United State*
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
EPA-600/2-80-045
February 1980
Reseercn and Development
Color Removal from
Kraft Mill Effluents by
Ultrafiltration
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-80-045
February 1980
COLOR REMOVAL FROM KRAFT MILL
EFFLUENTS BY ULTRAFILTRATION
by
Henry A. Fremont
David J. Striley
Champion International
Knightsbridge, Hamilton, Ohio 45020
and
Myles H. Kleper
Robert L. Goldsmith
Wai den Division of Abcor, Inc.
Wilmington, Massachusetts 01887
Grant No. S804312-01
Project Officer
Kirk Willard
Food and Wood Products Branch
Industrial Environmental Research Laboratory
Cincinnati, Ohio 45268
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor
does mention of trade names or commercial products institute endorsement or
recommendation for use.
11
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on our
health often require that new and increasingly more efficient pollution
control methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both, efficiently and economically.
A field demonstration of color removal from kraft mill effluent streams
by ultrafiltration is discussed in this report. Technical and economic .
assessments of caustic extraction filtrate, pine decker and hardwood decker
effluent treatment by both spiral-wound and tubular ultrafiltration modules
are presented. It is hoped that the results of this study will lead to
continued research in the field of color removal by membrane processes and
eventually provide the pulping industry with a cost-effective method for
resource recovery and pollution abatement with these complex effluent
streams.
The Food and Wood Products Branch of the Industrial Environmental
Research Laboratory should be contacted for further information on this
subject.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
m
-------
ABSTRACT
Color removal from kraft mill effluents by ultrafnitration (UF) has been
successfully demonstrated during this program. A 3-stage, nominal 37.9 m3/
day (10,000 gpd) UF pilot plant was operated on caustic extraction filtrate
for several months. Extensive evaluation of spiral-wound UF modules was
carried out prior to staged-system operation in single module tests. During
these tests feed pretreatment and prefiltration options were investigated and
the effects of a range of operating parameters on module flux performance
were studied. A second module configuration, tubular assemblies, was tested
in both 12.7 mm (0.5 in) and 25.4 mm (1 in) diameter tubes. All field tests
were performed at the Canton, North Carolina Mill of Champion International.
Non-cellulosic ultrafiltration membranes were evaluated in laboratory
tests before field trials were initiated. The preferred membrane was cast
from a polysulfone formulation.
Spiral-wound modules showed severe flux loss within a few hours exposure
to the waste stream. Membrane surface analysis identified the main stream
foulants as kaolinite clay, starch and titanium dioxide. These foulants were
the result of white water recycle from the paper mill back to the pulp mill.
Pretreatment and prefiltration techniques were ineffective in preventing
"slime" layer formation by these species.
Tubular modules exhibited high, stable process flux and recoverable
water flux characteristics. Membrane surface fouling was not observed with
tubular modules. These modules operate under more turbulent flow than
spiral-wound modules reducing "slime" layer formation. Average flux values
when processing caustic extraction filtrate were 2.87 m3/m2-day (70 gfd) at
a 1.2X concentration, 2.26 m3/m2-day (55 gfd) at a 10X concentration and
1.03 m3/m2-day (25 gfd) at a 50X concentration. These data were recorded at
50°C, 5.1 atm (75 psig) inlet pressure and a circulation flowrate through
25.4 mm diameter tubular assemblies of 136 m3/day (25 gpm).
Color removal by the non-cellulosic ultrafiltration membranes ranged
from 97% to 99% when calculated on a concentrate basis. Projections based
on caustic extraction filtrate process data indicate ultrafiltration would
result in an overall color reduction of 91% (mass basis) for this stream.
iv
-------
Conceptual designs and economic analyses were developed for treatment
systems with capacities of 3,790 m3/day (1 MM gpd) and 7,980 m3/day (2 MM
gpd). Both tubular and spiral-wound module systems were analyzed with pro-
cess streams of caustic extraction filtrate and, pine and hardwood decker
effluent. For tubular systems installed capital costs ranged to $4 MM.
Operating costs ranged to $2.52/metric ton of pulp produced per day (to
$2.27/ton). Spiral-wound module system costs were based on idealized
systems and are not achievable given today's technology. Capital costs for
spiral-wound module systems were projected to be/-v50% of tubular system
costs. Treatment costs were estimated to range to $1.89/metric ton of pulp
(to $1.72/ton).
Additionally during this program, caustic extraction filtrate and decker
effluent stream characteristics were monitored and qualitative assessments
of ultrafiltrate and UF concentrate recycle within a kraft mill were made.
This report was submitted in fulfillment of Grant NO. S804312-011 by
Champion International and the Walden Division of Abcor, Inc. under the
sponsorship of the U.S. Environmental Protection Agency. This report covers
a period from March 29, 1976 to August 25, 1978, and work was completed as
of September 29, 1978.
-------
TABLE OF CONTENTS
DISCLAIMER ii
FOREWORD , i1 i
ABSTRACT iv
FIGURES vix
TABLES V11
ENGLISH-METRIC CONVERSION TABLE XlV
ACKNOWLEDGMENT xv
1. INTRODUCTION 1
Background 1
Ultrafi1tration 4
Waste Stream Characteristics (North Carolina Mill) 5
Ultrafiltration Process Considerations 6
2. CONCLUSIONS 8
3. RECOMMENDATIONS 13
4. PROGRAM OVERVIEW 14
5. EXPERIMENTAL EQUIPMENT AND PROCEDURES 16
Pilot Plant 16
Single Module Test Stand 28
Tubular Modul e Test Stand 30
Membrane Cleani ng Procedures 30
51 mm Diameter Depth FiHer 32
Laboratory Ultrafiltration System 35
Stirred Cell Ultrafiltration System 35
Multiple Cell Tests 39
Membrane Casting Solution Preparation 39
Sampli ng and Analysi s 39
6. RESULTS AND DISCUSSION 43
Feed Character!' sti cs and Pretreatment 43
Selection of Preferred Membrane for Color Removal 51
Field Experience with Spiral-Wound Modules 71
Field Experience with Tubular Assemblies 96
Cleani ng Effecti veness 114
Materi al Balance 115
7. CONCEPTUAL DESIGN 125
Introduction 125
Details of Case 1 Design 125
Details of Case 2 Design 132
Details of Case 3 Design 133
Details of Case 4 Design 136
vii
-------
Details of Case 5 Design (Idealized Spiral-Wound
Module System) 136
Details of Case 6 Design (Idealized Spiral-Wound
Module System) 139
Details of Case 7 Design (Idealized Spiral-Wound
Modul e System) 139
Details of Case 8 Design (Idealized Spiral-Wound
Module System) 141
8. PROJECTED ECONOMICS FOR FULL-SCALE SYSTEMS 144
Introduction 144
Costs for Design Cases 1 Through 4 144
Costs for Design Cases 5 through 8 (Idealized
Spiral-Wound Module Systems) 168
Potential Credits for Water Reuse and Resource
Recovery 180
9. WATER REUSE POTENTIAL 181
Pine Bleachery Caustic Extraction Filtrate 181
Pine and Hardwood Pulp Washing Decker Effluents 182
REFERENCES 183
APPENDICES
A. MEMBRANE CLEANING MATERIALS AND TECHNIQUES 186
B. DERIVATION OF THE EQUATION RELATING INTRINSIC REJECTION
TO APPARENT REJECTION 187
C. ADDITIONAL DATA FROM MEMBRANE SELECTION STUDIES 189
D. FINAL REPORT - USE OF SULFONIC ACID MEMBRANES FOR
TREATMENT OF PULP AND PAPER WASTE STREAM 192
E. PRETREATMENT STUDIES 204
F. 3-STAGE PILOT SYSTEM COLOR REJECTION DATA 225
G. ANALYTICAL DATA FROM 25.4 MM DIAMETER
TUBULAR ASSEMBLY EXPERIMENTS 228
vm
-------
FIGURES
Number Page
1 Simplified flow schematic of Canton Mill pilot system 17
2 View of pilot system showing 1.9 m3 (500 gal) feed tank 18
3 View of pilot system showing Bauer Hydrasieve 19
4 View of pilot system showing Hoffman Vacu-matic
Vac-20 filter 20
5 View of pilot system showing Kisco deep bed media filters ... 21
6 View of pilot system showing spiral-wound module ultra-
filtration unit end view with filter location 22
7 View of pilot system showing spiral-wound module ultra-
filtration uni t 23
8 View of pilot system showing details of control panel
(right side) 24
9 View of pilot system showing details of control panel
(left side) 25
10 View of pilot system showing tubular ultrafiltration
test stand 26
11 Prefiltratibn section of pilot system 27
12 Design flow rates for 3-stage ultrafiltration pilot system .. 29
13 Flow schematic for 0.05 m diameter depth filter test
system 33
14 View of laboratory 51 mm (2 in) diameter deep bed filter .... 34
15 Simplified flow schematic of laboratory ultrafiltration
test system 36
16 Detail of stirred-cell testing apparatus 37
17 Flow schematic for stirred cell total recycle tests 40
18 Monthly average temperature and pH levels of kraft
pulp mill effluent streams 45
19 Monthly average total solids concentration for kraft
pul p mi 11 eff 1 uent streams 47
ix
-------
Number
20 Monthly average color concentration of kraft pulp
mill effluent streams 49
21 Monthly average suspended solids concentration of kraft
pulp mill eff 1 uent streams 50
22 Caustic extraction filtrate prefiltration data
(monthly averages) 52
23 Intrinsic color rejection versus flux during stirred cell
tes ts 55
24 Coated HFD membrane flux data obtained during
parametric studies 58
25 Coated HFD membrane flux and rejection characteristics
determined at constant operating conditions 59
26 The effect of polymer concentration on WRP membrane
water flux 63
27 The effect of polymer concentration on WRP membrane
process fl ux 64
28 The effect of polymer concentration on WRP membrane
col or rejecti on 65
29 Comparison of interpolyrner fixed charge and WRP membrane
f 1 ux dec! ine 68
30 Full-scale Vexar WRP11W33 spiral-wound module flux versus
time during laboratory total recycle test 70
31 "Typical" module flux performance during single module
tests 77
32 Module flux performance during total recycle tests 78
33 Module flux performance during pH adjustment tests 80
34 Module flux performance at low inlet pressure
(2.04 atm) [30 psig] 82
35 Module flux decline as a function of operating time 84
36 Module flux versus circulation rate 85
37 Flux history during 3-stage pilot system operation
(0 to 90 hours) 89
38 Flux history during 3-stage pilot system operation
(90 to 180 hours) 90
39 Flux history during 3-stage pilot system operation
(180 to 270 hours) 91
40 Flux history during 3-stage pilot system operation
(270 to 335 hours) 92
-------
Number Page
41 Performance characteristics of 12.7 mm diameter WRP
membrane assemblies 99
42 25.4 mm diameter tubular polysulfone membrane flux
versus time during 17% conversion test period 104
43 25.4 mm diameter tubular polysulfone membrane flux
versus time during 90% conversion test period 106
44 25.4 mm diameter tubular polysulfone membrane flux
versus time during 98% conversion test period 107
45 25.4 mm diameter tubular HFM membrane flux versus
time during 90% conversion test period 108
46 25.4 mm diameter tubular HFM membrane flux versus
time during 98% conversion test period 109
47 Tubular polysulfone membrane permeate quality and color
rejection during 1.2X and 10X concentration periods
(caustic extraction filtrate feed) 112
48 Tubular polysulfone membrane permeate quality and color
rejection during 50X concentration period
(caustic extraction filtrate feed) 113
49 Flow schematic for identification of samples from material
balance studies with caustic extraction filtrate
(25.4 mm diameter tubular polysulfone membranes) 116
50 Relative material balance for 50 times concentration on
pine bleachery caustic extraction filtrate using
25.4 mm diameter tubular polysulfone membranes 118
51 Analyses of pine caustic extraction filtrate at various
stages of ultrafiltration using 25.4 mm diameter
polysulfone tubular membranes 119
52 Ion concentration versus concentration ratio for ultra-
filtration of caustic extraction filtrate (25.4 mm
diameter tubular polysulfone membranes) 120
53 Ratio of ionic-chlorine/volatiles versus concentration
factor for ultrafiltration of caustic extraction
filtrate (25.4 mm,tubular polysulfone membranes) 122
54 Specific gravity versus concentration factor for ultra-
filtration of caustic extraction filtrate (25.4 mm
diameter tubular polysulfone membranes) 124
55 Proposed 3,790 m3 (1 MM gpd) UF system flow schematic 129
56 Proposed 3,790 m3 (1 MM gpd) system typical
subsystem outline drawing 130
XI
-------
TABLES
Number Page
1 Assays and Methods Employed During Experimental Program 42
2 Overall Summary of Feed Stream Characteristics 44
3 Results of Stirred Cell Screening Tests 54
4 Parametric Studies Test Matrix (25.4 mm Diameter x
1.52 m Long Tubular Membranes) 57
5 Performance of WRP Membranes (at 5.1 atm, 21°C) 61
6 Interpolymer Fixed Charge Membrane Test Results 67
7 Modules Employed During Single Module Tests 72
8 Chronology of Single Module Test Experience 73
9 Sequence of Cleaning Solutions Applied to 3-Stage
Pilot System After 320 Hours Operating Time 95
10 Summary of 3-Stage Pilot System Color Rejection and
Permeate Quality During Cuastic Extraction
Filtrate Processing 97
11 Initial Performance Characteristics of 12.7 mm
Diameter WRP Tubular Assemblies 100
12 Performance Characteristics of WRP Tubular Assemblies
wi th Tubulence Promoters 102
13 Flux Recovery for 25.4 mm Diameter Tubular
Polysulfone Membranes 110
14 Full-Scale System Design Cases 126
15 Ultrafiltration Section Design -- Case 1 128
16 Ultrafiltration Section Design ~ Case 2 134
17 Ultrafiltration Section Design — Case 3 135
18 Ultrafiltration Section Design — Case 4 137
19 Ultrafiltration Section Design — Case 5 138
20 Ultrafiltration Section Design — Case 6 140
xii
-------
Number Page
21 Ultra filtration Section Design -- Case 7 142
22 Ultra-filtration Section Design -- Case 8 143
23 Case 1 Design Capital Cost Summary 146
24 Case 2 Desi gn .Capi tal Cost Summary 153
25 Case 3 Design Capital Cost Summary 155
26 Case 4 Design Capital Cost Summary 157
27 Case 1 Design Operating Cost Data 160
28 Case 2 Design Operating Cost Data 161
29 Case 3 Design Operating Cost Data 162
30 Case 4 Design Operating Cost Data 164
31 Summary of Projected Economics for Design Cases
1 Through 4 165
32 Case 1 Design Capital Cost Summary with Future Reductions
in Ultrafiltration System Costs Considered 166
33 Case 1 Design Operating Cost Data with Future Reductions
in Ultrafiltration System Costs Considered 167
34 Case 5 Design Capi tal Cost Summary 170
35 Cast 6 Design Capital Cost Summary 172
36 Case 7 Design Capital Cost Summary 173
37 Case 8 Design Capital Cost Summary 174
38 Case 5 Design Operating Cost Data 175
39 Case 6 Design Operating Cost Data 176
40 Case 7 Design Operating Cost Data 177
41 Case 8 Design Operating Cost Data 178
42 Summary of Projected Economics for Design
Cases 5 Through 8 179
xi ii
-------
ENGLISH-METRIC CONVERSION TABLE
To convert from
To
Multiply by
Atmosphere
Cubic meter
Cubic meter
Cubic meter per day
Cubic meter per day
Cubic meter per sq. meter-day
Cubic meter per sq. meter-day
Ki1ogram
Kilo Watt
Meter
Meter
Square meter
Square meter
Pound per sq. inch 14-7
Cubic feet 35.31
Gallon 264.2
Gallon per day 264.2
Gallon per minute 0.183
Gallon per minute per sq. ft. 0.17
Gallon per sq. ft-day 24.39
Pound 2.205
Horsepower 1.341
Feet 3.281
Inch 39.37
Square feet 10.76
Square inch 1,550
xiv
-------
ACKNOWLEDGMENT
During the two years of the study, significant contributions were made
by a large number of vendors, EPA, Abcor and Champion International person-
nel. Exceptional contributions were made by those listed below:
Perry W. Bartsch, Vice President - Operations Manager of Champion Paper's
Canton Mill for giving permission to use the Canton Mill for this pilot
program.
.-t
Charles Seay, William Chapman and Herbert Pomfrey for operation of the
pilot system.
Ed Dyer and Dan Tate, Technical Control-Canton Mill for the daily ad-
ministration.
Don Grant, Arye Gollan, Leon Mir and Steve Jakabhazy of Abcor, Inc. for
many technical contributions.
Ed Hedrick (Purchasing) and Bobert Townsend and David Jesse (Accounting)
of Champion International for support which enabled the smooth operation of
the program.
JoAnette Coe and Marcia Smith of Champion International and Sharon
Collins and Jean Gilmartin of Walden Division of Abcor, Inc. for providing
the needed secretarial services throughout the program and final report
preparation.
Financial support for this program was provided through the Industrial
Environmental Research Laboratory of the U.S. Environmental Protection
Agency. The support and technical assistance of Kirk Willard, Ralph Scott,
John Ruppersberger and Jack Collins are gratefully acknowledged.
XV
-------
SECTION 1
INTRODUCTION
BACKGROUND
The 119 kraft pulp mills (1) in the United States produce about 85% of
the chemical wood pulp consumed. In an integrated pulping and paper making
operation, a substantial volume of waste water is discharged, typically
about 125 m3/metric ton of pulp (30,000 gal per ton of pulp). Of concern are
the pH, temperature, suspended solids, BOD and color loading of this
effluent. Conventional and generally available techniques are adequate in
most cases for waste treatment except for color removal. It may also be
noted that conventional waste treatment does not provide for water reuse,
and as such, is not conducive to eventual closed-loop operations.
Color bodies found in pulp and paper mill wastes are resistant to bio-
logical degradation. Consequently, new treatment techniques for color
removal are undergoing active development and actual plant scale demon-
stration. Processes developed include chemical precipitation (2-11),
including lime precipitation and alum precipitation, adsorption (12-15),
oxidation (16-19) and reverse osmosis and ultrafiltration (20-30). Rapid
infiltration (31) involving percolation of the effluent through the ground
has been demonstrated.
The reduction of effluent color by modification of pulping and
bleaching sequences has been the subject of extensive development.
Segregation of mill waste streams is often practiced and it is likely that
segregation of wastes by color content will eventually be required for
adequate waste treatment. For example, in Champion Papers North Carolina
Mill, about 60% of the total mill color effluent is present in about 3,790
to 7,580 m3/day (1 to 2 MM gpd) of the pine bleachery first-stage caustic
extraction filtrate. This flow amounts to about 2% to 5% of the total mill
effluent, yet removal of color from this stream could reduce total effluent
color by 60%.
The second most important controllable source of color in a kraft mill
is the pulp washing decker effluent. This waste stream is present in all
pulp mills, while the pine caustic extraction filtrate is found only in
mills producing bleached pulp. At the North Carolina mill, approximately
7,580 m3/day (2 MM gpd) of mixed pine and hardwood decker effluents are
currently discharged. This waste contributes about 20% of the total mill
effluent color.
-------
In an EPA-supported program (Project No. S800261) (32) Champion Inter-
national demonstrated an ultrafiltration process for kraft mill effluent
color removal on selected pulping and pulp bleachery process effluents. The
treatment of pine caustic extraction filtrate of a pulp bleachery and the
decker effluents from unbleached pulp washing were examined in pilot scale
studies. These two streams contribute about 80% to 90% of the color dis-
charged from an integrated kraft mill. The results of this study are
detailed in Champion's final report to EPA, dated May 1973 (32).
The major problem encountered in the original program involved membrane
cartridge reliability and life. The spiral wound cellulose-acetate membranes
employed were susceptible to cartridge plugging by suspended solids and
membrane surface fouling by fine colloidal matter. The plugging problem was
solved by use of cartridges with special flowchannel spacers. Trouble-free
operation was achieved with minimal feed prefiltration, at least from the
point of view of cartridge plugging.
More troublesome, however, was membrane surface fouling, which resulted
in low membrane flux (unit capacity) and necessitated frequent membrane
cleaning. Especially important in promoting membrane fouling were the feed
pretreatment steps required to protect the cellulose-acetate membranes from
chemical degradation. This pretreatment included temperature reduction
(from about 60°C to 38°C) and pH reduction (from about pH 11 to pH 7).
These two steps significantly reduced the stability of the colloidal material
in the influent, resulting in rapid membrane fouling.
It was concluded that a membrane which could operate at the raw in-
fluent pH and temperature would offer significant process improvements:
1) reduced membrane fouling, hence higher flux and less
frequent membrane cleaning; and
2) elimination of feed pretreatment, i.e. cooling and
neutralization.
These improvements would lead to major benefits in terms of both increasing
process reliability and decreasing process capital and operating costs.
After the original EPA-sponsored program, Champion conducted two
additional related studies. The first demonstrated the technical effective-
ness of spiral wound membrane modules with improved flow-channel spacers.
These membrane modules exhibited the ability to process the effluents of
concern without module plugging, and the accompanying mechanical failure,
observed in the prior EPA-supported program.
The second program was sponsored to evaluate new non-cellulosic
membranes developed by the Walden Division of Abcor, Inc. In this program,
it was clearly demonstrated that membrane materials suitable for treatment
-------
of the raw effluents,without pretreatment were becoming available. The
results on membrane flux, color body removal effectiveness and cleaning
ability were clearly superior to those obtained with the cellulose-acetate
membranes employed earlier. In a parallel membrane development program at
Columbia University, another membrane suitable for treatment of pulp mill
effluents had been developed.
Based on the prior EPA program results, the module mechanical studies
and the availability of these new membranes, Champion applied for a demon-
stration grant renewal, the results of which are reported here.
The program described in this report encompasses a two-year study for
the development and demonstration of ultrafiltration as a process for color
removal from kraft mill effluents. The site of the project was Champion
Papers' Canton* North Carolina Mill, which is a typical integrated kraft
mill.
The program included four major elements:
1. Evaluation and selection of non-cellulosic membrane
materials and module geometry.
2. Selection and evaluation of feed pretreatment alternatives.
3. Modification of the 37.9 m3/day (10,000 gpd) ultrafiltration
system originally built for EPA project no. S800261.
4. Evaluation of non-cellulosic membranes in the modified
37.9 m /day (10,000 gpd) pilot system.
Additional studies were undertaken to develop: estimated operation and
capital costs as well as space and energy requirements for several flow
size units; bases for disposal of the concentrated wastes; and, the
potential for reuse of the permeate water.
Briefly, the project objectives have been threefold.
1. To demonstrate with commercially available equipment the
effectiveness of ultrafiltration to reduce color in the
first-state pine bleachery caustic extraction filtrate and
pulp washing decker effluents to low levels.
2. To demonstrate the potential for reuse of purified effluents
and means of disposal of the concentrated wastes produced
by the membrane process.
3. To demonstrate that the estimated process economics, based
on non-cellulosic membrane materials, will be attractive in
comparison with other color abatement processes available
to the paper industry.
-------
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, a feed solution is fed into the membrane unit, where water and
certain solutes pass through the membrane under an applied hydrostatic
pressure (permeate). Solutes whose sizes are larger than the pore size of
the membrane are retained and concentrated (concentrate or retentate).
The pore structure of the membrane acts as a molecular filter, passing
some of the smaller solutes and retaining the larger solutes. The pore
structure of this molecular filter is such that it does not plug since
larger solutes are rejected at the surface and do not penetrate the membrane.
Thus, concentration of specific solution components can be achieved.
Considerations important for determining the technical and economic
feasibility of ultrafiltration as applied are the rate of solution trans-
port through the membrane (flux) and the separation efficiency (rejection).
Other important factors include the membrane fouling rate, membrane clean-
ability, membrane material of construction and its physical properties and
membrane geometry.
Ultrafiltration membranes that withstand almost any aqueous application
are now being produced or developed in the membrane industry (33). This is
significant in application of ultrafiltration to the process wastes under
discussion since the streams treated in this study are hot (38 to 60°C) and
have high pH (10-13). The availability of a membrane material which is
stable at these conditions can reduce the pretreatment costs for the
ultrafiltration processes.
Choice of the specific membrane material for use with pulp mill waste
is complex but, in general, is based on membrane pH and temperature
stability; controllability of pore size, to provide desired flux and
rejection; fouling characteristics; cleanability and available geometric
forms of modules.
At present, ultrafiltration membranes are available in five principle
geometric configurations: spiral wound, tubular, plate and frame, hollow
fiber, and dynamic membranes. The spiral wound unit is a double membrane
sheath which is wrapped around a central permeate removal tube. Various
types of separators (spacers) are in use to separate the membrane sheet to
allow for concentrate flow through the spiral in a longitudinal direction.
A tubular unit consists of a porous tubular substrate which is cast
coated, usually on the inside, with the membrane material.
-------
The plate and frame unit is similar in construction to the traditional
plate and frame filter. Effluent is pumped into a thin space between
membrane covered plates continuously. Permeate is removed from the plates
and concentrate from the space between plates.
Hollow fiber units consist of bundles of small diameter membrane tubes,
hence providing large membrane area per cubic foot of membrane module volume.
Flow of the material to be ultrafiltered is usually on the inside of the
tubes. Permeate is removed from the outside of the tubes.
The dynamic membrane differs from the previous geometries in that the
membrane is formed on a porous substrate by coating from suspension with a
hydrous oxide.
The choice of the membrane geometry is significant in determining both
the capital and operating cost of an ultrafiltration process. The advantages
and disadvantages of membrane configurations have been discussed elsewhere
(34). A review of cost-performance data has shown that spiral wound
membranes hold promise of being lowest in costs. Some process limitations
exist for membranes in this configuration. For this application, the
potential economic advantages make it desirable to learn how to design this
configuration so that it is adequate for processing both pine caustic
extraction filtrate and decker effluents.
Membranes of porosity similar to Kodak HT-00 have been found to yield
rejections in the 90-94% range for caustic extract, and 95-98% range for
the decker effluents.
WASTE STREAM CHARACTERISTICS (NORTH CAROLINA MILL)
The waste streams which have been examined are the pine bleachery
caustic extraction filtrate and the pine and hardwood pulp washing decker
effluents. These streams are all highly colored. At the Canton mill, the
caustic filtrate has an average color of about 20,000 c.u. (Color Units:
basis Pt-Co Standard), the pine decker about 6,000 c.u. and the hardwood
decker about 11,000 c.u.
Color in the deckers is predominately lignin materials which have been
dissolved from'the wood matrix in the pulping operation and has molecular
weights predominately in the 4,000 to 8,000 range. Color in the caustic
filtrate is due primarily to the shards of lignin and chlorinated lignin
produced through oxidation in the bleaching operation. These materials
appear to be predominately in the range of 400 to 4,000 molecular weight.
The pH's of the decker effluents are in the range of 10 to 11. The
caustic filtrate pH is 11 to 12. The temperature of all three effluents
are in the range of 38 to 60°C (100° to 140°F).
-------
The suspended solids content of the caustic extract varies from about
50 ppm to 500 ppm. Suspended solids content of the pine decker is 50 to
250 ppm and for the hardwood decker, it is 100 to 500 ppm. The suspended
solids content for all three streams is similar in that there are included
varying amounts of pulp and clay materials. About one-half of the suspended
solids are less than 10 microns in size.
There is another similarity in the solids content of these streams.
Each stream contains perhaps 100 ppm of a micellar polysaccharide material
which, under adverse conditions, can agglomerate and appear as a slimy
foul ant on the surface of membranes. This type of fouling can substantially
reduce the membrane flux and necessitates membrane cleaning operations.
In the past several years, as a result of the intensive pulp and paper
mill water conservation programs, this potential micellar fouling problem
has been aggravated. The North Carolina mill has traditionally operated on
an average of ten uses of water prior to discharge. It is now approaching
twelve uses of water. To accomplish this, there is increased replacement
of fresh water makeup for various operations by reuse of streams such as
paper machine white water. These streams introduce additional micellar
species such as starch and sizings which intensify the potential fouling
problem.
The total solids in the streams vary but, in general, are one-quarter
to one-third color body materials. The caustic extract has on average
about 7,000 ppm of total solids with the predominant ions being sodium,
calcium, hydroxide, chloride and sulfate. The pine decker has an average
total solids of about 2,400 ppm and the hardwood decker about 3,000 ppm. In
ultrafiltration of these streams, the multivalent ions such as aluminum and
sulfate are usually rejected to some small extent. However, this rejection
concentration is small enough that even at 100 times concentration of these
streams, the osmotic pressure increase is small and it does not interfere
with the low pressure requirements for membrane flux maintenance.
ULTRAFILTRATION PROCESS CONSIDERATIONS
The ultrafiltration processes for each of these pulp mill streams
consists of a pretreatment system, a membrane system and a disposal system.
The pretreatment system is designed to remove fibrous material and coarser
suspended solids from the stream to be treated. A number of commercially
available devices can be used for this purpose. With presently available
membrane materials, neither stream pH or temperature (below 85°C) need to
be controlled. Practical methods for eliminating slime forming material
from these streams have not been demonstrated.
Technical feasibility has been demonstrated on a pilot scale for the use
of pretreatment systems and membrane systems to remove color bodies by
ultrafiltration from bleachery pine caustic extraction filtrate and pulp
-------
washing pine and hardwood decker effluents. The pilot scale demonstrations
have been conducted at pressures of 3.4 to 6.8 atm (50 to 100 psig) with
properly pretreated flow streams. Membrane flux levels of 0.82 to 3.28 m3/
m2-day (20 to 80 gfd) have been demonstrated with color rejections in the
90-98% range at concentration ratios up to 100 times.
Use of ultrafiltration processes on decker effluents could permit
closed cycle operation of the pulp washers with removal of its color con-
tribution to total mill effluent. In addition, it would permit conser-
vation of energy, salts, and water. Disposal of the concentrate from the
decker effluents would be straightforward. Concentration to the proper level
of organic material would allow for disposal of the concentrate in the weak
black liquor streams. The inorganic chemical and organic material heat
value recovery and reuse from black liquor are fundamental to kraft process
economics. The permeate stream would be reused as make-up water in pulp
processes.
Disposal of the concentrate from the caustic extraction filtrate is
more complex because of the organic and inorganic chloride content. This
concentrate may be disposed of by incineration in the lime kilns or with the
black liquor, landfilled with lime sludge, or used for special chemical
properties. The permeate can be recycled to the bleachery substantially,
excess over this use can be used for low grade water on the mill site.
As described in detail in the following, the technical feasibility of
color removal from kraft mill process effluents by ultrafiltration processes
has been demonstrated. Projected costs and peripheral considerations of use
of such processing in existing mills are attractive vis-a-vis alternative
available technologies.
-------
SECTION 2
CONCLUSIONS
This demonstration program has shown that ultrafiltration is a techni-
cally viable treatment process for color removal from kraft mill effluent
streams. Processing takes place at natural stream temperature and pH levels
due to the development of polysulfone (non-cellulosic) membrane formulations.
Reuse of the ultrafiltered water (permeates) in pulping and bleaching oper-
ations appears practical and could provide water, material and energy con-
servation. The color concentrates (retentates) may be beneficially used
(decker effluents) or treated in existing disposal systems (caustic fil-
tration extract).
Tubular assemblies are the preferred ultrafiltration module geometry.
This configuration exhibited high, stable flux performance, was readily
cleanable and showed consistent color rejection. Spiral-wound cartridges,
on the other hand, were plagued by severe fouling due to slime layer forma-
tion, had poor flux recovery upon cleaning and exhibited widely varying
color rejection.
The economics of tubular ultrafiltration as projected, are cost-competi-
tive with alternative treatment processes.
These general considerations are supplemented by the following specific
findings:
1. FEED CHARACTERISTICS
- Solids content, color, pH and temperature of the caustic
extraction filtrate stream, pine decker effluent and hardwood
decker effluent varied randomly from month-to-month. Wide
fluctuations were observed on a daily basis. However, since
ultrafiltration systems are generally insensitive to shock
loading, and since field program data indicates no adverse
effects from day-to-day feed variation no feed equalization
is deemed necessary prior to processing.
- Due to equipment changes, process changes or water conser-
vation measures the nature of the three effluents has
changed over the past four to five years. These types of
changes can be expected to continue in the future but should
pose little, if any problem for a tubular ultrafiltration
system.
-------
2. NON-CELLULOSIC MEMBRANE DEVELOPMENT
The conclusions presented in this subsection are based on
laboratory studies of candidate membranes in flat sheet
configuration. Laboratory studies are useful in screening
membranes relative to each other with respect to various
parameters. Data from such studies can be extrapolated to
membranes in other geometric configurations.
- Polysulfone (WRP) membranes are preferred over specially
coated type HFD and HFM membranes (commercially available
from Abcor, Inc.) for this application. The best rejecting
HFM membrane exhibited 98.5% rejection with an associated
average flux of 0.86 m3/m2-day (5.1 atm) [21 gfd (75 psig)].
Two HFD membranes had 99% intrinsic color rejection but low
flux. The majority of WRP membranes had color rejections of
99% to 99.5% with flux levels as high as 1.97 m3/m2-day
(48 gfd).
- HFD and HFM membrane coating formulations degraded with
time. This factor reinforced the preference for WRP-
series membranes.
- High rejection-moderate flux ultrafiltration membranes can
readily be made from polysulfone-based casting solutions.
An entire series of membranes were cast from solutions con-
taining different polymer concentrations with and without
non-solvent, and with and without surfactant in the gelation
bath. The preferred WRP membrane had a sol vent :non-solvent:
polymer weight ratio of 52.7:26.3:21.
- Interpolymer fixed charge membranes (prepared by H. Gregor,
Columbia University) were compared with WRP membranes in
both short and long term tests. The short term tests
(3-hours) showed no practical differences between these
membrane types. In long-term tests (65 hours) both membrane
types exhibited essentially identical rates of flux decline.
However, the WRP membranes showed higher color rejection and
higher flux levels. Thus, the WRP membranes were selected
for field evaluation.
3. FIELD EXPERIENCE WITH SPIRAL-WOUND MODULES
- Extensive processing of caustic extraction filtrate on a once-
through basis was performed with a single module test stand.
Sharp flux decline was consistently observed with 24-hour
flux levels in the 0.12 to 0.41 m3/m2-day (3 to 10 gfd)
range. Since no concentration occurred and since no increase
in pressure drop was observed across the modules the flux
decline pattern indicated a steady fouling of the membrane
surface by some species in the caustic extraction filtrate.
This foul ant was subsequently identified to be a "slime"
-------
layer composed of volatile organics, kaolinite clay, starch,
titanium dioxide and carboxylic acid salt. The major com-
ponents of the fouling "slime" layer come from recycle of
white water from the paper mill back to the pulp mill.
- Extensive testing in the areas of spiral-wound module pre-
treatment and prefiltration were ineffective in eliminating
the foul ants from the feed stream. Also, attempts to improve
spiral-wound module flux performance by exploring a range of
operating conditions proved unsuccessful. These results,
coupled with significant flux recovery problems upon module
cleaning, indicate that further spiral-wound module develop-
ment will be required before this is a viable ultrafiltration
membrane geometry for kraft mill effluent processing.
4. FIELD EXPERIENCE WITH TUBULAR ASSEMBLIES
- Polysulfone tubular assemblies in both 12.7 mm (0.5 in) and
25.4 mm (1 in) diameters exhibited high, stable flux perfor-
mance during several hundred hours of caustic extraction
filtrate processing. Operating on a once-through basis, flux
for a 12.7 mm diameter tube (WRP formulation) stabilized at
1.64 to 2.26 m3/m2-day (40 to 55 gfd) for over 450 hours.
No detergent cleaning or mechanical cleaning was necessary.
25.4 mm diameter tubular assemblies (commercial formulation)
operated on caustic extraction filtrate at 3 concentration
levels averaged these flux levels:
Concentration Average
factor Conversion m^/m^-day (gfd)
1.2X 16.7% 2.87 (70)
10X 90% 2.26 (55)
50X 98% 1.03 (25)
Again, flux was stable over hundreds of operating hours.
Flux recovery was rapid with a dilute detergent wash. The
improved flux performance observed with tubular assemblies
as compared to spiral-wound modules is a function of the
higher superficial velocity over the membrane surface achiev-
able with the tubular configuration. This leads to more
turbulent flow and minimizes the gel ("slime") concentration
layer at the membrane/liquid interface.
- Color removal by the tubular assemblies was exceptional,
averaging 97% to 99% on an individual stage basis. Overall,
on a mass discharge basis (kg color removed per day), 91%
color removal is projected for processing of the caustic
extraction filtrate stream to a 50X concentration.
10
-------
5. PROJECTED ECONOMICS FOR FULL-SCALE SYSTEMS
- Eight full-scale system design cases were analyzed in terms of
both capital and operating costs. For tubular systems treating
3,790 and 7,580 m3/day (1 and 2 MM gpd) of either caustic ex-
traction filtrate or decker effluent total capital investment
ranges from $2 to 4 MM. Treatment costs, based on 727 metric
tons of pine pulp produced per day (800 tons/day), range from
$1.51 to $2.52 per metric ton ($1.36 to $2.27 per ton) for
caustic extraction filtrate. The treatment costs range from
$0.76 to $1.26 per metric ton ($0.68 to $1.13 per ton) for
decker effluents on a 1,318 metric ton/day (1,450 ton/day) basis.
- Future technological advances are expected to reduce large-
scale tubular ultrafiltraton system capital costs. Using future
cost estimates, a 10% cost savings over current capital cost
projections was calculated. The uncertainties associated with
the future costs of labor, materials and equipment may, however,
reduce this projected savings.
- Economic projections for spiral-wound module ultxafiltration
systems were based on idealized systems not attainable with
today's technology. Using idealized design flux values capital
investments of $1 to $2.1 MM are projected. Treatment costs
are $1.09 to $1.89/metric ton ($0.99 to $1.72/ton) of pulp for
caustic extraction filtrate processing and $0.55 to $1.89/
metric ton ($0.50 to $0.87/ton) of pulp for decker effluent
processing.
6. WATER REUSE POTENTIAL
- The permeate from the ultrafiltration unit treating caustic
extraction filtrate will constitute about 98% of the feed
stream. This effluent will have low color, essentially no
suspended solids and will have very low heavy metal content.
In addition the permeate will have a high pH and be at process
temperature.
The permeate, with its physical and chemical attributes, should
be an adequate water makeup stream for use in the bleachery
processes. The high pH, and reduced buffering capacity should
allow for lower new caustic requirements. The high temperature
should reduce the system energy requirements. Because of the
absence of suspended solids and the decreased heavy metal con-
tent, the permeate should reduce spray head and other scaling
problems.
It is believed that at least half of the permeate can be used
in a bleachery recycle mode and that the savings in chemicals,
water and energy from such use will have beneficial effects in
reducing the net cost of the operation. Permeate which is
11
-------
excess can be admixed with mill input fresh water without dis-
cernable effects upon the fresh water quality. This is be-
cause the small permeate volume would be diluted 25 to 50 times
by the larger fresh water input.
- The concentrate from ultrafiltration of caustic extraction fil-
trate has no direct value in pulp mill operations. Disposal
of this material would be site-specific. In those installa-
tions equipped to remove chlorides from black liquor systems
this concentrate could be flowed to the weak black liquor
system with some small gain in energy recovery. Alternatively,
the concentrate could be combusted in a typical modern lime
kiln without noticeable effect, especially because of the low
sodium content. Some mills which do not have sufficient lime
kiln capacity dispose of the excess lime sludge off-site. Be-
cause of the high pH of this sludge and the relatively small
volume of the concentrate, the concentrate could be added to
the lime sludge and carried to landfill as insolubilized
calcium salts.
- For ultraf iltration of pine and hardwood pulp washing decker
effluents both permeate recycle to the pulping system for
makeup water and concentrate recycle to the weak black liquor
system are projected with concomittant cost reduction due to
chemical, water and energy recovery values.
7. COMMERCIAL RELIABILITY
Further information is needed in several areas to establish the
commercial reliability of the process:
- A larger scale process demonstration involving substantial
numbers of modules and long term continuous operation is
necessary to assess the manufacturer's capability to reproduce
modules with requisite characteristics, and also, to acquire
statistical data on module durability and useful life.
- Quantities of permeate and concentrate are needed to demon-
strate reuse or disposal of these streams on a reasonable scale.
