate flow from
Stage 5.
Measures permeate flow
rates from each of the
seven membrane assemblies.
Control panel
Feed tank (T-l)
Feed tank (T-l)
Measures flow rates of
mixed permeate from all
membrane assemblies.
Heat transfer surface
for heating or cooling of
process fluid,
Shuts down membrane system
and sounds alarm on low
level of process fluid
-------
TABLE A-1
Page 3
Code
Name
Identification
LS-P
M-l
P-l
M
O
N9
Permeate Sump
Level Switch
Gems level switch.
Model LS-1900
Location
Function
Feed Tank Mixer Lightnin mixer, Model
ND-2A
Stage 1 Feed Goulds multistage
Pump centrifugal pump, Model
MB-13400
Circulation Goulds multistage
Pumps for Stages centrifugal pumps/
2,3,4, and 5 Model MB5100
In permeate sump tank
(T-2)
Feed tank (T-l)
On bottom of module
On top right hand
side of: module
Shuts down permeate
pressurization pump (P-8)
on low level in permeate
sump. tank.
Mixes feed tank contents
allowing efficient
neutralization and temp-
erature control.
Pumps process fluid
through the automatic
backwash filter unit and
Stage 1 membrane assemblies
This pump also provides
feed pressurization in
subsequent membrane stages.
Boosts pressure of feed
sequentially to Stages 2,
3,4, and 5.
P-6
Feed Booster
Pump
Corcoran 1 hp
centrifugal pump,
Mounted on feed tank
panel
Transfers feed from feed
tank to suction of P-l.
Acid Pump
Precision Control
diaphragm pump, Model
11321-71
Mounted by acid drum
on mezzanine
Pumps acid from 55 gal.
drum to feed tank for
feed neutralize*"'on.
-------
TABLE A-1
Page 4
Code
P-8
Name
Identification
PHIS-1
Permeate
Pressurization
Pump
pH Indicator/
Controller
Location
Little Giant plastic
head pump, Model
1P732 (Grainger)
Leeds and Northrop
monitor, Model 7070-
02-3-107-6-04, with
pH electrodes and probe
Mounted on underside
of module tray
Monitor mounted in
in control panel; pH
probe assembly in
feed tank
Function
CO
o
u>
PHR-1
pH Recorder
Rustrak Instrument
millivolt recorder
Control panel
Transfers pressurized
permeate into membrane
modules to keep them
pressurized and wet during
system shutdown.
Measures and controls
pH of the; process fluid.
An electrode probe
assembly is mounted in the
feed tank and input is
transmitted to the control
panel. The low alarm on
the monitor is used to turn
the acid pump on and off
with a dead band range of
0.5 pH unit. The high
alarm of the monitor is
used to shutdown the system
and sound an alarm on
rising pH.
Records pH. The
output from the Leeds and
Northrop controller, serves
as the input to the
recorder. Full-scale on
the recorder (10 divisions)
corresponds to the pH
range 2-12.
-------
Code
PI-
PI
-Flj
-F2J
PI-6
TABLE A-I
Page 5
Name
Identification
Water Filter
Pressure Gauges
Broughton Pressure
Gauges
Location
Process Filter
Pressure Gauge
0-400 psi Pressure
Gauge
Function
On water filter
On bottom of module,
in the piping running
from P-l outlet to
process filter inlet
Measures pressure drop
across water filter (F-2).
High pressure drop during
the backwash cycle in-
dicates the need to "blow-
down" the water filter.
Measures pressure,, at the
filter inlet.
PI-1I
PI-IE
PI-2I
PI-2E
PI-3I
PI-3E
PI-4I
PI-4E
PI-5I
PI-5E
PS-F
Pressure Gauges
PS-2
0-300 psi Pressure
Gauges
Control panel
Filter
Differential
Pressure Switch-
Static 0-Ring
Pressure Switch,
Low Pressure
Switch
Static O-Ring
Pressure Switch,
Mounted in module
between inlet and
outlet piping of
process filter
Mounted in piping
manifold on suction
side of P2-P5; top
right-hand side of
module
Measures inlet and outlet
pressures to each of the
five membrane stages.
Senses pressure differen-
tial across the process
filter, triggering an
automatic backwash cycle
when the pressure drop
builds up to a specified
level (nominally 75 psi)
Shuts down pumps P2, P3
P4, and PS when their
suction pressure drops
below a preset value
(nominally 35 psi}
-------
TABLE A-I
Page 6
Code
Name
Identification
Location
Function
S-l
T-l
T-2
to
s
TIS-1
Strainer
Steam Trap
Feed Tank
Permeate Sump
Tank
Temperature
Indicator/
Controller
Common Y Type
Pipeline Strainer
Next!to upper left-
hand corner of con-
trol panel
On exit line of HE-1,
located on Feed Tank
500 gallon fiberglas
tank
5 gallon polyethylene
tank
United Electric, Model
1202 Temperature
Indicating/Controller
On top of module
tray
Mounted on Feed Tank
control panel
TR-1
Temperature
Recorder
Temperature
Indicators
Rustrak Instruments
Temperature Recorder
Temperature Gauges
Control panel
On module; in inlet
piping to membrane
stages 1,2,3/4, and 5
Removes suspended solids
before concentrate flow
regulator to prevent
plugging
Removes condensed steam
Surge for process fluid,
and mixing vessel for
control of pH and temp-
erature.
