WATER POLLUTION CONTROL RESEARCH SERIES • 12040 EEL 02/72
REVERSE OSMOSIS CONCENTRATION
OF DILUTE PULP
and
PAPER EFFLUENTS
M.S. ENVIRONMENTAL PROTKC T1ON AC.KNCY
-------
ERRATA
REVERSE OSMOSIS CONCENTRATION OF DILUTE PULP & PAPER EFFLUENTS 12040 EEL 02/72
Change table on Page 335 as shown below;
CM wash water 58.7 Tp_; 16.5 pounds per 1000 gallons
CM wash water FROH; 2.64 TO: 0.74 $/1000 gallons
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes the
results and progress in the control and abatement of pollution
in our Nation's waters. They provide a central source of
information on the research/ development, and demonstration
activities in the water research program of the Environmental
Protection Agency, through in-house research and grants and
contracts with Federal, state, and local agencies, research
institutions, and industrial organizations.
Inquiries pertaining to Water Pollution Control Research Reports
should be directed to the Chief, Publications Branch (Water),
Research Information Division, R&M, Environmental Protection
Agency, Washington, D. C. 20460
-------
REVERSE OSMOSIS CONCENTRATION OF DILUTE PULP & PAPER EFFLUENTS
Averill J. Wiley, George A. Dubey and I. K. Bansal
Ihe Pulp Manufacturers Research League
and
The Institute of Paper Chemistry
10U3 E. South River Street - P.O. Box
Appleton, Wisconsin
for the
OFFICE OF RESEAECH AND MOHITORIIG
ENVIROIMENTAL PROTECTION AGIICY
Project #1201*0 EEL
February, 1972
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C, 5B402 - Price $2.7«
-------
EPA REVIEW NOTICE
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents neces-
sarily reflect the views and policies of the
Environmental Protection Agency, nor does mention
of trade names or commercial products constitute
endorsement or recommendation for use.
ii
-------
.ABSTRACT
Adaptation of reverse osmosis as a method of concentration for dilute
effluents of pulping, bleaching, and paper manufacture was conducted
in laboratory, pilot scale, and in large 50,000 gallon per day field
demonstrations at pulp mills. Most of these dilute wastes at 1 percent
solids contained suspended particles, colloidal suspensoids, large
molecular-weight wood derived organics, and/or scale-forming inorganic
chemical residues. Tubular membrane systems capable of being operated
at self-cleaning velocities increasing beyond 2,0 feet per second,
as concentration advanced to 10 percent solids, were apparently best
adapted to processing these effluents at sustained high flux rates
and relatively free of fouling problems. Capillary fiber and spiral
wound sheet membrane systems required expensive clarification treat-
ment before and during concentration. Tubular systems studied were
subject to excessive failure rates in terms of life of membrane support
structures or to leakage of internal connections based on the support
structure. Feasibility of employing RO for concentration of dilute
pulping and bleaching effluents depends on developing routes to sub-
stantial improvement in life expectancy of RO equipment to maintain
high flux rates and rejections at much lower membrane maintenance and
replacement costs than prevailed with equipment available for these
studies conducted from 1967 through 1971- This report was submitted
in fulfillment of Project Number 120^0 EEL, Contract WPRD 02-01-68,
under (partial) sponsorship of the Office of Research and Monitoring,
Environmental Protection Agency.
lit
-------
CONTENTS
Section
I
II
III
I?
V
YI
VII
VIII
IX
X
XI
Conclusions
Recommendations
Introduction — Application of Reverse Osmosis to
Processing Pulp and Paper Mill Effluents
Principles of Reverse Osmosis as Applied to
Concentration Processing of Pulping Effluents
Laboratory Studies on the Reverse Osmosis
Concentration of Effluents from the Pulp and
Paper Industries
Design of Trailer
Five Field Demonstrations
- Ca Base Pulp Wash and Cooling Waters
- NSSC White Water
- NH3 Base Pulp Wash Water
also Ca Hypochlorite Bleach Effluent
- Kraft Bleach Effluent
also Kraft Rewash Water (Preliminary)
- Chemimechanical Pulping Wash Water
Engineering and Development Studies
- Development of Engineering Design Factors
- Pilot Studies on Membrane Fouling and
Concentration Polarization
- Computer!zed Mathematical Model
Membrane Module Life Studies
Process Economics
Acknowle dgments
1
3
5
9
15
63
77
77
111
137
166
202
231
231
252
260
297
331
339
v
-------
CONTENTS (2)
Section Page
XII References 3)43
XIII last of Publications
XIV Glossary of Terms Abbreviations and Symbols
XV Appendices - Additional Photos 355
XVI WRIC Abstract Form 359
vl
-------
FIGURES
No. Pages
1 Schematic Diagram of Reverse Osmosis 10
2a Photograph of Milton-Roy Duplex Pump Test Stand 17
2b Photograph of Pilot Unit 17
3 Evidence for Membrane Compaction 3k
h Flux Rates from Coated (Fouled) Membranes 39
5 Equipment for RO Processing with a Single Module UO
6 Attempts to Remove Foulant with Adsorbants — Effect on
Flux Loss U3
7 Effect of Sudden Reductions in Operating Pressure on
Flux Loss UU
8 Effect of Pressure Pulsing at Various Points in RO
System on Flux Loss ^6
9 Effect of Pressure Pulsing Cycle Rate on Flux Loss ^7
10 Processing of Coating Effluents by Reverse Osmosis 55
U Effect of pH on Acetate Ion Rejection 6l
12 Photo of Trailer Mounted RO Field Demonstration Unit 72
13 Photograph of RO Trailer — Rear Bank 73
lU Photograph of Interior of Trailer 7^
15 Flow Diagram of the Large-Scale Reverse Osmosis Unit 75
16 Flow Sheet for Calcium-Base Acid Sulfite Pulp Mill 79
17 Flow Sheet — Washing Cycle and Liquor Collection —
Ca-Base Wash Water 80
18 Flow Sheet - Small Pilot RO Unit for Ca-Base Wash Waters 83
19 Small Pilot Flux Rate History - Ca-Base Wash Waters 86-88
20 . Flow Sheet - Pretreatment and Trailer Operation Ca-Base
Wash Waters 96
vii
-------
FIGURES .(2)
No. Pages
21. Solids Concentration vs_ Specific Gravity Field Run on
Ca^-Base Wash Waters 101
22. System for NSSC Idnerboard Mill with RO Water 112
23. Pretreatment and RO Operations for Processing NSSC
White Water with Small Pilot Unit 115
2k Daily Flux Rates and Pressures for Small Pilot Unit 118-119
25 Variation in Flux Rates with Increase in Concentration —
NSSC White Water 120
26 Schematic of Reverse Osmosis Pretreatment and Trailer
Equipment 123
27 Flux (gfd) vs Operating Hours at Green Bay Packaging 13^
28 Flow System for NHs Pulping and Bleaching with Effluent
Collection Points for RO 139
29 Flux Rate vs Operating Hours — Ammonia-Base Acid Sulfite
Liquor
30 Kraft Mill Pulp Washing and Bleaching Schematic, iTQ
31 Flow Sheet — RO Processing of Alka.line EKbrac^iQn KBE 176
32 Decrease in Flux Rate with Increase in Solids
Concentration 188
33 Comparison of Flux Rates with and Without Pulsing 190
3U Osmotic Pressure of Alkaline Extraction KBE 192
35 Total Solids as Function of Specific Gravity of
Alkaline Extraction KBE 193
36 Flux Rate History While Processing Kraft Rewash Water 200.
37 Chemimechanical Pulping Flow Sheet 205
38 Schematic of Reverse Osmosis at Appleton Papers Inc. 209
39 Flux Rate History While Processing CM Wash Water 212
1*0 Osmotic Pressure of CM Pressate and Digester Liquors 219
viii
-------
FIGURES ,(3)
No» Pages
!*1 Specific Gravity — Concentration Relationship for
Press Liquors 220
42 Flux Rates During Concentration Runs with Fresh CM
Press liquors 221
^3 Aging Effect of Press liquors on Flux Rates 22U
UU. Reynolds Number of NSSC White Water 233
U5 Reynolds Number of Ammonia-Base Acid Sulfite Liquor 233
^6 Reynolds Number of Calcium-Base Acid Sulfite Liquor
1*7 Reynolds Number of Kraft Bleach Effluent
48 Frictional Pressure Drop in One 18-Tube Havens Module
for 10 Percent Solids Liquor 236
49 Pumping Energy in One 18-Tube Havens Module for 10.0
Percent Solids Liquor 237
50 Osmotic Pressure vs Concentration of the Liquor 240
51 Effect of Temperature on Flux Rates of Water and
Calcium-Base Acid Sulfite Liquor 242
52 Schematic Diagram of Experimental Set-Up 244
53 Permeation Resistance ys_ Reynolds Number-NSSC White Water 245
54 Effect of Velocity on the Flux Rates of Ca-Base Liquors
(No Precipitate) 247
55 Effect of Velocity on the Flux Rates of Ca-Base Liquor
(Precipitate) 2 1*8
56 Membrane Constant vs. Average Pressure (RO Studies) 251
57 Effect of Velocity on Flux Rate Declines Due to
Concentration Polarization and Fouling Factors for
Calcium-Base Acid Sulfite Liquor (RO Studies) 257
58 Effect of Velocity on Flux Rate Declines Due to Concen-
tration Polarization and Fouling Factors for "Aamonia-
Base Acid Sulfite Liquor (RO Studies) 258
ix
-------
FIGURES
No.
59 Velocity vs_Permeation Resistance — Calcium-Base.Acid
.Sulfite, NSSC and Kraft Bleach Effluent liquors 262
60 Flow Chart of Reverse Osmosis Optimization Studies 265-268
6l Schematic Diagram of Experimental Set-Up 279
62A Flux Rates During Concentration Runs of NSSC Liquor
(I Module Configuration) 285
62B. Flux Rates During Concentration Run of NSSC Liquor
(II Module Configuration) 286
63 Effect of Number of Modules in Series on the Capital
and Operating Costs of a Reverse Osmosis Unit
Processing 750,000 Gallons per Day of NSSC Liquor 29^
6k Pretreatment Section Adjacent to Trailer RO Field Unit,
Appendix A 356
65 Pilot RO Unit Operating on Caustic Kraft Bleach
Effluent, Northwest Paper Company, Cloquet, Minn. 357
66 Pilot RO Unit for Recovery and Processing of Chemi-
mechanical Pulp Wash Water at Locks Division,
Appleton Papers Inc., Combined Locks , Wis. 358
-------
TABLES
Ncu Pages
1 Effect of Membrane "Tightness" on Product Water Quality 21
2 Concentration with Tubular Units with. Three Different
Membrane "Degrees of Tightness" 22
3 Concentration with Tubular Units Containing Type 2A
(Open) and Type 3 (Intermediate) Membranes 23
k Processing Clarified and Unclarified Kraft Bleach Plant
Effluents Through Tubular Modules 25
5 Percent Rejection for Typical Pulp Wash, and Bleach
Effluents 29
6 Percent Rejection During the RO Concentration of
Condensates; from the Evaporation of Spent Sulfite liquors 29
7 RO Processing at Different Pressure Levels with Spiral
Wound Modules 33
8 BOD5 Rejection pn l8-Tube .Module Life Study - Effect of
Sampling Time on Rejection Value 37
9 Product Flux Rates vs Time Hi
10 Effect of Various Pretre fitments — RO Processing of a
Kraft Pulp Wash Water 1*2
11 Study of Membrane Fouling with a Single-Tube Havens
Module — Using Additives to Prevent Fouling 1*9
12 Study of Membrane Fouling with a Single-Tube Havens
Module — Cleaning Procedures After Fouling 50
13 Processing of Hypochlorite Bleach Plant Effluent by,
Reverse Osmosis with Pressure Pulsing 51
1^ RO Concentration of a Hypochlorite Bleach Plant Effluent
with Pressure Pulsing 52
15 Reverse Osmosis Processing of Barking Water 53
16 Processing of De-inking.Wastes by Reverse Osmosis 53
17 Reverse'Osiiosis Processing of Evaporator Condensates 56
xi
-------
TABLES (2)
'No. Pages
18 Reverse Osmosis Processing of Acid Sulfite Evaporator
Condensates at Two Pressure Levels 57
19 Reverse Osmosis Processing of Magnefite Evaporator
Condensates 58
20 Beverse- Osmosis Concentration of an Acid Sulfite
Evaporator Con dens at e. with pH Adjustment of the' Feed 59
21. Rejection of Acetate Ions at Various pH Levels 60
22 Reverse Osmosis Processing of a Calcium-Base Acid Sulfite
Evaporator Condensate After pH Adjustment 62
23 Performance of Reverse Osmosis at Various Rejection Ratios 69
2k, Flux Rate Summary — Small Pilot Bun on Ca-Base Wash Water 85
25 Average Rejection Ratios — Small Pilot Run on Ca-Base
Wash Waters 89
26 Analytical Data for Small Pilot Run Ca-Base Washjfeters 90-93
27 Module Bank Configuration — Trailer Unit Ca-Base Wash
Waters 95
28 Trailer Data Averages Ca-Base Wash Waters 97
29 Detailed Summary of Trailer Operations Ca-Base Water 98
30 Recovery Data for Trailer Unit — Ca-Base Wash Waters 103
31 Analytical Data for Trailer Unit — Ca-Base Wash Waters lQlf-106
32 Summary of Downtime — First Field Demonstration 108
33 Estimated Pollution Loads from Various Effluent Flows 110
3k Analysis of a Typical NSSC White Water Feed 113
35 Flux Rate Data — NSSC White Water with the Small
Pilot Unit 117
36 Analytical Data and Rejections — Processing NSSC White
Water with Pilot Unit 121-122
xii
-------
TABLES (3)
No.. Pages
37 Summary of Downtime for Second Field Demonstration NSSC
White Water 12?
38 Analytical Data and Rejections While Processing WSSC
White Water with Trailer Unit 129-130
39 Calculated Recovery of Feed Liquor Components Concen-
trate from Trailer Unit 131
to Summary of Operating Data RO Trailer 132
kl Comparison of Selected Data RO Trailer 133
^2 Summary of Flux Data from Re check Tests 135
^3 Pollution Loading of Pulp Wash Water from the Ammonium-
Base Acid Sulfite Mil
kh First Period Arrangement of Modules in Large-Scale
Reverse Osmosis Trailer Unit (June 9 to August 11, 1969)
U5 Second Period Arrangement of Modules in Large-Scale
Reverse Osmosis Trailer Unit (October 1-17, 1969)
^6 Flux Rate and Rejection Ratios of Pulp Wash Water —
Small Pilot Unit
Vf Summary of Trailer Downtime While Processing NHa-Base
Pulp Wash Water
Summary of Flux Rates and Operating Variables for
Base Acid Sulfite Liquor (June 9 to August 11, 19^9) 150
Rejection Data for NH3-Base Acid Sulfite Liquor (June 9
to August 11,, 1969) 151
50 Summary of Flux Rates and Operating Variables for
Base Acid Sulfite Liquor (October 1-17, 1969) 152-153
51 Rejection Data for NHa-Base Acid Sulfite Liquor
(October 1-17, 1969) 1?6
52 . Recovery Data for NHa-Base Acid Sulfite Liquor 157
53 History of Module Failures 159
xiii
-------
TABLES, 00
lo. Pages
5^ Summary of Flux Rates and Operating Variables' for
Hypochlorite Bleach Effluent l63
55 Rejection Data for Hypochlorite Bleach Effluent l6k
56 Ranges of Desired' Properties' of Process Water for
Kraft Pulping and Bleaching Operations 168
57 RO Feed Concentrations of Alkaline Extraction KBE 1J1
58 Variation in Analytical Characteristics of Different
Bleach Liquors 172
59 Outline of Operating Conditions-RO Processing of
Alkaline Extraction KBE 177
60 . Steady State Flux'Rates of Alkaline Extraction KBE 178
6l Average Rejections of Phase I Processing Alkaline
Extraction KBE • 179
62 Phase I — Performance Data Straight-Through Processing
of KBE " 180
63 Performance Data Summary — Phases I and II , I8l-l82
6k Phase III - Data Summary 185-187
65 Range of Recovery of Rejected Components at Various
Levels of Concentration 189
66 . Reverse Osmosis Processing of Combined First and Second
Stage.Bleach Plant Effluents . . 195
67 Rejection Data for Reverse Osmosis of Kraft Pulp Wash
Waters 196
68 Coa^arative Flux Rates"vs_ Solids Concentrations of
Kraft Wash Water 197
69 Analytical Data Pilot Unit Run on Kraft Pulp Wash Water 198
70 . Pollution Loading of Dilute Pulp Wash Water from a
Chemimeehanical Pulp Mill
71 Calculation of Press Efficiency 207.
xiv
-------
TABLES .(:5)
72 Concentration Levels and Rejections for CM ¥ash Water 211
73 Flux Rates and Rejection Ratios During Concentration
:Runs with Fresh Press Liquors 222
7k Effect of Press Mquors Aging 225
75 Indicated Findings and Goals for Concentration of
CM Effluents 227
76" Rejection Ratios for k Liquors at a Concentration
of 0.2-10 Percent Solids
77 Reynolds Number and Velocities of NSSC White Fater
Below Which Relatively Higher Permeation Resistances
Are Obtained 2l*6*
78 Effect of Pressure on the Membrane Constant and
Rejection Ratios 250
79 Effect of Velocity on Flux Rate Declines and Rejection
Ratios Due to Concentration Polarization and Fouling
Factors for Calcium-Base Acid Sulfite Liquor 255
80 . Effect of Velocity on Flux Rate Declines and Rejection
Ratios Due to Concentration Polarization and Fouling
Factors for Ammonia-Base Acid Sulfite liquor 256
8l Effect of Velocity on the Permeation Resistances and
Rejection Ratios — Calcium-Base Acid Sulfite liquor,
HSSC Liquor, and Kraft Bleach Effluent Liquor 26l
82 Flux Rates and Rejection Ratios During the Concentration
Run of Calcium-Base Acid Sulfite Liquor 281
83 Flux Rate and Rejection Ratios During Concentration Run
of ISSC liquor (I Module Configuration) 283
8k Flux Rate and Rejection Ratios During Concentration Run
of Liquor (II Module Configuration)
85 Optimization Criterion of Large Scale Reverse Osmosis Unit 288
86 Comparison Between Experimental and Computerized Mathe-
matical Model Flux Rates — Calcium-Base Acid. Sulfite
liquor 289
.xv
-------
TABLES (6)
No. Pages
8? Comparison Between Experimental and Computerized Mathe-
matical Model Flux Rates — NSSC liquor (I Module
Configuration) 290
88 Comparison Between Experimental and Computerized Mathe-
matical Model Flux Rates — NSSC liquor (II Module
Confi guration) 291
89 Effect of Number of Stages and Recovery Ratios of Water
on the Capital and Operating Costs of Reverse Osmosis
Unit 292
90 Effect of Number of Modules in Series of the Capital
and the Operating Costs of Reverse Osmosis Unit 293
91 Flux Rates and Permeate Quality with Spiral Wound
Modules on Life Study - (Acid Sulfite) 302.
92 Flux Rates and Permeate Quality with Spiral Wound
Modules on life Study - (NSSC) 3Qk
93 Flux Rates and Rejections for Seven-Tube Modules on
life Study 306
9k Module Replacement Schedule for the Seven-Tube life Study 308
95 Flux Rates and Module Replacement for Seven and Eighteen-
Tube Modules on life Study 311
96 Flux Rates and Module Replacement for Eighteen-Tube
Spiral Wound Modules on life Study 313
97 Flux Rates with Braided Fiberglass Tubular Modules 315
98 Permeate Quality from Braided Fiberglass Tubular Modules 316
99 Flux Rates and Permeate Quality with Fourteen-Tube
Modules on life Study ' 317
100 Flux and Permeate Quality with Fourteen-Tube Modules
on life Study 318
101. Flux Rates and Module Replacement for Sixteen-Tube
Braided Modules on life Study 320
102. Flux Rates and Rejections for Thirty-Six Tube.,Modules
on Life Study 322
xvi
-------
TABLES (7)
No* • Pages.'
103 life. Performance of 387 Modules on frailer-Mounted
Field Unit 32k
10U. Indicated Capital and Operating Costs of Reverse Osmosis
Units Based Upon Five Field Demonstrations — Calgon-
Havens 18-Tube Tubular Modules 333
105. Supplementary Capital and Operating Charge Estimates 337
xvii
-------
SECTION I
COICLUSOTS
The capabilities of reverse osmosis as a new tool for concentrating and
recovering the solutes in dilute pulp and paperraaking effluents have
been confirmed in intensive exploratory studies in laboratory and small
pilot-scale test programs. Optimum performance was "best achieved in
concentrating dilute feeds at about 0.5 to 1.5 percent solids by about
10 times to provide concentrates at 8 to 10 percent solids, and with
membrane rejections of 90 to 99 percent for most components in the feed.
Low molecular weight salts and volatiles were less well rejected.
Problems of concern and for which compensating operation parameters were
studied and developed included:
Fouling of the membranes by suspended particles, colloidal suspensoids
of large molecular weight organics, resins, pitch and the like could be
at least partially controlled "by pretreatment» by periodic pressure
pulsations believed to achieve backward osmotic flushing through the
membrane and by periodic washing of the membrane surfaces. But self
cleaning, high velocities of flow were found to be the most likely
route to maintaining high rates of flux through the membrane, and espe-
cially so with the newer, high performance, tight surface membranes be-
coming available for field tests in 1971. Minimum velocities of 2 feet
per second overcame concentration polarization, but 3-0 feet per second
were required to maintain adequate mass transfer rates. Osmotic pres-
sures ranging from 50 to 80 psia in bleach effluents and ehemiiaeehanical
pulp wash waters fed at 1 percent significantly reduced the effective
driving force as concentrations reached 10 percent solids and osmotic
pressures of 310 to 330 psia. Higher operating pressures were needed
to reach upper levels of concentration in those substrates. Concentra-
tion polarization did not appear to seriously affect performance in
these studies at operating pressures below 800 psig.
Larger confirming trials were conducted in field demonstrations ranging
from 5000 to 50,000 gallons per day on five waste flows of particular
concern to the industry. Concentration of the materials suspended or
dissolved in these wastes could be achieved at high levels of recovery
for all but the smaller molecular weight solutes and volatiles.
However, it was not possible to demonstrate sustained, long-term process
operating feasibility in the extended life performance tests of the
membrane equipment available for these demonstrations because of the
low levels of reliability for available membrane equipment in terms of
freedom from plugging of channels, freedom from failure of membrane
support structures, and freedom from serious leakages of internal con-
nections within the membrane module. Biose capillary fiber and spiral
wound sheet membrane systems tested were of excellent structural design
and stability but were subject to irreversible plugging by particulate
-------
matter contained In the feed or which developed during concentration.
Tubular systems operated at high velocities substantially solved plug-
ging and fouling problems, but none of the tested tubular designs were
free of structural failure or alternately of internal leakage problems.
Processing economics were affected most importantly by module failures,
and the resultant excessive charges for replacement and of maintenance
ranged as high as 60 to 80 percent of total operating charges as deter-
mined in computer based comparative cost studies.
Engineering studies for optimizing RO design indicate manifolding of
half inch diameter tubular systems might best be directed to limiting
installations of modules in series to a very few units (250 to 300
linear feet maximum) for each stage of concentration. The number of
stages seems of less concern as long as total holding in the system
does not exceed limits of about one hour. Hold up in the system should
not promote chemical precipitations or aging and break up of colloidal
systems, nor should it approach cell regeneration times for microbio-
logical slime growth. Straight-through operation of membrane systems
is indicated as a desirable goal with hold-up by recycling and in surge
systems limited to minimum periods of time.
Cost evaluations appear most sensitive to membrane module replacement
and maintenance charges in terms of sustained performance and life
expectancy; to membrane permeation rates as measured by the reference
NaCl flux rate; and to increase in osmotic pressures as concentration
proceeds. Large-scale commercial applications in the waste treatment
field cannot be expected until life performance of membrane equipment
has been improved far beyond the less than one year expectancy demon-
strated in these trials. Engineering design and manufacturing quality
control are problems under intensive development by suppliers and 'the
future of RO will depend on the success of these efforts.
-------
SECTION II
The chief roadblocks delaying practical application of RO to waste treat-
ment problems lie in the several causes for short life expectancy of the
membrane system. The less than 12-month life which could "be demonstrated
in these studies, in terms of either stress fatigue-related failures of
the membrane support structures, or alternately of sustained performance,
free of module plugging and internal leakage problems, was responsible
for excessive operational charges. Module maintenance and replacement
charges for short-lived equipment were calculated to range to as much
as $2 per thousand gallons of permeate water production.
Suppliers of membrane equipment, all of whom have been straining to per-
fect module design and manufacturing quality control to increase life
expectancy, should be encouraged in every way possible to attain goals
of a minimum 3-year average life, and the resultant reduction in main-
tenance and replacement charges.
Optimization of design of manifolding systems for large installations of
BO equipment should be verified with further mill trials under actual
plant operating conditions, and with minimum hold-up times, to reduce
degradation effects arising from aging of the feed waters.
Membrane development to increase capabilities for operating at wider
ranges of pH and temperature could substantially reduce operating charges,
Cooling of feed liquors can be a substantial expense, and higher tempera-
tures of operation could reduce or eliminate microbiological sliming.
Neutralization to suitable pH ranges for membrane processing can involve
substantial expense for reagents, and importantly also may involve
chemical or physical changes of phase, such as formation of precipitates
or break up of colloidal sols, with resulting fouling problems.
Dynamic membrane studies should be advanced to achieve higher -levels of
solute rejection without serious reduction in permeation rates. Some
waste flows have components, such as lignin, with capabilities for
dynamic membrane formation. Development of controlled conditions for
formation, removal, and reformation of such supplementary membrane
effects could substantially improve performance and cost parameters.
Promotion of turbulence of flows across the membrane surface has been
advanced with some success as a method of reducing power costs. Im-
provements are needed for designs tested in these studies to reduce
side effects, and especially of fouling in the presence of turbulence
promoters.
-------
SECTION III
lOTRODUCTIOI
An intensive program of screening, evaluating, and adapting of the rela-
tively new developments in membrane processes to the increasingly urgent
problems of treating dilute liquid wastes of the pulp and paper industry
was initiated in 1958 in the laboratories of the Pulp Manufacturers
Research League (located on the campus of The Institute of Paper Chemis-
try), llectrodialysis was first studied in detail tip through substan-
tial pilot-scale studies. Continuing development of other membrane
processes in the program of the Office of Saline Water, U.S. Department
of Interior, became of increasing interest, and additional League studies
on reverse osmosis (RO) and ultrafiltration were activated in 1965•
The first technical paper published in 1967 from these studies adapting
RO to the pulp and paper field1 attracted the attention of Federal pollu-
tion control authorities, and it was suggested that the program might
advantageously be extended and the development of practical applications
be accelerated by partial support from Federal research funding then
becoming available. An application to the Federal Water Quality Admin-
istration (now the Environmental Protection Agency) was formulated
and accepted as Research and Demonstration Grant 120^0 EEL for this
study on a total budget of $690,530.00, of which 70 percent, $U83,370,
derived from the Federal grant and the remaining 30 percent, $2Q7sl60,
was funded by the Research League and individual pulping concerns inter-
ested in field studies at their mill sites.
In support of such an application, a field demonstration program was
planned and substantial laboratory studies needed for design of Held
equipment were carried out during the 6-month period of final negotia-
tion for the grant.
Final specifications and contracting for construction of the large
trailer-mounted field unit could then be established promptly, after
the terms of the grant were finalized September 26, 1967. Final speci-
fications were developed for the field unit; bids were solicited; a
contract was negotiated; and final fabrication and assembly completed
in time for delivery of the unit October 9s 1968. Start-up tests were
completed in about three weeks and the first field denonstration was
gotten under way October 31, 1968.
Five field studies (three with the large unit at 50,000 gallons per
day and two with smaller field units at 1500 to 8500 gallons per day)
were conducted at intervals thereafter on (l) calcium-base acid sulfite
"digester cooling water," on (2) Neutral Sulfite Semichemical Machine
"White Water," on (3) NHs-baae acid sulfite pulp wash water, on (k)
kraft second-stage, alkaline extraction, bleach plant effluent (KBE),
and finally on (5) the wash water from a high yield chemimechanical
pulping process obtained by screw pressing high density pulp slurries.
-------
Much laboratory research on special problems and an extensive program
of careful analytical control were carried out concurrently on a smaller
scale at the mill sites and in the League's central laboratory in Apple-
ton.
Active laboratory and field studies were completed June 30, 1971 and
several months have been spent thereafter in collating and evaluating
the extensive backlog of data which accumulated during this 4-year
research and demonstration study.
The Pulp Manufacturers Research League merged into The" Institute of
Paper Chemistry effective April 1, 1970, but this research program
was continued without interruption by the League staff, which then
became the Effluent Processes Group of the Institute's DLvision of
Industrial and Environmental Systems.
These intensely interesting laboratory research and engineering applica-
tion studies for this membrane process were conducted during a period
of great competitive activity among many equipment supply firne each
striving to be first on the market with practical membranes and effi-
ciently designed modular membrane support equipment, pumps, controls,
and instrumentation as production items ready for commercial-scale
v&e. Development of reliable equipment with desirable life performance
expectations has been more difficult than anticipated, and has required
years rather than months to attain. Actuallys the work reported has
been conducted with a series of prototypes undergoing a continuous
process of improvement in design, construction, and performance.
Membrane equipment suppliers have supported an amazing and costly program
of research and development on new and improved membranes and support
structures. No authoritative sources are known on which to establish
firm estimates of the total research funding devoted to development
of reverse osmosis equipment over the past ten years, but there are
guesstimates that costs to industry, university, and governmental labo-
ratories and engineering research centers have ranged to the $100 x
106 level. This project has benefited greatly from the extensive inter-
est and cooperation from all of these sources. Commercial application
and use of reverse osmosis can be expected much sooner than would other-
wise have been possible.
Steady development of the proving out process and the eminent availa-
bility of practical RO concentration processing equipment has substan-
tially affected the thinking and planning for environmental engineering
studies on industrial waste treatment. New and improved routes to
compliance with new and more rigid standards of quality for industrial
process effluents are being sought. The availability of reverse osmosis
as a new tool for completing the closure of process water systems is
developing as another interesting area of application.
The first RO spin-off project directed to closing a pulp mill process
water system is under way at the ISSC mill of Green Bay Packaging Inc.,
-------
the site of the second field demonstration. This has advanced through
a first phase evaluation of membrane equipment available on the market
as of 1970-71 under a second Research and Demonstration Grant to that
company "by the Environmental Protection Agency.
A second application research program is in advanced stages of organiza-
tion on kraft "bleach effluent chemical recovery, which is expected
to be actively under way at The Institute of Paper Chemistry early
in 1972.
Additional application studies are in various stages of planning on
other waste streams.
This report is directed to delineation of the areas of possible use
for RO concentration processing equipment within the pulp and paper
industry. It will be interesting indeed to see the development of
pathways and the actual commercial-scale applications which develop
within the industry as a result of this pioneering effort.
Importantly, with respect to the identification of the exact source
or manufacturer of equipment tested and for which data are reviewed,
the reader's understanding is solicited on the format and design of
this report.
Every effort has been made to evaluate the advantages inherent in the
basic classifications of the many membrane formulations and the wide
variety and conformations of modular membrane support structures. Most
of the studies were conducted on the cellulose acetate class of mem-
branes ? with lesser studies on the nylon product. This report has
nxrtj been directed t;o identifying individual formulations (usually pro-
P:rietaj?y) for diacetate, triacetate, and polyamide products. The studies
Save covered substantial eowpa^atiye evaluations of three basic RO
design of membrane systems, including the capillary fiber, sheet mem-
brane pack, and tubular conformations. The text avoids identification
and discriminatory discussion of proprietary designs under continuing
development by individual manufacturers. An exception concerns purchase
of equipment with project funds. Such equipment, and particularly
that mounted on the trailer field unit, is specifically identified.
Equipment was purchased after careful evaluation of its relative merits
and advantages at the time of initiating a specific study. Changes
and improvements in design have substantially altered the status quo
on a month by month basis.
Great monetary benefits to the project arose in the supply without
cost of numerous samples and test modules of membranes and equipment
for life testing, as described in Section IX. This has often been
at substantial cost to the suppliers cooperating in these evaluation
and exploratory research studies. This contribution and excellent
cooperation is gratefully acknowledged. A list of cooperators is
supplied in Section XI.
-------
SECTION IV
PRINCIPLES OF REVERSE OSMOSIS AS APPLIED TO
CONCENTRATION PROCESSING OF PULPING EFFLUENTS
Knowledge of osmotic phenomena dates "back more than two centuries — the
experiments of the Abbe Nollet on diffusion through animal membranes
were published in 17^8. It was over a hundred years later, however,
that experiments with artificially prepared membranes were successful
(by Traube in 1867). In 1877 Pfeffer made the first quantitative mea-
surements, using a membrane consisting of copper ferrocyanide precipi-
tated in the pores of porcelain. A good review of early work on osmosis
was written by Findlay .
This section contains a description of the phenomenon of osmosis, a
brief development of the thermodynamic theory, a discussion of concentra-
tion polarization and fouling of'"the membrane, and some general consider-
ations of how these principles will effect the large-scale application
of reverse osmosis (RO) for concentrating pulp and paper industry dilute
effluents.
Description of Osmosis
Osmosis depends on the existence of a membrane that is selective in
the sense that certain components of a solution (ordinarily the solvent)
can pass through the membrane, while one or more of the other components
cannot do so. Such a selective device is called a semipermeable membrane.
When a dilute solution is separated from a concentrated solution by
a semipermeable membrane, there will be fluid transfer from the dilute
to the concentrated stream until an equilibrium pressure exists on
both sides of the membrane. This equilibrium pressure (actually, the
equilibrium pressure difference between the solvent and solution phases)
is called the "osmotic pressure." Osmotic pressure is a colligative
property of the solution and cannot depend in any way on the membrane,
so long as the latter has the necessary property of semi permeability.
Application of a positive pressure on the concentrated side of a mem-
brane, equal to the osmotic pressure plus a small, positive differential,
will cause the fluid to flow from the concentrate to the dilute stream.
This phenomenon is called "reverse osmosis" and the amount of water
flowing in the direction opposite to the osmotic flow is directly pro-
portional to the differential pressure applied for any given membrane
porosity (Fig. l).
The performance of a reverse osmosis membrane is generally characterized
by the flux rate and rejection capabilities of the membrane. The flux
rate at any concentration of the liquor can be related to the osmotic
pressure of the liquor by the following equation:
-------
OSMOSIS
When fluids of different con-
centrations in a vessel are
separated by a membrane, the
dilute solution will flow
through the membrane into
the concentrated solution
FHESH
WATER
[
SIA
WATER
'MEMBRANE
OSMOTIC PRESSURE
The level of the dilute solu-
tion drops and the level of the
concentrated solution rises
until an "equilibrium" is
reached. The pressure dif-
ference between these two
levels is the "osmotic
pressure,"
OSMOTIC
PRESSURE
1
FRESH
WATER
SIA
WATER
MEMBRANE
REVERSE OSMOSIS
If a pressure in excess of the
osmotic pressure is applied to
the concentrated solution, the
flow is reversed from the con-
centrated solution to the dilu-
ted solution.
osmosis."
This is "reverse
PRESSURE
350
psig
FRESH
WATER
SEA
WATER
L
MEMBRANE
Figure 1. Schematic Diagram of Reverse Osmosis
10
-------
F - A (Ap - Air) (l)
•where
P = flux rate through the membrane, gfd
A = water permeability coefficient, gfd/psig
Ap = difference between the applied pressure and the delivery
press-ore of product water, psig
ATT = (difference between the osmotic pressure of the liquor
and product water) + (osmotic pressure increase due to
concentration polarization and fouling effects), psig
The product water is delivered at atmospheric pressure. Since the
osmotic pressure of the product water is usually very small compared
to the osmotic pressure of the liquor, the former term can be ignored.
In the case of zero concentration polarization and fouling effects,
the driving force (Ap - Aw) "becomes equal to the difference between
the applied pressure (P^) and the osmotic pressure of the liquor (ir).
Therefore, equation (l) becomes:
p = A(PA - TT) (2)
From equation (2), it is apparent that the higher the osmotic pressure
of the liquor the lower the flux rate for a fixed applied pressure.
The osmotic pressure of a solution depends on the concentration, activity
coefficient, degree of ionization of the various components in solution,
and the temperature.
In equation (2)
Tf = iRTC (3)
where
tr = osmotic pressure, atm
i = Vant Hoff's factor, which takes into account the degree
of ionization and activity coefficient
R = gas constant = 82.1 atm/g mole cm3 °K
T «' temperature, °K
C = concentration, g mole/cm3
For an ideal dilute solution, i has a constant value. Therefore, the
osmotic pressure, TT, is directly proportional to the temperature and
concentration of the solution. In the case of pure water, IT is equal
to zero.
The rejection ratio is an important parameter in the design of a reverse
osmosis unit for the purpose of concentrating the feed over a wide
range or for chemical fractionation systems. For membrane concentration
processing, the rejection ratio is defined as the ratio between the
concentrations at both sides of the membrane at a certain spot:
11
-------
R = (i _ -L) x 100 (10
where
R = percentage rejection ratio
= concentration of the permeate
= concentration of the feed to the module
It is important to mention here that both of these concentrations should
be considered from the standpoint of being measured instantaneously.
Concentration Polarization and Fouling of the Membrane
Most of the dilute wastes of pulp and paper industries are complex
in character, containing significant amounts of colloidal and fine
particulate suspended solids. These colloidal and suspended solids
have a tendency to "coat" or "foul" the membrane surfaces, thus result-
ing in poorer long-term flux rate and rejection characteristics of
the membranes.
Coating or concentration polarization implies the accumulation of solids
at the membrane surfaces due to the bulk movement of liquor toward
the membrane and rejection of solids at the membrane liquor interface.
The solids concentration at the membrane liquor interface increases
until back diffusion caused by the concentration difference balances
the corrective flow of solution to the membrane and that which leaks
through the membrane in the product water. A good review of concentra-
tion polarization and its effects on flux rate have been described
extensively in the literature3*1*. Fouling of the membrane surfaces
by microbiological growth is often observed in sustained operations
with wastes containing nutrients capable of promoting growth of bacteria,
yeasts, and molds.
One of the methods of minimizing concentration polarization and fouling
of reverse osmosis membranes has been to maintain adequate degrees
of turbulence and mixing within the membrane tube. The concept of
higher velocity is very useful in sweeping the membrane clean, thus
maintaining economically feasible long-term, steady flux rates. However,
it should be kept in mind that, although higher velocities across the
membrane surface can reduce the physical accumulation of solids near
the membrane, they cannot eliminate chemical or electrical affinity
which the solids may have with the membrane. Special problems of pre-
treatment may arise if such affinities are apparent. A detailed de-
scription of the effects of velocity on the flux rate and rejection
performance of RO membranes is discussed in the later sections of this
report.
12
-------
Application of RO Principles to Concentration
Processing of Dilute Pulping Effluents
In this modification of RO for concentration processing and fractionation
to achieve pollution control and other routes to effluent treatment,
the -water contained in relatively dilute process streams is forced
under pressure through a membrane in conventional RO modules originally
designed for salt water conversion applications. However, the objectives
in this program call for a much higher degree of water recovery ranging
from JO percent to more than 95 percent of that contained in the dilute
feed waters and also for much higher concentration of the dissolved
solids rejected "by the membrane system than is normally practiced in
salt water and "brackish water conversion. Here the degree of pretreat-
ment required, the' quality of the concentrate and of the water permeate
become matters of added concern, since these products are to be subject
to reuse or final disposal without pollution of the water, land, or
air environment.
An RO plant will mainly consist of:
(l) A pump for raising the pressure to the chosen operating
pressure.
(2) The reverse osmosis membrane unit, in which the feed
liquor is separated into water permeate and concentrated
liquor.
(3) the recycling pumps for overcoming the frictional pressure
drop and for maintaining adequate velocities across the
membrane surface.
The minimum energy can be calculated thermodynamically2; it is simply
the product of the osmotic pressure and volume of the solution. In
actual use all processes, including reverse osmosis, use much more
energy than this. There are several reasons for the extra energy needed
in reverse osmosis. In the first place, although any pressure exceeding
the osmotic pressure will cause reverse osmotic flow, achieving a prac-
tical rate of flow may require a much higher pressure, perhaps several
times as high as the osmotic pressure. Second, as the product water
is removed from the dilute pulping liquor feed, the concentration,
and therefore the osmotic pressure of the remaining liquor is increased.
Finally, there is a significant increase in the membrane fouling as
well as frictional pressure drop with increase in the concentration
of the liquor.
Complicating the above considerations is the relation of plant cost
to operating pressure. A high pressure will increase the product water
rate and so decrease the membrane area. This will tend to reduce the
cost of the reverse osmosis unit. Against this saving will be the
increased cost of energy and pumping equipment.
Another important factor to be considered in the RO plant design is
the concentration polarization and fouling of the membrane surfaces.
The build-up of fouling materials near the membrane surfaces must be
13
-------
removed by diffusion "back into the bulk of the solution; this can be
done much more effectively if diffusion is aided by turbulent flow.
It has been found experimentally that higher degrees of turbulence
and mixing is advantageous in minimizing the" concentration polariza-
tion and fouling of reverse osmosis membranes.
These general principles will be frequently referred to in subsequent
chapters describing the adaptive research program covered in this report,
-------
SECTION V
LABORATORY STUDIES ON THE REVERSE OSMOSIS CONCENTRATION
OF EFFLUENTS FROM THE PULP AND PAPER INDUSTRIES
This section deals with the early evaluation studies conducted in the
laboratory as a necessary step preliminary to selecting equipment
best suited for laboratory, pilot, and field studies. Subsequent labora-
tory programs were required to develop answers to problems arising
in the larger scale work in pilot and field studies. Equipment"of
various types available from the several suppliers active in the field
had to be evaluated in terms of specific process variables, and especial-
ly with capability for handling industrial waste flows containing sus-
pended material, colloidal suspensoids, and large molecular weight
organics which could foul the membrane assemblies.
INITIAL LABORATORY STUDIES
Equipment
Test Stands — Pumps and Controls
The early exploratory studies on reverse osmosis processing for concen-
tration and fractionation of dilute effluents from the pulping industry,
were conducted with a small desk model of a laboratory unit supplied
by Gulf General Atomic and designated as their Mark II lab unit. This
equipment was designed by the manufacturer to test spiral wound ROGA
membrane modules, but in this project, was later adapted to testing
other types of small membrane assemblies. This unit had certain limi-
tations, and particularly with the limited rate of flow from its small
pump, but it was used effectively and is still in service as problems
arise within its capabilities.
For testing larger reverse osmosis membrane modules, a number of-Hypro
rotary piston pumps were acquired for test stands delivering from 2.5
to 20 gallons per minute at pressures to 600 psig.... These units were
effective, but were expensive to maintain, since they were not designed
for continuous, round-the-clock service required for module life studies.
Poor lubricating characteristics of the pulping effluents under test
tended to accelerate wear and scoring of the pistons and the need for
frequent and costly replacement.
These rotary pumps were in turn replaced with more reliable equipment
for sustained operations undertaken within this research and demonstration
grant project. Variable stroke reciprocating piston pumps (Model C,
Milton Roy, Types MR-1 Simplex and MR-2 Duplex) were mounted on heavy
frames with four rubber-tired wheels for ready portability from one
site to another. Each portable test stand included one-gallon accumu-
lators for each individual pump, together with supporting pressure
15
-------
gages, "back pressure regulator valves (Victor Acme K-20), and instru-
mented controls and relays for shutting the entire unit down if the
desired ranges of liquid level, pressure and temperature were exceeded.
These units were also equipped with timers and mechanized valves de-
signed to pulse the unit by bypassing the back pressure control valve
to reduce the pressure to zero for brief intervals to achieve normal
osmotic flow back through the membrane in handling fouling problems.
A good deal of exploratory laboratory and small pilot study was accom-
plished with one Simplex pump unit and one Duplex pump unit acquired
for preliminary studies conducted prior to undertaking the work under
the Federal Research and Demonstration Grant 120^0 EEL covered in this
report. After starting the grant studies, two additional field test
stands of this type, one Simplex and one Duplex, were constructed to
extend the program for preliminary field evaluation of various wastes
at individual mills. This gave a total of six pumping units on four
test stands, all of which have been in continuous service for some
four years in and out of our laboratories. One of these units in field
service is shown in Figs. 2a and 2b. Individual flow sheets incorpo-
rating these units in various demonstration studies are provided in
Section VII.
Modular Membrane Equipment
The selection of membrane equipment suitable for pulp and paper effluents
was a critical item throughout these studies. The various processing
parameters affecting the rate of solvent (water) permeation or "flux"
through the membrane were especially critical to the success of this
project and were dependent upon maintaining clean hydraulic flows across
the surface of the membrane to avoid fouling of the membrane surface.
Not all types of equipment on the market could meet the requirements
for being able to handle small amounts of suspended solids, and espe-
cially of fiber contained in most mill effluents. Again, the membrane
system design had to be capable of maintaining sufficient turbulence
or velocity across the surface to avoid deposition of precipitates
and crystalline material which might be thrown down or which might
form slimes or scale during the course of concentrating a mill waste
flow. Problems were also apparent in maintaining the membrane equip-
ment free from fouling caused by pitch and other resinous and gummy
materials released from colloidal solutions during the course of
processing. A breakdown of colloidal suspensoids of pitch and resin
particulate matter apparently takes place with changes in pH, tempera-
ture, and pressure. The data seemed to indicate that much less fouling
was experienced when the breaking of the colloidal sols could be avoided
or reduced.
These problems were factors in the selection of equipment to be used
from the various types which were submitted for testing. Capillary
fiber RO membrane equipment has been shown to have interesting capa-
bilities in the field of processing clear brackish waters, but in these
studies could only be used for limited purposes on pulp and paper
16
-------
Figure 2a. MLlton Roy Duplex Pump on Wheel Mounted Pilot-Scale Teat
Stand, Complete vith Motor, 2 Accumulators, Automated On-Off
Controls for Temperature and Liquid Level, and Motorized Valve for
Pulsing on Programmed Time Schedule
Figure 2b. Photograph of Pilot RO Unit Complete with Instrumented Pump
Test Stand, Bank of Modules, and Feed Tank Set up for Continuous
Operation on Recycle for Life Tests
17
-------
wastes, such as the processing of clear evaporator condensates having
no carry-over of colloidal material or of suspended matter which could
plug the fiber bundles.
Restrictions to free flow were also apparent in the membrane separation
structures of sheet membrane pack systems, such as the spiral wound
modules which were available for testing in this grant study. Such
restrictions made the sheet membrane systems susceptible to plugging
with suspended matter much as with the capillary fiber system, though
to a lesser extent. However, several promising applications for the
sheet membrane (spiral wound) module assembly became apparent during
these grant studies, and the suppliers have been modifying these membrane
pack assemblies to reduce this problem. The greater ratio of membrane
surface area to modular volume should substantially reduce capital
and replacement costs, and it is likely that such sheet membrane assem-
blies may prove economically advantageous in the processing of clear
effluents, and for early stages of concentration of very dilute flows
up to the point where precipitates and crystalline material may arise
as concentration proceeds.
As the picture developed in the preliminary studies and advanced with
pilot and field research, it became more and more apparent that of
the membrane systems available at that stage of development, tubular
systems were best adapted to the overall requirements and objectives
for this project. "Hie 1/2-inch tubular systems of Calgon-Havens (Havens
International), Aqua-Chem, Westinghouse, American Standard, and Enviro-
genics (Aerojet-General), have been tested and found to have capabilities
for handling the plugging and fouling problems attendant to processing
of pulp and paper industry wastes. Tubular systems utilized in this
project all had 1/2-inch diameter tubes fabricated with various metal,
fiberglass, and resinous materials for the membrane support structures.
Cellulose acetate membranes and various formulations are cast in place
or inserted into these tubular supports.
Types and Grades of Membranes
Selection of the membranes most effective for these studies was also
a critically important problem. Each of the suppliers provided various
cellulose acetate formulations and membranes were cast and tempered
in various grades of permeability. These grades of permeability tend
to be available from different suppliers in from 3 to 5 degrees of
"openness." The system of classifying the membranes was not uniform
between the various manufacturers, but in general, a No. 1 or No. 2
membrane was most open, and membranes with the highest numerical designa-
tion had the greatest degree of tightness and of solids rejection.
For control purposes these membranes are usually graded in terms of
rejection of NaCl from standard solutions with one gram or five grams
of salt per liter. For the purpose of evaluating membrane equipment
for processing of pulp and paper mill effluents, these studies tended
toward a 3-step standardized system of control tests; first with control
tests for flux rate and rejection using city tap water (or distilled
18
-------
water in closely controlled laboratory studies), then with 0.5 percent
NaCl solution, and finally with performance tests on a 1 percent "stan-
dard" solution of Ca-base spent sulfite liquor solids having various
inorganic and organic components in small and large molecular weight
classifications. This "lignin liquor" product is available on the
market throughout the U.S. as a spent sulfite liquor (SSL) concentrate
with 50 percent solids and also as spray dry solids. A 1 percent solution
of Ca-base SSL solids has the advantage of providing a solution of
colloidal lignin, of various wood-derived sugars and of other character-
istic pulping process organics along with* inorganic pulping chemicals.
A minor disadvantage in preparation and use of the "standard pulping
test liquor" arises in that some volatiles as acetic acid and methanol
present in the original digester liquors in amounts usually less than
1 percent of the total solids are lost in evaporation and spray drying.
These constituents are usually of little significance for the type of
tests reported, but they can be added or their absence allowed for if
close control of BOD 5 or similar problems are being studied. Tests
with the 1 percent Ca-base SSL "standard test solution" will often
be referred to in the text and tables of this report.
METHODS AND EXPERIMENTAL PROCEDURES
Samples for laboratory studies were brought in from the mills in quanti-
ties suitable for the individual study under way. Routine screening
of modules and membranes was usually carried out with use of the 1
percent Ca-base SSL test solution made up from the 50 percent concentrate
or of spray dried material from the same source which were available
through the courtesy of the Appleton Division of Consolidated Papers.
Similar spent liquor products commercially available from other sources
around the country have generally proven to be equally suitable for
this purpose.
The development of test data on a wide variety of other mill effluents
from pulping, bleaching, and papermaking generally involve samples
of one gallon to 50 gallons for laboratory study, and even larger
quantities were required for sustained module and membrane life studies .
Samples for membrane testing were clarified if necessary through
mesh screening prior to use in tubular assemblies. More complete clari-
fication with 3 to 50 IM filtration was practiced for capillary fiber
and spiral wound membrane pack processing equipment.
Provision was made for adjustment of pH and the control of temperature
as needed for these laboratory test programs . More elaborate pretreat-
ment was seldom resorted to for routine testing, although there were
a few instances where various chemical methods of precipitation and
flocculation and physical methods of centrifuging, ultrafiltration,
and the like, were tried as possible answers to difficult problems.
In general, the programs were carried out with little more than screening
in the belief that effective membrane processing should be capable
of handling industrial waste flows with a minimum of costly pretreatment.
19-
-------
ANALYTICAL FOB EVALUATION OF PROCESS AND EQUIPMENT
Experimental Procedures
Various analytical procedures were used to evaluate the performance
of the reverse osmosis concentration and fraetionation systems on pulp
mill effluents . • Complete evaluation involved the analysis of the feed
liquor to the RO process 5 of the permeate (product water); and of the
rejected solutes in the concentrate stream. Boutine assays included
total dissolved solids (2k hours at 103~105°C); neutralized 72-hour
solids in cases •where acid volatiles were present; chemical O3Qrgen demand
(COD), NaCl (as chloride); lignin (optical density at 28l run); 5-day
biochemical oxygen demand (BODg); pH and color (Co-Pt units). Other
specialized assays were used at times.
CRITERIA FOR SELECTION OF -
EXPERIMENTAL DEVELOPMENT
A first consideration in selecting RO equipment for these studies was
based upon performance capabilities of a membrane system for concen-
trating the solutes in dilute pulp and paper effluents at feed concen-
trations on the order of 1 percent total dissolved solids (IDS) by a
factor of 10 times or more to achieve intermediate concentrates at about
10 percent solids suitable for economic final concentration or disposal
by conventional evaporation and combustion systems.
' Selection of Membrahe_!Iy'pes and Grades (Degree of Rejection
or Mambrane "Ti ghtness " )
The rate of solvent (water) permeation (flux rate) and the impermea-
bility to solutes (the rejection ratios) are dependent upon the "tight-
ness" or "openness" of the membrane at any given pressure level. A
variety of different membranes from manufacturers or from research
centers were evaluated in our early laboratory studies of reverse osmosis ,
A series of experimental runs with controlled tests for comparison of
performance of various types and grades of membranes was undertaken.
Table 1 summarizes the data derived from membranes in spiral wound "BOGA"
modules from Gulf General Atomic (now Gulf Environmental Systems) in
three levels of rejection (tight, intermediate, and open membranes)
utilizing a first-stage chlorination effluent from bleaching Ca-base
acid sulfite pulp.
Flux rates progressed from 6.1 gfd for the tight membrane to 10.9
for the open ' jBsmbrane i and the corresponding rejections were on the
order of 50 percent or less for the open membranes and 90 percent or
more for the tight membrane. For the purpose of reusing the .clear water
recovered at the 80 percent level in the permeate, the intermediate
membrane grade apparently performed with the best economy to give satis-
factory quality of water for reuse.
20
-------
TABLE 1
EFFECT OF MEMBRANE "TIGHTNESS" ON PRODUCT WATER QUALITY
Chlorination stage sulfite bleach effluent
Spiral wound modules
1*50 psig - 25°C
1.8 gpm - Feed rate
80 Percent water recovery
Flux,
Sample gfd
Process feed
Product water
Tight 6.1
Intermediate 7'0
Open 10.9
Concentrate
Tight
Intermediate
Open
Solids ,
mg/la
I960
1U
72
822
12120
61.UO
5160
Optical
COD, Chloride, Density .
mg/la mg/la PH at 281 nm°
1*30 2.2 8.75
100 36 2.6 0.191*
132 251 2.2 0.261
516 1*00 2.3 3.63
6655 3260 2.2 80.1*
1*366 196U 1.9 M.8
3U05 852 2.2 31.7
^Based on composited samples for each recovery level.
A measure of the lignln content.
A second and similar test of tubular membrane equipment using higher
grades of membrane tightness available from Havens Industries (Type
3 — intermediate; Type h — moderately tight; and Type 5 ~ tight) con-
firmed this picture and extended the data at various levels of 32,
60, and 85 percent water recovery (Table 2).
A third test (Table 3) in this series compared the open Type 2A Havens
membrane with the intermediate Type 3 membrane at HO, 60, and 90 percent
water recovery. The flux rates for the open membrane remained high,
but the solids and chloride rejections were reduced below the 50 percent
level as the water recovery advanced to the 60 and 90 percent levels.
These data have been confirmed repeatedly in an extensive continuing
evaluation on other waste flows, and were the base for establishing
equipment specifications for this project at the intermediate degree
of membrane tightness to achieve economic feasibility in terms of
21
-------
TABLE 2
CONCENTRATION WITH TUBULAR UNITS WITH THREE DIFFERENT
IffiMBRANE "DEGREES OF TIGHTNESS"
Ca-Base acid sulfite first stage bleach effluent
650 psig Operating pressure
Temperature 25°C
k gpm Feed rate - recycle flow
Havens
Membrane
Process feed
Product water
Type 3
Type 1*
Type 5
Concentrate
Product water
Type 3
Type I*
Type 5
Concentrate
Product water
Type 3
Concentrate
Flux,
gfd
For
ll*.7
8.7
8.1
For
13.5
10.2
9.0
For
8.3
Solids ,
mg/1
1586
32 Percent^ Water
21
16
13
180 k
60 Percent. Water
' 61*
1*8
31
312k
05 Percent Water
368
9239
COD,
mg/1
988
Recovery
10 1*
6k
88
117 fc
Recovery
102
T6
T8
1687
Recovery
JM
5070
Chloride,
ag/1
370
63
16
19
551
83
IT
13
850
198
2131
Optical
Density
at 281 ni
7. It
0.090
0.063
0.051
9-U
0.102
0.065
0.066
1U.7
0.211
»»5.9
A measure of the lignin content.
22
-------
TABLE 3
CONCENTRATION WITH TUBULAR UNITS CONTAINING TYPE 2A (OPEN)
AND TYPE 3 (INTERMEDIATE) MEMBRANES
Ca-base acid sulfite first-stage bleach effluent
650 psig Operating pressure
25° C Temperature
1» gpm Feed rate - recycle flow
Sample
Process feed
Product water
1
2
3
Concentrate
Process feed
Product water
1
Concentrate
Water
Recovery,
percent
Flux Solids, COD,
gf d nig/1 ing/I
For Type 2A Open Membrane
655
Chloride,
mg/1
162
313
370
U66
1*0 20.9 602
60 30.7 816
90 22.6 1038
6928 5655
For Type 3A Intermediate Density Membrane
2200 1200
90
8.0
22
53300
25100
15
12U90
Optical
. Density
pH at 281 nma
2.3
2.8
2.7
2.5
2.3
2.1*
2.3
7.0
0.88
1.0U
1.68
30. 1»
12.0
0.067
2.U 250.0
A measure of the lignin content.
practical rates of flux and of recovery of dissolved components in
the concentrate and relatively clean reusable permeate waters. Almost
all studies thereafter were conducted with membranes of intermediate
levels of tightness equivalent to Havens Type 3, Occasional trials
were made from time to time with new membranes coining from a continuing
program of research and development conducted at universities and by
individual suppliers. Late in 1970 reports became current of substan-
tial improvements becoming available in the flux rates for tighter
membranes from several suppliers, and such membranes in single test
modules did become available for preliminary testing in the Spring
months of 1971, as reported in Section IX, but this project was at
that time in the final phases of completion, and a significant program
of testing could not be undertaken for review in this report.
23
-------
MODULE EESIGFS
The first modular equipment available for laboratory studies was of
the spiral wound design. This system had a number of engineering design
and economic advantages for processing large quantities of dilute efflu-
ents , namely :
a. The large area of membrane per unit of equipment volume
would result in reduced overall space requirements,- compared to otter
styles of aodule arrangement, tubular or plate-frame."
b. The diameter (and area) of the modules could be adjusted
during manufacture to fit any size pipe.
c. Bie modules could "be readily placed in series or parallel
operation.
d. The use of sheet membranes and larger membrane surface
to module volume ratio has significant cost advantages both for capital
cost' and also the important operating charge category of loembrane main-
tenance and replacement.
The flow pattern through thin separator channels between membranes
was, however, more subject to plugging by feed liquors containing par-
ticulate matter, or from which precipitates formed during concentration.
In the processing of mixed first and second stage bleach effluents
(kraft or sulfite), for example, precipitates slowly formed and settled
within the module flow paths and on the surfaces of the membrane even
after filtration through a 50 ]M filter. Periodic baekwashing cycles
to remove the precipitated, crystalline scale, partly alleviated the
problem, but this procedure would add markedly to the cost of commercial
scale operations .
The clear hydraulic flow pattern of tubular-type modules has been found
more adaptable to the processing of pulping effluents containing partic-
ulate matter (fibers) and constituents with low solubility products
( calcium sulfate and calcium sulfite) which have caused module plugging
in the more restricted flow channels of other module designs.