- The design and operation concepts used in this study were
simplified for demonstration of the process on a limited scale.
Validation of the process system projections is necessary using
a prototype installation which models a full scale configur-
ation, controls, operating protocols, cleaning systems, dis-
posal systems, shock loading, etc. al.
- Operation of a prototype demonstration plant could provide in-
formation needed to "harden" the present projections of capital
and operating costs, space requirements, operating and mainten-
ance manpower, disposal of permeates and concentrates, shock
response and the host of other considerations on which ex-
panded data is needed to firm the concepts of commercial
reliability.
12
-------
SECTION 3
RECOMMENDATIONS
The major recommendation resulting from this study is field demon-
stration of tubular polysulfone membrane treatment of kraft mill effluent
streams on a significant scale. A 190 to 379 m /day (50,000 to 100,000 gpd)
staged pilot unit should be installed at a pulp mill and operated for at
least 12 months. Circulation flowrate and inlet pressure requirements for
each stage should be optimized to reduce overall system power consumption.
Cleaning frequency and duration should be detailed and membrane life
(mechanical failure, flux recovery with time) carefully monitored.
Full-scale system control and monitoring requirements should be con-
sidered as part of this demonstration and innovative engineering techniques
for tubular ultrafiltration systems of this magnitude should be explored. In
addition, the conceptual design and economic projections presented in this
study should be updated.
Product streams, both permeate and concentrate, should be recycled within
the mill to verify their projected reusability. Detailed chemical analyses
for material balance studies should be continued.
Additional, more basic recommendations are:
Study of tubular module design to produce lower cost, more compact
systems. For systems of the magniture required for kraft mill effluent
streams - 3,790 to 7,580 m/day (1 to 2 MM gpd) - significant cost
savings may be realized.
Spiral-wound module development should parallel tubular module
design studies. Improvement in spiral-wound module feed-side
spacer designs could increase turbulent flow and reduce (or elimi-
nate) slime layer build-up. In such a case lower-cost, more com-
pact ultrafiltration treatment systems would become available.
13
-------
SECTION 4
PROGRAM OVERVIEW
The previous EPA sponsored documentation study report (32) on the use
of ultrafiltration processes to control kraft mill effluent color presented
a number of recommendations for subsequent work. In that study the techni-
cal feasibility and economic promise of such processes was demonstrated.
The major recommendations emphasized the need for studies to improve and
verify the economic bases for the process as new membrane capabilities
were developed.
Studies conducted subsequent to 1973 demonstrated that the development
of non-cellulosic membranes by the industry was at a point that promising
non-cellulosic membranes were becoming commercially available and would be
of value in this application.
With the support of EPA this program was undertaken as described in
the foregoing.
The levels on which the program planning was done were principally:
1. Non-eellulosic membranes were commercially available which
could operate with long life and provide:
- operation at process stream temperatures and pH thus
reducing costs of acid and cooling for pretreatment and
also reducing the potential for stream micellar
agglomeration to foul.
- operation at fluxes 2 to 4 times those obtained with
previously used cellulose acetate membranes thus pro-
viding the basis for a plant of given size flow capacity
at a small fraction of the required membrane area. Thus,
even though the cost per unit area of the new membrane
was high, the capital cost of the plant would be main-
tained at that projected in the previous estimates
despite inflation.
2. The commercially available non-cellulosic membranes could
be available in spiral wound configurations with open
channel spacers which would allow:
- low cost per unit area membrane modules.
- low pumping energy requirements compared to other
membrane configurations.
14
-------
- reduction of fouling due to higher superficial velocity
over the membrane surface.
- ease of cleaning.
The program was initiated with a laboratory screening of a number of
non-cellulosic membrane systems which could be commercially available. It
was decided to primarily study the pine bleachery caustic extraction filtrate
as the feed to be processed because it constitutes the major color con-
tributor in a kraft pulp mill system and, also, because all previous studies
have taught that the pulp washing decker effluents performed in a similar
manner as regards membrane flux but with even more effective color rejection.
A commercially available membrane system was selected and demonstrated
in the laboratory. When the system was installed in the mill using a "live"
feed, however, severe fouling problems were observed.
It was determined that these fouling problems were a result of two
interrelated factors:
1. As is increasingly true of most mills, in the interests of
water conservation the Canton mill was using increasing amounts
of "dirty" water (paper machine white water) for replacing
fresh water makeup in the pulping and bleaching system. This
introduced a greater amount of fouling material which the pre-
treatment system, as defined, did not remove.
2. Open spacer spiral-wound modules were not commercially avail-
able and the available Vexar spacer spiral-wound modules tended
to build heavy slime-like polarized films that rapidly reduced
the effective membrane flux.
Because of the very attractive economic advantages inherent in the use
of spiral-wound membrane systems it was decided to study the possible
methods of reducing the foul ant potential of the feed streams to a level
where the available spiral-wound membranes could operate effectively.
Extensive studies were undertaken on available mechanical filtration equip-
ment singly and in combination for this purpose. Studies were also con-
ducted on combinations of these equipments together with a number of natural
and synthetic flocculants and agglomerants. It was concluded after months
of study that no practical solution would be available, within-the time
constraints of this program, which would permit economic control of this
fouling potential in the mill stream.
The remainder of the program was devoted to demonstration of the poten-
tial of these non-cellulosic membranes in tubular form (commercially
available) for kraft mill effluent color removal. The effectiveness of
this geometric form of the selected membranes was demonstrated to be
practical both technically and on the basis of the projected economics.
The following report sections present discussions of the scope and
details of this study.
15
-------
SECTION 5
EXPERIMENTAL EQUIPMENT AND PROCEDURES
PILOT PLANT
General
The pilot plant for kraft pulp mill effluent treatment studies was
designed as a 3.79 m3/day (10,000 gpd), three-stage system utilizing Vexar
spacer spiral-wound membrane modules. A simplified system flow schematic
for the pilot plant is shown in Figure 1. The pilot system contained two
major areas: a pretreatment and prefiltration section and a 3-stage ultra-
filtration section. Photographs of various system components are shown
in Figures 2 through 10. The operation of these components is discussed
below.
Pretreatment and Prefiltration Section
The pretreatment and prefiltration section of the pilot plant (see
Figure 11) was designed to remove suspended solids from the feed streams.
Suspended solids (including fibers) reduction was necessary to prevent
plugging and reduce fouling of the spiral-wound membrane modules. Pretreat-
ment (investigated during a portion of the field program only) consisted
of polymer addition to the feed stream to flocculate suspended particles
and aid in their removal by the system prefilters. Components of the pre-
treatment system were a polymer solution tank, a Milton-Roy metering pump,
a mixing tank, and a Hoffman Vac-20 Vacu-matic filter. The polymer solution
was pumped by the metering pump to the mixing tank where it was combined
with the feed stream. Flow from the mixing tank to the Hoffman filter was
by gravity. The Hoffman filter contained a pump to transfer the filtrate
to the system feed tank.
The Hoffman Vac-20 Vacu-matic filter is a flat-bed vacuum filter. It
consists of an endless metal conveyor belt which supports the filter's media
and passes it over a vacuum chamber. During filtration, as the filter cake
increases in thickness and filtrate flow decreases, the Vac-20 automatically
indexes fresh filter media. The filter media are supplied in roll form and
are available with a wide range of effective pore sizes.
16
-------
Hoffman
Vac-20
filter
Stage 3
Permeate
3-stage ultrafiltration system
Figure 1. Simplified flow schematic of Canton Mill pilot system.
17
-------
Figure 2. View of $f?«t systew showing
.9 ®3 (500 gal) feed tank.'
18
-------
Figure 3. yiew of pilot system showing
Bauer Hydrasleve.
-------
ro
o
View of pilot system showing
Hoffman ¥acu-matic Vac-20
filter.
-------
View of pilot system showing
K1sco deep bed media filters
-------
¥1ew of pilot sysfc^
spiral-wound iwwlwle ultra
ffltrstion unit end view
22
-------
Im Pressure Shut
off Switch
Figure 7. View of pilot system stowing
spira1-»ourui module uttra-
fnitration unit.
t
-------
Concentrate Sole
noid Valve Timor
Figure 8. ••• View of pilot system
details of control panel
(right side).
24
-------
Feed and Redrcu-
Flow Meters
Module Inlet
Outlet
Gauges
t, 1ft«an tf -fllot §ystero
details of control
{left side).
25
-------
Single Stage
f wises in Serie
Figure. 10.- View of pilot system showing
tubular ultroffltr'ation test
stand.
26
-------
ro
•> Backwash
Shallow well pump connectec
to Kisco valves to
pressure to keep them
closed
*v a$
1 I
Figure 11. Prefiltration section of pilot system.
1
Velmac
FTTter
_7\-
To flow
meter and
first stage
Cuno pump
FiTEiF
Broughton
Filters
Solenoids and pressure
switch (PS) connected
to backwash controls.
Backwash Drain
-------
The equipment in the prefiltration area consisted of a Bauer Hydrasieve,
a system feed tank, a Kisco deep bed media filter, a Broughton basket type
filter, a Velmac disc filter, and a Cuno cartridge filter. The process flow
pattern was designed to utilize these filters in series (in the order listed
above). The Bauer Hydrasieve removed fibers from the feed stream and was
located upstream of the system feed tank. The Hydrasieve was used for part
of the program but was discontinued when Kisco personnel indicated that the
fiber mat build-up on the Kisco media would aid in small suspended particle
removal.
The 1.9 m3 (500 gal) system feed tank was equipped with a level control,
temperature control, and a mixer. Process stream temperature for the total
system was controlled at this point. From the feed tank the process stream
was pumped to the Kisco deep bed media filters. These filters contained a
0.76 m (30 in) bed of Filter AG (granular non-hydrous aluminum silicate)
media. Two Kisco filters were piped in parallel; one filter on the process
stream and one backwashing or on standby. The effluent from the Kisco
filters was passed through the Broughton filters, the Velmac filter and the
Cuno filter, in series. While the Broughton filters were used throughout the
program, the Velmac and Cuno filters had only limited use. When not in use
the Velmac and Cuno cartridges were removed from their housings. Following
the Cuno filter the process flow entered the ultrafiltration section of the
pilot system.
Ultrafiltration Section
The ultrafiltration section of the pilot system contained three re-
circulated stages in series. Each stage consisted of a module housing,
valves, piping, and a pump. Two Vexar spacer spiral-wound membrane modules
could be positioned in each housing. These 0.1 m (4 in) diameter x 0.9 m
(36 in) long modules each contained approximately 3.7 m2 (40 ft2) of
membrane area. Feed recirculation was required to maintain sufficient flow
rates through the modules. Also, a single pass system would not give
adequate process stream circulations (see Figure 12 for design flow rates).
The circulation pumps on each stage were piped with a recycle line,
high and low pressure switches, and valves for individual stage control.
The total flow through each stage, ratio of fresh feed to recirculated feed,
and pressure to the stage were all controlled. The high pressure switches
protected membrane modules from over pfessurization. The low pressure
switches shut the system down for pump protection in case of system upset.
As stated earlier the feed tank in the prefiltration section had temperature
control and a high temperature shut-off switch.
SINGLE MODULE TEST STAND
The single module test stand was designed to test pretreatment and
prefiltration efficiency on the process feed stream and to evaluate module
performance characteristics and cleaning requirements. The test module
28
-------
38.2 m /day
Feed 40.9
111 (7.5
2.7 m3/day
(0.5 gpm)
m /day ^
gpm) (1
\
(
/ ^
4 gpm) '
54.5
m3/day
(10 gpm)
IJ7.2
' N^
10.5 gplfQ
(7 gpm)
First Stage
1 24
^ (4
51 .8 m3/day
(9.5 gpm)
Second Stage
13.1
^ (2-!
54.5 m3/day
(10 gpm)
Thi**H ^ttano
2.3-
V $•*
54.5
(10 gpm)
5 m3/day
.5 gpm)
68.1
jp3/day N
(12.5 gpn
5 m3/day
5 gpm)
54 .'9
m3/day
(10.07 gprr
i m3/day
»3 gpm)
/
0
0.38
0 (o.
16.4 m3/day
(3 gpm)
^
07 Ifpm)
Figure 12. Design flow rates for 3-stage ultrafiltration pilot system.
29
-------
shell held a single 0.1 m (4 in) diameter spiral-wound module. Mill effluent
was processed through the module on a once-through basis exposing the
module to high volumes of fresh feed. The stage-1 pump was used as the feed
pump for the test module stand. Temperature, pressure and flow rates
similar to proposed pilot system operating conditions could be readily met
with this system.
TUBULAR MODULE TEST STAND
The tubular module test stand served two functions:
— Determination of tubular membrane performance characteristics; and,
— Operation as a single stage recirculating pilot system.
To evaluate tubular membrane fouling properties the system was operated as
a single pass system. The main pilot system prefiltration system was used
and the Stage 1 pump was the feed pump for the test stand.
The single stage recirculating pilot system consisted of 2, 3, or 4
tubular membrane assemblies in series. Both 12.7 mm (0.5 in) diameter and
25.4 mm (1 in) diameter tubular assemblies were tested. The 12.7 mm (0.5 in)
diameter tubes were tested in various lengths, with and without turbulence
promoters. The 25.4 mm (1 in) diameter tubes were 1.5 m (5 ft) long and
contained 0.1 m^ (1.1 ft^) of membrane surface area.
The tubular system was operated at 136 m3/day (25 gpm) and 109 m^/day
(20 gpm) recirculation rates. At a 1.2X concentration this single stage
tubular membrane test stand had a 27.3 m3/day (7,200 gpd) capacity.
MEMBRANE CLEANING PROCEDURES
MRP Membrane Spiral-Wound Modules
The cleaning procedure for WRP membrane spiral-wound modules found to
be most effective consisted of three steps:
Step 1: Flush the system with 49°C (120°F) water.
Step 2: Prolonged recirculation of a cleaning solution through
the system.
Step 3: Flush the system with water.
A description of each step follows:
Step 1: The flushing of the system with 49°C (120°F) water was an
important part of cleaning. The fouling layer was substan-
tially removed by this warm water flush. When cold water
was used, the fouling layer apparently set-up and was much
harder to remove in Step 2. Typically, this warm water
flush was done with a 1.7 atm (25 psig) pressure on the inlet
30
-------
side and 27.3 to 43.6 m3/day (5 to 8 gpm) flow rate. The
49°C water was fed from the cleaning tank and sewered after
going through the system. The operating procedure was to
fill the 0.38 m3 (100 gal) cleaning tank with water, heat to
49°C with live steam, and then pump through the system until
the cleaning tank went dry. This gave a 10-15 minute water
flush.
Step 2: The most effective cleaning composition was a solution of
0.5% NaOH and 0.25% EDTA. This cleaning solution was pumped
through the system and totally recycled back to the cleaning
tank. The normally used cleaning conditions were:
a) 1.7 to 3.4 atm (25 to 50 psig) inlet pressure
b) 43.6 to 65.4 m3/day (8 to 12 gpm) recirculation rate
c) 49 to 54°C (120 to 130°F)
d) one hour duration
The cleaning system pump was incapable of pumping the 43.6 to
65.4 m3/day (8 to 12 gpm) desired recirculation rate. The high
recirculation rate was obtained by using the Stage 1 pump in
series with the cleaning system pump.
Step 3: The cleaning solution was flushed from the system with warm
water. Normally 49°C water was used, but cold water seemed
to be just as effective. The final flush was to remove all
the dirty cleaning solution from the system. The flushing
water was pumped from the cleaning tank through the system and
then sewered. Typically, this was done at 27.3 to 43.6 m3/day
(5 to 8 gpm) with an inlet pressure of 1.7 atm (25 psig).
After cleaning the membrane modules, a water flux was run. If the flux
was less than 1.64 m3/m2-day (40 gfd) Steps 2 and 3 were repeated. A 1/2%
solution of Ultraclean was usually effective at this point in the repeat of
Step 2.
Appendix A contains a summary of cleaning materials used.
WRP Membrane Tubular Assemblies
An almost total lack of fouling problems in MOOO hours operation with
the WRP membrane tubular assemblies resulted in little cleaning information
for this membrane geometry. The cleaning procedure used was essentially the
same as for the spiral wound modules:
Step 1: Flush system with 49°C (120°F) water.
Step 2: Recirculation cleaning with 0.5% Ultraclean detergent
for two hours .
Step 3: Flush system with water.
31
-------
In step 2, the system was recycled at 136 m3/day (25 gpm) flow rate with
a pressure of 3.4 atm (50 psig) and a temperature of 49°C. Recirculation
time was two hours but probably was in excess of the time necessary to clean
the membranes.
This three-step cleaning procedure resulted in water flux at 49°C of
5.33 to 6.97 m3/m2-day (130 to 170 gfd) for the clean WRP membrane tubular
assemblies.
51 MM DIAMETER DEPTH FILTER
A 51 mm (2 in) diameter laboratory-scale depth filter was tested at the
Canton Mill to assess the suspended solids removal efficiency of various
filter media. The media evaluated were:
Media Bed depth
- Anthracite coal 0.46 m (18 in)
Silica sand 0.25 m (10 in)
Garnet sand 0.05 m ( 2 in)
- Garnet sand 0.76 m (30 in)
- Filter Ag 0.76 m (30 in)
- Manganese green sand 0.76 m (30 in)
- Granular PVC 0.76 m (30 in)
A flow schematic of the depth filter test system is shown in Figure 13. A
photograph of the depth filter system is shown in Figure 14. A slip stream
of caustic extraction filtrate was fed to a 54.5 m3/day (10 gpm) Bauer
Hydrasieve to remove fiber. The Hydrasieve underflow was collected in a
surge tank and transferred to the uppermost portion of the column by a
metering pump. The column was constructed of translucent polycarbonate to
allow visual inspection of the media. Surface caking and/or stream
channeling could thus be observed and corrected.
The feed percolated through 0.76 m (30 in) of filtering medium. When
anthracite coal and silica sand were employed the intermix zone was 82 mm
(3.25 in). In all cases the column backflush expansion height was 0.17 m
(6.7 in). The media was supported by a fine-mesh screen and a perforated
plate used to evenly distribute the backwash flow. The inlet pressure to
the filter bed and the filtrate suspended solids content were measured.
The flow rate through the column was maintained at 29.4 m3/m2-day (5 gpm/
ft2). The feed solution was at actual temperature (52-57°C) and pH (pH =
11.5).
During regeneration, filtrate was fed through the base of the column,
and the media bed was expanded to the full column height. The backwashing
was typically performed for 5 to 8 minutes at a flow rate of 88 m3/m2-day
(15 gpm/ft2).
32
-------
Caustic
Extraction
Filtrate
FI
i—txh
-M-
CO
CO
eed
Metering
Pump
Screen
Backflush Expansion
Area
1-
0.5 m of 1.8 mm -
diameter anthracite
coal'
Backflush
Collection
Tank
m of 0.55 mm
diameter silica sand
_L
Backflush distribution
Plate ir
-M-
Backflush
Metering
Pump
Filtrate
Holding
Tank
Figure 13. Flow schematic for 0.05 m diameter depth filter test system.
-------
Uboratory 51 m ll 1i>j
Deep jfe-
-------
LABORATORY ULTRAFILTRATION SYSTEM
A simplified flow diagram of the UF test system employed for tubular
assembly parametric tests and module screening tests prior to mill evalu-
ation is shown in Figure 15. A centrifugal booster pump (Dayton Model
6K507) was used to provide sufficient pressure to pass the feed through two
400 y stainless-steel strainers, in parallel, for removal of gross solids.
A centrifugal circulation pump (Worthington Model D-820) was used to
pressurize the feed and pass it through the membrane module. The flow rate
and pressure were controlled by the pump bypass valve (V-8) and the concen-
trate throttle valve (V-7). A low pressure switch (LPS) protected the pump
from running dry. The concentrate could be recycled either to the feed tank
or to the suction of the circulation pump. A temperature controller
(United Electric, Type 1200) and heat exchanger were used to control the
temperature at a predetermined level. The permeate and concentrate flow
rates were measured, and the feed flow rate was calculated (sum of concen-
trate and permeate flows). The feed pressure and pressure drop across the
module wei*e also determined.
The test system shown in Figure 15 was typically operated in the total
recycle mode. In this operating mode both the concentrate and permeate
streams are returned to the feed tank and the feed is time-invariant.
STIRRED CELL ULTRAFILTRATION SYSTEM
Single Cell Tests
The screening tests for evaluation of non-eellulosic membranes were
performed in stirred cells operated as shown in Figure 16. The membrane to
be tested was placed in the stirred cell (Amicon Model 202) and the cell was
filled with 180 ml of caustic extraction filtrate at ambient temperature.
The magnetic stirrer was then started and the cell was pressurized to
5.2 atm (75 psig). The first 50 ± 2 ml of permeate was collected and dis-
carded while the second 50 ± 2 ml was collected and analyzed for color. A
raw feed sample was also analyzed for color. During the test the volume
processed with time was carefully monitored.
Following the test with caustic extraction filtrate the membrane was
subjected to a salt rejection test under the same operating conditions using
a solution of magnesium sulfate as feed. The first 10 ± 0.5 ml of permeate
was discarded and the next 20 ± 1 ml collected and analyzed for conductivity.
Membranes were evaluated in terms of average flux, apparent color
rejection, intrinsic color rejection and salt rejection. An explanation
of each of these evaluation criteria follows:
- Average flux: The average flux was obtained by dividing the total
volume permeated in the test (100 ml) by (1) the time required to process
the permeate and (2) the total surface area of the flat sheet in the
stirred cell (31 cm*).
35
-------
CO
V-11
Pump
Cleaning
Tank
Permeate Return
•>«-
V-5
-M-
S.S. Screen
Circulation
Pump
UF Permeate to
Collection Tank
Legend
DV - Drain Valve
FI - Flow Indicator
IPS - Low Pressure Switch
P - Pressure Indicator
SV - Sample Valve
TIC - Temperature Indicator/Controller
Figure 15. Simplified flow schematic of laboratory ultrafnitration test system.
-------
CO
Pressure Regulator
Compressed Air
Membrane
Porous Membrane-
Support
Pressure Relief Valve
Stirred Cell
Test Solution
Stirring Bar
Magnetic
Stirrer
I
Permeate
Figure 16- Detail of stirred-cell testing apparatus.
-------
- Apparent color rejection: The apparent color rejection was
obtained using equation 1:
R(%) = 1 - r2" x 100 0)
\ Lf/
where
C = Concentration in the permeate
Cf = Concentration in the feed
- Intrinsic Rejection: Membrane rejection is an intrinsic property and
may not be derived directly from test data under varying feed concentration
conditions. Using certain assumptions (i.e. rejection is independent of
concentration and osmotic pressure effects are insignificant) the intrinsic
rejection can be predicted from the following equation which is derived in
Appendix B:
(l-Y,)^~Ri - (l-Y0)^~Ri 1
R _ i 1 2 I /2\
where
R, = Apparent rejection
a
R.. = Intrinsic rejection
Y.J = Conversion at beginning of sample (50 ml permeated)
Y£ = Conversion at end of sample (100 ml permeated)
Conversion is defined as follows:
y =
- V (3)
Vo
where
Vo = initial volume in cell
V = volume in cell at time t
- Salt rejection: The salt rejection was obtained using equation (1)
where
C = Conductivity of the permeate
Cf = Conductivity of the feed
38
-------
MULTIPLE CELL TESTS
When several membrane types were being compared a multiple stirred cell
test system was employed. This system is shown schematically in Figure 17.
Caustic extraction filtrate was charged to the feed tank and pumped through
4 ultrafiltration stirred cells (Amicon Model 202), in series, by a piston-
type positive displacement pump (Chemtrix Model 7800C). An accumulator
dampened pressure fluctuations and a pressure gauge indicated the inlet
pressure to the first test cell. A back pressure regulator maintained the
operating pressure at 5.2 atm (75 psig). The permeate from the four cells
and the concentrate from the fourth cell were all returned to the feed tank
(i.e., total recycle operation). These tests were performed at ambient
temperature.
MEMBRANE CASTING SOLUTION PREPARATION
The WRP casting solutions used to prepare flat-sheet membranes for
laboratory evaluation were formulated by mixing 500 grams of the required
chemicals in a one liter sample jar, sealing the jar, and placing the
solution on a roller apparatus until a solution which appeared to be homo-
genous was obtained. The following chemicals were used in preparing the
solutions:
Polymer: Polysulfone (Union Carbide #P3500)
Solvent: N-methyl-pyrroli di none
Non-Solvent: Tetrahydrothiophene-1,1-dioxide (Eastman #P9323)
Mixing times of 2-14 days were required to obtain solution homogeneity.
Once the solution was homogenous its viscosity was measured and the membrane
was cast. Casting was performed by drawing the solution over a sheet of
Remay (which was taped to a glass plate) with a Gardner knife set to yield
a solution thickness of 10 mils above the backing. The backing was then
removed from the glass plate, held in a vertical position for 30 seconds
and gelled in a room temperature bath containing either deionized water or
a 0.01 weight % solution of dodecyl sulfate, sodium salt (Aldrich #75,192-2)
in deionized water.
SAMPLING AND ANALYSIS
Samples were collected for two purposes during the program. First,
samples of the three process streams-of interest; hardwood and pine pulp
washing decker effluents and pine bleachery caustic extraction filtrate,
were analyzed at least daily to provide statistical data on composition
and variations in composition on the various streams. Secondly, the
operational streams were sampled during test module stand or pilot plant
operation. These samples were taken of the feed stream, recycle stream,
39
-------
-£»
O
Concentrate Return Line
Back Pressure
Regulator
Key
ACC-Accumulator
P -Pressure Indicator
Stirred Cell Assemblies
Return Line
Figure 17. Flow schematic for stirred cell total recycle tests,
-------
concentrates (from each stage) and permeates (individual module and com-
posites), All samples were taken in clean glass bottles rinsed several
times with the stream to be analyzed after purging the sample lines.
The routine tests run on all samples were:
- temperature,
- PH,
- color,
- suspended solids, and
- total solids.
Standard methods for these analyses are listed in Table 1; brief descriptions
follow.
Temperature: The temperature was measured with a mercury ther-
mometer on the stream during sample taking.
pH: The pH was measured with a Beckman Zeromatic pH
meter with standard glass and calomel electrodes.
Color: Sample had pH adjusted to 7.6, filtered through a
0.8 micron filter disc, and absorbance measured at
465 mu on a Spectronic 20 spectrophotometer.
Standard Pt/Co solution used for calibration.
Suspended solids: (Gravimetric) - measured volume of sample vacuum
filtered through dried, preweighed 0.4 micron
filter disc. Discs plus sample dried 4 hours at
105°C, cooled in a dessicator and reweighed.
Total solids: (Gravimetric) - Weighed sample dried at 105°C for
4 hours, cooled in a dissicator and reweighed.
Composite samples of the feed stream, the concentrates, and the permeates
for 1.2X, 10X and BOX concentration during tubular membrane pilot operation
were collected. These samples were submitted to Galbraith Laboratories, Inc.
for the following analyses:
- Specific gravity
- Total solids
- Volatile solids
- Ash
- Calcium
- Iron
- Aluminum
- Sodium
- Sulfate
- Chlorine
- Ionic chloride
Standard methods employed for these assays are also given in Table 1.
41
-------
TABLE 1. ASSAYS AND .METHODS EMPLOYED .DURING EXPERIMENTAL PROGRAM
Constituent Assay Method Reference
Aluminum Atomic absorption SM 301A*
Calcium Atomic Absorption SM 301A
Chlorine Titration SM 409E
Color Colorimetric TAPPI Method
Ionic chloride Titration Manufacturer's Manual
Iron Atomic absorption SM 301A
pH Meter reading SM 408C
Sodium Flame photcanetric SM 320A
Sulfate Gravimetric SM 427B
Suspended solids Gravimetric SM 208D
Total solids Gravimetric SM 208A
Volatile solids Gravimetric SM 2Q8 E,G
* SM 30A (etc.) refers to procedure number in "Standard Methods for the
Examination of Water and Wastewater," 14th Edition, APHA, 1975.
42
-------
SECTION 6
RESULTS AND DISCUSSION
FEED CHARACTERISTICS AND PRETREATMENT
The characteristics of the feed stream and the effectiveness of pre-
treatment processes are important considerations in the operation of spiral-
wound ultrafiltration modules. As such, the caustic extraction filtrate, and
decker effluents were sampled on nearly a daily basis throughout a 21 -month
period (about 600 samples for each stream). Grab samples of the untreated
streams were analyzed for:
-- pH;
-- Temperature;
-- Total solids;
~ Suspended solids; and,
-- Color.
Several pretreatment options were studied during this program, however,
routine samples were collected of the Kisco depth filter effluent alone.
These samples were only taken of the primary stream studied, caustic ex-
traction filtrate. The pretreated stream samples were analyzed for suspended
solids.
A summary of the mill effluent stream characteristics is presented in
Table 2. A discussion of these overall summary data and average monthly
data follows.
The overall average pH of caustic extract, pine decker and hardwood
decker effluents was 10.8, 10.9, and 9.9 units, respectively. Monthly
average pH values are shown in Figure 18. The hardwood decker stream
averaged a pH of 10 during nearly half of the sampling period. Excursion
from this pH level were consistently on the low side until the final three
months of the program. These latter readings are, however, closer to the
10.0 to 10.6 pH range observed for this stream in a prior investigation (32).
43
-------
TABLE 2 . OVERALL SUMMARY OF FEED STREAM CHARACTERISTICS
Caustic extraction filtrate Pine decker effluent Hardwood decker effluent
Parameter range average range average range average
pH (units) 9.1-12.3 10.8 10.0-12.5 10.9 8.2-11.4 9.9
Temperature(°C) 29.4-52.2 40.8 30.0-65.0 48.3 30.0-61.7 43.9
Color (units) 4,670-29,300 18,450 666-13,000 4,740 2,370-34,000 13,600
Total solids
(tng/1) 2,370-10,600 6,390 686-5,600 2,090 1,060-9,690 3,610
Suspended solids
(mg/1) 10-253 62 13-427 67 40-660 251
Suspended solids
following depth
filter pre-
treatment
(mg/1) 2-124 29
-------
11
10
9
8
1 ' -"I | ••"-" 1 1 1
- *^__^ x>---A A^«^^X»\ Xj\ <^* ,A''*~
XA-" ~ •
W- f'~^
,'» --T ~- V V -T T, .JF----V T>, ,,-T
*•• v-
_ 0—0—0 Caustic extraction
A— A--A Pine decker
•...v. ..« Hardwood decker
1 1 1 1 1 1
j— •
4rzl-
--T — _
» —V
filtrai<
1
en
55
5Q -
r45 -
o
<-?
UJ
C£
|40h
•35
^
June
1976
1 1 1 I
,^ A '/^"A v A^
*-^' ^^s' "^tf/ ~y\^^' v>
•-AC/ —
i i i i
Sept. Dec. March June
1977
DATE
1 1
•— •— • Caustic extraction
A-^A-.-A pine decker
»...T-...T Hardwood decker
I |
Sept. Dec.
I
A
//L
filtrate
5
|
March
1978
Figure 18. Monthly average temperature and pH levels of kraft pulp mill effluent streams,
-------
The caustic extract and pine decker pH levels varied randomly during the
nearly 2-year sampling period. The average values for the caustic extract
stream were also below previously recorded levels. In 1972-73 the caustic
extract stream pH ranged from 11.5 to 12 (32), as compared to a 1976-78
range of 9.1 to 12.3. The pine decker stream exhibited a wider pH range
during the latter sampling but during both periods the average pH was about
10.9.
The wide fluctuations observed in the pH of these streams is probably
due to recycling of other mill effluents back to the bleachery and decker
operations. This recycling may also account for the lower pH levels in the
caustic extract and hardwood decker streams.
Temperature
The monthly average temperatures of the streams of interest are also
plotted in Figure 18. Cyclical fluctuations due to seasonal variations in
temperature were not observed. Also, average temperature for all streams
were below previous recorded values (48 to 57°C) (32). The overall average
temperatures observed in this program were 40.8°C for caustic extraction
filtrate, 48.3°C for pine decker effluent and 43.9°C for hardwood decker.
The lower observed temperature for the caustic extraction filtrate may be
due to the long uninsulated line (122 m) [400 ft] from source to test
position.
Total Solids
Overall total solids contents averaged 2,090 mg/£ for the pine decker
effluent, 3,610 mg/£ for the hardwood decker effluent and 6,390 mg/£ for
the caustic extraction filtrate. Monthly average values are given in
Figure 19. The pine decker effluent exhibited random variation in total
solids loading and had a range of daily readings from 686 mg/£ to 5,600
mg/£. The hardwood decker total solids content appeared to follow a
seasonal pattern: low levels in winter, rising through spring then re-
maining at the average level through summer and into fall. However, the
number of cycles observed was too few to verify this trend. The total
solids range for this stream was 1,060 mg/£ to 9,690 mg/£. The caustic
extraction filtrate total solids level remained higher than average during
the final six months of the sampling period. This suggests the expansion of
water conservation measures at the mill bleachery. The total solids
loading for the caustic extraction filtrate stream ranged from 2,370 mg/£
to 10,600 mg/£.
Both the wide range of total solids readings and the average loadings
observed for the three streams are comparable with prior data.
46
-------
10
9 —
8
x 7
CO
o
o
o
'—•—• Caustic extraction filtrate
k—^—A Pine decker
. _, Hardwood decker
Overall average
concentration = 6,390 mg/1
Overall average
concentration = 3.610 mg/1
Overall average A
concentration = 2,090 mg/1
June
1976
Sept.
Dec.
March
1977
DATE
June
Sept.
Dec.
March
1978
Fi"ure 19. Monthly average total solids concentration for kraft pulp mill effluent streanr.
-------
Color
Plots of the monthly average color concentrations of the caustic ex-
traction filtrate and decker effluents are shown in Figure 20. The pine
decker streams averaged 4,740 color units and remained fairly stable on a
monthly basis throughout the 21-month sampling program. The hardwood decker
effluent varied widely in color concentration. Daily readings ranged from
2,370 to 34,000 color units. The pattern of monthly changes in color
concentration for the hardwood decker closely parallels the pattern observed
for the total solids content of this stream.
The stream with the highest color content is the caustic extraction
filtrate. This stream averaged 18,500 color units with a range of 4,670
to 29,300 color units. The color content remained above average for the
final six months of the sampling period, again signalling increased water
conservation measures.
During the 1972-73 sampling program the average color contents of these
streams were 28,000, 6,000 and 11,000 color units for the caustic extraction
filtrate, pine decker and hardwood decker, respectively. For the caustic
extract stream the current average color loading represents a 50% decrease
from the past level. The current pine decker color concentration is about
25% lower than the previously measured level. Conversely, the hardwood
decker color loading has increased by almost 20%. All of these comparisons
point out the continually changing nature of the mill effluent streams.
Suspended Solids
Plots are shown in Figure 21 of the monthly average suspended solids
concentrations in the three streams. An average loading of 62 mg/£ was
recorded for the caustic extraction filtrate. This stream ranged on a daily
basis from 10 to 253 mg/£ suspended solids. The pine decker effluent had a
similar average, 67 mg/&, and a somewhat wider range: 13 to 427 mg/&.
Monthly fluctuations in the suspended solids content for these streams were
relatively minor.
The stream with the highest suspended solids loading and hence, poten-
tially most troublesome for spiral-wound module ultrafiltration is the
hardwood decker effluent. The spread in suspended solids readings for this
stream was 40 to 660 mg/£ with an average level of 251 mg/£. The variations
in suspended solids loading for the hardwood decker do not correlate over the
entire sampling period with the total solids and color content variations
discussed above.
Caustic Extraction Filtrate Pretreatment
The effectiveness of depth filtration in removing suspended solids from
caustic extraction filtrate was monitored throughout this program. Samples
of the caustic extract stream were analyzed for suspended solids before and
48
-------
22
20-
18 -
16 -
CO
O
- 14
X
§ 12 _
10
•a:
or
O
o
S 6
, , !
Overall average concentration = 18,500 units
mf \ i v • \
\ /' '•
V \ ,*. ,T
>'
\ / Overall average
,' concentration = 13,600 units
V r
1 /
Overall average
concentration = 4,740 units
•—•—• Caustic extraction filtrate
A ± A Pi ne decker
--T Hardwood decker
_| I
June
1976
Sept.
Dec.
March
1977
DATE
June
Sept.
Dec.
March
1978
Figure 20. Monthly average color concentration of kraft pulp mill effluent streams.
-------
en
o
500
400 —
S
UJ
o
g
o
o
en
UJ
a.
(/>
300
200
100
T
• —•—• Caustic extraction filtrate
A--A—A Pine decker
V....W...JT Hardwood decker
Overall average
concentration - 251 mg/1
Overall average pine decker
concentration = 66
\/
Overall average caustic
extract concentration = 62 mg/
I I I
June
1976
Sept.
Dec.
March
1977
June
Sept.
Dec.
March
1978
DATE
Figure 21. Monthly average suspended solids concentration of kraft pulp mill effluent streams
-------
after treatment by the Kisco depth filter. The results of these analyses are
plotted in Figure 22 along with the calculated suspended solids removal
efficiency for the filter. The average feed suspended solids loading was
60 mg/fc; the average filter effluent suspended solids loading was 29 mg/£.
An average removal efficiency of about 50% was achieved. This average re-
moval efficiency is somewhat deflated due to periodic breakthrough in the
filter column. Disregarding the April 1977 and February 1978 average figures
(see Figure 22) the depth filter removal efficiency for suspended solids had
a mean value of 57%.
Summary
In general, solids content, color, pH and temperature varied randomly
from month-to-month. Wide fluctuations were observed on a daily basis. A
seasonal pattern to effluent temperature changes was not observed. A cyclic
pattern to the hardwood decker total solids and color concentrations began
to take shape but would require further study to verify.
In terms of total solids and color the caustic extraction filtrate
stream has the highest loadings. The suspended solids content of the hard-
wood decker effluent is about four times as high (251 mg/£) as the pine
decker (67 mg/£) or caustic extraction filtrate (62 mg/Ji) streams.
Due to equipment changes, process changes or water conservation measures
the nature of the three effluents has changed over the past four to five
years. These types of changes can be expected to occur in the future but
should pose little, if any, problems for an ultrafiltration treatment system.
SELECTION OF PREFERRED MEMBRANE FOR COLOR REMOVAL
Introduction
The initial phase of this program involved selection of the preferred
non-eellulosic membrane for color removal from kraft pulp mill effluents.
The preferred membrane was selected from laboratory tests with the following
candidate systems:
- Commercially available UF membranes coated with cross-linking
agents to improve rejection characteristics;
- Non-commercial polysulfone-based membranes coated with cross-
linking agents to improve rejection characteristics;
- Non-commercial polysulfone-based membranes tailored specifically
to color removal applications (non-coated); and
- Non-commercial interpolymer fixed-charge membranes.
All candidate membranes were evaluated in flat-sheet form in stirred-cell
tests (Screening Tests). Several of the coated commercially available
membranes were further evaluated in the tubular geometry (Parametric Tests).
51
-------
TOO
80
60
40
20
^ L A
en
ro
TOO
80 —
60 -
S
2 40
o
to
o.
CO
CO
20 _
—
June
1976
1 < 1 I 1
• — • — • Before Kisco depth filter
• — a — 41 After Kisco depth filter
A/W./\ /
x / V \/
\ ^ ^ • •
• /•-•-^ -•--• -"-.^ \ ^^i
- V ^^
""-••"
i i i i i
Sept. Dec. March June Sept.
1977
DATE
I 1
-y_
/ ^*m-m
/
i — — m
I i
Dec. March
1978
Figure 22. Caustic extraction filtrate prefiltration data (monthly averages).
-------
Following selection of the preferred membrane, a full-scale spiral-wound
module was fabricated and laboratory tested before module preparation for
on-site piloting.
All membrane selection tests were performed with caustic extraction
filtrate shipped from the Canton Mill. This stream contains lower molecular
weight color bodies than the decker effluents and, as such, is a "worst
case" feed stream.
Membrane Screening Tests
The first series of membrane screening tests were performed with Abcor,
Inc. types HFD, HFM and WRP membranes. The type HFD and HFM membranes are
commercially available, non-eellulosic ultrafiltration membranes. WRP
membrane is a research polysulfone-based material. Interpolymer fixed-
charge membranes (provided by H. Gregor, Columbia University, New York) were
not available for the initial screening tests.