Serves as reservoir for
permeate for system pres-
surization during shutdown
Measures and controls temp-
erature of process fluid
in feed tank. Electronics
control operation of V-9
allowing either steam or
cooling water to pass
through HE-1 when called
for by the sensing element.
Continuously records feed
temperature in Feed Tank.
Indicates temperature of
feed to each- of the five
membrane stages
-------
TABLE A-1
Page 7
Code
V-1I
V-2I
V-2E
V-3I
V-3E
V-4I
V-4E
V-5I
V-SE;
Name
Identification
First Stage
Control Valve
Location
1" globe valve
Control Valves 1/2" globe valves
Lower left-hand
front of module
Control panel
Function
Throttles feed flow from
P-l.
To control flows and pres-
sures through membrane
stages 2,3,4, and 5; to
isolate any of stages 2,3,
4 and/or 5, as desired.
V-1SI
V-1SE (
V-2S1 7
V-2SE
V-3SI
NV-3SE
.*V-4SI
V-4SE
V-5SI
V-5SE/
\
V-1SPA
V-1SPB
V-1SPC
V-2SP.
V-3SP
V-4SP
V-5SP J
V-1R
V-2R
V-3R
V-4R
V-5R
Feed Sample
Valves
1/4" Petcocka
Control Panel
Permeate Sample
Valves
1/4" petcocks
In permeate piping on
right hand side of
Module
To sample inlet and outlet
feed flows to each of the
five menbrane stages.
Valves for collection of
permeate samples from each
of the seven membrane
assemblies.
Pressure Relief
Valves
Brass Pressure
Relief Valves
In piping to inlet of Pressure relief valves
each of membrane stages preset at 175 psig to
1 through 5 prevent over-pressurization
of any of the five membrane
stages should system mal-
function occur
-------
TABLE A-I
Page 8
Code
Name
Identification
Location
Function
V-PR
V-6
V-7
V-8
V-9
Permeate Pressure Brass Relief Valve
Relief Valve
Check valves
Sample Valve
Float Valve
Drain Valve
Steam
Solenoid
Valve
Check-all 1/2"
Union Check valves
1/2" globe valve
Fisher Controls
float operated valve,
Type 17IF
2" ball valve
Automatic Switch
Co. Steam Solenoid
Valve, Model 8222B24
In permeate line from
Stage 5
In piping on right
hand side of module
On feed inlet piping
to Feed Tank
Feed tank
On tank outlet
In feed tank on inlet
to HE-1
Pressure relief valve
preset approximately at
50 psig,+*> relieve prensur
if overpressurization
should occur in any of
the pern.eate piping.
To allow pressurized
permeate to pass into the
membrane stages upon shut-
down, and to prevent
process fluid from enterin
the permeate sump tank whe
the system is operative.
To draw off raw feed for
sampling or drainage.
To maintain feed tank
at the desired level so
long as process feed is
available.
Drain valve to drain tank
contents
To control flow of steam,
or cooling water to HE-1,
actuated by TIS-1.
-------
TABLE A-I
Page 9
Code
V-10
Name
Identification
Location
Function
Check Valve
V-ll
First Stage
Pressure Relief
Valve
8
V-12
V-13
Check Valve
First stage
bypass valve
V-14
Ball Valve
1" Union Checkall
Check Valve
D'Este Type V
Direct-acting back-
pressure regulator
with external pressure
sensing
On suction of P-l
Behind process filters.
1" Checkall Union
Check Valve
1" globe valve
In piping between
filter outlet and
Stage 1 inlet
At top of control
panel
Brass ball valve
On suction side of
booster pump (P-6);
on feed tank
Prevents flow reversal
into the feed tank when
the system is shutdown
and operating under
permeate pressurization.
Maintains constant pressure
at filter outlet. If '
pressure drops at the
filter outlet due to cake
build-up on the filter,
V-ll closes thus reducing
the amount of bypass arounc
the filter, and increasing
the inlet pressure to
the filter.
Prevents flow backing-up
into filters during system
shutdown and pressurizatior
with permeate.
To control recirculation
rate around stage 1. In
normal operation it is
anticipated that this
valve will be closed. This
is to be confirmed in
actual operation.
Isolation valve, for service
of P-6.
-------
TABLE A-1
Page 10
Code
V-15
Name
Identification
Location
Function
Concentrate
Solenoid
Valve
1/4" Asco solenoid valve,On module,in concentrate This normally-closed
Model 8262A233
line just before FI-C
V-16
Solenoid Valve
1" normally-closed
Asco solenoid valve,
Model 8211B27
In piping on moduli
just before FM-1
V-17
Bypass valve
On outlet line of
HE-1 (on tank)next to
ST-1
V-18
Transfer valve
1" brass globe
valve
Feed Tank, on booster
pump outlet
solenoid valve is open in
operation, allowing
concentrate to be removed
continuously from the
system. When the system
is shutdown the solenoid
valve closes, thus
preventing pressurized
permeate removal from
the system through the
concentrate line.
To prevent continued
drainage of feed tank
contents into the membrane
system during automatic
shutdown. This solenoid
valve will open only when
the system is operational.
This valve is to be opened
when cooling water is
passed through HE-1. This
permits water flow to pass
to drain without having
to pass through the steam-
trap (ST-1).
This valve allows removal
of feed tank contents
after neutralization and
temperature control to
another system,if desired.
Feed material can be
pumped out this line for
pilot filtration tests.