•First and second stage kraft "bleach effluents and mixtures of the two
could "be. processed without clarification for the production of high
quality water and a four- to fivefold concentration of the' solids (Table
Tubular modules from five manufacturers have been evaluated in our
laboratories along with two spiral wound designs and a hollow fiber
system* Ibe excellent area/ volume ratios previously referred to for
the spiral wound modules are markedly extended^ In the hollow fiber
concept. A pressure vessel 0.5 inch in diameter and five feet long
can accommodate 18 square feet of membrane surface. At the same time,
the plugging characteristics of the thin-flow patterns have been inten
sified'by the natural filtering capacity of the fiber ''bundles.' In
this experience, feed streams containing particulate matter required
21,
-------
TABLE It
PROCESSING CLARIFIED AND UNCLARIFIED KRAFT BLEACH PLANT
EFFLUENTS THROUGH TUBULAR MODULES
Type 3A Membrane
U50 psig Operating pressure
Temperature 25°C
2 gpm Feed rate - recycle flew
Sample
Water
Recovery,
percent
Flux,
gfd
Clarified Chlorination Stage
Solids,
mg/1
3120
BOD,
mg/1
570
COD,
mg/1
2050
Chloride.
mg/1
11UO
Optical
Density
at 281 nm
9-7.
Product water
1 20
2 UO
3 60
It 80
Concentrate
5-7
6.3
6.7
6.5
Unclarified Chlorination Stage
90
86
100
131
33>»5
166
800
590
200
262
2U2
269
8625
2130
UO
35
72
60
5025
975
0.157
0.167
0.1U6
0.167
23.0
9-7
Product water
1
2
3
1*
Concentrate
Clarified Alkaline
Product water
1
2
3
it
20
Uo
60
80
Wash
20.
UO
60
80
5.27
6.0
6.5
6.5
5.5
5-8
6.8
7-1
Concentrate
Unclarified Alkaline Wash
52
50
57
75
15U75
1335
25
' 2U
2U
30
5350
1230
62
1000
60
222
228
228
231
8700
550
18
18
17
19
1608
550
20
28
23
35
U900
335
12
12
9
16
1U12
320
0.072
0.058
0.067
0.07U
0.15
0.036
0.029
0.033
0.039
0.19
Product water
1 - 20
2 UO
3 60
U 80
Concentrate
U-7
5-7
6.2
6.5
U8
32
23
32
U300
lU
625
25
22
17
12
1078
12
20
8
21
1368
0.03U
0.027
0.02U
0.028
0.79
A measure of the lignin content
25
-------
careful filtering (5 ym or less) prior to processing in the capillary
fiber system.
In some cases, such as with evaporator condsnsates, the feed streams
are relatively free of suspended materials and low solubility constit-
uents . Capillary fiber systems were tried with evaporator eondensates
and tended to confirm these expectations, but still with practical
problems apparently arising from uncontrolled microbiological growth
and the tests were terminated.
.PULP AID PAPER PROCESS
Acid •Sulfite Pulping
Ca-Base 'Standard' Test Liquor
Ca-base "lignin liquor" has been adopted for the purpose of this project
to fill the need for a standard liquor upon which comparative control
tests could be established throughout the entire program of 3-1/2 years
of research and field studies. Calcium-base acid sulfite digester
liquors at 10 to 3A percent solids are evaporated to 50 percent solids
concentration and the resultant "lignin liquor" is marketed commercially
in that form nationwide, and also as a spray dry product. This product
was usually diluted with tap water to 1 percent concentration as a
standard for the frequent control tests.
Ca-Base Aci d__Sulfite_Fulp_W.ash_Water
This product is available from advanced stages of washing or "cooling"
of the', pulp and was tested as received at about 1 percent solids +_
0.5 percent.
Acid Sulfite Evaporator Coride nsate
Many acid.sulfite mills evaporate their strong liquors. Volatile acids
as; S02, acetic and formic acids, and other volatiles are contained
in the evaporator eondensates and comprise a major source of BOD in
the total effluent flow from an acid sulfite pulp mill.
AddlSulftte 'Bleach 'Plant Effluents
Two such effluents, one a first stage chlorination effluent and one
a single-stage Ca hypochlorite effluent were tested.
'HHs-Base 'Acid;Sulfite Wash Waters
Similar to wash waters from Ca-base acid sulfite wash waters.
26
-------
Mg and Na-Base... High-Yield Bisulfite Wash Waters
Similar to Ca and MS base, but usually have higher pH levels due to
the cooks with high levels of bisulfite to achieve higher pulp yields.
Neutral Sulfite Semi chemical "White Water"
The pulp wash water from on-machine washing of paperboard pulp and
related products is referred to as "NSSC white water" in common with
the terminology for the fiber containing effluent draining from paper
machines .
Kraft (Alkaline Sulf ate) Pulping
Kraft Pulp Wash or Rewash Waters
These effluents may exist as residues in older kraft mills, but in
most cases are being systematically eliminated as mills are modernized
or replaced with new continuous digesters having in-digester washing.
Nearly all residual waters are recycled and pass on to the evaporators
and chemical recovery systems. These flows no longer appear to be
a large problem needing study by the RO route. Study of a rewash water
was initiated early in this project but the work was terminated short
of a full-scale field demonstration.
Kraft Bleach Effluents (KBE)
The bleach plant effluents of kraft pulping comprise a major source
of the remaining incompletely solved water pollution problems in the
kraft pulping branch of the industry. High levels of dilution are
characteristic (10,000 to 30,000 gal. /ton bleached pulp), and although
BODs loads are relatively low (30-50 Ib/ton pulp production), the color,
the content of resistant organics with high COD, and the substantial
level of inorganic salts (NaCl, NazSOtJ are of concern. Usually there
exists a multiplicity of effluents from the various bleaching operations
or sequences such as in (CEDED), in which the first stage of chlorina-
tion (C) (low pH) is followed by two stages of alkaline extraction
(E) alternating with two stages of low pH chlorine dioxide (D) bleaching.
Tests conducted within this RO research and field demonstration grant
project included substantial trials and a field test on second-stage
caustic extraction effluent and with lesser studies on the first stage
chlorine (C) effluent.
Kraft 'Evaporator Condensates
These condensates can be substantial air and water pollution problems
from kraft recovery systems , but preliminary testing was not promising
and these wastes were not included in the program here reported.
27
-------
Barking Waters
The "barking of wood is usually conducted with late-stage recycle waters
which derive from various mill sources low in BODg and chiefly comprise
problems of dealing with suspended solids and fines in primary clarifi-
cation systems. Preliminary tests for evaluation of possibilities
for concentration of low levels of solubles were conducted in this
study, but the importance and feasibility of such processing remained
in question, and large-scale tests were not attempted within the budget
limitations of this project and availability of time of the staff.
Paper Ml 1 .Effluents
Various small-scale tests on paper machine dye wastes, coating wastes,
and machine vhite waters were conducted, but the limitations of cost
for processing such extremely dilute (0.1 percent TDS or less) effluents
could not justify RO concentration processing by factors which would
range upwards of perhaps 100 times or more. Other methods of disposal
processing of such dilute waste flows appear more practical at this
stage of developing the membrane processes.
Deinking Wastes
This type of waste high in content of coating clays, hydrocolloids,
and printing ink residues is a critical problem for which membrane
processing has been considered and tested in a. preliminary way. Results
have been marginal. Further study may change this picture.
technically, the testing program has shown that hig^i quality water
for recycle and reuse can be recovered from all of these wastes by
reverse osmosis under properly controlled conditions of fluid velocity,
pE, temperature, and control of changes in state for colloidal materials,
Commercial feasibility as a means of•concentrating dissolved pollutants
is quite another matter, and the prospects of large-scale application
are limited to those effluents-having a suitable range of concentration
of dissolved materials and of effluent volumes upon which capital and
operating costs can be justified. Pulp wash waters at about 1 percent
solids concentration from which chemical values, clean reusable waters
for recycle, and major pollutants can all be recovered have been of
first concern and have received priority in this study.
Rejection Batios
The degree to which an RO membrane is capable of rejecting and concen-
trating solutes from the feed liquor to produce clean reusable permeate
waters is a primary consideration in close association with the permeate
flux rates. The rejection ratio is calculated by the formula:
C
Percent Rejection = (l - -f-) x 100 (1»)
28
-------
where
C = concentration of the constituent in the product water
Crp = concentration of the same material in the original feed
The rejection ratio ranges summarized in Table 5 have been achieved
for the pulp wash waters and bleach effluents tested and to lesser
extent in such cases as with volatiles, which pass freely through the
membranes at low pH levels, as summarized in Table 6. Acid volatiles
may be neutralized and concentrated with technical success, but the
economics have yet to be developed on this type of waste flow.•
TABLE 5
PERCENT REJECTION FOR TYPICAL PULP WASH AND BLEACH EFFLUENTS
600 psig Operating Pressure
Temperature 25°C
Membrane Type
Constituent Intermediate Tight
Solids,, total 75-98 95-99
Optical density (color) 98-99.8 99-99.8
Biological oxygen demand (BOD) 80-95 90-99
Chemical oxygen demand (COD) 85-98 90-99
Inorganics 60-95 9^-99
TABLE 6
PERCENT REJECTION DURING THE RO CONCENTRATION OF CONDENSATES
FROM THE EVAPORATION OF SPENT SULFITE LIQUORS
600. psig Operating Pressure
Temperature 25°C
Intermediate Membrane Tight Membrane
Constituents pH 2.7 pH I*.8 pH 6.1 pH 2.7 pH U.8 pH 6.1
Solids, total 27 57 70 ^5 8l 95
Optical density (color) 66 55 ^8 79 71* 80
Biological oxygen demand 26 k$ 58 52 62 82
Chemical oxygen demand 19 1*3 57 37 6"2 8l
Volatile acids (as
acetic) 7 1*0 65 26 70 95
29
-------
'Process Variables
Teii|ger_ature_
Processing temperature was from the first considered to be an important
parameter of operation. Prior experience had indicated' an increase
in temperature from 10 to 31°G would approximately double the permeation
rate for a given membrane at a single pressure level1*5. Later studies6
with a calcium-base acid sulfite liquor in the concentration range
of 12 to 10^ grains solids per liter (osmotic pressure 125-200. psi)
confirmed the temperature effect to be on the order of 2.1 percent
per degree centigrade at the 12 g/1 concentration and was not signifi-
cantly different at the IQl* g/1 level, A search for membranes capable
of operating at l4Q°C or above continued throughout the project not only
to achieve advantage from higher flux rates but also to reduce require-
ments for expensive cooling pretreatment for warm or hot feed liquors.
Fluid Velocity
Concentration polarization, the formation of stagnant layers of fluid
at the membrane surface, produces several effects detrimental to the
efficiency and economics of membrane separation processes;
1. The osmotic pressure of the stagnant layer increases with rising
solute concentration and the effective pressure for the perme-
ation of solvent through the membrane decreases, thereby in-
creasing the pumping power requirements.
2. Concentration polarization decreases the quality of the perme-
ating solvent (water), since the increase in the solute concen-
tration at the membrane surface increases the transport of
solute through the xaembrane,
3. Die decrease in permeate quality may adversely affect the
membrane and cause accelerated deterioration; concentration
polarization could aggravate this effect.
k. Concentration polarization may result in precipitation
of marginally soluble solutes, thereby causing scaling
on the membrane surface and resulting in a loss of
solvent permeation.
The first technical papers describing results of this project1*5 pre-
sented data experimentally derived on the linear velocity required to
process various effluents in one-half inch inside diameter tubular mod-
ules. Velocities necessary to minimize concentration polarization were
on the order of 35 cm/sec at 7 grams solids per liter, 60 cm/sec at 36
g/1, and 80 cm/sec at 95 g/1; or in all cases Reynolds numbers in excess
of 5000,
30
-------
A third paper6 related the viscosity of the various process streams to
Reynolds numbers for various solids levels and temperatures indicated
that turbulent flow was considered to occur at Reynolds numbers above
1*000, A velocity of approximately one foot per second (30 cm/sec) should
be .sufficient to produce turbulent flow in the temperature range of
25-35°C for solids concentrations of about:
2 percent or less for neutral sulfite semichemieal white water.
3 percent or less for ammonia-base acid sulfite and second-stage
kraft bleach effluents.
k percent or less for calcium-base acid sulfite liquors or wash
waters in 0.5 inch ID tubular nodules.
For higher concentrations» a velocity of one foot per second may or
may not be in the turbulent range, depending upon the temperature and
viscosity of the process feed.
A marked decline in flux with increasing feed pH was observed in early
laboratory'trials. As a general rule, a change in the pH away'from
the normal level of pH for the waste stream resulted in a decrease in
the transport of water through the membrane. This did not appear to
be due to membrane changes per/se, as there was no significant difference
in the flux rate with a water feed between the pH of 3 and 8.
The pH change necessary to avoid hydrolysis of cellulose acetate mem-
branes increased the possibility of the formation of insoluble materials
in some feed streams. In these cases the adjustment of the pH in efflu-
ents containing suspensoids in the colloidal, semiflocculating, or other
dispersed states apparently resulted in coating of the membranes and
a loss in flux rate.
Little, if any, effect upon rejection ratios was apparent in changing
the pH for most of the wastes under investigation, but the effects were
substantial and quite important in the case of evaporator condensates.
Neutralization of low molecular weight volatile acids as acetic acid
converted these to larger molecular weight salts that were well rejected.
Pressure '
Although the equation:
F - aE (Ap-Tr) (5)
where
Pw - rate of water permeation, in gfd
a = membrane area, sq ft
31
-------
E = membrane constant
Ap = pressure differential across membrane, psi
TT = osmotic pressure of process steam minus product water, psi
generally holds for the overall processing of the pulping waste streams,
there are other pressure related factors which must be taken into account,
such as membrane compaction which may produce changes in the rejection
ratio, as well as in flux rate.
The effect of pressure on the flux rates, while processing bleach efflu-
ents, was the subject for study early in this project and was found
to deviate from the usual directly proportional relationship. In the
case of studies with the spiral wound modules (Table 7) data indicated
a deviation from linearity which approached a 2:1.5 relationship between
pressure and flux.
Some of this deviation was traced to membrane compaction while processing
distilled water with old and new tubular modules (Fig. 3). In later,
carefully controlled studies with a large number of l8-tube modules,
this was shown to be due to an almost linear decrease in the Membrane
Constant (E) with increases in the applied pressure.
While it would appear from the data in Table 7 that there was an in-
crease in the rejection ratios with increased pressures, this has not
been borne out in large-scale operations to date. After the initial
compaction of the membrane, there appears to be an increase in the
membrane rejection roughly proportional to pressure increases for some
of the effluent constituents (chlorides). However, this relationship
does not extend to all components, such as color bodies in the waste
stream.
Process Modification Studies for Operating Problems
Encountered in Field Studies
After most of the process variables had been identified, their effects
determined, and all possible means for their control had been developed
to maintain the process under adequate control for obtaining good re-
search data, search began for process modifications which would permit
sustained operation for the reverse osmosis concentration of the broad
range of pulping effluents. In some laboratory and field trials un-
expected problems were encountered which raised questions as to the
usefulness of RO for the concentration of certain of these process
effluents.
The laboratory studies have since developed answers to most of these
problems and with suitable modifications in pretreatment or in operating
procedures, it appears that almost all pulp and pauer waste streams
studied can be successfully concentrated on a sustained continuous
basis.
32
-------
TABLE 7
RO PROCESSING AT DIFFERENT PRESSURE LEVELS WITH SPIRAL WOUND MODULES
Kraft Second-Stage Bleach Effluent
Temperature 25-32°C
Operating
Pressure, Flow Rate, ml/mln Chloride Concentration, ing/1
psig Concentrate Product Water Feed Concentrate Product Water
First Trial
w 100 370 10,2 310 310
250 338 33.0 330 90
400 305 51.0 400 80
Second Trial3
100 397 12.8 339 350 10
250 375 30.5 360 9
400 348 46.4 370 8
a Second trial started 24 hours after first trial.
-------
to
18
16
12
10
20
Type 3 Membrane (New)
Type 3 Membrane (Old)
60
80
100
120
160
Hours
s 5
o
1
a
m
To
Type 5 Membrane
lid
•*
20
60
K) 100
Hours
120
Figure 3. Evidence for Membrane Compaction (Distilled Water at 600 psig and 25°c)
-------
Microbiological Slimes
Many of the pulp and paper process effluents subject for study in this
project contain wood sugars and other fermentable carbon compounds
derived from the wood and also other nutrients which allow substantial
microbiological growth. Slime formation can "be a serious problem. Long
holding times or recycle at certain temperatures and pH levels promote
such growth and membrane process systems must be properly designed
to reduce or eliminate this problem.
The RO equipment available for early studies was not designed to that
end and experience quickly showed up these deficiencies. The problem
was especially apparent in the capillary fiber and spiral wound assemblies
when operated on a recycle feed system.
Tubular RO systems capable of being operated with high velocities,
to achieve a flushing action across the surface of the membrane, reduced
the microbiological sliming problem substantially but even so there
was much evidence of unacceptable levels of flux loss due to fouling
with microbiological growths. In certain cases, such as with the module
life studies, where the process liquors were under continuous recycle
conditions for extended periods of time, there were even problems with
slime pads blocking the operation of the pressure regulating valve.
A "ball-flushing" procedure proposed by .Dr. Ldeb of the University
of California was first used in the life studies to maintain clean
membrane surfaces. This consisted of passing a slightly oversize (3A-
inch) ball of soft polyurethane foam through the 1/2-inch ID tubes
of a module at atmospheric pressure.
While this helped to maintain the flux rates at high levels, there
were several difficulties:
The construction of the modules was such that it was
difficult to pass the ball through a number of modules
in series and this limited the usefulness of the pro-
cedure for large installations such as the trailer unit
(387 Modules).
It was feared that the presence of hard sharp scaling
particles on the surface of the membrane might result
in scratching of the active surface of the membrane,
resulting in leaks and loss of rejection.
Later trials indicated that a number of procedures could be
used to reduce or eliminate the formation of slimes in the
process liquor, including:
35
-------
a. The use of biocides such as aUcyldimethylbenzyl-
ammonium chloride (Zephiran) was helpful in systems
which required recycling.
b. Straight-through operation with low holding time and
high velocities permitted slime-free operation of
modules which had been thoroughly clean at the
start of the process period.
c. Operation of the system at elevated temperatures
above (^5-55°C) reduced or eliminated microbio-
logical growth problems in all cases where the
membrane could withstand such temperatures in
sustained operation. Cellulose acetate membranes
are not usually recommended for temperatures above
UO°C and some manufacturers limit use of their
products to 35°C which is ideal for microbiological
growth.
d. The use of chlorination or ultraviolet light for
the sterilization of the incoming feed stream has
been tried with some success to clear, colorless
process streams. Ultraviolet light is, however,
absorbed in colored pulping effluents and can be
expected to have limited use in such cases.
Low BOD Rejection of Low Molecular Weipfot
Ferment able Materials
In the module'life studies, which entailed a feed recycling system
with fresh liquor makeup every other day, a wide range of BODS rejection
values were noted. These rejections were lower than those achieved
in single pass operations with the same feed and module system.
The only readily apparent difference between the two systems was the
time interval between sampling of the feed and product water streams.
. In the single pass unit, the interval between the time that the feed
was sampled and the composite product water sample was taken was usually
not more than three hours. In the recycle system, however, the feed
sample was taken immediately after makeup and the product water was
sampled the following morning to permit equilibration of the process.
Tests with dyes and tracers indicated that at a 3 gpm flow rate and
600. psig operating pressure with the Type 3 modules, the system could
be brought to equilibrium within an hour or less. With this factor
under consideration, a series of BODs analyses were set up on sampling
periods of 1-1/2 hours, as well as the 2^-hour sample. Data in Table
8 indicate a marked difference in the rejection values for the two
sampling periods. The a, b, c, etc., samples, which were taken 1-
1/2 hours after feed makeup, had BODs rejections in excess of 98
36
-------
8
BODs REJECTION ON 18-TUBE MODULE LIFE STUDY-
EFFECT OF SAMPLING TIME ON REJECTION VALUE
BOD5t mg ./l
Sample
Fa
Pa
Fb
Pb
Fc
PC
Fc1
Pc1
• Fd
Pd
Ff
Pf
Ff
Pf
Fg
Pg
Fh
Ph
Fi
PI
'pi1
Pif
PJ '
PJ.
F - feed
P - perinea
Date
9/6
9/6
9/9
9/9
9/11
9/11
9/12
9/12
9/13
9/13
9/18
9/18
9/19
9/19
9/20
9/20
9/23
9/23
9/25
9/25
9/26
9/26
9/27
9/27
te
First*1' Second11
Sample Sample
1785
93
2018
95
2110
80
21k)
3l*0
2085
92
2268
127
2168
300
2210
111
2505
109
21*55
180
21*55
251
2275
128
BODs
Rejection, pe
95
95
96
81*
96
91*
86
95
96
93
78
•9>
First samples taken 1-1/2 hours after fresh feed make-up.
1»
Second samples taken 2^ hours after fresh feed make-up.
37
-------
percent, while those taken the following day (a1, b', c', etc.) were
generally below 85 percent. Sugars readily rejected could be degraded
to nonrejected volatiles.
Fouling by Organic and Inorganic Coatings, Colloidal Suspensoids,
and Crystalline Scale Deposits
Rapidly decreasing flux rates with tubular modules, while processing
volumes (50-100 gal.) of certain wastes, indicated some factor other
than osmotic pressure or membrane compaction might be responsible for
the loss. This was especially apparent when these same effluents could
be concentrated in small volumes without marked flux losses as long
as the flow paths were not plugged.
Data obtained in processing 50-gal. quantities of second-stage kraft
bleach effluent (caustic KBE) using old and new Type 3 and Type 5
modules, are plotted in Fig. U. The flux decline for the older Types
3 and 5 modules, which had lower initial flux rates, was less than
for the new Type 3 membranes, the flux losses of which were significant
in the first ten hours and were still declining after 118 hours of
operation.
Tubes from these modules at the end of the trial had a coating on the
membrane surfaces. These coatings had not appeared, or were not present
in observable amounts, when smaller volumes were processed. Therefore,
some waste streams under investigation probably contained small quanti-
ties of materials (not microbiological slimes) which could coat the
membranes over a period of time.
During several trials with acid sulfite pulp wash waters, pitch deposits,
as well as calcium sulfate scaling of the membrane surfaces were en-
countered. Other process streams also contained materials, either
in colloidal suspensions or of limited solubilities, which if processed
to high concentrations would precipitate and coat the membranes.
Water Washing
In a field trial with a wash water from calcium acid sulfite pulping
of sprucewood large quantities of "pitch" deposited on the walls of
the feed tank in the foam layer produced by the splashing of the in-
coming liquor.
After collecting some of this deposit, a laboratory study was made
of a water suspension of the pitch with a single tube module. This
module was constructed from an old 7-tube module by repositioning the
end caps to provide straight-through flow in a single tube. The flow-
sheet (Fig. 5) details the basic design and modifications of this setup
for processing small volumes of material.
38
-------
to
vo
16
£ 12
-------
Modification 2
Automatic Valve
Modification 1
Manual Valve
i
Pressure
Regulator
nil
Concentrate
Product Water
Reservoir
1-50 gal.
Relief
Valve
Feed,
X
Drain
Valve
Air Vent in
Outside Casing
1967 Model Havens RO
Module with Type 3 (Med.)
or Type 5 (Tight) Membranes
O Gage
03 Pulsation Damper
\/Accumulator
High Pressure
Pump
aA one quart size Greer Olear "bladder type (water service) accumulator.
Figure 5. Equipment Set-up for Reverse Osmosis Processing
with a Single Module
-------
The results (Table 9) for the processing of this reconstituted deposit
in water, show a progressive decline in the flux rate with time. The
flux rate could, however, be partially restored by a brief water washing.
Periodic washes or flushing with clear tap water appeared to be one
route for control of certain types of membrane fouling and to maintain
high flux rates for water transport through the membrane.
TABLE 9
PRODUCT FUUX RATES VS TIME
Reconstituted Pitch From Filters at Consolidated
Single Tube Havens Module - Type 3
500 peig Operating Pressure
Temperature 25-35 C
Product Water Flux, corrected to 25°C
Time, minutes
0
60
120
180
240
300
360
420
1410
Pretreatment for Removal of the ''Fouling1''' Agent
A loss in flux rate was noted while processing NSSC pulp wash water
in studies of the optimization of the fluid velocity across a membrane
at different concentrations. This was traced to micelles (ray cells)
in this waste stream which were 'deposited on the membrane surfaces whe
their colloidal state was destroyed.
Pretreatment studies based on the removal of these materials by the
addition of filter aids and filtration were partially successful in
reducing, but not completely eliminating, flux losses. The addition
of filter aid or carbon to adsorb the fouling agent, without removal
by filtration, was completely unsuccessful in removing the "fouling1
constituents from a kraft rewash water (Table 10).
Ul
ml/min.
27.4
22.3
22.0
23.6
23.6
23.2
23.4
23.4
19.5
gfd
11.1
9.1
8.9
9.5
9.5
9.4
9.4
9.4
7.9
Percent of Initial
100
82
80
86
86
84
85
85
71
-------
TABLE 10
EFFECT OF VARIOUS
REVERSE OSMOSIS PROCESSING OF KRAFT PULP WASH WATEB
Calgon-Havens Single-Tube Unit at 600 psig
Flux Rate
Pretreatment
None
Aerated and
mechanically
stirred'
a
Mechanically
stirred onlya
Aeration without
mechanical
stirringa
None but "hard
pressure-pulsed"
30 sec./hr.
0-Hour»
gfd
15.6
15.6
15-8
Ik.6
22-Hours,
gfd
6.3
10. k
10.8
10.9
Loss,
percent
1*6
32
31
31
1U.3 2.0
13.2 at k6 hr. 9.6
11.9 at 51* hr. 18
aFoam skimmed periodically for 1 hour.
Pressure Pulsing
During the trials with both treated and untreated effluent samples,
we had observed an increase in the flux rates after short periods
of pressure reduction to atmospheric levels (Fig. 6'and 7). Further
development of this "pressure pulsing" technique resulted in a process
modification which permitted sustained operation, even in the presence
of fouling agents. Flux rates were sustained at 85-99 percent of
the initial water permeation rate. Experience with pressure pulsing
in these studies when concentrating pulp and paper effluents was more
favorable and consistent than the experience reported by Kopecek and
Sourirajan7, who favored extended treatment with elevated back pressure
to restore flux rates. Apparently, the pressure pulsing applies mqre
to flushing out of foulants while back pressure treatment relieves
sustained membrane compaction problems.
-------
Heutral Sulfite Semichemlaal White ¥ater
Single Tube Module
"type 3 Membrane
500 psig Operating Pressure
Initial Flux (gfd)
#1 = 12.0 (Celite filtered through Sparkler}
§2 - 11.6 (Celite filtered through Millipore)
#3 = 11.6 (Activated carbon without filtration)
100
90
80
3
70
60
50
I 1*0
30
20
10
Headings after
a - 1/2 hour at
atmospheric
pressure
b - 20 minute
water wash
Effect of
Cleaning
|
357
Hours of Processing
22
Figure 6. Attempts to Remove Foulant with Adsorbants — Effect on Flux Loss
-------
Neutral Sulfite Semichemical White Water
Single Tube Module
Hype 3 Membrane
500 psig Operating Pressure
Initial Flux, gfd
#1 - 11.9 (Variable Pressure Reductions)
#2 « 11.6 (Pressure REJduced to 50 psig 30 sec/30 min)
3 5
Hours of Processing
#1
Reading After Pressure:
a. Reduced to 1*00 psig for 30 sec
b« Reduced to 300 psig for 30 sec
c. Reduced to Atmosphere for 30 sec
Reading Before Pressure Reduced (d)
Effect of
Cleaning
22
Figure 7. Effect of Sudden Reductions in Operating Pressure on Flux Loss
-------
The following curves illustrate the research phases in the development
of "pressure-pulsing" and an attempt to detail some of the variables
that might "be used with different processes to maintain high water pro-
duction during sustained operation.
'Figure 7. The effectiveness of reduction in operating pressures (500
psig) to kOO, 300,, or to 50 psig in maintaining flux rates was studied
by partially opening the bypass valve between the pump and the modules
(the relief valve in Pig. 5) for 30 seconds at regular intervals. This
produced not only a reduction in the pressure, but at the same time
a drop in fluid flow through the modules. Pressure reductions om the
order of 100-200 psig failed to maintain a high flux. Reductions to
50 psig also failed to eliminate declines in the flux rate but did mark-
edly decrease the rate_ of flux loss.
Figure 8. With provision for complete depressurization, periodic pres-
sure reductions (pulsations) to atmospheric pressure levels by opening
the relief (bypass) valve between the pump and the modules, resulted
in maintaining an average flux level between 95 and 103 percent of the
initial flux rate (Curve l). Reductions in the pressure produced by
opening a valve between the modules and the regulating valve (Modifica-
tion 1 on Fig. 5) resulted in pressure reduction, and at the same time
a very rapid increase in the velocity of fluid through the modules as
the accumulator discharged its volume of fluid. While this "washing"
had little apparent effect on the overall flux rate, there was evidence
that the variations in flux rates were greater (Curve 2),
Since flux values plotted in Curves 1 and 2 of Fig. 8 were from readings
after pressure-pulsing, we had no data on the flux rates between readings;
i.e., the "low" values obtained during each 30-minute period. A trial
"was made to determine the fluxes' before and after the pressure reductions .
Besults of this study (Curve 3) demonstrate flux variations between
5-8 percent of the initial value for 30-minute cycles between depressur-
izations and an increase in the magnitude of the variations when the
processing time between "pulses" was increased.
Further studies with automatically timed, rapid opening ball valves
in-line between the module and the pressure regulating valve (Modifica-
tion 2 in Fig. 5) and to, 80,. and 100-minute pressure-pulsing cycles
resulted in maintaining flux levels between 90-100 percent of the initial
during sustained operations (Fig. 9).
Extension of these studies to single and multimodule installations have
resulted in several additional modifications of the pressure-pulsing
procedure. In certain waste streams the pressure pulsing cycle can
be as long as k to 6 hours, while in others a shorter cycle may be re-
quired to maintain optimum flux rates.
During the pressure reduction cycle, a back-flow of permeate was observed,
and this is believed to permit "osmotic" rinsing of the membrane. It
-------
tJeutral Sulfite Semi chemical White Water
Single Tube Module
Type 3 Membrane
500 psig Operating Pressure
Curve 1. Pressure Reduced by
Opening Bypass Valve Before Module
Curve 2. Pressure Reduced by
Opening Bypass Valve After Module
0)
•H
4i
•<-(
S
0)
u
f-l
V
PH
100
90
80
Curve 3. Pressure Reduced by Opening
Bypass Valve After Module (Readings
before and after pressure reduction.)
100
90
80
tx
; Reading before pressure reduction
a Pulsing increased to 60-min cycle
J_
I
I
J
I
79 ±
processing
Figure 8. Effect of Pressure Pulsing at Various Points in RO System on Flux Loss
3 5
Hours of Processing
3 5
Hours of Processing
-------
100
90
80
Neutral Sulfite Semicaemieal White Water
Single Tube Module
lype 3 Membrane
500 psig Operating Pressure
Initial Flux Rates = 6.k gfd
Curve 1« Pressure Seduced to Atmospheric 60 Seconds Every Uo Minutes
Curve 2. Pressure leduced to Atmospheric 60 Seconds Every 80 Minutes
Qanre 3. Pressure Reduced to Atmospheric 60 Seconds Every 100 Minutes
100
90
80
I
5 7
Hours of Processing
22
Figure 9. Effect of pressure Pulsing Cycle Bate on Flux Loss
-------
is necessary to maintain the permeate discharge lines submerged in a
vessel containing sufficient permeate so as to keep the modules filled
during this back-flow cycle.
In additional laboratory and field trials, we found that a rapid depres-
surization of the system (from operating pressure to atmospheric) was
more effective in restoring the flux rate than was a slow pressure re-
duction coupled with flow stoppage; these procedural modifications have
been termed "hard" and "soft" pressure pulsing, respectively. However,
the capabilities of membrane support structures to withstand the added
stress of hard pulses has been a question yet to be resolved. Much
experience has been gained in some four years of testing, but this must
be recognized as a difficult design and manufacturing problem. High
velocity is considered to be at least a partial alternative to hard
pulsing.
Additives
During the period of development of the mechanical method for maintain-
ing clean membrane surfaces» a study was also being undertaken to
chemically block the deposition of "fouling agents" by the addition
of various dispersants and other fouling depressants to the feed stream.
The use of additives, such as cationic polymers, ethylenediamine tetra-
acetic acid (EDEA), aryl alkyl sulfonates and polyphosphates showed
little promise of inhibiting membrane fouling (Table 11). High levels
of these additives at excessive cost were indicated to be required,for
significant reductions in fouling effects.
Chemical Cleaning
Module cleaning procedures, which included washing with chemicals, such
as EDTA, detergents, fractionated liquor products, acid or alkali solu-
tions, and enzymatic laundry detergents did show promise for removing
fouling agents from the surfaces of the membranes (Table 12).
In the concentration of a calcium acid sulfite wash water from 1 percent
solids to 10 percent solids, a precipitation and deposition of calcium
sulfate (or sulfite) was often encountered. "By careful control of the
concentration levels, velocity, and other operating parameters the depo-
sition of this scale material could be minimized or eliminated. If,
however, the membranes did become coated with calcium salts, a wash
with a 1 percent solution of EICEA removed the material and restored
the flux rate if the scale had not been allowed to dry or become heated
to form the dehydrated salt.
In most of the other processing test runs a flush with a 1.5 percent
solution of an enzymatic detergent (BIZ) for 20 minutes, followed by
a rinse with pH 3 water to remove the alkaline detergent, has proven
adequate for restoring the flux.
1*8
-------
TABLE 11
STUDY OF MEMBRANt FOUL1HG WITH A SINGLE-TUBE HAVENS MODULE USING
ADDITIVES TO PREVENT FOULING
Kraft Pulp Kash Water
500 psig Operating Pressure
pH 7.0 Temperature 35-40°C
Hours of
Operation
No additive (control)
1
7
22
Flux Rate,
gfd,
at 4Q
-------
TABLE 12
STUDY OF MEMBRANE FOULING WITH A SINGLE-TUBE HAVENS MODULI
CLEANING PROCEDURES AFTER FOWLING
Kralt Pulp Wash Water
500 psig Operating Pressure
pH 7.0 Temperature 35-40°C
Hours of
Operation
Ho ftetreatment
0.5
2.5
5.5
22
30
46
50
Flux Rate,
gfd
at 40°C
18.9
18.5
17.0
13.1
11.1
8.4
8.4
Washed with 2 Percent Polytergent B-300
52
68
9.4
6.9
Hashed with 2 Percent Versene-100 (BDTA)
69
70
18.5
19.1
No Pretreatment
0.5
22
18.9
9.2
Washed with 1/2 Percent Versene-100
23 15.7
Washed with 1 Percent Versene-100
25 17.5
Washed with 2 Percent Versene-100
27 18.9
Hashed with 1,5 percent BIZ (enzymatic detergent)
29 18.8
Hours of
Operation
Ho Pretreatment
1
22
Flux Rate,
at 40°C
14.4
5.2
Hashed with 2 Percent Aidonolig
(Fraction of SSL)
23 7.4
Washed with 5 Percent Aldonolig
24 11.4
Hashed with 2 Percent Versene-100
26 14.7
Ho Fretreatmgnt
1
22
16.2
7.9
Hashed with 5 Percent Aldonolig
25 12.2
Hashed with 2 Percent Polytergent B-300
27 12.2
Washed with 2 Percent Versene-100
29 16.4
Mo Pretreatment
1
22
15.9
7.1
Hashed with 1.5 Percent BIZ (enzymatic
detergent)
23 15.7
Hashed with 2 Percent Veriene-100
24 15,7
-------
Hypo chlorite Bleach Effluent
A calcium hypochlorite bleach effluent was processed in one of the small
laboratory units using Type 3 membranes in tubular nodules at flux rates
of 6-9 gfd and rejections of 8l percent for solids, 78 percent COD,
and Jit percent for chlorides (Table 13).
TABLE 13
PROCESSING OF HYPOGHLOUTE BLEACH PLAHT BY
OSMOSIS WITH PULSIIG
Havens single-tube module
I^pe 3 membrane
600 psig Operating pressure
Pressure pulsing 30 sec/30 min
Stream Solids, mg/1 COD, mg/1 Chloride, mg/1
Peed 2726 1#1 673
Product water 521 106 172
Rejection, percent 8l 78 7*t
Flux rates = 6-9 gfd.
The ratherilow values for both flux rate and rejections reflect the
higher inorganic salt content having high osmotic pressure. In field
tests, without pressure pulsing, scaling of the membrane surfaces with
a material which analyzed 19-3 percent calcium and 37-^ percent ash
was encountered. While the fouling agent was organic in character,
there was evidence that calcium deposits were aggravating the problem.
This was substantiated by the ease with which the membrane surfaces
could be restored by flushing with a solution of IDTA.
A second trial with the same waste as used for the field test, but with
pressure pulsing, was successful in removing 90 percent of the water
without a narked decline in the flux rates other than the decline to
be expected with the increased osmotic pressure (Table Ik).
Barking Waters
The removal of bark from pulp logs prior to chipping for the making
of pulp requires large quantities of water and results in a waste stream
having a dilute suspension of fine particles and dissolved substances.
51
-------
TABLE lU
RO CONCENTRATION OF A HYPOCHLORITE BLEACH
EFFLUENT WITH PRESSURE PULSING
Single-tube Havens module
Type 3 membrane
600 psig Operating pressure
Pressure pulsing 30 sec/hour
Water
Recovery, Flux, Solids, COD, Chloride,
Stream percent gfd mg/1 mg/1 mg/1
Tap water 1.6.6
Feed 3010 502 677
Product water
1 to 10.1 120 20 206
2 60 8.7 lU8 UO 220
3 90 6.5 157 H3 222
Concentrate 28650 U633 U?22
Tap water 16.1
A test sample of a "barking water which had been clarified by screening
and sedimentation was successfully concentrated by RO to about 2.0 per-
cent total solids and an almost odorless product water with 71 mg/1
dissolved solids, zero suspended solids and no measurable color (Table
15) was also produced.
Concentration to a solids level which could be utilized by burning (or
other disposal) equipment could require the removal of 99-99-8 percent
of the water. Technically this appears possible but the economics of
such high levels of concentration have not appeared favorable. Special
conditions may alter this situation and permit development of practical
applications.
Deinkirig aiid Other Paper Recycling Waste Waters
In the recycling of paper there are process effluents from deinking
of printed paper and cleaning of fibers from coated clay-filled paper,
which have critical disposal problems not yet completely solved. Highly
colored paper machine white waters have been successfully processed
in the laboratory RO systems at high flux rates using Type 2 (open)
52
-------
membranes with unexpectedly high.rejections of solids (90 percent),
COD (97 percent), color (99-*- percent), and phosphates (96 percent) (Table
16). The high phosphate rejection could also be of importance in pro-
cessing other phosphate-containing effluents from the pulp and paper
industry. It is probable that use of the tighter, Type 3 menbrane would
raise the rejection of the phosphates to the order of 98 percent or
better.
TABLE 15
PBOCESSING OF BARKIHG WATER
Spiral •wound module
Type kA-2 membrane
^•50 psig Operating pressure
Water Flux Rate, Optical
Recovery, gfd Solids, Density a
Stream percent at kQcC g/1 at 281 nm
Feed 0.370 1-36
Product water
1 60 7.2 Q.Ote O.l81f
2 80 7.8 0.050 0.219
3 90 8.2 0.071 0.279
Concentrate 1.98 7-20
A measure of the lignin content.
TABLE 16
PROCESSING OF DEINKJNG WASTES BY REVERSE OSMOSIS
Single-tube module
600 psig Operating pressure
Pressure pulsing 30 sec/30 min
Stream Solids» mg/1 COD, xag/1 PO^, mg/2
Feed 3870 22?»0 17-7
Product water 389 123 0.68
Rejection, percent 90 9^5 96.2
Initial product water flux = 16.3 gfd.
Flux range for 8-hour run = 15.7-16.U gfd.
Final product water flux = 2.6,k gfd.
53
-------
Paper Coating Effluents
A study was also made of the possibility of concentrating paper coating
effluents for reuse. A test was undertaken with a starch-filler-bright-
ener material from a coating operation. Titanium dioxide (white) and
clays were the chief components in the feed liquor.
Using the reverse osmosis process, with the pressure-pulsing modification,
concentration of these waste streams with tubular modules was carried
out with manually controlled pressure reductions of 30 seconds per hour
(Fig. 10). Overnight, unpulsed, operation and cleaning steps are in-
cluded to show the flux losses without pressure pulsing and methods
that can be used to reestablish the water transport rates.
Products #3 and #k and a mixture of all four products (#1-U), were suc-
cessfully concentrated with flux rates on the order of 85-101 percent
of the initial values as long as the pressure pulsing schedule was main-
tained and to TO percent of the initial if "pulses" were discontinued
overnight.
Products #1 and #2, which contained large percentages of alum, had fluxes
which decreased throughout the run, even with pressure pulsing at 30
seconds per hour. An increase in the pulsing cycle to half-hour inter-
vals might have decreased the flux declines, but this has not been inves-
tigated.
Evaporator Condensates
Early in 1966 studies were conducted on RO processing of the condensates
from two different evaporator systems. Observations at that time were:
1. Ninety to ninety-five percent of the water could be recovered
from an acid sulfite condensate as a colorless, odorless product with
high electrical resistance (low concentrations of ionizable components).
2. Sixty percent of the water could be recovered with a 3-fold
reduction in both Chemical Oxygen Demand (COD) and Volatile Acids (VOA),
but there was a marked increase in the transport of materials contrib-
uting to the COD and VOA content of the permeate water as the concen-
tration and water recovery continued above the 60 percent level.
3. The product water and concentrate pH and resistance values
indicate that the main constituents of the product waters were organic
acids and those of the concentrate were inorganic salts and ligninlike
materials with "high color" (Table IT).
A substantial program of laboratory study was undertaken with these
condensates since they comprise a major proportion of the BODs-load
remaining in total pulp mill effluents after disposal processing of
the digester liquors has been undertaken. These studies included:
-------
100
Si 9C|
M
Sin
c
Single Tube Module
3 Membrane
P-.
60-
500 psig
Corrected to 25°C
Initial Flux, gfd
Headings After;
• a. - Atmos. Pressure for 30 Sec
"b. - Attnos. Pressure for Additional 30 Sec
c. - ^fe,ter Washed
d. - Ball Pliished
100
#L-k (Mixed)
Figure 10.
5 T 9
Hours of Processing
Processing of Cpating Effluents by Keverse Osmosis
22
-------
TABLE 17
REVERSE OSMOSIS PROCESSING OF EVAPORATOR CONDENSAfES
Spiral Wound Module
SA-2 (Old-Dense) Membrane
450 pslg Operating Pressure
Product
Water fotal
Recovery, Volume, time,
percent liters hours
Resistance,
ohm-era
flux,
Total
Total Grams Grams
pH Vol. Acids COD
330
4.0
299
790
60
75
99
one.
60
75
99
1
18.9
24,3
35
3158
2973
2652
38
3.0
3.0
2.9
4,0
4.1
4.2
5.2
96
--
139
15
280
395
547
41
220
4.0
341
394
60
70
95
60
70
95
23.3
27.3
40
2431
2323
2062
2.8
2.8
2.8
3.6
3.6
3.8
100
155
208
107
143
214
Cone.
66
3,8
38
115
1. The effects of operating pressure on rejection (Table 18).
2. Stripping to remove the sulfur dioxide followed by processing
with and without pH adjustment (Table 19).
3. The adjustment of the pH to 7-0 and processing at four levels
of solids concentration (Table 20).
k. Processing of acetic acid solutions in a hollow nylon fiber
module (above studies were with tubular systems) at higher
pH levels than could be used in the cellulose acetate designs
(Table 21 - Fig. ll).
5. And finally sustained operation with the spiral wound modules
for continuous concentration of an acid sulfite condensate at
pH 6.0-6.8 (Table 22).
56
-------
TABLE 18
REVERSE OSMOSIS PROCESSING OF ACID SULFITE EVAPORATOR
CONDENSATE AT TW3 PRESSURE LEVELS
Seven-Tube Module
Type 3 Membrane
Water
Recovery, flux,
Sample percent
450 psig (pH Ad Justed to 5.0 wlth NaOH)
Feed
Product Water
1 40
2 60
3 80
Concentrate
250 psig (pH Adjusted to 5.0 with NaOH)
Process Feed
Product Water
1 40 3.0
2 60 3.0
3 80 2.4
Concentrate
Volatile
Acids
8/1
14.7
15.6
4.17
4.44
4,29
32.20
pH
5.0
5.5
5.7
5.5
3.88
4.06
4.17-
44.84
4.0
3.9
3.0
5.4
4.9
3.8
3.8
3.9
5.2
57
-------
TABLE 19
REVERSE OSMOSIS PROCESSING OF MAGWEFITE EVAPORATOR
CONDENSATE
Spiral wound module
Type l*-A-2 membrane
1*50 psig Operating pressure
Sample
As Received
Process feed
Product water
1
2
3
Concentrate
Water
Recovery,
percent
60
80
90
Flux,
gfd
6.0
10.9
12.2
Solic
8/3
8.62
7. 8U
7.51*
7.89
11.28
pH Adjusted with NaOH
Process feed
Product water
1
2
3
Concentrate
Steam Stripped only
Process feed
Product water
1
2
3
Concentrate
Steam Stripped & pH
Process feed
Product water
1
2
3
60
80
90
60
80
90
Adjusted with
60
80
90
5-0
6.8
7-5
11.7
11.5
12.2
NaOH
6.U
6.8
6.1
11.55
7.U7
7-39
8.57
29.UO
It. 51
1*.29
U. 30
1*.70
5.20
Solids, Volatile Acids,
6/1 PH
5.11
It.11
It.02
It.97
10.69
l*.66
2.20
2.21*
3.57
12.61*
I*.25
3.85
3.89
l*.2l*
6.22
1*.76
3.36
3.71*
5.18
1.65
1.75
1.75
1.72
1.58
6.61
6.69
6.71
6.70
6.65
2.85
2.82
2,85
2.82
2.78
Concentrate
lit. 13
-------
TABLE 20
HEVHRSE OSMOSIS COMCEHTRAHOJI OP AS ACID SULFEEB EVAFOMTOB
COHBEHSAIE OTTH pH ADJUSIMSKD OP IBS FSTO
l8-Tube module
£00 psig Operating pressure
Rejection * 1 -
fype 3 Membrane T,
Concentration,
g/1 solids
a
14
19
28
pH
Feed
2.7
4.8
6.1
3.2
4.3
7-0
3.0
4.8
6.5
3.5
5.4
6.8
Solids
0.27
0.57
0.70
0.45
0.57
0.75
0.53
0.63
0.70
0.67
0.74
0.72
OD
0.66
0.55
0.43
0.8l
0.66
0.62
0-73
0.69
0.57
0.69
0.66
0.54
BOD
0.26
0.45
0.58
0.39
0.53
0.63
0.43
0,55
0,67
0.69
0.72
0.68
COD
0.19
0.43
0.57
0.41
0.53
0.60
0.49
o."59
0.62
0.54
0.63
0;59
tt,
0.34
0.34
0.34
0.57
0.57
0.48
0.63
0.67
0.65
0.72
0.6l
0.63
Conduc-
tivity
0.58
0.65
0.70
0.72
0.69
0.61
0.78
0.75
0.73
0.80
Q.?4
0.75
Volatile
Acids
0.07
o.4o
0.65
0.22
0.47
0.68
0.23
0,50
0,68
0.26
O'.6o
0.56
Solids
0.45
0.8l
0.96
0.63
0.84
0.97
0.78
0.91
—
0.80
0.96
0.97
OD
0.79
0.74
0.90
0.86
0.85
0.87
0.84
0.84
0.80
0.87
0.88
0.76
?pe 5 Membrane
BOD
0.42
0,62
0,82
0.54
0.73
0.87
0.63
0.79
0.86
0.82
0.83
0.81
COD
0.37
0.62
0.81
0.54
0,71
0,84
0.62
0.77
0.84
0.69
0.86
0.85
HH,
0.72
0.78
0.72
0.74
0.80
0.70
0.82
0.83
0.77
0.86
0.86
0.79
Conduc-
tivity
0,69
0,95
0.00
0.91
0.96
0.97
0.92
0.97
0.98
0.95
0.98
0.99
Volatile
Acid's
0.26
0.70
0.95
0.4o
0.70
0.97
0.4o
0.73
0.98
0.45
0.85
0.98
8
14
19
28
I/I
19
28
Pt
Concentrations in Feed
aag/1 ing/1 dhm cm
7500
no
280 9700
500 10300 16600
690 11700 23000
18
35
60
72
158
90
76
52
efi
4.2
^
6.3
-------
TABLE 21
REJECTION OP ACETATE IONS AT VARIOUS pH LEVELS
Hollow fiber module (Mo. l)
Acetic acid solution — 5 g/1
pH Sample
3.0 P-l
P-2
P-3
P-l*
P-5
P-6
h.o P-l
P-2
P-3
F-U
P-5
P-6
5.0 P-l
P-2
P-3
p.I»
P-5
p-6
6.0 P-l
P-2
P-3
P-l*
P-5
P-6
8.0 P-l
P-2
P-3
P-l*
P-5
P-6
11.0 P-l
P-2
P-3
P-l*
P-5
P-6
Percent Conversion
Percent Rejection = 100 11 -
Conversion,
percent
12.8
13.0
13.2
13-5
13.8
m.o
13.1
13.2
13. U
13.5
13.6
lU.l
12.9
12.9
13-0
13.2
13.1
13.2
11. T
11.8
12.1
12.0
12.0
11.9
11.9
11.9
12.1
12.1
12.2
12.2
13.0
12.6
12.6
12.7
12.8
13.0
ml permeate x 100
Reje
Bange
23.3
12.2
8.7
11.0
10.7
13. fc
28.0
20.9
21.1
21.1
20.5
21.1
57-2
1*7-5
1*7.3
1*6.8
1*6.8
1*7-3
71*. 6
73.2
73.0
72.1
73.0
72.7
85. k
82.0
79-8
82.0
83.8
81.3
81*. k
82.3
83.2
89.0
81*. 1*
88.1
Rejection, percent
Average
mlpermeate + ml reject
11
21
UT.
73
82
85
, Feed
Pressure
360
360
360
360
358
358
363
362
361
360
362
362,
365
366
362
361*
365
361
360
360
360
360
360
360
360
362
361
361
361
31*1
31*1
C • Cone, permeate Cf
Cone, feed
60
-------
90
80
70
60
S 50
c
o
-------
TABLE 22
REVERSE OSMOSIS PROCESSING OF A CALCIUM-BASE ACID SULFITE
EVAPORATOR CONDENSATE AFTER pH ADJUSTMENT
Spiral wound modules
pH Adjusted with NaOH
600 psig Operating pressure
3 gpm Flow rate
System Flow
Pattern
Straight
through at
pH 6.7
Recycle for
concentration
at pH 6.4
Recycle for
concentration
at pH 6.0
Percent Solids
Concentration
0.1*8 to 0.51
0.55 to 2.46
0.39 to 5-59
Percent
Water
Removal
5.7
78.0
93.0
Rejection, percent
BOD5 VOA COD
77-4 95.0
.U
98.6 95.6 91-3
92.8 96.0 96.U
Flux
Rate,
gfd
8.3
2.1
The studies were conclusive in showing membrane systems available for the
study on this project were not capable of rejecting low molecular weight
organic acids such as acetic, inorganic acid volatiles such as SOj
and also organic neutrals such as methanol. Conversion of the acid
volatiles to salt forms altered the rejection capabilities and this
neutralization route remains an interesting possible method of effec-
tively processing pulp mill evaporator condensates by membrane concen-
tration. However, the expense of neutralization is high, and this
membrane route is unlikely to be widely used for disposal processing
but rather for recovery of values whenever a market can be developed
with a return sufficient to balance the neutralization and subsequent
concentration costs in competition with synthesis of acetic acid and
related volatiles from petroleum sources. Other, more economic routes
to disposal processing of this waste flow are being sought and the future
of not a few of the older acid sulfite pulp mills hinges on successful
outcome of this overall research effort toward practical routes for
disposal if not for utilization.
62
-------
SECTION VI
OF THE LARGE-SCALE TRAILER UNIT FOE OSMOSIS
OF PULPING EFFLUENTS
The review and discussion in this section extends upon a prior publica-
tion7 describing the development of the engineering design for large-
scale applications of reverse osmosis (EO) as a method, for concentrating
dilute wastes without phase change. In this modification of BO for
pollution control and effluent treatment, the water contained in rela-
tively dilute process streams is forced under pressure through a membrane
in conventional RO modules originally designed for salt water conversion
applications. However, the objectives for this demonstration program
called for a .much higher degree of water recovery ranging from 70 percent
to more than 95 percent of that contained in the dilute feed waters
and also for much higher concentrations of the dissolved solids rejected
by the membrane system than is normally practiced in salt water and
brackish water conversion. This points to a growing need for designing
membrane systems especially to meet the requirements for concentrating
systems. This is indeed taking place, and much ingenuity is apparent
in equipment designs becoming available from suppliers. The choice
of tubular membrane systems from among the available RO conformations
has been described in the foregoing pages as a necessary first step
in planning the large-scale demonstration studies and design of the
large field unit.
The degree of feed pretreatment required, the quality of the concentrate,
and of the water permeate become matters of added concern, since these
products are to be subject for reuse or for final disposal without pol-
lution of the water, land, or air environment.
Considerable time and effort in the Effluent Processes Group laboratories
of The Institute of Paper Chemistry has been directed to investigation
of the potential applications of the' RO process for concentrating the
dilute effluents of the pulp and paper industry as a preliminary justi-
fication for these more detailed studies on design of the field unit.
Those preliminary studies developed.the basic information- for-planning
the program of application studies for the research and demonstration
grant as described in the various following sections of this report.
The data developed in this section were the' next step toward design,
construction, and operation of the large trailer-mounted field demon-
stration unit for processing specific waste streams at individual pulp
mills at feed rates in the range of 20,000 to 100,000. gallons daily.
The development of the principal and more important design criteria
followed studies in several areas of engineering evaluation.
Selection of specific types of RO equipment.
Module configuration
63
-------
Type of membrane
Methods of feed -pretre at merit
Development of membrane flux rate and rejection parameters for unit
design
Hydraulic parameters for design of the system
Velocity of flow across the membrane
Pressure drop
Final development of the flow sheet and detailed design of the unit
OP OSMOSIS EQUIPMENT
One of the first questions to be answered when designing a reverse osmo-
sis unit concerns the selection of module configuration and the type
of membrane to be used. The selection of module design is an important
factor from the standpoint of membrane fouling whereas selection of
the type of membrane is affected by the nature of the effluent water,
the required pretreatment of the feed; and the rejection characteristics
of various components contained in the feed. The selection of module
design and of the type of membrane are discussed separately below.
Module Design
The configurations of BO equipment available at the time of initiating
these design studies included:
Tubular Design
A number of suppliers have for the past 10 years been progressing in
development of tubular membrane support configurations based upon tubes
constructed of molded or extruded'resins, resin-bonded glass fiber,
solid or perforated metals, and more recently a design with tubes sup-
ported and surrounded in resin bonded sand or other aggregates. The'
tubular systems provide for clean hydraulic flows at high velocities
which have proved to be advantageous for concentrating effluents con-
taining suspended particles. Colloidal solutions and true solutions
which throw down precipitates or crystalline scale-forming deposits
on concentration are advantageously processed in the tubular design.
Tubular designs may have ratios of membrane surface to module volume
in the 50 to 200 square feet per cubic foot range, which is relatively
low in comparison to other configurations and handicaps the competitive
development of low cost modular systems accordingly. However, this
disadvantage can be outweighed by the'basic hydraulic operating credits
of a well-designed tubular system.
61*
-------
Spiral Wound Modules
The "jelly roll" configuration of the spiral wound module design provides
much more membrane area per unit of module volume than can be had for
tubular designs having the usual inside diameters in the range of lA
to ,2 inches. This advantage of the spiral wound design can be on the
order of 10 to 100 times that of the tubular equipment, and the cost
advantage is significant and well developed for processing clear brackish
waters and the like which are free of suspended matter and which do
not throw down precipitates or crystalline deposits during concentration.
However, the experience in preliminary studies of pulping effluents
as previously described indicated the tolerance for even small quantities
of suspended matter to be critically low for the spiral designs available
for laboratory and field phases of this study in 1968-69.
Hollow Fiber Configuration
This configuration is still more advantageous in terms of the high ratio
of up to 10,000, square feet of membrane surface per cubic foot of module
volume. But the even greater sensitivity of the bundles of microfibers
to plugging with suspended matter has proven to be a substantial road-
block to large-scale application studies to concentration processing
of the dilute industrial effluents covered in this report.
Type of Membrane
The commercial production of RO equipment at the time of design and
construction of the trailer unit was still in relatively early stages
of development. The membrane modules available on the market were de-
signed for operating in the range of 500 to 1000 psig, but relatively
little data were yet available in terms of module life and performance
based upon sustained continuous operation.
RO equipment commercially available has been almost exclusively based
upon two types of membranes — cellulose acetate and nylon, and upon .
various modifications in the:formulation and casting of those membranes.
The rate of hydrolysis for cellulose acetate, as reported in the liter-
ature, has been reported to be lowest at a pH of about k.5. Deteriora-
tion of cellulose acetate membranes accelerates as .the pH ranges above
or below this point. -For most applications, a lower pH limit of 2.5
to 3.0 and an upper limit of: 7.0 to 8.0 seems to be.acceptable. Nylon
membrane systems can stand feed solutions of pH 2.0 to 12.0,. and'.there-
fore might be selected for processing solutions in a broader'range of
applications to reduce the need and the expense for neutralization of
acidity or alkalinity.
Operation at the pH level of the feed solution as received may .reduce
the problems arising from membrane, fouling. This reduced level' of mem-
brane fouling seems to.be due to the: fact that most- feed, liquors .tend
to be in. a, state of equilibrium at.a particular pH, but if the pH is
changed, the, chances for precipitation, scale formation, and the like,
increases and may result in fouling of the membrane.
65
-------
In addition to the pH effect, the deterioration of the membrane is also
affected by higher temperatures of solution. For both cellulose acetate
and nylon membranes, an upper limit of *tO°C has usually "been specified
by the membrane manufacturer.
The flux rate and rejection characteristics of a membrane can be varied
over a fairly vide range, depending substantially upon the "tightness"
of the membrane. These operating characteristics may in the case of
cellulose acetate be substantially controlled by the degree of heat
treatment and in the solvent formulation used in fabrication of the
membrane.
In the pulp and paper industry, most of the dilute wastes contain low
molecular weight organics, such as wood sugars, which affect the BOD
and COD. Removal of a minimum of 75 percent of those components requires
a fairly tight membrane. Different types of 1/2-inch tubular modules
having cellulose acetate membranes were tested. Those with rejection
performance similar to Havens Type 3 membrane proved to be the most
satisfactory of those available in 1968-1969. Tests covered the range
equivalent to Havens coarse Type 2 to very tight Type 5 membranes. Other
manufacturers had membranes with a similar range, but with proprietary
differences in their individual numerical designations of types and
grades. Specifications established for the trailer unit called for
performance equivalent to 7000 square feet of membrane with moderately
tight rejection capabilities equivalent to the No. 3 membrane.