Thirty-five different membrane-coating combinations were prepared from
the basic HFD, HFM and WRP membranes. The coating formulations were based
on proprietary Abcor technology with chemical cross-linking agents. The
listing of the membranes prepared and their stirred-cell performance
characteristics are given in Table 3. As pointed out below coated membranes
failed to maintain their performance characteristics and were not selected
for pilot testing.
General trends can be observed from these screening tests (see Table 3).
First, WRP membrane rejection of color bodies exceeds that of HFD and HFM
membranes. No HFM membrane/coating combination provided intrinsic color
rejection as high as 99%. The best rejecting HFM membrane (HFM-GH 500)
exhibited 98.5% rejection with an associated average flux of 0.86 m3/m2-day
(5.1 atm) [21 gfd (75 psig)]. Two HFD membranes reached or exceeded 99%
intrinsic color rejection (HFD-FH 500 and HFD-GH 500) however these membranes
had flux levels of 0.26 m3/m2-day (6.3 gfd) and 0.53 m3/m2-day (13 gfd),
respectively. 'Nine of the 13 WRP membranes had intrinsic rejections exceed-
ing 99% with several membranes showing 99.5% rejection or better. Flux
levels for these WRP membranes were >0.62 m3/m2-day (15 gfd) in 6 cases, and
as high as 1.97 m3/m2-day (48 gfd).
Clearly, the most important characteristics in determining membrane
performance are rejection and capacity (flux). The preferred membrane would
have a high rejection and a high capacity. However, in actual practice it
is found that high rejection membranes have relatively low fluxes. This is
shown graphically in Figure 23 in which the intrinsic color rejection is
plotted versus membrane flux. Obviously a trade-off exists between
rejection and flux. The data in Figure 23 indicate that the coated WRP
membranes achieve significantly higher flux levels at the same degree of
color removal than either the coated HFD or the coated HFM membranes. This
53
-------
TABLE 3 . RESULTS OF STIRRED CELL SCREENING TESTS
Membrane
HFD
HFD-FN250
HFD-FN500
HFD-GN250
HFD-GN500
HFD-FH50
HFD-FH250
HFD-FH500
HFD-GH50
HFD-GH250
HFD-GH500
HFM
HFM-FN250
HFM-FN500
HFM-GN250
HFM-GN500
HFM-FH50
HFM-FH250
HFM-FH500
HFM-GH50
HFM-GH250
HFM-GH500
WRP
WRP-FN50
WRP-FN250
WRP-FN500
WRP-GN50
WRP-GN250
WRP-GN500
WRP-FH50
WRP-FH250
WRP-FH500
WRP-GH50
HRP-GH250
WRP-GH500
Average f 1 ux
m3/m2-day (gfd)
3.36(82)
3.89(95)
1.60(39)
3.44(84)
2.46(60)
1.52(37)
1.27(31)
0.25(6.3)
1.80(44)
1.23(30)
0.53(13)
4.26(104)
3.81(93)
2.91(71)
3.49(85)
2.34(57)
1.56(38)
2.01(49)
1.39(34)
3.73(91)
1.76(43)
0.86(21)
1.60(39)
1.60(39)
1.60(39)
1.48(36)
1.07(26)
1.97(48)
1.07(26)
0.62(15)
0.38(9.3)
0.36(8.7)
0.98(24)
0.78(19)
0.17(4.1)
Apparent color
rejection, %
84.0
82.2
89.0
90.3
92.2
96.8
98.1
98.3
96.0
97.0
98.2
78.0
84.3
88.0
87.8
92.0
96.6
93.9
94.8
89.1
95.4
97.3
94.4
95.6
98.3
97.6
97.9
98.4
98.7
98.5
99.3
99.5
99.1
99.1
99.1
Intrinsic color
rejection, %*
90.3
89.2
93.5
94.3
95.5
98.2
98.9
99.1
97.7
98.3
99.0
86.4
90.6
92.8
92.7
95.4
98.4
96.5
97.0
93.6
98.0
98.5
96.8
97.5
99.0
98.6
98.8
99.1
99.3
99.2
99.6
99.7
99.5
99.5
99.5
Salt1"
rejection, %
0
0
12
3
12
11
35
42
24
6
32
0
0
12
12
24
22
12
12
12
5
41
14
41
47
32
33
35
47
62
70
45
30
47
50
Calculated
t Using MgS04
54
-------
in
en
99.9
99.8
99.6
99.4
22. 99.2
5 99
UJ
•-3
cc.
o
98
96
94
92
90
60
40
20
10
"T
20
30
40
PERMEATE FLUX
50 60
T
HFM
I
70
T
80
T
90
TOO
no
120
Caustic extraction filtrate
Temperature: 21°C
Pressure: 5.1 atm (75 pslg)
J_
0.43 0.86 1.30 1.73 2.16 2.59
PERMEATE FLUX (M3/M2-DAY)
3.02
3.46 4.32
Figure 23. Intrinsic color rejection versus flux during stirred cell tests.
-------
suggests that the WRP membrane is probably the preferred substrate for the
treatment of caustic extract effluents; however, parametric studies were
performed with selected HFD and HFM membranes to verify this finding.
Parametric Tests
The purpose of the parametric tests was to determine the effects of
temperature, pressure and feed flow rate on the performance of HFD and HFM
membranes with different coatings. Three HFM and two HFD membrane/coating
combinations were tested in 25.4 mm (1 in) diameter x 1.52 m (5 ft) long
tubular assemblies. Only one of each type of treated HFD and HFM tubular
membrane was made. WRP tubular membranes were not available for testing at
this time. A listing of the parametric tests conducted is given in Table 4.
Flux data for the coated HFD membranes throughout tests 1 to 6 are
shown as a function of cumulative operating time in Figure 24. A general
trend of increasing membrane flux with time can be observed. Also, within
each total recycle test a flux increase with time typically occurred.
Looking specifically at the data from tests 4 and 6: a 3-fold increase in
HFD - FH 250 flux and a 2-fold increase in HFD - GH 500 flux took place
between these two tests. Both tests were conducted at 60°C (140°F) and
3.4 atm (50 psig) inlet pressure. The feed flow rate was 163.5 nvVday (30
gpm) for test 4 and 109 m3/day (20 gpm) for test 6. On the basis of the
lower feed flow rate a reduction in membrane flux would have been expected
in test 6 as opposed to test 4. Since the opposite effect occurred, and
because of the generally increasing flux performance the final three para-
metric tests were conducted at the same conditions as test 1 (49°C, 3.4 atm,
163.5 m3/day) [120°F, 50 psig, 30 gpm].
Both HFD membrane flux and membrane color rejection are plotted as
functions of cumulative operating time in Figure 25 for tests 1, 7, 8 and 9.
Flux increased with time while color rejection decreased. Similar trends
were observed for the coated HFM membrane series (see Appendix C). These
results indicated that the membrane coatings were degrading (possibly being
physically removed) as a function of operating time. In view of these
results and the excellent data obtained with the WRP membrane series in the
initial screening tests further evaluation of coated HFD and HFM membranes
was deferred.
Optimization of the WRP Membrane Series
Introduction--
Two conclusions were evident from the initial screening tests and the
parametric studies: The WRP membrane was preferred over the HFD and HFM
membranes and the membrane coating formulations were subject to degradation
with time. The next step in selecting the preferred membrane for color
removal from kraft pulp mill effluents was to determine the preferred
uncoated WRP membrane formulation.
56
-------
TABLE 4 . PARAMETRIC STUDIES TEST MATRIX
(25.4 MM DIAMETER X 1.52 M LONG TUBULAR MEMBRANES)
Test
1
2
3
4
5
6
7
8
9
Operating
temperature,
°C (°F)
49 (120)
49 (120)
60 (140)
60 (140)
60 (140)
60 (140)
49 (120)
49 (120)
49 (120)
Operating
pressure, atm
(psig)
3.4 (50)
5.1 (75)
5.1 (75)
3.4 (50)
5.1 (75)
3.4 (50)
3.4 (50)
3.4 (50)
3.4 (50)
Feed
flow rate,
m3/day (gpm)
163.5 (30)
109 (20)
109 (20)
163.5 (30)
163.5 (30)
109 (20)
163.5 (30)
163.5 (30)
163.5 (30)
Test
duration
hours
24
25
32
20
19
22
24
115
21
57
-------
2.4
2.0
1.6
1.2
0.8 -
oo
X
UJ
£
UJ
UJ
°- 0.4
0.2 -
O HFD-GH500
0-0
0.12
TEST
1 |
TEST
C 2 i '
TEST
™*Z. S
TEST
^4 5
5
TEST
" 6, .'
20 40 60 80 100
CUMULATIVE OPERATING TIME (HOURS)
120
60
50
40
30
20
m
m
cr
x
10o
9 S
8
140
Figure 24 • Coated HFD membrane flux data obtained during
parametric studies.
58
-------
O HFD-FH250
D HFD-GH500
III I III
88
4.8
4.0
<=? 3.2
CJ
CO
2.4
UJ
O.
0.8
I i
O HFD-FH250
D HFD-GH500
)^=5r
i | i i | ill
-Q-
4 6 8 10 20 40 60 100 200
CUMULATIVE OPERATING TIME (HOURS)
100
80 |
60 p
a
x
e>
40 3
20
400 1000
Figure 25. Coated HFD membrane flux and rejection characteristics
determinetl at constant operating conditions.
59
-------
WRP membranes can be formed using different casting solutions containing
the same base polymer, polysulfone. Differences in the casting solutions
(and casting conditions) result in the formulation of membranes with
different structures and hence different performance characteristics. The
variables examined in tailoring the WRP membrane to mill effluent treatment
were:
- polymer concentration in the casting solution;
- non-solvent concentration in the casting solution; and,
- surfactant presence in gelation bath.
The following method of labelling the membranes was derived:
Membrane designation WRP __ _ __
/Weight of non-solvent in casting
solution
\Weight of non-sol Vent + solvent /
in the casting solution
W - gelled in a surfactant solution
N - gelled in deionized water
30-100X (Weight fraction of polysulfone
in the casting solution)
As an example WRP10W33 was formed using a casting solution containing 20
weight % polysulfone 26.7 weight % non-solvent, and 53.3 weight % solvent
that was gelled in a bath containing surfactant.
Results and Discussion—
A summary of the membrane performance characteristics is given in
Table 5. The data shown for each membrane type are based on test results
(5.1 atm, 21°C) with four membranes. An explanation of the calculation
methods employed in determining the numbers presented is as follows:
1• Water flux: The four values obtained were averaged.
2. Process Flux at 3 Hours: The data from each of the four tests were
plotted on semi log paper (linear flux; log time) and a least squares
best fit was performed on each set of data. The flux level for each
membrane at three hours was read from the graph and the resulting
numbers were averaged.
3. Magnesium Sulfate Rejection: The four values obtained were
averaged.
4. Color Rejection: The highest and lowest rejection values obtained
were dropped and the remaining rejection values were averaged.
60
-------
TABLE 5. PERFORMANCE OF WRP MEMBRANES (AT 5.1 ATM, 21 °C)
Membrane
WRP 00 N 00
WRP 00 W 00
WRP 05 N 00
WRP 05 W 00
WRP 06 N 33
WRP 06 W 33
WRP 08 N 33
WRP 08 W 33
WRP 10 N 00
WRP 10 N 33
WRP 10 W 00
WRP 10 W 33
WRP 11 N 33
WRP 11 W 33
WRP 11 W 50
WRP 14 W 50
WRP 17 W 67
Water flux,
nr/m -day (gfd)
0.07 (1.6)
0.09 (2.1)
0.68 (16.5)
0.71 (17.4)
2.39 (58.4)
1.80 (43.9)
2.72 (66.4)
3.05 (74.4)
3.77 (92.0)
3.18 (77.5)
3.63 (88.6)
3.90 (95.2)
4.26 (104)
3.44 (83.8)
4.35 (106)
5.95 (145)
5.21 (127)
Process flux
@ 3 hours,
m3/m2-day (gfd)
0.03 (0.8)
0.05 (1.2)
0.38 (9.3)
0.33 (8.0)
0.70 (17.0)
0.57 (13.8)
0.62 (15.0)
0.63 (15.4)
0.76 (18.6)
0.67 (16.4)
0.78 (19.1)
0.70 (17.0)
0.85 (20.8)
0.80 (19.4)
0.93 (22.6)
0.98 (23.9)
1.00 (24.4)
Magnesium
sulfate
rejection ,%
__
18
32
26
32
26
33
25
31
26
34
32
30
34
34
26
35
Color
rejecti on, %
94.2
82.8
96.6
96.5
93.3
96.8
96.6
97.8
97.1
97.2 -
96.3
97.6
96.5
97.0
97.0
92.7
91.8
61
-------
The best results were obtained with the WRP08W33 and the WRP11W33
membranes. Membrane WRP08W33 yielded a color rejection of 97.8% and a
process flux after three hours of 0.63 m3/m2-day (15.4 gfd) when processing
caustic extraction filtrate. Membrane WRP10W33 yielded a color rejection
of 97.6% and a process flux after three hours of 0.7 m3/m2-day (17.0 gfd).
The effect of polymer concentration on water flux and process flux is
shown in Figures 26 and 27. As expected, flux declined with increasing
polymer concentration. A linear relationship is observed in both cases.
Little or no effect on membrane flux is evident from the presence of non-
solvent in the casting solution or from surfactant in the gelation bath.
The effect of all variables examined on membrane color rejection is
given in Figure 28. Rejection with respect to polymer concentration went
through a maximum in all cases in the range of 20% to 24% by weight polymer.
The best rejections were obtained with membranes cast from solutions contain-
ing non-solvent and gelled in a bath containing surfactant. It should be
noted that most of the membranes cast with non-sol Vent in the casting
solution had a solvent to non-solvent ratio of 2:1 (i.e., the effect of non-
solvent was not thoroughly investigated).
It can be concluded from these data that high rejection-moderate flux
ultra-filtration membranes can readily be made from polysulfone based casting
solutions. Color removals from caustic extraction filtrate in excess of
97.5% at flux levels of 0.7 m3/m2-day (17 gfd) were attained at 5.1 atm
(75 psig). These results can be compared to 90.3% color removal at 1.6 m3/
m2-day (39 gfd) for one HFD flat sheet membrane and 86.4% color removal at
2.0 m3/m2-day (49 gfd) for the HFM membrane.
The preferred WRP membrane was determined to be WRP09W33. This mem-
brane was not cast during these tests but was selected based on the data
plots of Figures 26, 27 and 28. Theoretically this membrane would have
higher flux than WRP08W33 (>0.63 m3/m2-day) [15.4 gfd] and higher
rejection than WRP10W33 (>97.6%). WRP09W33 would be gelled in a bath
containing 0.01 weight % dodecyl sulfate, sodium salt and be cast on a
Rernay backing using the following solution:
Solvent:* N-methyl-pyrrolidinone 52.7 grams
Non-Solvent: Tetrahydrothiophene-1,1-dioxide 26.3 grams
Polymer: Polysulfone 21.0 grams
Interpolymer Fixed Charge Membranes
Introduction—
Interpolymer fixed charged membranes were prepared by H. Gregor
(Columbia University, New York) and tested at both Columbia and Abcor. All
of these membranes were made in accordance with U.S. Patent 3,808,305 and
62
-------
CTl
CO
5.6 -
4.8 -
4.0
1
3.2
ac
S3
U_
Of.
2.4
1.6
0.8
Q No surfactant or non-solvent
O Surfactant; n6 non-solvent
V Non-solvent; no surfactant
A Non-solvent and surfactant
Pressure: 5.1 atm (75 psig)
Temperature: 21 °C
least squares
best fit
I
I
12 14 16 18 20 22 24 26
POLYMER CONCENTRATION («)
28
30
140
120
100
80
73
ff>
60 3.
40
20
32
Figure 26. The effect of polymer concentration on WRP membrane water flux.
-------
1.2
en
-p.
I
CM
o
x
to
CO
LU
I
a.
0.8
0.4
28
24
20
D No surfactant or non-solvent
O Surfactant; no non-solvent
V Non-solvent; no surfactant
A Non-solvent and surfactant
caustic extraction filtrate
Pressure: 5.1 atm (75 psig)
Temperature: 21°C
least squares
best fit
-o
73
O
O
m
oo
12
O
OO
12 14 16 18 20 22 24
POLYMER CONCENTRATION (%)
26
28
30
32
Figure 27. The effect of polymer concentration on WRP membrane process flux.
-------
cn
8 -
-------
consisted primarily of sulfonic acid polymers. Seven flat sheet (7.6 cm x
7.6 cm) membranes were tested at Abcor. The results of these studies are
presented below. Tests performed by Dr. Gregor are documented in his Final
Report, "Use of Sulfonic Acid Membranes for Treatment of Pulp and Paper
Waste Streams". This report is presented, in its entirety, in Appendix D.
Short Term Tests--
A summary of the short term test results obtained with the interpolymer
fixed charge membranes is given in Table 6. All membranes had only letter
designations (e.g., A, B ... G), as received. Details on the membranes of
most interest (A through E) can be found in Dr. Gregor's report.
In these 3-hour tests with caustic extraction filtrate the interpolymer
fixed charge membranes displayed color rejection and process flux character-
istics similar to the WRP membrane series. Process flux after 3 hours
ranged from 0.42 to 0.71 m3/m2-day (10.2 to 17.4 gfd). Color rejection
varied from 90.7% to 97.8%. Best results were obtained with membrane B
which exhibited a process flux of 0.7 m3/m2-day (17 gfd) and a color
rejection of 97.8%.
Long Term Tests--
The interpolymer fixed charge membranes were predicted (by their
developer) to be non-foul ing in nature. Since short term tests showed no
practical differences between the WRP and the interpolymer fixed charge
membranes, long term tests were conducted to determine the flux decline
characteristics of each membrane type. The membranes selected for testing
were those which had demonstrated superior performance in previous tests.
These were interpolymer fixed charge membranes A and B, WRP10W33 and
WRP11W33.
The flux decline curves for the polysulfone-based and sulfonic acid-
based membranes are shown in Figure 29 for the 65 hour comparison test.
Membrane flux decline rates were essentially identical even though the WRP
series membranes showed intrinsically higher flux characteristics. The
difference (i.e., increase) in WRP flux over previous trials is attributed
to variation in the feed solution. In this test the WRP and the inter-
polymer fixed charge membranes were evaluated simultaneously giving a true
comparison of both intrinsic membrane flux and rejection properties and
flux decline characteristics.
Membrane color rejections during this test were:
Membrane A - 93.4%
Membrane B - 96.7%
WRP10W33 - 97.2%
WRP11W33 - 98.1%
66
-------
TABLE 6 . INTERPOLYMER FIXED CHARGE MEMBRANE TEST RESULTS
Membrane*
A
B
C
D
E
F
G
Water
9?1ux,
m3/m2-day (gfd)
1.54 (37.6)
1.26 (30.8)
1.75 (42.8)
0.54 (13.1)
1.03 (25.1)
0.40 ( 9.7)**
0.38 ( 9.3)**
Magnesium
sul fate
rejection,%
48
50
36
54
35
52
25
Process flux
at 3 hrs,
m3/m -day (gfd)
0.63 (15.3)
0-70 (17.0)
0.71 (17.4)
0.42 (10.2)
0.59 (14.4)
0.52 (12.6)
t
Color
rejection,
%
96.5
97.8
93.3
97.2
90.7
93.9
t
* Designation of membrane as received. See Appendix D.
** Membrane may not have been properly "wet" before testing.
t A mechanical leak in the stirred cell occurred during the test.
67
-------
en
CO
1.6
1.2
i
ro
0.8
S 0.4
T
O WRP 10 W 33
E Interpolymer fixed charge membrane A
A WRP 11 W 33
V Interpolymer fixed charge membrane B.
Caustic extraction filtrate
Inlet pressure: 5.1 atm (75 psig)
Temperature: 21°C
40
30
-o
m
J
I
4 6 8 10 20
CUMULATIVE OPERATING TIME (HOURS)
40
20
en
-n
a
10
I I
60
80
Figure 29. Comparison of interpolymer fixed charge and WRP membrane flux and decline.
-------
Thus the WRP membranes showed higher color rejection and higher flux levels
than the interpolymer fixed charge membranes. These superior performance
characteristics, coupled with the essentially identical rates of flux de-
cline, indicated that WRP membranes were preferred for color removal
applications at Kraft pulp mills.
WRP Spiral-Wound Module Laboratory Test
Based upon the WRP/interpolymer fixed charge membrane test results and
the dependence of casting dope viscosity on polymer concentration it was
decided to prepare continuous flat sheet membrane from the WRP11W33
formulation. This flat sheet membrane was fabricated into two 0.1 m (4 in)
diameter x 0.91 m (36 in) long spiral-wound modules. Each module contained
approximately 3.72 m? (40 ft2) of active membrane area.
One module was tested in total recycle with caustic extraction fil-
trate for 140 hours. The module flux performance during this test is shown
in Figure 30. The flux was high, 2.05 to 2.67 m3/m2-day (49°C, 4.4 atm,
43.6 m3/day recirculation flow rate) [50 to 65 gfd (120°F, 65 psig, 8 gpm)L
and stable. The higher flux observed in the module as compared to stirred-
cell test results is partially a function of increased feed temperature and
improved hydrodynamics across the membrane surface.
Two other factors potentially affecting membrane flux performance are
changes in membrane properties associated with the adaptation of casting
procedures from the laboratory-scale to production-scale, and changes in the
feed stream due to aging. These two factors also contributed to a reduction
in membrane rejection for color bodies. The one set of samples collected
during this run (taken at 24 hours cumulative operating time) showed a
module color rejection of only 86% as compared to WRP11W33 flat-sheet color
rejection of >97%. The degree to which scale-up and feed degradation each
affected membrane performance could not be determined quantitatively. It
was assumed, however, that the necessity of shipping the feed solution for
this test by truck and the lack of feed refrigeration caused the higher
molecular weight color bodies to settle-out resulting in an unrepresentative
feed sample*.
At this point laboratory testing was completed. Further evaluation of
WRP spiral-wound modules was performed at the Canton Mill where fresh feed
would be available on a continuous basis.
Samples of caustic extraction filtrate used during stirred-cell tests were
air-shipped in 0.02 m3 (5 gal) drums from Canton, North Carolina and
refrigerated upon arrival at the test laboratory.
69
-------
CO
o
o
3.6
3.2
2.8
2.4
2.0
1.6
1.2
0.8
0.4
n
1 i 1 1 | 1 1 1 1 | |
— _
— -
Q . _
°-
Caustic extraction filtrate
_ Inlet pressure: 4.4 atm (65 psig)
Circulation flowrate: 43.6 m3/day (8 gpm)
Temperature: 49°C
— -
- —
— —
| I i I 1 i i i | 1 i
100
90
80
70
-------
FIELD EXPERIENCE WITH SPIRAL-WOUND MODULES
Introduction
Before initiating studies with the 3-stage pilot system, testing of
single spiral-wound modules was performed. In this manner the effectiveness
of various pretreatment options, the preferred conditions for module
operation and the degree of module flux recovery following cleaning could
be determined in more controlled experiments. Also, a significant level of
experience would be gained under actual field conditions while awaiting
fabrication of the remaining spiral-wound cartridges. At the conclusion of
the single module tests, the 3-stage pilot system was operated for a 3-
month period. Essentially all of these tests with spiral modules were
performed with the caustic extraction filtrate stream.
Throughout all of the spiral cartridge tests the main indicators of
module performance were membrane flux, flux recovery and color rejection.
Single Module Test Results
Six different spiral cartridges were employed during the single module
tests. The module designations and their descriptions are listed in
Table 7. Both the HR and LR modules became partially dry during shipment
and were rewetted by soaking in an isopropyl alcohol and surfactant
solution.
Single module experimentation focused on:
-- membrane flux dependency on time;
— effects of feed pH adjustment, feed temperature and
inlet pressure on module flux decline;
— module pressure drop as a function of time and feed
flowrate; and,
-- module cleaning requirements.
Additionally, sodium hexametaphosphate addition for colloid stabilization was
performed, general observations on cartridge mechanical integrity were made
and module color rejection was monitored. At this point in the program
color rejection was not a major concern since the extensive laboratory
screening tests had shown membranes could be "tailored" to this application.
A summary of all single module tests is given in Table 8. Except for
a lone 7 hour test (#9) with pine decker effluent all tests were performed
with caustic extraction filtrate as the feed stream. All flux data shown
in Table 8 have been temperature corrected to 50°C using a ratio of the
viscosity of water at the actual process temperature to the viscosity of
water at 50°C. The process flux data were generally recorded at 38°C to
50°C; water flux measurements were generally read at 20°C to 40°C. Over a
71
-------
TABLE 7 . MODULES EMPLOYED DURING SINGLE MODULE TESTS
Module
designation Description
HR WRP11W33, 60 mil vexar spacer; incurred partial
drying during shipment.
LR WRP11W33, 60 mil vexar spacer; incurred partial
drying during shipment.
WRP-1 WRP09W33, 60 mil vexar spacer.
WRP-2 WRP09W33, 60 mil vexar spacer.
0-PS Polysulfone membrane, 30 mil vexar spacer;
manufactured by Osmonics, Inc.
0-CA Cellulose Acetate membrane, 30 mil vexar spacer;
manufactured by Osmonics, Inc.
72
-------
TABLE 8. CHRONOLOGY OF SINGLE MODULE TEST EXPERIENCE
CO
Tot
no.
1
2
3
4
S
6
7
8
9
10
11
u
11
todult
00.
m
IR
U
HR
HUM
HR
Ml
MR
m
KR
HR
HS
KR
Test
duration,
hrs
20. >
15.5
18
«
16
22
21. 5
4.S
7
22
19
S.i
21.$
Initial
process
mW-day (ofd)"
0.78 (19.1)
0.67 (16.4)
0.95 (23.2)
1.16 (28.4)
4.22 (103)
1.12 (27.4)
0.62 (15.2)
0.30 (7.2)
0.45 (10.9)
0.44 (10.8)
0.80 (19.5)
0.62 (15.1)
0.5! (13.3)
Final
process
flux
m3/m2-dav (ofdf
0.19 (4.6)
0.21 (S.I)
0.09 (2.1)
0.34 (8.4)
1.06 (25.8)
0.14 (3.3)
0.09 (2.1)
0.23 (5.7)
0.18 (4.3)
0.89 (2.3)
0.34 (8.2)
0.34 (8.4)
0.23 (5.7)
Pre-Ust
water
mW-dav (cfdl*
2.69 (65.5)
0.88 (21.4)
0.80 (19.5)
1.78 (43.4)
4.80 (117)
3.26 (79.6)
1.23 (30.1)
0.32 (7.7)
0.84 (20.4)
0.95 (23.2)
1.24 (30.3)
0.68 (16.5)
2.16 (S2.7)
Post-test
water
flux.
n3/m*-d>v (ofdl*
0.80 (19.5)
0.55 (13.4)
2.33 (56.9)
1.79 (43.6)
1.23 (30.1)
0.32 (7.7)
0.84 (20.4)
0.95 (23.2)
1.24 (30.3)
0.68 (16.5)
2.16 (52.7)
1.18 (28.8)
Range of
color
rejection. I _
50-60
40-50
„
56-79
72-94
71-94
35-85
82-90
54-73
65-74
78-92
89-92
82-87
Coments
Module changed without final water flux.
Initial process flux reading at 21°C.
Possible pin-hole leaks In nodule account-
Ing for Msh 1nttt>1 flux and low Initial
rejection.
Pine decker effluent.
Caustic extraction filtrate adjusted to
pK 7.
Caustic extraction filtrate adjusted to
14
0.76 (18.C)
0.41 (10.5)
1.18 (28.8)
pH7.
88-93
(continued)
-------
TABLE 8.
(CONTINUED)
no.
15
16
17
18
19
20
21
22
23
24
25
26
27
28
2»
Moduli
no.
URP-2
WSP-2
KRP-2
wW-2
I*
L*
LR
LR
LI)
LR
LR
0-fS
0-PS
URP-2
0-CA
Test
duration.
hrs
22
19.5
21
4
5
23
7.5
7.5
52
25
27
23
21
22
20
Initial
process
flux.
n,W-d«y (afdl*
1.30 (31.8)
1.11 (27.0)
0.80 (19.6)
0.81 (19.7)
1.12 (27.2)
0.96 (23.5)
2.64 (64.3)
0.84 (20.4)
0.92 (22.5)
0.42 (10.3)
0.58. (14.1)
4.31 (105)
1.99 (48.5)
2.52 (61.4)
2.35 (57.3)
Final
process
flux,
u3/m2-day (afdl*
0.36 (8.7)
0.40. (9. })
0.39 (9.5)
0.59 (14.5)
0.47 (11.4)
0.12 (2.9)
1.19 (29.1)
0.76 (18.S)
0.32 (7.9)
0.19 (4.7)
0.23 (5.6)
0.60 (14.7)
0.62 (15.0)
0.61 (14.8)
1.44 (35.0)
P re- test
water
BrVm*-dav (afd)»
1.02 (24.8)
1.79 (43.6)
—
..
1.53 (17.2)
—
..
mm
..
0.59 (14.3)
.„
6.07 (148)
2.38 (58.0)
3.87 (94.5)
2.77 (67.()
Post-test
water
«3/&-daV («fd)«
1.79 (43.6)
—
—
..
—
-.
..
—
0.59 (14.3)
-
..
2.38 (58.0)
.1.32 (32.3)
2.19 (53.3)
2.43 (59.2)
Range of
color
rejection,!
66-73
44-63
51-65
58-63
84-90
18-75
(21-24)
67-74
66-82
59-63
58-76
56-61
54
79-86
71-86
ComwntS
Test rur at 2.04 atm (30 pslg).
Test rut: at 2.04 atm.
Test run at 2.04 atm.
Test run at 2.04 atm.
Test run at 60*C.
Module run 1n first stage of three stage system.
Data indicate probable leak.
Module run in first stage of three staoe system.
Module run 1n first stage of three stage system.
Module run 1n first stage of three stage systes.
Test run at 60°C. Feed adjusted to pH 12.
Module run 1n first stage of three stage system.
.Feed adjusted to pH 12.
Module had new outer wrap and new brine seal .
Feed adjusted to pH 9 during first 4.5 hours
of run. Module rinsed with water 3 tines
during run.
(continued)
-------
TABLE 8.
(CONTINUED)
en
Test
no.
30
31
32
33
34
35
36
3)
»
39
40
41
42
43
Module
no.
0-CA
0-CA
0-CA
0-CA
URP-1
HO
URP-1
URP-1
UtP-
URP-
WKP-
HRP- .
URP-
KRP-1
Test
duration.
hrs
26
21
21 .S
17.S
19
21. S
18
3
22
18.5
22
14.5
75
S3
Initial
process
fluxi
2.19 (S3.4)
1.23 (30.0)
1.27 (31.0)
1.05 (25.5)
2.15 (52.5)
2.32 (56.6)
2.19 (53.4)
2.29 (55.8)
2. SI (61.3)
1.87 (45.6)
1.47 (35.8)
2.21 (53.8)
1.82 (44. S)
1.31 (32.0)
Final
process
flux4
ui3/m2.day (ofd)*
1.24 (30,3)
0.40 (9.8)
0.25 (6.2)
0.34 (8.3)
0.84 (20.4)
0.29 (7.0)
0.62 (15.1)
1.41 (34.3)
0.63 (15.3)
0.58 (14.2)
0.52 (12.7)
1.15 (28.0)
0.55 (13.3)
0.76 (18.5)
Pre-test
water
flux.
n3/mZ-day (afd)*
2.43 (59.2)
1.51 (36.8)
1.59 (38.8)
1.38 (33.6)
3.94 (96.1)
4.08 (99.6)
2.70 (65.9)
3.21 (78.4)
3.42 (83.5)
2.49 (60.7)
2.30 (56.0)
3.09 (75.3)
3.33 (81.3)
2.33 (56.9)
Post-test
water
flux.
B3/m2-dav (afdj*
1.51 (36.8)
1.59 (38.8)
1.38 (33.6)
1.64 (39.9)
2.70 (65.9)
..
3.21 (78.4)
3.42 (B3.-5)
2.49 (60.7)
2.30 (56.0)
3.09 (75.3)
3.33 (81.3)
2.33 (56.9)
2.62 (61.5)
Range of
color
rejection, S
74-79
45-84
53-59
56-68
86
84
90
79
85-92
79-88
76-87
72-81
63-83
70-84
Contents
Test terminated when caustic line went down.
Test run In total recycle mode.
Test run 1n total recycle mode. Sodium
hexametaphosphate added to caustic extraction
filtrate In attempt to stabilize colloidal
natter.
Test run In total recycle mode.
All flw readies temperature corrected to 50'C.
-------
wide temperature range (i.e., 25°C to 50°C) the viscosity-ratio method may
be slightly inaccurate; however, for ease of test-to-test data comparison a
single reference point (50°C) was deemed desirable.
Using Table 8 for an overview, specific data observations will be
discussed below.
Typical Module Flux Performance—
In order to more fully appreciate the effects of the various techniques
employed for flux improvement, a "typical" flux decline curve will be
examined initially. The flux versus time plot for such a case, test #6, is
shown in Figure 31. In this test the HR module processed caustic extraction
filtrate on a once-through basis for 22 hours. Initial module water flux
was 3.26 m3/m2-day (79.6 gfd). Process flux at time zero, 2.37 m3/m2-day
(57.9 gfd), represents dilution of the caustic stream by water from the
backflush depth filter. In 0.3 hours, as the system was purged of water,
process flux declined by over 50% to 1.12 m3/m2-day (27.4 gfd). The module
flux continued to decline rapidly over the next 2 to 3 hours. A leveling
off begins to occur at ^.41 m3/m2-day HO gfd). Overnight operation of the
system resulted in a gradual loss of flux down to 0.14 m3/m2-day (3.3 gfd).
Again, it must be noted that no appreciable concentration occurred
during this test: the caustic extraction filtrate was processed on a once-
through basis. Also, no increase in pressure drop was observed across the
module. The flux decline is therefore related to operating time rather than
concentration effects (increased solids loading) or module plugging by
fibers (reduced membrane area and lower average pressure).
A flux pattern of the nature observed in test #6 indicates a steady
fouling of the membrane surface by some specie(s) in the caustic extraction
filtrate stream. At this point in the program the foulant(s) had not been
identified. It was clear, however, from both technical and economic stand-
points that either process operating conditions had to be modified or
additional feed pretreatment exercised if satisfactory flux levels were to
be achieved and maintained.
Total Recycle Flux Performance—
The flux vs. time curves for two total recycle tests (tests #41 and 43)
with the WRP-1 module are given in Figure 32. In both instances approxi-
mately a 40% flux decline occurred during the first 5 hours of processing.
The flux loss tapered-off slightly between 5 and 10 hours processing and
then stabilized at 1.15 m3/m^-day (28 gfd) and 0.82 m3/m2-day (20 gfd) for
tests #41 and 43, respectively. By contrast, the "typical" once-through
process test (see Figure 31) showed nearly an 80% flux loss in the first 5
hours with a subsequent steady flux decline down to 0.12 m3/m2-day (3 gfd).
Operation of a single module in total recycle at the laboratory had
shown stable flux performance on the order of 2.05 nr/mZ-day (50 gfd) (see
Figure 30), whereas total recycle operation in the field resulted in stable
76
-------
System startup
with water in
HR Module, Test #6
caustic extraction filtrate
sand filter
Feed flowrate: 49 m3/day (9 gpm)
Inlet pressure: 5.1 attn (75 psig)
Flux temperature corrected to 50°C
c:
i—
m
8 10 12 14
CUMULATIVE OPERATING TIME (HOURS)
Figure 31. "Typical" module flux performance during single module tests.
-------
2.4r
60
oo
• WRP-1 , Test 141
• WRP-1 , Test #43
caustic extraction filtrate
Feed flowrate: 54.5 m^/day
Inlet pressure: 5.1 atm (75 psig)
Flux temperature corrected to 50°C
gpm)
50
40
cn
20
10
10 15 20 25 30
CUMULATIVE OPERATING TIME (HOURS)
35
40
45
50
Figure 32. Module flux performance during total recycle tests.
-------
module flux only after initial steep flux declines. Changes in the feed
stream (i.e., aging, changing mill conditions) are believed to account for
these different flux patterns.
The stabilization of module flux in the total recycle mode suggested a
finite quantity of membrane foulant(s) was present in the caustic extraction
filtrate. Once the fairly rapid foul ant/membrane interaction was complete,
steady flux was maintained. Again, control of the foulant(s) on a continuous
basis was required.
Module Flux with pH Adjusted Feed—
The average pH of the caustic extraction filtrate was 10.8. Upward
adjustment of the pH could potentially stabilize colloids in the feed, while
downward pH adjustment was expected to agglomerate colloidal matter. In
the latter case, some fraction of the suspended solids would be removed by
the system prefilters.
Figure 33 shows the flux versus time curves for tests at pH 7 (tests
#10 and 13) and at pH 12 (test #25). In all three experiments initial
process flux was somewhat low (0.41 to 0.82 m3/m2-day) [10 to 20 gfd].
While this may be partially due to low pre-test water flux levels in tests
#10 and 25, the pre-test water flux of 2.16 m3/m2-day (52.7 gfd) in test
#13 belies complete dependence on this factor.
For the tests at pH 7 a 50% flux loss occurred within 5 hours operating
time. This is similar to the results of the total recycle tests. From this
point the flux decline was slow, but constant, with final flux readings of
0.08 to 0.25 m3/m2-day (2 to 6 gfd).
The test conducted at pH 12 indicated slightly less severe fouling with
a 32% flux decline in 5 hours. After five hours, however, the slope of the
flux versus time curves for the pH 12 and pH 7 tests were essentially
identical.
In summary, no noticeable improvement in module flux performance re-
sulted from feed pH adjustment.
Module Flux at Elevated Temperature—
The transport of water (flux) across a membrane surface increases with
increasing temperature. Increased process temperature has also been shown
to alter the fouling characteristics of certain feed streams (35). Thus,
the caustic extraction filtrate stream was processed at 60°C in two single
module runs (tests #19 and 24). In test #24 the feed pH was also adjusted
upward to pH 12.
The duration of test #19 was only 5 hours, however in that time a 58%
flux decline occurred. Test #24, conducted for 25 hours, had an initial
module flux of only 0.5 m3/m2-day (12.1 gfd) (60°C) and a final flux level
79
-------
00
o
• HR, Test #10, pH=7
• HR, Test #13, pH=7
A LR, Test #25, pH=12
caustic extraction filtrate
Inlet pressure: 5.1 atm (75 psig)
Flux temperature corrected to 50°C
pH=12
9 12 15 18 21 24
CUMULATIVE OPERATING TIME (HOURS)
Figure 33. Module flux performance during pH adjustment tests,
27
60
50
40
30
cz
X
CD
-n
o
20
10
30
-------
of 0.23 m3/m2-day (5.5 gfd). Based on the results of these two experiments
processing at elevated temperature does not reduce membrane fouling and,
hence, is not recommended.
Module Flux with Sodium Hexametaphosphate Addition-
One experiment was performed with sodium hexametaphosphate (SHMP) added
to the feed to reduce the precipitation of colloids onto the membrane surface.
The SHMP was added to the caustic stream at a 10 ppm concentration and a
75 hour total recycle test was conducted (test #42). Within 5 hours a 54%
flux loss, from 1.82 to 0.84 m3/m2-day (44.5 gfd to 20.5 gfd), was observed.
Thus the addition of SHMP had no effect on the startup performance of the
module. After 30 hours total recycle the flux stabilized at 0.53 to 0.62
m3/m2-day (13 to 15 gfd) for the duration of the test. Since the SHMP
showed no initial flux improvement, stability probably resulted from the
total recycle nature of the test rather than from colloid stabilization.
Module Flux with Lower Inlet Pressure--
Membrane compaction is a common, and sometimes highly significant,
phenomenon in the operation of reverse osmosis membrane modules. While
ultrafiltration is performed at much lower pressures than reverse osmosis,
the potential exists for UF membrane compaction leading to flux decline and
inability to recover initial module flux. To investigate this possibility
four tests (Tests #15 through 18) were conducted with the WRP-2 module
operated at 2 atm (30 psig) inlet pressure. This module had not been
operated at all prior to these tests.