-------
TABLE A-I
Page 11
Code
Name
Identification
V-19
V-20
V-21
to
M
o
V-23
V-24
V-25
V-26
V-27
V-28
V-29
V-30
V-31
V-FIA")
V-F1B (
V-F2A >
V-F2BJ
Location
Function
Isolation Valve 1" brass ball valve
Flush valve
1/2" brass globe valve
Permeate by-
pass valve
1/4" needle valve
Isolation
Valves
1/4" needle valves
Solenoid Valves
1" Asco 2-way normally-
closed brass solenoid
valves, Model 8211B27
On outlet line of
booster pump (P-6),
on feed tank
In concentrate line
from stage 5; on
module behind water
filter (F-2)
On bottom side of
module tray
At inlets to pressure
indicators; on rear
side of control panel
On piping to process
fluid filters
System isolation, as
needed.
When open, this permits
rapid flushing of all five
membrane stages since the
flow regulator (FI-C) can
be bypassed
To allow continuous bleed
from the permeate pump
(P-8) back to the permeate
sump tank (T-2), to avoid
overheating P-8 when
pressurized permeate is
not being transferred to
the membrane system.
To isolate pressure'gauges
should removal be required
during opera*-ion.
To isolate process filters
from feed during the
backwash cycle.
-------
TABLE A-1
Page 12
Code
Name
Identification
Location
Function
V-F3A)
V-F3B (
V-F4A f
V-F4B )
V-F5A")
V-F5B /
V^F6A \
V-F6B .)
V-CSA ^
V-CSB y
V-CSC {
V-CSD }
Solenoid Valves
Check Valves
Petcocks
2" Asco 2-way
normally-closed brass
solenoid valves,
Model 8211B82
1" and 2" brass check
valves
1/4" Petcocks
On piping to process
fluid filters
Installed in piping to
process fluid filters
Controls flow of backwash
water through the filter
during backwash cycle.
These four solenoid valves
are on the water lines to
the process filters.
Insures positive seating
of solenoid valves under
reverse pressurization.
V-CSA on feed tank in- To isolate the composite
let line, V-CSB on sampler from the four
inlet to P-l above sampling points as
P-l barrel, V-CSC on needed.
concentrate line
behind process filters,
V-CSD on mixed per-
meate line behind
filters
-------
A similar system was used to control and record pH. A
probe was installed in the tank and connected to a con-
trol system (PHIS-1). On rising pH the "low" contact
actuated an acid pump (P-7) which transferred sulfuric
acid from a drum to the surge tank. This same contact,
utilizing a "dead band" of 0.1 pH unit, stopped the acid
pump on falling pH. A "high" alarm, set at a pH above
the normal operating range, would shut down the system
and sound an audible alarm if pH in the Feed Tank rose
above a preset limit.
Feed from the Feed Tank was pumped through a five-stage
ultrafiltration system which contained spiral-wound mem-
brane modules. Feed from the feed tank was pumped by a
small booster pump (P-6) through a cumulative flow meter
(FM-1) to the suction of the Stage 1 pump (P-l). Flow
from the pump was through a throttle valve (V-II) and a
back-flushing filter system (F-1A and F-1B) for removal
of residual suspended solids, prior to introduction in-
to the ultrafiltration unit. Filter inlet and outlet
pressures were measured (PI-6 and PI-II). An automatic
backwashing cycle was triggered by a differential pressure
switch (PS-F) installed between the filter inlet and out-
let lines. A pressure relief valve (V-ll) installed on
the upstream side of the filter, but controlled by a sen-
sing element on the downstream side, maintained a constant
pressure at the filter outlet. Recirculation of feed
through the back pressure regulator was to the Stage 1
pump suction. After the filter, flow was through a check
valve (V-12) into three parallel passes of membrane cart-
ridges. Membrane Assemble 1-A initially contained East-
man Kodak spirals (3) and assemblies 1-B and 1-C contained
TJ Engineering spirals (3 each). Permeates from each pass
were collected separately and their flows were measured
and sampled individually.
On the inlet side of Stage 1, pressure (PI-1I) and tem-
perature (TI-1I) were measured, and a sample valve (V-1SI)
was used for collection of Stage 1 feed. A pressure re-
lief valve (V-1R) was provided as a safety device. The
concentrate was sampled through a sample valve (V-lSE)
and its flow rate (FI-1C) and pressure (PI-IE) measured.
A line was provided to recirculate part of the concen-
trate through a "bypass valve" (V-13). Stage 1 could
operate with or without recirculation through V-13.
Operation of the subsequent membrane stages was always
with recirculation, since this was required to maintain
212
-------
a feed flow rate through the membrane cartridges at or
above about 4 gpm. Flow on a "once-through" basis would
fall below this level, and recirculation was required
to achieve satisfactory operation.
Flow to Stages 2 through 5 was introduced into the suc-
tion side of the circulation pump for each stage. A low
pressure switch (PS-2) was installed to shut down these
stages should an upset in flow occur.
For each stage (Stages 2 through 5 are identical) flow
from the booster pump passed through a throttle valve
into the membrane cartridges. Pressure and temperature
were measured on the inlet to the membranes and pressure
and flow rate on the outlet. Pressure relief valves were
provided as safety devices should overpressurization oc-
cur. Sample valves allowed collection of feed and con-
centrate samples for each stage. Permeate flows were
individually sampled and collected, and their flow rates
measured. Each stage had an internal recirculation loop
which permitted maintenance of the desired flow rate
through the membrane cartridges. Initially, Stages
2 and 3 contained TJ Engineering cartridges (3 each) and
Stages 4 and 5 contained Gulf Environmental Systems cart-
ridges (2 each) .