PRETREATMENT OF FEED
Pretreatment of feed is important from the standpoint of membrane-module
life, and in maintaining steady long-term flux rates with minimum loss
of operating time for shutdowns and washups of the system. For the
tubular cellulose acetate modules finally selected, planning called
for minimum levels of pretreatment to maintain temperatures in the range
of 20 to hO°C and pH levels between 3.0 and 7-0. Gross quantities of
pulp fiber occasionally escaping into the feed and which could accumulate
and plug check valves and gages were to be taken care of by passing
the feed through a 100-mesh screen. This may be finer than actually
necessary since a more coarse JlO-mesh appeared fully adequate in later
tests during Demonstration No. 2 on NSSC liquor. A side hill screen
proved adequate in some of these studies but, in others, occasional
heavy slugs of fiber in the feed required use of a vibrating screen
loaned especially to this project by Sweco. This commercial unit (k
foot diameter model Sweco screen) provided excellent and fully adequate
service throughout the series of demonstrations, and greatly contributed
to reduction of operating problems.
Other suspended material in the feed has a strong tendency to foul the
membranes, thus resulting in reduced long-term flux rates. Efforts
in these studies were made to stay away from expensive ^pretreatment
operations such as reported for waste treatment in other industries
in which elaborate pretreatment systems have been employed ahead of
66
-------
RO processing. lime and ferric chloride treatment, addition of poly-
electrolytes, coagulation, filtration, and. air flotation are examples
of such advanced pretreatments. In contrast, pretreatment at all
demonstration sites in this project study consisted of no more than
a 1*0 to 150-mesh screening with pH and temperature adjustment. However,
advantage was taken of the fact that higher degrees of turbulence and
mixing at high velocities across the membrane surface in tubular systems
help in minimizing concentration polarization and fouling effects.
Initial, plans fpr temperature adjustment were based on reaction rates
within the limits of acid hydrolysis above pH 3.0 and of alkaline hy-
drolysis on cellulose acetate membranes. However, additional criteria
were developed on need for maintaining the highest possible operating
temperature consistent with obtaining maximum flux rates through the
membrane and with maintaining temperatures wherever possible at levels
which would help to inhibit microbiological growth and slime formation
on the membrane surfaces. Temperature inhibition of microbiological
growth can be expected above 1|.00C» but attempts to operate above hO°C
were strictly limited by accelerated chemical hydrolysis reactions on
the CA membranes. Later trials have shown nearly all evidence of micro-
biological fouling ceased when operating temperatures were purposely
raised to the k5°C level with the more resistant membranes which have
become available after completing design, construction, and operation
of the equipment used in these studies from 1968 through mid-1971. Mem-
branes suitable for higher temperature operation are being sought.
FLUX RATE AND REJECTION CHARACTERISTICS DURING
MEMBRANE CONCENTRATION PROCESSING
The initial flux rate and rejection characteristics of a membrane with
specific cellulose acetate formulation depend upon the thickness of
the active layer and the porosity of (the membrane as controlled by the
solvents used in casting and the degree of heat treatment in the fabri-
cation of the equipment. The flux rate performance of a membrane
during .concentration processing of any dilute waste is governed by the
osmotic pressure and membrane fouling characteristics of the solution. .-',-
The osmotic pressure of a solution increases significantly with increase
in the contents of low molecular weight inorganics. The rate of mem-
brane fouling is low for clean streams, and it becomes severe for solu-
tions containing substantial quantities of finely divided suspended
matter in the nature of colloidal s us pens oids . Fiber and other larger
aggregates of suspended solids are of less concern in tubular systems
if they do not accumulate and plug orifices in valves and pumps. . The
rejection ratio is an important parameter in the design of a reverse
osmosis unit for the purpose of concentrating the feed over a wide range
or for chemical fractionation systems. For membrane concentration
processing, the rejection ratio is defined as the ratio between the
concentrations at both sides of the membrane at a certain spot, and
it is important here that both of these concentrations should be con-
sidered from the standpoint of being measured instantaneously.
67
-------
Percentage recovery of water and chemicals in a reverse osmosis system
having a given membrane area is determined from the overall average
values of flux and rejection ratio obtained during the concentration
run. The usual concept of these "overall average values" can "be mis-
leading since they are generally calculated as the arithmetic average
without taking into account the removal of water taking place as the
concentration process progresses. In a multistage RO unit designed
to concentrate dilute effluents by 10 to 20 times the concentration
of the original feed liquor, a large portion of the water is removed
in the early stages of concentration, and therefore it is important
here to consider the percentage removal of water at various concentra-
tion levels in the determination of the overall average values . The
overall average value can be mathematically determined as follows:
I Wi Xi
Xn, = i" (6)
UA V TT-
I Wl
i
where
X = overall average value of X
Xi = average value of X in the i-th concentration stage
Wi = percent removal of water in the i-th concentration stage
= summation of all the concentration stages
In the above expression, X can represent a number of important parameters,
such as flux rate, rejection ratio, or solids concentration in the per-
meate. It is important that the calculations should be carried out
over a large number of concentration stages in order to determine a
true overall average value.
The chemical recovery ratio in a concentrating system represents the
amount of dissolved components recovered in the final concentrate. This
is of importance in case the concentrate contains valuable chemicals,
or if retention of pollution materials in the concentrate is desired.
It is calculated as follows :
TT~) (7)
Fi
where
R = chemical recovery ratio
Fj_ = feed flow (input)
F2 - concentrate flow.(output)
GI = concentration of feed
C2 = concentration of final concentrate
R = rejection of the membrane
68
-------
A quick estimation of overall quality of total permeate and its accept
ability. to meet 'antipollution quality standards or, alternately, its
suitability for reuse can lie expressed as follows:
- (8)
An example -of the kind of information which can be easily obtained from
the above equations is presented in Table 23. The calculated data show
the final concentrate volume (F2/Pi) ; the water to be removed (F]_ -
F2^/F2' the recovery of chemicals by the system T?; and the average con-,
centration of the permeate, Cp obtained at various rejections when con-
centrating 1 to 10.
TABLE 23
PERFORMANCE OF REVERSE OSMOSIS AT VARIOUS REJECTION RATIOS
R, rejection in percent 99 90 70 50
Cg/C^ (Concentration Factor) 10 10 ; 10 10
F2/F1 (Volume Reduction) 0.10 0.08 0.0*1 0.01
(F^FgJ/Fg (Water Removal) 0.90 0-92 0.96 0.99
R" (Chemical Recovery) 0.98 0.77 0.38 0.10
C /C^ (Concentration Permeate) 0.02 0.2U 0.6~3 0.90
A number of concentration runs were made using a small-scale unit on
various pulping effluents prior to the design of a large-scale trailer
unit. The results of these small-scale studies indicated that an overall
average flux rate of 7-5 gfd Gould, at that time, be obtained -with the
Havens Type 3 membranes then available when concentrating a 1.0 percent
solids feed to 10 percent solids concentrate at 25-35°C and 600. psig
pressure. The average rejection of solids for these modules were found
to be above 90 percent, whereas BODg rejections ranged from 75 to 90
percent. It was also noticed that the membrane rejection of various
components did not change significantly with increase in concentrations
for most of these pulping effluents .
SYSTEM HYDRAULIC PARAMETERS
The next step in the design of the large-scale trailer unit was to study
system hydraulic parameters , mainly velocity and pressure drop. The
importance of maintaining higher velocities across the membrane surface
69
-------
to reduce concentration polarization and fouling effects has "been empha-
sized previously. In addition, .Telocity and pressure drop parameters
play an important role in the determination of the "best module configur-
ation for a large, multistage concentrating system. Both of these param-
eters are discussed separately below.
In order to OTercome the limitations occasioned by reduction in the"
long-term flux rate and the rejection characteristics of the membrane
it is necessary to maintain adequate degrees of turbulence and mixing
necessary to minimize concentration polarization and fouling of reverse
osmosis membranes.
In the design of a demonstration unit the following assumptions for
the minimum velocity were used:
Concentration gal./min ft/sec
less than 2 percent solids 1.5 2.5
2 to 6 percent solids 2.0 3-3
6 to 12 percent solids 2.5 k.I
This concept of higher velocities is necessary in keeping the" membrane
clean and thus maintaining economic levels of long-term, steady flux
rates. However, it should be kept in mind that although higher velocities
across the membrane can reduce the physical accumulation of solids near
the membrane, it cannot alter any chemical or electrical affinity which
the solids may have with the membrane. Special problems of pretreatment
may arise if such affinities are apparent.
Die concept of minimum velocity at various solids levels as concentration
proceeds is also an important factor in the determination of the best
arrangement of modules. The maximum number of modules which can.be
put in parallel and the minimum number that have been put in series
for each concentration stage is determined on the basis of the minimum
velocity required. For given" pumping capacities of the pressurizing
and booster pumps, the number of modules in parallel decreases' propor-
tionally while designing for an increased velocity, while the number
in series increases proportionally.
Pressure Drop
The pressure drop in a system depends on a number of factors . First
of all, it depends on turbulence and frictional losses related to the
velocity in the module. We found for a module of 1968 Havens design
having eighteen 1/2-inch tubes, a tube length of 87 inches, and a total
membrane area of 17 square feet, that the total hydraulic resistance
could be formulated as follows :
70
-------
AP = A Y' (-9)
where
&P = frietional pressure drop, psi
A = 2.2
V = Telocity, ft/sec
In this particular design of module, a substantial percentage of'pres-
sure drop was found tc take place in the seTenteen l80° close return
turnarounds»
Pressure drop is one of the'important design factors of concern in the
selection of the number of modules to be used in series. The loss in
volume as concentration proceeds in. a series of tubes or modules is
also'to be considered. The maintaining of a desired ndnlmum level of
velocity at any point in the system'when connecting a large mmber.of -
modules in series requires' control of the total flow rate going into
the' system, allowance for the loss in volume, and also' consideration
of the pressure loss occurring in proportion to the number of modules'.
Under such conditions, it becomes necessary to add-booster'-pumps to
overcome-the pressure drop. Therefore, one should optimize the''pressure
drop against the total flow rate, while selecting the number of modules
to be connected' in series.
•PETAL OF THE -
The study of various design features,; as discussed in this- report led
to establishing'Specifications for.the finalj design of a .large-scale
demonstration unit,- .Pie specifications were submitted'.with Invitations
to "bid to tea concerua np|iye in supplying membrane equipment, Tro
quotations'were received'." Die" unit accepted" from these'bids was designed
in detail by the Havens International staff,' and built into a UO-foot
trailer'for mobility, as-shown-;in Fig. 12, .13, -lJ*.. All;.the hardware,-
except for the modules, was designed for handling up to 1QO.,QOQ gal./day
and a maximum-operating-pressure/-of,"1000s~ psig. The original equipment - '
included 396 Calgon-Havens 18-tube modules having 6f32^ square feet--of '
membrane, plus a reserve of 20 modules. The unit as delivered"had a
nominal capacity of 50,000 gal./day at -an overall average flux rate
of 7.5 gfd at 25-35°C and 600 psig pressure. . The maximum allowed oper-.
ating pressure at-that time'for the Calgon-Havens modules'supplied'was
600, psig, although the unit was required to pass acceptance tests under
full load at 1000, psig. Much credit must be given to the" staff of
Havens International for the' great amount of detailed engineering going
into final design and construction of this unit \sider the pioneering
conditions established in the' specifications.
figure 15 provides' the flow diagram for the' large-scale demonstration
unit along with the arrangement of modules in each concentration stage.
In order to concentrate 50,000 gallons per day of dilute feed from
71
-------
3 rtst WIT
Utufry ttwl*
Figure 12. Photo of Trailer-Mounted RO Field Demonstration Unit in
Operation (Field Test No. l) at Appleton Division, Ca-Base
Acid Sulfite Pulp Mill of Consolidated Papers, Inc. Feed
Storage to Trailer Itoit in U5,000 Gallon Tank at Left
-------
... :y:-r™
Figure 13- Photograph of RO Trailer. Banks of Tutmlar Modules Shown
as Mounted in Rear Half of Trailer. Similar Banks and Full
Access Doors Mounted in Front Half. Access to Control Panel,
Pumps and Valving Through Center Side Door
73
-------
Figure ik. Photograph of Interior of Trailer-Mounted Field Demonstration
Unit. One of 210 Ball Valves Being Operated to Isolate Individual
Manifolded Section of 6 to 15 Modules. Trailer Equipped with 38?
Modules Designed for Continuous Operation While Cleaning or
Replacing Modules Within Individual Sections
-------
Flow Diagram: Mode 1
Feed \- ^*"1 1 „
-^ — I t *
Main Pump
la
n>
i
* rv
* LA
PUBID A '
•n
Pump B
Rasp C
Concentrate
Back Pressure Valve
Jl
How Diagram: * Mode 2
r— — —
Feed [**~ — i
|_ | '* . la ID "
Main Pw?> " • h
— — . -i t i
-. . *
Main Pimp;
Max. Capacity 70 gpm
Operating P I*500 J>siS
4P - 1000 psig
Bank la Ib j
Modules Parallel 1-8 12 i<
Series 5 . 3
. . rotesl . '• 90. 36 3
-f4i ;- • "" '• -::-'-- :
M in J
1 ' " " i
».:. t f hd Conceatrate
Back Pressure Valve
Pun?! A: Puaip B: ' Pump C:
50 gpm 100 gpm 50 gpm
720 psig 720 psig 720 psig ..
50 psig 50 psig 100 psig
i in; iv v
jaw. 21
3 3.. 3 3 . •
3 2^- 1*4-4 63
?igure- 15. Flow Diagram of the Large-Scale Beverse Osmosis Unit
-------
1.0 to 10.0 percent solids under straight-through operations, it was
necessary to put quite a number of modules in series. Only a small
number of modules could be put in parallel due to need for maintaining
velocity with limited pumping flow rates. Under the setup of Fig. 15,
namely, 20 modules in series in 5 concentrating stages, a large pressure
drop of 1200. psig at 3-5 ft/sec was anticipated. This problem could
have been solved by adding twelve 100-psig booster pumps. But this,
of course, was impractical and not economical, and therefore the system
was designed which could be operated with a minimum of two or three
recycle loops. Thus, a total of 5 concentrating stages were installed
and the pressure drop at no point in any stage was more than 100 psig
below the maximum operating pressure. In the. case of high feed intakes
(at high flux rates), Banks II and III could'be operated without recycle
(Mode l) to satisfy minimum flow requirements; otherwise, II and III
could be operated on a recycle basis (Mode 2).
The main pump was a triplex reciprocating Manton Gaulin pump driven
by a direct current motor with electronic s.peed control to provide flows
in the range of 17 to 70 gal./min at pressures up to 1000 psig. A pres-
sure accumulator was provided to reduce hydraulic pulsations in the
system to not greater than plus or minus 0.5 percent of the average
system pressure. The recycle pumps were Goulds high pressure centrifu-
gal process pumps. The booster pumps available for installation origi-
nally employed packed sealing glands for the first demonstration and
were used in systems with a maximum operating pressure of 720 psig.
After careful study of operating conditions supervised by Goulds engi-
neers, these pumps were later replaced by'new pumps with mechanical
seals especially designed to operate at 1000. psig. The pumps were
equipped with John Crane packing single balanced mechanical seals placed
in combination with two flushing systems, one at each side of the seals.
All metal components in contact with the process solutions were ,3l6 ...
stainless steel.
The unit was equipped with a pH meter and control system (Universal
Interlock). The unit was also equipped with safety switches for auto-
matic shut down when pH, temperature, or pressure exceeded set limits.
Thus, there was only a minimal need for supervision approximating 1/2
man day per day during normal operation and sampling of the system.
After delivery and startup tests were completed, the trailer unit was
equipped with a "soft" system for pressure pulsing operation. This
was designed to accomplish the periodic cleaning of the membrane surfaces
in accordance with the principles developed in the preliminary studies.
The soft "pulsing" was found to be advantageous in minimizing the mem-
brane fouling caused by colloidal and fine particulate suspended solids
of the liquor stream, especially at lower fluid velocities.
-------
SECTION VII
FIVE FIELD DEMONSTRATIONS
FIELD DEMONSTRATION NO. 1
Concentration Processing of Calcium-Base
Acid Sulfite Wash Waters
This first field demonstration of reverse osmosis as a concentrating
system was conducted on dilute Ca-sulfite pulp wash waters at the Apple-
ton Division (interlake Mill) of Consolidated Papers, Inc. at Appleton,
Wisconsin. The demonstration with the trailer mounted semicommercial
unit followed a substantial laboratory and pilot scale testing program
(Section V). Additional studies for engineering design and optimization
of large installations were conducted later (see Section VIII). The
basic data described in Sections V, VI, VII, and VIII are evaluated
in Section X for developing the apparent economic picture on use of
RO in its present stage of development as a method for concentration
processing of dilute effluents of the pulp and paper industry.
Pulping Operations and Methods of Collecting the
Wash Water Used in This Demonstration
The Interlake mill pulps Canadian sprucewood by the calcium-base acid
sulfite process which was developed more than 100 years ago. The site
on the Lower Fox River in Appleton was first used for a small paper
mill built in 1853. The present sulfite pulp mill was built in 1891
and the digester system was rebuilt in 1929- This is exclusively a
pulping operation and no paper is made. The Mitscherlich horizontal
digesters produce a high-grade, long fiber, market pulp which is subse-
quently fully bleached, for use at other mills in the manufacture of
finer grades of business and writing papers and of glassine papers.
Many of the older pulp mills in the U.S.A. and Canada employing this
Ca-base, batch-type cook are considered to be outdated, have trended
toward becoming noncompetitive against newer continuous methods of pulp-
ing wood, and have been or are being phased out of business. The
difficult-to-solve pollution problems of this acid sulfite pulping process
has accelerated the demise of these older mills.
However, there are exceptions to this trends and continuing efforts
under way by the Consolidated staff at this mill, and also in similar
research, engineering and commercial studies at a substantial number
of other acid sulfite mills, have been directed to modification and
installation of improvements to the basic process for the combined
objectives of improving the economic competitiveness and to eliminate
the pollution problems of acid sulfite pulping. .This pulping process
still can be considered to have substantial advantages in producing
a premium grade, easily bleached pulp, and additionally it produces
high yields and modest dollar values of marketable lignin and
11
-------
carbohydrates from the other half of the pulpwood, a fraction which
is normally burned for low value heat recovery in the newer pulping
processes. However, the economics of complete treatment of all process
effluents for pollution control at this mill remains a formidable
problem.
A large bibliography9 with more than 8,600 references, records a century
of progress (and of many disappointing failures) deriving from a contin-
uing industry program of chemical research, engineering development,
and commercial application of modifications and improvements to solve
the pollution problems deriving from these acid sulfite process effluenta.
No one method of solving all of these problems has been applicable,
but much has been accomplished.
Figure l6 presents a flow sheet based upon the overall Ca-base acid
sulfite pulping process for producing marketable pulp and spent liquor
products, and also shows the various effluents from this process which
relate to the serious environmental problems in maintaining operation
of acid sulfite mills. These effluents include barking and wood prepar-
ation wash waters, dilute pulp wash waters, evaporator condensates,
wood fines, pitch and ray cell effluents from fractionators, and also
bleach plant effluents. Installation of the evaporation plant at the
Interlace mill in 1953 permitted highly effective processing of all
of the strong digester liquors that could feasibly be collected and
led to production of molasseslike concentrates and spray-dried lignin
liquor products which are marketed as dispersants, binders, adhesives,
and the like. About 70-75 percent of the dissolved spent liquor solids
deriving from the wood cook in the digester can be feasibly collected
and processed in this existing system. Development of an economically
feasible route for processing the remaining 25-30 percent of the spent
•digester liquor solids contained in the dilute wash waters has been
the subject for much study by the mill staff, and this trial of the
reverse osmosis membrane concentration system was designed to establish
its possibilities as another alternative route to that end.
A second more detailed flow sheet (Fig. 17) shows the flow within the
pulp washing areas of the Interlace mill in the successive steps of
draining the strong digester liquor and washing of the pulp prior to
final refining in the bleach plant.
All strong digester liquors and early stage wash waters are drawn from
the No, 1 liquor tank for concentration by evaporation and spray drying.
The dilute wash waters of concern to this demonstration derive from
the fifth stage of washing the digester stock in a batchwlse counter-
current washing system. The first four washes are collected and re-
.turned for use in the next washing cycle and eventually go to the
evaporators. The remaining fifth wash is referred to as "cooling water"
by the mill staff. It was found necessary for this demonstration to
run some of the fifth wash to the mill sewer because of its high teuper-
ature and in order to avoid need for installation of an expensive heat
78
-------
\
Sprucewood - 300 tons/day
Evaporation
Plant
T
#2
Vasl
r^
«
WasJ • Mashed
Percent IDS Pulp
>n
Wes
(cooli
tk Wash
raters
.ng vater)
Fract
ate
1
1
1
Bleach Plant
Bleached
I Pulp
Condensate
Acetic Acid
and Other
Volatiles
Spent Liquor Products -
100 tons/day
Fractionator
Effluent
Fines and
Pitch
OOO—,
OOO
Bleach
Effluent
White Water
Pulp to
Market
150 ton/
day
To reuse •<
a.) Pulp washing
b.)
c.)
Sewer
Figure 16. Flow Sheet for Calcium-Base Acid Sulfite Pulp Mill
-------
Horizontal Digester
(Mitscherlich System)
Digester
Liquor
Liquor
Tank
\
—\ ' ' J
#2
Recovery
Tank
#3
Recovery
Tank
1
r
:
i
RO
Feed
Tank
General
White
Water
Chest
12-15 Percent
TS to Evaporator
Spray Dryer
Market
f
i
r
r
To RO Pretreatment
System .
rH
/
Washing Procedure:
(A) Blow strong digester liquor to No. 1 liquor tank
(B) 1st Wash with water from No. 2 recovery tank. Drain to No. 1
weak liquor tank (8500 gal./cook)
(c) 2nd Wash with water from No. 3 recovery tank. Drain to No. 1
recovery tank (8500 gal.)
(D) 3rd Wash with water from general white water chest. Drain to
No. 2 recovery tank (8500 gal./cook)
(E) 'fth Wash with water from general white water chest. Drain to
No. 3 recovery tank (8500 gal./cook)
(F) 5th Wash with water from general white water chest. Drain to
Parshall flume until temperature less than 95°F, then to RO
feed tank. Total (12,000 gal./cook)
Figure 17. Flow Sheet - Washing Cycle and Liquor Collection Calcium-Base
Acid Sulfite Pulping. Volumes Given per 1^ Ton Cook (10-12
Cooks/Day)
-------
exchange system to bring the temperature down to less than hQ°C as needed
for these "brief trial runs. The'remaining flow of dilute wash, waters
contained an average total solids concentration of about 1,1 percent,
and was well suited as a feed stream for the purpose of demonstrating
the capabilities of this reverse osmosis membrane equipment for concen-
trating by a factor of 10 times to give concentrations at 10 percent
total solids which could be more economically processed and to provide
the .minimum volumes which would fit the limited capacity of the existing
evaporation plant in any commercial application to be made at this mill.
Full-scale commercial design would advantageously include all of the'
fifth wash waters, but pretreatment for reducing the temperature might
be required. Need for expensive heat exchange equipment might be reduced
or eliminated by auxiliary pretreatment steps such as flashing off of
the acid volatiles in a combined cooling and BOz recovery system. This
step to reduce the volatile acid content could also eliminate need for
pH adjustment. Additionally, membranes capable of operating at higher
temperatures and lower pH levels have become available in 1971 as sub-
stantial and important advances over the temperature and pH sensitive
reverse osmosis equipment that could be supplied in constructing the
trailer-mounted unit for these demonstrations in 1968 through 1970.
DESIGN OF THE EXPERIMENTAL PROGRAM
Labprat ory Fh as es
Section V described the extensive program of preliminary studies upon
which these small pilot and large field demonstration trials were based.
However, -continuing small-scale studies in the laboratory were needed
to develop answers to operating problems of pitch fouling, CaSOif scaling,
and the like, as they arose in the field-test.program. Then, too, equip-
ment suppliers for pumps, valves, membrane modules, and the like, were
developing improvements in the course of this program which necessitated
the continuance of special tests in addition to the extensive analytical
control program conducted in the Appleton laboratory in support of the
field trials.
Small Pilot^Scale Field Testing
Operation of the large, trailer-mounted field demonstration unit, de-
signed to process flows ranging from 20,000. to 1QQ.,OQO. gallons per day,
was preceded by six months of preliminary study in the mill with a small
field unit equipped with 2k.modules (7-tube) for processing 1500. gallons
per day. The small unit, with about 165 square feet of membrane area,
•provided"important experience in developing performance indices and
information on equipment life heeded for design of the large-scale runs.
However, straight-through operation on dilute feed at 1 percent solids
could not sustain reliable and substantial levels of concentration with
use of the pumping equipment available and. with that limited amount
of membrane area. These preliminary studies with the small unit were
advantageous in studying the response of the RO system to variables
81
-------
in terms of temperature, pressure, velocity, pH, and especially for
preliminary evaluation of fouling problems such as with the pitch con-
tained in these softwood pulping wash waters .
Equipment and Liquor Collection for Small Pilot Runs
Figure 18 shows the schematic diagram for the pretreatment and RO opera-
tion. The pulp wash water from the fifth cooling stage was drawn from
the digester ahead of the Parshall flume and pumped to a Sweco vibrating
screen employing a 125-mesh screen. The screened wash then flowed to
the first of two ^000-gallon stainless steel holding tanks. For daily
batch- type operation, 50 percent caustic was added manually to adjust
the pH to above 3» and the solution was then transferred to the second
^000-gallon storage tank by a centrifugal pump. The second storage
tank with a cone bottom permitted settling out of suspended solids . The
clarified pulp wash water could then be drawn from the upper levels
by a continuously operating centrifugal pump providing feed liquor to
the RO unit, A portion of the flow recirculated to the feed line for
the second storage tank to avoid freezing during severe cold weather.
The feed tank level was controlled by a float valve. A low level float
switch, designed to shut down both the Milton Roy high pressure pump
and the Hypro recycle pump was also mounted in this feed tank. The
main Milton Roy duplex pressurizing pump had 1-1/2 inch diameter plungers
operating on an adjustable 3-inch stroke in a Carpenter 20 liquid end
equipped with Hastelloy C ball valves which delivered from 0.3 to 6
gallons per minute at pressures to 1135 psig maximum. One-gallon,
bladder- type accumulators on each pump served to dampen pressure fluctu-
ations and 0 to 1000 psig bourdon tube pressure gages were installed
to measure pressure controlled by 1/2-inch Victor-Acme Type K-20 back
pressure regulator valves of bronze construction. '
One module bank employed 12 modules with Type 3 membranes, while the
second used 6 modules with Type 5 membranes followed by 6 with Type
3. By operating a Hypro pump with a capacity of 3-3 gpm, a feed and
bleed recycle system was established, whereby higher concentrations
could be obtained.
Summary of Small Pilot Operations
There were three separate phases in the small pilot unit experimental
program.
Phase I ( 1st Through Hh Week) Operation with. Old Modules
After a short period of operation with new modules, the presence of
pitch in the feed was observed, and also a decline in the flux rate
occurred. Operations with the new modules were discontinued and an
older group of modules was substituted until it was determined that
the presence of pitch did not damage the membranes or irreversibly affect
membrane performance. The new modules were reinstalled after it became
clear that the flux rate of the fouled modules could be restored to
82
-------
CD
U>
KaOH
Y
Gal.
Storage
Tanks
Drain
Vibrating Screen
Float
Switch
Float
Valve
Feed Tank
(pH adjusfanent)
High Pressure
Duplex Punrp
BO Modules
12 x 7 Tubes in Series
RO Modules
12 x 7 Tubes In Series
( Digester )
Eecycle Pump
5th (Cooling) Stage
Wash Water
JParshall
j Flume
9
9
Accumulator
Pressure
Regulator
Bypass
Valve
Permeate
Product
TT7\
Drain
Figure 18. W.ow iJheet-Small Pilot BO Unit for Processing
Ca-Base Acid Sulfite Ptilp Wash Waters
-------
the initial performance "by flushing with tap water. Because of the
uncertainty of the condition of the older modules, results of this phase
are not included in the discussion of data and results.
Phase II (5th Through 19th and 23rd Week)
The new modules, eighteen Type 3 and six Type 5» were put back in opera-
tion on a straight-through basis.' Velocity, washing frequency, and
manually operated pressure pulsing techniques were varied for the purpose
of observing and minimizing the fouling of the membranes.
'Phase.'.Ill {20th Through 22nd Week)
The concentrate was recycled to determine system performance at a level
of about 5 percent solids concentration.
Dataand Results—Small Pilot Operations
Table 2k shows the flux rates of the dilute feed, which ranged from
11 to 19 g/1 solids during the run, could be maintained at 9-8 gfd flux
and 5.8 gfd flux at 600 psig and 35°C for the Type 3 and Type 5 membranes,
respectively. These average flux rates were maintained with an inlet
velocity of approximately 2 ft/sec (outlet 1«5 ft/sec), and with pressure
pulse of U-5 minutes duration twice a day. These pulses, effected by
manually shutting down the main pressurizing-pump, were conducted at
8 am. and *t pm each day. During the 16 hours between pulses, the flux
rate dropped to as low as 6 gfd for.the Type 3 membrane,, while the Type
5 exhibited much less fluctuation with a normal minimum of 5 gfd. Ref-
erence to engineering design and optimization"studies described in Section
VIII would seem to Indicate that concentration polarization could account
for only a 10 percent drop in flux after 100 hours of operation, and
the remaining loss in flux apparently was due to fouling of the membrane
surface. The pulsing and washing studies showed that this reduced per-
meation effect could be washed away by normal,osmotic "back flow"
during a pressure pulse or by water flush of the membrane surfaces .
Table 2k also provides a summary of weekly average flux rates. Figure
19 gives the detailed flux rate history for Phases II and III of this
small pilot run.
Rejection Ratios
Average rejection ratios as determined by composite sampling from each
of the 16 weeks of operation on a straight-through basis and for the
3 weeks on a recycle basis are summarized in Table 25. More detailed
data are provided in Table 26.
Rejection ratios were excellent in all categories for the tight Ho.
5 membranes, and still very good for the higher flux No. 3 membranes.
Solids, color as measured by optical density, COD and Ca were all re-
jected at levels of 9^-5 percent or higher; BODS at 89-97 percent; and
those components (electrolytes) contributing to electrical conductivity
81*
-------
TABLE 2k-
FLUX RATE SUMMARY
34ALL PH.OT ROT ON Ca-BAS! ACID SUIiPITl WASH WA3HR
Feed
Operating Hours, Concn., Flux Rate, gfd
5th Week
6
7
8
9
10
11
12
13
Average 5-13th week
14
15
16
17
18
19
23
Average 14-19 and 23
20
21
22
end of the week
426
489
592
644
670
766
870
964
1060
1150
1259
1349
1416
1515
1617
1904
week
1718
1813
1833
g/1 solids
12.8
12.03
11.05
12.72
11.75
18,43
16.3
13.5
13.2
13.5
11.5
12.2
14.7
16.3
11.7
12.5
18.9
14.0
47.3
54.7
51.0
Type 3
10.4
10.8
7,7
10.0
8.8
8.5
6,6
8,1
9.3
8.7
8.8
9.9
9.1
10.6
9.7
9.4
10.6
9.8
9.1
7.5
6.8
Type 5
7.0
6.0
5.6
6.3
6.2
5.8
5.3
5.6
5.7
5.9
5.9
5.8
5.9
5.9
5.4
5.1
6.4
5.8
4.7
5.1
5.0
85
-------
93
*
&
VI
15
OS
ON
% i
S1 i
* g-
5 3
»
w
H-
VI
o
» -4
fvn
5S
I
H-
oo
H
<$ ^
to. M
P»
CO
3
ri-
m
rt
(B
CD O>
&
8^
p
CD
o
Flux Rate,
M i_i J_i M
=—T p-™S ^—f—r
overnight
Water Wash - down for weekend
ra
down 1 day
Water Wash - down for weekend"
to
£:
Water Wash - down for weekend
ro
01 03 <
0 to"
O l"4 <
/
Wat/er Wash - down for weekend
to
w ro {B V
-------
16
o "type 3 Adjusted to 600 psig + 35°C a>
x Type 5 Adjusted to 600 psig + 35°C ^
S
12
I9
\.
v-
1.7 ft/sec
*—*
\.
-x—x
av
1.7 ft/sec
av
1.7 ft/sec
av
1.7 ft/sec
I
x—
av
1.7 ft/sec
av
1.7 ft/sec
12 \ (961*3 13 (1060) ll* (1163) 15 (1259) 16
Weeks of Operation - Total Operating Hours End of Week ( )
Figure 19 (Continued). Small Pilot Flux Rate History
Ca-base Acid Sulfite Wash Waters
with Small Pilot unit
(13^9) 17 (1*16)
-------
CD
OO
18
16
Ik
yt 12
-------
TABLE 25
AVERAGE REJECTION RATIOS - SMALL PILOT RUN
Ca-BASE ACID SULFITE HftSH WMTERS
Re J e cti on Rat i QS> perc en t
Straight-Through Feed
11-19 g/lTS
Recycle Feed Approx.
50 g/1 TS
Solids
OD
BOD
COD
Ca
Conductivity
No. of detn.
Type 3
Av
86
97
74
88
95.5
76
(16)
Type 5
Av.
95
99
89
94.5
98
85
(16)
Type 3
Range"
90-94
99
77-86
90-94
96-98
72-77
(3)
Type 5
Range
98-9
99
91-97
98-99
99
82-91
(3)
at 85 to 91 percent in the straight-through runs and also in the recycle
runs with the No. 5 membrane. Color rejection (OD) remained in the 97-
99 .percent range for the Type 3 membranes; solids rejection 86 to 9^
percent; COD 88-9^ percent; Ca 95.5-98 percent; while BODS dropped to
7^-86 percent and conductivity .72-77 percent.
These data were the base upon which decision was made favoring use of
Type 3 (Havens) membranes then available as 'the most practical choice
for larger scale studies throughout this Research and Demonstration Grant
project. Economic advantages were apparent for the-Ho.1 3"membrane in
providing higher flux rates coupled with adequate and practical levels
of rejection for all important components dissolved in the feed liquors
in terms of quality of permeate waters for recycle back to the pulp mill
for reuse.
Phase III — Small PilotL_j^ncentrat^g Run
During a period of three weeks, the unit was operated at a concentration
of about 50 g/1 solids by means of recycling the concentrate. Average
flux rates were 7,5 gfd for the Type 3 and k.95 gfd for the Type 5 mem-
brane at 600 psig and 35°C.
89
-------
TABLE 26
ANALYTICAL DATA FOR SMALL PILOT RUN
Ca-BASE ACID SULFITE PULP WASH WATERS
vo
o
Sample No.
1
2
3
4
5
6
No. of
Operating
Hours
Feed
Permeate 3
Concentrate
Rej. percent
Feed
Permeate 3
Concentrate
Rej. percent
Feed
Permeate 3
Concentrate
Rej . percent
Feed
Permeate 3
Concentrate
Rej. percent
Feed 426
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed 489
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Solids,
s/i
15.45
2.04
87
13.52
2.08
84
14.61
1.84
87
18.47
2.80
85
12.80
1.84
1.17
86
95
12.033
2.70
0.654
14,535
84
95
Ca,
mg/1
640
30
790
96
600
30
760
95
626
24
648
96
798
42
882
95
582
32
20
643
94
97
630
40
10
94
98
Optical
Density,
281 nm
78
2
101
97
72
2
84
98
74
2
74
97
76
2.8
93
96
78
2
2
86
97
97
90
4
2
96
98
BOD,
mg/1
3887
1252
4628
68
3778
1102
4502
71
5013
1048
5028
79
6848
1906
7298
72
4015
1170
730
4315
71
82
4640
1720
347
63
92
COD,
mg/1
14.590
1,990
18,940
86
14,690
1,882
17,640
89
16,010
1,346
17,360
92
20,120
3,024
22,840
85
14,720
2,176
1,368
16,580
85
91
17,200
2,846
944
84
94
PH
3.72
3.89
3.78
2.98
3.20
3.08
3.15
3.16
3.12
2.72
2.80
2.75
3.09
3.30
3.11
3.10
3.25
3.40
2.95
Specific
Gravity
1.005
1.006
1.004
1.006
1.005
1.006
1.007
1.009
1.004
1.005
1.007
Specific
Resistance,
ohm- cm
278
1600
327
79
297
1180
248
75
263
1100
257
76
321
750
211
57
303
1230
1580
280
75
81
293
1050
1650
73
83
-------
TABLE 26 (Continued)
ANALYTICAL DATA FOR SMALL PILOT RUN
Ca-BASE ACID SULFITE FDtP WVSH WATERS
Sample No.
7
VO
10
11
12
Feed
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Peed
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Ho. of
Operating
Hours
592
644
670
766
870
964
Solids.
8/1
11.05
1.48
0.394
12.79
87
96
12.72
1.94
0.60
15.05
85
95
11.75
1.02
1.07
15.8
91
93
18.43
1.71
1.19
91
94
16.3
2.25
0.60
86
96
13.5
1.93
1.57
18.6
86
96
Ca,
ng/1
420
20
10
580
95
98
650
30
10
770
95
98
570
10
20
740
96
98
880
30
20
800
96
98
770
30
10
970
96
99
710
30
10
930
96
99
Optical
Density,
281 nm
64
2
0.6
73
97
99
71
2
0.9
86
97
99
71
1.0
1.0
93
99
99
104
2
2
123
98
98
92
2
0.81
133
98
99
78
2
0.7
112
97
99
BOD,
Wi
2870
740
240
3573
74
92
2895
1385
185
5770
53
94
2822
580
565
3385
79
80
5785
1255
760
5972
78
87
4860
1199
377
5810
75
92
3742
1043
424
4885
72
89
COD,
ng/1
11,880
1,460
456
13,600
88
96
1(5,760
1,708
604
12,980
84
94
13,500
1,004
1,075
17,540
93
92
19,800
2,080
1,332
22,680
89
93
17,920
2,125
656
23,080
88
96
11,800
1,785
1,366
21,240
84
88
PH
3.83
3.95
3.49
3.80
3.20
3.28
2.85
3.23
4.35
3.98
3.69
4.80
3.27
3.33
3.05
3.37
3.33
3.62
3.00
3.45
3.45
3.62
3.09
3.42
Specific
Gravity
1.003
1.004
1.005
1.007
1.005
1.007
1.070
1.095
1.007
1.010
1.005
1.007
Specific
Resistance,
ohm- cm
350
1555
3610
309
77
90
278
1055
1428
250
74
81
302
2290
1880
236
87
84
212
912
1130
188
77
79
230
860
1740
187
73
87
259
1120
2290
203
74
89
-------
TABLE 26 (Continued)
ANALYTICAL DATA FOR SMALL PILOT RUH
Ca-BASE ACID SULFITE PULP WASH WATERS
Sample Ho.
13
15
ro
16
17
18
Ho. of
Operating
Hours
Feed 1060
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed 1163
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed 1259
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed 1349
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed 1416
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed 1515
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Solids,
8/1
13.2
2.04
1.57
18.3
84
96
11.47
1.37
0.40
15.64
88
96
12.19
1.70
0.47
17.72
86
96
14.74
1.92
0.44
21.6
87
97
16.26
2.27
0.75
22.29
86
95
11.74
1.69
0.52
16.35
86
96
Ca,
rag/1
720
30
10
950
96
99
590
20
10
780
96
98
600
23
10
810
96
98
720
23
8.5
920
97
99
620
30
15
820
95
98
440
24
8.8
694
94
98
Optical
Density,
281 nm
78
2
0.8
111
97
99
69.75
1.54
0.69
96.25
98
99
71.2
1.52
0.58
101.2
98
99
85.0
2.0
0.8
117.25
98
99
97
2.20
1.25
134
98
99
76.0
1.64
0.61
206.3
96
99
BOD,
mg/1
4170
1333
586
5380
68
87
3633
958
328
4243
73
91
3418
724
305
4460
79
91
4658
1316
416
6245
72
91
5040
1352
510
6633
69
90
3412
956
372
4600
72
89
COD,
fflg/1
14,580
1,922
622
19,760
87
96
12,063
1,349
461
16,545
89
96
11,600
1,289
297
16,760
89
97
14,920
1,716
541
20,980
88
96
16,220
1,840
834
22,440
89
95
12,680
1,466
482
17,840
88
96
pH
3.50
3.71
3.18
3.46
3.90
3.97
3.48
3.82
3.82
3.80
3.20
3.72
4.88
4.08
3.38
4.80
3.57
3.63
3.15
3.60
3.60
3.73
3.07
3.57
Specific
Gravity
1.005
1.009
1.004
1.006
1.006
1.008
1.008
1.011
1.006
1.009
1.005
1.007
Specific
Resistance,
ohm- on
1100
2390
76
89
326
1400
3910
257
77
92
283
1110
2170
220
75
87
217
970
2930
180
80
93
242
930
1640
193
74
85
303
1130
1690
244
73
82
-------
26 (Continued)
AHALmCAL DATA FOR SWLL PILOT ROT
Ca-BASE ACID SULFITE PULP WASH WATERS
Ho.
19
20
u;
21
22
23
Feed
Permeate 3
Permeate 5
Concentrate
Rej, Type 3
Type 5
Feed
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
Feed
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
feed
Permeate 3
Permeate 5
Concentrate
Rej. Type 3
Type 5
1718
1813
1833
1904
0.93
4.73
47.33
90
98
3.29
0,23
54.69
94
99
3.67
0.44
50.99
93
99
12.33
1.52
0.37
18.87
88
97
Ca,
mg/1
600
30
10
740
95
98
90
30
2430
96
99
3?
10
2480
98
99
28
9
1688
98
99
529
23
6
785
96
99
Density,
281 nm
78.8
1.25
0.33
105.3
98
99
5.0
2.15
423,8
99
99
3.0
0,60
875
99
99
4,55
0.870
367.5
99
99
77,2
2.20
1.52
175
89
98
BOD,
mg/1
3750
1119
396
4930
73
90
2443
556
10,800
77
95
2025
390
13,600
85
97
1987
1034
14,150
86
93
3855
1152
351
5730
73
91
COD,
mg/1
13,340
1,614
266
17,920
88
98
4,365
982
46,450
90
98
4,061
602
62,150
93
99
2,770
728
45,108
94
98
14,506
1,808
470
22,090
88
97
PH
3.48
3.62
3.58
, 3,45
3.11
2.78
2.88
3.55
3.01
3.45
3.47
2.88
3.22
5.17
4.21
3.78
4.68
1.019
1.025
1.024
1.007
1.010
700
1100
195
72
82
730
1750
168
77
91
630
1220
147
77
89
323
1510
5050
260
79
94
-------
An observation of substantial interest to these overall laboratory,
pilot, and field studies for concentration and also fractionation
processing of a variety of pulp and paper waste waters--was especially-
apparent in these 3 weeks of sustained operation of the small pilot
unit at higher levels of solids concentration. The rejections for both
the moderately open No. 3 membranes and the tight" No. 5 membranes in-
creased 5 to 10 percent as the concentration progressed on a feed and
"bleed program. This has been observed repeatedly in processing sub-
strates with a range of molecular size in the' dissolved solids contained
in the original feed material. This increase in rejection observed
with concentration of mixed feeds of large and small molecular weight
compounds may seem to be counter" to the decreased rejection of small
molecular weight solutes such as NaCl that is commonly observed at in-
creasing levels of concentration of brackish water and the like.
In the case of these pulp mill effluents, the rejection increases in
later stages of the overall concentration as the low molecular weight
materials passing the membranes (e.g., sodium chloride, and especially
the low molecular weight volatiles such as methyl and ethyl' alcohol,
acetic acid and sulfur, dioxide) are being bled' from the system in the
first stages of a concentrating system. The remaining higher molecular
weight materials such as sugars and lignin continued to be rejected
quite well by the tighter RO membranes and the degree of membrane re-
jection increases as the content of these hi-gher molecular weight dis-
solved materials increases in the feed to the succeeding stages of
concentration. The. significance of these fractionation effects will
be discussed elsewhere in reviewing the several demonstrations on vari-
ous wastes studied in this overall report.
Lar fg-S c ale_ Fiel _d_S t udl es_with_th e_Tr at le r_ Mount e d
Demonstration Unit
The extensive laboratory and RO studies of concentrating the' spent wash
waters from this mill extending back for more than a year coupled with
the six-month runs on the small pilot unit as described above provided
a firm base of experience on which to design the experimental program
for the large trailer unit under construction during the summer and
fall months of 1968. The mill staff had substantially completed instal-
lation of the liquor collection, the pretreatment, and feed liquor pump-
ing and control system well in advance of delivery of the trailer-
mounted demonstration unit. This unit was delivered by the manufacturer
October 12, 1968 and was installed and given preliminary running tests
over a period of 2-1/2 weeks. The experimental program and collection
of data started Noveiaber 1, 1968.
Equipmenttfor Pret re atment
"Cooling" stage wash water from each digester was collected in an 1*5,000-
gallon tile lined storage tank, mixed with remaining liquor from previous
cycles, and pumped to the pretreatment facilities. The storage tank was
-------
normally operated nearly full and at the feed rate of ^5,000. to-50,000. .
gallons daily; this provided for an average hold time of 1.5 to 2 days.
A recirculating line at the storage tank served to mix the incoming
waters. Pretreatment consisted of screening through a 120-mesh screen
and pH adjustment. Much of the suspended solids apparently settled
out in the bottom of the storage tank since the small "basket screen
available for use apparently served the purpose without need for clean-
out during the 3-month run, although it was found to be coated with
pitch when dismantled at the end of the run. The pH was adjusted from
the raw feed level of between 2 and 2.5 to between k and U.5 by means
of a 50 percent caustic solution with automated pH control. Some dif-
ficulty with an old pH control unit during the first weeks forced the
installation of a new controller after which reliable performance was
achieved. Caustic was added to the first of two mix tanks. The pre-
treated water was pumped "by centrifugal pump to the inlet of the triplex
piston pump. Figure 20 shows the flow sheet for pretreatment and RO
trailer unit operations at this site. A photograph, Fig. 6k (Appendix
A), shows the pretreatment section in operation.
'Trailer Operations (Process Conditions)
The main pressurizing pump operated at discharge pressures between 550
and 600 psig to feed pH adjusted and screened dilute feed to Bank la
which was in series with Ib for the multistage system. Banks II and
III were operated in parallel throughout the demonstration, and fed
Bank IV and then the final Bank V stage. The configuration of the systeu
is summarized in Table 27. This table also gives the velocities esti-
mated to be required to overcome concentration polarization and fouling
as determined by preliminary field trials and laboratory work.
TABLE 27
MODULE BANK CONFIGURATION-TRAILER UNIT
Ca-BASE ACID SULFITE PULP WASH WATERS
No. of No, of
No. of Parallel Series
Modules Flow Paths Modules Minimum Velocity
Bank la/Ib 90/36 18/12 5/3 2.4 ft/sec (1.5 gpm)
Bank II/III 30/21* 10/.8 3/3 3-2 ft/sec (2 gpm)
Bank IV " 3M. U8 3 3-2 ft/sec (2 gpm)
Bank V 63 21 3 4.0 ft/sec (2.5 gpm)
A summary of water recoveries and average flux rates is given in Table
28, while a detailed summary of the process conditions' and resulting
flux rates for those times during which samples were taken is given
in Table 29.
95
-------
50 Percent Caustic from Bleach Plant
Pulp Wash
Waters
80,000
Tile Lined
Feed Tank
Controller
Caustj c
Hold
Tank
PH
Mix
Tank
Surge
Tank
r
&
Trailer Feed Pump
F = Feed to stage I
Cn = Concentrate from the
n-th stage
Pn = Permeate from the n-th
stage
Concentrate
to Liquor
Recovery
System
Stage 1
Bank IA
18 Parallel
Rows 5 Mod
Stage 2
Bank II
10/11 Rows
3 in Series
Stage 3
Bank IV
Stage
Bank V
21 11 Row:
Main
Pump
1
ur:
k
1
.zirij
Z
'
Un_ Series.
Each Row 1
Bank Ib
- 12 11 Rows
3 in Series
"* P
"1 ]
L
*
^r
_>
P2
|~ Bank Til ~~
_^ 8 11 Rows
3 in Series
__ .
P2
J
:2
V
h
W3 11 Rows
3 in Series
k
~~T~
5
" 3 in Seris s
\
D i -
<^/
C3
I Control Valve
"igure 20. Flow Sheet-Pretreatment and Trailer Operation
Ca-Base Acid Sulfite Pulp Wash Waters
-------
TABLE 28
TRAILER DMA AVERAGES
Ca-BASE ACID SULFITE PULP WASH WATER
\o
Period
rating Hours
0-100
100-200
200-268
268-400
400-500
500-587
Average
Intake
gal. /day
46,000
39,000
35,000
48,000
44,000
45,000
Average Recovery
of Product Water
90
85
90
Removal of
83
80
78
Average Concentration
Final Concentrate ,
solids percent
12
9
12
CaSO, in Bank V
6
6
5.5
Average
Temperatiire ,
u.
23
24
26
28
27
28
Average
Flux Rate,
gfd/day
6.9
5.2
4.8
6.2
5.7
5.7
587-690
52,000
Wash-up with Detergent
82 6
28
6.4
-------
IWLE29
penm.ro mmUM or miurorawEiin UOLE ERcsssnts COMSC ncnj starns tot* «ASE suasss
Set
1
1ft
2
1
k
s.
6
t
e
9
M
11
12
u
ik
w
16
IT
Hours
57
100
125.5
155
1T5.S
203
223
2»
325
3*. 5
379
k33
• k»5
5ZO
563.5"
619-5
6*5
666
5±.
22/515
21/525
Jl/535
25/5M
«/«•
WSlt
27/53T
35/505
28/525
M/5JQ
27AT5
29/^90
29/k90
U/500
IJ/5M
M/WJ
28/530
24/520
»UJ.,
Id
10
10
12
Ik
12
Ik
13.5
u:>
12
15
'13
Ik
11
11
ik
Ik
Ik
9M* I
n«^«,*
3.23
!*.*«
k.«
(T.aui
1.73
(1.05!
3,-kO
»."*»
(&-0VS
J.55
(J.32I
(9.001
5.MI
(T.U.)
k.OO
"n?^'
3.«
S.I
2.k
a.T
3.D
2.1
2.k
i.k
2.k
2.8
j.e
S.9
2.9
a.»
2,fl
J.O
a, a
2. a
!«!>/*'
Prattitni
22ASS.
Z2/S20
2W5J1
JJ/502
3S/WO
W/512
J7/53I
30/M5
25/501
2*/k6S
2J/k05
2S/U35
M/5DT
WMS
2W1.99
J^k27
ZW5rr
2W1.85
staid'
g/1
20
Ik
12
Ik
Ik
IS
16
16
it
1*
1^
- 15
15
I*
13
IT
l£
15
«** II/III
J, Flsx 3a*e,*
«.2S
(M.95)
6.18
(12.55)
«-83
(12.1B)
3.te
kl£
(7.T5)
3.VS
(*.«!
6.0J
(10.13)
5^23
(7. 35)
3.61
"S^f1
2,6
3.5
3.2
3.2
3.2
3.3
3.3
3,7
3.S
k.l
3.9
1.2
k.C
k.8
k.l
• '4.3
. b.e
k.2
s.
22/1,52
21/500
21/501
2S/k»0
26/151.
S5WOO
27/52T
3«/%aa
25/MJ
2J/18T
. 26/kOO
26/U22
19/505
Jl/W*
j4/fe92
32/^18
23/51T
!}/kBT
Sol Ida,
S/l
k6
kk
29
kZ
»
1.6
to
62
52
29
35
2S
i>2
»
2k
*
*£
M
smk iv
Fliut B««e,* »
ttt
U..5
•(2S.5f
B.B
(1^.5!
6.T
(W.5)
6. 3k
llz2
ti2.28)
5.5
(S.553
(12. T)
5.3.8
[8. *tV)
5-3T
S%i:r
3.1
3.1
3.»
3.1
3.1
3.1
3.1
3.1
1.2
1.1
3.1
3.1
3.1
3.1
1.1
3.1
3,1
3.1
SKI&
tfaa$lkv S«lld«( :
Prtssur* g/1
25/M2 »
25/^92 ?k
22/5S5 kk
27/505 Tl
!';kJ5 59
27/500 SO
29/522 100
V
nuj Bate," Velocity,
>gfS K/«ec
J-*
2.3
2.5
5.3
3.0
3.0
2.3
— Out af Safvlc* —
™ Ottt ef
23/(>5$ 51
2W35 58
2V*"55 53
3V5k! 78
31/MO 52
24/515 . k5
32/kn$ 62
3J/5kO 133
25/520 SI
Scrrte*-
9-S5 3.5
10.25 3-9
(20.J6)
11.25 3.9
(SO. 02)
k.ks 3-9
(J.52)
8.6k 3-»
(Ik. 18)
I.Ut 3.B
(11.301
T, M 3.9
< 13-65!
5.02 3.9
(6 61)
S.B2 3.5
Solids
9»
Tk
U
71
59
93
U»
62
52
51
58
53
18
52
k5
•62
133
31
o»«r»ia
Flux Bate,
i«
5-k
k.O
*'«
5.k
5.*
' k.k
6.k
5.3
5-0
5-1
6.Q
5-J
6.1
5-2
5.3
5-3
6.3
.1.4
£,£.
33T
1ST
38T
33T
3BT
387
32k
J2k
324
38T
367
357
381
381
331
391
361
381
*( ) mint
at Mo F'U wrt 9*e.
-------
Af\ber about 100. hours of operation, the appearance of calcium sulfate
precipitate in the' c on cent rat e in. Bank V began to affect the performance
of this "bank. After the -unit had operated for 268 hours, the problem
became severe as evidenced by a total flow rate -which had dropped sub-
stantially from the normal ^0 gpm. The pressure drop increased from
70 to 120 psi. The permeate rotameters for individual banks had not
yet been installed, but there were indications the flux rate in that
bank had dropped off. Up until this time, Bank V had been operated
to achieve concentrations as high as 16 percent total solids. Calcium
salts were observed.to be precipitating at that level of total solids.
Laboratory tests on the cleaning of CaSOi» were indicated as preiriously
described in Section V. Cleaning by EHFA solution was found to restore
the modules in Bank Y to the original conditions.
Polyphosphate Experiments
The concentration of dilute pulp wash waters from the 1 percent solids
level to 6 or 7 percent solids is a substantial achievement in itself,
but substantially less than the 10-12 percent desired for evaporator
feed. Since the membrane system seemed capable in all other respects
of achieving concentration in the 10 to 2.6 percent solids range, it
was decided to try to inhibit precipitate formation by the addition
of polyphosphates.
Three' laboratory experiments were conducted. In the first, a solution
of Orlene M, an organic heat-resistant polyphosphate, was added to Bank
V to a concentration of 50 mg/1, with Bank V total solids at 9-5 percent.
No precipitate was noticed in the system. However, a considerable drop
in flux rate from 7.5 gfd to 3-5 gfd was attributed to film formation
in the membrane by excess polyphosphate.
The experiment was repeated with 20 mg/1 polyphosphate at a total solids
of 8,5 percent. This was apparently satisfactory in that no precipitate
or drop in flux rate was observed. However, the experiment was pre-
maturely terminated by mechanical problems.
The third experiment at 20 mg/1 polyphosphate and a total solids of
12.5 percent again exhibited the scaling due to calcium sulfate.
Time did not permit further experiments during this field trial to.pin-
point the conditions necessary to obtain a final concentrate of 10 per-
cent or greater. However, it is apparent that precipitate formation
and scaling can be inhibited and that scale already formed can be removed
from the Membrane surface. Earlier experience in laboratory and pilot
studies indicating feasibility of concentrating to,10-12 .percent^ solids
was confirmed by later experience which indicated that optimization
of hydraulic system design with higher'velocities would permit concen-
trating the liquor to 10 percent solids or higher without need for
chemical additives.
99
-------
Since 80 percent of the water or better is removed in Bank I, II/III,
and IV at concentrations less than 7 percent, either of the above methods
applied to the 20 percent of remaining feed volume might be practiced
in a commercial system at reasonable incremental cost.
Fouling of Membrane Surfaces
After the first few days of operation at high flux rates, indications
of low flux rate in Banks I and II/III were observed. After the instal-
lation of permeate rotameters to monitor the flow rates of each of the
banks individually, the reduced flux rates in these banks became readily
apparent. The flux rate of Bank I had dropped to U.9 gfd while Bank
II/III was 11 gfd. Further washing tests were conducted on a laboratory
scale (see. Section V).
Clean-up with a 300 mg/1 solution of Polytergent B-300,. at a recireula-
tion rate of 1 to 1.5 gpm/module (1.6-2. If ft/sec) partially restored
the flux rates to 10.0 gfd at 600 psig and 35°C for Bank I and 12.7
gfd at 600 psig and 35°C for Bank II/III. This clean-up did not restore
the flux rate to the extent shown possible in laboratory tests. The
following weekend the modules were cleaned with the same solution at
recirculation rates between 2 and 2.5 gpm/module (3.2-lf ft/sec). This
clean-up followed after the last operations on liquor feed at this dem-
onstration site were concluded, and prevented recheck of the flux rate
with the wash water feed. However, a check with tap water showed the
following flux rates when adjusted to 600 psig and 35°C:
Bank V to Bank I 20.9 gfd
Bank II/III 23-9 gfd
Bank IV 20.0 gfd
Bank V 21.b gfd
This indicated the flux rates were well restored and the trailer unit
was moved to Green Bay for the second field demonstration on NSSC white
water as described in the following subsection.
Flux Rates
The results of the Appleton field demonstration run in terms of actual
flux rate determinations are compared with calculated values at standard
conditions of operating at 600 psig and 35°C in Table 29. These data
were for the period after the occurrence of calcium sulfate precipitate
in Bank V and the pitch and silicate problems in Bank I and II/III.
Calculated values are obtained from the measured permeation rate and
specific gravity of the liquor for each bank by determining the solids
from Fig. 21 and the osmotic pressure from the osmotic pressure curve
for Ca-base acid sulfite liquor (Fig. 50 of Section VIII). A temperature
correction of 2.1 percent/°C was applied, and the rate of driving force
at 600 psig to the driving force at the given operating pressure (ne-
glecting the osmotic pressure of the permeate) was used to convert the
readings to 600 psig and 35°C. The accuracy is dependent upon the
100
-------
l.o6r
1.05
o
a
8
1.03
o
•rl
-------
operating conditions but for those readings at pressures of about 500
and temperatures of 25-30°C, accuracy should be within 10 percent. This
may deteriorate to 20 percent for those readings more distant from the
reference point.
Recoverj.es_ and Rejections
Table 30 summarizes the recovery data for the run vith the large trailer
unit, while Table 31 gives the complete summary of analytical data.
The calculated rejection values would not be indicative of module per-
formance over a wide concentration range if the rejections varied signif-
icantly. In this case, however, performance at different concentrations
by individual banks does not vary significantly, and thus, those data
correlated at the average concentration do give representative data.
The recoveries are calculated from the rejections and water recovery
S= (F2/F1)1-R (7).
Pretreatment by Neutralization with NaOH for pH Adjustment
Adjustment of the feed liquor pH to a level in the range of 3-5 to U.O
was specified by the membrane equipment supplier as necessary to avoid
acid hydrolysis and damage to the cellulose acetate membranes in these
field trials. The cooking liquors have a pH of about 2.0 to 2.5 after
discharging from the digester, principally due to the content of free
and loosely combined sulfur dioxide. This sulfur dioxide is routinely
steam stripped in commercial practice such as in yeast plants and an
analogous stripping effect occurs in commercial evaporation of the liquors.