The flux history of WRP-2 during the low pressure tests is shown in
Figure 34. In each of the four tests significant flux loss is observed in
the initial portion of the test. The sharp flux decline is followed by a
gradual drop in flux throughout the remainder of each test. These results
are similar to the flux patterns developed at 5.1 atm (75 psig) inlet
pressure. It can also be noted from Figure 34 that detergent cleanings
could not recover process flux to its initial value (1.3 m3/m2-day) [32 gfd].
In fact, a 40% decline in initial process flux is observed by the fourth
test.
The membrane fouling occuring during caustic extraction filtrate
processing thus appears independent of inlet pressure in the range of
2 to 5.1 atm (30 to 75 psig). As such, membrane compaction is not
suspected to be of any significance for this application.
Comparison of Individual Module Flux Performances—
In addition to varying operating conditions and feed pretreatments to
isolate the factors contributing to poor membrane flux performance,
modules other than the WRP series were investigated. In this manner it could
be determined if the major foulant(s) were preferentially affecting the
WRP-type membrane. The membrane modules evaluated as alternatives to the
WRP modules were polysulfone and cellulose acetate modules commercially
available from Osmonics, Inc. These modules were designated 0-PS and 0-CA,
81
-------
2.4
T
T
60
2.0
WRP-2 Module
caustic extraction filtrate
Tests #15 to 18
Feed flowrate: 43.6 to 62.7 m-Vday
(8 to 11.5 gpm)
Inlet pressure: 2.04 atm (20 psig)
Flux temperature corrected to 50°C
50
40
CO
ro
CM
-».
CO
QJ
1.2
o 0.8
0.4
o
cr
r—
m
30
c:
x
o
20
10
10
20 30 40
CUMULATIVE OPERATING TIME (HOURS)
50
60
70
Figure 34. Module flux performance at low inlet pressure (2.04 atm) [30 psig].
-------
respectively. The CA module was operated initially with a pH 9 feed stream
to hydrolyze the cellulose acetate to cellulose. It was then expected that
the 0-CA module could be operated at the natural pH and temperature of the
caustic extraction filtrate.
Flux decline is plotted in Figure 35 as a function of operating time
for the WRP-1, 0-PS and 0-CA modules. Even though the individual tests
were performed over a several month period almost identical flux decline
curves were obtained. Thus the membrane fouling and module flux decline
are not specific to the WRP spiral-wound modules but to other polysulfone-
based modules and cellulosic modules, as well.
Module Pressure Drop Considerations—
The measurement of pressure drop across a spiral-wound cartridge is
important for two reasons. First, an increase in pressure drop with time
(at constant flowrate) signals module plugging. This could indicate
incomplete pretreatment, especially in streams containing fibrous material.
Second, the power requirement for an ultrafiltration system is determined
almost entirely by the power input to the feed circulation pump. This power
input is directly proportional to the product of the volumetric output of
the pump and the pressure drop across the module system. An increase in
circulation rate causes a concomitant increase in pressure drop, greatly
affecting system power requirements. Therefore, a tradeoff exists between
improved flux and increased power costs as the circulation rate is raised.
Plugging of a spiral-wound module can have several deleterious effects:
reduction in effective membrane area due to feed channel blockage; lower
average driving pressure across the module, possibly reducing flux; and
reduction in module flux recovery with cleaning, leading to shortened module
life. In this program, no statistically significant data were obtained on
module pressure drop versus time because of mechanical problems (brine seal
failures), module changeover during the tests and use of uncalibrated
pressure gauges. However, the daily log sheet data suggested no build-up
of pressure drop with operating time. This is most likely the case since
inspection of the feed channel spacer within one module showed no entrapped
fiber or other solids. On this basis the pretreatment sequence, while not
able to reduce module fouling, was sufficient to prevent module plugging.
Module flux versus pressure drop (plotted as flux versus circulation
rate which was more accurately measured) is shown in Figure 36 for HR module
processing of caustic extraction filtrate over a 7 day period. The extreme
scatter in the data and the lack of a linear relationship (log-log plot)
indicate flux sensitivity to circulation flow is minimal. Thus, in the
normal operating range of Vexar-spacer spirals (27.3 to 81.8 mj/m^-day
recirculation [5 to 15 gpm] no improvement in flux with increased feed
superficial velocity is observed.
83
-------
00
Test No.:
Feed Flowrate(n)3/day):
Inlet Pressure (atm):
8 10 12 14 16
CUMULATIVE OPERATING TIME, HOURS
Figure 35. Module flux decline as a function of operating time.
-------
00
en
CIRCULATION RATE (6PM)
5
10
CIRCULATION RATE (M3/DAY)
Figure 36. Module flux versus circulation rate.
15
20
t. .0
2.4
2.0
1.6
>-
S
o
CSJ
^ 1.2
CO
4<^H
X
Z3
IJ-
y 0.8
Z3
a
0.04
r i T
HR module °
caustic extraction filtrate
Inlet pressure: 5.1 atm (75 psig)
— Flux temperature corrected to 50°C ° —
— GO —
_
O w
— .
— <5 —
<§> Q O
oO ®
O
-------
In summary, neither increased pressure drop at constant flow (plugging)
nor improved flux at increasing flow (reduced concentration polarization)
were observed during the single module tests.
Module Mechanical Problems--
The mechanical integrity of the spiral-wound modules was generally good
but two problems did occur with prolonged operation. One problem was that
initially the resin impregnated fiberglass outer wrap on the modules became
loose and could easily slide off. If not protected by a strong outer wrap
the module could be punctured or damaged during handling. This problem was
overcome by changing to an epoxy which was more chemically stable than the
original impregnating resin. Unfortunately, this improved resin must be
heat cured presenting the possibility of membrane drying during fabrication.
The second mechanical problem was brine seal failure. The seals lost
strength, occasionally tore and generally degenerated to the point where feed
flow bypassed the face of the module. Under these conditions pressure drop
measurements became meaningless and actual circulation rates through the
module questionable. No successful solution to this problem was identified.
Module Flux Recovery—
As observed in Table 8 the spiral-wound modules could generally be
cleaned. What cannot be discerned from this Table is the difficulty en-
countered in the cleaning operations and the unpredictability of the results.
The occasionally week-long cleaning cycles that were needed are clearly
impractical for large-scale system operation. It is because of the
difficulty of module flux recovery that the single module tests were extended
for the major portion of this program. In this way both cleaning procedures
and pretreatment techniques could be further investigated.
The cleaning procedures employed were described previously. The dis-
cussion of cleaning effectiveness will be deferred until all membrane flux
data have been presented.
Module Color Rejection--
Color rejection exhibited by the spiral modules was lower than would
have been expected based on the laboratory screening test data. In fact,
it was seldom that module color rejection exceeded 90% (see Table 8).
Because of the significant flux decline problems which occurred and the
emphasis placed on feed pretreatment and foulant(s) identification, improved
module color rejection was not sought. The experimental design was to
concentrate on module flux improvement first. Then, if warranted, con-
tinuous flat sheet WRP membrane could be recast with better color re-
jection characteristics and fabricated into spiral-wound modules.
Pretreatment Testing and Foul ant Identification Summary
Analysis of a fouled membrane surface produced the following analytical
results:
86
-------
Volatile* at 105°C 88.0%
Kaolinlte clay 5.6%
Starch 4.0%
Titanium dioxide 1.4%
Carboxylic acid salt <1.0%
The major components of the fouling "slime" layer thus come from recycle of
white water from the paper mill back to the pulp mill. It is now known when
mill operators began to recycle white water. However, now that this water
conservation measure is in effect, the pulp mill ultrafiltration system must
be capable of treating a stream containing both clay and Ti02-
Foul ant removal was attempted with
-- Various depth filter media;
-- Chemical coagulant addition followed
by depth filtration; and,
-- Chemical coagulant addition followed
by vacuum filtration.
The chemical coagulants tested (mainly in jar tests) were Nalco coagulants
GWP-827, 7132, 107, and 8103; Chitosan; Arquad 2HT75; animal glue; lime
and acid addition with the Nalco polymers.
No satisfactory pretreatment method was identified during this program.
Detailed results of the pretreatment studies can be found in Appendix E.
It is of note that when one module was cut open for examination the
slime layer was readily removed from the membrane surface by simple
wiping. This indicates that the foulants were not bonded to the membrane
but were rejected in the gel layer adjacent to the membrane surface.
Three-Stage Pilot System Test Results
An effective pretreatment technique had not been identified by start-
up of the 3-stage pilot system. Nonetheless, the staged-pilot system was
operated for 335 hours over a 3-month period. It was hypothesized that the
first stage modules might be preferentially fouled by the white water
species, thereby limiting fouling in the second and third stages. If this
^phenomenon occurred, meaningful flux data at concentrated feed levels could
be obtained.
As was the case with the single module tests, emphasis was placed on
module flux and flux recovery rather than color rejection. The feed stream
to the staged pilot plant was caustic extraction filtrate for all but a
9-hour period. During this brief time when the caustic line was down, pine
decker effluent was processed.
87
-------
Narrative of Pilot Plant Flux Performance—
Discussion of pilot plant flux performance will proceed chronologically
through the 335 hours of system operation. To aid discussion, flux versus
time data for the 3 system stages are plotted in Figures 37, 38, 39, and 40.
Typical overall system conversions are indicated in the Figures. Routinely
95% to 97% conversion (20 to 30-fold volume reduction) was maintained.
Pilot system startup was with one module per stage. During the first
90 hours (see Figure 37) only limited flux data were available for stages 1
and 2 as permeate production was below scale on the rotometers. Initial
process flux levels were 0.82 to 1.0 m3/m2-day (20 to 25 gfd) for all three
stages. The stage 1 and 2 flux levels dropped off scale rapidly while the
stage 3 flux followed the pattern observed in the single module tests:
rapid initial flux loss then a more gradual decline over the next 20 hours.
A single 2.5 hour caustic/EDTA wash was performed after 25 hours exposure
to the waste stream. Based on third stage process flux of 0.62 m3/m2-day
(15 gfd), the cleaning was only partly successful.
A much more gradual decline in process flux is observed in the next
processing period: 26 to 48 hours. Here a 12-hour wash encompassing 2
caustic/EDTA cycles and a 4.5 hour ultraclean cycle, recovered stage 2 and
3 water flux (darkened symbols) to acceptable levels. Reintroduction of
the caustic extraction filtrate brought stage 3 flux quickly to 0.51 m3/m2-
day (12.5 gfd). Flux then gradually declined to approximately 0.41 m3/m2-
day (10 gfd). An ultraclean/caustic wash at this point was ineffectual,
limiting process data over the subsequent 10 hour period.
After a total of 81 operating hours the stage 3 module was removed and
stages 1 and 2 received a 2-hour ultraclean/EDTA wash. Flux recovery was
0.62 m3/m2-day (15 gfd) for stage 1 and 0.49 m3/m2-day (12 gfd) for stage 2.
Six hours following return to the process stream, stage 2 flux began
to rise and its permeate became very dark in color. This module was removed
and replaced. No module failure autopsy was performed.
Pilot system operation continued (see Figure 38) with the original
module in stage 1, a new module (previously in stage 3) now in stage 2 and
no modules in stage 3. Two ultraclean washes (one with EDTA) produced flux
levels of 0.90 m3/m2-day (21.9 gfd) and 1.76 m3/m2-day (43.0 gfd) in stages
1 and 2, respectively. An uninterrupted 44-hour processing period showed
continually decreasing flux for both stages with the stage 1 permeate output
approximately 50% below that of stage 2. This inversion may indicate
greater deposition of white water foulants on the stage 1 membrane surface.
A 2-hour ultraclean/EDTA wash brought stage 1 flux to 0.41 m3/m2-day
(10.0 gfd) and stage 2 flux to 1.21 m3/m2-day (29.5 gfd). Process flux
immediately fell to <50% of the water flux for both stages and another
2-hour ultraclean wash was initiated. Flux recovery for stage 1 improved
slightly to 0.49 m3/m2-day (12 gfd); for stage 2 only 0.82 m3/m2-day
88
-------
CO
csi
LU
C9
a:
UJ
o.
1.2
1.0
0.8
0.6
0.4
0.2
0
I 1 1 1 (31.1) A 1 ' '
D
1 module
w per stage
f/
\/
\ Note: limited flux data for stages 1 and 2 during
9A first 90 hours of pilot system operation.
-OX
\
X
\
/TSft^ ^A^
X ^^ ^
% (^""-A ^^^-^A
_ j. /G(W\ "**" ~~
cxl |> \_/ ""
O Stage 1 CX3 A
D Stage 2 1
- A Stage 3 CX1
Darkened symbols indicate water flux
CXN indicates number of cleaning cycles
Inlet pressure: 5.1 atm (75 psig) Flux temperature corrected to 50°C
O Indicates typical overall system conversion
1 1 1 1 1 1 1
^•w
Stage 2 _
module
failed, \
replaced, v^
C
• f-
/
• /
X ~~
/ \^
a' xv
t ^
CXI
Stage 3
module removed
1
30
25
20
1
15
10
5
0
10
20
30 40 50 60
CUMULATIVE OPERATING TIME (HOURS)
70
80
-o
m
CO
90
Figure 37. Flux history during 3-stage pilot system operation (0 to 90 hours).
-------
1.2
1.0
~ 0.8
uj
l/J
on
LU
o.
0.6
0.4
0.2
/
/
• (43.0])
New module in stage 2.
A
~L
90
1
CX2
1 j
v
cxi
a
O o
1
r - 1
o stage i,
D Stage 2
?^kened symbols 1ndicate water
25
Inlet pressure: 5.1 atm (75 psig^
Flux temperature corrected to 50*
O indicates typical overall
system conversion
100
no
120 130 140 150
CUMULATIVE OPERATING TIME (HOURS)
160
20
°N
1 P1ne 1
N- decker - ^
effluent
processing
1
x
5
10
170
180
Figure 38. Flux history during 3-stage pilot system operation (90 to 180 hours).
-------
1
I
a:
LU
a.
x
1 .£.
1.0
0.8
0.6
0.4
0.2
o
till
/•>
°s
\
\
Caustic \ Removed one
- . extraction S modu1e from Sta9e 3
/ filtrate \ \
v 1
\t \ 1
\ *
V
\
O
\
Replaced stage 1
module one new / — v Q
module in stage 1 v^/ \
one new module \™ \__/ ^
added to stage 2. \ \
Two modules in ^ ta
stage 3 CX4 ^ \ x
Stage 2 and 3 pumps ^ >» n,^ -^ ^
— replaced "* v x **- »
\^. •^.
x ^
^ —D ^v X^
^5?) ^xss
5- -^ ^"^ XAV
CXI
1 1 1 1
180 * 190 200 210 220
. . -, 1 JU
1 ' O Stage 1
A D Stage 2
A A Stage 3
Darkened symbols indicate water
flux
CXN indicates number of cleaning- 25
cycles
Inlet pressure: 5.1 atm (75 psig)
Flux temperature corrected to 50°C
O indicates typical overall
system conversion
- 20
A
M
•
• A
•
-V \ D\ • " 10
x S"^D N\T~ "'•" \x
y xx° A~\-(t °vx D ^~~~~ ~~ -^
^ vfy cxi /— \xo °~~~ ~ - -*-s
T 4syf (95^
vA I 1 J
CX2
1 I 1 1 0
230 240 250 260 270
X
-o
CT
m
CUMULATIVE OPERATING TIME (HOURS)
Figure 39. Flux history during 3-stage pilot system operation (180 to 270 hours).
-------
ro
1 .£
1.0
0.8
.— X
>•
s
1
CM
^
<*>
* 0.6
LU
1
O _
\
\ •
\
\ \ A /-\
^ \ ^ ^ CX12 \gvfi
<^S\\ ^^^"x i
<^ S% """^>, ^B^^Q
^"^^ o^^ (m} "*^u t ~
^o ^ ^ ^ cxs
"* ^ *^o
III! II
0
5
0 P
X
m
—i
3
5 S
UJ
3
10
3
1
290 300 310 320 330 340 350 360
CUMULATIVE OPERATING TIME (HOURS)
Figure 40. Flux history during 3-stage pilot system operation (270 to 335 hours)
-------
(20 gfd) was recorded. Process flux for stage 1 was below 0.21 m3/m2-day
(5 gfd) within a few hours after washing necessitating a further cleaning
step. This time two ultraclean/EDTA cycles (4.5 hours total duration) were
employed. Essentially no change in either stage water flux was observed.
Process flux for both stages began at <0.41 m3/m2-day (10 gfd) and
continued to decline over the next 20-hour period. Two ultraclean cycles
(2.5 hours) and two ultraclean/EDTA cycles (6 hours) recovered stage 1
water flux to only 0.41 m3/m2-day (10 gfd) and stage 2 water flux to 0.77
m3/m2-day (18.7 gfd).
Pine decker effluent was now processed as the caustic extraction
filtrate line was down. The flux pattern remained unchanged and a single
2.5 hour ultraclean/EDTA wash produced the typically low water flux levels.
A return to caustic extraction filtrate showed routine flux losses
over the next operating period (180 to 197 hours, see Figure 39). By the
conclusion of this processing period final flux for stage 1 was only
0.09 m3/m2-day (2.1 gfd). System operation was now interrupted as new
circulation pumps were installed in the second and third stages. Three
ultraclean washes and one ultraclean/EDTA wash (9.5 hours, in total) were
then performed with flux recovery to 0.30 m3/m2-day (7.3 gfd) for stage 1
and 0.64 m3/m2-day (15.6 gfd) for stage 2.
The stage 1 module, which had experienced nearly 200 hours of exposure
to the mill effluents appeared to be irreverisbly fouled and was replaced
with a new module. The pilot system was now operated with 1 module in
stage 1, 2 modules in stage 2 and 2 modules in stage 3. The next 20 hours
operation period began with process flux levels of 1.18 m3/m2-day, 0.50 m3/
m2-day, 0.34 m3/m2-day (28.8 gfd, 12.1 gfd, and 8.3 gfd) for stages 1,
2 and 3, respectively. Flux for all stages declined but the final stage 1
flux was still relatively high (0.80 m3/m2-day) [19.6 gfd].
After a 1.5 hour ultraclean wash, one module was removed from stage 3
and the system returned to caustic extraction filtrate processing. For all
stages process flux began below 0.74 m3/m2-day (18 gfd) and measured 0.21 to
0.41 m3/m2-day (5 to 10 gfd) within 15 hours.
The subsequent 3 processing periods showed less than 0.41 m3/m2-day
(10 gfd) flux for each stage. Also flux recovery was consistently worse
for stage 1 than for stages 2 and 3.
A 2-hour ultraclean wash cycle was conducted both before and after a
pump failure at the 279-hour marks (Figure 40). Resultant process flux at
each stage exhibited some initial improvement but after 20 hours was 0.12
to 0.25 m3/m2-day (3 to 6 gfd) for all stages. A single 2-hour ultraclean
wash followed, producing low water fluxes. Subsequent process flux declined
typically.
93
-------
A series of 12 wash cycles, as listed in Table 9, recovered overall
system water flux to 1.42 m3/m2-day (34.6 gfd). These wash cycles were
mainly performed with ultraclean and totaled 85.5 hours exposure of membrane
to cleaning solutions.
It was noticed at this time that some improvement in water flux
occurred when one, rather than two modules was present in a housing. A
check of pressure drop at various flow rates through both a single module
and a pair of modules in series produced the expected results. This ruled
out the possibility of seal failure in the second modules allowing feed to
bypass the membrane surface and giving an overall lower flux. While this
problem was discussed extensively, it was not resolved.
A final 20-hour processing period was initiated with stages 2 and 3
process flux once again in the 0.21 to 0.41 m3/m2-day (5 to 10 gfd) range.
The ending water flux values, after 4 ultraclean washes (28 hours) and 1
caustic/EDTA wash (4.5 hours) were 1.69 m3/mz-day (41.2 gfd) for stage 1,
1.04 m3/m2-day (25.3 gfd) for stage 2 and 1.14 md/m2-day (27.9 gfd) for
stage 3.
Summary of Pilot Plant Flux Performance—
The several hundred hour pilot plant test reconfirmed the position that
spiral-wound module systems - modules and pretreatment - are not currently
suitable for integrated kraft mill effluent processing. In non-integrated
mills, where the pulping operation is independent of the paper making
process, severe membrane fouling may not occur since white water will be
absent. During this program, however, there was no mechanism for segre-
gating the bleachery and decker effluents from white water recycle to test
this hypothesis.
General conclusions which can be drawn from 3-stage pilot plant
operation on caustic extraction filtrate are:
— At constant conversion, module flux declines rapidly (within
24 hours) to economically unacceptable flux levels.
-- Except in instances where new modules were installed in a
particular stage, flux levels per stage were not significantly
different after a few hours operating time. No stage was able
to maintain a flux rate of >0.61 m3/m2-day (15 gfd) and typically
performance was 0.21 to 0.41 m3/m2-day (5 to 10 gfd).
— There is some evidence to suggest that the first stage modules
are more heavily fouled than the subsequent stages. Flux
recovery for the first stage being consistently below that for
stages 2 and 3.
~ System flux and flux decline were basically unchanged during a
brief processing period with pine decker effluent.
~ Individual module exposure time to the waste stream had a greater
influence on module flux than the degree of individual stage or
overall system conversion.
94
-------
TABLE 9 . SEQUENCE OF CLEANING SOLUTIONS APPLIED TO 3-STAGE
PILOT SYSTEM AFTER 320 HOURS OPERATING TIME
1.
2.
3.
4-
5.
6.
7.
8.
9.
10.
11.
12.
Cleaning solution
Caustic/EDTA*
Ul trad ean/enzyme
Ultraclean
Ultraclean
Caustic/EDTA
Ultraclean
Ultracl ean/enzyme
Ultraclean/EDTA
Ultraclean
Ul tracl ean/enzyme*
Ultraclean
Ultraclean
Double brine seals placed
module in stage 1
Cleaning
duration, hrs
2.25
1.75
15.75
1.5
3.0
16.5
2.0
3.0
14.75
4.0
3.5
17.5
on each module, only one
Overall system
water flux,
m3/m2-day (gfd)
—
0.32 (7.9)
0.65 (15.8)
—
0.80 (19.6)
0.78 (19.0)
__
0.82 (19.9)
0.89 (21.6)
0.91 (22.3)
0.96 (23.5)
1.04 (25.4)
1.42 (34.6)
* Stage 1 backwashed with ultraclean after this cleaning cycle.
95
-------
-- No consistently effective cleaning solution was identified.
— Daily cleaning appears necessary. The duration of each cleaning
cycle would be a minimum of 2 hours, most likely significantly
longer.
It is also important to note that no increase in pressure drop across a
stage was observed. This indicates module plugging was not a factor in
determining system performance.
Pilot Plant Color Rejection--
Color concentrations of the feed into the pilot system (after the
Broughton filters), the recycle stream within each step, the reject stream
from each stage and the permeate stream from each stage were routinely
recorded during the 335 hours of testing. However, the low rejection
performance of the spiral-wound modules (typically 70% to 90%) during the
single module tests foretold that meaningful color rejection data would not
be obtained during 3-stage pilot plant operation.
The complete color rejection data set is presented in Appendix F and
summarized in Table 10. The average feed color concentration was 18,000
c.u. Based on the feed, average rejections were 67.5% for stage 1, 63.6%
for stage 2 and 34.6% for stage 3. These values are quite low and would not
be acceptable for full-scale operation. Even on a concentrate basis
(rejection based on the stream the modules are actually exposed to)
rejections were only 73.6%, 84.7% and 91.1% for stages 1, 2 and 3, respec-
tively.
FIELD EXPERIENCE WITH TUBULAR ASSEMBLIES
Introduction
The inability to remove stream foul ants to levels suitable for success-
ful spiral-wound module operation led to an investigation of polysulfone
membranes in a tubular geometry. The basic considerations which identified
tubular assemblies as a viable alternative to spiral-wound modules were:
-- Higher superficial velocity over the membrane surface is
achievable with the tubular configuration. This leads to
more turbulent flow and could minimize the gel concentration
layer at the membrane/liquid interface.
~ Tubular assemblies are more readily cleanable than spiral-
wound modules, again because of increased feed superficial
velocity. Also, tubular assemblies can be cleaned by
mechanical means (sponge ball circulation through the tubes),
if necessary.
Three forms of tubular assemblies were evaluated:
-- 12.7 mm (0.5 in) diameter tubes;
— 12.7 mm (0.5 in) diameter tubes with volume displacers; and
-- 25.4 mm (1 in) diameter tubes.
96
-------
TABLE 10. SUMMARY OF 3-STAGE PILOT SYSTEM COLOR REJECTION
AND PERMEATE QUALITY DURING CAUSTIC EXTRACTION FILTRATE PROCESSING
Average
Stage Feed
stream
1 18,000
2 18,000
3 18,000
color concentration (color units)
Recycl e
stream
24,200
50,100
159,000
Reject
stream
28,200
61 ,400
206,000
Permeate
stream
6,080
7,210
16,300
Average rejection (%)
Feed Concentrate (recycle
basis stream) basis
67.5
63.6
34.6
73.6
84.7
91.1
-------
The 12.7 mm (0.5 in) diameter tubes were tested first since they could be
readily bob-cast and then inserted in epoxy-reinforced fiberglass support
tubes. Successful operation of these tubes made change over of the 3-stage
pilot system to tubular modules of prime interest. To accomplish such a
change over rapidly, and within prevailing budgetary constraints, required
use of existing pumps and piping. This, in turn required tubular modules
with turbulence promoting spheres since they could be operated at lower
flowrates than open-channel tubes. Before fabrication of tubular modules
with turbulence promoters, this concept was tested at the Canton Mill to
determine if its flux characteristics paralleled those of the open-channel
12.7 mm (0.5 in) diameter tubular membranes. Finally, the development of
polysulfone casting solutions which could be cast directly onto fiberglass
supports allowed the evaluation of 25.4 mm (1 in) diameter tubular assemblies.
12.7 mm Diameter Tubular Assemblies
General--
Four 12.7 mm (0.5 in) x 1.2 m (4 ft) long tubular membranes were pre-
pared from the same WRP membrane formulation developed for flat sheet casting.
Tubular assemblies were made by inserting the membranes in support tubes and
securing them with grommets and expanders. The developmental nature of
tubular polysulfone casting led to wide variation in the individual tube flux
and rejection characteristics. Initial performance data for these tubes is
given in Table 11. Water flux ranged from 1.0 to 5.49 m3/m2-day (24.6 to
134 gfd) (50°C) and total solids rejection ranged from 43% to 86%.
The single module test stand was modified to run the tubular assemblies
and the 4 tubes were connected in series; tube M4 in the lead position.
Because of either improper membrane insertion or jarring during shipment,
the M4 membrane was stripped from its backing with the initial flow of water.
This tube was removed, tube C2 was stored as a spare, and the test stand was
restarted with tubes C5 and 01 connected in series. These tubes were
operated for nearly 500 hours with no further mechanical problems occurring.
Performance Characteristics--
Figure 41 shows individual tube flux and color rejection as functions of
operating time throughout their field evaluation. Also plotted is concen-
trate flow rate versus time. Since the caustic extraction filtrate was
processed on a once-through basis, concentrate flow is equivalent to feed
circulation rate (i.e., a measure of feed superficial velocity).
Flux for the two tubes was vastly different: 0.62 to 0.82 m3/m2-day
(15 to 20 gfd) for tube 01, 1.64 to 2.26 m3/m2-day (40 to 55 gfd) for tube
C5. These data follow directly from the wide difference (3.28 m3/m2-day)
[80 gfd] in water flux for the 2 tubes. The flux trends for the 2 tubes are,
however, identical. An initially sharp decline over the first 20 hours
waste exposure followed by a stabilization of flux for the remainder of
the test (472 hours).
During the entire processing period, which stretched over 50 days, the
system was shut down many times. Following each shut down the system was
flushed with water. At np_ time was detergent or mechanical cleaning employed.
98
-------
700
VO
.8
§§16.4
o _i
O- C5
°- 01
caustic extraction filtrate
Inlet pressure: 3.4 to 4.1 atm (50 to 60 pslg)
Temperature: 38 to 51°C
200 250 300 350
CUMULATIVE OPERATING TIME (HOURS)
400
450
4.0^
3.0"
Figure 41. Performance characteristics of 12.7 mm diameter WRP membrane assemblies.
-------
TABLE 11. INITIAL PERFORMANCE CHARACTERISTICS
OF 12.7 MM DIAMETER WRP TUBULAR ASSEMBLIES*
Membrane
designation
C2
C5
M4
01
Water flux @
50°C,
m3/m2-day (gfd)
5.49 (134)
4.88 (119)
1.01 (24.6)
1.47 (35.9)
Total solids removal
efficiency**,%
43.2
53.6
85.5
65.2
~*Operating conditions were:Feed circulation rate: 16.4 to 21.8 m3/day
(3 to 4 gpm)
Inlet pressure: 6.1 atm (75 psig)
** Feed solution for removal efficiency measurement was a 2 wt. % solution
of Carbowax 6000M.
100
-------
The average flow through the membranes was 24.5 m3/day (4.5 gpm). This
corresponds to a superficial velocity of 2.26 m/sec (7.4 ft/sec). It can be
hypothesized that the flux decline at the start of the run (0 to 25 hours)
is associated with the drop in circulation rate below 21.8 m3/day (1.98 m/sec)
[4 gpm (6.5 ft/sec)] and a build up in the membrane gel layer. This theory
does not hold, however, since the flux recovered at the 25 hour mark while
the flowrate continued to decline (through 46 hours). Hence, if a critical
feed superficial velocity (below which fouling occurs) exists with tubular
assemblies it was not identified during this test.
Color rejections for the two tubular assemblies were not uniform. The
high flux tube, C5, showed 91% to 99% color rejection averaging 97% to 98%
rejection. The lower flux 01 tube color rejection varied from 77% to 95%.
Thus the WRP tubular assembly color rejection was superior to that of the
spiral wound modules in both cases. Also, the exceptional color rejection
of the C5 assembly equalled laboratory screening test results with this
formulation polysulfone.
12.7 mm Diameter Tubular Assemblies with Turbulence Promoters
It was of interest to test the 12.7 mm (0.5 in) WRP tubes with turbulence
promoters because the pumping capacity of the staged pilot system was too low
for open-channel tubular modules. The type of module available would have
had too high a pressure drop at flows of 21.8 m3/day (4 gpm) and greater.
Tubes packed with turbulence promoting spheres require lower circulation rates
to achieve turbulent flow. If successful operation was realized, the pilot
system could be outfitted with packed tubular modules without major
modification.
Two 12.7 mm (0.5 in) x 0.61 mm (2 ft) tubes with turbulence promoters
were prepared. While it is difficult to calculate superficial velocity levels
in a packed tube, experience has shown that a flow rate of 2.73 to 5.45 m3/
day (0.5 to 1 gpm) is comparable to a 24.5 to 27.3 m3/day (4.5 to 5 gpm)
open-channel flow rate.
Table 12 shows typical data for the tests with packed tubes. The total
flow rate to the system started at 4.9 m3/day (0.9 gpm) and during the 30
hour test declined to 2.73 m3/day (0.5 gpm). The pressure drop across the
2-tube module became so high that the flow could not be kept at its initial
level. Initially the pressure drop was 3.8 to 3.9 atm (56 to 58 psig).
Ending pressure drop was 4.6 to 4.8 atm (68 to 70 psig). Since at the lower
flowrate a lower pressure drop is expected, plugging of the packed tube was
evident and the test was terminated.
Even with this plugging problem, flux decline was not as severe as with
spiral-wound modules.
Color rejection for both WRP tubes was 88% to 93%.
101
-------
TABLE 12. PERFORMANCE CHARACTERISTICS OF WRP TUBULAR ASSEMBLIES WITH TURBULENCE PROMOTERS
o
ro
Cumulative
operating
time,
hours
0
0.6
1.6
2.6
3.6
4.6
5.6
6.6
7.6
23.3
24.3
25.3
26.3
27.3
28.3
29.5
30.3
Inlet
pressure,
atm (psig)
5.2 (76)
5.0 (74)
5.0 74)
5.1 75)
5.2 76)
5.2 (76)
5.2 (76)
5.2 (76)
5.2 (77)
5.4 (80)
5.4 (79)
5.4 (79)
5.2 (77)
5.2 (77)
5.3 (78)
5.3 (78)
5.4 (80)
Pressure Circulation
drop, atm flow rate,
(psig) m3/day (gpm)
Flux @ 50°C,
m3/m2-day (gfd)
Tube 1
Color
Tube 2 Tube
rejection (%)
1 Tube 2
3.9 (58) 4.9 (0.90)
3.8 (56) 4.8 (0.88)
3.8 (56) 4.7 (0.87)
3.9 (58) 4.7 (0.86)
4.5 (
4.5 (
4.5 (
4.5 (
56) 4.6 (0.85)
56) 4.4 (0.80)
56) 4.2 (0.77)
56) 4.1 (0.75)
4.5 (66) 4.1 (0.75)
5.0 (73) 3.3 (0.60)
4.7 (69) 3.3 (0.60)
4.7 (69) 3.3 (0.60)
4.6 «
4.6 ((
4.6 ((
4.6 (
4.8 ]
57 3.2 (0.58)
57 3.1 (0.57)
58 3.1 (0.56)
58) 2.9 (0.53)
rO) 2.9 (0.53)
1.60 (39.0) 0.69 (17.0) 89
88
1.64 (39.9) 0.64 (15.6)
1.60 (39.0) 0.62 (15.0) 90
1.60 (39.0) 0.56 (1
1.60 (39.0) 0.46 1
1.54 (37.5) 0.56 1
1.57 (38.4) 0.60 1
3.6
1.3 92
3.6
4.7)
89
—
89
—
—
1.54 (37.5) 0.77 (18.7)
1.23 (30.0) 0.52 (12.7) 93
1.23 (30.0) 0.52 (1
1.20 (29.3
1.20 (29.3
1.17 (28.5
1.17 (28.5
0.49 (1
0.49 1
0.49 1
0.49 1
2.7)
2.0)
2.0) 93
2.0)
2.0) 93
92
--
—
91
--
92
1.11 (27.0) 0.43 (10.6)
1.11 (27.0) 0.46 (1
1.3)
—
-------
25.4 mm Diameter Tubular Assemblies
General--
The development of 25.4 mm (1 in) diameter x 1.52 m (5 ft) long tubular
polysulfone membranes by Abcor occurred during the final months of this
program. These membranes were cast from a proprietary formulation not of
the WRP-series. Two membrane assemblies were tested, in series, on caustic
extraction filtrate at 3 concentration levels. The concentration levels,
simulating individual stage conversions, were:
Concentration
Factor Conversion
1.2X 16.7%
10X 90%
BOX 98%
Each concentration level was maintained for about a 1-week period. During
the 10X and BOX runs low-molecular weight cut-off HFM tubular assemblies
were operated in series with the polysulfone tubes. These membranes were
"tighter" than standard HFM membranes but unlike the membranes used in the
screening tests did not have any coating treatments.
Performance Characteristics--
Flux--Tubular polysulfone membrane flux versus time is plotted in
Figure 42 for the 1.2X (16.7% conversion) test period. The two tubular
assemblies showed similar flux patterns over the 132 hour run even though
one tube (tube B) consistently had 0.41 to 0.62 m3/m2-day (10 to 15 gfd)
higher flux than the other tube (tube A).
After a 10% flux loss in the first 5 hours of the run, flux stabilized
for 15 hours at 27.5 m3/m2-day (67 gfd) for tube A and 3.65 m3/m2-day (89
gfd) for tube B. Flux declined gradually over the next 55 hours to 2.71
m3/m2-day (54 gfd) for tube A and 2.79 m3/m2-day (68 gfd) for tube B. This
flux decline is associated with increased total solids levels (see Figure 42)
In the final 60 hours of the run very stable flux was achieved except for
one unexplained flux increase between the 75 and 85 hour mark. Once
stabilized average flux levels were 2.26 m3/m2-day (55 gfd) and 2.87 m^/m^-
day (70 gfd) for tubes A and B, respectively.
During the 1.2X concentration test the feed circulation rate was varied
between 109 and 136 m3/day (20 and 25 gpm). This resulted in a change in
feed superficial velocity from 3.1 m/sec (136 m3/day) [10.2 ft/sec (25 gpm)]
to 2.49 m/sec (109 m3/day) [8.17 ft/sec (20 gpm)]. The first time the feed
velocity was lowered permeate flux declined slightly (see Figure 42). Upon
increasing feed velocity again flux stabilized. A return to the lower feed
velocity for the final 20 hours of the test produced no effect on flux
stability. Overall minimal flux loss is observed at the lower flowrate,
however further tests will be necessary to fully document the effect of
lower circulation rates.
103
-------
109 m3/day
(20 gpm)
Tube A
Tube B
caustic extraction filtrate
Feed circulation rate: as noted
Inlet pressure: 5.1 atm (75 psig)
Permeate flux temperature corrected to 50°C
40 60 80 100
CUMULATIVE OPERATING TIME (HOURS)
Figure 42. 25.4 mm diameter tubular polysulfone membrane flux versus time during
17% conversion test period.
-------
At both 109 and 136 m3/day (20 and 25 gpm) the feed superficial velocity
is greater than the 2.26 m/sec (7.4 ft/sec) employed in the 12.7 mm (0.5 in)
diameter tubular tests. Thus, it may be possible to lower the circulation
rate still further without significant flux loss or membrane fouling result-
ing. This is a very important consideration since system power requirements
are greatly increased at higher flowrates. In fact, a reduction of 27.3 m3/
day (5 gpm) in feed circulation from 136 m3/day (25 gpm) to 109 m3/day (20
gpm) would result in greater than a 20% decrease in ultrafiltration system
pumping power.
Polysulfone membrane flux at a 10X concentration factor (90% conversion)
is shown as a function of time in Figure 43. Again the two tubes showed
slightly different flux levels while following similar flux curves. Gradual
flux loss is observed as the concentration within the membrane loop is
allowed to reach 10X. After this point, as a steady conversion was main-
tained, the flux stabilized at 2.05 m3/m2-day (50 gfd) for tube A and
2.49 m3/m2-day (60 gfd) for tube B. These flux levels were stable for the
final 45 hours of the test.
The 50X concentration test flux data are shown in Figure 44. Tubes A
and B exhibited essentially identical flux during this run. Therefore, for
clarity, the average flux of the two tubes is plotted in Figure 44. Flux
declined gradually over the 180 hour concentration period from 2.87 to
1.03 m3/mz-day (70 to 25 gfd). As the concentration was held steady at 50X
for the next 90 hours the permeate flux stabilized between 0.94 to 1.07
m3/m2-day (23 to 26 gfd).
Low molecular weight cut-off HFM tubular assemblies were run in series
with the polysulfone tubes in the tests at 90% and 98% conversion. Figure 45
shows the flux data for the 90% conversion run. One tube, tube D, failed
within 24-hours and was removed. Tube C was operated for the full 135 hour
test. Its initial flux was 9.02 m3/m2-day (220 gfd). This declined sharply
to 4.1 m3/m2-day (100 gfd) within 20 hours, stabilized for nearly a day's
time then began a constant, gradual decline. Final flux was 1.7 m3/m2-day
(42 gfd), approximately 30% below the average polysulfone membrane flux
during the same test.
HFM flux data during the 98% conversion test are plotted in Figure 46.
Flux began at 2.05 m3/m2-day (50 gfd) and declined to essentially zero in
200 hours. At this point no further HFM flux data were collected.
Flux Recovery—Table 13 summarizes polysulfone tubular membrane flux
recovery. After both the 1.2X and the 10X runs (133 and 268 cumulative
operating hours, respectively) water flux was recovered to at least initial
values with only a 2-hour ultraclean wash. The membranes were not washed
at the Canton Mill following the 50X concentration run. Rather, they were
partially filled with water, sealed and returned to Abcor. These membranes
were not used again for 3 months. At that time they received a 1-hour
105
-------
4.0
100
o
en
System
off/on
System
off^on
Caustic extraction filtrate
Feed circulation rate: 136 m-Vday (25 gpm)
Inlet pressure: 5.1 atm (75 psig)
Permeate flux temperature corrected to 50°C
0
Figure 43.
60 80 100
CUMULATIVE OPERATING TIME (HOURS)
25.4 mm diameter tubular polysulfone membrane flux versus time during
90% conversion test period.
140
m
73
-------
CVJ
X
5
Tubes A and B, average flux
caustic extraction filtrate
Feed circulation rate: 136 m3/day
(25 gpm)
Inlet pressure: 5.1 atm (75 psig)
Permeate flux temperature corrected
to 50°C
50X Concentration
Figure 44.