The final concentrate from Stage 5 was passed through a
flow control valve and flow meter (FI-C) prior to collec-
tion for incineration tests or discharge.
Each stage also had a means of introducing pressurized
premeate to the feed side of the membrane cartridges for
wet storage during system shutdown. The system for sup-
plying pressurized permeate to the membrane modules was
quite simple. Permeate from the five stages was manifol-
ded and introduced into a permeate surge tank (T-2) .
A permeate pump (P-8) automatically pumped permeate to
the membranes for each stage. Flow to the membranes was
through check valves, so that when the system was opera-
ting high-pressure feed could not be pumped into the per-
meate system. During shutdown the permeate surge tank
remained filled since all permeate fed to the membrane
system was returned as permeate. During operation, per-
meate passed through the surge tank and overflowed to
drain.
o
Membrane areas were approximately 300 ft for Stage 1
and 100 ft2 for each of Staqes 2 through 5. Total
213
-------
membrane area was approximately 700 ft2. At a nominal
membrane flux of 15 gal./day/ft2 (gfd) the plant had a
capacity of 10,500 gpd.
A revised pretreatment system flow schematic is given
in Figure A-2. The changes were made to permit instal-
lation of various filters, a cleaning system, and once-
through flushing of the membrane shells.
214
-------
•••
IV
-------
APPENDIX B
ULTRAFILTRATION PILOT PLANT
SPIRAL MEMBRANE CARTRIDGE IDENTIFICATION
217
-------
APPENDIX B Page 1
ULTRAFILTRATION PILOT PLANT SPIRAL MEMBRANE CARTRIDGE IDENTIFICATION
UF"^-^^^ Date
Cartridge^^^^^
Location ..
Stage
la
Ib
Ic
2
3
4
5
Cartridge
Position
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
8-19-72
Type
E
E
E
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
Gulf
Gulf
Gulf
Guli
«
1
2
3
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
8-22-72
Type
E
E
E
T.J.
.T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
Gulf
None
Gulf
Gulf
None
Gulf
*
1
2
3
16
17
18
13
14
15
7
8
9
10
11
12
1
2
3
4
8-30-72
Type
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
E
E
E
T.J.
T.J.
T.J.
*
4
5
6
1
2
3
19
20
21
8-31-72
Type
E
E
E
T.J.
T.J.
T.J.
Gulf
Gulf
T.J.
T.J.
T.J.
.T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
Gulf
Gulf
f
1
21
3
1
21
3
1
2
19
20
21
10
11
12
13
14
15
3
4
218
Eastman
Lot numbers of the first set of membrane
cartridges were not recorded, and they
are therefore numbered starting from
No. 1. Prom 11/1/72 all cartridges are
identified by manufacturer's lot number'.
Entries denote when cartridges were changed;
if no entry is given, then the cartridge was»not changed.
-------
APPENDIX B•
ULTRAFILTRATION PILOT PLANT SPIRAL MEMBRANE CARTRIDGE IDENTIFICATION
UF -^^^ Date
Cartridge"*—
Location --^
Stage
la
Ib
Ic
2
3
4
5
Cartridge
Position
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
11-1-72
Type
T.J.
n
it
n
n
n
n
ti
n
n
n
n
n
n
n
n
n
H
n
it
n
*
4713
4723
4724
4715
4709
4711
4725
4722
4719
4717
4716
4714
4708
4736
4712
4737
4735
4731
4718
4720
4721
11-3-72
Type
T.J.
T.J.
T.J.
*
4718
4720
4523
11-5-72
Type
T.J.
T.J.
T.J.
Gulf
Gulf
ft
3747
4725
4722
3
4
11-18-72
Type
T.J.
T.J.
T.J.
#
3752
4731
4735
E
Eastman
219
-------
APPENDIX B Page 3
ULTRAFILTRATION PILOT PLANT SPIRAL MEMBRANE CARTRIDGE IDENTIFICATION
UF"^^^ Date
Cartridge "~-^^^
Location ^^--^...
Stage
la
Ib
Ic
2
3
4
5
Cartridge
Position
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
12-8-72
Type
T.J.
n
ii
it
n
n
n
n
n
n
n
n
n
n
n
n
n
n
Gulf
Gulf
. #
1948
1970
3749
4715
4709
4711
3747
4725
4722
4717
4716
4714
4708
4736
4712
3752
4731
4735
3
4
12-19-72
Type
T.J.
n
n
n
' n
n
#
4715
4709
4711
1948
1970
3749
12-21-72
Type
T.J.
n
it
it
n
n
ii
it
it
n
n
n
it
n
it
Gulf
Gulf
Gulf
Gulf
#
4715
4709
4711
2026
4283
4395
3747
4725
4722
4717
4716
4714
4708
4736
4712
1
2
3
4
1-4-72
Type
T.J.*
T.J.*
T.J.*
f
4829
4828
4827
*Wide channel
corrugated spacer
cartridges
E = Eastman
220
-------
APPENDIX B Page 4
ULTRAFILTRATION PILOT PLANT SPIRAL MEMBRANE CARTRIDGE IDENTIFICATION
UF*" — "---^ Date
Cartridge "--v^^^
Location — ,.