In both cases, the pH rises to a level in the range of 3-5 to U.O. Steam
stripping could not feasibly be installed for these brief small-scale,
field trials but chemical neutralization was easily substituted by use
of approximately 0.5 g. NaOH per liter (U.3 lb/1000 gal.) throughout
these trials with the small pilot and with the large field demonstration
unit.
Adjustment of the pH to 7.0 was tried during one brief period of opera-
tion to determine the possible increase in BODg rejection that might
be achieved. This resulted in formation of a precipitate and a reduced
flux rate, but no significant change in BODs rejection was observed
(Sample 17, Table 31).
Enerffi Consumption
Power usage was calculated on the basis of the several conditions of
a 50,000 gpd input for the feed concentrations available at this mill,
an inlet pressure of 600 psig, 90 percent efficiency for the motor and
main high pressure pump, and a Uo percent efficiency for the centrifugal
pumps. Under these operating conditions energy consumption was found
to be IK 8 Kwh/1000 gal. of feed liquor for pressurizing the system with
the main pump and an additional 6.if Kwh/1000 gal. for the booster-recycle
operations. This amounted to a total of 11.2 Kwh/1000 gal. or 562 Kwh/day-
102
-------
TABLE 30
RECOVER* DATA FOR TRAILER UNIT WHILE PROCESSING Ca^BASE ACID SULFITE WASH WATERS
Le No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
IS
16
17
Water
Recovery ,
percent
90
75
85
82
87
90
75
70
75
75
75
75
73
75
75
90
66
Recovery, j?ercent
Solids
95
97
98
95
96
96
93
95
92
93
89
89
92
SI
91
87
89
BOD5
85
89
87
87
87
85
84
86
81
85
74
70
81
85
93
69
83
COD
94
95
95
94
95
94
91
94
92
93
88
80
93
93
93
37
91
Ca
98
99
99
97
98
98
99
99
97
97
98
96
98
98
98
96
95
OD
99
99
99
99
99
99
99
99
99
98
98
99
98
99
98
96
95
Range 66-90 87-98 69-89 87-95 95-99 95-99
During the last two weeks when the average intake was about 50,000 gpd, •
we found an actual average power consumption of 6^0 Kwh/2^ hours, Ihis
also includes power for instrumentation, lights, and the centrifugal
pump ahead of the main pressurizing pump. Although the power consumption
"by the electric space heater was also measured by the sane Kwh meter,
it was not included in the last figure, due to the fact that the warm
liquor in the modules and manifolds heated the trailer automatically
to a temperature well above the. preset.temperature on the thermostat,
so the heater was not activated during RO operations. All in all, both
power consumption figures seem to be in good agreement.
103
-------
TABLE 31
ANALYTICAL DATA FOR TRAILER UNIT WHILE PROCESSING
Ca-BASE ACID SULFITE PULP WASH WATERS
Sample
Set
Feed No, 1
Permeate
Concentrate
Rejection, percent
Feed No. 2
Permeate
Concentrate
Rej.,, percent
Feed No, 3
Permeate
Concentrate
Rej., percent
Feed No. 4
Permeate
Concentrate
Rej., percent
Feed No. 5
Permeate
Concentrate
Rej •> percent
Feed No. 6
Permeate
Concentrate
Rej., percent
Feed No. 7
Permeate I
Perm. II/III
Perm, IV
Perm. Total
Cone . I
Cone. II/III
Cone. IV
Rej. I
Rej. II/III
Rej. IV
Feed No. 8
Permeate
Concentrate
Rej . , percent
Feed No, 9
Permeate
Concentrate
Rej., percent
Neutralized Calcium
Solids, g/1 rag/1
8.05
1,03
84,30
97,8
8.02
0.58
42.89
97.8
9.40
0.82
74.89
98.9
10,21
0.86
60.00
97.55
9,85
0.85
79.49
98.1
11.14
1.16
102,56
98.0
13.04
1.14
1.22
1.87
1.67
16.52
19.69
66.32
92.3
93.3
95.7
12,58
1.31
53.63
96.0
11.09
1.86
50.80
93.9
430
20
5040
99.3
440
10
1480
99.0
530
20
4280
99.2
510
20
1650
98.15
530
16
4300
99.3
570
19
4500
99.3
630
11
12
15
14
760
1090
3720
98.4
98.7
99.2
620
10
1850
99.2
600
30
2560
98.1
OD at
281 nm
48.7
1.6
99.5
45.9
0.80
99.2
51.5
1.76
99.2
57.7
1.11
99.4
57.6
1.10
99.6
66.2
1.43
99.6
78.4
1.06
1.25
1.80
1.66
70.2
1.50
98.9
60.20
1.46
99.1
BODB , COD,
mg/1
2970
924
93.6
2870
589
89.4
3355
841
93.1
3705
811
92.8
3755
901
93.5
4090
1153
92.9
4445
814
839
1079
1074
4270
927
87.7
3440
1123
84.5
mg/1
8960
1306
97.4
9140
838
95,6
11,120
1119
97.6
11,930
1293
96.6
11,540
1260
97.4
13,308
1689
97,2
14,300
1148
1227
1846
1663
13,540
1391
94.6
11,580
1742
93.7
CoPt
Units pH
350
<5
230
<5
230
<5
270
<5
270
<5
320
<5
350
<5
<5
<5
<5
350
<5
250
<5
5.00
4.61
4.60
4.87
4.40
4.20
3.83
3.69
3.72
3.92
3.88
3.92
3.87
3.63
3.35
3.91
3.95
3.35
3.88
3.83
3,90
3.68
3.62
3.24
3.26
3.51
4.1
4.1
4,2
3.90
3.25
3.75
SpGr
1,005
1.041
1.005
1.022
1.005
1,035
1,006
1.029
1.007
1.039
1.007
1.048
1.007
1.009
1.011
1.031
1.007
1.027
1.006
1.024
-------
TAH.I 31 (Continued)
Sanple
Set
Feed No. 10
Permeate
Concentrate
Rejection, percent
Feed No. 11
Perm. I
Perm, II/III
Perm. IV
Perm. V
Perm. Total
Cone. I
Cone. II/III
Cone. IV
Cone, V
Rej. I
Rej. II/III
Rej. IV
Rej. V
Feed No. 12
Permeate I
Perm. H/III
Perm, IV
Perm. V
Perm. Total
Cone. X
Cone. II/III
Cone .IV
Cone. V
Rej. I
Rej. II/III
Rej. IV
Rej. V
Feed No. 13
Permeate
Concentrate
Rej., percent
Feed No. 14
Permeate
Concentrate
Rej.» percent
Feed No. 15
Permeate
Concentrate
ReJ«t percent
Neutralized
Solids, g/l'.
13.99
1.99
59.50
94.6
13.57
1.40
1.37
1.56
4.00
2.10
15.97
17.59
31.64
55.39
90.5
91.8
93.7
90.8
13,25
1.29
1.40
1.78
7.18
2.65
15.4
17.95
42.76
82. 561
91.0
91.6
97.1
88.5
13.57
2.0
55.23
94.2
10.43
1.36
17.93
90.4
15.1
2.29
65.0
93.5
Calcium
wg/l
680
30
2720
98.2
670
10
10
10
80
30
760
820
1390
2650
98.6
98.7
99.1
96.0
686
12
17
16
243
45
746
854
1740
4260
98.3
97.9
98.8
91.9
650
27
2450
98.3
490
19
1900
98.4
690
27
3260
98.6
OD at
2ol tun
74.40
2.78
355.2
98.7
67.2.0
1.32
1.49
1.63
4.55
2.07
98.2
98.1
98.5
97.7
70.0
1.09
1,66
1.54
18.80
5.41
98.5
98.0
98,8
99.3
68.0
2.ilO
98.7
58.0
1.29
99.1
80.8
2.85
98.7
BOIL,
mg/I
4870
1324
87.6
4220
1068
1086
892
2280
1292
75.9
77.2
86.0
74.2
4220
761
818
992
3585
1440
82.6
82.9
82.2
57.1
4210
1221
85.4
3185
802
88.0
5325
594
95.8
COD,
ng/1
16,430
1956
94.9
14,400
1282
1298
1416
3850
2020
91,6
92.2
93.8
89.6
14,050
1118
1252
1540
7670
2562
92.4
92.3
93.8
80.8
15,620
1632
95.4
11,560
1380
94.7
16,750
2232
95.4
CoPt
Units
300
<5
220
<5
<5
<5
<5
<5
300
<5
<5
<5
<5
<5
350
<5
250
<5
400
<5
PH
4.52
4.02
5.30
4.05
3,85
3,95
3.91
4.02
4.02
4.12
4.15
4.25
4.28
3.75
3.58
3,55
3.68
3,78
3.30
4.05
4.30
4.25
4.05
4.32
3.68
4.52
4.32
3.82
4-. 45
4.18
3.90
4.38
SpGr
1.007
1.029
1.004
1.0055
1.0065
1.013
1.0245
1.003
1.006
1.0065
1.019
1.036
1.003
1.024
•1.004
1.020
1.004
1.0285
105
-------
TABLE 31 (Continued)
Feed
Perm. I
Perm. II/III
Perm. IV
V
Total
I
II/III
IV
Sample
Set
Ho. 16
Perm.
Perm,
Cone,
Cone.
Cone.
Cone.
Kej.
ReJ.
a*j.
lej.
neutralized
Solids, g/1
11.9
0.95
1.00
1,68
5,44
1.80
Calcium OD at
mg/1 28l an
TOIL, COD, CoPt
mg/1 mg/1 Units pH SpGr
V
I
II/III
IV
V
16.
18,
47.
133,
93,
94.
94.9
94.0
.3
,7
.2
,5
,3
.3
Feed
Permeate
Concentrate
Re,)., percent
No. 17
10.0
1.37
29,2
90.6
580
9
10
14
86
22
780
880
2490
5660
98.7
98.8
99.2
97.6
250
19
720
96.1
62.0 3650 12,280 300 4.15 1.003
1.32
1.68
2.40
7.75
2.80
98.2
98.1
98.6
98.0
588
600
1038
3425
1188
06.0
88.0
88.2
72.1
862
1026
1449
5278
1775
94.0
94.0
95.6
91.3
<5
7
<5
10
<5
4.08
4.15
4.07
4.14
4.25
4.50
4.50
4.60
4.65
1.006
1.007
1.021
1.060
62,0
5.05
95.6
3185
822
84.8
10,860
1610
91.8
300
25
6.18 1.004
5.66
5,55 1.014'
Trailer Operations — Mechanical Performance
During this first demonstration, start-up problems associated with the
development of a new unit operations process hindered the gathering
of process data, tost of the problems were with associated equipment,
especially the mechanical seals on the centrifugal booster pumps and
for the pH equipment, and not with the modules themselves,. The modules
performed well mechanically in these first months of their overall
service life. Eight modules out of 38? had to be replaced during the
690 actual operating hours at the demonstration, site. Six had to "be
replaced due to fiberglass support tube rupture, one due to failure
of tube seals (inserts), and the eighth was replaced because it -was
so badly plugged with calcium sulfate that the scale could not be re-
moved without damaging the membrane. This was then tabulated as a
mechanical failure. One failure occurred in Bank I, one in Bank II/III,
two in Bank IV, and three in Bank V. Based on eight replacements in
690 hours , the unit averaged about 86 hours between module failures
in this first three month trial • The manifolding system on the trailer
was designed with cut-out valves for each five nodules, and this per-
mitted maintaining the unit in operation without need for shutdown to
replace individual modules.
Ope rat ing Hours — Downtime
schedule called for operation from early Monday morning until late
Friday afternoon each week, or about 100 hours per week. Thus, during
the 15 weeks of the demonstration, 1386 hours were available and the
unit was actually in operation for 690 hours, resulting in an operating
efficiency of h6 percent. However, during the last two weeks after
the startup mechanical problems were taken care of', an operating
106
-------
efficiency of 100 percent was achieved. Prime causes of downtime are
listed in Table 32. Individual mechanical problems resulting in down-
time - are. discuss ed'ln the" 'following paragraphs'.'
The pH controller initially used for the liquor pretreatment system
in the mil' was an old model' purchased in 19^5 and for which repair
parts were difficult to obtain. A new pH meter, controller, and trans-
ducer were purchased by The Institute of Paper Chemistry, for temporary
use' on this 'project,' and after' its installation, no further' problems
were encountered. The' pH meter' control system installed' on the trailer
worked well and provided automatic shutdown to protect the membrane
system when the feed pH control system failed in the mill.
Initially, the three centrifugal pumps, used as recycle pumps, were
supplied with a type of packing gland which did not seal reliably at
the system pressures of MDGrfSoO psig. These glands were replaced with
mechanical seals having a flush system. This arrangement proved satis-
factory for Pumps A and B, circulating liquor at concentrations below
7 percent > but at higher concentration with the attendant calcium sul-
fate precipitate, the spring in Pump C became clogged with calcium
sulfate. It operated well when the final concentration was reduced
to 7 percent solids for the final run.
The CaSOi, scaling problem was described earlier in the report and re-
sulted in system outage while the Bank V modules were being- cleaned.
Iqulpment to accomplish pressure pulsing as described in Section V was
installed oh the trailer during this*first run. Design criteria were
needed as based on actual^ operation prior to fabricating the pressure
pulsing system. Several" days were required for installation and testing
but were not charged to'the downtime summary.
DISCUSSION OF'THE DATA FOR NO. 1
During the 690 hours of actual operation, the semicommercial demonstra-
tion unit treated about 1,200 .,000. gallons of dilute, sulttte wash waters.
Msre than k$ tons of solids ;were concentrated at recoveries ranging
upward of 85 percent in the form of a 6-l6 percent solids ;concentrate,
and were sent to existing evaporators for further concentration pro-
cessing. At the same time, between 80-90 percent of the feed liquor
was recovered as a clear and colorless renovated water free of pitch,
microbiological growth, and foam problems which might affect reuse in
the mill. The high quality of this water for reuse in the mill and
the important pollution control capabilities were well demonstrated
in this extended run with indicated recovery of.69-89 percent of the
BODsr 87-95 percent of the COD, 87-98 percent of the total solids^ 95-
99 percent of the calcium, and an average of 99 percent of the color
in the concentrate. The flux rates averaged between 5 and 6.5 gfd for
the overall concentration at temperatures between 25 and 35°C, and pres-
sures between too and 550 psig, and with 'final total solids concentra-
tions ranging from 6 to 12 percent.
107
-------
OD
TABLE 32
CONSOLIDATED (APPLETON DIVISION)
RO TBAILER OPERATION - DOWN TIME
Period
From
10/31/68
ll/25'68
12/9/68
12/30/68
1/20/69
1/31/69
Total
Percent of
Percent of
To
11/25/68
12/6/68
12/20/68
1 *5 /**A l£.ft
1.1.1 JU/OO
1/16/69
1/31/69
2/14/69
available
down time
Total
Available
364
199
198
180
240
205
1386
100
Hours
Operating
102
101
91
76
126
194
690
49.8
Hours
Lost
262
98
107
104
114
11
696
50,2
100
Pumps
244
--
78
Down for
9
--
3
334
24,1
48.0
Hours Lost
pH Scale
__
82
2 27
66
32 8
5
121 101
8.7 7.3
17.4 14.5
Other
18a
I6b
—
29C
74d
3e
140
10.1
20.1
Ruptured
Modulesf
0
1
I
1
2
1
6
High pressure cut-off.
^High pressure cut-off.
c
Frozen pipe, power failure, high pressure cut-off.
Installation of pressure pulsing equipment, power failure, oil in effluent guage.
eHigh temperature cut-off.
System not shutdown for module replacement.
-------
The permeate water was remarkably- clear, colorless and low in dissolved
materials. Complete reuse to replace fresh water at a number of process
points within the mill would be desirable. Volumes of this permeate
at kO gpm were insufficient for conducting significant mill tests in
the time available, but the mill operating staff were of the opinion
this clear, clean water could be recycled without problems of build-
up of color, scale, slime, or other undesirable components. Stripping
of sulfur diojd.de and other volatiles, such as acetic acid, might be
desired to obviate any build-up of such materials in some areas of reuse,
but this would be of less concern in other recycle streams.
The feed liquor used in these demonstration trials in laboratory, small
pilot, and in the large trailer-mounted field unit was available at
between 1 and 2 percent total solids, but was otherwise representative
of the total pulp mill effluent being discharged to the river in the
form of digester cooling and pit waters totaling 1,200,000 gallons daily
at about 0.68 percent total solids. The feasibility of RO concentration
processing of the total volume of digester cooling and pit water cannot
be seriously considered.
The cooling wash water which was subject for study in this demonstration
had a total solids content averaging 11. U8 g/1 and a BODg of 3.87 g/1
from 17 weekly assays during the 3 months of operating the field unit
at this mill. The exact volume of "cooling" water which might be re-
covered for RO processing at those concentrations has been difficult
to determine because of its mixture with pit waters in a common effluent
line from a number of digesters, but separation of these flows would
seem to indicate a volume in the range of 1*00,000 to 700,000 gallons
daily, and an assumption of 500VOOO gallons may be reasonable for pur-
poses of the following estimations. Recent assays of the total of all
mill effluents,' of the combined digester cooling and pit waters, and
estimates calculated on the basis of 500,000. gallons of cooling water
flow provide a base for evaluating the possibilities for substantially
reducing the pollution problems by use of RO concentration processing
of those mill effluents (Table 33).
From these data it seems possible to consider RO as a means of processing
the "cooling" wash water to achieve as much as 1/3 reduction in soluble
solids and in BODs as based on the total liquid effluent discharges
from all operations of pulping, bleaching and evaporation at this mill.
The 50,000 gal./day of RO concentrate produced in a volume reduction
by a factor of 10 to. 1 could conceivably be processed with nominal expan-
sion of the existing evaporation, spray drying and marketing program
for spent liquor solids.
Module maintenance and replacement costs which developed unfavorably
in following field demonstrations as discussed in subsequent sections
of this report raised serious questions on RO operating charges. This
factor substantially affects feasibility for achieving these projected
advantages under practical conditions. Disposal problems might be
109
-------
substantially alleviated "by sale of the' concentrated and spray dry spent
liquor products in the long-term and successful program practiced for
nearly 20 years at this mill.
TABLE 33
ESTIMATED POLLUTION LOADS FROM VARIOUS EFFLUENT FLOWS
Estimated
Combined Values for
Total Mill "Cooling" Percent Digester Percent
Flow All and of Total "Cooling" of Total
Effluents Pit Waters Mil Flow Water Only Mil Flow
Volume,
gal./day 10,650,000.. 1,180,000 (ll.l) 500,000 (If.5)
Total soluble
solids,
tons/day
BOD5,
tons/day
70,35
23.70
29.21 (lH.5)
Ik.l6 (60,6)
2k.Q (3U.O)
8.1? (3*. 5)
In any case, the demonstration provides an alternative to other possible
routes under consideration for reducing the substantial losses of pulp-
ing liquor solids and importantly reducing the total pollution loading
of this mill. State pollution control surveys of November-December,
1970 indicate the total digester pit and cooling waters account for
^.5 percent of the suspended solids, k6 percent of the dissolved solids,
1*8 percent of the total suspended and dissolved solids, and 59 percent
of the BODg in the total of all effluent waters including the fraction-
ator effluents and evaporator condensates. If the fractionator effluents'•
were to be effectively processed in a clarifier installation, the capa-
bilities for reducing the mill environmental problems by HO processing
of the digester cooling and pit waters could increase to 18.8 percent
reduction for suspended solids, 52 percent reduction for the total of
soluble solids and suspended solids, and 68 percent reduction for the
total BODg loading from this mill. The evaporator condensates and bleach
plant effluents would then become the principal remaining problems for
final solution of major water pollution problems at this mill.
The choice of methods which might" be used to bring the' effluent discharge.
loadings within acceptable limits becomes a matter of evaluating the
economy of the various possible processing routes in terms of capital
and operating charges'. The economics of RQ treatment are discussed
separately in Section X.
110
-------
FIELD DEMONSTRATION NO. 2
Concentration Processing of Neutral Sulfite Semichemical
White Water by Reverse Osmosis
This subsection describes field studies vith the pilot unit and with
the larger trailer-mounted demonstration unit at the neutral sulfite
semichemical pulp mill of Green Bay Packaging, Inc., in Green Bay,
Wisconsin. Laboratory work which preceded the demonstration is described
in Section V, while additional work to obtain engineering design infor-
mation optimizing a commercial-scale installation for this waste water
is described in Section VIII.
A research and demonstration grant to Green Bay Packaging, Inc., from
the Environmental Protection Agency has followed the completion of the
studies which are the subject for this report and are currently in
progress (December, 1971)- Full-scale plant design studies are resulting
from the new grant and, if the outcome is favorable, could result in
the first commercial installation of reverse osmosis in the pulp and
paper industry.
Pulping. Paper Machine Operations, and Wash Water Collection
This integrated pulp and paperboard mill produces 210 tons of unbleached
fiber daily by the neutral sulfite semichemical process employing batch
digesters. About 200,000 pounds of spent.liquor dissolved solids are
produced. Approximately 70 percent of the solids are contained in the
digester liquors separated from the pulp at the digester and the screw
presses. This fraction is further processed-for.recovery of pulping
chemicals by a FluoSolids combustion process10. Figure 22 is a schematic
representation of the pulping and paperboard operation. The cooked
wood chips, after pressing, are slurried in large volumes of water during
primary refining operation. The virgin pulp is then mixed with 60 tons
per day of corrugated waste clippings and subjected to a final refining.
The pulp is further diluted and cleaned in centrifugal cleaners, after
which it is sent to the paper machine. On-machine washing is practiced,
and as much as can be used of the machine white water is recycled to
the primary refining and cleaning steps described ajaove. This machine
"white water" comprised the feed to the reverse osmosis system for both
the pilot run and the trailer run, "White water" is a technical term
for the effluent waters containing fibers which drain from the wire
screens on the paper machine. In this mill the "white waters" are quite
dark in color. The overall objective vas to develop methods for closing
the pulp mill and the paper machine water systems by total recovery
of the dissolved solids in a concentrate stream and production of clean
water for mill reuse in place of fresh water. Concurrently with the
RO studies, the mill has been engaged in a program to increase direct
recycling so as to obtain the maximum water reuse consistent with product
quality and mechanical operations. The unmodified white water available
during most of the pilot and field studies covered in this report aver-
aged 0.8 to 1.0 percent total solids, but more recently the recycle
111
-------
Batch
Digesters
Screw
Press
Liquor to
EluoQollda Combustion
System
Primary
Refining
Finishing
Refiner
Impco
Thickener
T
Screened White Water
Feed to BO
Centrifuge!
Cleaners
Paper Machine
System
Fiber
Flotation
Unit
Kraft
Repulper
Kraft
Clippings
Wiite Water
Containing
Beeovered
Stock
J
White
Water
Chest
to Hhl
te Water
Jtion
Figure 22, Meutral Sulflte Semichemlcal Pulping
and Board Mill White Water System
112
-------
waters have had a higher solids concentration, averaging greater than
3.5 percent total solids as of December, 1971- Morris, et^al^.11 have
summarized this recycle development program to date, including projec-
tions of planning for future RO processing at this NSSC pulp Mil.
Small amounts of suspended fibers are present but these total solids
mainly include suspended "fines" and colloidal suspensoids deriving
apparently from ray cells with the dissolved solids. Analytical
characterization of the screened paper machine white water recovered
from the flotation unit and Impco thickener as feed to the RO unit at
the time of this demonstration is summarized in Table 3^-.
TABLE 31*
ANALYSIS OF A TYPICAL NSSC WHITE WATER FEED
Total solids, g/1 9.0
Sodium, mg/1 1075
BOD, mg/1 23UO
COD, mg/1 9930
Optical density at
281 nm 51.8
Specific resistance,
ohm-cm 350
Small Pilot Phit Study
Description of Equipment and Pretreatment Procedure
The small scale lab and pilot equipment for development of procedures
for pretreatment and for reverse osmosis were similar to that used on
Ca-base acid sulfite wash water at the' first demonstration site. Nine
new, Type 3, 18-tube modules, manufactured by Havens International and
delivered early in 1968, were used to process NSSC white water obtained
as filtrate from the thickener. The paper machine white water used
as feed to the thickener was the underflow from the flotation unit.
This fraction of the total white water, after clarification by passage
through the mat of pulp on the Impco thickener, was collected for feeding
to the RO process.
The thickener filtrate was discharged to a UO-mesh Sweco vibrating screen,
with the filtrate running to a 150-gallon stainless steel tank, and
then was pumped by centrifugal pump through a stainless steel heat ex-
changer, which cooled the filtrate from M3°C to 35°C for RO processing.
The pH of this feed liquor ranged around 7'0 and further adjustment
113
-------
was not necessary. Temperature and tank level svitches were mounted
on the feed tank to shut off the pressurizing pump in case of high
temperature or low level.
The pressurizing pump was a simplex, positive displacement, piston pump
with infinitely adjustable stroke rated at 0 to 5-1 gpm at 7^5 psig.
This pump discharged to the module system consisting of three banks,
each with three 18-tube modules in series. The flow pattern was set
up so that the flow from the simplex pump was fed into two of the banks
in parallel, with the concentrate from these two banks combined to serve
as the feed to the third bank of three modules. For the first part
of the run with straight-through operation, the concentrate from the
last of the modules in this bank was wasted to the sewer. To obtain
a higher concentration in the last part of the run, a recycle feed and
bleed system was set up, with the concentrate returned to the feed tank
with a measured flow being drawn off from an adjustable weir device.
The last bank was not used during the final 3 weeks of operation. Pres-
sure fluctuations were dampened by a 1-gallon bladder-type accumulator
at the pump discharge. System pressure was manually controlled by a
spring-loaded pressure regulator. The system could be pressure pulsed
with a preset time cycle by an electrically operated valve which re-
lieved system pressure to atmospheric at the inlet side of the module
system. Figure 23 is a schematic of the pretreatment and RO system
for this phase of the study.
Data and Results — Preliminary Small Scale Runs
Reverse osmosis performance at feed solids concentrations between 1.0
and 9.3 percent was determined in two phases. The unit was operated
first on a straight-through basis with a raw feed concentration of
approximately 9.6 g/1 for five weeks. Phase II followed with nine weeks
of testing at various recycle rates to give performance indices for
solids concentrations between 33-93 g/1. All tests were conducted with
the feed solution of about 35°C, pH around 7, and pressures between
500-600 psig. Some of the modules developed leaks which decreased or
ceased when the pressure was reduced. Later experience would seem to
indicate that these leakages may have occurred at the sleeve-type seals
for the individual tubes. Rupture of several modules occurred in the
final weeks at 600. psig with pressure pulsing. The flux rate results
are reported at the conditions of measurement and for comparative pur-
poses, and also after adjustment to 600 psig and 35°C.
Inlet velocity to each of the two parallel banks was 2.k ft/sec while
operating with dilute feed on a straight-through basis. The loss of
velocity within each stage of the system varied as the flux rate varied
at the different concentrations. For the higher concentrations in Bank
III, the velocities were between 3.3 and U.3 ft/sec.
Throughout the run, the system was pressure pulsed for about 1 minute
each hour. This pulse apparently provided a cleansing reverse flow
back through the membrane by normal osmosis from the dilute permeate
llU
-------
Cooling Water
Feed From
Impco thickener
I« J
lirBc
Reject
Sewerec
f-;
150- Gallon
Surge Tank
Straight-through Operation
Heir Set to Maintain a Given
Solids Concentration
In the Feed Tank
Pressure Regulation Valve
Permeate
Sewered
Permeate
Sewered
J
Permeate
Sewered
PRV is a Pressure Relief Valve
PG is Pressure Gage
Ace is Bladder-Type Accumulator
Figure 23. Pretreatment and EO Operations-for Processing
IfSSC White Water with Small Pilot Unit
-------
side to the concentrate side. The permeate rate was measured 10 minutes
before the pulse and 10 minutes after the pulse. The flux rate dropped
about 12,percent between pulses, and average flux rates given are the
arithmetical average of these two determinations. Samples taken at
the time of each reading were composited over a week's operation.
Flux Rates and Rejections'— PhaseI —
Straight-Through Operation
Measured flux rates averaged 11.3 gfd at inlet pressures between 1+91
and 500 psig for the dilute feed with an average solids concentration
of between 11-1^ g/1. This corresponded to a flux rate of 13.3 gfd
when adjusted to 6*00, psig and 35°C. Flux rate averaged 10.h gfd before
pressure pulsing and 12,3 gfd after pressure pulsing for the five-week
run. The weekly average flux rates for the run are shown in Table 35«
Rejections were excellent, averaging 98.7 percent for total solids,
97-5 percent for sodium, 96.^ percent for BODs, 98.^ percent for COD,
99 percent for optical density at, 281 nm, and 9^-5 percent based on
conductivity. The only pretreatment required was screening through
a ^0-mesh screen and temperature adjustment to safe operating range
for the cellulose acetate membranes (below ^0°C). The modules were
washed with water to protect the membrane while standing idle over the
weekend after five or six days of operation, but the daily flux rate
history (see Fig. 2k) indicates that no irreversible flux loss occurred
over a weeks operation. The reader is referred to Section VIII for
a subsequent study on optimizing the BO system while processing NSSC
white water,
Flux Rates and Bejections — Phase II —
Recycle Operation
The flux rate data at feed concentrations between 38 and 103 g/1 total
solids are plotted in Fig. 25, The flux rate of 13.2 gfd at 600 psig
and 35°C at 12 g/1 solids decreased to h.k gfd at 600 psig and 35°C
at 100. g/1 solids. Most of the water (80 to 90 percent) to be removed
overall was processed at relatively high flux rates ranging above 7
gfd. Bejections as shown in Table 36 remained uniformly high at all
concentrations studied. This indicates the recycle system at higher
concentrations was representative of a system processing this water
in a staged, straight-through system such as would probably be used
in large-scale operations .
• •Field Demonstration Studies 'with. Large Trailer-Mounted RO Unit .
The large 50,000 gallon per day trailer unit was operated on white water
for a total of 593.5 hours, liquor processing and equipment maintenance
problems caused many delays and seriously hampered process evaluations.
Interpretation in terms of performance which might be expected in a
commercial unit processing this waste water is best obtained from evalu-
ating operating data from the overall studies on both, the small pilot
116
-------
TABLE 35
FLUX RATE DATA
NSSC WHITE WATER WITH THE SMALL PILOT UNIT
Av. System
Concentration
Week No. g/1 solids
1
2
3
4
5
Av. Straight -
Through Flow
6
7
8
9
10
11
12
13
14
11
12
14
11
14
12
58
75
60
74
78
44
102
38
98
Feed
Temp,
°C
35
35
35
36
36
35.5
36
37
37
35
35
35
36
34
35
Average
Inlet
Pressure,
psig
491
515
540
560
550
531
546
561
536
576
583
596
585
612
620
Before Pulse,
gfd
10.5
10.6
10.2
11.2
9.5
10.4
6.0
5.3
6.6
5.4-
5.4
7.2
3.7
8.0
4.4
After Pulse,
gfd
13.1
12.4
12.1
13.2
10.8
12.3
6.8
6.0
7.4
5.8
6.2
8.4
4.3
8.9
4.9
Average
gfd
11.6
11.5
11.2
12.2
9.9
11.3
6.4
5.7
6.9
5.5
5.8
7.8
4.0
8.4
4.6
Av. Adj.a,
gfd
14.6
14:2
13.1
13.4
11.5
13.3
7.5
6.3
8.4
6.1
7.0
8.0
4.3
8.4
4.4
Adjusted to 35°C and 600 psig inlet pressure.
-------
H
OO
12
„ 10
1
&
x 8
O Before pulsing
• After pulsing
ISO
200
300
too
500
600
700
800
900
Operating Hours
Figure 2k. Daily Flux Rates.and Pressures for Small Pilot Unit ISSC White Water
-------
M
H
DO
T-I
CD
P<
'600
£550
09
CO
•o—o-
-o—o
I6r-
12
10
I
O Before pulsing
• After pulsing
g 36-53 g/1
63-85 g/1
TDL
I
I
I
900
1000
1100
1200
1300
1400
1500
1600
1700
Figure 2k (Continued)
Operating Hours
Daily Flux Rates and Pressures for Small Pilot Unit NSSC White Water
-------
16-
ik
12
Sio
bO
ti 8
K
H 6
u.
X
V.
I I i i I I I I I I I
0 10 20 30 to 50 60 70 80 90 100 110 120
Solids Concentration, g/1
Figure 25. Variation in Flux. Rates with Increase in
Solids Concentration Processing NSSC
White Water with Small Pilot Unit
and the larger field units. In Tooth cases, product vater quality was
excellent, with exceptions occurring when faulty individual modules
were obviously degrading the total flow. The final concentration during
most of the runs was about 6 percent solids, and fell short of the
desired 10 percent, but experience with both the pilot unit and short
runs with the trailer unit shows that a 10 percent concentration can
be reached feasibly. The overall flux rate for the trailer unit dropped
to as low as 5-1 gfd. This indicates fouling occurred. However, subse-
quent studies with the trailer operating at higher velocities showed
that overall flux rates of 7 gfd while concentrating from 1-10 percent
solids could be maintained. These studies conducted at the Institute
are reported in Section VIII,
Equipment Description
A schematic of the auxiliary equipment installed for the trailer is
shown in Fig. 26. The white water feed was obtained after passing
through the thickener mat on the repulper system. This was then pro-
cessed through a Uo-mesh Sweco vibrating screen for final fiber removal
as described for the small pilot study but with larger screen, tanks
and piping. In initial tests, 100-mesh screens were tried, but these
blinded frequently from slime growth. The 1000-gallon storage tank
120
-------
TABLE 36
ANALYTICAL DATA AND REJECTIONS WHILE PROCESSING
SEMICHEMICAL WHITE WATER WITH PILOT UNIT
Sample Solids,
No. g/i
1 Feed 8.65
Permeate 0.18
Concentrate lU.06
HeJ. , percent 98.!*
2 Feed 9-31
Permeate 0.16
Concentrate 15-79
ReJ. , percent 99
3 Feed 10.85
Permeate 0.18
Concentrate 17-50
ReJ. , percent 98-7
1* Feed 8.2l»
Permeate 0.12
Concentrate 1U.21
Rej., percent 98-9
5 Feed. 10.70
Permeate 0.23
Concentrate 16.5!*
ReJ . , percent 98. 3
6 Feed 50-98
Permeate 0.81
Concentrate 65-0
ReJ. , percent 98.6
7 Feed ' 60.60
Permeate 1.06
Concentrate 90.81*
ReJ. , percent 98.6
8 Feed 51.17
Permeate 0.68
Concentrate 67.87
ReJ, , percent 98.9
9 Feed 63-35
Permeate 1-08
Concentrate 85.2!*
ReJ. , percent 98.5
Na,
mg/1
1020
36
19 hO
97-6
1250
33
2190
98.1
li*20
1*0
2290
97-8
1050
29
1770
97-9
10l»0
50
161*0
96.3
7700
233
9500
97-1
8200
281*
12,100
97-2
7300
191*
9 too
97.7
10,000
275
12,300
97-5
BODs,
mg/1
2120
61*
2920
97-5
1705
1*8
2780
97-9
2395
110
361*5
95
1770
92
2570
95-8
21*20
157
3770
96
11,200
1*85
15,950
96.U
lit, 1*50
61*1*
.20,950
96.1*
12,1*50
1*21*
16,200
97
15,500
590
20,750
96.8
COD,
mg/1
9120
210
ll»,860
98.2
9990
191
16,560
98.6
11,620
2 1*0
18,21*0
98.1*
8680
168
lit, 660
98.7
12,000
291*
17,660
98.0
55,000
791*
81,300
99-0
67,750
1061
105 ,600
98.8
56,650
665
7l*. 500
99-0
70,950
835
88,1*00
99-0
OD,
201 nm
1*7-0
0.7>*
125.0
99-1
50.0
0.60
82.0
99-1
58.0
0.7
96-5
99-0
Ul*
0.1*6
78
99.2
60.0
0.90
89-0
98.8
320
1.70
too
95-5
350
1.82
575
99-6
255
1.52,
1*00
99.5
371
0.99
1*75
99-8
Specific
Resistance,
ohm- cm
630
7800
31*8
93.7
385
7600
250
95-8
1*12
7300
292
95-1
590
7800
31*2
91*. o
387
5500
258
9l*. 1
137
1560
97
92.5
ll*2
11*20
127
90.5
155
1950
131*
92.6
156
I960
139
92-5
pH
7.1*8
6.70
7-05 •
6.35
6.56
6.1*5
6.28
6.58
6.60
6.70
6.00
6.1*5
6.70
5.61
6.50
5.1*0
5-30
5-50
5.88r
5-65
5-95
5-95
5.59
6.10
6.30
6.20
6.55
121
-------
TABLE 36 (Continued)
ANALYTICAL DATA AND REJECTIONS WHILE PROCESSING
SEMICHEMICAL WHITE WATER WITH PILOT UNIT
Sample
Nc.
10
11
12
13
.e Solids ,
g/1
Feed 72 . 79
Permeate 0.96
Concentrate 83.21*
Rej. , percent 98.8
Feed 36-97
Permeate 0.1+4
Concentrate 53.32
Rej . , percent 99
Feed 92.62
Permeate 1.22
Concentrate m . 57
Rej. , percent 98.8
Feed 32.97
Permeate 0.37
Concentrate 1*2.65
Rej. , percent 99.0
Feed 92-92
Permeate 1.25
Concentrate 10**. 19
Rej. , percent 98.7
Na,
mg/1
9500
259
11,500
97-5
1*100
107
5700
97-8
13,500
293
13,500
97-8
1*200
125
5200
97-3
10,000
303
11,720
97.2
BOD 5 ,
mg/i
ll*,250
570
18,500
96.5
8675
252
13,300
97-7
18,1475
829
20,850
95.8
6850
216
8975
97-3
19,375
920
31*, ooo
95.8
COD,
mg/1
71,1*50
868
88,500
98.9
39,900
1*18
57,600
99-1
106 ,700
1279
117,800
98.9
37,900
l+ll*
51,200
99-1
101,500
131*6
117,200
97-7
OD,
281 nm
385
1.09
1*90
99.8
206
0.65
291*
99-7
581
1.05
61*1
99.8
195
0.67
222
99-7
1*90
1.15
555
99.8
Specific
Resistance,
ohm— uro
215
1750
182
88.7
372
3650
350
89.8
1*2
978
1*2
95-7
96
3100
82
97.1
1*2
81*2
1*2
95.0
PH
6.35
5-50
6.55
6.53
5-72
6.57
6.85
5.81
6.87
6.62
5.88
6.82
6.28
•5-35
6.25
was equipped with a low level capacitance probe interlocked with the
trailer controls and with a "bubble pipe connected to a valve on an
emergency make-up line providing white water clarified by the flotation
unit.
The feed was pumped from the tank with a Worthington centrifugal pump
which brought the pressure to about 37 psig; the feed then made a single
pass through the heat exchanger. City (Lake Michigan) water was the
coolant; the discharge temperature was sensed by a capillary bulb and
fed to a pneumatic transmitter, which in turn supplied a standard
recorder-controller. The latter actuated the cold water supply valve.
The feed pressure to the trailer was about 23 psig.
A second tank was utilized for collection of the concentrate (effluent)
from the trailer; on occasions when the system operated smoothly and
the concentrate solids content was high, this stream was pumped to
the pulp mill liquor collection system for eventual incineration in
the FluoSolids recovery plant.
The trailer described in detail in Section VI required only connections
for power and the five liquid streams — feed, concentrate, product
122
-------
Cooling Water
S-J)-*^ CJontroller
Emergency Makeup
ro
uo
1000 Gallon
Feed Tank
Bubbler Type s*
Level Probe -^
Capacitance
Level Probe •—
jlinpco
White
Jump
ToJPulp Mill
Liquor
Collection
System
b-
Heat
Exchanger
Main
Pressurizing
Pump
Concentrate
Collection
j Tank
Bank la
18 parallel rows
5 modules in series
each row
Bank Ib
12 parallel rows
jwdules.
CBank II
10 parallel rows
3 modules in series
Bank III
8 parallel rows
modules _in_ series _j-
L^"
"U
Bank IV
48 parallel rows
3 modules in series
*•
\ Bank V 1
• 21 n#ra1 lf»1 fows ^
jf
\ r*
5T
To Sewer
.01
so
a
m
JL
Air Loaded
Back Pressure
Regulator
Figure 26. Schematic of Reverse Osmosis Pretreatment and Trailer Equipment
-------
(clarified) water, city water, and floor drain. The 38? Type 3 modules
in the trailer were arranged in five stages.with a particular pattern
throughout the tests. Banks'II and III were operated in parallel', al-
though the piping arrangement permitted an alternative series configura-
tion. For almost all tests, the full complement of 387 modules (6579
square feet of membrane surface) was used.
A Manton-Gaulin piston pump with a variable-speed drive provided initial
pressurization. Centrifugal pumps in each of the last 3-stages were
utilized to restore pressure losses and to provide for the' high velocity
reeirculation necessary to prevent fouling on the membrane surface.
The centrifugal pumps were conventional with the exception that the
casings were built for a high inlet pressure and specially designed
mechanical seals were required in place of the usual packing on the
shaft.
The restriction routinely in use for controlling back pressure at the
concentrated end of the process flow was a manually-adjusted ball valve;
a gas-loaded back pressure valve was also available but not used because
of slower, less sensitive response during frequent startups and shutdowns
The product -water from each row flowed to a common header' for the partic-
ular stage; the flows were then combined before being discharged.
All liquor lines were equipped with flowmeters of the magnetic follower
type, having a spindle-mounted rotameter plug inside the pipe; these
were quite inaccurate since the presence of slime introduced an obvious
but indeterminate error. Pressure gages and temperature devices were
located throughout the system but the capillary lines were subject to
pluggage because of the suspended solids content of the liquor. Visual
rotameters were installed on the product water lines and proved satis-
factory.
The electrical system usually worked satisfactorily. A timer system
had been incorporated to permit a "pulse" or rest period in which the
unit would be depressurized. Because the small pilot unit of 1968 had
many ruptures which may have been attributable to the "hard" abrupt
pulse, and for mechanical simplification, the' trailer was provided with
a system in which depressurization was relatively gradual. Shock and
hammer were not apparent in the system, ,
Operations
The' unit was operated at pressures over 300 psig for 623-75 hours, of
which 593,5 hours were on white water. There were five runs on white
water including h principal periods of operation and one brief run;
interspersed were periods of cleaning and testing.
121+
-------
Testing - 2/2U.- 3/10
Checkout, modifications and "brief runs on city water and white water
during startup.
Run #1 - 3/11 - 3/15
the first two days various problems, e.g.., ruptured modules, caused
interruptions . During the next three days the flux rates were declining
sharply. Plugged rows were encountered in Bank V. Since the plugging
or fouling was continuously causing increased pressures , it was decided
to terminate the run after 88.75 hours.
Testing - 3/16 - U/7
A program of evaluation and cleaning took place. After flushing the
system with a detergent solution, the flux rate for every row was checked
with city water. Since low flux rates indicated possible plugging on
"the liquor side, all 117 rows were checked for liquor flow. The unit
Was then operated with a 500 mg/1 salt (NaCl) solution for feed to check
for abnormal salt content in the product water for each row. Simulta-
neously the flux rates were measured on individual rows.
Since the flux rates were still low, cleaning was performed by circu-
lating a neutralized solution of the sodium salt of EDTA (ethylenediamine
tetraacetic acid). Following this treatment of all banks, a flux rate
check was made with water; this indicated further cleaning was required.
The cleaning was repeated on Banks I and IV, following which the flux
rate check indicated that most modules were again clean.
Run #2 - U/8 - Vl8
A new run began with white water, but again minor problems — ruptures,
"temperature surges, mill downtime, continuous pressure rise, etc. —
caused many interruptions . There was little information on steady state
operations; consequently, the unit was shut down until personnel were
available for 2U-hour coverage. Operating time on white water was 217-5
hours during this run,
Testing - U/15 - U/27
Maintenance was performed on the Manton-Gaulin pump, and other minor
problems were corrected. Flux rate check was made with city water.
Run #3 - V28 - 5A
With personnel assigned to every shift, the unit was operated on white
water for 1^2 hours, of which 121,25 hours were in a continuous run.
One interruption occurred that was later determined to be a problem
in the main pump electrical drive system. Other minor disturbances,
such as ruptured modules, were corrected as they arose. During the
long run, the flux rate again deteriorated.
125
-------
Testing - 5/5 - 5/18
Minor maintenance problems were corrected. One module vas removed from
each row in a section of Bank la (l8 total) and from each row in Bank
V (21 total) to reduce the pressure drop in later tests. A flux rate
check was again made with city water.
Run #U - 5/19 ~ 5/25
Again with personnel on all shifts, the unit was started up on white
water and operated 139 hours. To determine the effects of slime, veloc-
ity, and flux regeneration resulting from downtime, the run was divided
into three major sections:
With some rows blocked off to increase velocity through the
remaining rows and with 20 mg/1 slimicide present in the
feed, the unit was operated for ^.25 hours; a failure in
the main pump electrical drive system occurred after 25.75
hours, however. The run was terminated when a booster pump
seal failed. The unit was flushed with water and remained
down until repairs could be made on the pump.
Operations were resumed with the same rows still blocked off
to maintain the high velocity, but with no slimicide in the
feed. This run proceeded for hk hours.
Operations continued with the same number of rows in
operation, but about every 50 minutes the rows in each
stage were rotated, i.e., idle rows were put into service
and a comparable number were taken out of service. Again,
no slimicide was used in the feed. This method of opera-
tion continued for ^9-25 hours.
The unit was run for an additional 1.5 hours with all rows
in service (no rotation) in an attempt to obtain higher
concentrations.
The unit was then shut down and flushed.
Run #5 - 5/2?
A brief run (5-75 hours) was made with white water to obtain samples
at a high concentration, following which the unit was flushed.
Clean Up - 5/28 - 6/5
The flux rate was again checked with city water, following which the
39 modules removed for the final tests were reinstalled. The entire
unit was then flushed with an EDIA solution and after checking the flux
rate with city water, and further flushing with a solution with 20. mg/1
algacide, it was disconnected and sent to the next demonstration site.
126
-------
A detailed accounting of downtime is provided in Table 37 and shows the
unit was operating on NSSC white water for 60.2 percent of the avail-
able operating hours.
TABLE 37
SUMMARY OF DOHN TIME FOR SECOND FIELD DEMONSTRATION
NSSC WHITE I»fER
Period
From
3/6/69
3/17/69
4/8/69
4/19/69
4/28/69
5/5/69
5/19/69
5/27/69
Total
Percent
Percent
Total Hours
To Available
3/15/69
4/7/69
4/18/69
4/27/69
5/4/69
5/18/69
5/25/69
--'
of available
of downtime
190
136
238
40
142.5
80
152.5
6
985
100
Operating
89
0
217.5
0
141
0
139
6
539.5
60.2
Hours Lost Due to
Lost Pumps Liquor" Wash0
101 -- 94 3
136 -. -- 104
20,5 -- 11
40 16 -- 8
0.5 -- ' — ' '•--
80 - -- 24
13.5 13
0
391.5 29 105 139
39.8 2.9 10.6 14,3
100 7.4 26. 8 35,5
Other
4d
32*
9.5d
16*
0.5«
56h
0.3*.S
118.5
12.0
30.3
Ruptured
Modulesc
7
1
9
0
6
0
7
30t
High temperature of mill shutdown,
Includes flux checks with city water and 5000 ppm NaCl.
failure did not cause system ihutdown unless indicated,
Module rupture.
Meeting of staff and awaiting supplies,
nodule tltortage-
^Trouble in main pump electrical system.
h
Rearranged module flow pattern to reduce pressure drop,
Does not include 24 "fouled" or "plugged" modules.
J5Sto_...and Results
taken for laboratory analysis were limited in number since
the trailer was not operated over a wide range of conditions. The
soalyses generally were comparable to the information obtained during
the pilot runa» but did not exhibit -the same low variance from the
^an values, due to interruptions an operation of the unit and to
Problems arising with the greater number of modules having various
127 '
-------
degrees of leaking, plugging, etc. The analytical data and rejections
calculated by assuming that the concentration in the system is equal
to the average of the feed and final concentrate are given in Table
38. In those cases where the concentration in the: final concentrate
was not determined analytically, it was approximated. These results
agree closely with the data obtained from the pilot run. Table 39 gives
recoveries of the constituents in the final concentrate as calculated
from the rejections and recoveries provided in the previous table.
In general, the unit was operated at 525-550 psig on the main pump dis-
charge, with a feed temperature of 32-3^°C, producing a concentrated
stream of about 5.1 percent solids, and with an overall flux rate of
6.0 gfd (see Table to).. Limited data indicated a flux rate of it.9 gfd
in concentrating a feed of 1.7 percent solids to 9.5 percent. The flux
rate was substantially below that anticipated and was the subject of
much investigation.
Deterioration of the flux rates, not readily apparent due to frequent
shutdowns in early stages of operation, were plainly evident in the
longer runs. However, the flux rates recovered quickly after a shutdown
of 30 minutes or more. See Tables ^0 and ^1 for the summary of operating
data and Pig. 27 for a chronological plot of flux rates. The reasons
for flux rate recovery were not readily apparent, and were cause for
much speculation. A 30-minute shutdown might, for instance, be suffi-
cient for microbiological gas production to lift slimes from the membrane
surface. Much more likely was the probability of normal osmotic flow
taking place back through the membrane from the permeate side to the
concentrate side to flush away accumulations of fine fiber and of large
molecular weight organic solubles such as lignin from the membrane sur-
face matrix. Recovery from membrane compaction was also considered
to adequately fit the picture.
After the early runs in which the flux rate deteriorated so quickly,
a sample of the fouling material was taken from a defective module.
Chemical analysis showed a very high percentage of calcium. It was
speculated that a skin of fouling was present on the membranes when
the trailer was received from the calcium-based Interlake Mill, and
that new fouling was accumulating on this "prefilter." Consequently,
all .modules were cleaned with a solution of EDTA, a chelating agent
known to be effective on calcium ions. This cleaning resulted in an
improvement in the water flux rate to a value close to the original.
Immediately after restarting on white water, however, the flux rate
again declined; it was partially recovered either by a down period of
several hours in which it was postulated biological action could take
place or by a longer rest (several days) with city water containing
20 mg/1 fungicide.
Several techniques, such as variation in the frequency and duration
of the pulse, produced no improvements. The frequent interruption of
operations by ruptured modules, etc., prevented determination of the
rate of deterioration and whether or not a plateau of stable operation
128
-------
TABLE 38
ANALYTICAL DATA AND REJECTIONS WHILE PROCESSING
SEMICHEMICAL WHITE WATER WITH TRAILER UNIT
Water
Sample Recovery
No. Percent
1 Feed 8l
Permeate
Concentrate
ReJ . , percent
2 Feed 7)4
Permeate
Concentrate
ReJ . , percent
3 Feed 87 .
Permeate
Concentrate
ReJ . , percent
1* Feed 80
Permeate
Concentrate
ReJ . , percent
5 Feed ?6
Perm. I
Perm. II /III
Perm. IV
Perm. V
Perm. Total
Cone . I
Cone. II /III
Cone. IV
Cone. V
ReJ. I
ReJ. II/III
ReJ. IV
ReJ. V
6 Feed 72
Permeate
Concentrate
Re j . , percent
7 Feed 79
Permeate
Concentrate
ReJ . , percent
8 Feed 87
Permeate
Concentrate
ReJ . , percent
Solids.
8/1
10. Y5
0.68
57-57
98
8.9!*
0.63
3l».03
97.1
9-1*8
0.58
78.86
98.6
9.08
0.36
>*5.59
98.7
9-35
0.15
0.21
0.1*1
0.98
0.1»1
li*.i*6
17.01
29.72
38.39
99-2
98.8
98.6
97-5
10.08
0.32
36.52
98.6
7.91
0.36
38.3U
98.1*
7-38
0.1*2
58.66
96.7
Ha,
mg/1
1210
186
2820
87.7
1030
181*
9000
96.3
101*5
11*3
71*1*0
96.6
1075
92
5800
97.3
1125
37
51*
101
257
116
1760
2080
31*1*0
1*680
97.1*
97.1*
97.1
9l*. 5
1225
10 1*
MOO
96.3
925
93
1*1*1(0
96.5
905
106
6100
97-0 .
BOD,
mg/1
2938
276
96.8
2975
1*88
92.5
271*5
312
97-1
2230
2ll*
96.6
301*0
10U
ll*2
310
61*2
261*
1*313
501*6
9030
io;6oo
97.2
96.7
96.6
93.9
2908
288
95.1*
2075
265
95-6
1975
282
96.1*
COD,
mg/1
12 ,010
680
98.2
10,11*0
680
97.1
10,692
560
98.7
10,090
377
98.8
10,510
191*
261
1*67
893
1*37
11*, 920
18,1*00
31* ,720
1*1,280
98.5
98.2
98.7
97.8
11,310
367
97-8
9070
1*16
98.1*
81*20
1*39
98.8
Color
10,000
5
99.9+
5000
10
99.8
5000
7
99-9
3500
10
99.9+
5000
5
7
5
8
8
6000
8000
22,000
27,200
99-9
99.9
99- 9+
99.9+
5000
50
99-0
5000
50
99-0
5000
50
99-0
Specific
Resistance, Specific
ohm-en pH Gravity
212
1338
51.7
90.1
2Uo
1322
1*0
89.1*
257
1921
1*8.9
92.0
261
2671*
66.20
93.9
227
6281*
1*228
2189
936
2ll*5
172
ll»9
9lt.lt
76.7
96.8
96.5
95.7
91.8
3l»l
3068
88.6
93.0
317
2812
80.6
92.7
31*3
21*38
60.8
91.7
7.00
5-78
7.08
5-75
5.50
6.05
7.22
5-75
7.00
7.1*0
5.72
6.70
6.18
5.18
U.90
1».72
U.92
U.75
6,32
6.07
6.05
6.13
6.72
MB
6.25
5.85
5.38
5.83
6.03
U. 97
6.03
l.OOl*
1.028
1.005
1.039
1.005
1.033
l.OOl*
1.021
l.OOl!
1.007
1.008
i.oiU
1.018
l.OOl*
1.017
1.003
1.017
1.003
1;027
129
-------
TABLE 38 (Continued)
Sample
No.
10
11
ANALYTICAL DATA AND REJECTIONS WHILE PROCESSIHG
SEMICHEMICAL WHITE WATER WITH TRAILER UNIT
Water
Recovery Solids,
Percent g/1
Feed
Permeate
Concentrate
Rej., percent
Feed
Perm. I
Perm. II/III
Perm. IV
Perm. V
Perm. Total
Cone. I
Cone. II/III
Cone. IV
Cone. V
Rej. I
Rej, II/III
Rej. IV
Rej. V
Feed
Perm. I
Perm. II/III
Perm. IV
Perm. V
Perm. Total
Cone. I
Cone. II/III
Cone. IV
Cone. V
Rej. I
Rej. II/III
Rej. IV
Rej. V
86
91
92
8.90
0.53
62.37
98.5
Ha,
ng/1
1070
1U6
7860
96.7
BOD,
mg/1
2230
1(98
93.3
COD,
mg/1
10,020
735
98.1
Specific
Resistance, Specific
Color ohm-cm pH Gravity
5500
5
99-1
5.72
0.15
0.13
0.22
1.11
0.3li
7.53
10.26
28.1(9
61.3
97-7
98.7
99.2
98.2
660
30.5
3U.5
56
281.5
82.5
1280
1320
3160
60 1*0
96.8
97.1*
98.2
95.3
ll»55
98
103
163
610
215
i860
2560
7510
15,630
9U.1
90.6
97.8
96.1
6530
187
169
260
929
1*72
8020
10,920
28,600
78,100
97.1*
98.5
99-1
98.8
1(000
35
8
8
25
15
6500
10,000
20,000
1*0,000
99-1
99-9
99.9
99.9
268
1839
5U.9
91.2
8601*
7886
I*6Ul
917
3028
352
267
112
61. U
95.3
96.6
97.6
93.3
8.52
O.ll(
0.18
0.1(7
1.03
0.39
12.76
17. te
53.32
.Ok.99
98.6
99
99-1
99-0
1170
36.5
»*7.5
125.5
269
107
I960
2180
6160
11,580
97.7
97-8
98.0
97-7
2 1*29
118
11.6
353
601
292
3160
3960
13,520
27,850
95.8
96.3
97. >*
97-8
8670
198
226
500
939
1*39
13,31*0
18,560
69,050
133,1*00
98.2
98.8
99.3
99-3
6500
3
5
10
15
10
10,000
10,000
30,000
85,000
99.9+
99.9+
99.9+
99. 9+
295
7526
5753
2033
959
2U16
220
158
65
1*0
96.6
97.2
96.8
95.8
5.1(2 l.OOl*
U.75
5.90 1.020
6.77
5.95
,60
33
,80
.65
6.1.5
6.32
6.28
6.23
58
80
55
35
55
• U5
6.U5
,81
86
1.002
,0025
,001*
,015
1.025
1.000
001
006
022
6.03 1.01(5
could be reached. Hence, it was necessary to have technical personnel
present to operate the unit around the clock for a week. This run was
accomplished with only one interruption and did demonstrate the rate
and extent of flux decline. During the run it was decided to "rest"
or take out of service briefly every section of modules to see if the
flux rate would be improved; this proved effective but again left un-
answered the reason for the improvement.
A final run, with 2l*-hour coverage, was made to determine the effects
of slime, velocity, and of module rotation (see Table Ul). The flux
rate again declined but much less sharply; it was evident that as many
as four factors may have influenced the rate of fouling, including solu-
ble organic foulants, CaSOi, scaling, microbiological sliming and mem-
brane compaction. The tests were too brief for determining the ultimate
flux rate with the improved techniques, but they did demonstrate that
the fouling could probably be controlled.
After all tests were completed, the trailer was again cleaned with EDTA,
but the flux rate was not restored to the original value. This might
have been expected because the fouling accumulation since the first
130
-------
TABLE 39
CALCULATED RECOVERY OF FEED LIQUOR COMPONENTS
CONCENTRATE TRAILER UNIT
Sample
No,
1
2
3
U
5
6
7
8
9
10
11
Water Recovery
Recovery in Concentrate, percent
in Permeate, percent
81
7^
87
80
76
72
79
87
86
91
92
Solids
97
96
97
98
98
98
97
97
97
96
97
Sodium
82
95
92
95
95
95
95
9k
93
91
9k
BOD5
95
89
9U
9«*
95
9^
93
93
87
89
92
COD
97
96
97
98
98
97
98
98
96
96
97
Color
99+
99+
99+
99+
99+
99+
99+
99+
99+
99+
99+
Range
72-92
96-98
82-95
87-95 96-98
99+
E3DTA cleaning was probably not due to calcium deposits, and the ehe-
lating agent would not affect it much.
Samples from some of the ruptured tubes were saved and surface accumu-
lations were microscopically inspected. The brief examination indi-
cated the presence of very fine pulp fiber bonded or reinforced with
slime accumulations.