120 160 200
CUMULATIVE OPERATING TIME (HOURS)
25.4 mm diameter tubular polysulfone membrane flux versus time during
98% conversion test period.
280
-o
m
70
m
CD
-n
o
-------
o
o>
Tube D failed;
removed
System
off/on System
off/on
A Tube C
Tube D
caustic extraction filtrate
Feed circulation rate: 136 m3/day (25 gpm)
Inlet pressure: 5.1 atm (75 psig)
Permeate flux temperature corrected to 50°C
60 80 100
CUMULATIVE OPERATING TIME (HOURS)
Figure 45. 25.4 mm diameter tubular HFM membrane flux versus time during 90% conversion
test period.
-------
100
o
VO
A. Tube C
caustic extraction filtrate
Feed circulation rate: 136 m3/
day (25 gpm)
Inlet pressure: 5.1 atm (75 psig]
Permeate flux temperature
corrected to 50°C
80 120 160 200
CUMULATIVE OPERATING TIME (HOURS)
Figure 46. 25.4 mm diameter tubular HFM membrane flux versus time during 98% conversion
test period.
-------
TABLE 13. FLUX RECOVERY FOR 25.4 MM DIAMETER TUBULAR POLYSULFONE MEMBRANES
Cumulative
operating
time, hrs
0
133
268
541
Inlet
pressure, atm
(psiq)
6.1 (75)
6.1 (75)
6.1 (75)
3.4 (50)
Water flux @
Tube A
5.33 (130)
5.21 (127)
5.45 (133)
1.76 (43)
50°C, m /m -da,
Tube B
5.86 (143)
6.72 (164)
7.05 (172)
2.38 (58)
y (gfd)
Cleaning conditions
New membranes.
2-hour ultraclean wash.
2-hour ultraclean wash.
Membranes stored for 3-
months without washing.
Then washed for 1-hour
with ultraclean/caustic.
Partial membrane drying
suspected.
Note: lower operating
pressure.
-------
ultraclean/caustic wash. Resultant water fluxes were 1.76 m3/m2-day (43 gfd)
for tube A and 2.38 m3/m2-day (58 gfd) for tube B. However, these data were
recorded at only 3.4 atm (50 psig) operating pressure. Also, partial membrane
dry-out is suspected.
HFM tubular membrane flux averaged 24.6 m3/m2-day (600 gfd) initially.
Subsequent to the 10X concentration run tube C flux was recovered to only
47.6 m3/m2-day (116 gfd) with ultraclean. This membrane assembly was not
cleaned after the Canton Mill tests. Tube D failed during the 10X concen-
tration period.
Color Rejection—The polysulfone membranes in 25.4 mm (1 in) diameter
tubular assemblies produced permeate of acceptable quality at all three
system conversions tested. Permeate color concentration and membrane color
rejection during these tests are shown in Figures 47 and 48. In the 1.2X
concentration (16.7% conversion) runs the average permeate color concentration
{both tubes A and B) was 709 color units (c.u.). Throughout this test the
feed averaged 20,200 c.u. and the concentrate averaged 25,400 c.u. On a
feed basis, average color rejection was 96.5% to 97%. This range of rejection
was consistently maintained throughout the run as Figure 47 indicates.
The 10X (90% conversion) period data are also graphed in Figure 47.
Average permeate quality after reaching a 10X concentration was 1,870 c.u.
Membrane color rejection (feed basis) was consistently 88% to 90%. During
this test the average color concentrations were 19,100 for the feed stream
and 117,000 for the concentrate.
The 50X (98% conversion) data are plotted in Figure 48. Permeate quality
degrades as the concentration with the membrane loop builds up to 50X. After
reaching 50X permeate quality averages 10,000 c.u. Mean feed color concen-
tration was 23,400 throughout the test. Concentrate color, after reaching
50X, averaged 927,000. While the membranes rejected 98% to 99% of the color
they were exposed to (concentrate basis), actual color rejection on a feed
basis ranged from only 50% to 60%.
Projected Color Discharge; Mass Basis—Using the data developed during
these tests, projections can be made of color discharge on a mass basis (kilo-
grams of color per day) from each stage of a full-scale system. In develop-
ing these projections the following data (and assumptions) are used:
- The UF system consists of 3 stages. Conversions per stage
are ]&7%, 90% and 98%, respectively.
- 3,790 m3/day (1,000,000 gpd) of caustic extraction filtrate
are processed.
- Average feed color concentration is 20,000 color units.
- Average permeate color concentrations are:
m
-------
100
CQ
I—
O
UJ
<-}
UJ
a:
a:
o
_i
o
o
^n—D-O-oa
Rejection calculated
on feed basis
2.5
ro
Values shown
are average
of tubes A and B
CUMULATIVE OPERATING TIME, HOURS
Figure 47. Tubular polysulfone membrane permeate quality and color rejection
during 1.2X and 10X concentration periods (caustic extraction filtrate feed).
-------
Values shown
are average of
tubes A and B
CUMULATIVE OPERATING TIME, HOURS
Figure 48. Tubular polysulfone membrane permeate quality and color rejection
during SOX concentration period (caustic extraction filtrate feed).
-------
Stage Color Concentration (c.u.)
1 800
2 2,000
3 10,000
- 8 color units are equivalent to 1 mg/1 color discharged.
The stage 1 discharge of color is thus:
670,000 gal
day
3.785 1
gal
800 c.u.
mg
8 c.u.-l
kg
mg x 106
= 253.6 kg/day
Stage 2 color discharge is:
230,000 qal
day
3.785 1
gal
2,000 c.u.
mg
8 c.u.-l
kg
mg x 106
= 217.6 kg/day
And for Stage 3:
80,000 gal
day
3.785 1
gal
10,000 c.u.
mg
8 c.u.-l
kg
mg x 106
= 378.5 kg/day
Using a similar calculation, if no treatment were employed for color removal
9,462.5 kg/day (20,365 Ib/day) of color bodies would be discharged. The
overall color reduction in the caustic extraction filtrate stream by employing
tubular polysulfone membranes under the assumed conditions and discharging all
the permeate to sewer would therefore be:
9,462.5 kg-(253.6 kg + 217.6 kg + 378.5 kg) Q1 n
-------
The typical WRP membrane spiral wound modules used for this experimental
work had initial water flux values at 21 °C of 3.28 m3/m2-day (80 gfd) After
use and cleaning a "clean" membrane had flux values between 1.64 to 2*05
m3/m2-day (40 and 50 gfd). This value was reproducible over the number of
cleaning cycles done on the various membrane modules. The individual modules
Had insufficient time on stream to predict membrane life or long term
cleanability.
Flux decline and fouling conditions were not observed with the poly-
sulfone tubular membranes. Between experimental runs (typically one week in
duration) the tubes were cleaned with the cleaning solution found optimum for
WRP spiral wound modules. This cleaning recovered water flux values
equivalent to the initial values.
The flux decline, fouling problems and cleaning problems encountered
with the WRP spiral wound modules were not found with tubular membranes.
While no problems were observed and none are predicted from the operating
experience, long term cleanability data are not available for tubular mem-
branes operated on pulp mill and bleachery effluents.
MATERIAL BALANCE
In the course of the four week pilot run with 25.4 mm diameter tubular
membranes, samples were collected for material balance analyses. Samples
of feed, permeate and concentrate were collected for each assembly at
concentration ratios of 1.2X, 10X and 50X. Color and total solids analyses
were conducted on the as-collected samples in the Champion Laboratory.
Composited samples were analyzed chemically at Galbraith Laboratories,
Knoxville, Tenn. These analyses are presented in Appendix G.
A flow schematic for identification of the samples collected from this
system is presented in Figure 49. A material balance for this system
consists of the comparison of the mass flow and constituent composition of
the feed material - in this case, pine bleachery caustic extraction filtrate -
to the sum of the mass flows and constituent compositions of the final con-
centrate and the composited permeates from each stage.
There are a number of non-controllable parameters in operating a pilot
system on a "live" operating plant effluent. For example, the feed
composition can vary on an hour to hour and week to week basis in an un-
predictable manner. In addition, the substrate for analysis is a very
complex mixture of organic and inorganic dissolved, nuclear and agglomerated
materials which renders precise chemical analysis difficult. In addition to
these considerations, for pragmatic purposes, it was necessary to reduce the
total number of samples for analysis by preparing composites of a number of
like samples collected over a period of time. The material balance obtained
under these constraints is valuable in providing a set of guidelines and an
understanding of the system but it is not interpretable as a precise de-
lineation of the system in all respects.
115
-------
Concentrate #2 (C2)
Concentrate #3 (C,)
Feed (F)
Volume 100
Sp. Gr. 1.001
,x
X
X
\
* Volume 98 '
Sp.Gr. 1.001
X*
Permeate #1 (P,)
Volume 2.0
f Sp.Gr. 0.999
»^
1.2
_x
X
X
>
\
* Volume 9.8 £
Sp.Gr. 1.005
7"
Conrpnf ratinn fartor 1
Permeate #2 (P»)
Volume 88.2"
f Sp.Gr. 1.000
f
«w
1 >
o
^
X
X
>
~^ Volume 2.0
Sp.Gr. 1.04
X*
Conrpntratinn
factor 50
Permeate 13 (?%)
Volume 7.8
f Sp.Gr. 1.000
Permeate
Total volume 98
Sp.Gr. 1.000
Figure 49. Flow schematic for identification of samples from material balance studies with
caustic extraction filtrate (25.4 mm diameter tubular polysulfone membranes).
-------
Relative Material Balance
A relative material balance for the system during concentration to a SOX
concentration ratio is presented in Figure 50. In this figure each component
is considered as 100% in feed concentration. The amount of the component
relative to the feed concentration is presented for the final concentrate and
the composited total permeate from the system. As may be noted, the color
constituents are retained in the concentrate (92.6% of the feed). The
majority of the total solids are contained in the permeate (64.8%) as might
be expected due to the large amount of ionic and small molecular weight
organic materials in this feed stream. AIT metal ions show some rejection
over the membrane varying from rejections of 100% for aluminum to a
rejection of 7.6% for sodium.
The sulfate and chloride ion distributions can be interpreted to in-
dicate that these species are expedited in passage through the membrane
because they are identified at low levels in the concentrate. This is
discussed in more detail below. It is of interest to note that the organic
chloride is predominately in the permeate. This indicates that a large
amount of this chlorine is associated with low molecular weight materials
like chloroform which pass through the membrane easily.
A graphical display of the detailed chemical analyses by ultrafiltration
stage is presented in Figure 51. This figure provides information on the
measured values of various constituents. Examination of this data will
provide information which would be too voluminous to include here. For
example, from the €3 concentrate data one may obtain a rough equivalence
figure for conversion of color as measured in color units to actual parts per
million by weight content; e.g., dividing the color (927,000 c.u.) by the
solids content (102,800 ppm) indicates that roughly 9 color units are
equivalent to 1 ppm solution solids by weight.
Ion Rejection by the Membrane
The concentration of various elemental materials in the ultrafiltration
concentrates obtained by operating at several concentration ratios is
presented in Figure 52. The lower chart presents data for aluminum, calcium,
iron and sodium. The ion concentration ratio is the ion content in the
concentrate relative to that in the feed. If a metal were rejected 100%
by the membrane, then at a 50X concentration ratio the ion concentration ratio
should be 50X. As can be observed this is the case for aluminum. Iron is
rejected about 92%. Calcium is rejected about 65% and sodium is rejected
about 7.6%.
These rejections may be related to charges on the species and a like
electrostatic charge on the membrane surface. The rejection may also be due
to chemical complex formation by the multivalent ions with other materials
in the system. Irrespective of the explanatory mechanism, this metal
behavior has economic importance. The removal of the metals improves the
quality of the permeate and allows for broader utility of this stream for
117
-------
FEED
Water
Color
Total Solids
Aluminum
Calcium
Iron
Sodium
100
100
100
100
100
100
100
00
Sulfate 100
Ionic Chloride 100
Organic Chloride 100
CONCENTRATE
TOTAL PERMEATE
Water
Color
Total Solids
Aluminum
Calcium
Iron
Sodium
Sulfate
Ionic Chloride
Organic Chloride
2
92.6
29.9
100
68.4
96
7.6
0.4
1.4
32.3
%
Water
Color
Total Solids
Aluminum
Calcium
Iron
Sodium
Sulfate
Ionic Chloride
Organic Chloride
98
11.2
64.8
10.1
20
56.6
66.6
94.2
79.2
Figure 50. Relative material balance for 50 times concentration on pine bleachery caustic
extraction filtrate using 25.4 mm diameter tubular polysulfone membranes.
-------
Feed:
IL
Total Sol Ids, % 0.782
Color, c.u. 25,400
Aluminum, ppm 3
Calcium, ppm 45
Iron
, ppm 3
Sodium, ppm 1840
Sulfate, ppm 32
Ionic Chloride, ppm 1558
Organic Chlorine, ppm 800
Total Solids, %
Color, c.u.
A1 utnl nun) , ppm
Calcium, ppm
X I O*( 9 PP
Sodium, ppm
Sul fate , ppm
Ionic Chloride, ppm
Organic Chlorine, ppm
0.717
20.893
3
36
2 ^
1790
oc
CO
1542
501
r
./"
/^
concentration
factor 1.2 •*
s
L» permeate _]_
Total Sol Ids, % 0.425
Color, c.u. 709
_2
1.91
116,700
16
241
18
2650
13
1370
2117
r concentrate ,
jr concentration
./^ factor 10
l_^ permeate _2_
0.441
1762
Aluminum, ppm <1 <1
Calcium, ppm 2
Iron, ppm 0.4
Sodium, ppm 1360
Sulfate, ppm 15
Ionic Chloride, ppm 1510
Organic Chlorine, ppm 212
Concentration Ratio 1.2
3
0.4
960
18
1457
396
10
10.28
927,200
152
1180
92
6510
4.6
1015
7751
concentrate
concentration
factor 50
permeate
0.867
9974
<1
12
0.51
1790
15
1777
558
50
Figure 51. Analyses of pine caustic extraction filtrate at various stages of ultrafiltration using
25.4 mm diameter polysulfone tubular membranes.
-------
o
o
o 0
60
50
o
t—i
140
z:
o
I—I
5 30
20
o
o
2io
o
10 20 30
CONCENTRATION FACTOR
40
50
Figure 52. Ion concentration versus concentration ratio
for ultrafiltration of caustic extraction filtrate
(25.4 mm diameter tubular polysulfone membranes).
120
-------
recycle. The low rejection of sodium is important in ultrafiltration Re-
ducing the sodium level minimizes the osmotic pressure build-up problems on
concentration and consequently allows for operation of ultrafiltration at
high concentration levels and relatively low system pressure.
The upper chart in Figure 52 presents ion concentration data for sulfate
and chloride ion as a function of concentration factor. As presented, it
appears that these ions are removed from the concentrate as the concentration
factor is increased and also that the sulfate (with the higher charge) is
expedited out of the concentrate more readily than the chloride. The sulfate
appears to be about 20% of what would be anticipated and the chloride about
67%.
This phenomenon is not explainable with the data at hand. There are,
however, a number of possible explanations. First, in this complex system
it is possible that the analytical techniques employed are not adequate for
the purpose. Secondly, it is possible that in the concentrated material
the sulfate ion especially might be tied into a complex with other materials
and hence effectively removed from the analyticalsarrpLe, A third possible
explanation may be posited on the basis of membrane surface charges and their
influences in expediting passage of anions into the permeate.
Organic Chlorine
Analyses were made on each sample for ionic chloride, total chlorine and
also-for volatile materials. The difference between the value for a sample
of total chlorine and ionic chloride is assumed to be chlorine bonded to
organic material. The volatile content of the sample solids is assumed to be
a measure of the total organic content. The ratio of the non-ionic chloride
to the volatile content is then a measure of the chlorine content of the
organic materials. In previous work it has been demonstrated that for pine
bleachery caustic extraction filtrate color bodies the chlorine content is
about 8%.
In Figure 53 the calculated ratios of non-ionic chlorine to volatiles
(% chlorine content of the organic materials) is presented for the permeates
and concentrates as a function of concentration factor.
The organic chlorine content of the materials in the permeate appear to
rise rapidly in the initial stages of ultrafiltration concentration and then
decrease as concentration is continued. The organic chlorine content of the
material in the concentrate monatonically decreases with concentration and at
50 times concentration are approaching the anticipated value.
Removal of the low molecular weight chlorinated materials - especially
materials like chloroform - is felt to be the basis for the results obtained.
In the feed substrate there are a number of chlorinated materials possible
which should be capable of free passage through the membrane, all with organic
chlorine contents which are high (e.g. chloroform at 89.1%) compared to the
121
-------
50
20 30
CONCENTRATION FACTOR
Figure 53. Ratio of non-ionio-chlorine/volatiles versus concentration
factor for ultrafiltration of caustic extraction
filtrate (25.4 mm tubular polysulfone membranes).
122
-------
color bodies (about B% chlorine). As these small, high chlorine content
materials are removed in the permeate, the type of results obtained would be
expected.
Specific Gravity
In the course of the analyses the specific gravity for each of the
samples was measured. These results are presented in Figure 54. As pre-
sented, the specific gravity of the permeates to 50 times concentration is
constant at 1.000. The specific gravity of the concentrate increases with
increasing concentration.
The inherent density of the solids in the concentrate have been
calculated from this data and appear to be about 1.43 gms/ml.
123
-------
1.050
1.040
1.030
1.020
LU
O-
CO
1.010
1.000
0.900
20 30
CONCENTRATION FACTOR
40
50
Figure 54. Specific gravity versus concentration factor for ultra-
filtration of caustic extraction filtrate (25.4 mm
diameter tubular polysulfone membranes).
124
-------
SECTION 7
CONCEPTUAL DESIGN
INTRODUCTION
Eight design cases have been reviewed in developing economic projections
for ultra-filtration of kraft pulp mill effluents. These 8 cases, summarized
in Table 14, encompass the 3 streams of interest: caustic extraction
filtrate, pine decker and hardwood decker; two module geometries: tubular
and spiral-wound; and, two system capacities: 3,790 and 7,580 m^/day (1 MM
and 2 MM gpd). The 3,790 or 7,580 nwday (i Or 2 MM gpd) caustic extraction
filtrate would come from bleaching 727 metric tons/day (800 tons/day) of
pine pulp. The 3,790 or 7,580 m3/day (1 or 2 MM gpd) decker effluent would
come from washing 1,318 metric tons/day (1450 tons/day) of mixed pine and
hardwood pulp. The system of most potential interest, a 3,790 m3/day (1 MM
gpd) tubular system treating caustic extraction filtrate (Case 1), will be
discussed in detail. The remaining seven cases will be discussed as variants
of the Case 1 design.
The flux levels used for design of the tubular systems were derived from
the experimental data. The same flux levels were used in designing the
spiral wound systems in the anticipation of the development of improved
module spacer designs.
DETAILS OF CASE 1 DESIGN
Summary
The Case 1 design bases are as follows:
— 3,790 m3/day (1 x 106 gpd) pine caustic extraction filtrate
processed.
~ tubular membrane assemblies used. Each assembly 25.4 mm (1 in)
diameter x 3.05 m (10 ft) long, interconnected to form a 6.1 m
(20 ft) length. Each parallel pass of membranes containing 8
6.1 m (20 ft) lengths of membrane connected in series with U-
bends. Total membrane area per parallel pass is 3.27 m^ (35.2
ft2).
-- prefiltration consists of a hydrasieve for fiber removal and a
backflushable 5-10 u sock filter.
125
-------
TABLE 14. FULL-SCALE SYSTEM DESIGN CASES
PO
Case
number
1
2
3
4
5
6
7
8
Feed
stream
Caustic extraction
filtrate
Caustic extraction
filtrate
Pine and hardwood
decker
Pine and hardwood
decker
Caustic extraction
filtrate
Caustic extraction
filtrate
Ptne and hardwood
decker
Pine and hardwood
decker
..Flowrate,
m3/day (MM gpd)
3,790 (1)
7,580 (2)
3,790 (1)
7,580 (2)
3,790 (1)
7,580 (2)
3,790 (1)
7,580 (2)
Quantity of pulp
produced, metric tons/day
(tons/day)
727 (800)
727 (800)
1,318 (1,450)
1,318 (1,450)
727 (800)
727 (800)
1,318 (1,450)
1,318 (1,450)
Module
type
tubular
tubular
tubular
tubular
spiral -wound
spiral -wound
spiral -wound
spiral -wound
Overall
system
conversion
98%
98%
95%
95%
98%
98%
95%
95%
-------
-- the ultrafiltration system consists of 3 stages. These stages,
detailed in Table 15, contain a total of seven subsystems.
-- a 10% excess has been calculated into the membrane area require-
ments as a safety factor.
— a single^cleaning station is included with the UF system. This
station is capable of automatic cleaning and rinsing of any one of
the subsystems at any time.
— the UF permeate is collected in a holding tank (1 hour residence).
From here it is recycled within the mill, used for system cleaning
operations or sewered.
— the UF concentrate is returned to the weak black liquor, to the
lime kiln or to landfill, or mixed with lime sludge.
Flow Schematic Description
A flow schematic of the Case 1 system is shown in Figure 55.
A flow of 3,790 m3/day (700 gpm) of first stage pine caustic extraction
filtrate is continuously fed into a hydrasieve" for fiber removal. While
3,790 m3/day (700 gpm) is the average feed flow rate, it is expected that the
system can handle reasonable flow fluctuations (±10%) with little, if any,
difficulty. Should a serious mechanical problem occur downstream of the
hydrasieve, a bypass to sewer is provided. Also provided is a port for
permeate feed to the hydrasieve to maintain the entire system in recycle
operation should the caustic line go down.
Fiber in the feed stream will be removed by the hydrasieve and recycled
to the tower. If necessary, the fiber-laden stream can be sewered. The
hydrasieve screening will have a flowrate of 91 to 364 nvVday (1,000 to
4,000 gph).
The underflow from the hydrasieve will be passed through a backwashable
finger filter by a 3,790 m3/day (4.1 atm head, 1750 RPM) [700 gpm (60 psi
head, 1750 RPM)1 centrifugal pump equipped with a 27.8 kw (40 hp) motor.
A spare 3,790 nH/day (700 gpm) pump is provided to maintain system oper-
ability during maintenance periods. The finger filter consists of
cylindrical stainless steel screens outfitted with polypropylene "socks".
Solids on the order of 5 to 10 y will be retained by the filters which are
automatically backflushed with UF permeate when a critical pressure
differential between the inlet and outlet headers is reached. Two
3,790 m3/day (700 gpm) filter systems are provided; a clean system being on
standby at all times.
The effluent from the sock filters flows into the first stage of the
UF system. This stage is made up of 3 identical subsystems each containing
327 m2 (3,520 ft2) of membrane area. A conceptual design of a subsystem is
shown in Figure 56. In its simplest terms a subsystem consists of membranes,
a membrane support rack and a pumping station.
127
-------
TABLE 15. ULTRAFILTRATION SECTION DESIGN — CASE 1
Stage number
Item
Total membrane
area, m2 (ft2)
No. of subsystems
Membrane area per
subsystem, m2 (ft2)
1
981
(10,560)
3
327
(3,520)
2
477.5
(5,140)
2
238.8
(2,570)
3
327
(3,520)
2
163.5
(1,760)
No. of parallel
membrane passes
per subsystem
100
73
50
Circulation flowrate
per subsystem, m3/day (gpm)
Pressure drop per
parallel pass,atm (psi)
Motor power, kW (hp)
Efficiency, %
Brake power
per subsystem, kW (hp)
3 2
Design flux, m /m -day (gfd)
Typical conversion,
feed basis, %
13,516
(2,480)
2.38
(35)
74.6
(100)
80
58.9
(79)
2.87
(70)
67
9,946
(1,825)
2.38
(35)
55.9
(75)
74
47.0
(63)
2.05
(50)
90
6,813
(1,250)
2.38
(35)
37.3
(50)
70
33.6
(45)
1.03
(25)
98
128
-------
ro
10
zvvs&e*** i?A&
•^soaRif
Figure 55. Proposed 3,790 m3 (1 MM gpd) UF system flow schematic.
-------
CA)
O
Figure 56. Proposed 3,790 m3 (1 MM gpd) system typical subsystem outline
drawing.
-------
The polysulfone membranes are in a tubular geometry. Two 25 4 mm
(1 in) diameter x 3.05 m (10 ft) long (current technology) membrane sections
are interconnected to form a 6.1 m (20-ft) membrane length. Threaded U-bends
"turn-around" the flow to the next tubular assembly in series. Eight
membranes (and 7 U-bends) comprise one parallel pass. The pressure drop
across a pass is 2.38 atm (136 m3/day recirculation, 54°C) [35 psig (25
gpm recirculation, 130°F)]. With a 5.1 atm (75 psig) inlet pressure, the
average operating pressure is thus 3.91 atm (57.5 psig).
Each pass of 8 membranes provides 3.27 m2 (35.2 ft2) of membrane area.
As needed, parallel passes can be added to the membrane rack to provide the
desired subsystem membrane area.
The membrane support racks are rigid, steel units which support the
weight of the membranes and the fluid being processed. A central distri-
bution header carries the feed stream to the membrane inlets. Two inlet
nipples are situated side-by-side enabling one parallel membrane pass to be
supported on each side of the central distribution header. A single-nipple
header on each side of the rack collects the concentrate and returns it to
the suction of the circulation pump. The individual tube permeates are
collected at the opposite end of the rack and manifolded. Permeate flow is
measured by a flow meter.
A skid-mounted pumping station capable of providing 136 m^/day (25 gpm)
recirculation flow (at 5.1 atm) [75 psig] through each parallel membrane
pass is provided. The centrifugal circulation pump is interlocked with a
temperature indicator-controller and equipped with high and low pressure
shutdown switches with audio and visual alarms.
The concentration bleed from the stage 1 subsystems is combined and
fed into the second stage of the ultrafiltration system. Stage 2 consists
of 2 subsystems, each containing 239 m2 (2,570 ft2) of membrane area (73
parallel passes of membranes). These subsystems are similar in design to
the stage 1 subsystem.
In a like manner, the Stage 2 concentrate will flow into the third
stage of the system for further concentration. This stage has 2 sub-
systems, each with 163.5 m2 (1,760 ft2) of membrane area (50 parallel
passes).
Typical conversions across each system stage are:
Stage Conversion
1 67%
2 90%
3 98%
The final concentrate (from stage 3) is bled off at a controlled rate by a
flow ratio controller. This controller receives signals of the feed
flowrate and ratios the final concentrate flow to maintain the desired
131
-------
overall system conversion. At a feed flowrate of 3,790 m3/day (700 gpm),
76.3 m2/day (14 gpm) of concentrate would be discharged. This flow would
be directed to either the lime kiln, to the weak black liquor, or admixed
with lime sludge.
The permeate produced by each stage is combined and flows into a 189.5
m3 (50,000 gal) holding tank. This tank provides a residence time of approx-
imately 70 minutes. The permeate is used for several functions in the
operation of the overall treatment system:
— make-up water for UF system cleaning solutions;
— rinse water for UF system cleaning;
— backflush water for sock filters;
— feed water to maintain system operation in recycle if the caustic
extraction filtrate line is down.
These uses, however, require only a fraction of the permeate flow. Some
portion of the permeate (5% to 50%) may be recycled to the bleachery. The
remainder may be blended with the plant treated water supply. Alternatively,
the permeate may be sewered.
Cleaning Sequence
Each of the 7 subsystems is cleaned independently, as needed. A single
cleaning station consisting of a 7.58 m3 (2000 gal) cleaning tank, 2.27 JIH
(600 gal) rinse tank, a cleaning pump and associated solenoid valves,
timers and piping is supplied with the ultrafiltration system.
It is anticipated that each subsystem will require one cleaning cycle
per two week period. The length of each cleaning cycle will probably vary
per stage and range from 1 to 3 hours. The conversion across each stage
will "swing" as the individual subsystems are being cleaned. This "swing"
will be handled automatically by the flow ratio controller receiving signals
from the Stage 1 feed inlet and stage 3 concentrate outlet. During the
cleaning "swing" periods the second stage of the UF system is pressed upon
to work hardest. For this reason the stage 2 membrane area is ^50% over
that needed if no subsystems were ever shutdown for cleaning. However,
even with this increase in membrane area, stage 2 will not be able to handle
the infrequent stage 1 subsystem cleaning cycles without a "swing" in
system conversion.
Overall, the system will have one subsystem in a cleaning mode less than
4% of the time.
DETAILS OF CASE 2 DESIGN
Summary
The Case 2 design bases are as follows:
132
-------
-- 7,580 m3/day (2 x 106 gpd) pine caustic extraction filtrate;
— tubular membrane assemblies;
— hydrasieve and sock filter prefiltration;
— three-stage UF system (see Table 16).
Description
The increase in system capacity from 3,790 to 7,580 m3/day (1 to 2 MM
gpd) has no effect on overall system operation. In fact, the only changes
which do occur are in the number of subsystems per stage and the capacity of
each subsystem. For the Case 2 design, 10 identical subsystems are
employed. Each subsystem contains 327 m2 (3,520 ft2) of membrane area (100
parallel passes of 8 6.1 m tubes, in series). There are 6 subsystems in
Stage 1, 2 subsystems in Stage 2 and 2 subsystems in Stage 3. As with
design Case 1 a 10% excess membrane area is provided as a safety factor.
However, unlike Case 1, the Case 2 system does not have the extra "swing"
capacity built into Stage 2. The individual stage conversion will still
shift as a subsystem is being cleaned, but a loss in overall conversion may
occur more frequently. The impact of less membrane area in Stage 2 will,
however, be relatively minor since with 6 subsystems in Stage 1, a smaller
percentage of membrane area will be off line during any one cleaning
operation.
It is important to note that many alternative designs exist for both the
3,790 and 7,580 m3/day (1 and 2 MM gpd) systems. For example, a 7,580 m3/
day (2 MM gpd) system may preferably consist of 2 identical, independent
3,790 m3/day (1 MM gpd) systems. In that way, a prolonged shutdown of any
one subsystem would have less of an overall effect. Clearly, a signifi-
cant engineering effort will be required to detail the optimum design for
any of the 8 cases under discussion in this section.
DETAILS OF CASE 3 DESIGN
Summary
The Case 3 design bases are:
— 3,790 m3/day (1 x 106 gpd) pine or hardwood decker effluents;
— tubular membrane assemblies;
— hydrasieve and sock filter prefiltration;
— three-stage UF system (see Table 17).
Description
The Case 3 system is designed to process either pine decker or hardwood
decker effluent at a rate of 3,790 m3/day (1 MM gpd). Since the decker
effluents can be recycled to the weak black liquor with less concentration
133
-------
TABLE 16. ULTRAFILTRATION SECTION DESIGN — CASE 2
Stage number
Item
Total membrane
area, m2 (ft2)
No. of subsystems
Membrane area per
subsystem, m2 (ft2)
No. of parallel
membrane passes
per subsystem
Circulation flowrate
per subsystem, m3/day (gpm)
Pressure drop per
parallel pass, atm (psi)
Motor power, kW (hp)
Efficiency, %
Brake power
per subsystem, kW (hp)
Design flux, m3/m2-day (gfd)
Typical conversion
(feed basts), %
1
1,962
(21,120)
6
327
(3,520)
100
13,516
(2,480)
2.38
(35)
74.6
(100)
80
58.9
(79)
2.87
(70)
67
2
654
(7,040)
2
327
(3,520)
100
13,516
(2,480)
2.38
(35)
74.6
(100)
80
58.9
(79)
2.05
(50)
90
3
654
(7,040)
2
327
(3,520)
100
13,516
(2,480)
2.38
(35)
74.6
(100)
80
58.9
(79)
1.03
(25)
98
134
-------
TABLE 17. ULTRAFILTRATION SECTION DESIGN -- CASE 3'
Item
Total membrane
area, m2 (ft2)
No. of subsystems
Membrane area per
subsystem, m2 (ft2)
No. of parallel
membrane passes
per subsystem
Circulation flowrate
per subsystem, nvVday, (gpm)
Pressure drop per
parallel pass, atm (psi)
Motor power, kW (hp)
Efficiency, %
Brake power
per subsystem, kW (hp)
•2 o
Design flux, m /m -day (gfd)
Typical conversion
(feed basis), %
Stage number
1
981
(10,560)
3
327
(3,520)
100
13,516
(2,480)
2.38
(35)
74.6
(100)
80
58.9
(79)
2.87
(70)
67
2
477.5
(5,140)
2
238.7
(2,570)
73
9,946
(1,825)
2.38
(35)
55.9
(75)
74
47.0
(63)
2.05
(50)
90
3
117.8
(1,268)
2
58.9
(634)
18
2,453
(450)
2.38
(35)
19.6
(25)
60
14.2
(19)
1.64
(40)
98
135
-------
than required for caustic extract, an overall system conversion of 95% is
acceptable. The lower conversion in the third stage results in a higher
projected average flux,J.64 m3/m2-day (40 gfd), and a reduced membrane
area requirement, 118 nf (1,268 ft2). Except for the stage 3 differences the
Case 3 and Case 1 system designs are alike.
DETAILS OF CASE 4 DESIGN
Summary
-- 7,580 m3/day (2 x 106 gpd) pine or hardwood decker effluent;
-- tubular membrane assemblies;
— hydrasieve and sock filter prefiltration;
-- three-stage UF system (see Table 18).
Description
The Case 4 system for processing decker effluents parallels the Case 2
design for caustic extraction filtrate. The difference in these design
cases is the reduction in overall system conversion to 95% for the decker
effluent processing. The effect of this conversion change is observed in
the third stage of the ultrafiltration system where membrane area is
reduced to 255 m2 (2,746 ft2).
DETAILS OF CASE 5 DESIGN (IDEALIZED SPIRAL-WOUND MODULE SYSTEM)
Summary
The Case 5 design bases are as follows:
— 3,790 m3/day (1 x 106 gpd) pine caustic extraction filtrate;
— spiral-wound membrane modules with Vexar spacers. Each module is
0.1 m (4 in) in diameter x 0.91 m (36 in) long and contains
approximately 2.97 m2 (32 ft2) of membrane surface area. Three
modules would be connected in series and housed in a single shell.
— prefiltration consists of a hydrasieve for fiber removal, a back
flushable 5-1Oy sock filter and disposable string-wound cartridge
fi1ters.
~ the ultrafiltration system consists of 3 stages. These stages are
detailed in Table 19.
Description
The design change from tubular assemblies to spiral-wound modules has
minimal effect on overall system operation. Three stages in series are still
employed for the ultrafiltration system with interstage and system con-
version being similar to those for tubes. Feed recirculation in each stage
136
-------
TABLE 18. ULTRAFILTRATION SECTION DESIGN -- CASE 4
Item
Total membrane
area, m2 (ft2)
No. of subsystems
Membrane area per
subsystem, m2 (ft2)
No. of parallel
membrane passes
per subsystem
Circulation flowrate
per subsystem, m3/day, (gpm)
Pressure drop per
parallel pass, atm (psi)
Motor power, kW (hp)
Efficiency, %
Brake power
per subsystem, kW (hp)
3 2
Design flux, m /m -day (gfd)
Typical conversion
(feed basis )^_%
1
1,962
(21,120)
6
327
(3,520)
100
13,516
(2,480)
2.38
(35)
74.6
(100)
80
58.9
(79)
2.87
(70)
67
Stage number
2
654
(7,040)
2
327
(3,520)
100
13,516
(2,480)
2.38
(35)
74.6
(100)
80
58.9
(79)
2.05
(50)
90
3
255.1
(2,746)
2
127.6
(1,373)
39
5,314
(975)
2.38
(35)
37.3
(50)
65
28.3
(38)
1.64
(40)
95
137
-------
TABLE 19. ULTRAFILTRATION SECTION DESIGN — CASE 5
Stage number
Item
Total membrane
area, nr (ft^)
No. of subsystems
Membrane area per
subsystem, m^ (ffc2)
No. of parallel
membrane passes
per subsystem
Circulation flowrate
per subsystem, m^/day (gpm)
Pressure drop per
parallel pass, atm (psi)
Motor power, kW (hp)
Efficiency, %
Brake power
per subsystem, kW (hp)
3 2
Design flux, m /m -day (gfd)
Typical conversion
(feed basis), %
1
989.9
(10,656)
3
329.9
(3,552)
37
2,017
(370)
2.04
(30)
14.9
(20)
60
9.7
(13)
2.87
(70)
67
2
481.6
(5,184)
2
240.8
(2,592)
27
1,472
(270)
2.04
(30)
11.2
(15)
60
7.5
(10)
2.05
(50)
90
3
338.9
(3,648)
2
169.4
(1,824)
19
1,036
(190)
2.04
(30)
7.5
(10)
60
4.8
(6.5)
1.03
(25)
98
138
-------
is 54.5 m3/day (10 gpm). Permeate and concentrate quality is unchanged as
are the reuse applications for these streams. On the pretreatment end of the
system an additional filter unit is provided to further protect the spiral -
wound modules. This unit consists of string-wound cartridge filters. These
5y filters are disposable.
Membrane area requirements for the spiral systems (Cases 5 through 8)
remain unchanged from those developed for tubular systems. That is, the
same flux levels have been assumed in each comparative case. In this manner,
"ideal" spiral-wound system case analyses will be developed. "Ideal" spiral-
wound module performance may not be achievable given the current state-of-
the-art of membrane technology. However, if future spiral systems are to be
of interest to the pulp and paper industry long-term reliable operation must
be demonstrated. This will require improved module designs, potentially
having flux levels equivalent to tubes. Again, the spiral-wound module
system designs and economic projections presented in this report reflect
"ideal" cases and are not achievable given today's technology.
DETAILS OF CASE 6 DESIGN (IDEALIZED SPIRAL-WOUND MODULE SYSTEM)
Summary
-- 7,580 m3/day (2 x 10^ gpd) pine caustic extraction filtrate;
-- spiral-wound modules;
-- hydrasieve, sock filter and cartridge filter pretreatment;
-- three-stage UF system (see Table 20).
Description
Case 6 is analogous to Case 2 in that a 7,590 m3/day (2 MM gpd) caustic
extract stream is being processed. The first stage of the spiral-wound
system consists of 3 subsystems, each with 660 m? (7,104 ft2) of membrane
area. The second and third stages are nearly identical. These stages each
have two subsystems with ^325 m2 (^3,500 ft?) of membrane area. Other than
the 10% safety margin in membrane area, no excess capacity has been designed
into Stage 2. Therefore, during cleaning cycles (which are potentially of
a longer duration than for tubular systems) the overall system conversion
may fall below 98%. If the bi-weekly cleaning cycles for each of the Case 6
spiral-wound subsystems should exceed 5 hours, then the ultrafiltration
system would be in a cleaning mode more than 10% of the time. Under these
circumstances a design change to incorporate more membrane area, and/or
more subsystems may be required.
DETAILS OF CASE 7 DESIGN (IDEALIZED SPIRAL-WOUND MODULE SYSTEM)
Summary
— 3,790 m3/day (1 x 106 gpd) pine or hardwood decker effluent;
-- spiral-wound membrane modules;
139
-------
TABLE 20. ULTRAFILTRATION SECTION DESIGN — CASE 6
Stage number
Item
Total membrane
area, m2 (ft^)
No. of subsystems
Membrane area per
subsystem, m2 (ft2)
No. of parallel
membrane passes
per subsystem
Circulation flowrate
per subsystem, m^/day (gpm)
Pressure drop per
parallel pass, atm (psi)
Motor power, kW (hp)
Efficiency, %
Brake power
per subsystem, kW (hp)
3 2
Design flux, m /m -day (gfd)
Typical conversion
(feed basis), %
1
1,980
(21,312)
3
659.9
(7,104)
74
4,033
(740)
2.04
(30)
26.1
(35)
65
17.2
(23)
2.87
(70)
67
2
642.1
(6,912)
2
321.1
(3,456)
36
1,962
(360)
2.04
(30)
14.9
(20)
60
8.9
(12)
2.05
(50)
90
3
659.9
(7,104)
2
329.9
(3,552)
37
2,017
(370)
2.04
(30)
14.9
(20)
60
9-7
(13)
1.03
(25)
98
140
-------
-- hydrasieve, sock filter and cartridge filter pretreatment;
-- three-stage UF system (see Table 21).