Stage
la
Ib
Ic
2
3
4
5
Cartridge
Position "
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
2-8-73
Type
T.J.
it
it
ti
ii
ti
n
ii
n
n
n
n
n
n
ii
H
n
n
n
n
it
* .
1970
3748
2209
2026
4283
4395
3751B
3782
3751A
4829
4828
4827
4708
4736
4712
4731
4721
4737
4735
4720
4718
2-10-73
Type
T.J.
T.J.
T.J.
*
3749
3782
3751B
2-14-73
Type
T.J.
T.J.
T.J.
*
1950
4722
4714
Type
*
E = Eastman
221
-------
APPENDIX C
DETAILS OF FILTERS USED IN PILOT PLANT PROGRAM
223
-------
Can save by recovering solids
formerly lost in effluents
Reduces operating costs and
improves efficiency in process
separations and in pol Iution
control
Minimal maintenance and no
power costs—there are no
pumps, motors, nozzles or
moving parts
Provides fast return on invest-
ment (often within months)
Unique screen is self-cleaning,
non-clogging and puncture-
proof . Needs I itt le or no
attention
Easy to install. Saves space
The patented C-E Bauer
Hydrasieve™ is a simple, highly
efficient screening device for
removing solids from low con-
sistency slurries.
Exclusive design features and
the unit's ability to operate con-
tinuously for extended periods
without attention make the rugged
Hydrasieve a reliable profit builder
in a broad range of fluid removal
applications. Hundreds of units
are now in use.
Typical liquids/solids separa-
tion operations include dewater-
ing, thickening, recovering usable
solids, classifying, and
fractionating.
Installations. Waste water treat-
ment and pollution control sys-
tems. Municipal sewage plants.
Chemical, plastic, and ceramic
classification. Synthetic and
natural fiber recovery. Salvaging
rubber fines. Processing soup in-
gredients, fish, citrus fruits, etc.
Recovering hog hair and other val-
uable solids for meat and hide
processors. and similar operations.
C-E Bauer developed the Hydra-
sieve originally for the pulp and
paper industry. In high density
pulping systems, it pre-thickens
pulp ahead of the Bauer Helipress"
HOW IT WORKS
Gravity feed
of liquids/solids I
Self cleaning.
non clogging stainless
steel screen for
continuous dewatering
,
Rugged all stainless
steel or fiber
glass construction
minimizes maintenance
_. Headbox
Alternate
feed inlet
Removed or
recovered
solids
Note how "Coanda" effect strips liquid from
bottom of the stream flowing down the
screen. This "wall attachment" effect
accelerates fluid removal action.
224
•The Hydrasieve is manufactured under U.S. Letters Patents No. 3451.555 and 3.452.876. and patents pending
Canadian patent No 835737; Great Britain 1.196.303/1.196.304. Menico 99.076. France 1.529.448: Belgium
729 813. Other patents are pending in the United States, and foreign countries.
"Cleaned" effluent improves plant
pollution control efficiency,
lowers B.O.D.. reduces sewage rates
Downward curve of Bauer screen bars
divides the flow of slurry into separate
streams between the vertical supports thus
preventing clogging or blinding.
-------
the revolutionary
HYDROMATION IN-DEPTH FILTER
225
-------
THE SPARKLER MCRO FILTER
IS A REVOLUTIONARY DESIGN
WITH THE EMPHASIS ON
SIMPLICITY AND EFFICIENCY
OVERHEAD SUSPENSION DESIGN
Engineered for heavy duty service
Rugged structural steel frame
Rigid support for the tank and cover
Stationary cover - - unnecessary to break
inlet and outlet piping to open filter
Retractable tank with 4 point suspension
Perfect meshing of tank and cover — posi-
tive seal
• No cake disturbance in opening - - plates
remain stationary
• Internal self-sluicing
• Dry cake removal either by retracting tank
or internal conveyor screw
• While in operation Vz of frame space free
for traffic
• Simple to completely automate
• Capacity 10 to 300C sq. ft. filter area
226
-------
F.O.
ITEM
NO.
1
2
3
4
5
6
7
8
9
to
II
PAR Jo. PER CONSTRUCT. MAT'L.
DWG . NO .
3002-000-
3001-001 -
3000-01 1 -
3001 -014-
-015-
-016-
-017-
-018-
3001 -019 -
250?
£75-
BRONZE
-001
-001
-00^
-004
-001
001
-004
SEE
0-RING MATERIAL
NITRILE BUNA N
TEFLON COATED
300
MA.
MA)
CUJ
ADC
CU!
BRC
BA!
DA
-019-001
X.OPERATIN
<.OPR.TEMF
;TOMFR
STN.STL STEEL
-002 -003
-00 -001
-005 -006
-004 -004
-002 -003
002 003
-005 -006
•uuo uuc
TABULATION
< c
A &
AVAILABLE
VITON TEFLON
TEFLON COAT
3001-01
G PRES
'ERATUR
)RFSS
5T. ORDER
3UGHTON OF
5KET FILTEf
TE SHIPPED
Nn
ID. No,
R MEDIA
•
9-002 3001-0
SURE
F
QUAN DESCRIPTION
FILTER ASSEMBLY
BASKET ASSEMBLY
•'OP HOUSING
SAFETY LOCK
"OP CAP
CASING
INLET CONNEC"ION
RPlTTAM
0-RING
'2. ££/)AJ?
z /ita*/1 A/-/"/*^"
TEFLON
ENCAPSULATED
19-003 3001-019-004
PSI
•F
BAJ
473 W
WELLE
U—
^X
0ZZ&
fA
5
v^/^v. C^
SEC!