Flux rate checks were made on several occasions with city water feed;
8elected rows and nodules-were monitored'on each check to: determine
variation in fouling, fable 1+2 is -a summary of the test flux data
at various intervals. ¥ithin the banks, there is evidence that
fouling was related to the flow pattern of the lines distributing •
liquor. On an overall basis, there were several incidents of
individual rows and modules fouling nonuniformly or even improving
•*a flux despite a decrease in the overall flux. It seems significant
that Bank 1.0 o^, operating at high velocity had consistently high flux
rates. Because of the difficulty in developing adequate answers to
the fouling problem during this second field demonstration, further
studies were carried out as reported in Section VIII.. Higher operating
velocities were ultimately found to suppress the fouling by NSSC-white-
water. ;
131
-------
n
TABLE UO
SUMMARY OF OPERATING DATA REVERSE OSMOSIS TRAILER
Average of Data for Dates Shown
3/11 - 3/15
Condition:
No. of measurements
Main pump pressure,
psig
Flux rate, gfd
Bank I
Bank II/III
Bank IV
Bank V
Overall
Feed rate , gpm
Cone, rate, gpm
Feed, solids, percent
Hone. , solids , percent
Irregular
Running
5
506
6.7
5.9
5-1
5-7
32.7
6.6
0.95
5.2
it/8 - U/16
Irregular
Running
8
529
6.5
6.7
5.2
5^8
31.3
5-0
1.03
6.0
it/28 - 5/3
Continuous
(Personnel
on Shift)
35
5U8
6.5
5-6
2\Q
5-1
29.0
5.9
0.93
it-5
5/19 ~ 5/25
Continuous
Evaluation of
Velocity, etc.
35
532
7. it
7.6
7.2
U.6
7.0
28.2
U.2
0.78
5.7
3/11 - 5/25
Weighted Average
of all Readings
83
537
6.9
6.6
5.8
3-9
6.0
29.1
5.1
0.88
5.2
Single Reading
5/27
Special Run
for High Solids
1
560
7. ^
8.7
5.8
3.6
6.4.
29.0
2.6
0.90
9.5
-------
TABLE 1+1
COMPARISON OF SELECTED DATA REVERSE OSMOSIS TRAILER
Conditions U/28 - **/29 5/19 ~ 5/20 5/21 - 5/22 5/23 - 5/2U
Velocity (see below) Normal High High High
Slimicide None 20 mg/1 None None
Rotation of rows No No No Yes
Data
Hours kk 39-25a ^5-25 U?.50
No. data measurements 11 9 12 12
Main pump pressure, psig 537 51*? 529 51*?
Flux rate, gfd (Average)
Bank I 6.8 7.8 6,3 8.1
Bank II/III 6.2 8.5 7 .'5 6.9
Bank IV It.6 7.8 7-3 6.6
Bank V 3-3 5-3 1*.3 U.3
Overall 5.U 7-6 6.6 6.8
Loss in flux rate/2l4 hr
Bank I 1.1 1.0 1.1 0.2
Bank II/III 1.1 2.9 0.6 1.1
Bank IV 0.5 -O.h 0.5 0.8
Bank V ,0.7 1-0 0.2 O.U
Overall 0.8 0.7 0.7 0.6
Velocity, gpm/rov
Bank I 1.63 1-75 1-70 1.83
Bank II/III 2.57 2.92 2.78 2.69
Bank IV 1.92 2.23 2,ll* 2.13
Bank V 2.36 2.70 2.68 2.73
a •
Control failure occurred after 2k hours; first two readings after failure omitted.
b •
The flow measurement devices are known to be inaccurate, but, the relative values
should be reliable.
Pumping Energy
power' consumption factor was checked during several periods of
steady operation. The power required for the pumps and controls aver-
aged about 13-1 Kwh per 1000 gallons of feed in this field demonstration
°n NSSC white water.
:*teghanical Failure
Curing this demonstration a total of thirty modules (7.7 percent) failed
by rupture out of 38? in operation; i .e. , there' was a failure in the
"iberglass support structure resulting in a massive leak. These rupturee
occurred at varying intervals and averaged once each twenty-one hours .
'!1he distribution of failures was proportionate to the number of modules
133
-------
T.Q -
3* 5-0
3.0
2.0
9.0
a.o
a 7.0
3* 6.0
m 5,0
3.0
10,0
9.0
a 8,0
at7-01
A 5»o
4.0
3-0
- 10.0
9.0
»; 8-°
gf 7.0
6.0
5.0
9.0
8.0
5.0 _
•*. \ • N..
* ^ '. -X*
H v
/
ii
^>
•rl 1
l!
°!
Dates
C
• '•' '«"••• »t-r»i
t
Normal
velocity -
rest «TT
sections
oace for
50 Bin.
3/U-3/15 t/3-k/lfl V83-5/3
' i I l 1 1. 1 I I l l i i i ii i* ' i i i i i
) 20 >ta 6o 80 120 iw 160 180 200 220 ate 26) 28o 3oo 3» 3Uo 3&) 380 too teo
Operating Hours at Greea Bay
l
t
Control
failure
do wo 30
Bin.
i i
Wo 1*60
k
Control fstilure ~ ^3 mi a.
"t ft !f 1
» slianlciae So sltalcide No slimicide
High velocity High velocity High Telocity
Rotating (rest )
Of sections - 50 mir .
Rotating
. ., 5/19-5/25 Stopped
1 l' , , , , i ',
W30 500 580 jltO 5fc 580 600
Figure 27. Flux (gfd) vs Operating Hours at Green Bay Packaging
-------
TABLE kS
SUMNARX1 OF FLUX DATA J1CM IBSTS
3/05/69 - Overall flux on trailer, with city water, was
15-9 8« (25*0) at 550 pstg
All remaining tests are gfd at 530 pilg and 25 °C
Unit Measured
Trailer — Overall
A — All at once
B — Computed from each
bank at 530 pslg
Bank I
Sow 1«Q4~ bottom
Inlet nodule
2nd
3rd
4th
5th
Row 1,081, ^fcP
Inlet nodule
2nd
3rd
B«* Jl/ni
Hoir 2.09 fop
Inlet nodal*
2nd
3*4
Bank i?
Row 4.02 bottom
Bow 4.03 top
Row li.07 bottom
Row l».n middle
Inlet nodule
3rd
Row 4,15 top
Inlet nodule
2nd
3rd
Sow h.ifi top
3*nk V
Row 5.01 bottom
"ow 5.05 bottom
3/19/69 -
3/25/69 -
4/04/69 -
4/07/69 -
V25/69 -
5/16/69 -
5/28/69 -
6/04/69 -
3/19
10.1
12.1
9.9
15-2
11,3
10.5
8.7
8.0
9.0
6.6
5.2
5.5
5-5
8.1i
10.0
6.5
City water, after
89.25 hours
on white water
500 mg/1 BaCl solution
City water, after
City water, after
City water, after
City water, after
City water, after
City water, after
3/25 fc/4
11.7
10.2 11.5
12.3 12.3
10.0
11,3 12.5
9.3 10.3
7.8 9.0
8.1 9.9
7.8 7.7
13.7
llt.fi 15.7
9.3 12.8
13.1 15.7
10, 8 13. 3
10,2
8,8 li',l
U.I 2».0
8.0 9.7
9.1
11.1
7-7
7.2
5-5
6.2
6.7
6.6
fc.2 6.k
fc.fi 5.8
7-9 10.7
5.6
10.6 12.3
12.4 13.7
8.0 11.5
Versene wash of all banks
Versene wash (2nd) of Banks I & TV
217.5 hours
359.5 hours
504.25 hours
Versene wash
4/7
(12-9)
From 4/4
and It/7
14.0
15.1
13.0
10.9
11.5
11.1
17.0
14.1
16.7
12.1
12.0
15.5
13.1
8.7
10.8
9.8
8.8
8.5
13.6
12.1
on white water (since
on white water
on white water
of all banks
4/25 5/16
10.5 9-5
10.5 11.1
11.9 10.6/11.3
22.2
7.7
7-3
7.3
6.5
16,3
13.5
17.1
11.4 11.9
13.7
13.4
12,6
9.7 U.4
11.8
15.1
11.4
7.8
10.5
8.2
6.1
8,2
12.9
8.8
9-1 8.7
12.5
9.0
Veraetie)
5/28
8,1
8.2
8.4/8.7
Out of
•errtce
4.0
3-3
3.3
3.7
12.6
12.3
15.6
10.1
10.9
11,2
11.8
7.8
9.6
10,8
7-3
8.0
9-7
5.8
3-3
6.3
11.7
7.8
7.8
12.1
6.8
6/4
9.8
10.6
9i 8/10.0
12.7
5.6
4.9
5.3
5.3
14.1
13.0
16.2
11.6
13.8
14.1
14.5
11.7
12.6
14.8
12.8
13.3
13-6
10. li
7.1
10.4 '
14.6
10.7
8.9
12.1
9.5
135
-------
in each stage. By a distinct margin, more failures occurred in the
top rows of the"-sections than in the' middle or "bottom rows. Many of
the' modules were in tight'.groups of serial numbers, implying quality
variations between lots of modules during manufacture. The soft pulse
cycle in the trailer unit did n6t seem to reduce this module failure
rate appreciably from the rate on the small units operated with a sharp
"hard" pulse. However, most of the ruptures in the trailer did occur
on repressurization after a pulse. Sustained mechanical stress or fa-
tigue at points of imperfect manufacture probably accounted for these
pressurising•failures.
There were some twenty-four incidents- of plugged or severely fouled
modules or rows of modules'. Stripped membrane was apparently the primary
cause of this type of failure.- Very few plugged modules occurred in
the bottom rows. Proportionately by stages» the fewest were found in
Stage I and the most in Stage IV. A concentration effect may have con-
tributed to the plugging problem.
SUMMARY AID CONCLUSIONS TO THE SECOND FIELD
DEMONSTHATIQI ON NSSC WHITE WATER
The overall objective of this demonstration was to establish technical
feasibility for employment of reverse osmosis as a key concentration
step in closing the pulping and the paperboard machine effluent systems
at this mill. Technical feasibility of this new and relatively untried
unit operation could be considered well established. The demonstration
further provided a base for advancing the development of answers to
practical operating problems and for establishing the economics of
reverse osmosis concentration;
The high quality levels of recovered water were well established and
indicated the water can readily be recycled back to the mill for-effec-
tive, trouble-free use in a closed recycle system. Clear, colorless
water, free of foaming, sliming, and scaling problems of concern in
mill operations can also be cited as having acceptably passed taste
tests to many persons who have sampled the permeate.
Flux rates of water permeating through the membrane are a critical test
of economic feasibility. A rate of 7 gallons per square foot of membrane
surface per day was established in these demonstration trials. lew
and improved membrane equipment becoming available for tests as this
project'terminated in mid-1971 indicate substantially'greater flux rates
on the order' of 10 gfd overall in concentrating from 1 percent solids -
to 10 percent solids seem attainable. It should be noted that 80 to
90 percent of the water to be removed in the 1 to 10 percent co»cen»-
trating range is readily removed at high flux rates in the early stages
on dilute feed. The flux rate decreases with increasing levels of con-
centration, but minimum flux rates of 5 gfd for the final 10 to 20 percent
of the -water recovery were demonstrated in these runs.
136
-------
Flux rates were indicated to be dependent upon fouling characteristics.
Techniques to reduce fouling problems developed during and after this
demonstration depended especially upon maintaining certain minimum
velocities, proportionately "higher with increasing concentration level;
upon periodic pressure pulsing to permit cleaning of the' membrane
by normal osmotic back flow.
Mechanical problems and liquor supply problems other than membrane ecjuip-
^snt failure encountered during the demonstration reduced productivity
of data and the attainment of desired, long continuous .runs free of
interruptions. Satisfactory answers to the mechanical operating problems
seem capable of being developed in the normal course of perfecting the
equipment used.
However, membrane life and reliability of membrane module supporting
8tructures remain as the critical problems to be solved in establishing
economics of the process.
objectives of this mill program for recycling are discussed in more
. The economics of processing the liquor are further developed
in Section X.
FIELD TO. 3
CONCENTRATION OF AMMONIUM-BASE ACID SULFITE PULP
WASH WATERS BY REVERSE OSMOSIS
subsection describes laboratory and field studies for Demonstration
3 at the Oconto Falls, Wisconsin mill of Scott Paper Company. This
conducted as a feasibility study for concentration processing of
&1fflnonium-base acid sulfite pulp wash water with principal data and con-
tusions 'based upon use of the large-scale reverse osmosis trailer unit
Processing 50,000 gallons of pulp wash water per day.
A substantial amount of laboratory and pilot-scale field studies on
the various types of dilute waste flows at this mill preceded the large-'
scale field demonstration. These preliminary tests evaluated possibil-
ities for effectively processing the pulp wash waters, evaporator con-
^ensates, and the Ca hypochlorite bleach plant effluents collected from
Piping and bleaching operations of the mill. It was concluded from
these preliminary studies that the pulp wash water carried a relatively
large portion of the total pollution load from the Oconto Falls mill.
Affective complete treatment processing first required concentration
the dilute solutes. Subsequent studies were concerned especially
development of methods to process that flow by reverse osmosis
produce a concentrate for economic final processing and to recover
reusable water for recycle to the nil 1.
subsection includes the following areas of study:
137
-------
Nature of the Ammonium-base Acid Sulfite Pulp Wash Water at
Oconto Falls.
Possible Alternatives for Treatment of the Ammonium-base Acid
Sulfite Pulp Wash Water.
Design of the Experimental Program.
Laboratory Phases
Pilot-Scale Field Studies
Large-Scale Trailer Unit Field Studies
Data and Results
Pilot-Scale Field Studies
Large-Scale Trailer Unit Field Studies
Summary of Large-Scale Trailer Unit Field Studies
Module Life Experience
Technical Feasibility of Reverse Osmosis Concentration
Supplementary Studies on Concentration Processing of Hypochlorite
Bleach Effluent
Nature_of Ammonium-Base Acid Sulfite Pulp
Wash Water at Oconto Falls
Figure 28 provides a flow sheet for ammonium-base acid sulfite pulping
and bleaching operations of the Scott Mill at Oconto Falls showing points
of collection for effluents feeding to RO test equipment and indicates
several possible areas for recovery and use or recycle of RO concentrates
and permeates.
The Scott pulp mill at Oconto Falls produced 120 tons per day of bleached
pulp (70 percent softwood:30 percent hardwood) at the time of making
these field studies. Pulp wash water from this mill comprised a princi-
pal portion of the total pollution load on receiving waters as summarized
in Table ^3, Based upon the volumes and loadings indicated, the Mil
discharged 125,^00. gallons per day of wash water containing 35,000. pounds
of dissolved solids, 7800, pounds of BOD5, and 750 pounds of ammonia
daily. At this concentration the wash waters were found to have a tem-
perature range of h3 to U8°C and a pH range of 2.8 to 3-0. Both pH
and temperature may require some adjustment for concentration processing
in the membrane equipment, which was limited to operation at temperatures
not exceeding UO°C and within a pH range of 3-0 to 7.O.. The feed liquor
showed no evidence of containing pitch or scale-forming precipitates
138
-------
MHj-Base Cooking Acid
RO Concentrate._8 ffl3B
80 gpm Clear Permeate to W&sh Recycle . _ . ,
Figure 28. Plow System for Pulping and Bleaching with Effluent Collection Points for
10 Tests and Possible Routes to Beeovery and Recycle NHj-Base Acid Sulfite Pulp Mill
-------
which were of concern in processing Ca—base softwood liquors in the
first.demonstration in Appleton. However, foaming was a serious problem
in conducting the pretreatment steps for this ammonium-base liquor and
careful design of mechanical handling equipment may be required in such
steps as.screening fiber from the' liquor. Foam did not present a prob-
lem within the EO system.
TABLE k3
POLLUTION LOADING OF PULP WASH WATER
FROM THE AMMOIIUM-BASE ACID SULFITE MILL
Discharge volume = 125,itOO gallons per day
Discharged per ton of pulp = 10^5 gallons
Pulp Production = 120 tons per day
SL
Concentration
Constituent
Total solids
BOD5
COD
ra3
Temperature = 1*3 to k8°C (110 to 120°F)
pH = 2.8 to 3.0
mg/1
33,000
7,500
50,000
720
lb/1000 gal.
275
63
U17
6
Based on average volume and 3.3 percent
solids.
Possible Alternatives for Treatment of Ammonium-Base
Acid'Sulfite Pulp Wash Water
Unbleached pulp washings are a pollution control problem for most older
pulping mills, but especially for those in the sulfite industry. In
the acid sulfite ad.ll, washings may originate from operations conducted
in the digester, in blow tanks, and in subsequent washing and refining
operations, The term "washings" in this mill refers to weak drainage
produced after" the strong liquor has been displaced in the blowpit dif-
fusion washing step.
Biese trials for EO processing of pulp wash waters were undertaken at
a time when various alternatives for treatment have been under consider-
ation for development of practical answers to the pollution problem
of this outdated and relatively small pulp mill. A preferred route
-------
for treatment involves reducing the quantity of pulp wash water as in
newer modern pulp mills which usually involves large capital investment
for installation of one of the' various possible modifications of a staged
°*" count ere urrent washing system. • A survey of suitably modified pulp
washing systems is understood to "be actively under way, and one or the
°ther of these routes will probably be the chief competitor to possible
^Mployment of an RO concentration system. Much depends upon capital
investment, and importantly also upon greatly reducing the operating
charge for maintenance of the BO equipment , and importantly also upon
greatly reducing the operating charge for maintenance of the RO equip-
ment , as will be further discussed in Section X of this report.
Another alternative route to processing the pulp wash waters has been
extensively studied in the microbiological oxidation processes by trick-
ling filter, activated sludge, aerated lagoon, or even anaerobic treat-
ment. Reduction of the BOD is, of course, the chief problem of immedi-
ate concern, but there are also other components in these pulp wash
vaters which could be expected to require further treatment within the
next few years and beyond the capabilities of the biological oxidation
Processes in terms of removal of resistant organics and inorganics which
be incompletely treated, if at all by biological methods .
Search for other alternatives which could keep this mill in operation
^e understood to be under way. Many of the older' acid sulfite mills
facing this identical situation have already been shut down for lack
°f feasible answers in the past five years or so , and every effort is
made by mill management to avoid a final decision in that direction.
Design of the Experimental Program
Ig-boratory Phases
exploratory studies on RO processing of this pulp washing
effluent were conducted in the central laboratory of the Effluent
^ocesses Group at The Institute of Paper Chemistry prior to undertaking
the field demonstration. As a further step in preparation for the field
Besting with the large 50,000 gallon per day unit, small pilot and the
large-scale field demonstration units were installed and tested at Oconto
Palls. Laboratory control and development studies were conducted con-
fidently with the field runs in Oconto Falls and also in confirming
studies after these trials were completed. These laboratory studies
directed to first establishing the degree of pretreatment needed
successful operation of the reverse osmosis equipment:
Gross amounts of fiber were removed by a screening step ahead
of the membrane system. Preliminary work has shown that screens
having a mesh of kO to 120 per inch seemed to be adequate for
keeping the check valves on the pumps and also the back pressure
valves for the entire system' from plugging by gross amounts of
suspended matter. Small quantities of cellulose fiber had no
apparent deleterious effect on operation of the tubular membrane
-------
system. This was in contrast to the requirements for careful
clarification and micro filtration steps required to eliminate
plugging of capillary fiber and spiral-wound reverse osmosis
systems which were conducted in preliminary tests in the
Appleton laboratory as described in Section V.
Adjustment of the pH was necessary to have the feed liquor
within the safe operating range of U.O to 7.5 for cellulose
acetate membranes. Hydrolysis of the membranes may occur
above and below that pH range. In order to prevent membrane
hydrolysis, the feed liquor was neutralized to a pH of k.O
to 5.0 with a commercial grade of sodium hydroxide.
Temperature adjustment was also required before reverse
osmosis processing of the feed. When necessary» the feed
liquor was cooled below 35°C with use of a tubular heat
exchanger.
Foaming of the liquor was a substantial problem in the
makeshift equipment available for pretreatment and required
careful agitation and mixing of the liquor in the neutral-
ization and fiber screening steps.
Fouling of the membrane surfaces continued to be an important
concern in maintaining high-permeation rates of clean water.
NHa-base acid sulfite pulp wash water did not contain appre-
ciable amounts of resinous or colloidal particulate matter.
Therefore, no need was found for other pretreatment steps
ahead of the reverse osmosis system. However, high degrees
of turbulence and mixing were maintained across the membrane
surfaces to minimize concentration polarization and membrane
fouling effects,
Pilot-Scale Field Studies
Principal objectives in conducting the preliminary studies were to evalu-
ate Type 3 Havens modules under conditions of straight-through continuous
feeding, and at the same time to conduct controlled comparative study
of flux rates for Type k and Type 5 membranes over an extended period
of time. These studies became important because Type 3 modules installed
for life study in Appleton laboratories showed a rapid decrease in flux
rate under pressure. But the more dense structure of the Type U and
Type 5 was reported to be less subject to compaction and usually did
not show such a large initial reduction in product flux rate.
Fourteen Havens modules (four dense Type 5, one less dense Type U, and
nine relatively porous Type 3) were mounted on one of the pilot-scale
Milton Roy pumping units. The Milton Roy main pressurizing pump was
a duplex unit with each of the two pumps having an adjustable stroke
length providing a variable flow of from zero to 173 gallons per hour
and up to 1100 psig pressure. The pulp wash water was made iip
-------
continuously "by, diluting digester strength spent liquor with well water
to a, level of about 1.0 percent solids. The temperature of the wash
^"ater was 12-lU°C. The pH of the wash water was adjusted prior to re-verse
osmosis processing. Die flux rate studies were made with straight-
through continuous feeding. Both the concentrate and permeate were
allowed to flow to the sewer after sampling, limitations on pump size
and the number of modules did not permit undertaking continuous concen-
tration studies in these first tests.
Jjarge-Seale Trailer Unit Field Studies.
The objective for the larger scale studies was to demonstrate the capa-
bilities of reverse osmosis systems in concentrating dilute pulp wash
water to 10 percent or higher levels of solids under conditions of mini-
mum recycling, and with the shortest possible holding periods of the
liquor in process .
large-scale trailer unit used for these studies was designed to
Process 20,000 to 100,000 gallons per day at a maximum operating pressure
of 1000 psig. For this demonstration the trailer-mounted unit had 387
Havens 18-tube modules set up in five concentrating banks' for processing
flows at 50,000 gallons per day. Table kk lists the arrangement of
Modules in each concentration bank for the demonstration at this mill.
TABLE kk
FIRST PERIOD ARRANGEMENT OF MODULES IN LARGE-SCALE
REVERSE OSMOSIS TRAILER UNIT
(June 9 to August 11, 1969)
'Number of
Modules in
Parallel Rows Series in Each Total
Banks' in Each Bank Parallel Bow Modules'
la 18 5 90
Ib 12 3 36
II 10 '3 '30
III 8 3 21*
IV W 3 Ikk
V 21 3 63
Operation of the trailer unit at Oconto Falls was beset by numerous
Difficulties. Aging and failure of the membrane modules was a growing
Problem in the seventh month of trailer operation. During the first
Period of kj2 hours, some 63 of 38? modules were removed and replaced
1U3
-------
due mostly to tube failure and leakage. The trailer had been purchased
with a one-year warranty, and accordingly the manufacturer arranged
to repair and replace the original complement of modules. Trailer opera-
tion was halted on August 11, 1969 for removal of the modules. These
were crated and returned to San Diego where new improved tubes and mem-
branes were installed. The trailer unit was started again October 8,
1969 with 238 rebuilt modules. The modules configurations for the second
period runs made during the month of October are given in Table kj.
TABLE ^5
SECOND PERIOD ARRANGEMENT OF MODULES IN LARGE-SCALE
REVERSE OSMOSIS TRAILER UNIT
(October 1-17, 1969)
Number of
Modules in
Parallel Rows Series in Each Total
Banks in Each Bank Parallel Row Modules
la 15 fc 60
Ib 10 2 20
II 8 2 16
III 8 2 16
IV 30 3 90
V 12 3 36
The arrangement of modules in each bank was modified for the purpose
of maintaining adequate velocities at various levels of solids concen-
tration. A successful run of 211 hours was possible with this modified
set up of new modules within the 3-month contractual period of study.
However, excessive failure rates were quickly apparent with these new
modules after the first few days of the final period of operation.
Data arid Results
The small pilot-scale field studies had been mainly concerned with the
performance of Havens modules under conditions of straight-through con-
tinuous feeding. There was no attempt to study continuous concentration
of the liquor because of limitations on size and availability of pumping
equipment. Flux rate-concentration studies could be made with straight-
through flows in the larger field demonstration trailer unit under con-
ditions of minimum recycling and short holding periods for the liquor
in process. The data and results of these comparative studies in the
small and large units are reviewed in the following discussion.
-------
Pilot^Scale Field Studies
small pilot studies at Oconto Falls employed fourteen Havens modules
°f Type 3, Type k} and Type 5 membranes. Pulp wash water at about 1.0
Percent solid concentration was processed through the modules at 12
"to ll|°C and at 500 psig pressure. The temperature of pulp wash water
Was low ."because the feed liquor was made up by diluting digester-strength
spent liquor with very cold well water. This reduced the original flux
rates , since there was no arrangement for preheating the wash water
before reverse osmosis. Velocities on the order of 3.0 to 3.5 feet
Per second were maintained throughout these studies. Readings of product
flux, temperature, and pressure were taken daily. Samples of product
water from each type module, as well as feed, were taken daily and compos-
ited weekly for analysis at the Appleton laboratories. These samples
Were analyzed for solids, biochemical oxygen demand (BODs), chemical
°xygen demand (COD), ammonia-nitrogen (NHs-N), calcium (Ca) and color.
Percentage rejection ratios (R) were calculated using equation (U) in
Section IV.
Table k6 gives the average flux rate and rejection data of pulp wash
Vater for three weeks of continuous operation. Control tests permitted
conversion of observed data to establish standards for temperature and
Pressure. Table W> shows that the flux rates for relatively more porous
Type 3 membrane modules averaged about 2^ gfd when calculated at 35° C
and 500 psig pressure. The flux rates of dense Type h and Type 5 modules
varied between 9.0 and lU.U gfd at 35°C and 500 psig pressure. The
Dejections of all components for Type 3 membrane were lower than the
corresponding rejections of Type k and Type 5 membranes . Therefore ,
the higher flux rates in the case of the Type 3 membranes were at the
expense of rejections.
Since the feed to this system was an ammonium-base sulfite wash water,
it is interesting to note that very little ammonia was being transferred
"trough the membrane. Most of the ammonia appeared to be bound to or-
Sanics of sufficient molecular size to be well rejected. The relatively
high concentrations of calcium in the feed reflect calcium carry-over
from the wood and also the use of mineralized well water in the cooking
^d washing operations. All three types of membranes provided rejections
above 95 percent for all the dissolved solids components ,• except for
B°D5, which ranged from 80 to 90 percent.
The small pilot unit installed at Oconto Falls encountered some problems .
vith reciprocating pump failure due to piston scoring. This scoring
aPpeared to be accelerated by the poor lubrication quality of the low
PH wash water. There were several module failures in this straight-
trough feeding system, but all of them occurred in the older Type 3
Modules. Most of these Type 3 modules were given to us during a period
°* module supply shortages and these had been used previously on high-
^scosity tomato wastes. In addition, these modules had been trans-
ported between field trials in zero weather in an uninsulated truck
-------
TABLE fc6
FLUX RATE AMD REJECTION RATIOS OF PULP WASH WATER
SMALL PILOT UH1T
Total Solids of Feed Llfoor » 1.5 percent
Average Pressure =» 500 psig
Peed Liquor Temperature « 12-lU C
Average Flow Rate = 3.0-3.5 ft/sec
H
.*-
cr\
1st Week
Feed
Product water
Type 3
Type 5
2nd Week
Feed
Product «ater
Type 3
Type k
Type 5
3rd Wfeek
Feed
Product water
Type 3
Type k
Type 5
Flux
Rate Corrected
at 35°C.,
fffd
25.2
llf.lf
9-0
Ik.2.
9-0
23-0
lfc.0
3.8-
Coneentration,
Solids
12,060
563
too
387
11,320
580
1*09
382
11,000
501*
383
BOB
5
3555
510
326
3820
672
512
3195
605
336
COD
1^,753
719
352
15,160
838
567
1^,260
7^9
1*76
1%-S
280
2.8
1.1
0.7
220
2.6
1.1
0.7
2%6
3.0
1.0
Ca
91
2
2
2
au
2
2
2
81.
2
2
2
Color
2500
15
7-5
5
2500
10
5
5
2750
7
5
5
pH
3-5
3^9
3-9
3.6
l* .2
3.8
3.9
i*.0
1*.2
i*.2
3.9
Average Rejection, percent
t-N Ca Color
Solids BOI^ COD
9^
96
96
9%
96
96
95
96
96
.3
.7
.8
.9
A
.6
A
.5
.9
85
88
90
82
86
87
81
87
89
.5
.6
.8
A
.6
.6
.1
.7
.5
95
96
97
9^
96
96
9^
96
96
.1
-7'
.6
.5
•3
A
.8
.If
.If
99
99
99
98
99
99
98
99
.0
.6
.8
.8
.5
.7
~8
.6
97
97
97
97
97
97
97
97
97
.8
.8
.8
.6
.6
.6
.5
.5
.5
99
99
99
99
99
99
99
99
99
.k
.7
.8
.6
,8
.8
.7
.8
.8
Hbte: Color measureiaents vere made on Kellige Aqua Tester (Color Comparator) .
There, vere a total of iH- modules — 9 modules of Type 3, 1 module of Type
and k modules of Type 5.
-------
and there was evidence of ice formation in the water, contained in the
nodules, in transit.
results of this pilot study indicated that relatively high-flux
rates, averaging about 2k'. gfd at 35°C and 500. psig pressure, could be
obtained during processing of 1.0 percent solids wash water through
Havens Type 3 membrane. The flux rates were actually measured at 12
"to lk°C and the data were corrected to 35°C and 500. psig with use of
established conversion factors. The rejections of Type 3 membrane were
found to be above 95 percent for all components, except for BODg, which
Varied between 80 and 85 percent. It was concluded from these studies
"that Type 3 Havens membrane could be used effectively for the concentra-
tion of pulp wash water.
'large-Scale Trailer Unit Studies
The' larger field demonstration studies at Oconto Falls were primarily
concerned with obtaining data from straight-through concentration of
pulp wash water to 10 percent or higher levels of solids in the multi-
stage reverse osmosis trailer units. The first two months of operation
were interrupted by the previously described mechanical and module fail-
ure problems. As a result of these problems, the data and observations
are divided into two periods .
First Period — June 9 to August 11, 1969
The trailer arrived at Oconto Falls and was spotted adjacent to the
sulfite mill on June 9, 1969. By June 17 substantially all equipment,
including pumps and modules, were tested and ready for preliminary opera-
tion of the unit . The trailer unit was run for a total of kfZ hours .
that brief period, some 63 out of 387 modules failed and were removed.
Sweco vibrating screen was overloaded and damaged, allowing a wood
chip to pass through into the feed liquor. This was cause for blowing
the packing in one cylinder of the feed pump . In addition , there were
a number of brief shutdowns due to abnormally high feed liquor tempera-
ture, exceeding the capacity of the heat exchanger, failure of liquor
SuPPly from the mill, plugging of caustic lines, and failure of the
pump which supplied caustic for adjusting the pH of the feed liquor.
Wide variations were experienced in quality of the pulp washings, and
this resulted in unusually high concentrations of total solids and in
high. temperatures' of the feed stream at times. Overflow of hot, strong
Digester liquors into the wash water supply system was responsible for
these feed quality problems, and, since such discharges sometimes lasted
*"or most of a day, they forced the reverse osmosis unit to .be shut down
for corresponding periods (Table kf) . However, the most frustrating
aspect of operations was the increasingly critical problem of module
failure from leakage of seals and rupture of the membrane support struc-
tures after some 7 to 8 months of operation during Demonstrations 1,
2 .and 3-
-------
TABLE 47
SUMMARY OF TMILER DOWNTIME WHILE PROCESSING NH^-MSE PULP WSH HRTER
Period Total Hours Hours Lost foe
_ Ruptured
From To Available Operating Lost Pumps Liquor* pH Wash Other Modules
6/19/69 7/12/69 358 171 187 4 11 16 — 156C 7
7/14/69 7/18/69 - - - - Liner Sleeves for Cylinder Block in Main Pump Installed at Factory - - - -
7/19/69 8/1/69 332 203 129 73 13 1 15 2?d 26
8/1/69 8/12/69 151 92 59 3 11 -- 22 23e 22
8/12/69 10/7/69 Modules of Series 1000 and 1100 Replaced with Rebuilt Modules -
10/7/69 10/17/69 226 211 15 — 5 — — 10* 14
£. Total 1067 277 390 80 40 17 37 216 698
Percent of available 100 63.4 36.6 7.5 3.7 1.6 3.5 20,3
Percent of downtime 100 20.5 10,2 4.4 9.5 55.4
aHigh temperature or shortage of liquor..
System not shut down for module failures unless indicated.
cSearch underway for new field engineer.
^Hew screen on pretreatment line and module failures,
Module failures.
Electrical problems,
®Does not include "leakers" (48) and "plugged" (3) modules.
-------
Because .of these frequent shutdowns, it ."became difficult • to .perform
systematic studies, and few' flux rate measurements could be'made from
sustained' operation. . Table W summarizes the key flux rate.and oper-
ating variables' in the' data during this, first'period of operation. These
data indicate that the average flux rates of the' order'of 8.0 to 10.0
gfd corrected to 35°C and 600. psig pressure, were obtained while concen-
trating the feed liquor from 1-2 percent to about 6. percent solids con-
centration. The average percentage of water removal during these con-
centration runs ranged between 66 and 7^ percent.
Table ^9-giires the' rejection data for four samples collected during
this first period of operation. The percentage rejection ratios are
"based on the.average value of the concentrations of feed and concentrate.
It is noted from Table 1+9 that the rejections were all above 9^ percent,
except for BQDs, which varied from 16 to 95 percent.
During this period of June 9 to August 11, the' trailer .was operated
for about kj2 hours and experienced a number of shutdowns, mostly due
to.module failure, little in the way of sustained'and reliable data
could be obtained as a result of these frequent shutdowns. It was,
therefore, decided to replace all the original complement of modules.
The trailer was shut down on August 12 and all the modules were shipped
to the Havens factory in San Diego. Membrane tubes and seals in all of
these old modules were replaced under the original warranty agreement.
Shrouds, caps, and turnarounds which appeared to be in good condition
Were reused again in these modules. The newly refurbished modules
arrived back in Oconto Falls in the first week of October.
Second Period — October 1 to October IT, 1969
During this second- period, performance of a total of 238 rebuilt modules
"Was compared to the original setup of 387 modules used in the first
period of field trial, at Oconto Falls and in" the first two demonstrations.
The arrangement of these rebuilt modules was modified in each staged
'bank of the' trailer for the purpose of maintaining adequate velocities
necessary to minimize concentration polarization and fouling of the
surfaces. By October 8, all the .modules were tested and were
for systematic concentration studies. These rebuilt modules had
relatively high starting flux rates during the initial period, during
"Which the' membrane was compacting.
During this second period we had a successful run of 211 hours "with
Relatively few'module failures. Table 50 summarizes- the flux rates
and operating variables for each of the five concentration banks in
tersB of the number of modules, operating pressure, temperature, velocity,
and overall percent recovery of water. Figure 29 plots corrected flux
rate 'versus operating hours for these five banks operated'as four stages.
Banks II and III were run in parallel throughout these' concentration
-------
TABLE U8
SMWRY OP FLUX ABB OPERATORS VARIABLE
FOR HHj-BASE ACID SULFITE LIQUOR
(June 9 to August 11, 1969)
H
O
Operating
Hours
171
273
^72
Average
Operating
Pressure,
?sig (P)
too
kQo
590
Average
Feed Liquor Average
Temp. C
(I)
23
29
26
Velocity,
ft/sec
2.5-^.0
2.5-4.5
2.0-4.5
Concentration, g/1
Peed
21.9
12.4
14.0
6_
Concentrate
64.4
60.2
54.0
Becovery
of ¥ater»
Percent
66
79
74
Average Flux
Rate at P,
psig asd T C
gfd
k.k
5.5
6.3
Average Flux
Rate at 600
psig and 35°C
gfd
10.0
8.5
8.0
Note: Tap water flux rate at zero hours = 13 gfd at 35 C and 600 psig average pressure.
-------
TABLE
REJECTION DATA FOR W3 -EASE ACID SULFITE LIQUOR
(June 9 to August 11, 1969)
Sample
I Peed
I Permeate
I Concentrate
Rejection ratio, percent
II Feed
II Permeate
II Concentrate
Rejection ratio, percent
III Feed
III Permeate
III Concentrate
Bejection ratio, percent
TV Feed
TV Permeate
TV Concentrate
lejection ratio, percent
Concentration^ g/l
Neutral
Solids
21.93
1.7%
6%. 39
96.0
12. %3
l.%0
60.23
96.2
10.31
0.66
%%.70
97.6
18.80
0.68
63. %3
98.%
BOD
5
3.%6
1.60
—
76.5
3-73
1.31
—
88.0
2.02
0.53
—
85.0
5-76
0.65
...
9%. 6
COD
32.50
1.36
- __
97-9
26.70
2.05
__
97.%
l%.95
1.07
__
97-3
25.00
1.08
__
98.0
Calcium
0.091
O.OOO
0.22%
100.0
0.070
0.001
0.32%
99*5
o.o%%
0.00%
0.320
97.80
0.093
0.001
0.361
99.6
ffi^-N pH
0.576 %.3
0.02% %.%
1.959 %-2
98.1
0.236 3.%
o.0%3 3-9
l.%73 2.9
95.0
0.200 %.3
0.023 %-5
1.270 %.l
96.9
0.%3l %.3
0.023 %.8
i,%90 %-9
97-6
Specific
Gravity
1.006
—
1.028
--
1.002
--
1.022
1.001
--
1.015
—
1.005
—
1.023
—
Color
%500
15
__
99-9
3500
25
—
99-6
3500
15
—
99.8
5500
25
—
99.7
Optical
Density
at 281 nm
220
2
590
99-5
126
21
562
93-9
113
10
%20
96.3
205
7
626
98.3
-------
TABLE 50
SIWMART OF FLUX AHB OPSRAHB3 VARIABLES
FOR HH-BASE ACID SULFITB 1IQUGR
(Octoter 1-17, 1969}
H
vn
PO
Operating
Hours
10
Ik
IT
36
109
187
205
211
Av operating pressure,
psig (P)
Av feed temp, "C (T}
Av percent solids of feefl
liquor
Av velocity, ft/sec
Av flux rate at 600 psig
and 35 °C
Overall percent of recovery
water, percent
15BIe~con,tlnued next pageT"
Bank la and Ib
$3. of
Modules in
Operation
61*
80
80
80
80
80
80
80
80
80
Inlet
Velocity,
ft/sec
3,7
3.5
—
—
2.7
1-9
2.6
2.5
—
515
27
Measured
Flux Rate
at P pstg
and 1°C,
gfd
9-3
8.6
8A
7.2
?A
7.0
—
5.6
5.6
6,3
Corrected
Flux Rate
at 600 psig
and 35 "C,
gfd
12.9
12.1
11.8
10,1
IDA
9-9
..
T-9
7.9
8.9
Jfo. of
Modules in
Operation
32
32
32
32
38
32
32
26
36
36
Bank
Inlet
Velocity,
ft/sec
3.0
..
..
3.7
«.*»
2,7
a, 9
537
2?
it-m
Measured
Flux Bate
fet p. psig
and T°C,
gfd
8.2
7.2
7-7
7.2
7.2
6A
k.8
•t .2
%.9
k.J
Corrected
Flux Bate
at 600 psig
and 35°C,
gfd
U.1
9-7
10 A
9-7
9-7
8.6
6A
5.6
6,6
6.3
3.1
9.0
a.9
7-0
3%
-------
50 (Continued)
SUMMARr OF FLUX BATES AND OPERATING VARIABLES
FOR fiff, - BASE ACID SOtFITE LIQUOR
^(October 1-17, 1969)
vn
U)
BANK IV
Operating
Hours
4
10
Ik
17
36
in
109
187
205
211
flo. of
Modules in
Operation
81
90
90
8l
81
81
90
90
90
'90
Inlet
Velocity,
ft/ sec
3.9
4.8
—
—
5.0
5-0
4.1
4,0
--
Measure
Flux Bate
at P psig
and
-------
16
12
10
as
13
Curve Bank
2
3
I*
II, III
IV
V
Operating
Pressure,
psig
600
600
600
600
Feed
Temp, °C
35
35
35
35
Average
Velocity,
ft/sec
3.1
2.9
Average
Percent
Solids
of Liquor
3,1-14.9
It. 9-7. 5
7.5-8.9
Indicated performance curve for overall
concentration from 1.0 to 10.0$ solids
6_
I
I
I
I
80 120
Operating Hours
160
200
Figure 29. Flux Rate vs Operating Hours — Ammonia-Base Acid Sulfite Liquor
15
-------
During this run we again had substantial variation in the concentrations
of incoming feed liquor to the trailer unit. At tines. the' concentra-
tion -of the' feed' was as high as 5 percent solids. It is noted' from
the' data of Table 50 and Fig. 29 that average flux rates of the order
°f 5.5 to: 9.0 gfd (or an overall flux rate of 7.0 gfd) were maintained
at 600, psig and 35°C over a period of 211 hours . During this period,
the average concentration of the feed' liquor actually available was
3 percent solids and the feed' was concentrated to an average level of
9 percent solids.
Figure 29. also' gives the indicated performance for overall concentrations
from 1 to 10 percent solids (the range, intended for study in the original
experimental design). This performance curve was determined by taking
*nto consideration the percentage removals of water at various concen-
tration levels , and shows an overall flux rate of about 8.0 gfd in the
range of 1 to 10 percent solids concentrations. This was higher than
that of 7 gfd in the range of 3 to 9 percent solids. The higher rate
°f flux at lower starting concentrations is to "be expected since 80
Percent of the water to be removed in the 1 to 10 percent solids range
°ccurs in the early stages of concentration "below 5 percent solids.
(It should be noted that the average tap water flux rate of these Type
3 Havens modules was 13 gfd at 35°C and 600 psig pressure, }
51 gives the rejection data for two brief periods of higher level
concentration during this second period of operation. These two sets
of samples were collected when feed concentration was as high as 5 per-
cent solids, and the concentration of corresponding concentrates was
about 13 percent solids. Here the rejection ratios are again based
the average value of the concentrations of feed and concentrate.
data in Table 51 indicate that the rejections of solids, COD, calcium
and NHg-H are all above 92 percent, whereas BOD 5 rejections average
about 81* percent.
ignmary of Iarge-Scaler.; Trailer_iaiit ;Fteld Studies
Curing the first and second periods of trailer operations the unit was
operated for a total of 683 hours. During that period about 1.2 million
gallons of feed liquor was processed to produce too.,000. gallons of con-
centrate and 800,000 gallons of clean, reusable product water, The
average flux rate throughout the concentration runs was about 7 gfd
vhile concentrating the feed liquor from 3 to 9 percent solids and with
"5 percent overall recovery of water. These flux rates were determined
at 35°C and 600, psig average pressure. (The initial tap water flux
*"ate of both old and new Type 3 Havens modules was 13 gfd at 600. psig
and 35°C.)
Table 52 gives the recovery data corresponding to each of the six samples
collected during the first and second period of operation. The data
In Table 52 indicate that the percentage recovery of solids, COD, calcium,
NH-3-H, and color were all above 90 percent, whereas the' percentage recovery
of.BQDs averaged about 85 percent.
155
-------
TABLE 51
REJECTION DATA FOR iSH -BASE ACID EULFtfE LIQUOR
(October 1-1?, 1969)
t-*
v/i
ON
Samples
V Feed
V Permeate
¥ Concentrate
Rejection ratio, percent
VI Feed
VI Permeate
"TO Caoceatrate
Bejection ratio, percent
Concentration, g/i
MeutraL
Solids
50.1*8
4.13
135.91
95.6
47.23
6.44
137.35
93-1
BOD
s
13-83
3.41
32.37
85.4
12.45
4.30
3S.34
83.1
COD
69.TO
6 AT
176.20
94.7
65.90
9.60
185. to
92. 4
Calcium
0.210
0.016
0.528
95.7
0.218
0.032
0.580
92.0
NHg-N
1.436
0.106
4.022
96.1
1.578
0.205
4.463
93-7
pH
3-70
3-75
4.10
—
4.45
3-92
4.12
-_
Specific
Gravity
1.018
—
1.054
—
1.017
—
1.054
—
Color
15,000
500
—
98,2
12,500
1,300
__
94.7
Optical
Density
at 28l tan
450
a
1220
91.9
436
92
1256
89.1
-------
H
VJ1
TABLE 52
RECOVERY DATA FOR MHg-BASE ACID SULFITE LIQUOR
(June 9 to August 11 and October 1-17, 1969)
Concentration,
Sample
1
2
3
4
5
6
HQ+ f * ~PO ••
Operating
Hours
0-171
172-273
273-300
300-472
472-572
572-683
rr»*»n-h TV»r?r>vprv
g/1 solids
Feed
21.93
12.43
10.31
18.80
50.48
47.23
•ra-hin = (
Concentrate
64.39
60.23
44.70
63.43
135 .91
137.35
r Concentrate flow
Solids
96
94
96
97
95
93
rate x
Recovery Ratio, percent
BOD
5
77
85
8l
94
85
COD
97
95
96
97
--
83 92
concentrate
Calcium
100
98
96
99
94
NKj-N
98
92
95
96
96
92 93
concentration^ .
Color
100
99
99
99
98
94
i nn
Feed flow rate x feed concentration
-------
The osmotic pressure of the NHa-tase acid sulfite liquor was relatively
high compared to those of calcium-base acid sulfite and ISSC liquors,
and -varied from 125 psia at 10 grams per liter to 350 psia at 100. grace
per liter solids concentration. The effectire driving force determined
from osmotic pressure data averages about 375 psia at 600 psig operating
pressure within the range of 3 to 9 percent concentrations. Based on
the average tap water flux rate of 13 gfd at 600, psig this effective
driving force at 375 psi would have resulted in an average flux rate
of at least 8 gfd. Hie overall average flux rate obtained during the
entire operating period was 7 gfd which indicates that there probably
was some fouling of the membrane surfaces. However, the unit was "pulsed"
during the entire period by the automatic pulsing system programed
to shut down 2 minutes out of each hour. The pulsing proved to be help-
ful during this demonstration period, but it did not completely solve
the fouling problem.
Fouling of the membrane was always a matter of concern in these pilot
and trailer field studies. Therefore, special studies were made for
determining the required degrees of turbulence and mixing necessary
to minimize concentration polarization and fouling effects. These studies
were made in the Appleton laboratories after the trailer unit was moved
from Oconto Falls. The results of these studies have been discussed
separately in Sections v* and VIII which deal with reverse osmosis design
and engineering studies. The later data indicate that although veloci-
ties of 2.0 ft/sec may satisfy the theoretical requirements for adequate
turbulence and mixing with NHa-base wash water concentrations up to
levels of k.Q percent solids, the actual experimental evidence points
to a need for minimum final velocities of 3-0 ft/sec at any point in
the system to also satisfy the mass transfer requirements. Inlet veloc-
ities of not less than 2.7 ft/sec were employed in this NHa-base demon-
stration program but the average and final velocities were.substantially
less than optimum at times and may have resulted in accumulative fouling
affecting all banks of the trailer unit and especially in the final
banks working at the higher levels of concentration.
' Module Life Experience
Life performance for the original coat of 396 modules, with which the
trailer was originally equipped in Hovember 1968, proved to be about
7 to 8 months, as based upon experience at the end of the second demon-
stration. At the start of Demonstration Ho. 3, the" failing mechanical
strength of the membrane support tube apparently became a limiting
factor. The evidence of module and tubular failures appeared to increase
and to be critically dependent upon the operating pressures at 6*00, psig
or higher, It would appear that stress and fatigue effects in the tubu-
lar support structures after six months of operation were responsible
for the increasing number of module failures in the third demonstration
at Oconto Palls, Later experience was pertinent during subsequent
velocity-foialing studies conducted with the rebuilt modules over an
18-month period in the large-scale trailer unit at Appleton, when the
module life was observed to be relatively much better at pressures of
158
-------
WO-5GQ pslg. At tines the unit was operated successfully with rebuilt
nodules at the lower pressures for as.long as eight months without occur-
rence of any module failures,
Table 53 gives the history of module failures during the different period
°f trailer operation in this demonstration run at Oconto Falls. During
first period of June 9 to August 11, the trailer was operated for
total of ^73 hours, during which some 62 out of 387 nodules were re-
and replaced due to leakage, usually following tube rupture. On
August 11, 1969 fourteen tube ruptures occurred in 12 hours of actual
operation. Decision to return all modules to the factory for rebuilding
w&s made at that time and the unit was shut down August 12. The newly
refurbished nodules arrived back in Oconto Palls during the first week
of October.
TABLE 53
HISTORY OP MODULE FAILURES
frailer Unit 'Module'Failures'
Operating Plugged &
Period Hours Ruptured Leaked
June 9-July 18 0-171 6 1
July 19-25 172-273 12 0
July 26-August 11 27^72 37 6
August U-October 1 ----- -Shut Down ------
October 1-17 1»73-683 lA kk
Curing,the second and concluding period of the demonstration from October
7 to October 17, 1967, 211 hours of continuous operation were achieved
with fourteen module failures due to tube rupture out of 238 modules.
However, kk of these new nodules were removed and replaced due to leak-
aSe. Most of these leaks were found to be due to faulty sealing by
the plastic sleeves which act as a gasket or seal in the tubular connec-
tions to the end caps and turnarounds. The failure of the plastic
sleeves" proved a major problem with the rebuilt modules.
Module life is understandably a very important factor in the economics
of reverse osmosis systems. The membrane replacement costs for equip-
•fcent marketed in 1970-71 account for as much as kQ percent of the' total
Operating cost. Modules with.a minimum life expectancy of 12 months,
snd preferably 18 to 21» months" can be anticipated as, a minimum require-
ment for economic feasibility in concentration processing of dilute
Pulp and paper effluents having little or no recovery of values to help
support the'operating charges.
159
-------
Technical arid Commercial Feasibility of
Reverse Osmosis Concentration
The overall purpose of this field demonstration was to determine whether
reverse osmosis can "be feasibly developed as one of the alternatives
for concentration processing of Ms-base acid sulfite pulp wash water.
This demonstration has provided us with the first available operating
data from sustained runs on membrane processing of this type of pulp
wash water. The second period of this large-scale demonstration along
with special velocity-fouling studies described in Section VIII have
helped to determine the required degrees of turbulence and mixing for
maintaining sustained and practical levels of flux rates. No elaborate
systems or procedures for pretreating the pulp wash water were required
for this NHa-base pulp wash water beyond normal adjustment of pH, tem-
perature, and screening (100 mesh) to remove gross amounts of fiber
and particulate matter.
The economics of reverse osmosis depends first upon recovering the efflu-
ents from the pulp mill in reasonably high concentrations to provide
RO system feed liquors in sufficiently small volume to maintain low
levels of capital charge for the membrane equipment. Secondly, the
membrane module must have a minimum life expectancy of 12 months, and
preferably 18 to 2h months.
The third important condition depends on maintaining permeation of clean
water at economical flux rates. In this field trial the average feed
concentration was 3 percent solids, which was considerably higher than
the planned level of 1 percent dissolved solids. Average flux rates
of 7 gfd were maintained in concentrating this pulp wash water within
the range of 3 to 9 percent solids concentrations. For the purpose
of engineering cost calculations, an overall flux of about 8 gfd might
be expected in the range of 1 to 10 percent solids concentration, since
80 percent of the water is removed in the early stages of concentration
below 5 percent solids. However, improvements in membrane performance
from new types of equipment becoming available in 1971 can probably
be counted upon to substantially improve the cost picture (Section X).
The ammonia-base acid sulfite mill at Oconto Palls, with a rated produc-
tion of 120 tons per day, discharged a calculated average of 125,^00
gallons per day of wash water, with a BOD load of 63 pounds per ton
of product, and at 33 grams dissolved solids per liter (275 pounds per
1000 gallons) discharged about 38,533 pounds of spent liquor solids
daily in this effluent. At 90 percent recovery of these solids in an
overall concentration by reverse osmosis and by evaporation, some 17
tons of additional spent liquor solids might be recovered and marketed
in the chemical sales division or alternately be burned for heat value.
Utilization routes to disposal of the concentrate should substantially
improve the economics of reverse osmosis in the prevailing, early com-
mercialization stages of developing applications for RO in the pulp
and paper industry.
160
-------
economies of reverse osmosis processing of this ammonia-base liq.uor,
are the subject for detailed study in a comparative evaluation based
on all'five .demonstations as described in Section X and •will not be
discussed'..further in this chapter. However, the overall picture for
the demonstration provides a substantial base for serious consideration
of reverse osmosis as an effective and competitive route to concentration
9nd recovery of heat or other values in the liquor solids and with elim-
ination of one of the principal effluents contributing substantially
to the total pollution load from this mill.
Supplementary Studies
large-scale field demonstration at Oconto Palls was preceded by
a substantial amount of laboratory and pilot scale field studies on
various types of dilute wastes. The wastes included pulp wash water,
evaporator condensate, and hypochlorite bleach effluent collected from
various pulping and bleaching operations of the mill. It was concluded
from these studies that the pulp wash water comprised a relatively large
Portion of the pollution load from the Oconto Falls mill and therefore
the' large-scale field demonstration was conducted' on pulp wash water
alone. During the initial pilot-scale field trials, a 9-^eek study
Was carried out to determine "Hie feasibility for membrane concentration
Processing of a single-stage hypochlorite "bleach effluent. It was the
first of two field demonstrations on bleaching effluents and was carried
°ut prior to Demonstration Mb, k on kraft bleach liquors at Cloquet,
Minnesota. This supplementary section discusses the flux rate and re-
jection data obtained during this pilot-scale field study on hypochlorite
bleach effluent.
Concentration 'Processing of Hypoehlorite BleachJSffluent
Hie' Oconto Falls, mill discharged 3 million gallons of bleach -effluent
per day (equivalent to 25,000. gallons per ton of bleached pulp).
•& typical analysis of this dilute bleach effluent follows:
Solids = 1.5 g/1
BQD5 * 125 Mg/1
COD = 500 mg/1
Calcium « 300, mg/1
Chlorine = U50 ng/1
Color « 100
Optical density at 281 nm = 2.1-2.5
pH B 6.7-8.2
Temperature * 12-27°C
'the concentrations of the hypochlorite bleach effluent were substantially
below what might be considered to be an economical level of concentration
(5.to 10 grams per liter), but there was no practical way to.obtain
500 to 1500, gallons daily of a more concentrated hypochlorite bleach
effluent without substantial process modification and.serious disruption
161
-------
of mill operations. This small pilot trial vas, therefore, carried
out on the available supply of very dilute bleach effluent with the
use of Type 3 Calgon-Havens l8-tube modules manufactured in 1968. A
number of continuous flux rate runs were made under conditions of both
straight-through operation at low levels of concentration and also
by a recycle run to achieve higher levels of concentration. The
straight-through operations involved using nine modules set up in staged
system with 2 parallel rows of 3 modules in series followed by one
row of 3 modules in series. For recycling conditions, six modules
were connected in two parallel rows.
These bleach effluent tests were conducted in the period from January -
28 to March 6, 1969. During this time we had a total operating run
of 807 hours.
Table 5^ presents the data on the flux rates and operating variables.
Table 5^ indicates that the flux rates under straight-through conditions
on a feed liquor having 1.0-1.6 grains per liter solids were low and
leveled off at about 5.0 gfd at 600 psig and 35°C temperature. Under
recycling conditions the flux rates dropped even further because of
higher feed concentration. The concentrations during the recycling
were increased by a factor of about six times and reached 9 grams per
liter. The flux rate at this highest concentration was only 1.9 gfd
at 35°C and 600 psig pressure.
Low flux rates were apparent and indicated serious fouling problems
were being experienced at velocities of 2.9 ft/sec. Increasing the
velocities to as high as 3.^ ft. per second did not improve the flux
rate. One module was ball flushed at about 800 operating hours and
the fouling material was analyzed. The fouling material contained
19-3 percent calcium and 37-^ percent ash. Therefore, it seems the
membrane fouling was due to the accumulation of a mixture of calcium
salts and organics on the membrane surface.
Subsequent experience indicated that this fouling might have been over-
come by maintaining velocities higher than k.Q ft/sec. However, in
addition to the fouling problems, the high osmotic pressure of the
liquor would be expected to adversely affect the starting flux rates.
This is understandable because the osmotic pressure of a liquor in-
creases rapidly in the presence of low molecular weight inorganic
materials. This hypochlorite bleach effluent contained about 67 percent
inorganics in the form of calcium chloride, and the organics represented
only a small portion of the total material.
i
Table 55 gives the rejection data for hypochlorite bleach effluents.
Percentage rejections for all components are fairly good as summarized
below:
Solids 87-90
BOD5 70-95
COD 85-99
162
-------
Average
Pressure,
psig
(P)
Average
Temper-
ature,
(T)
Average
Tnlet
XKLLC O
Velocity,
jj
ft/sec
Average
Flux
Rate at
P, psig
and T'C
gfd
Flux Rate
Corrected
to 600
psig and
35 °C
TABLE 51*-
SUMMARY OF FLUX RATES AND OPERATING VARIABLES FOR HYPOCHLORITE BLEACH EFFLUENT
Average
Concen-
tration
Operating of Feed,
Hours g/1
Straight-Through: 2 Parallel Rows of 3 Modules* in Series Followed by
1 Row of 3 Modules in Series
0-101 1.6 610.0 13-0 2.7 5.3 7.5
102-202 1.0 615.0 13.0 2.7 k.2. 6.0
803-232 1.5 610.0 12.0 2.7 3-3 ^.8
Recycle: Six l8-Tube Havens Modules - 2 Parallel Rows of 3 Modules in
Series in Each Parallel Row
833- 3W
3^9-508
509-556°
557-705
706-7^3
3.6
*.5
5.0
9-1
6.U
•6.0
580.0
570.0
580.0
560.0
565.0
570.0
23.0
26.0
23.0
27.0
26.0
2»f.O
3-*
3.*
2.7
2.7
2.7
2.7
2.5
2.2
—
1.5
1.9
3-1
3.3
2.8
—
1.9
2.5
u.o
water flux rate = 20 at 35 "C and 600 psig average pressure.
is average inlet velocity to Banks I/II. Average Inlet velocity to Bank III
Varied lf.6-5.0 ft/sec . depending on flux rate.
unit was down briefly due to module rupture in the 509-556 operating period.
At "jkk hours, the modules were ball flushed before the flux rate run.
Calcium 82-92
Chloride 80-90
Color 9^-99
rather remarkable phenomenon was observed with the reduced pH of
permeate. The pH of the permeate for the first sample was 2.7 ,
U&its lower than the pH of feed and concentrate. The difference in
PH became less with increasing operating time, and finally there was
tto difference at about 600 operating hours . This might indicate a
8low buildup of fouling materials to provide a dynamically formed mem-
brane with greater levels of rejection for components affecting the
PH than that achieved by the cellulose acetate membrane alone.
163
-------
1ABLI 55
HJICKEON BA2A. FOH HXPOCHLOEEEE BLEACH ElfWJElH?
Bejeetioa Ratio, percent
Sample
13.