Description
The spiral-wound system for treating 3,790 m^/day (1 MM gpd) of decker
effluent is designed to have an overall conversion of 95%. It is identical
to the Case 5 design for caustic extraction filtrate except that the third
stage contains less membrane area (125 m2) [1,344 ft2].
DETAILS OF CASE 8 DESIGN (IDEALIZED SPIRAL-WOUND MODULE SYSTEM)
Summary
~ 7,590 m3/day (2 x 106) pine or hardwood decker effluent;
— spiral-wound membrane modules;
-- hydrasieve sock filter and cartridge filter pretreatment;
-- three-stage UF system (see Table 22).
Description
The final design case considers treatment of 7,590 m3/day (2 MM gpd)
of decker effluent by spiral-wound modules. Overall conversion is 95%.
System pretreatment and operations are the same as described for Case 5.
141
-------
TABLE 21. ULTRAFILTRATION SECTION DESIGN -- CASE 7
Stage number
Item
Total membrane
area, nr (ft^)
No. of subsystems
Membrane area per
subsystem, m^ (ft2)
No. of parallel
membrane passes
per subsystem
Circulation flowrate
per subsystem, m3/day(gpm)
Pressure drop per
parallel pass, atm (psi)
Motor power, kW (hp)
Efficiency, %
Brake power
per subsystem, kW (hp)
3 2
Design flux, m /m -day (gfd)
Typfcal conversion
(feed basis), %
1
989.9
(10,656)
3
329.9
(3,552)
37
2,017
(370)
2.04
(30)
14.9
(20)
60
9.7
(13)
2.87
(70)
67
2
481.6
(5,184)
2
240.8
(2,592)
27
1,472
(270)
2.04
(30)
11.2
(15)
60
7.5
(10)
2.05
(50)
90
3
124.9
(1,344)
2
62.4
(672)
7
382
(70)
2.04
(30)
5.6
(7.5)
40
3.0
(4)
1.64
(40)
95
142
-------
TABLE 22. ULTRAFILTRATION SECTION DESIGN — CASE 8
Stage Number
Item
Total membrane
area, m2 (ft2)
No. of subsystems
Membrane area per
subsystem, m2' (ft2)
No. of parallel
membrane passes per
subsystem
Circulation flowrate per
subsystem, m3/day (gpm)
Pressure drop per
parallel pass, atm (psi)
Motor power, kW (hp)
Efficiency, %
Brake power per
subsystem, kW (hp)
3 2
Design flux, m /m -day (gfd)
Typical conversion
(feed basis), %
1
1,980
(21,312)
3
659.9
(7,104)
74
4,033
(740)
2.04
(30)
26.1
(35)
65
17.2
(23)
2.87
(70)
67
2
642.1
(6,912)
2
321.1
(3,456)
36
1,962
(360)
2.04
(30)
14.9
(20)
60
8.9
(12)
2.05
(50)
90
3
267.6
(2,880)
2
133.8
(1,440)
15
818
(150)
2.04
(30)
7.5
(10)
50
3.7
(5)
1.64
(40)
95
143
-------
SECTION 8
PROJECTED ECONOMICS FOR FULL-SCALE SYSTEMS
INTRODUCTION
Capital and operating cost estimates have been prepared for each of the
eight design cases. Because of the significant cost differential between
tubular and spiral-wound ultrafiltration systems, the economic analyses
which follow will center on tubular systems (design cases 1 through 4) be-
fore discussion of spiral-wound systems (design cases 5 through 8) is begun.
Included with the tubular system cost estimates are projections of future
cost savings due to advances in membrane technology. Note, cost estimates
for spiral-wound systems are based on ideal system performance and are not
attainable given today's technology.
COSTS FOR DESIGN CASES 1 THROUGH 4
Bases for Capital Cost Projections
The capital costs for the full-scale treatment systems incorporating
tubular ultrafiltration systems were derived in the following manner.
First, all major auxiliary equipment and the ultrafiltration system were
sized. The auxiliary equipment (pretreatment system, permeate collection
tank, etc.) was then costed through vendor quotes and catalogs. The
ultrafiltration system costs were divided into hardware and membrane costs.
Hardware costs for the 3,790 m3/day (1 MM gpd) caustic extraction filtrate
system (case 1) were broken out as carefully as possible for this type of
engineering estimate. Hardware costs for the remaining 3 tubular systems
were derived as fractions of the case 1 hardware costs. In all case 1
through 4 estimates, membrane costs Were held constant at $293.5/m2 ($27.27/
ft2). This corresponds to $120 per 6.1 m (20 ft) tubular assembly.
The on-site engineering design and installation expenses were calcu-
lated as percentages of the equipment costs. Different multipliers were
used for the auxiliary equipment engineering and installation than for the
ultrafiltration system. This is because the ultrafiltration systems are
supplied as skid-mounted subsystems and all engineering internal to the
ultrafiltration system (inter-stage piping and electrical, etc.) will have
been completed by the vendor. The percentage increases for installation and
design used were:
144
-------
Auxiliary Ultrafiltration
equipment system
Installation 40% of equipment cost 15% of equipment cost
Detailed engineering 12% of installed cost 5% of installed cost
design
Costs for a building to house each treatment system were calculated by
estimating the area requirements (based on subsystem conceptual desiqn) and
applying a multiplier of $161.5/m2 ($15/ft2).
Installed equipment cost, engineering and building costs were totaled
and the following three categories and multipliers were used in sequence
to obtain the total installed cost.
Cost
element Multiplier
Administration and Supervision 2%
Contingency 10%
Inflation 10%
Design Case 1 Capital Cost
The equipment costs for the 3,790 m3/day (1 MM gpd) caustic extraction
filtrate treatment system are detailed below. The total installed capital
cost is calculated in Table 23.
Pretreatment--
Hydrasieve - 3,790 m3/day (700 gpm) fiberglass
frame, 0.76 mm (0.030 in) screen, two 1.83 m
(72 in) screens placed back-to-back $15,000
Fiber Pump - 16.4 m3/day (3 gpm), 0.37 kw
(0.5 hp) $ 700
Feed Pump - 3,790 m3/day (700 gpm), 4.1 atm
(60 psig) head, 29.8 kw (40 hp) motor, carbon
steel $ 2>500
Feed Pump (Spare) - same as above $ 2,500
Sock Filters (2) - 3,790 m3/day (700 gpm),
automatic controls for extended backwash,
0.15 m (6 in) stainless steel drain header,
polypropylene sock, 5 to lOy VH ,uuu
145
-------
TABLE 23. CASE 1 DESIGN CAPITAL COST SUMMARY
Pretreatment system
Permeate collection/
distribution subsystem
Concentrate collection/
distribution subsystem
Installation @ 40% of
total auxiliary equipment
cost
Auxiliary equipment
installed cost
Ultrafiltration system
(includes $524,160
for membranes)
Ultrafiltration system
installation @ 15% of
UF system cost
Ultrafiltration system
installed cost
Total equipment cost
Detailed engineering design
@ 12% of auxiliary
equipment installed cost
and 5% of UF system
installed cost
Building (372 m2 @ $161/m2)
[4000 ft2 @ $15/ft2]
Subtotal A
Administration and super-
vision (2% of subtotal A)
Subtotal B
Contingency (10% of
subtotal B)
Subtotal C
Inflation (10% of
subtotal C)
Total installed cost
84,200
53,800
10,100
148,100
59,240
1,349,000
202,350
207,340
1,551,350
1,758,690
102,448
60,000
1,921,138
38,423
1,959,561
195.956
2,155,517
215,551
$ 2.371.068
146
-------
Piping and Valves (rough estimate) - includes
305 m (1,000 ft) of 0.25 m (10 in) diameter
carbon steel pipe @ $32.8/m ($10/ft) $20,000
SUBTOTAL FOR EQUIPMENT $81,700
Freight-in transportation costs @ 3% of
equipment cost $ 2,500
TOTAL PRETREATMENT SYSTEM $84,200
(Excluding installation, engineering design,
equipment procurement and startup)
Ultrafiltration System—
A total of 1,785 m2 (19,200 ft2) of membrane will be used. The costs
are broken out by stages and major categories below. Costs shown for the
UF system include all factors except on-site engineering, on-site instal-
lation, on-site supervision and start-up.
Stage 1 Pumping Packages (3 required) -
include 13,516 m-Vday (2,480 gpm) pumps,
74.6 kw (100 hp) motors, all valves, piping
and mounting materials, engineering and
assembly @ $38,900 $116,700
Stage 1 Membrane Racks (3 required) -
include inlet and outlet headers, per-
meate collection headers, membrane
supports, all materials, assembly
labor @ $42,900 $128,700
Stage 1 Automatic Cleaning Packages
(3 required) - include piping, solenoid
valves, engineering and assembly @
$9,240 each $ 27,720
Stage 2 Pumping Packages (2 required) -
include 9,946 m3/day (1,825 gpm) pumps,
55.9 kw (75 hp) motors, all valves,
piping, mounting materials, engineering
and assembly @ $38,900 $ 77,800
Stage 2 Membrane Racks (2 required) -
similar to stage 1 membrane racks
@ $36,300 $ 72,600
Stage 2 Automated Cleaning Packages
(2 required) - similar to stage 1 units
@ $9,240 $ 18,480
147
-------
Stage 3 Pumping Package (2 required) -
include 6,813 m-Vday (1,250 gpm) pumps, 37.3
kw (50 hp) motors, all valves, piping,
mounting materials, heat exchangers in
circulation loop, engineering and assembly
@ $30,900 $ 61,800
Stage 3 Membrane Racks (2 required) -
Similar to stage 1 membrane racks @ $28,200 $ 56,400
Stage 3 Automatic Cleaning Packages
(2 required) - similar to stage 1 units
@ $7,920 $ 15,840
Cleaning Station - includes 7.58 m3 (2,000
gal) clean tank, 2.27 m3 (600 gal) rinse
tank, 3,270 m3/day (600 gpm) pump with 18.6
kw (25 hp) motor, tank and pump mounting,
engineering and assembly labor $ 29,700
Interstage Piping - for feed, permeate and
cleaning. Includes pressure control of
feed and flow ratio control of feed to con-
centrate to automatically control system
conversion. Materials, engineering and
assembly $128,000
Electrical - includes central control panel
with semi-graphic display, subsystem wiring
to common terminal box, materials, engineer-
ing and assembly $ 91,100
yitrafiltration Membranes - tubular, 25.4 mm
(1 in) diameter x 3.05 m (10 ft) long WRP
membrane assemblies. Including inter-
connectors, fiberglass backing, plastic shells
and U-bends. 8,736 required @ $60 each $524,160
TOTAL UF SYSTEM COST $1,349,000
(including all factors except on-site
engineering, on-site installation, on-
site supervision and start-up)
Permeate Collection/Distribution System—
The UF permeate will be collected in 189.5 m3 (50,000 gal) tank (^ 1
hour residence time). From 5% to 50% of the permeate will be returned to
the bleachery for reuse. The remaining permeate will be used for sock filter
backflush, subsystem cleaning blended with the mill water supply or dis-
charged to sewer.
148
-------
Permeate Collection Tank - 189.5 m3 (50,000
gal), construction on-site $ 30,000
Permeate Distribution Pump - 3,790 m3/day
(700 gpm), 18.6 kw (25 hp) motor, 2.72 atm
(40 psig) head, carbon steel $ 2,200
Piping and Valves (rough estimate) includes
305 m (1,000 ft) of 0.25 m (10 in) diameter
carbon steel pipe @ $37.8/m ($10/ft) $ 20,000
SUBTOTAL FOR EQUIPMENT $ 52,200
Freight-in transportation costs @ 3%
of equipment costs $ 1,600
TOTAL PERMEATE COLLECTION/DISTRIBUTION SYSTEM $ 53,800
(Excluding installation, engineering design,
equipment procurement and startup)
Concentrate Collection/Distribution System--
The UF system concentrate will be returned to the weak black liquor,
the lime kilm or admixed with the lime sludge at an average flowrate of
76.3 m3/day (14 gpm).
Concentrate Holding Tank - 1.9 m3 (500 gal) $ 1,500
Concentrate Transfer Pump - 76.3 m3/day
14 gpm), 2.7 atm (40 psig) head, 3.72 kw
5 hp) motor, carbon steel $ 800
Piping and Valves (rough estimate) $ 7,500
SUBTOTAL FOR EQUIPMENT $ 9,800
Freight-in transportation costs @ 3%
of equipment costs $ 3°0
TOTAL CONCENTRATE COLLECTION/DISTRIBUTION SYSTEM $ 10,100
Summary--
The total installed cost for the case 1 system is estimated to be
$2,371,068 (see Table 23). The ultrafiltration system accounts for 57%
of this figure.
149
-------
The ultrafiltration system cost consists of approximately $291/m2
($27.3/ft2) for membranes and $463/m2 ($43/ft2) for hardware. Overall, this
amounts to $753/m2 ($70/ft2) for the 178.6 m2 (19,220 ft2) ultrafiltration
system.
Design Case 2 Capital Cost
Pretreatment—•
The capacity of the Case 1 pretreatment system must be doubled to treat
7580 m3/day (2 MM gpd) of caustic extraction filtrate. The costs for the
increased capacity components are shown below. The general characteristics
of the components are the same as detailed in the Case 1 cost discussion.
Hydrasieve - 7,630 m3/day (1,400 gpm) $ 25,000
Fiber pump - 7,630 m3/day (6 gpm) $ 1,200
Feed pump - 7,630 m3/day (1,400 gpm) $ 4,000
Feed pump (spare) - 7,630 m3/day (1,400 gpm) $ 4,000
Sock filters (2) - 7,630 m3/day (1,400 gpm) $ 82,000
Piping and valves - $ 35,000
SUBTOTAL FOR EQUIPMENT $151,200
Freight-in transportation costs @ 3% of
equipment cost $ 4,500
TOTAL PRETREATMENT SYSTEM $155,700
(Excluding installation, engineering design,
equipment procurement and startup)
The Case 2 pretreatment system cost is 1.85 x the Case 1 cost estimate.
Ultrafiltration System--
The hardware cost for the ultrafiltration system is based on a system
consisting of 10 identical subsystems. Each subsystem would be the same as
described for stage 1 of the case 1 system.
UF system hardware - $1,317,600
150
-------
UF membranes - 3,270 m? (35,200 ft2) @
$293. 5/m* ($27.27/ft2) 960j000
TOTAL UF SYSTEM COST $2,277,600
(including all factors except on-site
installation, on-site supervision and
startup)
This ultra-filtration system provides a doubling of process capacity for
1 .69X the case 1 cost.
Permeate Collection/Distribution System--
For the 7,580 m3/day (2 MM gpd) systems (cases 2 and 4) it was decided
to retain the same size permeate holding tank as for the 3,790m3/day (1 MM
gpd) systems. This volume (189.5 m3) [50,000 gal] would still be sufficient
for flushing, cleaning and recycling operations. Holding tank residence
time would be reduced to 35 minutes.
Permeate collection tank- 189.5 m3
(50,000 gal) $ 30,000
Permeate distribution pump - 7,630 m3/day
(1 ,400 gpm) $ 4,000
Piping and valves $ 35,000
SUBTOTAL FOR EQUIPMENT $ 69,000
Freight-in transportation costs @ 3% of
equipment cost $ 2,100
TOTAL PERMEATE COLLECTION/DISTRIBUTION SYSTEM $ 71.000
(excluding installation, engineering design,
equipment procurement and startup)
Concentrate Collection/Distribution System—
Concentrate holding tank - 1.9 m3 (500 gal)
(same size as for Case 1) * '>500
Concentrate transfer pump - 153 m3/day
(28 gpm) * ]'oou
Piping and Valves $ 11>50°
SUBTOTAL FOR EQUIPMENT $ 14>000
151
-------
Freight-in transportation costs @ 3% of
equipment costs $ 400
TOTAL CONCENTRATE COLLECTION/DISTRIBUTION SYSTEM $ 14,400
(excluding installation, engineering design,
equipment procurement and startup)
Summary—
The case 2 capital cost is summarized in Table 24. Total installed cost
is $3,972,151. This cost is 1.675X the case 1 installed cost, the ultra-
filtration system accounts for 57%. The ultrafiltration system is $696/m^
($64.70/ft2) ($402.6/m2 for hardware and $293.5/m2 for membranes) [$37.4/ft2
for hardware and $27.3/ft2 for membranes].
Capital Cost for Design Case 3
The case 3 system is designed to treat 3,790 m3/day (1 MM gpd) pine or
hardwood decker effluent. The only major change from the case 1 design is
that processing is only necessary to 95% conversion in the ultrafiltration
system. Therefore the third stage of the ultrafiltration system contains
less membrane area than the corresponding stage under case 1.
Pretreatment System—
The pretreatment system is identical to that presented for case 1.
TOTAL PRETREATMENT SYSTEM COST $ 84,200
Ultrafiltration System--
The ultrafiltration system hardware cost is unchanged from case 1 for
stage 1, stage 2 and interstage connections. The stage 3 hardware cost will
be somewhat reduced. Overall hardware cost is calculated as the same hard-
ware cost factor ($463/m2) [$43/ft2] as for case 1.
UF system hardware - $729,624
UF membranes - 1,576 m2 (16,968 ft2) (3
$295.5/m* ($27.27/ft2) $462,760
TOTAL UF SYSTEM COST $1,192,384
152
-------
TABLE 24. CASE 2 DESIGN CAPITAL COST SUMMARY
Pretreatment system 155,700
Permeate collection/
distribution subsystem 71,100
Concentrate collection/
distribution subsystem 14,400
241,200
Installation @ 40% of
total auxiliary equipment
cost 96,480
Auxiliary equipment
installed cost 337,680
Ultrafiltration system
(includes $960,000
for membranes) 2,277,600
Ultrafiltration system
installation @ 15% of
UF system cost 341,640
Ultrafiltration system
installed cost 2.619.240
Total equipment cost 2,956,920
Detailed engineering design
@ 12% of auxiliary equipment
installed cost and 5% of UF
system installed cost 171,480
Building ($557 m2 @ $161/m2)
[6,000 ft2 @ $15/ft2] 90.000
Subtotal A 3,218,400
Admtnlstration and super-
vision (2% of subtotal A) 64.370
Subtotal B 3,282,770
Contingency (10% of
subtotal B) 328»277
Subtotal C 3,611,047
Inflation (10% of ,,., irw!
subtotal C) 36_uiOi
Total installed cost $ 3,972.151
153
-------
Permeate and Concentrate Collection/Distribution Systems--
No change from the case 1 study occurs for either the permeate or
concentrate collection/distribution system.
TOTAL PERMEATE COLLECTION/DISTRIBUTION SYSTEM $ 53,800
TOTAL CONCENTRATE COLLECTION/DISTRIBUTION SYSTEM $ 10,100
Summary--
Table 25 details the case 3 capital cost projection. Total installed
cost is estimated to be $2,137,736. Membrane system cost is 56% of total
cost. As with case 1, ultrafiltration system cost is $753.5/m2 ($70/ft2).
Capital Cost for Design Case 4
Design case 4 is for 7,580 m3/day (2 MM gpd) decker effluent treatment.
The costs parallel case 2 costs with a reduction in ultrafiltration system
costs consistent with the reduced membrane area.
Pretreatment System—
The pretreatment system is identical to the case 2 pretreatment system.
TOTAL PRETREATMENT SYSTEM COST $155,700
Ultrafiltration System—
UF system hardware - 2,871 m2
(30,906 ft*) » $402.6/m2 ($37.4/ft2) $1,155,900
UF membranes - 2,871 m2 (30,906 ft2)
@ $293.5/m2 ($27.27/ft2) $ 842,810
TOTAL UF SYSTEM COST $1,998,710
Permeate and Concentrate Collection/Distribution Systems--
Permeate and concentrate collection/distribution system costs are
unchanged from those presented for case 2.
TOTAL PERMEATE COLLECTION/DISTRIBUTION SYSTEM $ 71,100
TOTAL CONCENTRATE COLLECTION/DISTRIBUTION SYSTEM $ 14,400
154
-------
TABLE 25. CASE 3 DESIGN CAPITAL COST SUMMARY
Pretreatment system • 84,200
Permeate collection/
distribution subsystem 53,800
Concentrate collection/
distribution subsysten 10,100
148,100
Installation @ 40% of
total auxiliary equipment
cost 59.240
Auxiliary equipment
installed cost 207,340
Ultrafiltration system
(includes $462,760
for membranes) 1,192,400
Ultrafiltration system
installation @ 15% of
UF system cost 178.860
Ultrafiltration system
installed cost 1.371.260
Total equipment cost 1,578,600
Detailed engineering design
@ 12% of auxiliary equipment
installed cost and 5% of UF
system installed cost 93,444
Building (372 m2 @ $161/m2)
[4000 ft2 @ $15/ft2] 60.000
Subtotal A 1,732,044
Administration and super-
vision (2% of subtotal A) 34.681
Subtotal B 1,766,725
Contingency (10% of
Subtotal B) 176'672
Subtotal C 1,943,397
Inflation (10% of
subtotal C)
Total
installed cost $ 2,137,736
155
-------
Summary--
The case 4 design capital cost summary is shown in Table 26. Total
installed cost is $3,556,524. Again, membrane system cost is 56% of total
cost. As with case 2, the ultrafiltration system is $696.4/ir ($64.70/ft2)
of membrane area.
Bases for Operating Cost Projections
Operating costs fall into 3 main categories: materials, conversion
expense less depreciation, and depreciation. The only materials required
for system operation are cleaning chemicals for the membrane system. These
chemicals are caustic and EDTA. They were costed at bulk quantity prices.
Conversion expense (less depreciation) is made up of four major ele-
ments. The first element, labor, was assumed at 6 hours per day for all
systems. Current labor expense (including fringe benefits) is $10/hour.
Electrical power was calculated based on feed pump, ultrafiltration system
circulation pumps and permeate distribution pump brake horsepower. The
current Canton Mill power cost of $0.0225/kwh was used. Repair and
maintenance costs consist of materials and labor. Maintenance material was
calculated as 1.5% of the hard goods cost (total equipment cost less
membrane replacement cost). The maintenance labor cost was assumed to be
equivalent to the material cost. Thus, maintenance labor hours were back-
calculated using a rate of $15/hour. The final element for this cost
category is insurance and taxes. This operating cost was assumed to be
one-half of the maintenance material cost.
Depreciation expense has two elements: membrane replacement cost and
equipment amortization. The tubular membranes were depreciated over a
3 year life at their replacement cost, not their original cost. This re-
placement cost is $172.2/m2 ($16/ft2) when ordered in minimum quantities of
2,000 3.05 m (10 ft) lengths. This reduction in tubular assembly original
cost ($27.3/ft2) results from return (and recovery) of the permeate
collection shell and the fiberglass support tube. Also, replacement U-bends
are not required. The equipment (other facilities) was depreciated over
15 years on a straight line basis.
Throughout the operating cost estimates the plant was assumed to be
operating 24 hours per day, 365 days per year.
Operating Costs for Design Case 1
The case 1 operating costs are detailed below with sample calculations.
Materials--
The cleaning materials are 0.5% caustic and 0.25% EDTA. A cleaning
solution volume of 7.58 m3 (2,000 gal) per subsystem is required. This
156
-------
TABLE 26. CASE 4 DESIGN CAPITAL COST SUMMARY
Pretreatment system 155,700
Permeate collection/
distribution subsystem 71,100
Concentrate collection/
distribution subsystem 14,400
241,200
Installation @ 40% of
total auxiliary equipment
cost 96.480
Auxiliary equipment
installed cost 337,680
Ultrafiltration system
(includes $842,810
for membranes) 1,998,710
Ultrafiltration system
installation @ 15% of
UF system cost 299.807
Ultrafiltration system
installed cost 2.298.517
Total equipment cost 2,636,197
Detailed engineering design
@ 12% of auxiliary equipment
installed cost and 5% of UF
system installed cost 155,447
Building (557 m2 @ $161/m2)
[6,000 ft2 & $15/ft2] 90.000
Subtotal A 2,881,644
Administration and super-
vision (2% of subtotal A) 57,633
Subtotal B 2,939,277
Contingency (10% of subtotal B) 293.927
Subtotal C 3,233,204
Inflation (10% of ,9, ,9n
subtotal C) 323,320
Total installed cost $ 3.556,524
157
-------
volume is held constant even though subsystem size varies because the
difficulty of cleaning may increase in the higher conversion stages. One
cleaning per subsystem biweekly is projected.
For caustic,
2,000 gal
cleaning
3.77 kg
gal
1 cleaning^
2 days
0.5% caustic
= 18.8 kg
caustic/day
If a 50% caustic soda is used 37.7 kg/day (83 Ib/day) are required. At a
cost of $0.15/kg ($0.07/lb) of caustic soda a daily cost of $5.81 is incurred.
For EDTA,
2,000 gal
cleaning
3.77 kg
gal
1 cleaning
2 days
0.25% caustic _
9.44 kg EDTA/ day
At a cost of $1.95/kg ($0.885/lb) the daily charge is $18.41.
Total material cost is $24.227day.
Conversion Expense Less Depreciation—
Operating Laboi—
6 hr
day
Electrical Power--
% = $60/day
500 hp
.7457 kw
hp
24 hr
day
$0.0225
kwh
= $203.747 day
Maintenance Material--
1.5%
($1.758.690-307.200)
year
year
= $59.64/day
Maintenance Labor—
4 hr
day
Insurance and Taxes—
% = $60/day
0.5
day
= $29.82/day
Total conversion expense less depreciation is $413.207day.
158
-------
Depreciation--
Membranes--
1 ,784 m2
3 years
$172.2
m2
year
365 days
= $280.55/day
Other Facilities--
$2,063,868
15 years
.year
365 days
= $376.96/day
Total depreciation is $657.51/day.
Summary—
The case 1 operating costs are summarized in Table 27 on daily, yearly,
volume and percentage bases. Overall operating cost is $1,094.93/day
($0.29/m3) [$1.10/1000 gal]. For a 727 metric ton/day (800 ton/day) pulp
mill this amounts to a cost of $1.51/metric ton ($1.37/ton) of pine pulp.
The major cost factors are power - 18.2%, membrane replacement - 25.5%,
and facility depreciation - 34.5%. Total depreciation accounts for 60%
of the operating cost.
Operating Costs for Design Case 2
The same volume (7.58 rr)3) [2,000 gal] cleaning solution is used with 10
subsystems being cleaned per 2-week period rather than 7 as in case 1.
Material costs thus increase by 10/7.
All conversion expenses are figured in a like manner to the sample
calculations presented above for case 1. These expenses are summarized in
Table 28.
For this 7,580 m3/day (2 MM gpd) caustic extraction filtrate treatment
system daily operating costs are $1,835.03. This is equivalent to $0.24/
m3 ($0.92/1000 gal) treated or $2.52/metric ton (727 metric ton basis)
[$2.29/ton (800 ton basis)]. Major cost factors are again power, membrane
replacement and facilities depreciation. These factors account for 19.6%,
28.3% and 33.7% of the operating costs, respectively.
Operating Costs for Design Case 3
Case 3 operating costs are detailed in Table 29. Daily costs are
$998.01 for washing of 1,318 metric tons/day (1,450 tons/day) of mixed
pulp production. This translates to $0.26/m3 ($1.00/1000 gal) or $0.76/
metric ton ($0.69/ton) of pulp.
159
-------
TABLE 27. CASE 1 DESIGN OPERATING COST DATA
Cost Element
Material (for cleaning)
Caustic (50%)
EOTA
Total materials
Conversion expense
Labor (Including benefits)
Electrical power
Repair and Maintenance
Material 1.5X
cost
Labor
Insurance and Taxes 0.5
Total, excluding
Depreciation
Membranes
Other facilities
Total conversion expense
Total incremental cost
Quantity
37.6 kg/day
4.9 kg/day
6 hr/day
9,055 kwn/day
of total equipment less membrane
yearly - (1.5X x 1,451, 170}/yr
4 hr/day
x maintenance materials
depreciation
$ 307,200
$ 2,063,868
Unit Cost
$ 0.154/kg
1.95/kg
$ 10/hr
$0.0225/kwh
replacement
$ 15/hr
3-yr Hfe
15-yr Hfe
' $/day
5.81
18.41
24. ZZ
60.00
203.74
59.64
60.00
29.82
413.20
280.55
376.96
$ 1,070.^1
$ 1,094.93
$/year
2,121
6.720
8,841
21,900
74,365
21,768
21,900
10,884
150,817
102,400
137,591
$ 390,808
$ 399,649
$/m3
0.006
0.016
0.054
0.158
0.016
0.079
0.109
0.074
0.997
0.283
0.289
$/M-gal
0.03
0.06
0.20
0.06
0.06
0.03
0.41
0.28
0.38
1.07
1.10
% of Total
2.7
5.5
18.2
5.5
5.5
2.7
37.3
25.5
34.5
97.3
100.0
-------
TABLE 28. CASE 2 DESIGN OPERATING COST DATA
Cost Element
Material (for cleaning)
Caustic (50%)
EDTA
Total materials
Conversion expense
Labor (including benefits)
Electrical power 15,
Repair and Maintenance
Material 1.5% of total equipment
replacement cost yearly
Labor
Insurance and Taxes 0.5 x maintenance materi
Total, excluding depreciation
Depreciation
Membranes $
Other facilities $ 3,
Total conversion expense
Total incremental cost
Quantity
53.8 kq/day
13.5 kg/day
6 hr/day
929 kwh/day
less membrane
= (1.53! x 2,393
6.5 hr/day
al
563,200
408,951
Unit Cost
$ 0.154/kg
1.95/kg
$ 10/hr
$ 0.0025/kwh
,720)/yr
$ 15/hr
3-yr life
15-yr life
$/day
8.31
26.28
34.59
60.00
358.40
98.37
97.50
49.19
663.46
514.34
622.64
$ 1,800.44
$ 1,835.03
$/year
3,033
9,592
12,625
21,900
130,816
35,905
35,588
17,954
242,163
187,734
227,264
657,161
669,786
$/m3
0.005
0.008
0.047
0.013
0.013
0.006
0.088
0.068
0.082
0.238
0.242
$/M-ga1
0.02
0.03
0.18
0.05
0.05
0.02
0.33
0.26
0.31
0.90
0.92
% of Total
2.2
3.3
19.6
5.4
5.4
2.2
35.9
28.3
33.7
97.8
100.0
-------
TABLE 29. CASE 3 DESIGN OPERATING COST DATA
en
PO
Cost Element
Material (for cleaning)
Caustic (50%)
EDTA
Total materials
Conversion expense
Labor (including benefits)
Electrical power 8,
Repair and Maintenance
Material 1.5* of total equlprament
replacement cost yearly
Labor
Insurance and Taxes 0.5 x maintenance materl
Total, excluding depreciation
Depreciation
Membranes $
Other facilities $ 1,
Total conversion expense
Total Incremental cost
Quantity
Unit Cost
37.6kg/day $ 0.154/kg
4.9 kg/day 1.95/kg
6 hr/day $ 10/hr
124 kwh/day $ 0.0025/kwh
less membrane
« (1.5* x l,307,112)/yr
4 hr/day $ 15/hr
al
271,488 3-yr life
866,248 15-yr life
$/day
5.81
18.41
24. Z2
60.00
182.79
53.72
60.00
28.48
384.99
247.93
340.87
$ 973.79
$ 998.01
$/year
2,121
6,720
8,811
21,900
66,718
19,607
21,900
10,395
140,520
90,496
124,417
355,435
364,274
$/m3
0.006
0.016
0.048
0.014
0.016
0.075
0.102
0.066
0.090
0.258
0.264
$/M-gal
0.03
0.06
0.18
0.05
0.06
0.03
0.38
0.25
0.34
0.97
1.00
% of Total
3.0
6.0
18.0
5.0
6.0
3.0
38.0
25.0
34.0
97.0
100.0
-------
Operating Costs for Design Case 4
Table 30 presents case 4 operating costs. Total daily costs are
$1,663.24 or $0.22/m-5 ($0.83/1000 gal). Operating cost is $1.26/metric
ton ($1.15/ton) of pulp (1,378 metric ton/day basis) [1,450/ton day basis].
Summary of Capital and Operating Costs for Design Cases 1 through 4
A summary of the projected economics for treatment systems incorporating
tubular ultrafiltration systems is given in Table 31. Capital investment
for all cases ranges from $2 to $4 MM. Treatment costs, based on 727 metric
tons of pine pulp produced per day, range from $1.51 to $2.52 for caustic
extraction filtrate. The treatment costs range from $0.76 to $1.26 for
decker effluents (1,318 metric ton/day basis).
In the case 1 through 4 economics the membrane replacement cost is a
minimum of 25% of the operating cost. This is with a projected membrane life
of 3 years. If membrane life were, in fact, only 2 years the operating
costs would increase by $0.19/metric ton for case 1, $0.35/metric ton for
case 2, $0.09/metric ton for case 3 and $0.17/metric ton for case 4.
Future Costs for Tubular Ultrafiltration Systems
Future technological advances are expected to reduce large-scale
tubular Ultrafiltration system capital costs. These advances will take the
form of lower-cost, more-compact modules, possibly with insertable membranes.
The insertable membranes would have a lower replacement cost than current
membranes. Lower cost systems may be available in the next 2 to 5 years.
To assess the impact of reduced future Ultrafiltration system costs and
lower membrane replacement costs on overall treatment system capital and
operating costs the case 1, 3,790 m3/day (1 MM gpd) caustic extraction
filtrate system, has been reanalyzed with the following changes:
Capital costs
~ auxiliary equipment and building costs increased 20%;
— UF system cost calculated at $646/m2 ($60/ft2).
Operating costs
— cleaning chemicals expense increased 10%;
— labor expense increased 20%;
— electrical power costed at $0.025/kwh;
-- membrane replacement costed at $86.1/m2 ($8/ft2).
The revised capital and operating cost summaries are presented in Tables 32
and 33, respectively.
163
-------
TABLE 30. CASE 4 DESIGN OPERATING COST DATA
cr>
Cost Element
Material (for cleaning)
Caustic (5d%)
EDTA
Total materials
Conversion expense
Labor (including benefits)
Electrical power
Repair and Maintenance
Material 1.5%
repla
Labor
Insurance and Taxes 0.5 x
Total, excluding
Depreciation
Membranes
Other facilities
Total conversion expense
Total incremental cost
Quantity
53.8 kg/day
13.5 kg/day
6 hr/day
14,460 kwh/day
of total equipment less membrane
cement cost yearly « (1.5% x 2,141
6.5 hr/day
maintenance material
depreciation
$ 494,496
$ 3,062,028
Unit Cost
$ 0.154/kg
1.95/kg
$ 10/hr
$ 0.0025/kwh
,701)/yr
$ 15/hr
3-yr life
15-yr life
$/day
8.31
26.28
34.59
60.00
325.40
88.02
97.50
46.87
617.79
451.59
559.27
$ 1,628.65
$ 1,663.24
$/year
3,033
9,592
12,625
21,900
118,771
32,126
35,588
17,108
225,493
164,832
204,135
594,457
607,082
$/m3
0.005
0.008
0.043
0.012
0.013
0.006
0.082
0.059
0.074
0.215
0.219
$/M-gal
0.02
0.03
0.16
0.04
0.05
0.02
0.30
0.23
0.28
0.81
0.83
% of Total
2.4
3.6
19.3
4.8
6.0
2.4
36.1
27.7
33.7
97.6
100.0
-------
TABLE 31. SUMMARY OF PROJECTED ECONOMICS FOR DESIGN CASES 1 THROUGH 4
en
Design case number
Item
Stream
Flow, m3/day (MM gal /day)
UF membrane cost, $
UF system cost
(including membranes), $
Total installed cost, $
Daily operating cost, $
Cost per m3, $
Cost per 1000 gal , $
Cost per metric ton of pulp,
Cost per ton of pulp, $
1
Caustic extraction
filtrate
3,790(1)
524,160
1,349,000
2,371,068
1,095
0.29
1.10
$ 1.51
1.37
2
Caustic extraction
filtrate
7,580(2)
960,000
2,277,600
3,972,151
1,835
0.24
0.92
2.52
2.29
3
Decker
effluent
3,790(1)
462,760
1,192,400
2,137,736
998
0.26
1.00
0.78
0.71
4
Decker
effluent
7,580(2)
842,810
1,998,710
3,556,524
1,663
0.22
0.83
1.31
1.19
-------
TABLE 32. CASE 1 DESIGN CAPITAL COST SUMMARY WITH
FUTURE REDUCTIONS IN ULTRAFILTRATION SYSTEM COSTS CONSIDERED
Pretreatment system
Permeate collection/
distribution subsystem
Concentrate collection/
distribution subsystem
Installation @ 40% of
total auxiliary equipment
cost
Auxiliary equipment
installed cost
Ultrafiltration system
(includes $524,160
for membranes)
Ultrafiltration system
installation @ 15% of
UF system cost
Ultrafiltration system
installed cost
Total equipment cost
Detailed engineering design
@ 12% of auxiliary equipment
installed cost and 5% of UF
system installed cost
Building (372 m2 @ $194/m2)
[4000 ft2 @ $18/ft2]
Subtotal A
Administration and super-
vision (2% of subtotal A)
Subtotal B
Contingency (10% of subtotal B)
Subtotal C
Inflation (10% of subtotal C)
Total installed cost
101,040
64,560
12,120
177,720
71,088
1,144,000
171.600
248,808
1,315,600
1,564,408
87,057
72,000
1,723,465
34.469
1,757,934
175,793
1,933,727
193.372
$ 2,127,099
166
-------
TABLE 33. CASE 1 DESIGN OPERATING COST DATA WITH FUTURE REDUCTIONS IN
ULTRAFILTRATION SYSTEM COSTS CONSIDERED
Cost Element
Material (for cleaning)
Caustic (50X)
EDTA
Total materials
Conversion expense
Labor (Including benefits)
Electrical power 9,
Repair and Maintenance
Material 1.5X of total equipment
replacement cost yearly
Labor
Insurance and Taxes 0.5 x maintenance materl
Total, excluding depreciation
Depreciation
Membranes $
Other facilities $ 1,
Total conversion expense
Total incremental cost
Quantity
37.6 kq/day
4.9 kg/day
6 hr/day
OSS kwh/day
less membrane
- (1.5X x 1,410
4 hr/day
al
153,600
973,499
Unit Cost
$ 0.154/kg
1.95/kg
$ 12/hr
0.025/kwh
,808)/yr
$ 18/hr
3-yr life
15-yr life
$/day
6.39
20.26
26.65
72.00
226.38
57.98
72.00
28.99
457.35
140.27
360.46
958.08
$ 984.73
$/year
2,333
7,395
9,728
26,280
86,627
21,162
26,280
10,581
166,933
51,200
131.567
349,699
359,427
$/m3
0.007
0.019
0.060
0.015
0.019
0.008
0.1Z1
0.037
0.095
0.253
0.261
$/M-gal
0.03
0.07
0.23
0.06
0.07
0.03
0.46
0.14
0.36
0.96
0.98
X of Total
3.1
7.1
23.5
6.1
7.1
3.1
46.9
14.3
36.7
97.9
100.0
-------
The future case 1 system total installed cost becomes $2,127,099 a 10%
reduction from the estimated cost based on current technology. The future
operating cost is $984.73/day. This is equivalent to $0.25/m3 ($0.93/1000
gal) or $1.35/metric ton ($1.23/ton) of pulp. This also represents a 10%
cost savings over current projections.
The uncertainties associated with the future costs of labor, materials
and equipment make it difficult to extrapolate too many years ahead. It may
be that reduced ultrafiltration system costs and lower membrane replacement
costs will, rather than produce overall cost savings, just counteract other
increases to hold down costs to the projections shown in Tables 23 to 30.
COSTS FOR DESIGN CASES 5 THROUGH 8 (IDEALIZED SPIRAL-WOUND MODULE SYSTEMS)
Bases for Capital Cost Projections
Capital cost estimates for full-scale treatment system employing spiral-
wound modules were calculated using procedures similar to those described
for the tubular-based systems. Differences from the procedures described for
the tubular-based systems are listed below.
— System costs averaged $236.8/m2 ($22/ft2) of membrane area.
— Installation costs for the spiral-wound systems were calculated
as 40% of the equipment cost. These systems will be of similar
complexity to the tubular systems and will require similar
installation. Because they have a lower capital cost than
tubular systems they require larger multipliers to present real-
istic associated costs.
-- Detailed engineering design costs include 12% of the UF system
cost. The reasoning for this change is the same as described
above for installation costs.