=7.! 25 — -i
f)! 1
f .? /"
L t
jfaj
ft
\
••"
^
sf
>*•
— 1
|
43.
fltdt
1
52
94
a ' — ' 3.375
'ION VIEW
19
©
©
\
Q FILTER ASSEMBLY
}T WILSON
^lUilion Com/jam
/ j
ASHINGTON STREET
ISLEY, MASS. 021S1
017)237-1755
BROUGHTON CORPORATION
GLENS FALLS NEW YORK, 12801
3000 FILTER ASM-2" PLAIN END
DWG. BY: ^u,^
CK'D. BY; «r>/
SCALE:
^ DA
ow
TE: i
G. NO
?oo
£-*<>.*
2-000-00_
/r^
-------
li
li
14
li
!««• XH.T
COYK »JH»tT
V«LVC
rcc
IITI4-C
NOTE
WHtN ORDERING PARTS. LIST SERIAL
NUMBER AND DRAWING NUMBE R F-2916
SPARKLER MFG. COMPANY
CONROE TEXAS
STANDARD L-6 PARTS LtST
"GUARDIAN" TRAP FILTER
DATE
SCALE:
If 5 '64
ORN BV
CKD BY
WE J
BCD
F29I6
228
-------
TYPE IB
175 P.S.I. @ 250°F
3/4* NPT Connections
TYPE CT
300 P.S.I, (n 200°F
3/t" or 1" NPT Connections
1B1
1B2
Three-piece housing; centerpost construction. Available in one
and two-high cartridge models.
STANDARD — cast iron and steel housing, asbestos shell gaskets,
fiber cap nut gasket, drain plug, steel internals, or, 304 stainless
steel housing and 304 stainless steel internals. Mounting pads on
all models are drilled and tapped for mounting bracket.
DIMENSIONS-WEIGHTS
CT-101
CT-102
1B1
jrvj
3/«"
3/4"
12V
22%"
4l/4 "
4l/»"
9lbs.
12 Ibs.
Three-piece housing; ring nut construction holds sump to head.
Available in one and two-high cartridge models.
STANDARD — cast brass head and ring nut, drawn 304 stainless
steel sump, Buna N head gasket, 302 stainless internals. Heads with
H" connections have mounting pads drilled and tapped for
mounting bracket. Also available with a 304 stainless steel cast
head (ring nut nickel plated.)
DIMENSIONS-WEIGHTS
HOUSING CATALOG NUMBERS
Mounting bracket
#35581-01
Filter utilizes:
Micro-Klean Series G78 cartridge(s)
all Micro-Wynd cartridge(s)
Micro-Screen Series 52243 cartridge
Micro-Screen Series 52043 cartridge
Poro-Klean Series 50387 cartridge
HOUSING CATALOG NUMBERS
Mounting bracket
135581-03
Filter utilizes:
Micro-Klean Series G78
all Micro-Wynd carmdge(s)
229
ORDER CARTRIDGES SEPARATELY
-------
THE RIGHT FILTER FOR ANY FLUID
'-&
MICRO-KLEAN D
^4
4
3
•«
MICRO-WYND
Liquid and gas filters of AMF Cuno design are available from
your distributor in several types. Each filter tvpe is described
below with detailed information further in the catalog.
1 MICRO-KLEAN II... depth type, disposable fiber car-
tridge in micronic ratings. Ideal for removing contaminant
which is fibrous, abrasive or gelatinous in character. Typical
applications include gas, alcohols, glycols, coolants, fuels,
oils, lubricants, cosmetics, paints and varnishes, syrups, com-
pressed air, water. See pages 4 and 5. For Micro-Klean air
line filters see pages 14 and 15.
MICRO-WYND II . . . depth type, disposable wound car-
tridge in ratings from 1 to 350 microns. Intended for use where
Micro-Klean fibers or resin binders are not compatible with
the fluid, or, where a wound type cartridge is preferred.
Typical applications include plating solutions, organic sol-
vents, waxes, detergents, plastics, resins, deodorants, animal
and vegetable fats and oils, metal cleaning solutions. See
pages 6 and 7.
4
MICRO-SCREEN
MICRO-SCREEN . . . surface type, metal screen cartridge
—cleanable and reusable—in micronic and screen mesh
ratings. Ideal filter media for final protection in ultra-clean
fluid systems . . . offering complete freedom from media
migration. Also intended for applications involving high
temperature and corrosive conditions. Typical applications
include steam, strong acids, concentrated alkalis, strong
reducing agents and other chemicals which react with fiber
type cartridges. See pages 8 and 9.
PORO-KLEAN . . . depth type, sintered metal cartridge—
cleanable and reusable in 5, 10, 20 and 40 micron ratings.
Used in place of Micro-Screen when contaminant is gelat-
inous or fibrous in character. Also recommended for heavy
viscous fluids. Typical applications include steam, air, water,
solvents, polymers, acids, cellulose solutions, process gases.
See pages 8 and 9.