Feed
PH
Concen-
trate
Permeate
Average
Concentration
of Peed, g/1
Solids
BOD
5
COD
Calcium
Chloride
Optical
Density
at 28l rm
Color
Straight-Through
1
2
3
6.7
7-1
7-1
6.7
7-0
7.1
4.0
fc.l
5.8 .
I.Ik
i.to
l.Oit-
91.0
89.0
88.0
68.0
73-0
8^.0
85.0
88.0
85.0
92.0
89.0
90.0
88.0
89.0
83.0
99-0
97.0
97-0
96.0
95.0
9^.0
Recycle
fc
£
6
7
8
9
10
n
7.2
7-3
7.6
8.2
7-6
7A
7.0
7-7
7.1
7.2
7.7
8.8
7.6
7.%
7.0
7-7
6.2
6.7
7-1
7-5
7-6
7.2
6.6
7-3
l.^S
1.28
3.56
Ml
9-28
8.86
t.86
7.19
87.0
86.0
85.0
85.0
88.0
89.0
87.0
91.0
77.0
78.0
83.0
93.0
95.0
97-0
88.0
91.0
87.0
86.0
95.0
96-.0
99.0
99-0
97.0
9&.0
82.0
Slf.O
86.0
91.0
89.0
88.0
87.0
89.0
77-0
83.0
79.0
77.0
81.0
82.0
82.0
87-0
99-0
98.0
99-0
98.0
99-0
99.0
98.0
99.0
95-0
92.0
99-0
98.0
99-0
99-0
9^.0
98.0
-------
It can be concluded that membrane fouling was a serious problem during
these bleach effluent studies, and any further investigation of the
treatment of hypochlorite bleach effluent which might be undertaken
vith use of RO should be aimed at finding methods for reducing the degree
°f fouling. One of the methods worth considering is to maintain higher
fegrees of turbulence and mixing across the membrane surface. Another
Possible route may be the employment of pretreatment steps for the bleach
ef fluent feed ahead of the membrane system. lime and ferric treatment ,
Edition of poly electrolytes, coagulation, and filtration are examples
°f such pretreatment s which might be considered.
CONCLUSIONS
field demonstration of reverse osmosis concentration processing
°f ammonia-base acid sulfite pulp wash water operated at flow rates
°n the order of 50,000 gallons per day. The trailer was operated for
a total of 683 hours, during which about 1.2 million; gallons of liquor
Were processed to produce U00,000 gal. of 10 percent TS concentrate
800,000 gallons of clean reusable water.
Serious problems of membrane module failure were encountered throughout
this field trial.
Average flux rates of 7 gallons per square foot per day were maintained
in concentrating the pulp wash water in the range of 3 to 9 percent
solicis concentrations. For the purpose of engineering cbst calculations,
*& overall flux rate of at least 8 gallons per square foot per day might
be expected with use of membrane equipment available in 1968-70 to
Achieve a concentration in the range of 1 to 10 percent solids, as based
^pon 80 percent of the water being removed in early stages of concentra-
tion below 5 percent solids.
elaborate or expensive pretreatment of pulp wash water was required
d nominal amounts of temperature adjustment, pH control, and sus-
pended solids removal through a 100-mesh screen.
pulp wash waters processed in .this third demonstration contained
Relatively small amounts of colloidal and particulate matter. These
an>ounts were not shown to be of concern in these trials , since we were
&ble to control the fouling of the membrane surfaces by maintaining
Moderate levels of velocity of flow. Velocity of flow at higher levels
°f concentration was indicated to be an important means of maintaining
high overall rates of flux, however. RO processing of Ca hypochlorite
bleach effluent did not appear promising under the conditions of pro-
Cessing the very dilute feed liquor available for the test runs under-
taken.
165
-------
FIELD DEMONSTRATION NO. U
CONCENTRATION PROCESSING OF CAUSTIC EXTRACTION
KRAFT BLEACH EFFLUENT BY REVERSE OSMOSIS
This section describes laboratory and field studies conducted as Demon-
stration No. h in the series conducted under Federal Research and
Demonstration Grant 120^0 EEL to the Effluent Processes Group at The
Institute of Paper Chemistry. The demonstration field trial was con-
ducted, at the Cloquet, Minnesota kraft pulp mill of The Northwest Paper
Company Division of Potlatch Forests as a feasibility study for con-
centration processing of alkali extraction kraft bleach plant effluent
(alkali extraction KBE).
This demonstration was preceded by a substantial amount of laboratory
and small-scale pilot study on various types of bleach plant effluents
from various mills and various types and sequences of bleaching opera-
tions . (See Section V.) It has been concluded that the alkaline extrac-
tion KBE and particularly the second-stage effluent in bleach sequences
such as CEDED comprises one of the more serious pollution problems from
kraft mill bleach plants. A large part of the pollution loading from
bleaching operations from the Cloquet mill derives from this source
(secondary alkali extraction KBE). As such, it has been of particular
interest to develop new methods for treating this type of bleach waste
as alternatives to conventional disposal routes, such as by bio-oxidation.
Problems were recognized in undertaking these studies. High levels
of dilution have been of particular concern. The Cloquet kraft mill
was designed to produce fine quality grades of bleached kraft paper
by the best methods prevailing at the time of its construction.
Accepted good bleaching practice at that time called for using large
amounts of water for washing the pulp. The bleach plant effluent which
could be made available at this mill was, therefore, representative
of the very dilute wash water characteristic of this type of bleach
washing operation. These large volumes of bleach effluents, with low
concentrations of dissolved solids, are more dilute than would be desir-
able for a commercial RO operation for concentrating and disposing of
the bleach plant effluent.
More modern bleach plant design and bleaching practices are proving
that large volumes of total bleach plant wash waters can be substan-
tially reduced from the prevailing levels of 10,000 to 50,000 gallons
per ton of bleached pulp to volumes as low as 6000 gallons per ton.
With this in mind, planning proceeded for a field demonstration trial
of reverse osmosis processing of the alkaline extraction KBE, which
is presently available in volumes estimated to be on the order of 7000
gal./ton pulp withiri the.total volume of all bleach effluents at this
mill. The objective was to be directed to concentrating this dilute
effluent having less than 1 percent solids to an intermediate stage
concentrate at around 10 percent solids content in one-tenth or less
of the original volume. It could then be expected that the 10 percent
166
-------
solids concentrate could be feasibly processed for final disposal or
utilization to achieve much more complete pollution control than by
Mo-oxidation Of the dilute flows. Methods for final processing of
the 10 percent solids concentrate of "bleach liquors are known to be
developing as a new phase of research on bleach plant effluents and
could feasibly include further concentration of the 10 percent solids
Product to 50 percent or higher for combustion or other suitable methods
°f final disposal.
NATURE OF BLEACH PLANT TREATMENT PBOBLEMS
Bleach plant effluents from pulp mills represent a significant pollution
loading on receiving waters , Suspended fiber in these effluents first
requires clarification treatment. Dissolved solids may range to several
hundred pounds per ton of bleached pulp and the BODg from 10 to 50 pounds
Per ton. Special problems are apparent in respect to color, foaming,
content of resistant organics, inorganic salts, and also components
toxic to fish may be present.
Kraft mills , especially those in mild climates , may employ conventional
Primary clarification and secondary bio-oxidation techniques to treat
Dilute bleach wastes for removal of suspended solids and BODs. Problems
^ay arise If more complete treatment is required. A 1968 survey12
showed that 82 out of a total of 112 mills with combined kraft mill
effluents provided primary treatment, and 55 provided some form of sec-
ondary treatment, such as storage oxidation basins, aerated stabiliza-
tion basins, or activated sludge.
Biological oxidation does not significantly reduce the color of these
effluents. It is knowii that liine precipitation treatment will substan-
tially reduce the refractory organic colored components13. The massive
lime process which has been developed through larger scale pilot planting
is presently being tried commercially11*. Recently, work on coagula-
tion15!16 has shown that Al*8 and Pe*8 can be used to reduce the color
°£ both chlorination stage and caustic extraction bleach wastes . Ap
^ch as 96.5 percent of the color and 85 percent of the total organic
°arbon are removed under optimum conditions from the alkaline extraction
^SE. However, settling and dewatering of the resultant floe has been
ahown to be a problem.
1 1 has shown that activated carbon can be effective in reduction
of 1die color in caustic extraction KBE. This nay account for 60-80
Percent of the color in the total of all KBE streams .
0116 important objective in treating bleach effluents is to recover clear
Water for reuse, and particularly for recycle to final wash stages.
F°r this there is need to know more about the quality of water required
f°r use in bleaching and pulp washing. Smith and Berger18 have studied
^e route to this end. They show that reusable process water from total
Bleached kraft mill effluents could be obtained by a process comprised
of Primary clarification of suspended solids, massive lime precipitation
16?
-------
for color reduction, "bio-oxidation to reduce BOB, and a final treatment
"with activated carbon. This four-step processing was estimated to cost
lU.5^/1000 gallons. The addition of ion exchange demineralization might
add 25^/1000. gallons in the case of alkaline extraction KBE» "but removal
of chlorides from total "bleach effluents remains as a substantial prob-
lem. The degree of recycle of NaCl which can "be tolerated in a "bleach
system may be fairly high in some situations, less in bthers.
Water quality necessary in the production of bleached and unbleached
kraft pulp as compiled by Berger19 is given in Table 56.
56 :
RAISES OF PROPERTIES OP PROCESS FOR
KRAFT PULPING AND BLEACHIIG OPERATIONS
Unble ache d Ble ache d
Turbidity, units' 5-25 0-5
Color units 10-80 0-5
pH 6.5-8.0 6.8-7-3
Total alkalinity, mg/1 " 20-150 20-75
Hardness, (as CaCOs), mg/1 5-200 5-100
Dissolved solids, mg/1 50-500 50-250
'Chloride, mg/1 1Q-1501 "- -10-150
Iron, mg/1 0.5 max. 0.2 max.
'Manganese, mg/1 0.3 max. 0.1 max.
COD, mg/1 0-12 0-8
BOD5, mg/1 0-5 0-2
A large proportion of the existing bleach, plant operations employ large
volumes of water with effluent volumes ranging.from 20,000 to ltO,000
gallons per ton of product. Under these conditions, research and de-
velopment objectives as well as practical installations have understand"
ably been directed to appropriate disposal or reuse methods of - treatment
based on handling such large volumes of relatively weak effluents.
A variety of methods are being developed to accomplish volume reduction"
Continuous bleaching systems are a principal route in new plants-
Existing bleach plants may be modified to,various degrees for
168
-------
count ercur rent washing, for diffusion washing20, for extracting and
pressing of high consistency pulps, and by innovating in various other
. Indications are that a higher concentration may be obtained in
mills as an incentive for changing the bleach system in coming
years . It becomes reasonable to consider new, alternative methods for
complete treatment of bleach wastes where low volumes of flow in
range of 7000 gallons per ton of pulp are available21. Recovery
reuse of bleach chemicals as well as recovery and recycle of reusable
can be brought into focus as additional objectives for overall
'bleach effluent treatment with availability of concentrates at 10 percent
solids or higher. Spent bleaching chemicals in the form of 200, to 300
Pounds of sodium chloride per ton of bleached pulp could lead to recovery
°f as much as 50 tons of Nad per day in a 500-ton bleach plant.
CHABACTERIZATION OP KRAFT BLEACH EFFLUENT AT CLOQUET
studies in this demonstration trial of reverse osmosis were carried
°ut on alkaline extraction KBE from kraft pulping of pinewood. The
Wll is rated at 185 tons daily of softwood kraft pulp and 120 tons
°f hardwood kraft pulp.
30 provides a flow sheet for the kraft pulp washing and bleaching
°f the pine softwood pulps . Some special studies for concentration
Processing of the rewash water from the pine pulp mill are reported
ln a final section of this report, but this section is primarily con-
cerned with the alkaline extraction KBE from the second stage in the
Bleaching sequence. Dilute liquor from the seal box from the second-
stage washer was piped at about 0.25 percent solids concentration to
" reverse osmosis pilot unit installed at this mill, see' photo, Fig.
(Appendix B). Similar effluent was collected and sent to Appleton
laboratory studies conducted before, during, and after the pilot
in Field Demonstration No. k,
the demonstration run, about 1500 gallons of caustic extraction
effluent were processed daily by reverse osmosis. The mill staff calcu-
that about 30,000 gallons of total bleach plant effluent were
ved from each ton of bleach pulp production, and that the alkaline
accounted for about 7000 gallons of this total effluent from each
of pulp production.
57 provides average analytical data from 13 samples of the dilute
liquor collected during these field and laboratory studies on mem-
processing of the kraft extraction bleach plant effluent for
No. U. The dilution of this bleach plant effluent was
greater than desired, but represented the best concentration obtain
at this mill at the time of these studies. Solids concentration
at 2,6 grains per liter is about one-fourth of the desired 1 percent
which might be considered economically feasible for commercial
The BOD 5 and COD of this liquor were fairly high at 190 ,
mg/1, respectively. Color is also a characteristic of concern
169
-------
Fine Kraft Pulp Washing
From
Foam
System
(T"
Pulp
Washer
ll
I
r
P
1
r •
Pulp
tfashei
«
1
|"~
MP
1
|
Re-
Clear Water
to RO
to weak black
liquor stor-
age and
evaporatiToh
-8
aca
M
ce
Storage
figaxe 30* Karaft Mill Pulp Meshing and Bleaching Schematic
-------
pollution control. Much of the color deriving from the entire bleach-
operation can be found in the caustic extraction effluent.
TABLE 57
HO KRAFT CAUSTIC EXTRACTION KBE
FEED CONCENTRATIONS
Solids 2630 mg/1
BODs 190 mg/1
COD llHO mg/1
Optical density, 281 nm 18,2
Sodium 610 mg/1
Chlorides k$Q ng/1
Color 6000
*able 58 provides additional analytical information showing the Tariation
1R analysis for samples obtained from Stage 1 (chlorination) and Stage
(caustic extraction) for a variety of types and kinds of bleach
, as well as kraf t softwood caustic extract. The differences
especially dependent upon the degree of dilution characteristic
each different pulp mill and of the different types of pulp and stages
bleaching. The purpose in presenting Table 58 is to help in showing
importance of having a good- 'analytical characterization of the par-
ticular type of liquor to be studied in any individual mill operation,
to place the results of the studies reported for this demonstration
proper perspective.
dilution of the liquor available for feeding to the reverse osmosis
ern in these studies presented no technical problems. RO concentra-
tion "by factors as high as 60 times could be demonstrated in these studies.
Dilution is an economic problem in terms of capital cost and operating
es for concentration systems. Economic feasibility of any eommer-
operation to be undertaken at this mill would be substantially
if the feed liquor were reduced to about one-fourth of its
and then to employ a concentration factor of perhaps 15 times
the membrane processing step. This point will be discussed in
detail in the concluding section of this report.
171
-------
TABLE 58
VARIATION IN ANALYTICAL CHARACTERISTICS
OF DIFFERENT BLEACH LIQUORS
Solids BOD pH Temp.,
mg/1It/ton mg/1'To/ton Approx. C
STAGE 1 (Chlorinatlon)
Hardwood Hi 869 43 169 20 2 21
Kraft Low 141 8.3
Softwood Hi 11,600 214 294 25 2.1 21
Kraft Low 594 67 102 6
Hardwood Hi 1,180 245 159 24.9 2.2
Sulfite Low 1,000 149 102 20.1 1.96
Softwood Hi 2,280 321 331 28.1 2.52
Sulfite Low 640 94 38 7.6 1.82
STAGE 2 (Caustic Extraction)
Hardwood Hi 1,750 83 179 14 10 g5 5
Kraft Low 125 6 10
Softwood Hi 5,980 223 877 12,2 9.9 65-5
Kraft Low 1,800 40 46 5.7
Hardwood Hi 4,020 835 575 117.4 7.54
Sulfite Low 1,390 207 196 29.6 6.72
Softwood Hi 2,780 391 420 59.1 7.73
Sulfite Low 490 70 25 5.4 6.65
DD1TCT OF THE EXPERIMENTAL PROGRAM
Laboratory Phases
A great deal of laboratory work in the central laboratory of the Effluent
Processes Group at The Institute of Paper Chemistry was conducted prior
to undertaking the field demonstration run with equipment installed
at Cloquet. Laboratory studies were also conducted concurrently with
the field run in Cloquet and also in confirming studies after that run
was completed. These laboratory studies were directed first to estab-
lishing the degree of pretreatment needed for successful operation of
the reverse osmosis equipment. The findings were similar to those in
the previous demonstrations:
1T2
-------
Gross amounts of fiber were removed by ho to 100. mesh screening
ahead of the membrane system.
Adjustment of pH may be necessary to have the' feed liquor "within
the safe operating range of pH 3.0 "to 7.0 for cellulose acetate
Btetnbranes . Several sources of acid were evaluated in the labor-
atory studies for neutralizing the alkaline extract. For
Demonstration No. k sulfurie acid was used to neutralize to a
PH of about 7-0 +_ 0.2. There is still some question as to the
exaet level of pH adjustment which may actually be needed in
commercial operation. Manufacturers of reverse osmosis equip-
^nt using cellulose acetate membranes normally specify a range
Between K.O and 7-0, but in some cases pH levels substantially
above and below this range apparently have been tolerated
Without evidence of membrane hydrolysis. A pH of 7*0 seemed
to "be entirely satisfactory for all studies conducted under
this demonstration on caustic extraction KBE. A technical-
Srade sulfuric acid was used for the purpose of maintaining
reliable and close control in these studies, but it was also
'fewonstrated that other sources of acid, such as the chlori-
n«fcion stage KBE, could be used satisfactorily to achieve
the same purpose and at much lower cost.
Temperature adjustment was not required for processing the
Bleach plant effluents in this particular study, although
some bleach effluents have a relatively high temperature and
^ght require cooling below about Uo°C to avoid damage to
the membranes . On the other hand, some effluents are too
copl for obtaining good flux rates and might advantageously
"e warmed to above 25° C. Average operations for Demonstration
*°« k were carried out at a temperature range of 30 to 35°C.
problems have always been a matter of concern in
Developing procedures for processing each new type of waste
Ho need was found for other pretreataent steps ahead
the reverse osmosis system. However, careful evaluation
made of possible need for preventing fouling of the
during these studies. No serious problems resulted
after the initial laboratory work developed operating param-
eters to avoid fouling. At lower velocities and degrees of
turbulence, it was found advantageous to practice the pulsing
Procedure of operating the system under periodic release of
the membrane system from high pressure for periods of 30
s®conds to several minutes every hour or so. Later experience
v°uld Seem to indicate that if the velocity of flow is greater
than the levels required to prevent concentration polarization,
^e need for the so-called pulsing operation may be greatly
Educed. However, the periodic pressure reductions still
Seein to have advantage when carried out on a daily basis,
^d may be related to membrane compaction as well as to
c°ncentration polarization and actual fouling of the membranes.
173
-------
Foulings as such, did not appear to be a problem in this
field demonstration. Fouling can occur to different
degrees with different types of pulping and bleaching
liquors. Where fermentable materials, such as wood sugars,
are present in the spent liquor, steps should be taken to
prevent microbiological sliming of the membranes. Sliming
may occur rapidly during shutdown periods of a day or more
if liquor stands in contact with the membranes; whereas
continuous operation at sufficiently high velocities may
keep the membranes clean indefinitely-. No problems were
experienced with membrane fouling during continuous
operation in this field trial.
Pilot-Plant Field Study
A substantial amount of experience with operating conditions at the
Cloquet mill of The Northwest Paper Company had been gained in an earlier
study on concentrating kraft pulping rewash waters during the summer
months of 1968. The mill staff subsequently determined that planned
changes in the pulping process would eventually eliminate the rewash
waters as a waste flow from this mill. Plans for Demonstration Wo.
k were, therefore, changed to develop methods for concentration pro-
cessing of the caustic extraction KBE. For this modified demonstration
preliminary data and experience were needed with one of the smaller
pilot-scale field units operating at the 1500 gal./day level. It was
further decided to develop all data from this trial with the small
unit. Scale-up data were available from the first three demonstrations
using both the small and large 50,000 gal./day field units. Long delays
were being experienced in obtaining reliable and properly tested mem-
brane modules then being produced in early stages of commercial produc-
tion by the several equipment suppliers active in this new field of
membrane processing. It was not yet possible to Justify the large
capital expenditure necessary for replacing the early prototype modules
with which the large field unit had been equipped for Demonstrations
1, 2, and 3.
Under these circumstances, it was agreed to employ one of the smaller
pilot scale Milton-Roy Duplex pumping units with capacity up to 6 gal-
lons per minute, together with a membrane unit containing 20 modules
of the Type 3 configuration available from the larger trailer unit.
Later in the run a bank of five Type 3 modules of the mid-1970 design
were installed. This unit was placed on stream at Cloquet, Minnesota
effective September 15» 1970 and ran through December 15, 1970.
Description of Equipment and Fretreatment Procedure
The equipment for pretreatment and the reverse osmosis equipment were
substantially the same for each of the three phases of the overall
study. The second-stage washer effluent from the rotary vacuum washer
was taken directly from the recycle loop at the washer seal box, and
-------
was piped approximately 50 feet through flexible plastic hose to a
100-mesh nylon side hill screen. The screened Tbleach plant effluent
passed through to a llQ-gallon stainless steel mix tank where dilute
sulfur!c aeid was added under automatic control to adjust the pH to
7.0. A relatively constant temperature was maintained by means of
& cooling coil in the mix tank. The pretreated liquor was then pumped
to a 50-gallon polyethylene holding tank from which it flowed by equal-
ized level to two 50-gallon plastic feed tanks. These two feed tanks
Biade it possible to service two separate pressurized reverse osmosis
systems set up in either series or in parallel configuration. The
Wilton Roy main pressurizing pump was a duplex unit with each of the
tvo pumps having an adjustable stroke length rated from zero to slightly
less than three gallons per minute and up to 1100 psi. Draw-off of
the concentrate from the final feed tanks was accomplished by means
°f a Brosites metering pump. The peiteeate was allowed to flow to the
sewer after sampling. A flow diagram for the system is shown in Fig.
31.
DATA AND RESULTS
It was not possible to accomplish complete, straight-through concentra-
tion from 1 to 10 percent dissolved solids with the small-scale equip-
nient because of the limitation on the number of stages that could be
studied with a limited number of modules and with no more than two
Pumping stages. The program for the 3-Jnonth study was, therefore,
divided into three stages of concentration. Phase It for the period
September 21 through October 8, was a dilute phase study starting with
bleach effluent having a feed concentration of only 2 to 3 grams per
liter as received from the mill. This was concentrated to about half
the original volume and to a dissolved solids level ranging between
and 5 grams per liter. The second phase provided one month of study
concentrating in the range of about 1 percent solids to upwards
of 5 percent solids. The third phase carried the concentration on
^P to as high as 15 percent solids (160 grains per liter). An outline
°f the planned operating conditions for these three phases of study
*8 provided in Table 59, and more ..detailed operating data are provided
and described later in Tables 62, 63, and 6^.
Phase I — Data and Results
parallel systems were set up each with five modules in series under
A and B. Initially, Bank A was fed at a velocity of about 2.5
feet per second, while Bank B was fed at a lower velocity of about
I-7 feet per second. A steady state for the flux rate was established
&£ter about eight days of nearly continuous.operation. The flux rates
were fairly high in the first days but subsequently dropped substan-
tially over 8 days of operation due to the low feed velocities.
levels of concentration were achieved in the range of 3 to 3.5
per liter in the concentrate. Rejections were quite good in
175
-------
Wash Water from
Seal Box
Side htl
Screen
100 mesh
H
-J
a\
Acid
Supply
Tank
Mix Tank
100 Gal. S/S
50 Gal.
Plastic
Hold
Tank
50 Gal.
Plastic
Bank A
Feed
Tank
\
1
LC
I
**^k
1
50 Gal.
Plastic
Bank B
Vr,^A
j
Meter:
ng
-.
fe&)
H«-h
Tank
Pomp
Concentrate
Collection
Milton Roy Duplex Pump
0.3 - 6.0 gpia
1135 psi maximum
Figure 31. Flowsheet-RO Processing of Alkaline Extraction KBE
-------
TABLE 59
OUTLINE OP OPERATING CONDITIONS
RO PROCESSING OF ALKALINE EXTRACTION KBE
I-1
Date
Phase I
9/21 to 9/29
10/1 to 1013
Phase II
10/14 to 10/22
10/22 to 11/5
Phase III
11/5 to 12/3
12/3 to 12/16
Module
Bank
A
B
A
B
A
B
B
A
B
A
B
A
Stage
1
1
I
1
1
1
1
2
I
2
1
2
Recycle
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Solids
Concen-
trate
K/l
In
2
2
3
3
2.5
2.5
2.5
5
2.5
7
2.5
7
Out
3
5
4
4
40
35
6
82
22
130
20
160
Vel.,
In
2.6
1.6
4.2
3.2
3.2
4.2
4.2
4.2
3.2
4.2
2.2
3.2
ft /sec
Out
1.7
0.7
3.7
2.5
2.5
3.9
3.9
4.0
2.5
3.7
1.2
2.7
Pressure
In (max)
515
500
555
520
525
520
590
475
520
810
515
795
, psig Number
AP of Seri
35
35
160
55
85
100
185
100
130
230
30
155
5
5
5
5
5
5
5
5
10
5
9
5
and Type
es Modules
"used"*
"used"
"used"
"used"
"used"
"used"
"used"
"used
"used"
"used"
new
a "Ifeed" modules were Type 3 modules which had seen prior use both at previous demonstration site
for-pilot studies, and were prone to seal leaks or a tube rupture at pressures over 500 psig.
Sew modules are Type 3 incorporating redesigned heads and turn arounds and rated for service at
800 psig.
-------
the range of 90 percent or higher for most components, except BOD5 ,
which was about "J6 percent rejected in the first week.
In the second week the feed rates were raised to correspond to inlet
velocities of k.2 and 3.2 ft/sec. The rate and extent of the flux
rate decline decreased so as to give steady state flux rates of 9-1
and 9.0 gfd, respectively, when adjusted to 600 psig and 35°C.
Problems with module seal leakage became more serious as the inlet
pressures were increased to give the same average system pressure
psig) at the higher velocities . Problems with leaky modules were
apparent in the form of temporary development of a substantial amount
of color in the permeate when the average system pressure exceeded
about 500 pounds per square inch in these old modules. A shutdown
for a short period of 5 minutes resulted in closure of the leaky seals
in the heads of the modules a few minutes after operation was resumed.
Studies during the first month of operation confirmed the importance
of maintaining initial feed velocities to the modules above 2 feet
per second. A characteristic fall-off in flux rates occurred at lower
levels of velocity, but the flux rates rose substantially again when
the velocities were increased to 3.2 and h.2 feet per second for Banks
A and B in the later stages of Phase I. Comparison of steady state
flux rates obtained at various velocities of flow in the module are
shown in Table 60.
TABLE 60
STEADY STATE FLUX RATES OF
ALKALINE EXTRACTION KBE
(Without pressure pulsing)
Average
Velocity, Product Water Flux Rate
ft/sec at 1*85 psig and 35 °C, gfd
1.1 3.8
2.1 U.8
2.8 7-1
3.9 7-2
Initial 11
Table 6l presents data summarizing the average rejections for the dilute
feed as determined from nine samples taken during straight-through
operation. It was apparent that a substantial part of the dissolved
solids were of low molecular weight, since it is normal to expect the
1T8
-------
rejection of dissolved solids of pulp wash waters to exceed 90 percent
and usually better than 95 percent. Two low readings, apparently
caused "by seal leakage, indicated by a and **, significantly affect
the high average rejection of color and optical density.
TABLE 61
AVERAGE REJECTIONS OF PHASE I (STRAIGHT-THROUGH
PROCESSING) (PERCENT)
ALKALINE EXTRACTION KBE
Rejection, percent
(l-Cp/Cc)lOO,
Constituent av of 9
Total solids 85. h
BOD5 7^.2
COD 93. k
Optical density at 28l nm 97.7&
Sodium 77.3
Chlorides 6k. 9
Color 97«9b
9.
Seven determinations of nine ranged 98-5 to
99.2 percent.
Seven determinations of nine ranged 99.6 to
99'9 percent.
Detailed presentation of data accumulated in Phase I are shown in Table
"2. These data are significant in terms of showing no technical problems
°r roadblocks to membrane process concentration of very dilute bleach
Affluents. The problems to be faced in processing such dilute waters
economic in nature, and are associated with the expense of removing
quantities of water to achieve significant levels of concentration
°f the dissolved materials. It would be better to use the bulk of
'these dilute effluents as early stage wash waters in countercurrent
or diffusion washing systems prior to membrane concentration.
Description of Phase II Operation
studies were continued with the same equipment at higher concen-
trations in Phase II, as shown in Table 63. However, in Phase II the
concentrate was recycled back to the feed tank in order to reach and
179
-------
TABLE 62
PHASE I - PERFORMANCE DATA
STRAIGHT THROUGH HO PROCESSING 07 KSE
Solids, mg/i
Feed
Concentrate
Permeate
BeJ., percent
BOD, mg/1
Feed
Concentrate
Permeate
ReJ . , percent
COD, mg/1
Feed
Concentrate
Permeate
Re j . , percent
Sodium, mg/1
Feed
Concentrate
Permeate
ReJ . , percent
Color
Feed
Concentrate
Permeate
He j . , percent
OD § 281 nra
Feed
Concentrate
Permeate
BeJ . , percent
ll
Feed
Concentrate
Permeate
Besistanee ohm-em 8
Feed
Concentrate
Permeate
Temp. , C
9/22/70
Bank A Bank B
2044
2892 3648
134 227
93.4 88.9
170
174 212
40 40
76.5 76.5
1208
1658 2082
60 64
94,0 94.7
472
630 785
50 86
89.4 81.8
6000
8750 10,000
7 7
99.9 99.9
13.92
20.2 26.2
.110 .126
99.2 99.1
6.59
6.78 6.88
5.81 5.90
25°C
481
343 282
4028 2685
36 36
'Pressure, psig
(In/out) 500/465 500/465
Permeate, ml/mln
Velocity, ft/sec
KfdA85 PsiK « 36 C
fffd/600 psi* 8 35°C
385 369
2.6 1.6
8.65 8.15
10.7 10.3
9/24/70
Bank A Bank B
1984
3004 4628
155 258
92,2 87.0
140
208 278
44 46
68.6 67.1
1308
1912 2942
78 84
94 93.6
460
635 915
55 102
88.0 77.8
5000
7500 13,750
20 10
99.6 99,8
14.56
21.4 35.4
.130 ,129
99.1 99.1
6.65
6.82 6.92
5.80 5.80
457
329 223
3433 2042
33 33
515/480 490/460
393 425
2.6 1.6
9.0 11.4
11,1 12.8
9/29/70
Bank A Bank B
2884
3460
635
78.0
194
267
59
69.6
1596
1996
i 95
"* 94.0
•a
&
a 632
•* 790
S 224
s 64.6
"8
SB
a 7000
3 7500
3 30
* 99.6
«
«
"a, 20.64
§ 26.95
o '30
f 98.5
7.03
7.05
7.10
344
295
864
33.5
510/440
165
1.6
3.88
4.9
10/8/70
Bank A Bank B
3120
3512 3840
637 383
79.6 87.7
181
182
51 46
71.8 7>».6
1504
1714 1888
86 79
94.1 94.7
720
835 890
242 134
66.4 81.4
7000
7500 8750
25 15
99.6 99.9
18.8
21.5 24.0
.225 .140
98.8 99.2
7.00
7.16 7.23
6.70 6,49
291
196
813 1420
31 31
555/395 520/465
284 347
4.2 ' 3.2
7.7 8.5
8.9 10.3
180
-------
TABLE 63
PERFORMANCE DATA SUMMARX - PHASES I AND II
Phase I Phase II
Solids, mg/1
Feed
Concentrate
Permeate
Bej . , percent
BOD, «*5/l
Feed
Concentrate
Permeate
Re j . , percent
COD, ng/1
Feed
Concentrate
Permeate
HeJ . , percent
Sodium, mg/1
Feed
Concentrate
Permeate
ReJ, , percent
Color
Feed
Concentrate
Permeate
Hej , » percent
00 § 281 nm
Feed
Concentrate
Permeate
Re j . , percent
Hi
Feed
Concentrate
Permeate
Resistance ohm-cm 8
Feed
Concentrate
Permeate
Teas, , °C
Pressure, psig
(in/out)
Permeate, ml/min
Velocity, ft/see^
*fd/W5 isiK It 3'6°C
gfd/600 psiff 6 35"C
10/13/70
Bank A Bank B
2870
3146 3276
611 486
87.9 83.1
125
117 118
22 24
82,4 80.8
1332
1450 1450
116 144
91.3 89.2
696
795 , 820
207 163
70.2 76,6
4000
6250 6250
300 400
92,5 90,0
16.2
18.0 18.2
0,87 1.31
94.6 91.9
6.93
7.18 7,09
6.72 6,70
25°C
313
291 276
918 1260
30 30
535/380 520/420
256 232
4.2 3.2
7.0 6.2
8.5 7.5
10/15/70
Bank A Bank B
12,630 16,070
17,170 17,450
1533 1252
87.9 92.2
553 852
1255 1255
166 181
70,0 78.8
7530 9430
10,590 10,050
193 142
97.4 98.5
2840 3560
3900 4000
534 461
81.2 87,0
32,000 40.000
50,000 45,000
100 500
99.7 99.9
93.6 115.6
135.0 125.5
1.10 0.585
98.8 99.5
7.14 7.29
7.28 7.39
6.83 7.19
105 84
85 78
396 454
31 32
520/445 465/370
224 202
3.2 4.2
5.7 5.75
7.0 7.4
10/20/70
Bank A Bank B
35,230 41,310
31,130 48,000
1494 1051
95.8 97.4
1530 1605
1230 1855
74 52
95.2 96.8
27,300 27,260
24,580 31,020
221 126
99.2 99.5
5920 7760
5500 8800
486 364
91.8 95.3
150,000 130,000
137,500 150,000
200 600
99.9 99.9
364 360
322.5 400
1.595 0,631
99.6 99.8
7.3 7.39
7.32 7,40
7.49 7,14
61 45
66 42
245 553
31 33
525/440 520/420
178 173
3.2 4.2
4.5 4.4
5.7 5.7
181
-------
TABLE 6.3 (Continued)
Phase II
Solids, mg/1
Raw feed
Module feed
Permeate
ReJ , , percent
BOD, mg/1
Saw feed
Module feed
Permeate
Re j . , percent
COD, mg/1
Raw feed
Module feed
Permeate
Rej . , percent
Sodium, mg /I
Raw feed
Module feed
Permeate
lej , , percent
golor
Raw feed
Module feed
Permeate
ReJ . , percent
OD i 28l nm
Raw feed
Module feed
Permeate
He j , , percent
SS
Baw feed
Module feed
Permeate
Resistance ohm-cm 8
Raw feed
Module feed
Permeate
Temp. , C
Pressure., pa ig
(In/out)
Permeate, ml/min
Velocity, ft/sac
gfd/600 psig 8 35°c
10/22/70
Bank A
3 3"
Mod. Mod,
39,940
2398 822
94.0 97.9
1610
116 52
92.8 96,8
28,880
278 109
99,0 99.6
7040
852 292
87.9 95.8
160,000
225 35
99.9 99.9
380
1.90 0,34
99.5 99,9
7,12
7.15 5,60
25°C
49
252 678
33.5
525/445
137 223
3.2
5.2
Bank B
3 7
Mod, Mod-
34,700
560 1412
98.4 95,9
1385
46 66
96.7 95.2
24,480
89 147
99.6 99.4
6160
202 515
96.7 91.6
130,000
25 60
99.9 99.9
330
0.246 0.72
99.9 99.8
6.80
6.40 6.80
55
961 397
33
545/440
243 161
4.2
6.1
10/29/70
Bank B Bank A
5 Modules
2366
4638 20,448
332 1276
92.8 93.8
865
940 1568
48 80
94.9 94.9
1410
2810 14,660
60 125
97.9 99.1
542
1026 3580
121 500
88.2 86.0
6000
12,000 64,000
10 45
99.9 99.9
17.06
33.8 66.0
0.11 0.44
99.7 99.3
8,32
7.28 7.40
6,70 7.00
411
229 79
1583 423
34,5 34
455/400 473/375
322 285
4.2 3.2
10.7 10,0
11/5/70
Bank B Bank A
5 Modules
2610
5906 81,460
588 1845
90.0 97.7
990
1290 5290
58 116
95,5 97.8
1470
2740 71,680
69 181
97,6 99.7
594
1324 10,980
195 672
85.3 93.9
7000
16,000 350,000
10 100
99.9 99i9
19.78
44.2 902
0,14 0.94
99.7 99.9
9.22
7,20 7.38
6.70 7.00
357
178 36.7
975 309
31.5 33
590/405 465/385
220 130
4.2 ,4.2
6.3 7,4
-102
-------
the higher concentrations. This could only "be realized by-
re circulation in this size installation. The two systems operated
in parallel for the first week at inlet velocities of It. 2 feet per
second in Bank A and at a lower velocity of 3.0 feet per second in
Bank B. The feed concentration was maintained at ItO grams per liter
in "both banks.
After the first week, the modules in Bank B were operated on a feed
liquor ranging from 5-10 grams per liter of solids, while Bank A was
operated with feed concentrations in the range of 20-80 grams solids
Per liter (see Table 63). Product water flow and specific gravity
°f the feed were measured at least twice daily during this run. Solids
concentrations and eventually also the osmotic pressure were determined
experimentally and used to correct the flux rate data to a standardized
level of 600 psig at 35°C. These corrections in flux rate data to
a standard level were important in achieving close comparison with
data for the following Phase III operations. Additionally, these cor-
rections helped develop an understanding of the scatter in the data
observed for Phase II and III where a wide range of osmotic pressure
was occurring as the concentration increased from 1-10 percent solids.
'There were also problems apparent when attempting to interpret data
in such a small system where recycling to achieve high levels of concen-
tration resulted in low levels of draw-off of the concentrates in compe-
tition with some leakage of the pump seals.
It should be noted that the progression in concentration from low levels
°f concentration to high levels of concentration produced a character-
istic increase in rejections at the higher levels of concentration
Curing Phase I and II and progressing into Phase III. This observation
has been made on numerous occasions over the past several years in
these studies when concentrating these dilute pulp mill effluents.
It can be concluded that low molecular weight materials, and especially
Volatile low molecular weight materials, are not well rejected in the
£irst stage of processing at the lower levels of concentration. These
Materials are lost in the permeate from the first stage, and after
*hey have passed from the system the levels of rejection are then ob-
served to be markedly increased in later stages of concentration.
Phase III — Operations, Data and Results
six new Havens modules of the latest design became available in October,
in time for use and for comparison in the Phase III study. Five
these new modules were set up in series as a new Bank A to become
second stage in a 2-stage concentration system. The first stage
s made up of ten of the older modules, all ten of which were hooked
^ in series in a single Bank B. (It should be noted that the designa-
tion of Banks A and B, as reported for Phase I, were reversed in the
c°ncentration runs for Phases II and III. Bank B then became the first
stage and Bank A the second stage for the concentration runs.)
183
-------
Detailed data for the Phase III operations are presented in Table 6k.
Concentrate from the first stage was.about 0.7 percent to about 2 percent
solids, and the second stage (A) achieved concentrations ranging to
8-l6 percent solids. Rejections for all components were high.,' and
for mast determinations ranged upwards of 95 percent during these concen-
tration studies in Phase III.
A rather uniform drop in flux rate was observed as the concentration
increased from 1 percent solids to 11 percent solids. Data are tabu-
lated to form a straight-line function as shown in Fig. 32. These
flux rate data, when corrected to 35°C at 600. psig, were quite consis-
tent over a 6-week period. The data summarized in Table 65 further
develop the picture for the range of rejections which occur at various
levels of solids concentration when processing caustic extraction bleach
effluents from this kraft mill. These data result from careful contin-
uing studies conducted with assay of a fairly large number of single
and composite samples during this 3-month demonstration. The data
are important in demonstrating that sustained high levels of performance
can be expected in large installations. This conclusion has been dif-
ficult to prove out with most small laboratory operations on a single
module or a few modules which are subject to interruptions and incon-
sistencies.
Figure 33 summarizes flux rate of the first stage with and without
pulsing, and shows that rates on the order of 8.5 gfd can be maintained
on a:sustained basis for three weeks or more where the velocity approx-
imates 2.8 feet per second. However, during the last three weeks we
tried operating at lower velocity of 1.7 feet per second and found
that the flux rates dropped to about 6-7 gfd. Pulsing this system
periodically by reducing the pressure to near zero for '30 seconds or
longer at periodic intervals of several,hours or even at once per day
levels restored the flux rate. The new modules in the second section
(B) delivered a uniform flux rate with no need for pulsing during the
six weeks they were in operation.
The problems of achieving high levels of recovery of the rejected solids»
BODg, COD, sodium, and color in terms of optical density are' summarized
in Table 65« High levels of recycle are necessary in order to achieve
high levels of concentration in small systems with only "a few modules,
such as was the case in these runs. The recoveries may appear lower
than would be the case where a straight-through operation achieves
much of the water removal at the lower level of concentration, and
only relatively smaller amounts of water are removed as the concentration
rises above 8 percent solids. The' lower quality of the final effluents
in a recycle operation reduces the numerical average for quality param-
eters to a greater'extent than would be the case for analyses of samples
from straight-through operation. Also, in the case of large-scale
operations, the relatively small volumes of permeate in the final stages
of concentration above 8 percent solids might be recycled to the feed
if necessary to improve the degree of recovery of individual components
in the concentrate and of turning out a clean, total permeate.
iSk
-------
TABLE 6k
III • BAIA SUMHARt
Solids, ag/1
Bav feed
Module feed
Permeate
Rej., percent
SOD, rag/1
Saw feed
Module feed
Permeate
BeJ,, percent
MA ag/1
flaw feed
Module feed
Permeate
SeJ . , percent
Sodium, nig/1
«av feed
Module feed
Permeate
R*J., percent
Colp£
Rav feea
Module feed
Permeate
ReJ., percent
°L*t 281 nrn
Siv f ee3
Module feed
Permeate
^eJ«» percent
IS
Raw feed
Module feed
Permeate
SSSistanee oton-cm
tew feed
Module 'feed
Permeate
TteniD °f»
tCr*"ii__£
Assure, psig
Un/out)
*62$n»a f- « M! /*vt4 y\
-,*weat;ef ml/iain
llillLdfe
ssZlPpSigirs
ll/U/70
Bank B Bunk A
2726
7118 133,695
255 627
91.2 99.8
175
310 5200
. 44 163
85.8 96.9
1600
4360 117,000
95 375
97.8 99.7
604
1632 16,280
211 838
87.1 95.4
7000
20,000 650,000
10 200
99.9 99.9
19,4
48.8 1448
0.17 2.16
99.6 99,8
10.4
7.32 7.30
6.51 7.08
at 25°C
354
163 28
892 252
30 34.5
500/375 770/565
IS3 229
3.2 4.2
fr 6.4 4.0
11/16/70
Bank B
15,628
782
95,0
795
37
95,3
10,260
97
99.0
2728
495
81.8
40,000
20
99.9
138.4
0.232
99.8
7,22
6.61
96
771
30.5
520/390
206
39
,t
.1
Bank A
102,045
2008
98.0
5300
127
97,6
80,750
285
99.6
15,680
654
95.8
400,000
150
99.9 f
1100
1.62
99.8
7.39
6.92
29
307
35
730/535
233 ;
It. 2
•* * £
?4.7
11/19/70
Bank B
16,684
818
95.1
2120
71
96.6
12,500
107
99.1
2760
271
90.0
50,000
20
99.9
152
0.245
99.8
6.83
6.55
103
721
31
495/390
228
3.2
7.8
Bank A
100,220
2090
97.9
5480
141
97.4
78,265
255
99.7
15,040
710
95.3
400,000
125
99.9
1036
1.316
99.9
7,20
6.82
29
317
34.5
740/540
208
4.2
4,0
11/21/70
Bank_B
20,016
806
96.0
1620
78
95.2
16,100
111
99.3
3100
280
91.0
90,000
25
99.9
206.4
0.288
99.9
7.45
6.71
94.6
686
31
495/395
156
3.2
5.4
Bank A
81,365
2764
96.6
4740
180
96,2
67,000
295
99.6
11,080
740
93.3
350,000
125
99.9
896
1,45
99.8
7.41
7.04
35.2
274
34.5
710/530
ft
4.2
1.3
185
-------
TABLE 6U (Continued)
PHASE III - DATA SUMMARY
Solids, mg/1
Raw feed
Module feed
Permeate
ReJ . , percent
BOD, mg/1
Raw feed
Module feed
Permeate
ReJ.. , percent
COD, ng/1
Raw feed
Module feed
Permeate
Re j . , percent
Sodium mg/1
Raw feed
Module feed
Permeate
ReJ . , percent
Color
Baw feed
Module feed
Permeate
ReJ i , percent
OD 6 28l n»
Raw feed
Module feed
Pemeate
ReJ , , percent
IE
Baw -feed
Module feed
Permeate
Resi stance ohm-cm
Raw feed
Module feed
Permeate
Tenjb , °C
Pressure, ssig
(in/out)
Permeate | ml/min
Velocity, ft/sec
pfd/600 psig at ~35
11/23/70
Bank B Bank A
2646
22,018 118,480
685 2038
96.9 98.3
215
1272 8350
62 144
95.1 98,3
1584
17,380 100,000
104 335
99.4 99.7
608
3530 16,000
250 710
92,9 95.6
7500
100,000 500,000
25 150
99.9 99.9
20,4
232 1332
0.28 1.71
99.9 99.9
10,04
7.40 7.28
6.70 6.98
6 25°C
358
86.9 28.2
790 298
29 32.5
525/480 780/575
248 224
o_ 2.6 4.2
77 41
12/1/70
Bank B Bank A
2454
19,930 100,560
690 1789
96.5 98,2
220
925 4600
40 138
95.7 97.0
1266
15,920 77,500
120 315
99.2 99.6
604
3224 13,440
246 594
92.4 95.6
4000
72,000 320,000
20 125
99.9 99.9
15.56
196 1100,8
0.284 1.49
99.8 99.9
10.62
6.90 7.05
6.20 6.30
358
94 31
810 350
30 35
490/405 685/455
224 168
3.2 4.2
8.5 V4.6
12/2/70
Bank B Bank A
2636
9288 132,630
693 2833
92.5 97.9
185
512 6250
48 235
90.6 96,2
1228
5620 104,500
95 446
98.3 99,6
690
1948 19,040
248 974
87,3 94,9
5000
24,000 400,000
15 175
99.9 99.9
14.40
63.6 1395
0.201 2.31
99.7 99,8
10.02
6.75 6.58
6.20 6.42
278 .
130 24
805 258
31 35
475/395 810/580
274 205 •
3.2 4.2
9.5 '3.2
12/7/70
Bank B Bank A
3242
7384 159,432
647 3026
91.2 98.1
253
447 8150
61 210
86.4 97.4
...
5560 137,500
113 449
98.0 99.7
662
1380 20,120
234 1002
83.0 95.0
80QO
22,000 640,000
20 275
99.9 99.9
26.44
59.6 1872
0,21 2.80
99,6 99.8
9.52
7.30 7.18
6,70 7.10
359
166 25.7
816 219
29 32
480/450 795/650
277 170
2.2 3.2
6.1 *2.3
186
-------
TABLE 64 (Continued)
PHASE III - DATA SUMMARY
Solids tng /I
Raw feed
Module feed
Permeate
Re j . , percent
BOD, m$/l
Raw feed
Module feed
Permeate
Rej., percent
ggp. ing/1
Raw feed
Module feed
Permeate
Re j . , percent
Sodium mg/1
Raw feed
Module feed
Permeate
Re j . , percent
Color
Raw feed
Module feed
Permeate
Rfij., percent
SLat 281 ran
Kaw feed
Module feed
Permeate
**j.» percent
El
*av feed
Module feed
Permeate
SSSistarice ota-cm at 25"C
«w feed
Module feed
*ernieate
HE*sswre, pslp
\Wout)
OWieate. ml/min
iSKcity, ft'/sec
SJoTepb pslg at 358C
12/9/70
Side B Side A
2648
8688 116,602
619 2014
92.9 98.3
183
542 5825
58 132
89,3 97.7
1432
5635 91,600
91 259
98,4 97.7
640
1708 13,160
222 670
87.0 94.9
6000
26,000 440,000
10 100
99.9 99.9
17.96
68,4 1208
0.199 1.30
99.7 99,9
10.51
7.35 7.42
6.81 7.10
337
149 32.6
852 310
28 30
510/490 705/555
273 204
2,2 3.2
8.5 4.5
12/14/70
Side B Side A
93,156
3084
96.7
5350
145
97,3
41,600
266
99.4
11,840
830
93
240,000
70
99.9
1120
1.61
99.9
7.21
7.02
37.2
267
33
740/620
312
3.2
5.9
12/15/70
Side B Side A
62,548
1975
96.8
3600
111
96.9
42,400
208
99.5
8920
556
93.8
160, 000
53
99,9
656.8
1.00
99.8
7.26
6.95
44.6
439
33
755/630
312
3.2
5.75
12/15/70
Side B Side A
5010 30,436
442 7328
91.2 75.9
312 1612
48 82
84.6 94.9
2045 22,200
77 145
96.4 99.4
1155 5530
148 458
87.2 91.4
6665 144,000
25 65
99.6 99.9
33,95 298,4
0.195 0,560
99.6 99.9
7.10 7.30
6.36 7.10
205 69.5
1218 427
31 33
515/490 710/590
198 340
2.2 3.2
5.8 7,0
107
-------
Phase III Data from 2-Stage Concentration
Stage 1
Stage 2
CD
10-
9-
8--
"0"
6
5-
4"
3
2-
1
4-
Flux Rate at 600 psig — 35°C
©Data taken 11/5 to 11/23
-I-Data taken 12/6 to 12/15
10
20 30 40 50 60 70
Solids Concentration of Feed? g/1
80
90
Figure 32. "Decrease in Flux Rate with Increase in Solids Concentration
100
110
-------
TABLE 65
RANGE OF RECOVERY OP REJECTED COMPONENTS AT
VARIOUS LEVELS OP CONCENTRATION .
(Percent)
Solids, g/l 3.0-9-3 15.6-22.0 81.U-U8.5 132. 6-159. U
Number of
determinations ^57 3
Solids 91.2-92,5 95-96.9 96.6-98.3 97.9-99-8
B°D5 85.8-90.6 95.1-95-7 96.2-97.7 96. 2-97. li
°OD 97.8-98.1* 99.0-99^ 99.6-99.7 99.6-99,7
Sodium 83.0-87-3 81.8-92.9 93.3-95-8 911.9-95.11
Color 99.9 99.9 99.9 99.9
Optical density
at 280 nm 99.6-99.7 99.8-99-9 99.8-99.9 99.8-99.9
"Che overall recoveries of rejected material experienced in this 3-
JBonth field trial, and as reported in Table 65, are especially interest-
ing in -view of counter observations of relatively poor rejections,
which may often "be observed in early stages of concentrating wastes
containing low molecular weight or small molecular size material which
can pass the membrane. Similar loss in rejections may also appear
^t high levels of recycle in the upper limits of concentrating some
Wastes (often observed when the concentration reaches 15 percent solids
°r equivalent levels, with resultant high osmotic pressure). Low molec-
ular weight material, and particularly volatile low molecular weight
teterial, may pass through the membrane rather freely in early stages
°f concentration and may substantially reduce rejection of the system,
as is the case of the BOD 5 rejections dropping below 70 percent if
acetic acid, methanol, H^S, and similar components are present. However,
for most wastes studied this is a transitory loss in rejections which
lnay have small effect on a total concentration system. Elsewhere it
was noted that reverse osmosis of evaporator condensates, with high
levels of volatiles, is a notable exception, and in that case it is
Necessary to neutralize acetic acid to achieve good rejections of the
&cetate salts and. good BOD reduction.
in rejections when the osmotic pressure reaches high levels is
apparently a limiting factor on membrane performance and needs careful
s*udy and definition for each type of waste. In this study on the
caustic extraction bleach effluent, the limiting level may be in the
neighborhood of 15 percent solids, but the volume of permeate containing
189
-------
20.
O
o
w
CJ
o
1
to
•rl
M
P.
O
O
TS
to
a
os
ia -
9- -
Stage 1 flux rate adjusted to 600 psi and 35°C
-t-unpulsed readings
Q pulsed readings
Idle
01
j^tdle-
Inlet velocity = 3.2 ft/sec
- Outlet velocity = 2.5 ft/sec
Inlet velocity = 2.2 ft/se<.
Wutlet velocity - 1.2 ft/swr
I I lo ll 12 ^3 ta 1*5 ^6 I1? 18 19 ^0 A 2^2 2*3 2*4 25 ' 2^ 30 \ 2 3 I \> I, 7 8 I 10 II 12 h 1
November
14 15
December
DATE
Figure 33. Comparison of Flux Rates vitla and \iitiiout I1ul3i:.;';
-------
poorly rejected solids may be sufficiently low even at that level of
concentration so as to permit recycle of the permeate to earlier stages
°f feeding the system where rejections are high. These studies have
°ot yet been carried far enough to thoroughly test this concept for
Achieving the higher levels of concentration above about 12 percent
solids.
One of the criteria in achieving efficient operation of the equipment
and for interpreting the results of analytical studies has been based
in these studies on development of an osmotic pressure to solids rela-
tionship upon which observed flux rates could be corrected to a standard
level for comparative evaluation. Such a relationship is shown in
W-g« 3^» in which the applied correction used in these studies is indi-
cated as one of the four curves on this graph,
Development of operating parameters for reverse osmosis concentration
for this caustic extraction bleach effluent from kraft pulping required
& quick method of developing the solids concentration as based on deter-
Bination of specific gravity at 20°C. Availability of this curve (Fig.
35) helped substantially in operating the equipment efficiently and
consistently during the final stages of the 3-month .run at Cloquet.
Consistent and reliable performance of the equipment was achieved in
this run as the operating know-how developed. Hie demonstration has
shown that alkaline extraction KBE can be concentrated from 30 to 50
tiroes or more by reverse osmosis on the basis of continuous operation
Sttd without the drastic variations and declines in flux rate experienced
°h smaller-scale studies conducted in the early phases of this overall
Project in the laboratory and in shorter term pilot studies. The overall
operations on site at Cloquet and the various laboratory studies con-
ducted concurrently in Appleton clearly show that this bleach plant
effluent can be successfully concentrated at high levels of treatment
e£ficieney and with sustained operating efficiency'in a continuous
°peration of membrane equipment.
SUPPLEMENTARY STUDIES
than five years of preliminary laboratory and small.pilot-scale
studies of membrane processing of various pulp and paper mill effluents
Preceded this demonstration run and provided the base upon which a
successful demonstration could be planned and executed. It is not
purpose of this report to present a detailed review, but certain
of those studies provide back-up experience in the field of
Processing kraft pulping and bleaching effluents."
Processing Mixed First Stage Chlorine and Second-Stage
Alkaline Extraction Effluents
Reverse osmosis concentration processing of mixed first and second-
8tage bleach effluents presents obvious advantage in terms of reduced
P^etreatment costs for pH adjustment and perhaps also temperature
191
-------
/
EJ
500-
400-
•H
m
Pt
N
£ 300.
0)
u
o
i
o
200 ..
100 -
Applied Correction
4- Concentrated by recycle —
vapor pressure osmometer
Q Concentrated by recycle —
flux rate method
0 One shot concentration —
vapor pressure osmometer
120
0
Figure 3k,
160
Solids, g/1
Osmotic Pressure of Caustic Extraction Effluent
192
-------
U)
1.09 -
1.08 -
1.07 -
€ .
01 1.06 -|
*»
SI-05 -|
c
u t.04 -|
*H
4-t
•8
&1.03H
ta
1.02 -
1.01
1.00
10
30
"Too IW 120 130 140 So U
80 90
Solids, g/i
Figure 35- Total Solids as Function"of Specific Gravity of Alkaline Extraction KBE
160
-------
balancing. The" cellulose acetate membranes available for reverse osmosis
at the time of conducting these studies required adjustment of pH within
the range of 3.0 to 7.0,, and a temperature range of 35-^0°C has normally
been recommended for attaining maximum flux rates and to avoid heat-
accelerated membrane degradation "by hydrolysis and related reactions.
Unfortunately, the first-stage chlorine KBE available at the Cloquet
mill was highly dilute as shown by the following typical analysis:
Solids 1.303 g/1
BODS 102. mg/1
Chlorine 1*72 mg/1
Color (Co or Pt) 100
OD at 280 nm 2.1
pH 2.15-2.U
Temperature 1T-19°C
There seemed to be no practical way in which to secure needed quantities
of 500 to 1500 gallons daily of a more concentrated first-stage chlorine
KBE without substantial process modification and serious disruption
of mill operations.
The dilute chlorine stage liquor at one-half the solids concentration
of the second-stage caustic extraction effluent could not be used as
feed in our demonstration without serious cut-backs in the production
of higher concentrations of final product. Attempts to set up the
main demonstration on combined effluents had to be set aside in favor
of work only on the caustic extraction liquor.
However, a small-scale short-term test run was made using a single
2-tube module especially fabricated for tests on small volumes of a
few gallons, Chlorine stage KBE and also caustic extraction stage
KBE, which had been obtained by manually pressing high density pulp
from the towers, were mixed and processed in this special reverse osmo-
sis concentration run. The caustic extraction feed sample had a solids
concentration of U-57 grams per liter, and the chlorine bleach effluent
was still very dilute at 1.2U grains per liter. These were mixed in
the proportion of 3 parts of the first-stage chlorine effluent and
one part second-stage caustic extraction product to approximate the
volumes being sewered at this mill. The mixture had a pi of 2.0 and
a solids concentration of 2.1 grams per liter. Table 66 summarizes
the data obtained on this small-scale test run which was carried to
the point of concentrating more than 10 times to achieve about 2,5
percent solids (2k.6k grams per liter). At that level the osmotic
pressure, due to content of Nad, was found to be sufficiently high
to have reduced the effective driving force and to have caused a sub-
stantial reduction in flux rate of more than 50 percent from Ik. 8 gfd
to 6.k gfd at 600 psig. Other experience would indicate the drop
in flux rate might be taken care of with higher operating pressures
and higher velocities as the concentration increases, but this pre-
liminary run could not be extended to evaluate that point. The
-------
TABLE 66
REVERSE OSMOSIS PBQCESSIBG OF COMBIHED FIRST
MD SECOHD-SfAGI BtEACfl PLAHT EFFLUENTS
Chlorination
effluent
Caustic extract
effluent
Combined
vo effluents
\_n
Final
concentrate
Dissolved
I/I Be,).,
1.21*
!».5T
2.117
5.393
9.96
17.81*
2k. 6k
Solids BOD,.
pereent mg/1 BeJ.t percent
190
279.
85.6 219 21
91.5
98.5
89.7
88.8 922 78. k
Optical
Density
§ 280 am. ReJ . . percent
1».28
32.S
10.69
2k.k
50.0
99.5
Itt
96.1*
97.9
98.8
99.1*
99.5
Sodium
mg/1 Rej.^ percent pH
77.5 1.75
1188 10.98
339 71.7 2.03
2.20
2.65
3.22
15^0 31.6 3.22
flux Rate
6 600 psig
and l*Qoc
1U.8
11.5
9.5
7.5
6A
-------
advantages apparent in processing combined effluents remain as important
objectives for further study.