Design 5 Capital Cost (Idealized Spiral-Wound Module System)
Pretreatment--
Pretreatment requirements for design case 5 (and all spiral-wound
design cases) are identical to those described for tubular systems since
the additional prefiltration required (cartridge filters) is included in
the ultrafiltration system capital cost. The case 5 pretreatment is there-
fore analogous to the case 1 pretreatment.
TOTAL PRETREATMENT SYSTEM COST $ 84,200
Ultrafiltration System—
The compact nature of spiral-wound systems, as compared to tubular
systems, greatly reduces hardware costs. Overall hardware cost for the
case 5 ultrafiltration system is $164.7/m2 ($15.30/ft2).
168
-------
UF system hardware - $298 200
UF membranes - 1,810 m2 (19,488 ft2) @
$67.3/m* ($6.25/ft2) $121 )800
TOTAL UF SYSTEM COST $420,000
Permeate and Concentrate Collection/Distribution Systems-
No change from the case 1 study occurs for either the permeate or
concentrate collection/distribution system.
TOTAL PERMEATE COLLECTION/DISTRIBUTION SYSTEM $ 53,800
TOTAL CONCENTRATE COLLECTION/DISTRIBUTION SYSTEM $ 10,100
Summary—
Table 34 details the case 5 capital costs. Total installed cost is
$1,136,426. The ultrafiltration system cost makes up 37% of this total.
Design Case 6 Capital Cost (Idealized Spiral-Wound Module System)
Pretreatment System—
o
The pretreatment system is the same as for the 7,580 m /day (2 MM gpd)
case 2 system.
TOTAL PRETREATMENT SYSTEM COST $155,700
Ultrafiltration System—
UF System Hardware - 3,282 m2 (35,328 ft2) 9
$176.5/m2 ($16.4/ft2) $579,200
UF Membranes - 3,282 m2 (35,328 ft2) @ $67.3/m2
($6.25/ft2) $220.800
TOTAL UF SYSTEM COST J800.000
Permeate and Concentrate Collection/Distribution Systems—
These system components are the same as for the case 2 design.
TOTAL PERMEATE COLLECTION/DISTRIBUTION SYSTEM $ 71,100
TOTAL CONCENTRATE COLLECTION/DISTRIBUTION SYSTEM $ 14.400
169
-------
TABLE 34. CASE 5 DESIGN CAPITAL COST SUMMARY*
Pretreatment system 84,200
Permeate collection/
distribution subsystem 53,800
Concentrate collection/
distribution subsystem 10,100
148,100
Installation @ 40% of
total auxiliary equipment
cost 59,240
Auxiliary equipment installed
cost 207,340
Ultrafiltration system
(includes $121,800 for
membranes) 420,000
Ultrafiltration system
installation @ 40% of
UF system cost 168,000
Ultrafiltration system
installed cost 588,000
Total equipment cost 795,340
Detailed engineering design
@ 12% of auxiliary equipment
Installed cost and 12% of UF
system Installed cost 95,440
Building (186 m2 @ $161/m2)
[2000 ft2 @ $15/ft2] 30.000
Subtotal A 920,780
Administration and super-
vision (2% of subtotal A) 18.416
Subtotal B 939,196
Contingency (10% of subtotal B) 93.919
Subtotal C 1.033,115
Inflation (10% of subtotal C) 103.311
Total installed cost $ 1.136.426
* Based on idealized spiral-wound modules. Not attainable with current
technology.
170
-------
Summary--
The case 6 capital costs are summarized in Table 35. Total installed
cost is $2,061,239. The ultrafiltration system is 39% of this capital cost.
Design Cases 7 and 8 Capital Costs (Idealized Spiral-Wound Module Systems)
Capital costs for design cases 7 and 8, 3,790 m3/day (1 MM gpd) and
7,580 np/day (2 MM gpd) spiral-wound systems for decker effluents, follow
directly from the 6 cases previously discussed. These costs are developed
in Tables 36 and 37. Total installed costs for case 7 are $1,039,666; for
case 8, $1,867,716.
Bases for Operating Cost Projections
The operating costs for the spiral-wound systems differ from those of
the tubular systems in that:
- cartridge filter replacement costs are included in the materials.
A flowrate of 81.75 m3/day (15 gpm) per 0.51 m (20 in) cartridge
is assumed. Each cartridge is projected to have a 72-hour life.
Cartridges cost $4 each.
- module life (membrane replacement life) is 1 year.
- membrane replacement cost is $67.3/m2 ($6.25/ft2) ($200/module,
minimum order quantity is 250 modules).
Operating Costs for Design Cases 5 through 8
The operating costs for the treatment systems incorporating spiral-
wound modules are detailed in Tables 38 through 41. As with the tubular
systems, depreciation accounts for the major portion of the operating
costs. The 1-year module life results in membrane replacement costs
being >40% of operating costs. Other facility depreciation contributes
nearly 25% to total operating costs.
Summary of Capital and Operating Costs for Design Cases 5 through 8
A summary of projected economics for treatment systems based on spiral-
wound module technology is presented in Table 42. In reviewing this table
and all cost data presented in this section it must be remembered that
these costs are based on ideal spiral-wound ultrafiltration systems. While
useful for comparison purposes among themselves and with tubular-based
systems, the economics given in Table 42 have no intrinsic value. Spiral-
wound systems currently cannot process the pulp mill effluent streams at
the flux levels on which these economics are based.
171
-------
TABLE 35. CASE 6 DESIGN CAPITAL COST SUMMARY*
Pretreatment system 155,700
Permeate collection/
distribution subsystem 71,100
Concentrate collection/
distribution subsystem 41.400
241,200
Installation @ 40% of
total auxiliary equipment
cost 96.480
Auxiliary equipment installed
cost 337,680
Ultra-filtration system
(includes $220,800 for
membranes) 800,000
Ultrafiltration system
installation @ 40% of
UF system cost 320,000
Ultrafiltration system
installed cost 1.120.000
Total equipment cost 1,457,680
Detailed engineering
design @ 12% of
auxiliary equipment
installed cost and
12% of UF system
installed cost 174,922
Building (232 m2 @ $161/m2)
[2,500 ft2 @ $15/ft^] 37.500
Subtotal A 1,670,102
Administration and super-
vision (2% of subtotal A) 33.402
Subtotal B 1,703,504
Contingency (10% of subtotal B) 170,350
Subtotal C 1,873,854
Inflation (10% of
subtotal C) 187.385
Total installed cost $ 2,061,239
* Based on idealized spiral-wound modules. Not attainable with current
technology.
172
-------
TABLE 36. CASE 7 DESIGN CAPITAL COST SUMMARY*
Pretreatment system
Permeate collection/
distribution subsystem
Concentrate collection
distribution subsystem
Installation @ 40% of
total auxiliary equipment
cost
Auxiliary equipment installed
cost
Ultrafiltration system (includes
$107,400 for membranes)
Ultrafiltration system installation
@ 40% of UF system cost
Ultrafiltration system installed
cost
Total equipment cost
Detailed engineering design
@ 12% of auxiliary equipment
installed cost and 12% of
UF system installed cost
Building (186 m2 @ $161/m2)
[2000 ft2 @ $15/ft2]
Subtotal A
Administration and super-
vision (2% of subtotal A)
Subtotal B
Contingency (10% of
subtotal B)
Subtotal C
Inflation (10% of
subtotal C)
Total installed cost
84,200
53,800
10,100
148,100
59,240
370,000
148.000
207,340
518.000
725,340
87,041
30.000
842,381
16.848
859,229
85.922
945,151
94.515
$ 1.039.666
•••^•-^^^^••i ••• •! ifc^^^»
* Based on idealized spiral-wound modules.
technology.
Not attainable with current
173
-------
TABLE 37. CASE 8 DESIGN CAPITAL COST SUMMARY*
Pretreatment system 155,700
Permeate collection/
distribution subsystem 71,100
Concentrate collection/
distribution subsystem 14.400
241,200
Installation @ 40% of
total auxiliary equipment
cost
Auxiliary equipment installed
cost
Ultrafiltration system (includes
$194,400 for membranes)
Ultrafiltration system installation
@ 40% of UF system cost
Ultrafiltration system installed
cost
Total equipment cost
Detailed engineering design @ 12%
of auxiliary equipment installed
cost and 12% of UF system installed
cost
Building (232 m2 @ $161/m2) [2,500 ft2
-------
TABLE 38. CASE 5 DESIGN OPERATING DATA*
Cost Element
Material (for cleaning)
Caustic (50*)
EDTA
Cartridge Filters (0.76 m)
Total materials
Conversion expense
Quantity
37.6 kg/day
4.9 kg/day
16/day
Labor (including benefits) 6 hr/day
Electrical power 2,237 kwh/day
Repair and Maintenance
Material 1.5% of total equipment less membrane
replacement cost yearly » (1.5% x 673,
Labor 2 hr/day
Insurance and Taxes 0.5 x maintenance material
Total, excluding depreciation
Depreciation
Membranes $ 121,800
Other facilities $ 1,014,626
Total conversion expense
Total incremental cost
Unit Cost
$ 0.154/kg
1.95/kg
4/cartridge
$ 10/hr
$ 0.0025/kwh
540}/yr
$ 15/hr
3-yr life
15 -yr life
$
S
$/day
5.81
18.41
64.00
88.22
60.00
50.30
27.68
30.00
13.84
181.82
333.70
185.32
700.84"
789.06
$/year
2,121
6,710
23,360
32,201 '
21,900
18,360
10,103
10,950
5,052
66,365
121,800
67,642
255,807
288,008
$/rt)3
0.023
0.016
0.013
0.007
0.008
0.004
0.048
0.088
0.049
0.185
0.209
$/K-gal
0.09
0.06
0.05
0.03
0.03
0.01
0.18
0.33
0.19
0.70
0.79
% of Total
11.4
7.6
6.3
3.8
3.8
1.3
22.8
41.8
24.1
88.6
100.0
•Based on idealized spiral-wound modules. Not attainable with current technology.
-------
TABLE 39. CASE 6 DESIGN OPERATING COST DATA*
CT>
Cost Element
Material (for cleaning)
Caustic (50%)
EDTA
Cartridge Filters (0.76 m)
.Total materials
Conversion expense
Labor (including benefits)
Electrical power
Repair and Maintenance
Material 1.5*
rep la
Labor
Insurance and Taxes 0.5 x
Total, excluding
Depreciation
Membranes
Other facilities
Total conversion expense
Total incremental cost
Quantity
53.8 kg/day
20.8 kg/day
32/day
6 hr/day
3,811 kwh/day
of total equipmment less membrane
cement cost yearly * (1.55! x 1,236
3.5 hr/day
maintenance material
depreciation
$ 220,800
$ 1,840,439
Unit Cost
S/day
$ 0.154/kg 8.31
1.95/kg 26.28
$ 4/cartridge 128.00
$ 10/hr
$ 0.0025/kwh
,880)/yr
$ 15/hr
1-yr Hfe
15-yr life
162.59
60.00
85.75
50.83
52.50
25.42
274.50
604.93
336.15
$ L, 215. 58
$ 1,378.17
$/year
3,033
9,952
46,720
59,345
21,900
31,299
18,553
19,163
9,278
100,193
220,800
122,696
443,687
503,032
S/m3
0.022
0.008
0.011
0.007
0.007
0.003
0.036
0.080
0.044
0.160
0.182
$/M-gal
0.08
0.03
0.04
0.03
0.03
0.01
0.14
0.30
0.17
0.61
0.69
% of Total
11.6
4.3
5.8
4.3
4.3
1.4
20.3
43.5
24.6
88.4
100.0
Not attainable with current technology.
-------
TABLE 40. CASE 7 DESIGN OPERATING COST DATA*
Cost Element Quantity Unit Cost $/day J/year $/m3 $/M-gal
Material (for cleaning)
Caustic (S03S) 37.6 kg/day $ 0.154/kg 5.81 Z.1Z1
EDTA 4.9 kg/day 1.95/kg 18.41 6,720
Cartridge Filters (0.76 m) 16/day $ 4/cartridge 64.00 23,360
Total materials 88.22 32,201 0.023 0.09 " "
Conversion expense
Labor (including benefits) 6 hr/day $ 10/hr 60.00 21,900 0.016 0.06
Electrical power 1,915 kwh/day $ 0.0025/kwh 40.09 14,633 0.011 0.04
Repair and Maintenance
Material 1.556 of total equipmment less membrane
replacement cost yearly • (1.5* x 617,940)/yr 25.39 9,269 0.007 0.03
Labor 2 hr/day $ 15/hr 30.00 10,950 0.008 0.03
Insurance and Taxes 0.5 x maintenance material 12.70 4,636 0.003 0.01
Total, excluding depreciation 168.18 61,386 0.044 0.17
Depreciation
Membranes $ 107,400 1-yr life 294.25 107,400 0.078 0.29
Other facilities $ 932,266 15-yr life 170.28 62,151 0.045 0.17
Total conversion expense I 632.71 203,939 0.167 0.63
Total incremental cost $ 720.93 236,140 0.191 0.72
% of Total
12.5
8.3
5.6
4.2
4.2
1.4
23.6
40.3
23.6
87.5
100.0
Based on idealized, spiral-wound modules. Not attainable with current technology.
-------
TABLE 41. CASE 8 DESIGN OPERATING COST DATA*
OO
Cost Element
Material (for cleaning)
Caustic (50%)
EOTA
Cartridge Filters (0.76 m)
Total materials
Conversion expense
Labor (including benefits)
Electrical power
Repair and Maintenance
Material 1.5%
repla
Labor
Insurance and Taxes 0.5 x
Total, excluding
Depreciation
Membranes
Other facilities
Total conversion expense
Total incremental cost
Quantity
53.8 kq/day
13.5 kq/dav
32/day
6 hr/day
3,634 kwh/day
of total equipmnent less membrane
cement cost yearly = (1.5% x 1,123
3 hr/day
maintenance material
depreciation
$ 194,400
$ 1,673,316
Unit Cost
$ 0.154/kg
1.95/kg
$ 4/cartfidge
$ 10/hr
$ 0.0025/kwh
,280)/yr
$ 15/hr
1-yr life
15-yr life
S/day
8.31
26.28
128:00
162.59
60.00
81.77
46.16
45.00
23.08
256.01
532.60
305.63
J 1,094.24
$ 1,256.83
$/year
3,033
9,592
46,720
59,345
21,900
29,846
16,849
15,425
8,424
93,444
194,400
111,554
399,398
458,743
$/m3
0.021
0.008
0.011
0.006
0.006
0.003
0.034
0.070
0.040
0.144
0.166
$/M-gal
0.08
0.03
0.04
0.02
0.02
0.01
0.13
0.27
0.15
0.55
0.63
% of Total
12.7
4.8
6.3
3.2
3.2
1.6
20.6
42.9
23.8
87.3
100.0
* Basea on idealized spiral-wound modules. Not attainable with current technology.
-------
TABLE 42. SUMMARY OF PROJECTED ECONOMICS FOR DESIGN CASES 5 THROUGH 8*
•-4
VD
Item
Stream
Flow, m3/day (MM gal /day)
UF membrane cost, $
UF system cost
(including membranes), $
Total installed cost, $
Daily operating cost, $
Cost per m3, $
Cost per 1000 gal, $
Cost per metric ton of pulp,
Cost per ton of pulp, $
5
Caustic extraction
filtrate
3,790(1)
121,800
420,000
1,136,426
789
0.21
0.79
$ 1.08
0.99
Design Case Number
6
Caustic extraction
filtrate
7,580(2)
220,800
800,000
2,061,239
1,378
0.18
0.69
1.89
1.72
7
Decker
effluent
3,790(1)
107,400
370,000
1,039,666
721
0.19
0.72
0.57
0.50
8
Decker
ef f 1 uent
7,580(2)
194,400
700,000
1,867,716
1,257
0.17
0.63
0.99
0.87
Based on idealized spiral-wound modules. Not attainable with current technology-
-------
As observed in Table 42, capital investment ranges from $1 to $2.1 MM.
This is roughly 50% of the installed costs for the tubular-based treatment
systems. Treatment costs are $1.09 to $1.89/metric ton ($0.99 to $1.72/ton)
of pulp for caustic extraction filtrate processing and $0.55 to $0.95/
metric ton ($0.50 to $0.87/ton) of pulp for decker effluent processing.
POTENTIAL CREDITS FOR WATER REUSE AND RESOURCE RECOVERY
No credits have been applied to the treatment system operating costs
for water reuse or resource recovery. A qualitative discussion of these
potential credits is presented in the next section, entitled, Water Reuse
Potential.
180
-------
SECTION 9
WATER REUSE POTENTIAL
In the foregoing economic projections no credits have been applied to
the treatment system operating costs for water (permeate) reuse or concen-
trate reuse. It is expected that for a mill installation these streams will
be recycled to some extent and that operating cost credits will accrue.
Reuse applications and potential credits are discussed below.
PINE BLEACHERY CAUSTIC EXTRACTION FILTRATE
The permeate from the ultrafiltration unit treating caustic extraction
filtrate will constitute about 98% of the feed stream. This effluent will
have low color, essentially no suspended solids and will have very low heavy
metal content. In addition the permeate will have a high pH and be at
process temperature.
The permeate, with its physical and chemical attributes, should be a
superior water makeup stream for use in the bleachery processes. The high
pH, and reduced buffering capacity should allow for lower new caustic
requirements. The high temperature should reduce the system energy require-
ments. Because of the absence of suspended solids and the decreased heavy
metal content the permeate should reduce spray head and other scaling
problems.
No estimate is made here for the potential economic value of these
factors since the potential will be quite site-specific. It is believed
that at least half of the permeate can be used in a bleachery recycle mode
and that the savings in chemicals, water and energy from such use will have
substantial effects in reducing the net cost of the operation.
Permeate which is excess can be admixed with mill input fresh water
without discernable effect upon the fresh water quality. This is because
the small permeate volume would be diluted 25 to 50 times by the larger
fresh water input.
Recycling all or substantially all of the permeate as above would have
an additional benefit of reducing the flow of effluent to be treated in the
waste disposal area by 2-4% which is not quantifiable in a generic case.
181
-------
The concentrate from ultra-filtration has no direct value in pulp mill
operations but it will be small in volume and because of its high organic
solids content will be combustible. Disposal of this material would be site-
specific. In those installations equipped to remove chlorides from black
liquor systems, this concentrate could be flowed to the weak black liquor
system with some small gain in energy recovery. Alternatively, the concen-
trate could be combusted in a typical modern lime kiln without noticeable
effect, especially because of the low sodium content. Some mills which do
not have sufficient lime kiln capacity dispose of the excess lime sludge
off-site. Because of the high pH of this sludge and the relatively small
volume of the concentrate, the concentrate could be added to the lime sludge
and carried to land fill as insolubilized calcium salts.
Because the water reuse and resource recovery considerations for the
ultrafnitration permeates and concentrates are site-specific no formal
estimate of credits are projected here. It is felt that the reuse of
permeate will result in values in reduction of water, chemical, energy and
maintenance costs, and waste disposal system loading which will substantially
reduce the net cost of operation of the ultrafiltration unit.
The disposal of the concentrate as discussed above should result in at
least a break-even disposal cost.
PINE AND HARDWOOD PULP WASHING DECKER EFFLUENTS
In a previous study (32) it was demonstrated that ultrafiltration of
decker effluents for the specific cases considered would reduce the net costs
of color removal by ultrafiltration to break-even or even provide a return
on the investment. Calculations based on the present tubular membrane
economic projections indicate that recycling the permeates to the pulping
system for makeup water and the addition of the concentrates to the weak
black liquor system, although site-specific, still offer the potential for
break even or low net cost operation of an ultrafiltration unit. The cost
values here result from recovery and recycling of water, chemicals, energy
and reduction of waste disposal system loading by 2-4%.
Large scale demonstrations of reuse or disposal were not possible in
this study. The pilot quantities produced were small relative to the
actual mill material flows. A substantially larger prototype demonstration
plant could provide the quantities of permeates and concentrates required
for reasonable scale reuse or disposal systems testing.
182
-------
REFERENCES
1. Lockwood Directory 1977, Venice Publishing Co.
2. 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)
3. Spruill, Edgar L., "Paper Mill Waste: Treatment for Color Removal,"
Industrial Wastes. 15, 21-23 (March/April 1971)
4. Gould, Matthew, "Lime-Based Process Helps Decolor Kraft Wastewater,"
Chem, Eng.. 55-57 (January 25, 1971)
5. 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, 1931-1934 (October 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. I. Treatment of Caustic Extraction Stage Bleaching
Effluent," Technical Bulletin No. 239 (July 1970)
7. National Council for Air and Stream Improvement, Inc., "The Mechanisms
of Color Removal in the Treatment of Pulping and Bleaching Effluents
with Lime. II. Treatment of Chiorination State Bleaching Effluents,"
Technical Bulletin No. 242 (December 1970)
8. Davis, C.L., Jr., "Tertiary Treatment of Kraft Mill Effluent Including
Chemical Coagulation for Color Removal," TAPPI. 52., No. 11, 2132-2134
(November 1969)
9. Middlebrooks, E.J., Phillips, W.E., Jr., and Coogan, F.J., "Chemical
Coagulation of Kraft Mill Wastewater," Industrial Wastes Water and
Sewage Works Supplement, 7-9 (March 19691
10. Smith, S.E. and Christman, R.F., "Coagulation of Pulping Wastes for the
Removal of Color," Journal Water Pollution Control Federation, 41, No.
2, Part 1, 222-231 (February 1969)
11. "Projects of the Industrial Pollution Control Branch, July 1971,"
Water Pollution Control Research Series #12000-07/71, 5-22, 5-23,
5-24, 5-25
183
-------
12. Ibid., 5-13, 5-26
13. Proceedings of the TAPPI 8th Water and Air Conference, Boston, Mass.,
1971 paper entitled "Activated Carbon System for Treatment of Paper
Mill Washwaters."
14. Ibid.. "Color Removal from Kraft Bleach Wastes by Ion Exchangers."
15. "Carbon Treatment of Kraft Condensate Wastes," TAPPI, 51_, 241 (1968)
16. "Ozone Decolonization of Effluents from Secondary Treatment," Paper
Trade Journal (January 28, 1974)
17. "Ozone: A New Method to Remove Color in Secondary Effluents," Pulp
and Paper (September 1974)
18. "Photochemical Decolorization of Pulp Mill Effluents," TAPPI. 58., No.
2 (February 1975)
19. "The Effect of Gamma Irradiation on Pulp and Paper Mill Effluents,"
Applied Polymer Symposium, No. 28, 1321-1220 (1976)
20. Lacey, R.E. and Loeb, S., eds., Industrial Processing with Membranes.
Wiley-Interscience, New York, 223+ (1972)
21. 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)
22. Wiley, A.J., Dubey, G.A. and Bansai, I.K., "Reverse Osmosis Concen-
tration of Dilute Pulp & Paper Effluents", for the Environmental
Protection Agency, Program #12040 EEL (February 1972)
23. 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)
24. Bansai, I.K., Dubey, G.A., and Wiley, A.J., "Development of Design
Factors for Reverse Osmosis Concentration of Pulping Process Effluents",
Presented at Membrane Symposium, National Meeting of American Chemical
Society, Chicago, Illinois September 14-18, 1970
25. Beder, H. and Gillespie, W.J., "Removal of Solutes from Mill Effluents
by Reverse Osmosis," TAPPI. 53_, No. 5, 883-887 (May 1970)
26. Bregman, Jacob I., "Membrane Processes Gain Favor for Water Reuse",
Environmental Science & Technology. 4_, No. 4, 296-302 (April 1970)
27. Wiley, A.J., Dubey, G.A., Holderby, J.M. and Ammerlaan, A.C.F.,
"Concentration of Dilute Pulping Wastes by Reverse Osmosis and Ultra-
filtration," Journal Water Pollution Control Federation, 42, No. 8,
Part 2, R279-R289 (August 1970)
184
-------
28. Ammerlaan, A.C.F. and Wiley, A.O., "Pulp Manufacturers Research League
Demonstrates Reverse Osmosis Process," TAPPI, 52. (1969)
29. Ammerlaan, A.C.F. and Wiley, A.J., "The Engineering Evaluation of
Reverse Osmosis as a Method of Processing Spent Liquors of the Pulp
and Paper Industry," prepared for the New Orleans Meeting of A.I.Ch.E.,
March 17-20, 1969
30. Ammerlaan, A.C.F., Lueck, B.F. and Wiley, A.J., "Membrane Processing
of Dilute Pulping Wastes by Reverse Osmosis," TAPPI, 52, No. 1,
118-122 (January 1969)
31. Percolating Effluent into Ground Reduces Color at Missoula Mill,"
Pulp and Paper (October 1976)
32. Fremont, H.A., Tate, D.C., and Goldsmith, R.L., "Color Removal from
Kraft Mill Effluents by Ultrafiltration," for the Environmental
Protection Agency, Project #5800261 (May 1973) EPA-660/2-73-019.
33. Spatz, D. Dean, Friedlander, Richard H., "Rating the Chemical Stability
U.C. RO/UL Membrane Materials," Water & Sewage Works, 36-40 (February
1978)
34. "Industrial Ultrafiltration," in Membrane Processes in Industry and
Biomedicine, Plenum Press (1971)
35. Gollan, A.Z., et a!., "Evaluation of Membrane Separation Processes,
Carbon Adsorption, and Ozonation for Treatment of MUST Hospital Wastes,"
Final Report for USAMRDC Contract No. DAMD17-74-C-4066, August, 1976
185
-------
APPENDICES
APPENDIX A. MEMBRANE CLEANING MATERIALS AND TECHNIQUES
Three classes of membrane cleaning materials were investigated:
- detergents
- enzymes
- chelating agents
The following materials and combinations were investigated for membrane
cleaning effectiveness.
a. Ultraclean (Abcor)
b. Enzyme (Abcor)
c. Iron Chelating Agent (Abcor)
d. C-S Detergent (Osmonics)
e. Ultrazme (Osmonics)
f. Caustic and EDTA
g. Oxalic acid
h. Ultraclean and caustic
i. Ultraclean and EDTA
j. Ultraclean and Iron chelating agent
k. Ultraclean and Enzyme
During the membrane cleaning studies sufficient work was done with the
chelating agents to determine that the "fouled" membrane was not the result
of metal fouling.
Plugging of the spiral wound membrane modules was not a problem during
this program. The prefiltration system removed the suspended particles and
cleaned the stream sufficiently to prevent plugging. A "dirty" module was
the result of a formed boundary layer on the membrane surface that needed to
be "scoured" from the surface. Backflushing of the module was not at all
effective in cleaning a fouled membrane. Soaking of the module in water or
sufficient recirculation time with a cleaning solution to loosen the bound
foul ant layer proved the only effective means to clean the modules.
186
-------
APPENDIX B. DERIVATION OF THE EQUATION RELATING INTRINSIC REJECTION TO
APPARENT REJECTION
The intrinsic rejection (R.j) is defined as follows:
where
C = Concentration in the permeate
Cf = Concentration in the feed
The apparent rejection (Ra) is defined as follows:
where
C = Average concentration in the permeate
Cf = Original concentration in the feed (assumed equal to
o Cf in calculating Ra)
Equation 1 can be rearranged to yield
Cp = (l-R1)Cf (3)
The concentration in the feed as a function of volume processed can
be expressed as follows:
-R.
C = C (^ (4)
^f f \v /
where
Vo = Initial volume in the cell
V = Volume in the cell at the time t
V and Vo can be related to the conversion (Y) as follows:
v - Vo - V - i I- (5)
Y -- Vo -- ' Vo
187
-------
Equation 5 can then be rearranged and substituted into equation 4 to yield:
Cf - Cfo (l - V) ^ (6)
Equation 6 can then be substituted into equation 3 to yield:
-Ri
C = (1 - R1)Cf (1 - Y) n
or
-R
- R,)0 - V) f (7)
fo
Integration of equation 7 with respect to Y yields:
Y r Y
I 2 J^ dy = (1 - R.) / 2 (1 - Y)-Ri dy
/Y1 fo ^1
or
1-R, 1-R,
C
-, -,
(1 - Y,) 1 - (1 - Y2) 1
'
__ (Y2 - Y,)
o
Substituting equation 2 into equation 8 yields:
1-R, 1-R
188
-------
APPENDIX C. ADDITIONAL DATA FROM MEMBRANE SELECTION STUDIES
o
I
CM
SI
s:
x
ZD
U.
LU
LU
LU
O.
4.0
3.6
3.2
2.8
2.4
2.0
1.6
1.2
0.8
0.4
0 2
•• • 1—
" 1 — 1 1 1 , T
0 -
v— - -° _
°^n
_ tJ--^^*^
0 HFD-GH500 JQ ~
0 HFD-FH250 ^
/**
&
-
* **
o °b
^
Eg 0
. _ n ~
|Q- ^
~u -
-
~^-~-^~.Q
™* ™
. TEST 1 TEST , TEST TEST , , TEST
i ,r « , r ,3 JC« \A- 6, •
100
90
80
70
60
50
40
30
20
10
9
8
7
6
5
0 20 40 60 80 100 120 140
CUMULATIVE OPERATING TIME (HOURS)
Figure Cl. Coated HFD membrane flux data obtained during
parametric studies.
189
-------
O HFM-FH500
D HFM-GH500
A HFM-FH250
88
4.8
4.0
o
CM
2.4
UJ
£
UJ
o.
0.8
--o-
o HFM-FH500
n HFM-GA500
A HFM-FH250
iiii
i | i
°\
Inlet pressure: 3.4 atm
Circulation rate: 163.5 m3/day
Temperature: 49°C
ll
100
80
60
40
20
•O
m
m
en
TI
o
10 100
CUMULATIVE OPERATING TIME (HOURS)
1000
Figure C2. Coated HFM membrane flux and rejection characteristics
at constant operating conditions.
190
-------
8.0
4.0
3.6
3.2
2.8
I 2.9
CM
CO
2.0
1.6
S 1.2
UI
Q.
0.8
0.4
100
90
80
70
60 m
50
40
sn
Tl
o
30
O HFM-GH500
° HFM-FH250
20
TEST
TEST
TEST
TEST
TEST
10
20 40 60 80 100
CUMULATIVE OPERATING TIME (HOURS)
120
140
Figure C3. Coated HFM membrane flux data obtained during
parametric studies.
191
-------
APPENDIX D. FINAL REPORT - USE OF SULFONIC ACID MEMBRANES FOR TREATMENT OF
PULP AND PAPER WASTE STREAM - PREPARED BY H. GREGOR, COLUMBIA
UNIVERSITY, SUBCONTRACT TO CHAMPION, P.O. 165723*
PURPOSE
The purpose of this subcontract was to have Gregor, on the basis of his
laboratory tests, select sulfonic acid membranes which appeared to be suit-
able for the treatment of feeds supplied by Champion, using analytical
procedures prescribed by Champion. Selected samples were to be supplied to
Champion for further laboratory tests by them, and then Champion was to
select membranes to be supplied in 0.3 m x 1.8 m (1 ft x 6 ft) sizes for
pilot plant studies.
MEMBRANES AND SUPPORTS
The membranes of US Patent 3,808,305 can be cast on a variety of support
materials. Those which had been employed were examined and a support
membrane which was believed to be well suited for the Champion study was
selected. Its chemical stability was evaluated according to specifications
given by Dr. Fremont (Champion), by subjecting it to solutions at pH 12
(with sodium carbonate) at 57°C (135°F) for several weeks. It was found that
the material was quite stable under these conditions. Then the temperature
was increased to 90°C, much more extreme than that encountered in field use,
and the support membranes continued to be stable.
Membranes Employed
All of the membranes employed for this study were made in accordance
with US Patent 3,808,305, and consisted primarily of sulfonic acid polymers
to make them non-fouling, and were cured by chemical cross-linking. They
were made in a variety of porosities and were characterized by measuring
their hydraulic permeabilities or fluxes, usually at room temperature in
water and in the presence of the dye erythrocin at the 15 ppm level.
Membranes are characterized according to their hydraulic permeabilities in
microns per second - atmosphere (ysa) by their dye rejections as percent
rejection in the usual manner. Membranes of selected porosity and other
characteristics were employed for this study.
TEST EQUIPMENT
In order to meet the specifications set down for this study, the cells
employed were of the Gelman stainless steel variety as has been described
in the original proposal to Champion. These have a small stainless steel
*
This final report is reproduced exactly as received except for required
format changes (retyping, etc.).
192
-------
insert screwed into the upper plate of the cells so as to allow for the
control of convection across the face of the membrane as feed solution is
pumped through the insert, across the face of the membrane and out the exit
port. The distance of separation from the insert to the membrane surface
was 0.5 mm and the radius of average velocity was readily calculated to be
0.822 centimeters for the insert of this dimension. Thus, the area of flow
at the radius of average velocity was 0.26 cm?. These cells have an
effective membrane area of 13.5 cm?, so a 2 psa membrane at 50 psi would
provide for a flux of about 30 ml per hour. These cells are shown in
Figure D-l.
The rate of feed recirculation across the face of the membrane was set
by the voltage supplied to the circulating pump which operated quite inde-
pendently of the total pressure imposed on the system. Tests showed that
average circulation velocities of greater than 0.5 meter/sec did not lead
to higher fluxes, so a rate of 0.5 meter/sec was maintained.
These cells were mounted into a batch-type unit (see Figure D-2) con-
structed entirely of stainless steel, having a reservoir volume of
approximately 2.5 liters, a circulating pump allowing for a rate of convec-
tion across the membrane face which could be varied at will over wide
ranges, capable of operation at temperatures well above 100°C and at quite
high pressures. The residual volume of this device was 60 ml, so in a
single run at that volume of feed the device could be operated to approxi-
mately 94% water recovery.
The absorbance of feeds, concentrates and permeates was determined as
follows, as specified by Champion. First, the pH was measured and then the
sample was diluted with distilled water until the absorbance reached the
useful range of the instrument, which was either a Spectronic 20 or Beckman
DU at 465 nm. At this point the pH was adjusted with sulfuric acid to 7.6
and the final reading was made of the original absorbance of that solution.
The mixed permeate assays were also determined directly by collecting
all of the permeate at a given degree of water recovery and subjecting this
collected permeate to the same analysis and reporting the results in that
manner. One can readily calculate the absorbance of the mixed permeate
and concentrate by calculations from the point by point composition of the^
permeate. This was not done because Abcor preferred the direct determination.
The specifications for this project were ultimately defined by Dr.
Gollan (Abcor): All runs were to be made at 140°F or 60°C. As regards
fluxes measured at 5.1 atm (75 psi), a membrane would be considered poor
if its flux were less than 0.83 m3/m2-day (20 gfd) and the very good
membrane would have a flux greater than 3.28 m3/m2-day (80 gfd). With
respect to rejection, on a mixed permeate up to 98% water recovery, a
rejection of greater than 90% was considered desirable.
193
-------
Feed in
0-ring
SS or 1 ucit
Mpmhranp
Membrane backin
Porus SS suppo
i
: >
.j--'
i — «- -
-^ o
\
e insert — > £J
^\
, , . ^
g — » " "
KXKMWp099ft!
rt *T .
l_
T_
""'T -
1
1
1
k
f
r
1
— *—!
P^
x...^""
— — — — v^
<>-, S_.....
— • »• b
_... > Feed out
^-Top
r>
sSs^ Stc
^SNS^ men
±^ ^i
*hn,,,- - "I
'M&Kft&Qixtspk
t
/
• — *• ~ *
— r
i
i
»
_ j
n'nless steel
ibrane holder
Bottom
r
Product
0.5
7 mm cj>
Gelman cell:
effective membrane
area - 13.5 cm^,
maximum pressure-102 atm.
_Q
y —
[V
n
\tr- 32 mm
__ 44 mm
membrane
Figure Dl. Modified Gelman filter holder.
194
-------
SS Tube
heating
coils
Gel man
cell
gas tank
P = pressure guage
T = temperature guage
Circulating Pump
10.9 m3/day (?. gpm)
Figure D2. High pressure-high temperature UF/RO assembly.
195
-------
The specifications of this project were subject to some variation as the
management of the project changed.
MEMBRANE SAMPLES SUBMITTED
A number of membrane samples were submitted. The set labeled A-E had
the following characteristics:
TABLE D-l. MEMBRANE PROPERTIES
^HWBWIW*HI*IHIMBBBI»B~BBBII^^
Designation
A
B
C
D
E
J sa
10
5.0
2.6
0.80
1.3
%DR
25
68
97
80
The membrane types used in the runs of Tables D2-D5 are designated
(A - E) by membrane type. A number of different variants within each type
were prepared and tested; the variant was usually that of thickness. For
example, membranes of designation D all had high rejections of dye, and
making them in thinner forms gave films of higher flux, with little loss in
dye rejection.
RESULTS OBTAINED - DISCUSSION
A number of membranes were subjected to tests at 80°C at 6.8 atm
(100 psi). Membranes having a wide range of dye rejections from nearly
zero to 99% were tested; these had fluxes from 0.5 ysa as high as 30 ysa,
all in water at room temperature. Only in those cases when membranes gave
a reasonable flux and a good level of dye rejection was the run continued
to the end. A number of preliminary experiments were also performed with
the Champion feed at room temperature, but here it was found that the results
were not at all indicative of those obtained at higher temperatures. At
general, fluxes were lower as expected, but the color rejection in some
cases was erratic, possibly due to association of lower molecular weight
color bodies in the feed with other molecules of higher molecular weight
which never permeated the membrane. The color of the feed stayed constant
over the entire period of its use.
196
-------
A few experiments were also performed to ascertain the performance of
the membrane operating at different temperatures because of the change in
the specifications from 80°C to 60°C. It was found that above 50°C on the
Champion feed the membranes of our study gave rather similar fluxes and dye
rejections, with a 10% to 20% increase in the ultrafiltration rate at 80°C
as compared with 60°C. However, at that point the project was terminated
so no definitive tests were performed thereafter.
Tables D2 to D6 summarize the results obtained. Table D6 shows that
rejection in all cases was well above specifications, that the flux in one
case was at the minimal level, in others at the good level and in two at the
Very good level, being well above 3.28 m3/m2-day (80 gfd) at 5.1 atm
(75 psi).
The pH of the permeate (Table D2) was slightly higher than of the feed,
as is usually the case when material which is a salt of a polyacid is
ultrafiltered. A small degree of hydrolysis of the salt results because the
counterions (Na+) are forced through the membrane while the polyanion remains
in the feed. Electroneutrality is maintained by hydrolysis, producing NaOH
in the permeate and the acid in the feed. These kinds of pH shifts are
frequently found in treating natural feeds.
In Table D3, the pH of the permeate was less than that of the feed, and
this difference increased throughout the run. One could postulate that
polycations were present, but this is most unlikely at these high pH levels.
Table 04 shows that the feed was 2 pH units above the first permeate, and
this difference persisted. Table D5 also shows a much lower pH for the
permeate than for the feed. At the 25%, 50% and 75% treatment points, the
permeate was at pH 9.2 and the concentrate at 8.7-8.9, or the normal
increase in pH was found.
The explanation which suggests itself is that there are colloidal
aggregates present in the original feed which impart a higher pH (by
adsorption to the electrode, possibly), but that these settle out as the run
continues where the expected pH difference between feed and permeate is
observed. Concentrate compositions and pH levels were always taken from
the circulating feed stream, so if material settled out at the bottom of the
feed tank where it would not appear in the concentrate, the pH shifts
observed would result.
It was not possible for us to evaluate or discuss the fundamental
aspects of the results obtained because no information on the chemical
composition of the feeds was supplied. Two feed samples were supplied. The
first had a pH of 8.5, it was small in volume 3.785 liters [1 gal] and was
consumed in a series of preliminary runs. The second sample was larger and
had a pH of 10.3. The former sample was darker, with an absorbance of
5.460, while the latter sample had an absorbance of 4.375 ± 0.010. Both
197
-------
TABLE D-2. CHAMPION EFFLUENT - RUN #12
Soln
Feed
PI
P2
P3
C3
P4
C4
P5
C5
P6
C6
P7
C7
P8
pH
8.5
8.8
8.9
8.6
8.5
8.6
8.5
8.9
8.8
8.9
8.8
8.9
8.8
8.9
Abs
5.460
.104
.154
.216
9.50
.268
10.6
.328
12.6
.348
12.6
.364
13.1
.324
%Rej % Water Rec Flux, m^/m^-day (gfd)
_
98
97
96
-
95
-
94
.