PORO-KLEAN
AUTO-KLEAN
CUNO-PORE . . . depth type, disposable media in cartridge
and disc form. Will remove contaminant which is fibrous,
abrasive or gelatinous in character. Possessing natural dessi-
cating properties, it is ideal for removing trace quantities of
water from oil. Recommended for free-flowing liquids where
optical clarity is of primary importance. Cartridge type
filters are used on dielectric insulating oils, coolants, deter-
gents, cosmetics, light lubricants, fuels, degreasers and
chemicals.
See pages 16 -19.
AUTO-KLEAN . . edge type, all metal filter. One turn of
the handle cleans cartridge and restores full flow. Ideal for
removing particles as small as 36 microns (.0015"). Typical
applications include paint, adhesives, resins, greases, inks,
tar, cellulose solutions, waxes, soaps, hydrocarbon oils, fuels
and lube oils. See pages 20-23
230
-------
APPENDIX D
MECHANICAL PROBLEMS OF DIFFERENT
SPIRAL MEMBRANE CARTRIDGES
231
-------
APPENDIX D
MECHANICAL PROBLEMS OF DIFFERENT SPIRAL MEMBRANE CARTRIDGES
Date
11-5-72
12-7-72
2-7-73
tv>
U>
hi
2-13-73
8-21-72
11-15-72
12-7-72
Operating Information
Stage
la
la
la
la
Ib
Ib
Ib
Cartridge
Number
T.J. 4713
4723
4724
T.J. 4713
T.J. 4715
4709
4711
T.J. 1970
3748
2209
T.J.
T.J. 4715
T.J. 4715
Cartridge
Position
Inlet
Middle
Outlet
Inlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Inlet
Membrane Cartridge
Characteristics
(AP) , psi
50-60
50-60
50-60
Rejec-
tion
O.K.
O.K.
Low
Low
O.K.
O.K.
Problems
Brine
seals
loose fit
Brine seal
flipped
Brine
seals
flipped
O-ring
leak
Cartridge
failure*
Brine
seals
weak
Brine
seal
flipped
Comments
Reinforced brine seal
May have happened around
12-2 when AP was high
May have occurred around
1-4-73 when AP suddenly
dropped from a high value
to a low one
All three cartridges were
tested individually and
found to give good color
rejection
Membrane replaced on
8-22-72
Reinforced brine seal with
additional tapes
-------
APPENDIX D
(continued)
MECHANICAL PROBLEMS OF DIFFERENT SPIRAL MEMBRANE CARTRIDGES
Page 2
Date
12-20-72
8-21-72
9-13-72
N> 11-15-72
u>
u>
1-4-73
12-6-72
Operating Information
Stage
Ib
Ic
Ic
Ic
2
3
Cartridge
Number
T.J. 1948
T.J.
T.J.
T.J. 4722
T.J. 4717
4714
T.J. 4708
Cartridge
Position
Inlet
Inlet
Inlet
Outlet
Inlet
Membrane Cartridge
Characteristics
(AP) , psi
Rejec-
tion
Low
Low
O.K.
LOW
O.K.
O.K.
Problems
Cartridge
failure*
Cartridge
failure*
Brine seal
weak
0-ring
failure
Brine
seals
squeezed
out, flow
leaking
underneath
brine
seals
Brine seal
flipped;
other two
cartridges
were O.K.
Comments
Membrane replaced on
12-21-72
Membrane replaced on
8-21-72
Reinforced brine seal
with tapes
Replaced O-ring and
rejection became normal
May have happened
between 12-3-72 and
12-5-72 when the
(AP) was high
-------
APPENDIX D
(continued) Page 3
MECHANICAL PROBLEMS OF DIFFERENT SPIRAL MEMBRANE CARTRIDGES
Date
10-26-72
10-31-72
11-16-72-
NJ
*
12-19-72
12-21-72
2-7-73
11-1-72
Operating Information
Stage
4
4
4
4
4
4
5
Cartridge
Number
T.J.
T.J.
T.J. 4737
4735
4731
T.J. 3746
T.J. 4724
Gulf
T.J. 4721
Cartridge
Position
Inlet
Middle
Outlet
Inlet
Middle
InJ.et
Outlet
Membrane Cartridge
Characteristics
(AP), psi
Rejec-
tion
Low
Low
Low
O.K.
Low
Problems
Unknown
reasons
Brine seals
weak
O-ring
failure
Brine seal
flipped
O-ring
failure
Brine seal
weak
O-ring
leak
Comments
All three cartridges we're
O.K. when tested indivi-
dually later on. Replaced
cartridges.
(AP) rose to 50 psi before
it happened
Replaced all cartridges
by two Gulf cartridges
Brine seal of the second
Gulf cartridge was O.K.
Slot too wide.
Replaced cartridge.
-------
APPENDIX D
(continued)
Page 4
MECHANICAL PROBLEMS OF DIFFERENT SPIRAL MEMBRANE CARTRIDGES
Date
11-3-72
2-7-73
to
Ul
Operating Information
Stage
5
5
Cartridge
Number
T.J. 4718
4720
4523
Gulf
Gulf
Cartridge
Position
Inlet
Middle
Outlet
Inlet
Outlet
Membrane Cartridge
Characteristics
(AP), psi
Rejec-
tion
Low
O.K.
Problems
Cartridge
failure*
Brine seals
had loose
tapes
Continents
Cartridges 4718 and 4720
were found to be O.K.
during individual cartridge
screening tests. Replaced
all three T.J. cartridges
by Gulf cartridges
.