STUDIES ON KRAFT PULP WASH WATERS
Extensive laboratory and pilot and also field test runs were made on
concentration of the wash water from the countercurrent washer and
from the rewasher in the kraft pulping operations ahead of the bleach
plant at this mill. At the time these tests were initiated, this was
considered to be one of the most serious pollution problems prevailing
at the Cloquet mill, but later on in the course of these studies the
mill management decided to make process changes which would eventually
recycle and eliminate this type of pulping spent liquor. The preliminary
RO studies were terminated. A brief accounting of the rewash water
concentration trials is reviewed in the following paragraphs, since
the experience was in many ways valuable. The rewash water was an
especially difficult substrate for processing by reverse osmosis in
terms of apparent fouling and a resultant decrease in flux rates through-
the membrane. These results were, however, not consistent in all runs.
Some samples sent to Appleton were relatively easy to process, and
then other shipments would give serious levels of fouling and flux
reduction. The critical importance of maintaining optimum velocities
was not apparent at that time, and overall average velocities may have
been critically low with so many modules operating in series.
Preliminary studies using two different types of cellulose acetate
membranes showed excellent levels of rejection for all components
studied, as shown in Table 6j.
TABLE 67
REJECTION DATA FOR REVERSE OSMOSIS
OP KRAFT PULP WASH WATER
[Moderately Tight (3) and Tight (5) Membranes]
Type 3, Type 5,
percent percent
Solids 95 99
Sodium 96 99
OD at 280 nm 97 98
BOD5 90 95
COD 95 98
The moderately tight, cellulose'acetate membrane (Type 3)» normally
used in most of these studies, was compared with the tighter type No.
5 membrane from this manufacturer. The average BOD removal was above
196
-------
50 percent for both membranes. Tables 68 and 69 provide more detailed
operating data for a preliminary field run conducted at Cloquet during
the summer months of 1°63 for Drocessinp: these wat?h waters.
TABLE 68
COMPARATIVE FLUX RATES VS. SOLIDS CONCENTRATIONS
OF KRAFT WASH WATER
Operation of Preliminary RO Unit at
Northwest Paper Company, Cloquet, Minnesota
Period
No.
Operating
Time
Totalized,
hours
Flux Rate,
gfd adjusted to 35°
Concentration
Feed,
solids g/1
Type 3
Used
Modules
Type 3
New
Modules
C
Type 5
1
2
3
5
6
7
8
9
10
Feed from Countercurrent Washer
101 9.0 7.7
l6l 14.8 4.8
162 19.8 4.9
219 13.6 5-0
Feed from Revasher
61 1.05
232 0.95
325 0.9
'Recycle Operation
48»( 1.67 — .
576 2.6l
596 6.39
4.7
3.2
3.2
2.4
12.6
12.2
20.1-8.0
9.8-3.5
12.4-3.3
4.3-2.4
5.2-3.1
4.5-2.6 untreated module
equipment and the size of the operations were similar to the later
conducted on the alkaline extraction bleach effluent. These runs
totaled 815 hours, vith a number of interruptions when it was decided
^° return to the laboratory to study individual problems between runs
at Cloquet. Table 68 compares the flux rates observed at various levels
solids concentration and with Types 3 and 5 membranes in previously
and in new modules. The first section provides data for the liquor
from the countercurrent washer, which was diluted with river water
approximately 1 percent solids, while the second section provides
from processing rewash water, both straight-through with the con-,
Centrate recycled. Although wide variations in flux rate were evidenced.
197
-------
TABLE £9
JWAMTICM. MIA PILOT UBIT EUR OB
XJUtFf HOP WASH
Operating
Time, Solida ,
hours g/i
1 , Feed 101
Permeate 3
Permeate 5
Concentrate
R«J, Type 3, percent
R*J. Type 5, percent
2, Peed 161
Perneate 3
Permeate 5
Concentrate
Rej, fyp« 3, percent
ReJ. tyj>« 5, percent
3. Feed 1^
Permeate 3
Concentrate
Rej. Type 3, percent
Rej. Type 5, percent
It, reed 219
Permeate 3
Permeate 5
Concentrate
Rej, Type 3, percent
BeJ. Type 5, percent
5. Feed 26 J
Permeate 3
Concentrate
Rej, Type 3, percent
6. Feed 232
Concentrate
BeJ. Type 3, percent
7 T «<
*• *eea ««*
Jeroerte 3
Concentrate
Rej, Type 3, percent
fl. Teed . «8*
Ferae ate 3
Concentrate
Rej, Type 3, percent
9 V e*
Permeate 3, 5Tfi
Concentrate
ReJ- Type 3, percent
10. Feed
Permeate 3
Concentrate
Rej. Tyy* 3, percent
9.03
0.71
0.1(3
1*.*T
92.1
95.2
iV.Bl
0.32
0.16
19.61i
97.8
98.9
19.6k
1.73
0.7U
25.80
91.3
96.3
13.58
Q.J2
0.15
17.62
93.2
98.9
1.05
0,05
1.73
95.3
0.957
0.01(5
1.039
95.3
0,90
0.038
1.00
95.8
1,67
Q.05S
1.88
96.5
a, 61
0.067
2.78
97. k
6.»
0,152
6.9k
97,6
Sodium ,
1720
212
9
2500
87,7
99-2
291(0
79
29
3630
97.3
99,0
3380
1(30
201
1)1(80
87.3
2190
3te
20
3250
99.1
202
11
3*8
9».6
511*
lit
5lW>
97.3
197
9.8
217
95.0
1*00
16
l»70
96.0
kgo
15
1*90
96.9
1660
' 111
1990
97-5
BOB,,
m/i
2585
365
165
3910
85.9
93.6
3305
22k
17k
1(580
93.2
9k. 7
531*0
728
6*90
86.1*
91.1
23fcO
46k
169
3515.
80.2
92<8
160
25
81t.li
129
27
126
79.1
192
26
218
86.5
306
32
3*8 ,
89.6
212
15
216
92.9
6k$
35
6I*J
oi»,e
Optical
COB, Density
ag/1 at 281 nra
8.310
1*93
Ilt6
12,800
98.2
Ik ,290
391
286
18,600
97.3
$8.0.
18,631
1,199
689
23,850
93.6
96.3
10,620
815
231)
16,1(00
92.3
97-8
836
53
93-7
796
U6
9tO
9^.2
760
886
9«.i
1,1(90
28
1,730
• 98.1
396
39
1,278
90.2
2,1(93
86
2,665
§6.6
70
£.1(0
0.93
10k
96.6
98.7
115
3.10
2.58
150
97.3
97.8
160
5.26
It. 19
208
96-7
97.1)
82.5
It.lb
2.00
125
95.0
97,6
6.28
0.27
10,60
95-7
6.50
o.afc
7.10
96.3
5.8
0,20
6.2
96.5
11.1
0.36
12.3
96.8
10.0
0.26
10.0
97. »
20.0
0.1(8
21.0
' 97.6
pH
7.67
7.73
6.62
8.58
—
7.23
6.38
6.1*0
7.35
--
7.33
7-05
6.58
7.38
—
7.80
7.35
7-90
7. 1(2
—
—
6.26
7,60
' —
7.35
7.20
7.55
—
6.1(9
6.10
6.99
_
7.05
6.50
•7.02
• _.
7.25
7.79
7.25
..
8,10
7.25
7.30
—
Specific
233
1,1*0
11 ,600
—
98.0
283
2,830
6,600
2 1<0
90
97.7
113
600
Il330
1,103
81.2
91.5
310
1,200
11,000
228
7k. 2
97.2
1,1)60
21,000
960
93-0
2,1(20
18,000
2,160
86.6
2,650
32,000
2,1(50
91-T
1,285
16, EDO
1,160
93.1
865
I*, 500
61(5
-------
Figure 36 shows the substantial reduction in flux rates experienced
with the kraft revash waters over the 580 hours that this dilute wash
water was processed. This loss in flux rate was not affected "by veloc-
ities up to 2.5 ft/sec or pressure pulsing for 1 minute every hour.
Higher velocities have since been found effective in reducing such
fouling but the work was terminated short of a full trial on rewash
Water.
Table 69 provides rejection data for these studies on the countercurrent
Wash water and on the rewash water. All rejections were excellent
"throughout these runs , although no attempt was made to carry the concen-
trations to high levels of concentration in these preliminary runs .
Part of the concentrate was recycled while processing rewash water
to increase concentration from 0.1 percent solids to a more representa-
tive 0.5 percent solids. In between the field runs at Cloquet during
these pulping rewash water studies, a number of laboratory trials were
^ade in an attempt to develop methods of pretreatment which would improve
"the flux rate performance. Various additives and pretreatment methods
Were tested; some were found to be moderately successful in reducing
the fouling but no means of maintaining steady state flux rate had
been found at the time it was decided that rewash waters would be elimi-
nated.
DISCUSSION
demonstration provided the first available operating data from
sustained runs on mejtfbrane processing of bleach plant effluents. The
ei&phasis and most Of the data desired from the studies on the second-
stage alkaline extraction KBE are usually considered to be of most
concern in maintaining effluent quality standards. However, the pro-
cessing of .combined effluettts can in all probability be developed as
* feasible concentration process with membrane systems if the starting
feed volume can be reduced to practical levels below 10,000. gal. /ton
Pulp.
High levels of salt (Had) concentration may appreciably raise the
osmotic pressure, and therefore need may arise for maintaining higher
°Perating pressures in the system to override that higher osmotic pres-
sure. But extensive experience is available from worldwide salt water
conversion studies to lend confidence to the conclusion that higher
of salt can be successfully handled.
field demonstration helped to prove out the methods developed in
three years of preliminary studies for maintaining .sustained and
Practical levels of flux through the membrane based on:
Maintaining velocities and turbulence at levels sufficiently
high to avoid concentration polarization and fouling of the
membrane .
199
-------
O
O
It
10
3 6
3 -
90
130
170
210
Z50 290
Elapsed Time, hours
J70
410
450
g
c 3
48?)
560 600
Time, hours
Figure 36. Flux Rate History While Processing Kraft Rewash Water -with the Pilot Unit
-------
Pulsing (periodic pressure reduction} to clear fouling and
membrane compaction.
No elaborate systems or procedures for pretreating the bleach effluent
Were required "beyond nominal adjustment of pH and temperature and sereen-
*nS (50 pesh) to reinove gross amounts of fiber and particulate matter.
introduction to this report cited the importance of trends toward
substantial reduction in the volumes of water used in the" total process
°f bleaching and washing of "bleached pulps . Present stage development
°f reverse osmosis equipment still involves relatively high capital
operating charges . Pulp mills and bleach plants normally expect
cost of fresh process water to be in the range of 3 to 10 cents
thousand gallons. Reverse osmosis has "been aimed' toward costs
°f less than 50 cents per thousand gallons of water removed, and some
B3anufaeturers look forward to costs as low as 25 cents per thousand
gallons eventually, but present-day evaluations to be discussed in
itors detail in Section X still show costs to "be more realistically
in the range of 50 cents to $1.50 per thousand gallons at this early
stage of commercial development. Water recovered by reverse- osmosis
therefore costs several tines as much as fresh incoming treated process
^ater. Obviously, the cost of treating outgoing waters for pollution
control must bear most of the eost for complete treatment of the bleach
Plant effluents . Most kraft bleaching operations have a BOD5 output
^0- the range of 33 to 50 pounds per ton of product and substantial
color and salinity problems , Secondary treatment designed to reduce
"^e BQD 5 by the biooxidation route would cost about k to 5$ per pound
£QDS or in excess of $1.00, per ton pulp (or in a 5000,- gallon per ton
?lw, ebout 20 cents per thousand gallons). Color removal has been
90 especially difficult problem, with charges' indicated to be on the
°*der of $2.90 per ton, or nearly 60 cents per thousand gallons at
flcws of 5000 gallons per ton22. Of more importance eventually may
te the recovery of 200 to 300 pounds of Had from each ton of chlorine
Bleached pulp. A substantial credit for reverse osmosis concentration
ing may be possible by recovery of bleaching chemicals in addi-
to pollution control and water revise credits .
economies of applying reverse osmosis to processing of pulp and
mill effluents are discussed more fully in Section X.
conclusions
Substantial field trials of reverse osmosis concentration processing
of second-stage caustic extraction bleach effluents have been success-
^ly conducted at flow rates on the order of 1500. gallons daily for
a 3-aiorith period, , •
^^ J^ates of 7 to 8 gf d can apparently be maintained Indefinitely .
JQ concentrating this bleach effluent in the range of 0,2 to 10 percent
°Uds concentration.
201,
-------
No elaborate or extensive pretreatment of the bleach effluent beyond
nominal amounts of temperature adjustment, pH control, and suspended
solids removal (50-mesh screen) seem to be required.
Treatment of combined bleach effluents seems also to be a feasible
operation on the basis of smaller-scale trials, although higher levels
of NaCl content may raise the osmotic pressure of the concentrates,
and hence require somewhat higher (but still feasible) operating pres-
sures to perhaps 800 psig.
The future economics of more complete bleach effluent processing may
be competitive and favorable in terms of:
Recovery of reusable water.
More complete pollution control treatment to remove color, foam,
inorganics, and resistant organics, as well as BODg.
Recovery of NaCl from bleach effluents.
Possibilities for ultimate regeneration of Na and Cl bleach
chemicals.
FIELD DEMONSTRATION NO. 5
CONCENTRATION OF CHEMIMECHANICAL PULP WASH
WATERS BY REVERSE OSMOSIS
This subsection describes laboratory and field studies conducted as
Demonstration No. 5 under Federal Research and Demonstration Grant
120^*0 EEL. The demonstration field trial was conducted at the Locks
Mill of Appleton Papers, Inc., a subsidiary of the National Cash
Register Company, in Combined Locks, Wisconsin. The recently modernized
pulp mill is representative of one of the newer developments in high-
yield pulping employing a chemical pulping process in a combination
with mechanical pulping. The method of pulping is usually referred
to as the chemimechanical (CM) process. More than 90 percent of the
original wood raw material is recovered as usable fiber for high-grade
specialty papers, and especially for telephone directory papers. High-
yield pulping processes utilize much more of the wood raw material
than can normally be expected for conventional chemical pulping systems*
and with resultant substantial reductions in water, air, and solids
waste disposal problems .
However, the portion of the wood which is fully dissolved or partially
solubilized in this high-yield pulping process presents a substantial
disposal problem in the mill program of compliance with effluent quality
standards. This new CM process is still in early stages of commercial
development and pulp production methods remain to be standardized.
Procedures will have to be developed for solving the pollution problems*
This demonstration has been directed to evaluation of one possible
route for recovery and concentration of the dilute wash from secondary -.
refining and washing of the pulp. Such a concentrate could then be
processed for disposal together with the strong spent liquor from the
202
-------
1st stage of press afining of the pulp. The various stages of refining
and washing of the pulp on conventional belt filters and related washing
equipment results in substantial dilution of the wash water effluents .
first objective in undertaking this field demonstration was to
survey the various steps of refining and washing the pulp in this
recently revised pulp mill. It was desired to establish the most effec-
tive point for collecting most of the solubilized or partially solubi-
lized material being washed from the pulp in the least possible volume
which could be economically collected and fed to a concentrating system
based on RO.
The total volume of the very dilute wash waters presently being dis-
charged ranges upwards of 1,250,000 gallons per day, with a total solids
content of less than 0.2 percent. Preliminary sampling indicated the
principal flow of pulp wash waters, if isolated from other more dilute
flows, might have a volume of about 550,000 gallons. Analysis of small
samples of such wash waters is summarized in Table 70, and calculations
of the apparent pollution load are provided in the second and third
columns .
Further isolation and recovery of the strongest wash waters in a commer-
cial operation might be expected to further reduce the volume and still
collect as much as 80 percent of the solubles and semisolubles . The
resulting volume of several hundred thousand gallons of strong spent
liquor might then be expected to be especially suitable for RO concen-
tration, starting with a feed at about 1 percent solids and increasing
the concentration to about 10 percent solids. That intermediate concen-
trate at 10 percent solids from .the membrane process could then be
evaporated economically by conventional methods , along with the strong,
Digester liquors to yield- a final concentrate containing substantially
&H of the pollution load from the pulp mill. Recycle of the reclaimed
Permeate water recovered by the RO system back into the final stages
pulp washing could be expected to further improve the recovery of
solubilized and partially solubilized materials to the point of
Achieving nearly 100 percent closure of the pulp mill cooking and wash-
ing system.
This demonstration, No. 5,. was then designed with those objectives
in mind. The following pages of reporting cover the various stages
°f developing the individual unit operations in the attempt to achieve
a practical system of RO concentration for the dilute pulping wash
Caters from this pulp mill.
The Ghemimechanical Pulping Process
high-yield pulping process is characterized by mild soaking in
a continuous chemical cook to soften the chips, followed by mechanical
^fining to separate the fibers of lignocellulosic materials. Figure
37 provides a flow sheet of the CM pulping system for the Locks Mill,
inclusion of an RO pilot treatment system. This mill, at the time
203
-------
of making these field trials, was rated at 200 tons pulp per day (85
tons unbleached and 115 tons bleached). The mill pulps aspen "hard"
wood (often referred to as poplar), by an alkaline sodium sulfite con-
tinuous quick cook of the chips in a high consistency pulping operation
with primary and secondary refining, followed by a low consistency
cleaning and washing operation. The principal effluent streams from
the mill include the pressafiner liquor flow and a combined effluent
wash water from the final stages of washing and refining of the pulp.
TABLE TO
POLLUTION LOADING OF DILUTE PULP WASH WATER
FROM A CHEMEMECHANICAL PULP MILL
Discharge volume = 550,000 gallons per daya
Discharged per ton of pulp = 2750 gallons
Production = 200 tons per day
Concentration
Milligrams Pounds per Pounds per
Constituent per Liter 1000 Gallons Day
Solids 6120 51.0 28,000
BOD5 1980 16.5 9,070
COD 7050 58.7 32,300
Volatile acid 1369 11. U 6,270
Sodium 832 6.9 3,800
Temperature = 65-70°C
pH = 7.5-8.2
preliminary estimate of the volume of dilute pulp wash water does
not include the flow of digester strength spent liquor.
The pressafiner liquor samples contained about 7 percent of total
in terms of both the organic and inorganic materials. One route to
disposal might be based upon a system to concentrate and burn the com-
paratively small flow ( 1+3 j 000. gallons per day) of pressafiner liquor
containing about 25,000 pounds of total solids, 11,000 pounds of BODg»
and 5100 pounds of sodium. Preliminary tests indicated RO flux rates
for the further concentration of the pressafiner liquor from the 7
percent solids level might be very low, and would not be economically
feasible under' the conditions studied. Later in the study the pressa-
finer liquors were found to have high osmotic pressure due to content
of KaaSOi*. Further studies at higher velocities and higher operating
20U
-------
ro
o
V/I
Figure 37. Chemimechanical Pulping Flow Sheet
-------
pressure might alter this preliminary conclusion substantially and
thereby reduce evaporation costs.
De-watering of the high consistency pulp coming from the" secondary
refiners seemed a logical approach to the problem of isolating the
wash waters in the smallest practical volume. At this point the high
consistency pulp contained a relatively large portion of the total
pollution load from the pulp mill. Several samples of wash water,
which had been hand pressed from the high consistency pulp available
from the blend tank at this point in the pulp washing system, confirmed
these preliminary conclusions, For the purpose of conducting this
field demonstration, provision was made for setting up a Zenith screw
press previously used in the old deinking mill which had been shut
down and dismantled when the new chemimechanical pulp mill was placed
in operation.
The press waters from the Zenith Press were found to have a temperature
range of 65-70°C and a pH range of 7-5-8.2. The temperature was adjusted
for membrane concentration processing to Uo-^5°C and the pH range to
6.0-7.0. The feed liquor contained relatively large amounts of partially
solubilized material in the form of fine colloidal particulate organics,
and at times the concentration of these fines were as high as 25 percent
of the total solids (or 1.5 g/1 in 6 g/1 total solids). The presence
of these fine hydrocolloids in the feed liquor could be of special
concern in designing a concentration and disposal system free of fouling
and plugging problems.
After installation of the Zenith screw press, a sampling program was
undertaken for the purpose of evaluating the efficiency of this method
of dewatering the high consistency pulp. The analytical data from
two sets of samples are summarized in Table 71» together with the re-
sults of calculations to determine recovery efficiency in terms of
solids, COD, BODs, sodium and volatile acids (as acetic).
The data presented in the table derive from the analysis of solutes
recovered in the pressate from two runs at different press settings.
The press cake of moist pulp fiber from each run was then subjected
to controlled laboratory washing by filtration of the reslurried 7
percent consistency pulp through laton Dikeman 3fo. 6l5 filter paper.
The filtrate from the first wash corresponded to the effluent from, the
normal mill washing step. The total of the solutes contained in the
pressate, plus the solutes in the first stage wash, provided a measure
of the total pollution load discharged from existing washing practices
within this mill. A second washing through a lab filter, after again
reslurrying the fiber pulp to 7 percent consistency, provided a measure
of the solutes remaining in the pulp as it leaves the pulp mill for
production of paper. The total of the solutes in the pressate, plus
that in nitrates I and II, provide a measure of the total solutes
fed to the press. The press efficiency could then be calculated in
terms of the percentage of solutes actually recovered by the press
206
-------
from the total of the solutes normally removed in the total washing
operation (Pressate + Filtrate I).
TABLE 71
CALCULATION OF PRESS EFFICIENCY FOR COLLECTING CM WASH WATER SOLUTES
Basis; 1 Gram of Dry Pulp
Feed to the Zenith Press - 13.28 grams water
Cake from the Zenith Press « 1.86 grains water
Pressate from the Zenith Press * 11.42 grams water • 0.011 liters water
(assuming a specific gravity of presaate
- 1.0)
Constituent
Milligrams ger Gram of Drj Pulpin
Pressate Filtrate
II
Filtrate
Solutes
Removed
in Wash
Total
Solutes
to Press
Press
Effi-
^ 8
ciency,
percent
Set No. 1
Total solids
COD
BOD
Sodium
Volatile acid
Set No. 2
Total solids
COD
BOD.
Sodium
Volatile acid
65,5
77.4
22.4
8.7
19,0
80.6
82.5
30.6
5,4
21.2
12.7
16.0
4.0
1.7
3.7
9,9
11.4
3.1
1.2
4.1
6.3
8.2
1.9
0.9
2.3
4.9
6.6
1.8
0.6
2.9
78.2
93.4
26.4
10.4
22.7
90.5
93,9
33.7
6.6
25.3
84.5
101,6
28.3
11.3
25.0
95.4
100,5
35.5
7.2
28.2
77,5
76.1
79.1
77.0
76.0
84.5
82.0
86.2
75.0
75.1
6Pres« efficiency, percent - (Solutes In the pressate/total solutes fed to
the press) X 100
Attainment of the desired level of liquor collection at 80 percent
Recovery seemed to be well established from these tests on samples
Stained in sustained operation with use of commercially available
S^eas equipment. Furthermore, some initial concern over the effect
°f a screw press operation on pulp fiber quality was apparently without
foundation. However, the mill technical staff felt the recovery probler
be better taken care of with even higher levels of efficiency
fiber quality using more modern dewatering equipment.
20?
-------
The mechanical stresses of the screw press apparently did bring about
a side problem experienced with the actual operation of the press.
This was concerned with the apparent increase in the content of colloidal
organics in the pressate. It was suspected that the screw press served
to remove a substantial amount of the colloidal material from the fiber
which might result from hydrating and removing polysaccharides in a
colloidal form. This problem along with its effect on the membrane
performance is again referred in more detail in later discussion of
the data for this field demonstration.
Availability of the screw press, together with its installation and
testing by the mill staff, greatly accelerated progress in undertaking
and executing this RO field demonstration.
PROGRAM AND DATA OF THE FIFTH FIELD DEMONSTRATION
Equipment
Small pilot RO equipment, similar to that used for processing bleach
effluents in the fourth demonstration, was considered best adapted
for conduct of this field trial. Volume production of reliable membrane
modules suitable for re-equipping the large trailer unit had not been
proven in the intervening months since completing the third field demon-
stration, although the large unit remained in continuing operation
for developing engineering design data, with the less than reliable
supply of rebuilt modules delivered in September, 1968. Smaller numbers
of newly improved tubular modules were available from development opera-
tions at Calgon-Havens and were used in this trial. A duplex pump
unit, a module test stand complete with automated controls, plus tempera-
ture controlling equipment with a tubular heat exchanger and pH control
equipment, were moved into the mill and connected to the feed supply
system based on the Zenith Press installed by the mill staff, see photo,
Fig. 66 (Appendix C). A field engineer from the Institute staff super-
vised installation and maintained the unit in operation, with a careful
program of daily sampling. Most of the analytical control studies
were carried out in the Institute laboratory a few miles away.
A flow sheet provided in Fig. 38 provides a schematic presentation
of the various unit operations of pulping, pulp refining and washing,
liquor collection, and of RO concentration processing of the pulp wash
water after cooling and pH adjustment. The pulp slurry from the blend
tank after the secondary refiners was dewatered in the Zenith Press
as previously described. The pressate flow at about hd gallons per
minute was screened through a Sweco vibrating screen of 100 mesh, and
then was pumped to the main storage tank having a capacity of 5000
gallons, sufficient for 12 to lU hours of operation for the RO unit.
The tank was filled twice daily from short 2 to 3-hour runs of the
press. Feed liquor from the storage tank was cooled to ^0 to ^5°C
by passing through the tubular heat exchanger, and the pH was then
adjusted to less than T«0, with dilute sulfuric acid in an agitated
50-gallon mix tank. Overflow from the mix tank passed to the 50-gallon
208
-------
ro
O
1. Pulping Operations
2. Liquor Collection
Screw Prems
3, Reverse Osmosis Concentration
Priaary_ 11(1
Cleaners UU
Palp to further
cleaners and vasheirg
t
Precooler
i5° Max.
H*SO» 1 &
CD
To tank truck
for processing
at IPC in pilot
scale and in
large trailer
units
Reverse Osmosis
Peed Tanks
— 600
_Jpsig
-\ 6QO
f/fsatfeari ^ RO
<»fr»i»>Ar\ Con eeat rate
i k. gQ permeate Water
Main Pressurizing
Pumps
Figure 38. Schematic of Reirerse Osmosis at Appleton Papers Inc.
-------
RO feed tank and then to the Duplex Milton Roy pressurizing pump feeding
directly to the RO system. Each of the individual piston pumps in
the duplex pumping unit fed a separate, parallel "bank of three 18-
tube Calgon-Havens modules. The modules were manufactured and delivered
during the early Spring months of 1971- The pump unit and module test
stand, with automated controls, have been previously described. The
conductivity of the feed solution was monitored by a Foxboro dynalog
conductivity recording unit, as a continuous indication of the sodium
and indirectly of the solids content of the wash water being processed.
After the startup it was found necessary to add a slow speed, 70-inch
agitator operated by a 5 Hp mixer, to the 5000-gallon storage tank
to keep the colloidal floe in suspension.
Data and Results
The experimental program was conducted in two phases. The principal
operations with the pilot unit located at the mill comprised Phase
I and were directed to establishing a continuous, straight-through
processing of the dilute feed liquor to establish operating parameters
on a sustained basis. It was not possible to concentrate to 10 percent
solids with the small unit operated in that manner. However, on occasion
when the feed liquor supply was interrupted by pulp mill shutdowns,
several short concentrating runs were conducted at the mill by operating
the unit on a recycle pattern rather than straight through.
Phase II studies to further establish operating methods and data for
the higher levels of concentration to 10 percent solids were conducted
with small and large units located at the Institute in Appleton using
feed liquor supplies trucked from the mill.
Phase I RO Pilot Operations at the Mill
The RO unit at the mill was operated continuously in so far as possible
around the clock and through the weekends for the six-week period from
May 12, 1971 through June 30, 1971. However, the liquor supply was
interrupted at times by mill shutdowns. Brief 2 to lf-hour shutdowns
for module washups occurred at the end of each run. Under these condi-
tions, eleven runs were completed at the mill, for which analytical
data on membrane performance and rejections are summarized in Table
72, and flux rate performance is shown in Fig. 39 •
As the experimental program developed, a pattern of operating problems
developed in two chief categories. A serious loss in flux rates was
encountered as concentration of the wash waters advanced toward the
10 percent solids level. Advancing the velocity of flow through the
•tubular modules helped, but provided only a partial answer to the re-
duced flux rates. The first area for substantial improvement developed
late in the study of concentrating problems after it became apparent
that the osmotic pressure was quite high in this -type of pulping efflu-
ent due to the content of Ete^SOif. The osmotic pressure, which advanced
210
-------
DUttF 72
RO PERfQSMNf.E WHILE PROCESS IMC
CHEMIMECHAMOU. PULP mSH »TERS
Run
Solids, mg/1
U.
IB
IDA
108
IOC
11
Feed
Permeate
Rej,, percent
BODs. Bg/1
Feed
Permeate
Rej . , percent
COO, mg/l
Feed
Permit*
Rej., percent
Sodium, mg/l
Feed
Peimeote
Re] . , percent
VCA, mg/1
Feed
Permeate
Rej., percent
Optical Density
at 281 mi
Permeate
pB
Feed
Permeate
Temperature, °C
Pressure, psig
in
oat
Velocity, ft/sec
Flux, gfd d 40**
nine, hours
5062
104
97.9
1445
56
96.1
5830
159
97.3
1372
20
98.5
1443
59
95.9
0.545
6.7
6.1
3S.2
610
555
2.9
4.7-8.3
13
14.854
221
98.6
3654
92
97.5
1.06
6.6
6.1
36.0
600
445
4.45
2.3
35
8568
64
99.2
3058
32
99.0
138SO
126
99.1
743
16
97.8
2380
36
98.5
0.507
6.7
6.2
36.4
600
465
4.45
Z. 9-6.0
186
14,150 5718
97
98.3
1670
46
97.2
6660
144
97.8
752
17
97.7
1260
22
98.2
0.775
6.5 6.6
6.2
37.5 42.2
600 600
460 460
4.45 4.55
2.7 6.0
44 SO
6367
118
98.1
2216
56
97.5
7115
143
98.0
863
24
97.2
1440
63
95.6
0. 795
6.8
6.1
38.2
600
460
4.55
4.8-6.0
108
6377
270
95.8
1487
141
90.5
7417
269
96.4
886
62
93.0
1474
181
87.7
1.423
6.7
6.5
42.0
600
455
4.55
4.5-6
134
13,270
170
98.7
3240
112
96, S
15,620
229
98.5
1643
59
96.4
3127
142
95.4
0.933
6.5
6.6
43.0
700
545
4.55
,6 1.8-2.5
22
11,960
140
98.8
6010
38
99.4
13,730
154
98,9
1550
38
97,6
2844
249
91.2
0.697
6.5
7.1
43.9
600
450
4. 55
1.9
23
6630
245
96.3
1110
118
89.4
7400
228
96.9
894
S3
94,1
1830
152
91 7
0.848
7,3
6.9
41.1
590
455
3.3
31
4613
292
93.7
3060
196
93.6
10345
400
96.1
1288
107
91.7
2646
226
91.5
1.400
6.2
6.5
35.0
700
555
4.5
5.2
63
19,710
935
95.2
6095
232
96.2
10,975
699
93.6
2740
204
92.6
4417
376
91.5
2.105
6.2
6.6
35.0
700
550
4.5
3.1-5.2
6
41,175
410
99.0
17.700
134
99,2
45,200
356
99.2
5270
98
98.1
9240
218
97.6
1.20
6.6
6.5
38.9
700
550
4.5
1.3-2.0
3
6188
234
96.2
2548
103
96.0
7373
264
96.4
765
56
92.7
1429
110
92.3
1.17
6.4
6.6
41.8
600
470
4.5
2.8-6.4
54
-------
ChemiiMChwdcal Pulp Wash Water
Flux Sate tor Six Modules Coi-i-octed to Ioe
-------
above 500 psig in the concentrate, substantially reduced and nearly
eliminated the effective driving force of the RO system operating pres-
sure at 600 psig when processing at the higher levels of concentration.
Higher operating pressures to 800 psig could "be expected to effectively
answer this part of the problem.
A second problem more difficult to evaluate became increasingly apparent
as the project advanced. Fouling of the membrane by colloidal organics
in the feed -was observed to increase with operating time and varied
with the age of the feed liquor. Development of a colloidal floe was
observed during a few hours of holding in the 5000-gallon feed tank.
Installation of a slow speed agitator to keep the floe in suspension
noticeably improved the flux rate performance. However, freshly pre-
pared liquor was more easily processed at the mill with less fouling
"than aged liquor. The aging effect and fouling with reduced flux rates
Were even more apparent in liquor hauled to the laboratory where the
fouling problem further increased with holding time. A search was
conducted for evidence of a hydrating reaction on the polysaccharide-
derived colloids, and indications pointed to probable increase in the
Content of hydrocolloids. Minimizing the holding and processing time
°f the feed liquor to reduce the postulated hydration reactions, seemed
to be the most promising route to maintaining high flux rates during
concentration after allowing for the osmotic pressure problem.
A detailed description of the eleven experimental runs conducted over
the six-week period of operation for Phase I at the mill follows.
Each of the eleven runs of from 10 hours to one-week duration (Table
T2) was followed by a washup with BIZ detergent, and in all cases the
flux rate was restored, in terms of a control test with NaCl solution,
s-nd in the initial rates of flux for the next run on wash water feed.
It will be noted that Runs 7> 8, .and 10 were exclusively operated as
concentrating runs with recycle of the concentrate, and Runs 1 and
3 were also concluded with brief concentrating runs on recycle.
Run Ho. 1
The first run was conducted first for 13 hours on a continuous, straight-
through feeding operation, with feed liquor having an initial solids
concentration of ^.2 g/1 and increasing to 6.1 g/1. The flux rate
Cropped from 9-7 gfd to U.5 gfd in the first 7 hours of operation.
velocity was advanced to ^.5 ft/sec from the starting rate of 2.9
at 2k hours when the flux rate had decreased to 3.1 gfd, but
vith only a slight improvement to be noticed in the flux rate. A concen-
trating trial on recycle was undertaken as Run Ib at the 27th hour,
continued for 8 hours with a further drop in flux rate to 2.3 gfd.
operating pressures were 600 psig at the inlet and from 555 psig
outlet in Run la to kk5 psig in Run Ib. Rejections for both la
ib runs were excellent at levels well above 95 percent for all
categories measured, including BOD5- It was apparent that increasing
213
-------
the velocity failed to improve the flux rate significantly once the
membrane had been substantially fouled.
Run No. 2
Sustained operation was carried out for 189 hours in the second run.
An inlet velocity of h.5 ft/sec was maintained, which controlled much
of the loss in flux rate observed early in the first run. Each batch
of fresh feed from the press produced an immediate improvement in flux
rate. This flux improvement was temporary for a period of only a few
hours, but provided evidence for a liquor aging reaction, with apparent
production of a colloidal floe capable of fouling the membrane. The
floe could readily be observed during formation and settling in the
large storage tank. Its formation could not be prevented, but installa-
tion and use of the slow speed agitator in midrun served to keep the
floe in suspension with some improvement apparent in the flux rate.
Operation was continued through the weekend, but the supply of feed
wash water was limited, and to maintain operation the velocity of flow
had to be decreased to 3.8 ft/sec. The flux rate dropped off steadily
with the reduced velocity, and continued a steady decline for the next
four days, after which the unit was shut down and given a BIZ detergent
washup. The flux rate on water and on 0.5 percent NaCl test solution
was apparently restored fully by the washup.
Pressure pulsing which had alleviated fouling problems with the NSSC
liquor in the second field demonstration had no apparent affect on
the fouling produced in processing the high-yield chemimechanical (CM)
pulping wash water. Various pulsing schedules were tried for 1-minute
periods each hour to 3 hours, without apparent effect when the velocity
of flow was maintained above 2,5 ft/sec.
Run No. 3
A new start with dilute wash water feed in a continuous, straight-
through study of flux rates, with pressure pulsing and storage tank
agitation, was established, but the supply of feed liquor failed at
26 hours after a shutdown for motor trouble in the pulp mill. The
unit was converted to recycle operation for concentration and reached
a level of lU.l g/1 solids after 19 hours. The flux rate dropped
steadily from 5.95 gfd to 2.7 gfd. The run was terminated at 37 hours
and the unit given a BIZ detergent washup, which again restored the
water and NaCl flux rates to a normal level of 8.0 gfd. Fouling
appeared to be the chief factor causing the reduction in flux rate.
Runs No. U, 5» and 6
Three successive runs followed. On each of these the periods of achiev-
ing flux rates sustained above ^.5 gfd were of increasing duration.
These were conducted with continuous, straight-through operation utiliz"
ing agitation in the storage tank and pressure pulsing with velocities
-------
Maintained at 4.55 ft/sec, and with inlet pressure at 600. psig. Agita-
tion was added to the small feed storage tank in Run 4, "but the runs
were in other ways conducted "by methods similar to that in Run No.
2. Run 4 continued for 50 hours before the flux rate dropped below
the 4.5 gfd level. Run 5 held the flux for 100 hours and Run 6 for
120 hours before reaching the 4.5 gfd level. Each of these runs was
terminated with BIZ detergent washups , which restored the fresh water
and NaCl test water flux rates to the original level. The curve for
the falling flux rate was steeper on this wash water than for all other
pulping effluents tested in this project, with the exception of kraft
pulping rewash water. But indications in these test runs pointed to
capabilities for maintaining continuous operation on dilute feed at
rates above 4.5 gfd for practical periods of several days to a week
or more between short period washups .
A minor problem occurred in acidifying the feed for early stages of
Run 6. The analytical data show somewhat lower rejections, ranging
^elow the usual 95-99 percent level to the 88-95 percent level. Appar-
entlys some sodium acetate was hydrolyzed by the excess acid to release
free acetic acid which then passed through the membrane and was lost
in the permeate.
Runs 9 and 11
additional runs on dilute feed liquor with straight-through operation
were carried out with a modified hookup for the two banks of modules
to put the second bank in series after the first bank instead of two
^anks in parallel. The low level of feed concentration in the first
"bank was a minor factor in evaluating the results, which tended to
further verify the picture for early stage concentration of the wash
Water at moderate flux rates .
brief runs (ib, 3, 7, 8, and 10) were conducted at the mill as
Preliminary trials of concentration to higher levels reaching toward
the goal of 10 percent solids concentration. All were conducted with
Recycle, since concentration was not possible on a straight-through
operation with the limited equipment at hand. The flux rate dropped
Precipitously on all trials , with advancing levels of concentration
approaching 5 percent solids and higher. The flux rate was shown to
be readily restored by washing, and experiments indicated the wash
With BIZ detergent might be unnecessary. Warm water seemed sufficient
to restore the flux to nearly the original value for routine operation,
although the BIZ washup still seemed advantageous for close experimental
control of membrane performance in experimental studies .
Run 11 was made to compare the performance of Banks 1 and 2 under dif-
ferent conditions of washing after substantial fouling occurred. In
"both cases the initial flux rates at 5.6 and 5-8 gfd decreased over
a period of 24 hours to 4.1 gfd. Bank 1 was then washed with. water
Which brought the flux rate back to 6.4 gfd. Bank 2 was not washed
and the flux continued to drop from 4.1 to 2.8 on the second day of
215
-------
operation. Operation continued for another day and the' flux rate for
Bank 1 dropped.at much slower rates of fouling to 4.25 gfd.
Runs 7 and 8 were brief concentrating runs starting with feed liquor
aged' for less than one day. In each case the starting flux rate after
a washup was 7 gfd. In Run 7 "the concentration during recycle increased
from about 6 g/1 solids to lU g/1, and for Run 8 the concentration
reached 12 g/1. For Run 7 'the operating inlet pressure was maintained
at 700 psig. The flux rate-dropped to 1.8.in one bank and to 2,5 in
the second bank for an average of 2.2 gfd. In Run 8 the inlet pressure
was 600 psig and the flux rate dropped to an average of 1.9 gfd. Both'
runs were conducted with a velocity of U.55 ft/sec. These results
were disappointing, but another run at an operating inlet pressure
of 700 psig maintained a flux rate of 5-2 gfd for 63 hours at a starting
feed concentration of h.6 g/1 solids and which was carried to a concen-
tration of 19.7 g/1. The average flux rate at 5*2 gfd indicated the •
increased operating pressure might substantially care for the problem
of balancing the rising osmotic pressure as the concentration progressed.
The concentrating run was continued in Bun lOb for six hours, during
which the flux rate was maintained by Bank 1 at 3.1 gfd and at 5-2
gfd in Bank 2. The concentration was then continued to ^.11 g/1 in
Run lOc for another three hours of recycle operation at 700 psig. The
flux rate dropped to 1.3 in Bank 1 and 2.0 in "Bank 2.
These concentrating runs were apparently greatly handicapped by the
growing evidence for development of a fouling agent, probably in the
nature of a hydrocolloid of polysaccharide origin as a. time-based reaction
or aging effect during the recycle operations. It was apparent in
the straight-through operations that fresh liquor had an aging effect
in a period as short as 2-U hours after passing through the screw press.
These concentrating runs were necessarily carried out by recycle for
substantially greater periods of time, ranging above 20 hours to achieve
concentrations approaching 5 percent solids.
It is indeed disappointing that equipment could not be made available
for concentration runs on a straight-through type of operation, employing
a minimum of recycle with freshly prepared liquor. The holding time
for these concentrating runs on a straight-through flow system would
be expected to be, on the order of 30 minutes or less in commercial
operations. These runs provided growing evidence that" development
of hydrocolloids or gels occurs in this type of spent liquor, and espe-
cially after a dewatering treatment through a screw press.
Additionally, the screw press may have inherent a disadvantage in the
high pressures, and the resultant stresses and abrasion of the fiber.
Such action may serve to remove hydrocolloids which would otherwise
remain bound to the pulp fiber by other methods of less vigorous de-
watering known to be commercially available for this type of operation.
There was considerable evidence that the fouling effect and the aging
effect varied to some extent from sample to sample of the press liquor.'
This could be further evidence that the aging effect and the fouling
216
-------
may be substantially reduced by employing more suitable methods of
dewatering the high consistency pulp produced by this chemimechanical
(CM) pulping process.
Continuing Study of Methods for Concentrating
Above 5 Percent Solids
The problem of developing satisfactory methods of RO concentration
of the CM wash water at solids levels above 5 percent were not being
satisfactorily solved with the EO equipment installed at the mill.
Concurrent studies were undertaken with small and large equipment avail-
able at the Appleton facilities of The Institute of Paper Chemistry.
Fifty-gallon samples of fresh press water ranging about 5 to 7 grams
solids per liter were collected and conveyed from the mill with the
least delay en route. These press waters were promptly concentrated
about 10 to 15 times by reverse osmosis, under recycling conditions,
to remove 90 to 9h .percent of the water, and thus achieving 8 to 11
Percent total solids in the final concentrate. A final reverse osmosis
Product at that range of concentration could feasibly be used as feed
"to a standard evaporation system.
^fortunately , ye could not concentrate large volumes of the pressate
at sustained and reasonable levels of flux rates with minimum of re-
cycling under the desired conditions of operation in the large trailer
unit. Reactions on aging of the pressate apparently produced excessive
quantities of colloidal gels when truck loads of the pressate were
stored more than 20 hours in our 5000-gallon feed tank. However, we
^id have reasonably high flux rates for smaller drum quantities of
fresh pressate, for which aging was limited to not more than 15 to
20 hours during transport, storage, and processing. In order to over-
come this limitation, 50-gallon samples of fresh pressafiner liquors
vere collected daily from the mill at about 6 to 7 percent solids con-
centrations and were further concentrated by reverse osmosis to about
8 to 9 percent solids. Fresh pressafiner liquors showed relatively
less fouling of the membrane and their flux rates were somewhat better
even when the osmotic pressure reached 500 psia at higher concentrations
°f the liquors.
this study, we used the latest Mark III design of Ps-series 18-
tube Calgon-Havens modules with Wo. 310 membranes. The following para-
graphs deal with studies on concentration runs conducted on these samples
°f press and digester liquors.
Osmotic Pressures of Chemimechanical (CM)
Fressate and Digester Liquors
°ae of the first objectives was to determine the effective driving
force for different concentrations of the pressates. The flux rate
at any concentration of the liquor is difectly related to the osmotic
Pressure of the liquor. This pressure increases as the salt concentra-
t;I-on increases during RO processing. The higher the osmotic pressure
°f the liquor, the lower the flux rate for a fixed operating pressure.
217
-------
The dilute samples of the pressate were concentrated from 0.7 to 0..8
percent solids by RO, whereas the digester liquors were concentrated
from T.O to 10.0 percent solids. During these concentration runs,
feed, concentrate, and product samples were taken. These samples were
analyzed for solids and osmotic pressure. A Vapor Pressure Osmometer
(VPO), which operates on the principle of lowering of the vapor pressure,
was used to measure the osmotic pressure of each sample of the liquor.
An Had solution was used as a reference for the VPO.
Figure ho plots the osmotic pressures versus total solids for pressate
and press liquors. It is apparent from this curve that the osmotic
pressures for press liquors were high as compared to those for the
pressates, and increased linearly from 355 psia at 65 g/1 to ^50 psia
at 100 g/1. Apparently, there were more low molecular weight inorganic
materials, especially NazSOi,, present in the pressafiner liquors. The
osmotic pressures of the press waters also increased linearly with
the concentration, but were less than 300 psia for liquor solids concen-
trations up to 10 percent solids.
Figure Ul gives the specific gravity of the liquor at 20°C for various
concentrations of pressate and provides a quick method of estimating
the solids content for use in controlling the RO processing of the
pressate waters.
Flux Rates and Rejectic-ri RatiosDuring
Concentration Runs of Fresh Press Liquors
Two concentration runs were made with 50-gallon samples of fresh liquor
using one module at ^0°C and 680 psig average pressure. The press
liquor, as received from the mill, had 5 to 7 g/1 total solids, and
the average of the amounts of suspended solids in these pressates varied
between 20 and 25 percent of the total solids. In one concentration,
the pressate was processed without removing suspended material, whereas
for the second concentration run the pressate was centrifuged in a
Sharpies super centrifuge before RO to remove as much as possible of
the floe forming colloidal suspensoids. Higher velocities on the order
of.,U.6 feet per second were maintained throughout these concentration
runs on liquor pretreated to minimize concentration polarization and
fouling effects. During the concentration runs, feed, concentrate,
and product samples were taken. These samples were analyzed for total
solids, sodium, volatile organic acids, BODs, and COD.
Figure h2 plots flux rate versus concentration of the liquor for centri-
fuged and noncentrifuged pressates. The flux rates for noncentrifuged
pressate decreased from 8.2 gfd at 7 g/1 to 3.2 gfd at 107, g/1, whereas
for the centrifuged pressate the flux rates are somewhat higher at
all concentrations. An exception was apparent at 76 g/1 solids, at
which level the flux dropped to 1.5 gfd. This low flux rate may have
been due to the aging effect of the liquor at 33 hours for the centri-
fuged pressate as compared to 22 hours for noncentrifuged pressate.
In addition, the flux rates of both centrifuged and noncentrifuged
218
-------
hO
•H
(I)
oj
a
o
HI
h
3
(0
01
8
-p
o
o
600
500
300
200
100
I.
2.
RO Concentrates of Centrifuged Press Liquors
RO Concentrates of Screened Digester
Strength Liquors 2
\
0
20
1*0 60
Total Solids, g/1
80
100
Figure l»0. Osmotic Pressure of Pressate and Digester Liquors
S19
-------
o
o
Cvj
+J
flj
o
•rl
0>
P4
1.06
1.05
1
o 1-03
1.02
1.01
1.00
0
20
ItO 60
Total Solids, g/1
80
100
Figure 1*1. Specific Gravity-Concentration Relationship
for Press Liquors
220
-------
Average Pressure = 680 pslg
Feed Liquor Temp = 1*0°C
Average Flow Rate = 2,8 gpm
e k.6 ft/sec
10
8-
$
H
O Noncentrifuged Press Liquor —
RO Processing Time * 22 hours
A Centrifuged Press Liquor —
RO Processing Time = 33 hours
*%,
^eos
I
20
W 60 80
fotal Solids, g/1
100
120
Figure 1*2. Flux Hates During Concentration Buns with Fresh Press Liquors
'press ates are less than the theoretical flux rates as determined from
the osmotic pressure under nonfouling conditions. This indicates
that there has been constant fouling- of the membrane throughout the
Concentration runs . It was not possible to control this fouling of
the membrane even at velocities as high as 7 to 8 feet per second
(equivalent to Reynolds Number = 28000-32000). Th e was some specula-
tion that electrical attraction may have existed between the negatively
charged colloids and the .cellulose acetate membrane. The literature
snd small-scale exploratory tests indicate the charge on the membrane
is normally negative, but the Intensity of this charge is very small.
Time was not available for a thorough study of the effect of electrical
charges on the fouling of the membrane.
Table 73 lists the flux rates and rejection ratios for both centrifuged
^d non centrifuged press ate. The rejection ratios for centrifuged
Pressate are excellent and range between 95 and 99 percent for all
221
-------
components, whereas for noncentrifuged pressate the rejections vary
between 85 and 95 percent, Uiese relatively low rejections in the •
aase of noneentrifuged pressate may have "been due to a slightly "leaky"
module in this particular run.
TABLE 73
FLUX RATES AND REJECTION RATIOS DURING CONCENTRATION RUNS
WITH FRESH PRESS LIQUORS
Average Pressure » 680 psig
Feed Liquor Temperature » 40°C
Average Flow Rate * 2.8 gpm » 4.6 ft/sec
Total
Solids
g/1
, Flux Rate,
gfd
Re lection Ratios
Solids COD
Noncentrifuged Press Liquor
6.8
11.8
22,6
46.6
107.0
6.8
13.2
33.2
51.8
76.1
8.2
7.1
6.3
4.8
3.2
Cen^rifuged Press
9.7
7.5
5.7
5.0
1.5
92.7
93.4
94.2
94.1
94.9
Liquor
98.8
99.3
99.5
99.6
99.7
(Total RO
95.6
•
96.4
96.7
..
BOD5
Processing
_.
__
M m
--
--
(Total RO Processing
98.7
99.1
99.6
99.6
99.7
98.7
99.1
99.6
99.6
99.7
, percent
Sodium
VGA
Time « 22 Hours)
82.9
84.3
85.8
83.3
__
Time - 33 Hours)
96.4
98.8
99.3
99.3
99.4
91.5
*•«•
87.9
92.4
...
98.5
98.4
98.8
99.5
99.8
222
-------
It -was also observed that the RO concentrates from noncentrifuged pres-
sate were highly viscous compared to those of centrifuged pressate.
The RO concentrate obtained without centrifuging formed a gel at 10.0
Percent solids, whereas the pressafiner liquor did not have such a
highly viscous character even at 50 percent solids concentration. This
gel formation and the viscosity effect led strength to the hypothesis
that the screw press may remove a substantial amount of colloidal mate-
rial following an hydration reaction on the polysaccharides to give
the colloidal form. It was interesting to find the viscosity of this
10 percent solids concentrate was greatly reduced at kO to 1*5°C tempera-
ture. This could promise much less of a fouling problem, if and when
niembranes become available for operation at temperatures of k5°C and
above.
Aging Effect of Press Liquors on Flux
Rates and Rejection Ratios
Throughout this study processing difficulties had been observed which
Seem to be associated with aging of the liquor. Staff chemists working
in the field of wood chemistry postulate the probability that hydration
°f polysaccharides may be taking place in the pressing of the pulp
and during aging of the pressate to yield a colloidal form. These
hydrated polysaccharides may then be developing some electrical affinity
or attraction to the membrane. This affinity of the negatively charged
colloids might very well result in severe declines in flux rates, either
by increasing the thickness of the membrane or by plugging of the micro-
Porous structure.
Figure h3 gives the history of flux rates versus operating hours for
centrifuged and noncentrifuged pressates. All the flux rates were
nieasured at ^0°C and 675 psig average pressure. Reynolds Numbers as
high as 20,000 (equivalent to U.5 to 5.0 feet per second of velocity
in 1/2 inch tube) were maintained throughout the flux rate runs for
the purpose of controlling concentration polarization and fouling of
the membrane. The data in Fig. U3 indicate the flux rate remains almost
constant at 7.5 gfd on relatively fresh noncentrifuged liquors having
&n age of less than 17 hours. The flux rate decreased rapidly to about
1.0 gfd, with liquors stored longer than 17 hours. As a check, a fresh
centrifuged sample of pressate was mixed with a sample aged for ^0
hours after centrifuging. This mixture having a solids concentration
°f 62 g/1 was shown to produce a rapid decrease in flux rate to 2.0 .
gfd during the first k hours of RO processing. These experimental
studies further confirmed a deleterious effect of aging on flux rates
for both centrifuged and noncentrifuged pressates. Centrifuging may
Remove all or part of the gel-like floe or other suspended material
contributing to fouling. If so, it apparently does not bring the aging
Reaction to a halt and further foulants apparently continue to form.
Figure k3 also indicates that the flux rates of both centrifuged and
noncentrifuged pressates are restored immediately to their original
values each time after tap water flushing of the membrane.
223
-------
RO Feed = Noncentrifuged Fresh Press Liquor
t)
*H
M
,11. i* g/i
I
I
10
8.U g/1
8r
20 30 1»0
Hours of Operation
50
60
RO Feed = Centrifug.ed Fresh Press Liquor, 5-6 g/1
Centrifuged ^0 Hours Old RO Concentrate, 62 g/1
Average Pressure = 675 psig
Feed Liquor Temp = ilO°C
Average Flow Rate = 2.7-3.0
= 1».5-5-0 ft/sec
10
Figure 'i
20
50
30 »»0
Hours of Operation
Aging Effect of Press Liquors on Flux Rates
60
-------
Table 7** shows, the effect .of aging on pressates in terms of flux rates
&nd rejection ratios. Flux rates decline rapidly with time and with
increase in solids concentration. Rejections are rather uniformly
high but show a trend toward further improvement as aging progresses .
Low starting rejections are known: to improve substantially, however.
Kie results of these studies suggest that severe flux rate declines,
associated with the aging effect, may perhaps be due to a buildup of
a secondary membrane, such as might be formed by colloidal gels attracted
"to the surface of the cellulose acetate membrane rather than due to
the' plugging of the porous membrane. Although compaction of the membrane
is well known to affect membrane performance, this did not appear to
fce significantly involved under the conditions being studied.
Jk
EFFECT OF PRESS LIQUORS AGING 01 FLUX
AND REJECTION RATIOS
Average Pressure = 675 psig
Feed Liquor Temperature = hQ°C
Average Flow Rate = 2.7 gpia « U.5 ft /sec
Time of - • Total Flux * '•• . • • •
Operation, Solids, Rate, _ Rejection Ratios, percent _
hours g/1 gfd Solids COD BOD5 Sodium ?OA
N on centrifuged Fresh Press Liquor
2 6.0 7-8 98.0 — — 97-0 89.9
60 11. h l.o 98.2 — — 97.5
Centrifuged Fresh Press Liquor and Centrifuged
to Hours Old RO Concentrate
1 Q.k 8.3
12 36.6 1.5
27.5 8l.lt 1.1
99.6
99.6
99.6
98.9
99. i*
99-5
98.9
99. U
98.1*
99.5
99.3
99.8
99.7
study and proving will be required, but it is possible to postu-
late the aging effect of the liquor, which results in rapid flux rate
Declines under high levels of recycle and during long holding time
for the liquor in process, may become less significant in a large-
acale RO unit which operates on a straight-through processing cycle
vith minimum holding time in process. In a large-scale unit, the feed
*ould be concentrated on a straight-through basis to the desired solids
concentration using a large number of modules. The total holding' time
°f the liquor in straight-through processing could be 30 minutes or
225
-------
less and seldom more than one hour even if some recycling may "be re-
quired to achieve full concentration.
OF
The exploratory studies conducted during the six-week field trial and
in the supporting laboratory evaluation of the problems encountered
have served to establish RO as an interesting contender among the
alternative methods to be considered for processing the dilute wash
water effluents from this chemimechanical pulp mill.
Substantial operational problems and difficulties remain for further
study, and practical methods of operation must be developed before
it can be Known whether commercial-scale installation of an RO membrane
concentration system could be a feasible answer to the effluent treat-
ment problems at this mill. However, solid accomplishments derived
from this field demonstration of the capabilities for RO and in identi-
fication of those problem areas which remain.
A first objective of the study has been concerned with developing a
feasible method for recovery of a suitable RO feed liquor containing
a major portion of the solubillzed and partially solubilized wood com-
ponents and of the cooking chemical residues contained in Minimum volumes
of the wash water. The trials of the screw press indicated that de-
watering of the high consistency pulp immediately after the secondary
refiners does answer these requirements very well. Tests of two settings
of the screw press in operation showed 80 percent recovery of the total
wash water effluent solids was being readily achieved. Removal of
the major portion of the solids at that early stage of refining and
washing should substantially reduce water requirements for final stage
washing, and should greatly increase possibilities for recycle and
reuse of the final stage wash waters.
Such recycle and reuse of the final stage wash waters would pick up
the 20 percent of the wash water solids remaining in the wet press
cake of pulp leaving the screw press , Recycle of the wash water con-
taining these solids would bring them back to the press for recovery
in the next cycle. The mill would then be within reaeh of total treat-
ment and disposal of all pulping process effluent solids.
The indicated volume and solids concentration relationships for collec-
ting the chemimechanical (CM) wash waters are summarized in Table 75•
The table also provides related comparative data on findings from this .'
demonstration and on probable goals for design and operation of a eononer*
cial concentrating system. The wash waters from CM pulping are presently
discharged to the mill sewer with other dilute flews in about 1,250,000
gallons per day at a solids concentration of 0.1 to 0.2 percent. The i11*
dicated flow of wash waters which could be collected for RO processing W
pressing the high consistency pulp from the blend tank .after the secon-
dary refiners was estimated by the mill engineering staff to be about
550,000 gallons 3 with 0.5 to 0,6 percent solids at the time of making
226
-------
"tills study. Further improvement in design and operation of equipment
for dewatering the pulp could reduce the volume of the wash water feed
to the RO concentrating system to give volumes as low as 200,000 gallons
per day and solids concentrations at the 1.0 to 1.2 percent level.
TABLE 75
INDICATED FINDINGS AND GOALS FOR CONCENTRATION
OF CM EFFLUENTS
Collectible Effluent
Volume, gal. /day
Total solids, percent
jftj Concentrate
gal. /day
solids, percent
£ged Available for
Evaporation
gal . /day
Rates
(To concentrate from
1 to 10 percent solids),
gfd
Present
Situation
1,250,000
0.1-0.2
Indicated
Findings this
Demonstration
550,000
0.5-0.6
50,000
to
75,000
5 to 10
Probable Goals
for Mill
Installation
200,000
1.0-1.2
20,000
to
30,000
7 to 12
i*o,oooa
to
50 ,000
90,000
to
125,000
60,000
to
80,000
2.5-7.5
8 to 10
^gester liquor volume presently estimated at U3,000. gal./day at
7 percent solids.
o
•W-gester liquor plus RO concentrate.
^e second principal objective for Demonstration No. 5 on CM pulping
vash waters was directed to proving the capability of RO for concen-
^ating this solution of wood-derived organics, which are apparently
^stable due to continuing reactivity with residual pulping chemicals
227
-------
The data accumulating throughout this run provide much evidence .of
an aging effect probably due to reactivity in the form of hydration
of the' polysaecharides . This aging effect becomes apparent within
a few hours after collecting the liquor, and progresses actively in
the first 2k hours to yield gummy hydroeolloids which have been found
to foul the membranes and seriously reduce their performance.