94
-
93
-
93
_
6
16
25
25
34
34
43
43
53
53
59
59
70
_
1.23 (30)
1.31 (32)
1.31 (32)
1.31 (32)
1.68 (41)
1.56 (38)
1.52 (37)
1.56 (38)
P - Permeate
C - Concentrate
Feed: 1000 ml T : 80'C, 176'F
Cell: SS, 13.5 cm2 P: 100 psi
Memb: No. 27, 1.2 ysa.93% DR (Series D)
198
-------
TABLE D-3. CHAMPION EFFLUENT - RUN #16
Soln
Feed
PI
P2
P3
C3
P4
C4
P5
C5
P6
C6
P7
C7
P8
C8
P9
C9
PH
10.3
9.5
9.5
9.4
9.3
9.2
9.3
9.0
8.9
8.5
Abs
4.375
.727
.428
.321
4.51
.253
4.57
.219
5.28
.222
7.71
.327
10.25
.410
13.48
1.510
36.95
SRej
-
85
90
93
94
_
95
-
97
_
97
_
97
96
% Water Rec
_
11.0
21.0
33.5
45.0
34
55.3
43
73.1
53
83.1
59
93.1
98.0
Flux, m3/m2-day (gfd)
.
6.8 (167)
6.8 (167)
6.8 (167)
6.8 (167)
6.8 (167)
6.8 (167)
6.8 (167)
6.86 (167)
3.5 (85)
P - Permeate
C - Concentrate
Feed: 1000 ml T : 80*C, 176*F
Cell: SS, 13.5 cm2 P: 100 psi
Memb: No. 32, 1.5 ;ysa, 78% DR (Series E)
199
-------
TABLE D-4. CHAMPION EFFLUENT - RUN #17
Soln
Feed
PI
P2
P3
P4
C4
P5
C5
P6
C6
P7
C7
P8
C8
pH
10.3
8.3
8.4
8.5
8.5
_
8.4
_
8.7
.
8.6
-
8.5
™
Abs
4.375
.349
.316
.310
.364
9.50
.400
12.75
.541
15.60
.820
25.72
5.320
74.00
%Rej % Water Rec Flux, nvVm^-da
_
92
93
93
96
-
97
-
96
-
97
-
93
—
_
11.5
21.5
32.7
43.3
43.3
60.2
60.2
70.7
70.7
80.7
80.7
90.7
90.7
-
5.1
5.1
5.1
5.1
5.1
5.1
5.1
3.0
y (gfd)
(125)
(125)
(125)
(125)
(125)
(125)
(125)
(73)
P - Permeate
C - Concentrate
* Abs Is measured for appropriately diluted solutions at'pH 7.6 at 465 nm, and
reported as the calculated abs of that solution.
Feed: 1000 ml T : 80'C, 176*F
Cell: SS, 13.5 cm2 P: 100 psi
Memb: No. 31, 5.0psa, 99% DR (Series E)
200
-------
TABLE D-5. CHAMPION EFFLUENT - RUN #18
XRej
% Water Rec Flux, m3/m2-day (gfd)
Feed
PI
P2
C2
P3
C3
MP-25%
Conc-25%
P4
C4
P5
C5
P6
C6
MP-50%
Conc-50%
P7
C7
P8
C8
P9
C9
HP-75%
Conc-75X
P10
CIO
Pll
Cll
HP-90%
Conc-903!
10.3
9.2
9.2
9.2
9.2
8.9
9.3
9.5
9.2
9.2
8.8
9.2
9.2
9.0
9.2
8.8
9.0
9.0
_
™*
4.05
.051
.063
4.30
.073
4.50
.063
4.28
.077
5.30
.092
6.40
.128
6,95
.098
5.21
.150
9.60
.190
12.35
.311
16.55
.217
7.75
.076
66.50
2.88
84.00
1.82
24.5
""— •• ~ -•
99
99
98
98
99
99
98
98
98
98
98
98
99
97
98
" " • • • — — . .
10.5
20.5
25.5
25
25
35.7
45.7
50.7
50
50
60.7
70.7
75.7
75
75
85.7
90.. 0
90
90
. .I— — i
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
7.6
3.6
3.6
3.6
3.6
3.6
3.6
_. . _ .
(87)
(87)
(87)
(87)
(87)
(87)
(87)
(87)
(87)
(87)
(87)
(87)
(87)
(87)
(87)
(87)
(50)
P - Permeate
C - Concentrate
MP - % means mixed permeate collected to that % water recovery.
Feed: 1000 ml T : 80'C, 176'F
Cell: SS, 13.5 cm? P: 100 psi
Memo: No. 39, 8.0 jisa, 99% CR (Series E)
201
-------
TABLE D-6. SUMMARY OF RUNS
No
12
16
17
18
T in 0°C
80
80
80
60
PSI*
100
100
100
100
J-100 psi*
34
167
125
87
J-75 psi*
26
125
94
65
CR
94
96
93
98.5
J-75 values were calculated from J-100 data.
* 75 psi - 5.1 atm; 100 psi - 6.8 atm.
202
-------
samples were stored in the refrigerator, allowed to come to room temperature,
shaken vigorously and then decanted rapidly into the test cell. The pH of
these two feeds did not change over the period of use, and there was no evi-
dence of flocculation or settling during the period of our tests. No pre-
treatment was ever employed; the feed material as supplied was used directly.
There were no identification numbers of the samples as received, other than
the designation Dan C. Tate, 2-14-087.
203
-------
APPENDIX E. PRETREATMENT STUDIES
POLYMER PRETREATMENT OF FEED STREAMS
The rapid flux decline of spiral wound WRP membrane modules was a major
problem. The analysis of the foulants from the surface of a fouled membrane
gave the following analytical results (see attached Nalco analysis).
Volatile at 105°C -88.0%
Kaolinite Clay -5.6%
Starch -4.0%
Titanium Dixoide -1.4%
Carboxylic Acid Salt -<1.0%
These data show that the major components in the "slime" layer come
from recycle of white water from the paper mill back to the pulp mill. The
prefiltration filters were effective in removal of the large suspended
particles that would plug a module. However, the very small particles were
getting into the membrane module and contributing to the fouling problem.
A study was made to find a polymer treatment that would flocculate these
small suspended particles and enable the pre-filters to remove them.
Oar tests showed the following Nalco coagulants had the given order of
activity at 3000 ppm dosage: GWP-827 > 7132 > 107 > 8103.
Besides the Nalco cationic coagulants, the following materials were in-
vestigated with jar studies: Chitosan, Arquad 2HT75, animal glue, lime,
and acid addition with the Nalco polymers.
Based on the results of the jar studies, the Nalco polymers Were
investigated for efficacy with a 51 mm (2 in) diameter deep bed laboratory
filter. The results of these experiments was at best inconclusive. The
polymers form a gelatinous floe that is not completely removed by the deep
bed filter or else forms (or reforms) after this filter. This results in
high suspended solids test data. Results are in Table El.
In the jar tests, addition of polymer resulted in substantial color
removal when the floe settled. There was no evidence of this color removal
with the laboratory deep bed filter studies.
A full scale trial was run with the Nalco 107 cationic polymer. During
the trial, Nalco 7763 was added to aid the floe formation.
A flow schematic for the Nalco 107 polymer trial is shown in Figure El.
The operating procedure was as follows:
204
-------
ro
o
en
TABLE E-l
RESULTS OF 2" DEEP BED LABORATORY FILTER WITH CALCo POLYMERS AND CAUSTIC EXTRACT FILTRATE
Nalco
Nalco
Nalco
Nalco
Nalco
Nalco
Nalco
Nalco
Nalco
Nalco
107
107
107
6WP 827
6WP 827/677
6WP 827/677
8103
8103
8103/677
107/677
Polymer
Cone
3000 ppm
3000 ppm
1000 ppm
500 ppm
500/8 ppm
300/8 ppm
3000 ppm
1000 ppm
500/8 ppm
500/8 ppm
T.S.
6752
7248
5824
6444
6350
7376
6816
7280
7316
5436
Feed
D.S. S.S.
6656
7184
5784
6368
6172
7296
6680
7172
7232
5408
72
68
32
42
32
36
88
64
44
32
Color
19
21
17
18
16
15
28
21
19
15
,331
,298
,998
,332
,832
,998
,665
,331
,265
,832
T.S.
6840
7260
5196
6488
6216
7376
6788
7260
7244
5420
Filtrate
D.S. S.S.
6752
7104
5128
6368
6140
7308
6500
7072
7124
5412
48
36
12
100
60
36
224
152
80
12
Color
19,998
21,298
15,998
18,665
17,332
16,332
27,665
21,431
19,398
15,832
-------
5%
107
Caustic
Extraction
Filtrate
From "G"|
Line
Milton-
Roy Pump
Mixing
Drum,
(0.2 m3)
Hoffman
Vac-20 Filter
Flocculant
Feed
Tank
(1.9 n)3)
fi~^
Feed
Pump
Kisco
Filter
Broughton
Filter
Concentrate
Permeate
Figure El. Flow schematic for Nalco 107 polymer trial.
-------
A 5% solution of Nalco 107 was made in the 0.21 ra3 (55 gal)
polymer solution supply drum. The 107 polymer solution was
pumped from the drum with a Milton-Roy positive displacement
pump. The pump was calibrated and the flow setting adjusted
to supply the quantity of 5% polymer solution necessary to
give 3000 ppm polymer based on the flow of caustic extraction
filtrate to the mixing drum.
The flow of caustic extraction filtrate to the mixing drum was
adjusted by valve to match the flow of filtrate from the Hoffman
Vac-20 to the 1.9 m3 (500 gal) feed tank. The Vac-20 has a
built-in pump to pump filtrate. This flow was 21.8 m3/day (4 gpm)
during the trial. There was a stirrer on the mixing drum with a
2-inch diameter propeller rotating at 200 rpm. The level in the
mixing drum was kept between 0.11 and 0.15 m3 (30 and 40 gal)
for the trial.
The mixture of caustic extraction filtrate/polymer was gravity
fed from the mixing drum to the Hoffman Vac-20 filter. The pool
volume on the filter varied as the filter media was "blinded"
and fresh filter media indexed. An estimate of variation in
pool volume was from 0.004 to 0.02 m3 (1 to 5 gal). The supply
pipe from mixing drum to filter was 51 mm (2 in) in diameter and
6.1 m (20 ft) long. This contained less than 0.015 m3 (4 gal)
of liquid.
The filtrate from the Vac-20 was pumped to the 1.9 m3 (500 gal)
feed tank. The flow from the feed tank through the test of the
system was the normal system flow.
A flocculant, Nalco 7763, was added to the feed tank during the
last part of the trial. Once it was obvious the Nalco 107
coagulant was not solving the foul ant problem, 8 ppm of Nalco
7763 was added to the system. This level of addition was
sufficient during jar tests to aid the floe formation. The
Nalco 7763 was added as a 5% solution. An initial addition
sufficient to give 8 ppm in the 1.9 m3 (500 gal) feed tank was
made. Subsequent additions were made every thirty minutes to
maintain the 8 ppm level.
The stage 1 shell of the pilot system was used to hold the WRP #1
module for the trial since the single module test stand has no
recycle capability. Because the through-put of the Hoffman
Vac-20 filter was 21.8 m3/day (4 gpm), this was the maximum rate
of fresh feed to the UF system. The trial was run with 21.8 m3/
day (4 gpm) fresh feed and a 54.5 m3/day (10 gpm) recycle rate
for a total flow rate of 6.3 m3/day (14 gpm) through the membrane
module.
207
-------
The results of the Nalco 107 polymer trial were very discouraging.
While the Nalco 107 does cause coagulation of the suspended matter in the
caustic extraction filtrate, the floe formed has no tenacity. The least bit
of agitation causes the floe to break and pass through filters in the pre-
filtration system. The floe reforms when the stream comes to rest.
The data show extremely high suspended solids for samples taken after
the Hoffman Vac-20 and the Broughton filters. The samples when taken looked
very clear. The one-half hour time prior to testing was sufficient for
copious amounts of floe to form. Table El shows the pH, suspended solids,
total solids and color data taken during the trial.
The suspended solids data in Table El show the poor system operation.
The feed to the system was around 100 ppm suspended solids. Yet,
values from 1000-2800 ppm suspended solids Were found after polymer addition
and filtration through the Vac-20, Kisco, and Broughton filters.
Prior to the Hoffman Vac-20 filter, the system was only mildly mixed.
The mixing drum was very lightly stirred and flow from the mixing drum to the
Vac-20 was by gravity. However, the volumes involved were small so the
system was never at rest. The Kisco deep bed filters have a liquid volume
of 0.57 m3 (150 gal) above the bed. At a flow rate of 21.8 m3/day (4 gpm),
there would be an average residence time of 40 minutes while going through
the Kisco filters. This should be sufficient to allow the floe to reform
and be filtered from the system. The data for suspended solids show the
floe reforming in samples taken after the Broughton filters. Therefore,
most of the floe is making it through the Kisco filters.
The loose floe problem was discussed with the Nalco people and when told
that the Nalco 7763 flocculant was tried during the trial, could offer no
suggestions for improvement.
Based on the experience gained during the polymer trial, the only hope
for polymer addition to remove foul ants would be the use of large settling
tanks for floe removal. The ultrafiltration pilot area does not have this
capability. Also, this would not be a viable solution for a full-scale
unit.
The addition of Nalco 107 polymer had an adverse effect on flux rate
during the trials. The breaking of the floe, which then passed through
the total system, was the cause. Unless complete floe removal can be
obtained, polymer addition only compounds the flux decline problem. The
data in Table E3 show the rate of flux decline.
Figure E2 shows a comparison of the data in Table E3 versys flux decline
without Nalco 107 polymer. The loss of flux occurring faster when the
Nalco 107 was used.
208
-------
TABLE El. RESULTS OF 51 MM DIAMETER DEEP BED LABORATORY FILTER WITH NALCO POLYMERS AND CAUSTIC
EXTRACTION FILTRATE
ro
o
Polymer
concentration
Nalco
Nalco
Nalco
Nalco
Nalco
677
Nalco
677
Nalco
Nalco
Nalco
Nalco
107
107
107
GWP 827
GWP 827/
GWP 827/
8103
8103
8103/677
107/677
3000 ppm
3000 ppm
1000 ppm
500 ppm
500/8 ppm
300/8 ppm
3000 ppm
1000 ppm
500/8 ppm
500/8 ppm
T.S.
6752
7248
5824
6444
6250
7376
6816
7280
7316
5436
Feed
D.S.
6656
7184
5784
6368
6172
7296
6680
7172
7232
5408
S.S.
72
68
32
42
32
36
88
64
44
32
Color
19
21
17
18
16
15
18
21
19
15
,331
,298
,998
,332
,832
,998
,665
,331
,265
,832
T.S.
6840
7260
5196
6488
6216
7376
6788
7260
7244
5420
FiH
D.S.
6752
7104
5128
6368
6140
7308
6500
7072
7124
5412
•ra-f-p
S.S.
48
36
12
100
60
36
224
152
80
12
Color
19,998
21 ,298
15,998
18,665
17,332
16,332
17,665
21,431
19,398
15,832
-------
TABLE E-2
PHYSICAL DATA FROM NALCO 107 POLYMER TRIAL
(10/5/77 & 10/6/77)
Suspended Total
Sample
(11:30 p.m.
10/5/77)
Caustic Extract
After Vac-20
After Broughton
Permeate
Concentrate
(3:30 a.m.
10/6/77)
Caustic Extract
After Vac-20
After Broughton
Permeate
Concentrate
(10:00 a.m.
10/6/77)
Caustic Extract
After Vac-20
After Broughton
Permeate
Concentrate
(3:00 p.m.
10/6/77)
Caustic Extract
After Vac-20
After Broughton
Permeate
Concentrate
pH
10.0
8.6
8.5
8.6
8.5
11.0
9.1
8.9
9.0
9.0
_ _ _
7.5
8.8
9.5
8.9
10.5
9.2
10.4
8.4
8.2
Solids
(PPm)
64
2768
1124
4
1068
96
2828
1160
8
1024
116
2164
1000
36
972
___
916
1204
—
928
Solids
(ppm)
6810
8066
7622
5470
2686
6734
7710
7176
5298
7186
6938
8090
7370
6100
7322
8142
9464
8700
6206
8562
Color
17,332
5.999
6,632
1,733
66,332
17,498
5,999
6,499
1,667
6,366
19,998
3,000
6,499
4,700
6,666
17,332
8,832
8,333
3,066
7,833
210
-------
TABLE E-3
FLUX DECLINE VERSUS TIME FOR NALCO 107 TRIAL
Elapsed Running Time (Hrs.) Flux (gfd) Percent Flux Loss
0 38.5 0
1 25.2 34.5
2.7 14.7 61.8
3.3 11.4 70.3
4.1 8.6 77.7
5.1 7.6 80.3
6.1 6.2 83.9
7.1 5.2 86.5
8.1 4.8 87.5
9.1 3.8 90.1
211
-------
LU
z
1— I
_J
o
LU
o
X
13
_J
1 1
^
ro §
IN> QJ
a.
" 1
10
20
30
40
50
60
70
80
90
100
C
i ! T 1 — r~ 1 1 r~ — r ~~ ' i r "
\
\
VA 0 o WRP #2 without polymer (5/9/77 data)
i\
^D D a WRP #2 Nalco 107 polymer trial
"Nr^^^o-o^
\ °"°"^^-_
~~ ° -
""• — ^Q., rj
1 1 1 1 1 1 ! 1 1 ! 1
) 2 4 6 8 10 12 14 16 18 20 22
CUMULATIVE OPERATING TIME, HOURS
Figure E2 . WRP #2 module flux decline with and without Nalco 107 polymer pretreatment.
-------
Table E4 contains additional data from the trial. After 20 hours of
running, five hours after the Nalco 7763 flocculant addition was started, the
membrane module was given a one-hour cleaning with ultraclean solution. The
hoped for flux recovery was not obtained with this quick cleaning. The flux
of 0.29 m3/m2-day (7.1 gfd) rapidly decreased back to the 0.08 m-W-day
(1.9 gfd) value of before cleaning.
51 MM DIAMETER DEPTH FILTER PRETREATMENT STUDIES
A 51 mm (2 in) diameter depth filter was operated at the Canton Mill
to assess the performance of various filter media. Representative data from
runs with beds of garnet sand, anthracite coal and silica sand, filter AG and
granular PVC are given in Tables E5 through E8. No media was significantly
superior to the filter AG (used in the Kisco filters) in terms of suspended
solids removal.
All media tested in the 51 mm diameter column exhibited substantial
head loss over relatively short periods of time (1 to 6 hours). This
occurred even though the caustic extraction filtrate feed stream was pre-
treated by a hydrasieve for fiber removal.
Based on these tests no change in the Kisco filter media was made.
213
-------
TABLE E-4
OPERATING DATA FROM NALCO 107 TRIAL
Time
10/5 11:25 am
12:25 pm
2:05
2:45
3:30
4:30
5:30
6:30
7:30
8:30
9:30
10:25
11:30
10/6 12:30 am
1:30
2:30
3:30
4:30
5:30
6:30
7:30
8:30
10:00
11:00
12:00 pm
12:55
2:00
3:00
4:00
5:00
6:00
6:45
Elapsed Pressure Total
Running Flux Drop Across Flow
Time (Hrs.) (gfd) Module (psig) Rate (gpm)
0
1
2.7
3.3
4.1
5.1
6.1
7.1
8.1
9.1
10.1
11.0
12.1
13.1
14.1
15.1
16.1
17.1
18.1
19.1
20.1
System shut
0
1
2
2.9
4
5
6
7
8
8.75
38.5
25.2
14.7
11.4
8.6
7.6
6.2
5.2
4.8
3.8
3.8
3.3
2.9
2.7
2.5
2.4
1.9
1.9
2.1
1.9
1.9
down for cleaning with
7.1
4.8
3.8
3.0
2.9
2.3
2.1
2.1
1.9
1.9
32
41
21
15
32
32
34
34
34
34
34
35
34
34
34
34
34
34
35
35
35
ulatraclean solution
40
42
42
42
40
40
40
40
40
39
12
15
10
8
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
13
15
15
15
14
14
14
14
14
14
214
-------
TABLE E5. REPRESENTATIVE OPERATING DATA FOR 51 MM DIAMETER DEPTH
FILTER WITH 0.75 M BED DEPTH OF GARNET SAND*
Date
5/3
5/4
5/5
5/6
5/9
5/10
Cumulative
operating
time, hrs
0.7
3.5
4.5
5.5
0.5
2.5
4.5
1.5
2.5
4.5
1.5
2.5
3.5
6.5
1.5
2.5
4.5
5.5
1.0
2.5
- • • • — ....
Feed
suspended
solids, mg/1
48
48
44
40
64
40
30
46
40
38
40
38
38
34
50
48
34
24
54
58
11
Filtrate
suspended
solids, mg/1
26
18
20
18
18
14
8
22
18
18
22
22
22
22
28
30
26
80
52
30
1 ' i
Removal
efficiency,
%
45.8
62.5
54.5
55.0
71.9
65.0
73.3
52.2
55.0
52.6
45.0
42.1
42.1
35.3
44.0
37.5
35.3
—
4.7
31.0
Feed stream is caustic extraction filtrate pretreated by
a hydrasieve.
215
-------
TABLE E6. REPRESENTATIVE OPERATING DATA FOR 51 MM DIAMETER DEEP BED FILTER
WITH 0.51M ANTHRACITE COAL PACKED OVER 0.25M SILICA SAND*
Date
2/16
2/18
2/22
2/24
2/28
3/2
Cumulative
operati ng
time, hrs
0.33
0.08
0.75
0.33
1.08
2.13
3.08
4.08
5.08
5.88
6.88
7.88
8.88
9.88
0.5
1.5
2.5
3.5
4.5
.33
1.33
2.33
3.33
4.33
5.33
5.83
.67
2.67
2.75
3.67
4.67
Feed
suspended
solids, mg/1
48
76
36
72
78
84
56
42
64
116
104
56
84
84
82
60
52
70
66
106
98
100
106
98
100
80
52
58
60
44
50
Filtrate
suspended
solids, mg/1
20
18
14
36
42
36
28
12
18
52
48
16
32
38
38
30
26
66
52
34
40
30
62
52
48
44
38
28
30
14
28
Removal
efficiency,
%
58.3
76.3
61.1
50.0
46.2
57.1
50.0
71.4
71.9
55.2
53.8
71.4
61.9
54.8
53.7
50.0
50.0
5.7
21.2
67.9
59.2
70.0
41.5
46.9
52.0
45.0
26.9
51.7
50.0
68.2
44.0
Feed stream is caustic extraction filtrate pretreated by a hydrasieve.
216
-------
TABLE E7. REPRESENTATIVE OPERATING DATA FOR 51 MM DIAMETER DEEP BED FILTER
WITH 0.75 M BED DEPTH OF FILTER AG*
Date
3/7
3/8
3/10
3/14
3/16
3/24
Cumulative
operating
time, hrs
2.33
3.33
4.33
.25
1.25
2.25
3.25
.35
1.42
2.42
3.42
.58
1.58
2.58
3.58
4.58
5.58
.75
1.75
2.75
3.75
4.75
.58
1.56
2.58
3.58
4.58
5.58
6.08
7.03
7.53
9.03
10.03
12.03
Feed
suspended
solids, mg/1
46
64
86
66
64
72
80
90
82
84
84
40
42
40
48
66
42
68
64
52
36
44
80
76
74
64
42
64
64
180
70
62
54
76
Filtrate
suspended
solids, mg/1
22
22
44
18
14
40
40
42
46
50
48
32
28
12
16
16
20
36
16
4
4
4
44
30
32
16
12
20
22
36
26
20
22
26
Removal
efficiency,
%
52.2
65.6
48.8
72.7
78.1
44.4
50.0
53.3
43.9
40.5
42.9
20.0
33.3
70.0
66.7
75.8
52.4
47.1
75.0
92.3
88.9
90.9
45.0
60.5
57.8
75.0
71.4
68.8
65.6
80.0
62.9
67.7
59.3
65.8
" '
217
-------
TABLE E8. REPRESENTATIVE OPERATING DATA FOR 51 MM DIAMETER DEPTH FILTER
WITH 0.75 M BED DEPTH OF GRANULAR PVC*
Date
4/1
4/4
4/6
Cumulative
operating
time, hrs
1.11
1.67
2.17
2.67
3.17
3.67
4.17
4.67
5.67
6.67
7.67
8.67
9.67
11.17
12.17
13.17
14.67
15.67
17.58
18.58
20.08
21.08
22.08
23.91
25.41
26.41
Feed
suspended
solids, mg/1
74
82
72
70
68
102
80
64
62
74
60
52
60
96
118
96
80
40
62
56
52
48
64
76
76
72
Filtrate
suspended
solids, mg/1
24
36
40
64
42
50
32
20
24
36
40
28
24
40
46
64
40
8
4
2
12
2
8
16
18
20
Removal
efficiency,
%
67.6
56.1
44.4
14.3
38.2
51.0
60.0
68.8
61.3
51.4
33.3
46.2
60.0
58.3
61.0
33.3
50.0
80.0
93.5
96.4
76.9
95.8
87.5
78.9
76.3
72.2
*
Feed stream is caustic extraction filtrate pretreated by a hydrasieve.
218
-------
LETTER FROM J.A. NOWAK, NALCO
DATE: JULY 14, 1977
SUBJECT: PLUGGED CHAMPION PAPERS
The plugging of Champion Paper ultra-filtration membranes can be
attributed to particles of kaolinite and titanium dioxide which are coated
and agglomerated with starch. The agglomerated particles also contain a
carboxylic acid salt which may stabilize the colloidal nature of the
foul ant.
Scanning electron micrographs of a new membrane and a plugged membrane
are shown in Figures E3 and E4 respectively. The micrograph of the new
membrane shows that the membrane pore size is very small (less than 0.1
microns) and that the surface of the plugged membrane is completely
covered with the "slime" foulant. Figures E5 and E6 are microprobe results
which indicate the chemical composition of the "slime".
Transmission electron micrographs of the "slime" sample are shown in
Figures E7, E8 and E9. The large plate-like particles in Figure E7 are
identified as kaolinite clay, while the small dense particles are titanuim
dioxide. The micrograph also indicates the presence of a coating
surrounding the particles and holding them together into larger chumps.
Figures E8 and E9 further illustrate the degree to which the particles are
coated and agglomerated.
X-ray diffraction and X-ray fluorescence analysis of the dried slime
revealed it to contain kaolinite (Al2Si205(OH)4) and anatase (TiOz).
Infrared spectroscopy indicated that the major organic constituent was
starch. The presence of starch was confirmed by chemical spot tests. A
carboxylic acid salt was also found to be present at low concentrations.
The composition of the slime is best represented as follows:
Volatiles at 105°C - 88.0%
Kaolinite Clay - 5.6%
Starch - 4.0%
Titanium dioxide - 1-4%
Carboxylic acid salt - <1.0%
The combination of the small pore size of the membrane, together with
the starch coated and colloidal nature of the particles, would seriously
impair the performance of the ultra-filtration membranes.
Please contact me if you require additional information or work.
219
-------
Figure E3. New membrane 5000X.
Figure E4. Fouled membrane 5000X,
220
-------
198SEC 61838INT
VS:258e HS: 26EV/CH
•
1237iEDflX
Figure E5. Elemental analysis of new membrane.
t iiiSEC 72S89IMT
VS:25i§ MS: 28EV/CH
9t€?24 12373EDAX
Figure E6. Elemental analysis of fouled membrane,
221
-------
I
'
-------
•
-------
Figure £9.
TEM OF »$UME:»
32,000 X (3,090^/cw)
-------
TABLE FT. FEED, RECYCLE, REJECT AND PERMEATE COLOR CONCENTRATIONS DURING 3-STAGE
PILOT SYSTEM TESTS (COLOR UNITS)
-O
-a
ro
ro
en
Cumulative
operating
time (hours)
2
5.5
23
30
48
A9
69
75
83
90
97
120
139
141
142
148
167*
175*
177
197
198
206
217
219
232
235
239
246
249
255
275
282
300
320
325
335
Feed
(after
brouqhton)
23,331
18,665
17,332
20,731
17,299
18,665
17,332
14,999
17,332
17,998
16,665
18,332
17,998
15,065
18,665
3,400
3,166
14,979
19,998
14,832
--
16,332
13,499
25,331
24,331
22,331
14,965
20,665
20,998
25,997
16,336
21,998
16,332
26,997
17.932
Recycle
27,331
25,666
19,998
22,998
25,664
21,331
24,664
21,998
21,331
21,331
23,166
18,665
19,765
20,831
17,965
22,664
4,300
3,666
18,065
25,331
20,998
33,997
24,998
19,665
33,330
26,664
28,997
18,498
26,664
27,997
32.997
20,498
25,331
16,998
39,663
25.997
Stage 1
Reject
_.
—
-.
.-
-.
--
__
._
--
—
--
—
—
—
—
—
—
—
—
—
—
22,331
39,996
27,331
29,664
16,665
29,997
28,664
36,663
20,998
25,664
17,665
42,652
28,331
Permeate
3,966
4,900
3,100
2,633
6,166
3,666
5,533
5,133
6,399
5,999
5,000
5,000
4,500
5,299
6,066
7,666
1,033
1,966
7,099
13,665
3,500
4,500
3,033
2,666
8,333
9,066
5,966
4,366
7,433
7,499
9,499
4,500
8,166
10,832
9,666
5,966
Recycle
32,997
26,997
21 ,998
27,331
29,660
31 ,630
36,663
30,997
47,329
33,663
49,995
30,997
28,331
39,996
53,328
89,658
12,332
15,998
64,994
84,325
34,996
53,328
54,995
41,663
108,323
59,327
64,994
48,329
59,994
96,657
61 ,661
39,996
36,663
40,329
83,325
59,661
Stage 2
Reject
—
—
—
—
—
—
—
44,996
39,663
52,995
30,997
29,997
43,996
56,661
91 ,658
12,665
16,665
69,660
89,991
—
—
—
44,996
114,989
64,660
70,160
49,995
66,993
108,323
64,994
44,996
42,662
42,996
84,325
60,661
Permeate
4,666
4,433
1,633
3,400
4,300
3,400
5,333
5,000
5,666
20,998
6,333
3,666
2,400
3,666
5,666
7,966
467
833
5,633
8,632
5,000
8,366
7,266
4,600
13,332
8,166
8,832
5,099
8,166
11,832
8,666
6,633
12,299
7,666
15,498
10,999
Recycle
59,327
53,328
25,997
173,983
74,993
59,661
100,657
74,659
_„
—
—
—
—
133,320
227,311
206,646
154,985
—
219,978
266,640
219,978
219,978
109,989
198,314
141,653
193,314
146,319
316,635
273,306
Stage 3
Reject
70,993
59,661
27,330
189,981
74,973
40,663
106,323
79,654
._
._
—
—
—
—
__
166,317
256,641
256,641
158,318
996,567
222,978
283,305
236,664
226,644
139,986
216,645
143,319
209,979
161,651
329,967
293,304
Permeate
6,299
4,766
1,333
9,699
4,666
4,666
9,732
8,732
__
--
-_.
--
--
--
_.
__
25,331
30,664
22,664
11,832
59,994
20,665
14,999
16,665
12,832
10,999
11 ,166
5,066
39,663
30,664
- co
3
J»
en
-a
t— i
MV
~
o
CO
_^
CO
-H
m
M^i
3»
O
O
r~
•yo
70
rn
r .
m
o
— 1
1^*4
O
^•^1
§
3=»
-------
TABLE F2. WRP SPIRAL-WOUND MODULE COLOR REJECTION DURING 3 STAGE PILOT PLANT TESTS
INS
INS
Cumulative
operating
time (hours)
2
5.5
23
30
48
49
69
75
83
90
97
120
139
141
142
148
167*
175*
177
197
198
206
217
219
232
235
239
246
249
Stage
%
rejection
feed
basis
83.0
--
83.4
84.8
70.3
78.8
70.4
70.4
57.3
65.4
72.2
70.0
75.5
71.1
59.7
58.9
69.6
37.9
52.6
31.7
76.4
—
81.4
80.3
67.1
62.7
73.3
70.8
64.0
1
%
rejection
concentrate
basis
85.5
80.9
84.5
88.6
76
83.8
76.0
76.7
70.0
71.9
78.4
73.2
77.2
74.6
66.2
66.2
76.0
46.4
61.3
46.1
83.3
87.9
87.9
86.4
72.0
66.0
79.4
76.4
72.1
Stage
%
rejection
feed
basis
80.0
—
91.3
80.4
79.3
80.3
71.4
71.1
62.2
0
64.8
78.0
86.9
80.0
62.4
57.3
86.3
73.7
62.4
56.8
66.3
—
55.5
65.9
47.4
66.4
60.4
65.9
60.4
2
%
rejection
concentrate
basis
85.9
83.6
92.5
87.6
85.5
89.3
85.5
83.9
88.0
37.6
87.3
88.2
83.7
91.6
89.4
91.1
96.2
94.8
91.3
89.8
85.7
84.3
86.8
89.0
87.7
86.2
86.4
89.4
86.4
Stage
%
rejection
feed
basis
73.0
—
92.9
44.0
77.5
73.0
47.9
49.6
--
—
--
--
--
--
--
--
—
--
--
—
0
—
0
12.3
0
—
7.5
0
19.4
3
%
rejection
concentrate
basis
89.4
91.0
94.9
94.4
93.8
92.2
90.3
88.3
—
—
—
—
—
--
--
--
—
—
—
—
81.0
86.5
89.0
92.4
--
—
92.2
93.2
92.4
(continued)
-------
TABLE F2 (continued)
ro
ro
Cumul ati ve
operating
time (hours)
255
275
282
300
320
325
335
Stage
%
rejection
feed
basis
64.3
63.5
72.4
62.9
33.7
64.2
66.7
1
%
rejection
concentrate
basis
73.2
71.2
78.0
67.8
36.3
75.6
77.1
Stage
%
rejection
feed
basis
43.7
66.7
59.4
44.1
53.1
42.6
38.7
2
%
rejection
concentrate
basis
87.8
85.9
83.4
66.5
81.0
81.4
81 .6
Stage
3
%
-------
APPENDIX 6. ANALYTICAL DATA FROM 25.4 MM DIAMETER TUBULAR
ASSEMBLY EXPERIMENTS
1 ', 3ft. ','*/ - -v
.:•:•..•*..• ,:•.•: ' •••.^..&&.&'^
£><>£*§
&$sf^
V$,\ P. O. BOX 4187, 2323 SYCAMORE DR. KNOXVILLE. TENNESSEE 37921 / 615 546-1335
•:- CERTIFICATE OF ANALYSIS
,' PeeTScBwe and
Itchnohgj
B»-
Mr. H. A. Fremont
Champion International Corporation
Kinghtsbridge
Hamilton, Ohio 45020
APR 2 * 1978
-ApTil 20, 1978
Received: April 6th
m
Dear Mr. Fremont:
Analysis of your water samples gave the following results:
Your#, Ourl, Analysis,
AB
WB-5031
PI & 2 1.2X WB-5032
PI & 2 10X WB-5033
Total Solids 0.717 %
Total Volatile Solids 0.190 %
Sulfate (as S) 26 ppm
Chloride 1542 ppm
Aluminum 3 ppm
Calcium 36 ppm
Iron 2.0 ppm
Sodium 1790 p'pm
Ash 0.53 %
Chlorine 2043 ppm
Specific Gravity 1.001
Total Solids
Total Volatile
Solids
Sulfate (as S)
Chloride
Aluminum
Calcium
Iron
Sodium
Ash
Chlorine
Specific Gravity
Total Solids
Total Volatile
Solids
Sulfate (as S)
Chloride
Aluminum
0.425 %
0.059 %
15 ppm
1510 ppm
less than
2 ppm
0.4 ppm
1360 ppm
0.37 %
1722 ppra
0.999
0.441 %
0.078 %
18 ppm
1457 ppm
less than
1 ppm
1 ppm
ISl
V i.
V.:; V;-;V.:.fe:'V'- ^^UJ&lVv.tf... #i. V;:'->!:'')r VV^VvV.jX/; tlii^';
^&££?^V>^^^
228
-------
^Wr--'^'«^1^^ ...
P.O. BOX 4187, 2323 SYCAMORE DR., KNOXVILL6, TENNESSEE 37921 / 615 S46-1335
CERTIFICATE OP ANALYSIS
Mr. H.A. Fremont
Page 2
April 20, 1978
Your#, Our#,
PI & 2 10X WB-5033
PI & 2 SOX WB-5034
Concentrate WB-5035
1.2 X
Concentrate WB-5036
10X
Analysis,
Calcium
Iron
Sodium
Ash
Chlorine
3 ppm
0.40 ppm
960 ppm
0.36 %
1853 ppm
Specific Gravity 1.000
Total Solids
Total Volatile
Solids
Sulfate (as S}
Chloride
Aluminum
Calcium
Iron
Sodium
Ash
Chlorine
Specific Gravity 1.000
0.867 %
0.383 %
15 ppm
1777 ppm
less than 1 ppm
12 ppm
0.51 ppm
1790 ppm
0.48 %
2335 ppm
Total Solids
Total Volatile
Solids
Sulfate (as S)
Chloride
Aluminum
Calcium
Iron
Sodium
Ash
Chlorine
Specific Gravity
Total Solids
Total Volatile
Solids
Sulfate (as S)
Chloride
0.782 %
0.304 %
32 ppm
J558 ppm
3 ppm
45 ppm
3.0 ppm
1840 ppm
0.48 %
2358 ppm
1.001
1.91 %
1.17 *
13 ppm
1370 ppm
229
-------
SgBf
f§[^
i§S
V ." ~>T '-^=.'i
>*<>,•.
5'->: -^ri>.
\^A
u-'
-------
REPORT NO.
EPA-600/2-80-045
4. TITLE AND SUBTITLE
Color Removal from Kraft Mill Effluents
by Ultrafiltration
TECHNICAL REPORT DATA
If lease read Instructions on the reverse before completing)
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
February 1980 Issuing date
. PERFORMING ORGANIZATION CODE
j.A. Fremont, D.J. Strlley, Champion International
l.H. Kleper. R.L. Goldsmith. Maiden Division of Abcor.Inc
8. PERFORMING ORGANIZATION REPORT NO
ID ADDRESS
hampion International
Knightsbridge
lamilton, Ohio 45020
10. PROGRAM ELEMENT NO.
1BB610
1. CONTRACT/GRANT NO.
S804312-01
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Laboratory
Jffice of Research and Development
J.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final; 3/?q/7fi - Q/?Q/7«
14. SPONSORING'AGENCY
EPA/600/12
15. SUPPLEMENTARY NOTES
16.
Color removal from kraft mill effluents by ultrafiltration (UF) has been examined
during this program. A 3-stage, nominal 37.9 m3 (10,000 gpd) UF pilot plant was oper-
ated on caustic extraction filtrate for several months. Extensive evaluation of spiral-
wound UF modules was carried out prior to staged system operation in single module
tests. During these tests feed pretreatment and prefiltration options were investigatec
and the effects of a range of operating parameters on module flux performance were
studied. A second module configuration, tubular assemblies, was also tested. All fielc
tests were performed at the Canton, North Carolina Mill of Champion International.
Non-eellulosic UF membranes were evaluated in laboratory tests before field trials
were initiated. The preferred membrane was cast from a polysulfone formulation.
Spiral modules were severely fouled by species present in white water recycle.
Tubular modules, however, exhibited stable, economically-viable flux performance.
Color removal by the tubular UF membranes ranged from 97% to 99% when calculated
on a concentrate basis. Projections based on process data indicate UF results in an
overall color reduction of 91% (mass basis) for caustic extraction filtrate.
Conceptual designs and economic analyses were developed for treatment systems with
capacities of 3,790 m3/day (1 MM gpd) and 7,980 m3/day (2 MM gpd). Additionally,
caustic extraction filtrate and decker effluent stream characteristics were monitored
and qualitative assessments of ultrafiltrate and UF concentrate recycle within a kraft
rill
de
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
COS AT I Field/Group
Caustic extraction filtrate
Color bodies
-olor removal Membrane processes
Decker effluents Ultrafiltration
Kraft mill
Wastewater treatment
Water pollution control
13B
. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (',
Unclassified
247
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE.^
U.S. BOVEBHMBIt WIHTING OffKt 1*0 -657-146/5595
------- |