*Any one of the following reasons:
1. Leakage at the glue searn;
2. Leakage at the material fold
adjacent to the permeate
collection tube;
3. Membrane and/or backing material
wrinkles causing direct leakage;
4. Pinholes in membrane.
-------
APPENDIX E
SLIME ANALYSIS
237
-------
APPENDIX E
SLIME ANALYSIS
A sample of the material flushed from a fouled membrane
cartridge was analyzed. The material was de-cationized
using Amberlite IR-120 resin. The flow sheet showing the
make-up of the sample and the fractions obtained is as
follows:
The six fractions from the cation column were examined.
The results leave some strange unanswered questions.
Fractions "A" and "C" should be the same. When pyro-
lyzed, the chromatograms closely matched. However, the
TGA showed degradations at 270° and 520° with a 10%
white ash for "A" and degradations at 270° and 440° with
1.3% white ash for "C". The difference in ash content
is unexplainable; especially since neither sample should
contain any inorganic salts.
The chromatograms from the pyrolysis of these fractions
closely match the pyrolysis chromatogram for glucose.
The pyrolysis by-products can be grouped into three
groups. This is a 600° pyrolysis.
1. Very volatile - retention times 0.6 to 2.2 min.
2. Intermediate - rentention times 6.0 to 12.3 min.
3. High boiling - retention times 14.0 to 30.0 min.
Groups 1 and 2 closely match the pyrolysis of glucose;
however group 2 is definintely stronger than glucose.
When pyrolized at 950°, the chromatograms match glucose
even closer. Apparently some of this material is poly-
saccharides in nature. It should be noted that even
though a similarity exists to a pyrolysis of cellulose,
several peaks are different.
Fraction D. The 600° pyrolysis of "D" match as far as
groups 1 and 2 are concerned; however the high boiling
group of compounds are completely different. A sample
of the crystals from the concentrate was dissolved in
water and acidified. The precipitate was filtered.
The pyrolysis of this solid matches exactly fraction "D".
TGA showed degradation at 300, 500, and 650°. Ash was
25% white.
238
-------
Sample I 3 Membrane wash - pH 6.6
Added 20 mg NaOH AH-10.4 then added 400 ml MeOH
Passed solution through cation column as 1:1 MeOH
Effluent
Evap. to remove MeOH
Fpt. separated - filtered
Filtration very slow - no wash
About 75% of ppt.
Sample "A"
46% of 13
Cation resin column
Backwashed with water
Remainder of effluent
air-dried
Black granular crystals
washed with water - filtered
Covered with
acetone
1
I
Filtrate
1
Acetone extract
Sample "B"
Filtrate
Solid
Sample "C"
10% of 13
cation resin column
washed with 72.94 mg HC1
(not enough HC1
Effluent neutralized with
56 mg NaOH
Volume 190 ml
1
ppt.
Collected backwash
milky in appearance
separated on standing
filtered
1
40 ml Evap. to dryness
Solids calculated as Na
1.8% of 13
1
150 ml
separated
PPt.
Extracted with ethur
1
Solid Sample
11% of 13
I
Filtrate
discarded
I
Ether extract
"E"
I
Ag. sol.
acidified '
extracted with ether
I
I
Ether extract
Sample "F"
Ag. sol.
-------
Fraction B. The chromatogram of this fraction was very
poor. A group of 12 to 15 materials with an estimated
boiling range 230 to 300°. When reacted to form the tri-
methylsilyl derivative, some differences are noted. The
IR showed strong aliphatic and carbonyl. The extract
also contained some hydroxyl absorption; but spectra
quality is poor.
Fraction E. IR showed strong aliphatic/ especially CH2.
Also minor indications of hydroxyl (extremely weak),
carbonyl (one only), and aromatic (possibility of a
little phenylphthalein - used as an indicator). The GC
showed several distinct peaks and a group of lesser peaks.
Slight changes occur when the fraction is derivatized.
Fraction F. This fraction chromatographs similar to
fraction "B", except that a very large distinct peak is
found which matches one of the peaks in "E". This
material does not appear to form any derivative. These
materials would all be very high boiling. The IR showed
strong aliphatic (CH? and CH-,) . Also minor indications
of hydroxyl, carbonyl (2 bands) and aromatic. "E" appeared
to be more contaminated than "F", although the single
carbonyl indicates that at least one component is absent
in "E" . Other than the carbonyl difference, both "E"
and "F" present very similar spectra. The strong ali-
phatic indicates a hydrocarbon possibility such as kero-
sene. Disregarding the fact that such a material should
not be in "F", the GC would indicate that if hydrocar-
bons were present, it would have to be C-15 to 18.
140
-------
~ . Subjcrt Ftvld &,
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
5
Organization
Champion International Corporation
Knightsbridge
Hamilton, Ohio 45020
Tifle
COLOR REMOVAL FROM KRAFT MILL EFFLUENTS BY ULTRAFILTRATION
i Q |\ Authorfs)
H A. Fremont
D. C. Tate
R. L. Goldsmith (Abcor, Inc
Cambridge, Mass.)
16
Project Designation
S 800 261
Note
22
Citation
Environmental Protection. Agency report
number, JEPA-660/2-73-019, December 1973.
Descriptors (starred First) #puip anl^J**
SEND TO: WATER RESOURCES SCIENTIFIC INFOBMATI
U.S DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C 502 40
WR:I02
WRS1C
(REV. JULY
CEN
U.S. GOVERNMENT PRINTING OFFICE: !»*-S46-314:179
------- |