Performance of the RO pilot unit operating in the mil on a straight-
through flow shows the fresh pressed liquor could be processed at
relatively high rates of flux on the order of 7.5 gfd during the first
U hours, and at rates above 5 gfd for periods up to 17 or 18 hours
of holding time for the feed liquor. Fouling becomes severe after
a day of storage.
The materials fouling the membranes were readily removed with a water
wash. An enzyme-type detergent added to the water may have accelerated
the washing under difficult conditions. No permanent effects on mem-
brane performance were apparent during the 6 weeks of operation in
the mill and in studies carried out concurrently at the laboratories
in Appleton.
It was especially difficult to conduct sustained concentrating runs
at levels above 5 percent solids because of the reactive aging effect
leading to severe fouling of the membranes during recycle operations
extending beyond 17 hours. Equipment was not available for conducting
straight-through concentration runs on liquors which were fresh and
more easily processed in the mill. A number of small-scale runs were
carried-out to the 10 percent solids level in recycle runs of limited
duration. These runs provided evidence pointing to probabilities that
fresh liquor could be processed continuously to achieve solids concen-
trations at the 10 percent level with properly designed equipment opera-
ting to process liquor with holding times of an hour or less with a
minimum of recycle. Operation under such conditions would be normally
expected in large-scale commercial equipment. These findings were
the basis for further tabulation of indicated findings and g. als
provided in Table 75-
It should, therefore, be expected that engineering design and plant
operation of an RO concentrating system in this chemimechanical pulp
mill could be directed to collecting 80 percent or more of the wash
water solids in about 200,000 gallons of dewatering effluent. Reverse
osmosis concentration would be expected to reduce the volume by a factor
of about 10 times to give from 20,000. to 30,000 gallons of intermediate
concentrate .at 7 to 12 percent solids. This intermediate concentrate
could then be combined with about to,000 to 50,000 gallons of digester
effluent at 7 to 10 percent solids to give a final 50 percent solids
concentrate readily processed for disposal by combustion or alternately
for recovery of pulping chemicals and other possible values.
228
-------
A number of critical problems req.id.re further research and development •
°f engineering design factors "before an economically feasible commer-
cial operation can be assured and undertaken.
A nodule life study was not a'part of this demonstration. For the
'"ost part new modules were available and used for the 6-week period
of study. ITo major failures occurred and, although one module showed
evidence of a slight leak at one of the tube seals and was replaced,
this was later proven to be a transient leakage that disappeared and
the module was put back into use at a later tine, nevertheless, the
extensive module life studies conducted throughout the course of the
five demonstrations and extensive laboratory studies for this research
^d demonstration project all were plagued with module failure problems
after a few months of operation and showed the critical nature of the
^dule performance and life expectancy over a desired minimum period
°f 12 to 2k months life expectancy. This problem was considered to
^e a responsibility and concern of equipment suppliers in terms of
and of close control of manufacturing operations in fabricating
membrane module equipment. Discussion of the effect of life perfor-
on the economics for application of RO processing of effluents
°f the pulp and paper industry are the subject for further discussion
in Sections IX and X.
Next of importance to the life and performance expectancy'of membrane
eqiuipinent is the practical performance in terms of flux rates of permeate
"through the large and'costly areas of membrane required for processing
several hundred thousand gallons of effluent waters daily. This demon-
stration has indicated that flux rates on the order of 2.5 "to 5-0 gfd
^ight be expected under the conditions of these studies, and that better
Performance ranging aro-und 7-5 gfd might be expected overall with pro-
G(2ssing of freshly prepared feed liquors held for 'periods of one 'hour
°r less when concentrating in the range of 1 percent solids to 10 percent
solids. This level of permeation rat;e has been adequately demonstrated
in sustained operations elsewhere in this report as pertaining to mem-
Cranes manufactured in'the period'1967 to'1970. It should also be
noted that newly improved membranes which have become available during
1971 could double these rates. Rates on the order of 8 to 10 gfd seem
to be a likely goal for designing a commercial installation from this
Point onwards."
Nevertheless, the fouling experienced with the unstable and reactive
feed liquors utilized in this demonstration and also those of similar
Reactivity or instability in the white water neutral sulfite semichem
and the, rewash waters of kraft pulping all point to need for careful
control of processing conditions. Exploratory studies conducted during
this trial point to hydration reactions' on polysaccharides removed
from the pulp as being subject to time-based hydration reactions. There
^as some evidence that the screw press used in dewatering the pulp
to produce the BO feed liquor was probably responsible for abrasion
°f the' lignocellulose fibers, thereby removing more of the reactive
229
-------
polysaecharide material responsible for fouling than might otherwise
be expected in more modern and less severe conditions of dewatering
of the pulp. Development of better methods of dewatering to reduce
the fouling problem is an indicated area for further study. Another
engineering design problem is expected to arise within the development
of the pulp washing system, should the dewatering step with RO concen-
tration be undertaken. Bemoval of 80 percent or more of the wash water
solids from the high consistency pulp coming from the secondary refiners
would greatly reduce the amount of processing required in the final
stages of washing the pulp. Reduced volumes of wash water would be
required for those final stages. Complete recycle of the resultant
dilute wash stage effluents back into the main pulp refining and,pulp
washing operations could be expected to become a large step toward
achieving complete closure of the pulp mill effluent water system.
The remaining 20 percent of the solids collected in the initial dewater-
ing step would be recycled into the recovery system. Careful design
of the system should go far toward achieving this objective.
Remarkably high levels of rejection by the membrane system have been
observed throughout this demonstration study. Recoveries at the level
of 95 percent or better might be expected in processing these effluent
waters by BO. However, the membrane systems are not perfect and the
remaining fraction of 5 percent or less of solids passing into the
permeates could present a minor effluent problem if discharged to receiv-
ing waters. Use of these relatively'clean, reclaimed permeate waters
for final stage pulp washing would be another step toward closing the
system and achieving complete control of pollution problems in this
mill. It was not possible to evaluate the chemical buildup problems
which might result from recycle of low molecular weight materials which
might be passing through the membrane into these permeate waters during
weeks or months of operating a recycle system. Analytical evaluation
and study would be indicated as a final step in determining the amount
of the recycle of the permeate water which could be tolerated. Neverthe-
less, it is probable that most if not all of the permeate waters could
be recycled to some more tolerant point in the. mill system, if not
in final stages of washing the pulp.
The remaining sections of this report further develop the possibilities
for optimizing the engineering data of an RO concentration system and
of evaluating process economics.
230
-------
SECTION VIII
ENGINEERING AND DEVELOPMENT STUDIES
Engineering Design Factors
The material concerning development of design factors for this project,
as detailed on the following pages, has been previously presented in
part as a published paper6.
Prior sections in this series review the development, design, and opera-
tion of reverse osmosis equipment for laboratory, pilot, and five large
field demonstrations for concentration processing of pulp and paper
effluents. Design factors developed in the early phases of these studies
Were often based upon incomplete analytical characterizations of the
Vastes and their effect on performance of membrane equipment. The
resulting estimations for design purposes proved out remarkably well
in setting up the laboratory, pilot, and fairly large field demonstra-
tions at flow rates ranging to 60,000 gallons per day and more. However,
areas of need for more exact analytical and design data have become
aPparent as the trials proceeded and experience with field conditions
developed. Particularly, there has been need for more exact data cover-
ing changes in processing variables as concentration proceeds in the
range of 0.1 percent to 10 percent dissolved solids.
This discussion deals mainly with development of design data on calcium-
&nd ammonium-base acid sulfite, NSSC, and second-stage kraft bleach
effluent liquors in the following areas of study:
Determination of Reynolds Numbers .
Pressure drop and pumping energy requirements.
Determination of osmotic pressures.
Determination of rejection ratios .
Product flux rate -temperature relationship for calcium-base
acid sulfite liquor.
Effect of velocity on flux rates of Na- and Ca-base liquors .
Microbiological fouling and membrane compaction.
Determination of Reynolds Number
of the first objectives in these studies was to determine required
Degrees of turbulence and mixing necessary to minimize concentration
Polarization and fouling of reverse osmosis membranes. These measure-
Btents were developed as Reynolds Number, applicable at various tempera-
tures and concentrations of ihe different types of pulp liquors. This
231
-------
of the important factors of concern in avoiding membrane "fouling"
and for maintaining high flux rates. Concentration polarization pro-
duces several effects detrimental to the membrane separation process:
(l) The osmotic pressure that must be overcome is that
corresponding to the solids concentration at the
membrane surface. This is true, since concentration
polarization causes the effective osmotic pressure
of the bulk of the solution. For this reason, the
required operating pressure for the RO cell is in-
creased by the polarization effect, and the pumping
power requirements will also be increased.
(2) The concentration polarization may have a detrimental
effect upon the dissolved solids content of the product
water because the dissolved solids content 'of this per-
meate will increase as the solids concentration at the
membrane surface increases.
(3) The deterioration of the membrane may be hastened by
increased solids content of the product water, and
concentration polarization can aggravate this effect.
(k) Finally, excessive concentration polarization may cause
precipitation of solids at the surface of the membrane.
Therefore, it is important to reduce concentration polarization by
maintaining turbulent flow across the surface of the membrane. The
calculation of Reynolds Number (Hge)'as a measure of the degree of
turbulence requires the determination of the density and viscosity
of the liquor.
The viscosities of the liquor were determined using an Ostwald Viscom-
eter, whereas the specific gravities were determined using a Pycnometer-
The pH of each sample of the liquor was adjusted between U.0-4.5 with
sodium hydroxide or sulfuric acid, depending on the initial pH. Then
Reynolds Number for different liquors were determined using the experi-"
mental values of the densities and viscosities at different temperatures
and percentage solids of the liquor. The .data for NRe and temperatures
are fitted to a straight-line relationships using the method of least
squares at different percentage solids of the liquor.
The results of Reynolds Number for NSSC white water, ammonium-, and
calcium-base acid sulfite liquors, and second-stage KBE liquors, are
plotted in Fig. kb-hf at different percentage solids and temperatures
of the liquors. .Various linear velocities (u) within the 0.5 inch
diameter (D) tubes were tested."
Turbulent flow is considered to occur at Reynolds Number above 1*000.
From Fig. UU-Uj it is apparent that a velocity of about 1.0 foot per -
232
-------
Q)
CC
2
u=
0.5 INCH
I FT /SEC
25°°25 27.5 30 32.5 35 37.5 40.0
TEMPERATURE/C
Figure kh, Reynolds Number of HSSC White Water
D«0.5 INCH
u = IFT/SEC
% SOLIDS
4500
3000
25 27.5 30 32.5 35 37.5 40.0
TEMPERATURE, *C
1»5. Reynolds Humber of Araaonia-Base Acid Sulfite Liquor
233
-------
0)
tr
z
SECOND
3500
3000
I
25. 27.5 30.0 32.5 35.0 37.5 40.0
TEMPERATURE, °C
Figure ^6. Reynolds Numbers of Calcium-Base Acid Sulfite Liquor
D= 0.5 INCH
u= I FOOT PER SECOND
% SOLIDS
o
cc
z
3000
25. 27,5 30. 32.5 35.0 37.5 40.0
TEMPERATURE, °C
Figure 1*7. Reynolds Numbers of Kraft Bleach Effluent Liquor
-------
second should "be sufficient to produce turbulence at all temperatures
indicated for a solids concentration of:
2.0 percent or less of NSSC white vater.
3.0 percent or less of ammonia-base acid sulfite and second-
stage kraft bleach effluent liquors
k.O percent or less of calcium-base acid sulfite liquor
flowing in a tube of 0.5 inch inside diameter.
For higher concentrations, a velocity of 1.0 ft /sec may or may not
^e turbulent , depending on the temperatures of the liquor. The optimum
and maximum desirable velocities are, of course, at much higher levels,
which are discussed in the following paragraphs .
pressure Drop and Pumping Energy Requirements for a
Commercially Available Tubular-Type Module
A. next step in developing engineering design data for RO concentration
Processing of the four types of pulping and bleaching liquors involved
^termination of the pressure drop in a representative tubular module
(Havens Model J 18-tube module having about ihk linear feet of 1/2-
inch ID tubes in series ) . It was then possible to calculate the pumping
energy as Kwh/1000 gallons which would be required to overcome this
Pressure drop. This is one of the important design factors of concern
in the selection of the number of modules to be used in series. If
velocity is to tie held constant, then by connecting a large number
modules in series , we limit the total flow rate going into the system;
at the same time', the pressure drop increases in proportion to
number of modules .- ! Under such conditions, it becomes necessary
add booster pumps to overcome the pressure drop. Therefore, one
to optimize pressure drop against the total flow rate, while select-
the number of modules to be connected in series.
drops in the 18-tube modules used in this study were determined
different flow rates for h liquors and water.
Reynolds number, at different flow rates, were determined using the
ities and viscosities of each liquor and water. Then the data
the pressure drops and N^e were fitted to a log-log expression,
g the method of least squares. Finally, the pressure drops at
various values of NRe were calculated from this expression, and the
faults are given in Fig. U8 at 35°C for about 10.0 percent solids
concentrations of*1* liquors and water. The pH of ' each liquor was ad-
dusted to U.5.
Assure drop at Nge of ltf),000. is highest (= 166 psig) for NSSC white
vater, and lowest (= 9U psig) for kraft bleach effluent liquor. This
is true because viscosities of various concentrations of kraft bleach
235
-------
TEMPERATURE - 35.0 'C
pH OF LIQUOR. 4.5
t No-BASE LIQUOR
2. Co -BASE LIQUOR
3. NHj -BASE LIQUOR
4. KRAFT BLEACH
LIQUOR
3. WATER
eooo. teooo. 34000. 32000. 40000.
REYNOLDS NUMBER) (NR()
Figure 1*8. Frictional Pressure Drop in One 18 Tube
Havens Module for 10 Percent Solids Liquors
effluent liquor were found to be lower than corresponding viscosities
for NSSC white water at various concentrations.
From Fig. 1*8, the pressure drop may be observed to increase rapidly
with increase in Npe. This is true because the pressure drop is directly
proportional to (RRe)n, where n varies from 1.75 to 2.00. The pressure
drop in three identical 18-tube Havens modules connected in series
was found equal to 3 times the pressure drop in a single l8-tube module.
The higher the pressure drop, the lower the driving force, and the
lower the flux rate. For example, in the case of three 18-tube modules
connected in series, the pressure drop at Npe = 1*0,000 for 103 g/1
concentration of NSSC white water was 1*98 psig. So the average pressure
of 351 psig for an inlet pressure of 600 psig can result in about 55
percent lower flux rates than the flux rates at 600 psig, even at 1.0
percent solids concentration of NSSC white water. Therefore, the pres-
sure drop is a very important design factor in the selection of the
number of modules to be connected in series.
Figure 1*9 gives the calculated pumping energy as Kwh per 1000 gallons
of liquor to overcome the frictional pressure drop at various values
of Npe. The pumping energy curves follow a pattern similar to the
236
-------
1.75
1.50
-------
AP = difference between the applied pressure and the delivery
pressure of product water, psia
ATT = (difference between the osmotic pressures of the liquor
and the product water) + (osmotic pressure increase due
to concentration polarization and fouling effects, psia)
The product water is delivered at atmospheric pressure. Since the
osmotic pressure of the product water is usually very small compared
to the osmotic pressure of the liquor, the former term can be ignored.
In the case of zero concentration polarization and fouling effects,
the driving force (AP - Air) becomes equal to the difference between
the applied pressure (P^) and osmotic pressure of the liquor (IT). There-
fore, equation 10 becomes:
F = A(PA - TT) (11)
From equation 11, it is apparent that the higher the osmotic pressure
of the liquor, the lower the flux rate for a fixed applied pressure.
In case of liquors having osmotic pressures higher than the applied
pressure, there is osmotic flow across the membrane.
A vapor pressure Osmometer (VPO) , which operates on the principle of
vapor pressure lowering, was used to measure the osmotic pressures
of each sample of liquor. An NaCl solution was used as a reference
for the VPO. However, the VPO was useful only for determining the
osmotic pressures of second-stage kraft bleach effluent liquor, and
of the NSSC white water.
The osmotic pressures of calcium- and ammonia-base acid sulfite liquors
could not be obtained by VPO, probably because of association and dls-
association properties of lignosulfonates in these spent sulfite liquors*
and were determined instead by measuring the flux rates at the different
concentrations of each liquor.
These liquor flux rates were then compared with the flux rates of
chloride solutions of known osmotic pressure. All the flux rate runs
were made at a higher velocity to minimize any increase in osmotic
pressure due to concentration polarization and fouling. T5y solving
equation 11 and the sodium chloride flux rate equation, Fs = A(PA - irg)»
the following relationship is obtained for the osmotic pressure of
the liquor:
where
7T = osmotic pressure of the liquor, psia
P. = applied pressure, psia
238
-------
F = liquor flux rate, gfd
Fs = sodium chloride solution flux rate, gfd
TT - difference between the osmotic pressures of sodium
chloride solution and the product water, psia
By substituting the flux rates of sodium chloride and liquor at 600
psig, and the osmotic pressures of sodium chloride in equation 12,
the osmotic pressures of calcium- and ammonium-base acid sulfite liquors
were determined. The results are given in Fig. 50 at 25°C and at dif-
ferent concentrations of the ^ liquors.
From Fig. 50, it is apparent that the osmotic pressures for second-
stage kraft bleach effluent liquor are very high compared to other
liquors, and it increases linearly from 8^.0 psia at 10.0 g/1 to 59*4.0
psia at 100.0 g/1. This is understandable, because there is more in-
organic material, especially NaCl, present in second-stage kraft bleach
effluent liquor as compared to the other 3 liquors. The osmotic pres-
sures of calcium-base acid sulfite liquor and NSSC white water increase
linearly with the concentration, and are less than 300 psia for liquor
Solids concentrations up to 100 g/1. Ammonia-base acid sulfite liquor
was found to have osmotic pressures greater than for calcium-base acid
sulfite liquor and NSSC white water. In addition, the osmotic pressures
of ammonia-base acid sulfite liquor do not vary linearly with the con-
centration.
Rejection Ratio
Dilute samples of each of the h liquors were concentrated from 1.0 to 10.0
Percent solids by reverse osmosis. During these concentration runs,
feed, concentrate, and product samples were taken. These samples were
analyzed for solids, optical density (OD), biological oxygen demand
(BOD5), and chemical oxygen demand (COD). Rejection ratios (R) were
calculated using the following formula:
R = (1 - Cp/Cc)iOO (13)
vhere
Cp = concentration of product
Cc = concentration of feed to the module
Ihe results are given in Table 76" at different concentrations of each
liquor. Rejection of color based on OD measurements at 281 nm is good
and ranged above 98.0 percent except in ammonia-base acid sulfite liquors.
Ammonia-base sulfite liquors contain relatively low molecular weight
colored materials, which are not completely rejected by Type 3 Havens
cellulose acetate membrane. The rejection of solids is above 95-0 .
Percent and range upward of that value in those liquors having good
color rejections. BODs rejections vary between 85.0-95.0 percent,
and were observed to be significantly higher at upper levels of solids
concentration for all types of spent liquors in these studies. COD
239
-------
700
600
0.
a? 50°-
o
g
^ 400
Ul
I 30°
Ul
DC
O.
O 200.
100.
TEMPERATURE- 25.0*C.
pH OF LIQUOR = 4.5
I KRAFT BLEACH
LIQUOR
4. NH3-8ASE
LIQUOR
2< NO-BASE
LIQUOR
3.
CO-BASE
LIQUOR
0. 25. 50. 75. 100.
CONCENTRATION OF LIQUOR, G./L.
125.
Figure ?0. Osmotic Pressure vs. Concentration of the
Liquor
rejections were found to be "better than BODs rejections and solids
rejections, and were above 97.0 percent in most of the observations.
The apparent anomaly with BODs rejections being high at advanced stages
of concentration, in some cases could be explained by permeation of
low molecular weight organics, such as acetic acid, which are high
in BOD5 in early stages of concentration.
Product Flux Rate — Temperature Relationship
The effect of temperature on product flux rate was studied at 12.0
and 10U g/1 solids concentration of calcium-base acid sulfite liquor.
Figure 51 shows the percentage change is flux rate vs_ temperature between
20.0 and U3.0°C. For calcium-base acid sulfite liquor of concentration
12.0 g/1, there was 2.1 percent increase in flux rate for every rise
in degree centigrade of temperature. The percentage increase in flux
rate did not change significantly at IQlt.O g/1 concentration of the
liquor. Flux rate variation with temperature is higher for water and
is about 2.8 percent rise per °C rise . According to Kopecek and
Sourirajan , water flux rate increases due to increase in the membrane
constant, which is related by the following expression at a given pres-
sure .
A/yw = constant
-------
TABLE 76
REJECTIOa RATIOS (FERCESt} FOB 4 LIQUOBS AT A
COHCBNTRATIQN OF 0.2 - 10 PERC2OF SOLIDS
Inlet Pressure = 550-600 pslg
pH of Liquor - 4.5 (Adjusted with SUSCk or HmOH)
Islet Velocity = 4-5 ft/sec
Type 3 Hfcvens Modules
Sanple
K>
-pr
I-1
1. Feed
2. Concentrate
Samples No. 1.
2.
3. Permeate
Samples No. 1,
' 2,
4. Rejection
ratios (percent) Bb. 1.
1. Feed saaple
2. Concentrate
Samples No. 1.
2.
3- Permeate
Sanples Bo. 1.
2. '
lj,_ Bejeetton
ratios (percent) Hb. 1.
2.
2lt-Hour
Ifeut.
Salids,
S/l
Optical
Itensity
at 281
tnn
Calcium-base aetd
11.4
39-5
91.6 '
1.23
2.93
96.90
96.81
aa
267
675
3-9
8.8
98.5**
98.70
Amorxia-tase acid
10.8
38.9
115.9
0.96
2.80
97.5%
97-59
102
323
995
12.5
28.5
96.13
97- lh
B005,
nig/1
COD,
fflg/1
24-Hour
Hsut.
Solids,
g/1
sulfite liquor
2820
107to
2!Q5Q
364
2078
91.96
91,1*6
13910
W120
113630
1515
3356
96,85
97.05
l.
2.
1.
2.
1.
2.
sulfite Utuor
3210
9650
28050
809
1910
91.62
93-19
13620
1*9280
1^3800
oAao
3812
97.12
97-35
1.
2,
1.
2.
1.
a.
11.0
57-1
101.7
1.47
1.95
97-42
98.08
Second
2.8
19-4
86,1
0.84
5.13
95.65
94.04
Optical
Eensity
at 281
fjlYi
i<e
58
298
585
1.3
1.8
99-55
99 .T°
stage kraft
16
117
642
0.5
1.1
99-60
99-B3
BOD5,
rag/1
water
JB70
9475
15900
854
822
90.90
94.84
bleach
189
955
4295
103
256
89.21
94.05
COD,
fflg/1
6220
58200
127000
1298
1670
9? .77
98.69
effluent liquor
1375
11120
47840
186
413
98.33
99-14
-------
eo.r-
i 5°
s
i
B
-\
.2.
I. Co-BASE LIQUOR, I2.0G./L., pH-4.3
2. Co-BASE LIQUOR, 104-0 G./L., pH - 4.5
3. WATER
25.
30. 35.
TEMPERATURE,
Figure 51. Effect of Temperature on Flux Rates of
Water and Calcium-Base Acid Sulfite Liquor
where
A = pure water permeability constant, g-mole
water/sq cm sec atm
liw = viscosity of water, centipoise
According to equation 1^, the increase in flux rate with increase in
temperature is a function of the decrease in viscosity of the solution.
The larger the percentage change in viscosity of the liquor, the higher
the percentage rise in flux rate. However, it is observed experimentally
from Reynolds Number studies that the average decreases of viscosity
per °C rise are 3.0 and 2.1 percent at 1.0 and 10 percent solids concen-
tration of the liquor, respectively. This flux rate increase did not
occur at the expense of hydrolysis of the cellulose acetate membrane.
Hydrolysis of a membrane is a long-term effect, whereas flux rate-
temperature effect is instantaneous. Of course, the rate of hydrolysis
has been found to vary as the reciprocal of the absolute temperature25
and it can result in increase in flux rate at the expense of percentage
rejection.
Figure 51 shows that calcium-base acid sulfite liquor flux rate at
HO.O°C will be 25-0 percent higher than the flux rate at 28,0°C.
-------
There fo re , the higher the temperature, the higher the flux rate. We may
conclude that temperature is an important parameter in the design of
a reverse osmosis plant.
Effect of Velocity on the Flux Rates of NSSC White Water
and Calcium-Base Acid Sulfite Liquor
Permeation Resistance vs Reynolds Number — NSSC White Water
effect of velocity on flux rate is expressed in terms of permeation
resistance, which in turn is determined by calculating the membrane
constant. The membrane constant was determined by using the osmotic
pressure and flux rate data of a known solution of sodium chloride.
permeation resistance (P^) is calculated using the following equation
derived and discussed in the section on osmotic pressure determination:
(15)
1 - ==• 1 - TT^
above equation, Pp, is the sum of the osmotic pressure of the liquor
and the osmotic pressure increase due to concentration polarization
and fouling effects. The effect of velocity on flux rate is expressed
in terms of permeation resistance because the permeation resistance
becomes almost independent of applied pressure and velocity when running
sodium chloride and liquor flux rates under identical conditions of
velocity, pressure, and temperature.
A schematic diagram of the experimental setup is shown in Fig. 52.
^e feed was pumped at about 20 psig from a 500-gallon plastic tank
"to a main piston pump by a centrifugal feed pump. The main pump is
a triplex reciprocating, positive displacement Manton Gaulin pump with
a direct current motor and an electronic variable speed drive. For
s study, the main pump discharged through 11 Havens modules (modified
contain only two tubes) at a pressure of 500 psig and flow rate
ying from 1.2-U.3 gpm in each module. A pressure gage was installed
at the inlet of each module. The flow rate of permeate was measured
from each module. The total permeate of 11 modules was collected and
B&xed in a small tank and then drained to a 50-gallon plastic tank
^derneath the trailer. The permeate from a 50-gallon plastic tank
vas mixed with the concentrate via a small centrifugal Eastern pump.
^e, re combined concentrate and permeate is returned through a heat
exchahger for cooling and then recycles to the 500-gallon feed tank.
different concentration runs of NSSC white water were made at
Velocities of 2.0-7.0 feet per second. The sodium chloride solution
flux rates were measured at 5.0 feet per second for each of the 11
^o^tube modules before each concentration run. Each flux rate run
*as made at 500 psig and 35°C. The results of the permeation resistance
calculated by using equation 12 are shown in Fig. 53. Because of the
^certainty as to accuracy of the osmotic pressures determined for
-------
ro
'E
F= FEED
E-CONCENTRATE
P= PERMEATE
FROM ONE
18-TUBE OR 2-TUBE
HAVENS MODULE
f»t m TOTAL
PERMEATE OF
II OR 22
IB-TUBE
HAVENS
MODULES
MANTON-GAULIN
PUMP
-me*
>ro
z>
8-t*M 10 !«*•»*-
E *P F ®
SHXH "' 6 K****-*-
E *P F
S-tXJ-j 8 ]«OO«-
^_
" 7.
F
? _
c
EASTERN PUMP
-rfrh 1
:
£
;
BACK
PRESSURE
REGULATOR
13
.PIPE
E + P
1
W
O
W
m
.
*
mw
X I
O t)
I>
> to
Ztn
®T
™*
3> J»
-)
COOLING
' WATER
(IN)
-m
2"S,S. PIPE
COOLING
WATER
= (OUT)
I
o
w
m
j
FEED PUMP
5OO GALLON
PLASTIC TANK
figure 52.. Schematic Diagram of Experiment, aL Set.-Uip
-------
26
230
u
u
200-
VI
£
f .70
there is.a uniform error in
the permeation resistances and so the absolute values of the permeation
Resistances at the highest velocities of each concentration is different
?*"om the osmotic pressures of ISSC white water as shown in Pig. 50.
®f course, the relative variations of the permeation-resistances with
velocities are quite accurate.
^he permeation resistance was determined by measuring the flux rates
°f each module and then the average permeation resistance was calculated.
Fig. 53f three sets of curves represent permeation resistances at
values of Npe for three different concentrations of NSSC white
At the same concentration, the permeation resistance increased
vtth decrease in NRe due' to increase in concentration polarization.'
*& lower velocities, the increase in permeation resistance was quite
significant. At lower concentrations of NSSC white water, the permeation
Assistance was lower because the osmotic pressure of the solution was
stibstantially less for that type of feed liquor high in content of
Dissolved salts.
Velocities below which relatively higher permeation resistances were
°^tained, were estimated from the data shown in Fig. 53 and are given
ift Table 77. It is seen from fable 77 that the higher the concentration,
-------
the higher the velocity "below which higher permeation resistance was
observed. This is true because there is an increase in concentration
polarization and fouling with increase in concentration.
TABLE 77
REYNOLDS NUMBER MD VELOCITIES OF NSSC WHITE WATER
BELOW WHICH RELATIVELY HIGHER PERMEATION
RESISTANCES ARE OBTAINED
Concentration, Velocity in 1/2-Inch
g/1 NRe Tube, ft/sec
12.7 16000 3.00
3^.3 19000 U.OO
56.0 22000 5.00
Concentration Polarization and Fouling Study as a Function of
Velocity for Calcium-Base Acid Sulfite Liquor
The effect of velocity on flux rate is expressed in terms of percentage
decreases in flux rate from the starting flux rates . Here the same
schematic diagram of the experimental setup as shown in Fig. 52 was
used. A number of continuous flux rate runs were made at different
concentrations of the liquor under controlled conditions of 35°C and
500 psig pressure. Before the start of each run, the modules were
washed with high velocity water and with detergent BIZ solution of 15
g/1 concentration. The percentage decrease in flux rate observed was
less than 2.0 percent over a continuous run of 97 hours at 3.0 ft/sec
and for 118 g/1 concentration of the liquor. As it became very difficult
to obtain fouling in case of a two-module setup, we decided to put nine
18-tube Havens modules in 3 manifolds, each manifold containing 3 modules
in series. The remaining 8 manifolds had 2-tube modules. The flux
rate of each of the nine 18-tube Havens modules were measured at dif-
ferent operating hours of the continuous run. The percentage decrease
in flux rate from starting flux rate was determined and then the average
percentage decrease in flux rate was calculated.
Figure 5^, gives the average percentage decrease in flux rate vs hours
of operation at two velocities of the liquor for concentrations above
10.0 percent solids of nonprecipitated calcium-base acid sulfite liquor.
From Fig. 5^, it is noted that the average percentage decrease in flux
rate at 70 hours of operation is reduced from 8.0 to ^,0 percent by
increasing the velocity from 0.8 to 1.2 ft/sec.
Figure 55 shows the effect of velocity on flux rate for calcium-base
acid sulfite liquor in which there was a significant amount of precipi-
tated calcium sulf ate solids. The average percentage decrease in flux
in this precipitated liquor is a strong function of velocity and it
2U6
-------
Ul
x
111
cr
U
UJ
UJ
Q.
UJ
I
UJ
APPLIED PRESSURE « 500 PStG
TEMPERATURE - 35.0 °C.
pH OF LIQUOR = 4.5
8.0
6.0
4.0
2-0
-,rt
0,0
I. CONCN,« 127.0 G./L,
VEL.-0.8 FPS
2. CONCN.*II7.0 G./L.
= 1.2 FPS
1
1
I
0. 20. 40. 60.
HOURS OF OPERATION
80,
Figure 5^4. Effect of Velocity on the Flux Rates of Ca-Base
Liquors (Wo Precipitate)
becomes less than 12.0 percent at 1.8 ft/sec over 50 hours of continu-
ous operation, Hie conclusion is that velocity is an important param-
eter in controlling the decrease in flux rate which may result from
scaling and fouling of membrane tubes.
it is noted that we have been able to control concentration
Polarization at velocities even below 1,5 ft/sec. Fouling has not
been apparent in 70-hour runs at these low velocities. However, it
is not economical to have such low velocities because the absolute
"value of flux rate increases with increase in velocity at a rate pro-
. ,2 *
Portional to V . At higher velocities, it becomes necessary to
optimizse this increase in flux rate against the loss of flux rate due
"to frictional pressure drop and the cost of pumping energy.
Microbiological Fouling
Fouling of the membrane surfaces by microbiological growth is often
observed in sustained operations with wastes containing nutrients capa-
ble of promoting growth of bacterial yeast and molds. Where growth
has developed significantly, flux rates may be restored by removing ,
the growth with flows at high velocities, with the aid of detergents,
21*?
-------
33
30
IS
10
5 -
CONCN-8B.OG./L.
VEL.-1 2 FPS
2. CONCN.- 900 G./L.
VEL -1.5 FPS
APPLIED PRESSURE-500 PSIG
TEMPERATURE • 3S.O *C.
pH OF LIQUOR-4.50
CONCN-1220 G./L.
VEL.-I 6 FPS
T).
10.
20. 30. 40.
HOURS OF OPERATION
50.
Figure 55. Effect of Velocity on the Flux
Rates of Ca-Base Liquor (Precipitate)
and by flushing plastic foam balls through the system periodically
at intervals of several days to a week or more depending upon the degree
of growth experienced. We have been concerned with finding ways to
prevent or inhibit such growth, and high velocity appears to be an
especially effective method of keeping the system clean or at least
reducing the frequency of need for^cleanups. Despite the extensive
experience with microbiological fouling on sustained runs, we have
been unable to observe and to maintain and to analyze fouling rates
satisfactorily under carefully controlled conditions ranging to runs
of as much as 120 hours of continuous operation with our standard test
solutions made up of a calcium-base feed sulfite liquor. Significant
increases in the resistance to permeation has been observed at lower
velocity with neutral sulfite semi chemical white water. These studies
are being continued to more closely develop knowledge of the fouling
rates and conditions for preventing the development of microbiological
fouling.
Effect of Pressure on Membrane Compaction
Membrane compaction at elevated pressures should not be confused with
Microbiological fouling. At higher pressures, there is compaction
of the porous membrane layer which results in a decrease in the meiribran
constant and hence the flux rate. The flux rate decline due to reduced
-------
membrane constant does not seem" to "be at the' expense- of rejections .
The' rejection ratios for spent sulfite liquor do not show any signifi-
cant increase with an increase in the membrane constant.
For this study, we set up 3 parallel rows of 3, Type 310 Ps-Series,
Havens 18-tube modules in our large-scale reverse osmosis trailer unit.
A number of continuous flux rate runs were made at different inlet
pressures, 500-800 psig using 1.0 percent solids concentration of NSSC
liquor. Each flux rate run was made for about 2k hours under controlled
conditions of temperature, 35°C, and various inlet pressures. Before
the start of each liquor run, the modules were washed with high velocity-
water and with detergent BIZ solution of 15 g/1 concentration. Higher
velocities, U.2-U.6 feet per second, were maintained throughout all
these runs to minimize any concentration polarization and fouling of
the membrane. The flux rates of individual modules were measured at
the end of each 2k hours continuous run. Then the membrane constant
fA" was determined using the osmotic pressures of NSSC liquors.
Table 78 gives the effect of pressure on the membrane constant and
Dejection ratios for three Havens modules operating under identical
conditions. Figure 56" plots the membrane constant versus average oper-
ating pressure for these three modules. From Fig. 56 » it is noted
that the membrane constant decreases almost linearly with increase
in pressure for all the three modules , and the rate of decrease in
the membrane constant "A," as determined from the' slope of the straight
line, varies "between 1.05 x 10" 5 and 1.15 x 10" 5. The' membrane constant
and the flux rate decreases "by about 17 percent due to membrane compac-
tion with an increase in the operating pressure from 500 to 800 psig,
whereas the rejection ratios do not show any significant change.
Finally, it is, noted that the flux rate loss due to membrane compaction
is very significant and one should include this effect while optimizing
the increase in flux rate at higher pressure against the capital cost
of modules, and the cost of pumping energy and membrane replacement.
CONCLUSIONS
A velocity of 1.0, ft/sec is sufficient to produce turbulent flow at
temperatures to 35°p and at solids concentrations up to 'if percent for
all four liquors. For high concentrations , a velocity of 1.0 ft/sec
or may not be turbulent depending on the temperatures of the liquor.
Pressure drop and pumping energy for the four pulping and bleaching
effluents studies were maximum for NSSC white water and minimum for
kraft bleach effluent liquors.'
Kraft bleach effluent liquors had the highest osmotic pressures which
increased linearly from Bk.O psia at 10.0 g/1 solids to 59^.0. psia
at iQO,,0 g/1 solids. The' osmotic pressures of the other' three liquors
Banged to about 300 psia. at 10.0 percent solids concentration.
-------
TABLE 78
EFFECT OF PRESSURE ON THE MEMBRANE CONSTANT
A AND REJECTION RATIOS
Feed Liquor Used = NSSC Liquor
Feed liquor Temperature » 35 C
Average Flow Bate - 2.5 a.8 gpa - 4.2-4,6 ft/sec
Membrane Constant,AjQ-O
t\3
\fl
o
Average
Pressure,
pslg
483
583
683
783
Concentration
of Liquor,
aA
10.6
9-9
10.0
9.8
Flux Rate, gfd
Module
Mo. 1
9.8
11.2
12.5
13.0
Module
No. 2
8.9
10.4
11.9
12.6
Module
Mb. 3
8.5
10.0
11.3
11.7
Percent Rejection Ratios of Consp. Permeate
Optical tensity
at 281 nm
99.2
99A
99-3
99-5
Solids
97-9
97.6
95-5
97.6
COD
97.7
97. b
95.7
97-5
BOIL
96.7
93.6
88.8
93-8
Sodium
96.7
97.2
95.6
97-3
gfd/psig
Module
So. 1
2.17
2.02
1-91
1.72
MOduie
No. 2
1.97
1.88
1.81
1.67
Module
So. 3
1.88
1.80
1.73
1.55
-------
2.4
if
c
n
c
o
o
v
c
X
2.2
2.0
1.8
1.6
1,4
1.2
Module
1.
Membrane Compaction Rate
as Slope or Rate of Change
In Membrane Constant "A"
1,15 x
1.05 x
10
10
1,10 x 10
••5
-5
I
I
1.0
400 500 600 700 800
Average Pressure, psig
Figure 56. Membrane Constant vs Average Pressure (BO Studies)
Rejection of OD at 281 am was above 98 percent for all liquors except
for ammonia-base acid sulflte liquor, which averaged 97 percent. The
Dejection of solids was above 95-0 percent, whereas BQDS rejections
varied between 85-95 percent. COD rejections ranged higher than for
and for solids rejections.
increase in flux rate per °C rise was 2.1 percent for calcium-
base acid sulfite liquor in the solids concentration range of 1.0-
lO.Qxpercent. The flux rate increase with increase in temperature
higher for water, and was about 2.8 percent for each °C increase
temperature.
251
-------
Reynolds Number of th order of 16000-22000 (equivalent to 3 to 5 ft/sec
in a tube 0.5-inch inside diameter) may "be necessary to prevent concen-
tration polarization and fouling in 12-56 g/1 concentrations range
of KSSC white water. For 10 percent solids calcium-base acid sulfite
liquor, the average decrease in flux rate over 70 hours of continuous
operation was reduced from 8.0 to ^.0 percent by increasing the velocity
from 0.8 to 1.2 ft/sec. This was true for liquors in which calcium
salts had not precipitated. The flux rate decline observed in liquors
in which precipitation of calcium salts occurred was observed to be
relatively more dependent on the velocity.
The flux rate decreased by about 17 percent due to membrane compaction,
with an increase in the operating pressures from 500 to 800 psig.
CONTROLLED STUDY OF MEMBRANE FOULING
AND CONCENTRATION POLARIZATION
Membrane fouling problems have been a chief operating problem throughout
the laboratory, pilot, and field demonstration studies conducted for
this research and demonstration project. Such problems are well known
to be of concern in other membrane research projects, such as in the
saline water conversion field and various methods of reducing or elimi-
nating the fouling effects have been developed. However, this project
concerned with concentration processing of wood pulping and bleaching
effluents has been especially involved with progressive loss in flux
rates as the concentration process advances above 5 percent solids
to desired levels of 10 percent solids for use as evaporator feeds .
The problems have, therefore, been somewhat specialized and unique
to this industrial waste treatment field. Progress has been made toward
developing operating procedures to reduce the fouling effect but much
has remained to be learned.
A series of carefully controlled studies of sufficient magnitude to
provide a good statistical base (eleven sets of single modules in paral-
lel and eleven sets of two in parallel) were conducted with close contro-
of velocity of flow, pressure, and temperature and of solids concentra-
tion. These studies permitted a much more careful analysis than had
heretofore been possible for the effect of velocity of flow across
the membrane surface in association with osmotic pressure effects
occurring as the concentration increases. The study has gone far toward
developing adequate answers to questions, and problems arising throughou
the 1^-year study, and have resulted in development of operating param-
eters to optimize the membrane concentration processing of the pulping
and bleaching effluents.
The concentration polarization effects arising from low levels of veloc-
ity and turbulence can be associated with several other, and apparently
different sources or causes for fouling. A secondary "dynamically11
formed membrane or film develops in the processing of some types of
large molecular weight organics, such as the lignins contained, in
252
-------
pulping liquors and has "been much studied23. Plugging of the micro-
porous structure of the basic cellulose acetate membrane "by penetration
of foulants is believed to "be another separate and distinct cause of
fouling, A sharp increase' in osmotic pressure is especially apparent
in concentrating substrate bleach liquors and pulping wash waters con-
taining salts, such as Had and NaaSO^. Osmotic pressures in the upper
levels of concentration of these substrates above 5 percent solids
may reach 300 psig or even higher, and substantially reduce or even
eliminate the effective driving force when operating pressures are
in the 500 to 600 psig range, with resultant sharp fall-off in flux
rates. Membrane compaction is another cause for reduction in flux
rates which has been much studied. Evidence that membrane compaction
can be a problem at elevated' pressure above TOO psig has developed
at times in this study, but this seems to be a matter of less concern
than are fouling and concentration polarization at pressures below
700 psig.
In order to develop a better knowledge of the causative factors and
to formulate operating parameters for optimum processing of these pulp-
ing and bleaching wastes, a series of systematic studies in moderately
sized equipment were planned and initiated as reported on the following
pages.
These studies were conducted in the Effluent Processes Group laboratories
at The Institute of Paper Chemistry using Havens modules in the pilot-
scale units and also in the' large-scale trailer-mounted field demonstra-
tion unit which had been moved to the Institute campus. The feed liquors
employed in these studies were prepared at various concentrations in
50 to 500 gallon batch quantities by diluting various evaporated concen-
trates , digester liquors and bleach effluent concentrates with tap
water. The concentration of the feed was kept constant throughout
each of these studies by recycling both the concentrate and permeate
back to the feed tank.
This report describes two areas of study, first on the concentration
and fouling relationships in Ca-base and Ma-base pulp wash waters,
and secondly on the permeation resistance ys velocity relationship
for Ca-base acid sulfite, HSSC and for KBE effluents; concentration
polarization and fouling as functions of velocity for Ca- and 1H|-
base acid sulfite liquor.
The objective of this study was to determine the velocity required
to overcome concentration polarization and fouling of tubular reverse
osmosis systems. This is one of the chief factors of concern in main-
taining high flux rates and of minimizing capital and operating costs.
For this study we used the' large-scale trailer unit equipped with from
11 to 22 Type 310 R-Series Havens l8-tube modules. This number of
Modules was sufficient to achieve a statistical base- for interpreting
results. The feed liquor was maintained at about 10 percent solids
and was prepared by diluting evaporated Ca-base SSL concentrate with
253
-------
tap water. The experimental setup for the trailer unit was the same as
described previously for the velocity studies and as shown in Pig. 52.
A number of continuous flux rate runs were made at different velocities
using 95-130 g/1 concentrations of Ca and lHa-base liquors. Each flux
rate run was conducted continuously for periods ranging from 70 to
100. hours under carefully controlled conditions of 35°C and 500 psig
pressure. Before the start of each run, the nodules were washed with
high velocity water and with detergent BIZ solution of 15 g/1 concentra-
tion.
The flux rates of each of the modules were measured and expressed in
terms of the percentages of the initial starting flux rate at different
operating hours of the continuous run. Then average percentages of
initial starting flux rates were determined at various hours of opera-
tion. The flux rate velocity studies were made for two different module
configurations :
Eleven 18-tube Havens modules — 11 parallel rows with one
module in each parallel row.
Twenty-two l8-tube Havens modules — 11 parallel rows of
two modules in series in each parallel row.
Tables 79 and 80 give the effects of velocity on flux rate declines
and rejection ratios due to concentration polarization and fouling
of the membrane for Ca- and NHg-base acid sulfite liquors. Figures
57 and 58 plot average percentage of initial flux rate versus hours
of operation for a setup of 11 Havens modules at various velocities
and at about 9-13 percent solids concentrations of the liquors. From
Fig. 57 it can be noted that the average percentage of the initial
flux rate at 72 hours of operation for Ca-base liquor increases from
88 to 92 percent by increasing the velocity from 2.k to 1*,5 feet per
second. For Ms-base liquor, Fig. 58 shows that the average percentage
of the initial flux rate at J2 hours increases from 90 to 97 percent,
with an increase in velocity from 1.8 to k.5 feet per second. Tables
79 and 80 list the average percentages of the initial flux rate at
2k, kB, and 72 hours, along with the average percentage rejection ratio3
of solids, optical density, COD, BODg, calcium and nitrogen for various
velocities of Ca- and Ms-base liquors. For 22 modules set up in 11
parallel rows of 2 modules in series, the average percentage of the
initial flux rate at 72 hours and at 1.7-1.8 feet per second becomes
83 percent for Ca-base liquor and 89 percent for NHa-hase liquor.
rejection ratios of solids, OD, COD, calcium and nitrogen are all
96 percent, whereas the BODs rejections vary between 85 and 95 percent*
The rejection ratios and the average percentage of the initial flux
rate are higher at higher velocities, which indicates that there is
decrease in concentration polarization and fouling of the membrane
with increase in the velocity.
-------
TABLE 79
EFFECT OF VELOCITY ON FLUX RATE DECLUIES AMD REJECTIOH RATIOS DUE TO CQNCEHTRATIQI
POLARIZATION AND FOULING FACTORS FOR CALCIUM-BASE ACID SULFITE LIQUOR
Inlet Pressure = 500 psig
Concentration of Liquor = 120-127 g/1
Feed Liquor Temperature = 35°C
pH of Feed Liquor « U.5
ro
V/l
Average
Velocity in
1/2" Tube,
ft/sec
Average Percent of Initial
Flux Rate
Average
Flux Rate,
gfd 2¥ Hours 48 Hours 72 Hours 281 nm Solids COD BODs Calcium
Average Percent of Rejection Ratios
OD at
Eleven 18-tuhe Havens modules. — 11
Jrows- of- one module-
2.4
3.5
4.5
1.8
1.7
5
5
5
u
1*
.0-6.6
.5-7.7
.0-7.1
Twenty-two
.0-6.0
.0-5. 4
in series
92
97
98
in each
90
94
96
. — f ..... —
parallel
88
__
92
row
99
99
99
l8-tube Havens modules — 11 parallel
in series
90
88
in each
88
85
parallel
85
83
row
99
99
.6
.7
.3
rows
.2
.3
98.6
98.9
98.6
97.4
96.2
98.2
92.7
92.0
95.4
99-6
99.7
99.5
of 2 modules
98.3
98,3
96.5
96.7
88.0
88.5
99.4
99.5
-------
TABLE 80
EFFECT OF VELOCITY ON FLUX RATE DECLINES AND REJECTION RATIOS DUE TO CONCENTRATION
POLARIZATION AND FOULING FACTORS FOR AMMONIA-BASE ACID SULFITE LIQUOR
Inlet Pressure = 500 psig
Concentration of Liquor = 95-106 g/1
Feed Liquor Temperature = 35°C
pH of Feed Liquor = U.5
ro
v?
OS
Average
Velocity in
1/2" Tube,
ft/sec
1.8
2.U
3.6
U.5
1.8
1.7
Average
Flux Rate,
gfd
Average Percent of Initial
Flux Rate
Average Percent of Rejection Ratios
OD atNitrogen
2U Hours 1*8 Hours 72 Hours 281 run Solids COD BOD5 Ammonia
Eleven l8-tube Havens modules. — 11 parallel rows of one. module
in series in each parallel row
U.6-6.6
3.7-6.0
5.7-7.1
U.5-7.3
97
97
98
99
93
95
95
97
90 99.1 98, U 95-7 89.9
93 99.5 99. ^ 98. U 91.5
95 99.2 99.0 9^.3 89.5
97 99. ^ 98.9 97.3 90.5
Twenty-two 18-tube Havens modules —11 parallel rovs- of -2 modules
in series in each parallel row
5.0-6.7
H.5-6.5
96
95
92
90
90 99.9 98.3 97.7 85.5
89 99.9 98.3 97.7 86.0
97-9
96.3
98.0
97.8
95-8
96. U
-------
99
97
4J
H
fcl
fi
95
93
91
89
87
85
Set up:
Eleven ifl-tube Havens modules — 11 parallel
rows of one nodule in series in each
parallel row
Cone - 123 g/l
Vel - 3.5 ft/sec
Gone - 120 g/l
Vel - 4.5 ft/sec
Cone » 127 g/l
Vel -2.4 ft/aec
_L
20
40 60 80
Hours of Operation
100
Figure 5J, Effect of Velocity on Flux Bate Declines Due to
Concentration Polarization and Fouling Factors
for Calcium-Base Acid Sulflte Liquor (RO Studies)
25?
-------
4)
O
Oi
&
0)
99
97
95
fn
^ 93
«
91
89
87
85
Set up: Eleven 18-tube Havens modules — 11 parallel
rovs of one module in series in each
parallel row
Cone
Vel
-106 g/1
A.5 ft/sec
Cone • 99 gA
Vel - 3.6 ft/sec
Cone - 103 g/1
Vel - 2.4 ft/sec
95 g7l
1.8 ft/sec
l
I
I
20
40 60
Hours of Operation
80
100
Pigxare 58. Effect of Velocity on Flux Bate Declines Due to
Concentration Polarization and Fouling Factors
for Ammonia-Base Acid Sulfite Liquor (RO Studies)
258
-------
Finally, it.is noted that we were apparently able to operate with free-
dom from concentration polarization and fouling at velocities even
below 2.0 feet per second. We had little or no evidence of fouling
of the'membrane surfaces by microbiological growth in 70-100. hour runs
at these low velocities. However, it would not be economical to employ
such low velocities because the absolute value of flux rate increases
significantly with increase in velocity. This effect of velocity on
the absolute values of flux rates is discussed in the next section
of this report.
' Permeation Resistances — Velocity Relationship for Calcium-Base
Acid Sulfite, NSSC, and Kraft Bleach Effluent liquors
The objective of this study was to determine the effect of velocity
on the absolute values for the flux rate. It has been observed by
Aggarwal and Sourirajan21* that the flux rate increases with increase
in velocity at a rate proprotional to V0*6, This is true because the
thickness of the concentration boundary layer decreases, and so the
mass transfer coefficient increases with increase in velocity. At
higher velocities, it becomes necessary to optimize this increase in
flux rate against the lose of flux rate due to frictional pressure
drop and the cost of pumping energy. Therefore, this study is important
in determination of the optimum velocity, not from the point of view
of overcoming concentration polarization and fouling, but from the
point of view of increasing the rate of mass transfer.
The effect of velocity on flux rate is expressed in terms of permeation
resistance, which in turn is determined by calculating the membrane
constant. The membrane constant was determined by running sodium chloride
solution flux rates. The permeation resistance is calculated using
equation (15).
We used our pilot-scale Milton Boy duplex test stand with pumping capacity
Up to 6 gallons per minute. The feed liquor was pumped from a 50-
gallon plastic tank and its concentration was kept constant "by recycling
both the concentrate and permeate back to the feed tank.
In this study, we used a single Havens module modified to contain only
two tubes at a pressure of 500 psig and velocity varying from 0.2 to
7.8 feet per second. A number of continuous flux rate runs were made
for about 8 hours at different velocities and at about 1.0 and 10.0 ,
percent solids concentrations of the liquors. The sodiua chloride
solution flux rates were measured at k.O feet per second before each
new concentration run. All the flux rate runs were made under constant
concentration conditions at 35°C and 500 psig pressure. Before the
start of each liquor run, the module was washed with high velocity
water and with detergent BIZ solution of 15 g/1 concentration. The
Module was also subjected to a "hard pulsing" treatment during all
the liquor .runs to minimize the fouling effects.
259
-------
Table 81 gives the effect of velocity on permeation resistances and
rejection ratios for about 1.0 and 10,0 percent solids concentrations
of Ca-"base acid sulfite, NSSC and kraft bleach effluent liquors. Figure
59 plots permeation resistance versus velocity at various concentrations
of the liquors. From Fig. 59§ it is noted that the permeation resis-
tances decrease considerably with increase in the velocity. The permea-
tion resistances decrease by about hO-JQ percent up to a velocity of
2 to 3 feet per second. The rejection ratios also increase with increase
in the velocity of the liquor. Above a velocity of 3 feet per second,
the permeation resistances do not show any significant decrease which
probably indicates that the' mass transfer' rate becomes constant above
this velocity,
Therefore, it may be necessary to maintain a minimum velocity of 3
feet per second (2 gallons per minute) for maximizing the mass transfer
rate and hence the flux rate. But one must optimize this increase
in flux rate against the loss of flux rate due to frictional pressure
drop and the cost of pumping energy.
CONCLUSIOIS
Minimum velocities as low as 2 feet per second may be required to over-
come concentration polarization and fouling effects up to 10 percent
solids concentration of Ca and NHa-base acid sulfite liquors. However,
it may not be economical to have such low velocities from the point
of view of maximizing the mass transfer rates. It may at times be
necessary to maintain a minimum velocity of 3 feet per second for maxi-
mizing the mass transfer rates and hence the flux rates.
DEVELOPMENT OF A COMPUTERIZED MATHEMATICAL FOR THE
OPTIMAL DESIGN OF LARGE-SCALE REVERSE OSMOSIS SISTEME
A third area for supplementary engineering and development for this
project has "been directed to a systematic study of variables for the
optimum design of large-scale reverse osmosis units to be used for
concentrating dilute effluents of the pulp and paper industry. One
of the first and important objectives was to determine the number of
membrane modules' and their best configuration for large, multiple-
stage concentrating systems. This is one of the important factors
in optimizing the capital and operating costs for the RO process.
For this study, a computerized'mathematical model was formulated using
design data developed with the pilot plant and large-scale trailer
units. Parameters were established for verification of the mathematical
model, and confirming trials were conducted for processing the various
types of effluents on which demonstrations were conducted within this
project. The formulations developed were then utilized in developing
SOUK of the economic data and conclusions reported in Section X.
260
-------
TABLE 81
EFFECT OF VELOCITY 01 III PERMEATIOH RESISTANCES AMD
REJECTION RATIOS - CALCIUM-BASI ACID SULFIT1 LIQUOR, NSSC LIQUOH,
AND KRAFT BLEACH EFFLUENT LIQUOH
Inlet Presiure * 500 psig
Feed Liquor Temperature • 35°C
Average
Velocity in
1/2" Tube,
ft/sec
0.2
0.5
0.9
1.7
2.5
Liquor
Flux Rate,
gfd
Calcium-base
6.5
7.1
7.7
8.0
8.1*
Permeation
Resistance,
psia
aeid sulfite lie'
211
126
93
7k
56
Percent Rejection Ratios
OD at
281 nm Solids
COD BODs
Calcium
or Sodium
lor — concentration =» 13 eA
98.1 95. k
98.5 97.8
08.8 98.3
98.9 98.5
99.0 98.6
Calcium-base acid sulfite liquor — concentration
o.t*
0.9
1.8
2.5
3.3
2.3
3.7
5.0
5.U
5.7
1*10
31*5
290
270
260
99.1 97.3
99.3 97.7
99.2 98,3
99.2 98.5
99.2 98.6
95.2 89.9
97.7 9k. 2
98.2 96.2
98.5 95.0
98.5 96.2
- 128 g/1
96.8 87.3
97.1 88.2
98.0 9k. 9
98.3 95.6
98.1* 9U.8
98.3
99.2
99.2
99.7
99. k
99.2
99. k
99.6
99.6
99.7
NSSC liquor — concentration « 10 g/1
0.3
0.5
0.9
1.7
0.2
1.1
2.0
2.9
5.1
0.3
1.2
3.0
5.1
7.8
3.3
k.i
5.3
6.3
Kraft bleach
6.8
8.2
10.6
12.1
12.7
Kraft bleach
1.7
3.3
5.1
5-5
5.8
368
329
278
235'
effluent liquor
285
21*0
163
112
95
effluent liquor
1*1*7
39k
3l*0
325
317
97.1 82.2
97.5 ' 83,9
97.9 86.0
98.3 88.7
— concentration »
99.8 97.5
99.9 98.5
99.9 98.7
99.9 98.9
99.9 99.1
— concentration »
99.8 96.2
99.9 97.1
99.9 97.5
-- __
99.9 97.5
81.2 51.2
82 .'8 52.6
88.6 55.9
89.0 6l. k
10 R/l
95. U 89.7
97.0 9k, 1
99.7 95.1
99.8 95.0
99.8 96.9
115 g/1
86.7 81*. 3
91.0 90.0
92,7 92.1
— -—
92.3 92.2
81*. 6
85.2
89.3
91.2
„_
__
__
™
__,
__
_ _
__
26'1
-------
440
400
360
320
n
•
P.
8 280
«t
a*
5 240
I
200
160
120
eo
Calcium Base Liquor - Cone. - 13 g./l.
Calcium Baae Liquor - Cone. • 128 g./l.
NSSC Liquor • Cone, - 10 g./l.
K.B. Effluent Liquor - Cone. - 10 g./l.
E. K.B. Effluent Liquor - Cone. - 115 g./l.
A- I
234
Average Velocity! ft./aec.
Figure 59. Velocity vs_. Permeation Resistance — Calcium-Base Acid
NSSC and Kraft Bleach.Effluent Liquors (RO Studies)
2(52
-------
DEVELOPMENT OF THE DATA AND THE MODEL
Basic Formulation and Cong)uter Utilization in
Setting Up a Mathematical Model
The basic formulations used in developing a mathematical model of an
RO concentrating system are two simple equations of flux rate and rejec-
tion ratios. For a semipermeable membrane, the flux rate and rejection
ratio are given by equations (.10). and (13). Here the rejection ratio
is defined as the ratio between the concentrations at both sides of
the membrane at a certain spot and it is an important parameter in
characterizing the quality of permeate for the processes of concentrating
dilute feed or separating chemicals.
Manual computations of equations (10) and (13) for a multistage reverse
osmosis system requires a substantial number of man hours. Also, the
flexibility of the RO system is greatly limited by the speed of manual
computation. In addition, many complexities are introduced in the
solutions of the above equations by the following factors:
Effect of temperature and higher operating pressure on the
membrane constant.
Variation of the osmotic pressures of the liquors with the
concentration.
Need of higher Velocities to overcome concentration polariza-
tion and fouling effects; Higher velocities cause large
pressure drops\ thus dScre'asihg the driving force across the
membrane.
In order to overcome this limitation, a computer program was written
in Fortran IV language2. Such a program is designed to calculate the
number of the following important factors for concentrating a certain
Volume of feed with a given feed concentration and a given percentage
recovery ratio of water in more than one concentrating stage:
Number of modules and their arrangement in each concentrating
stage.
Feed recycle ratio, inlet and outlet velocities, and percent
recovery ratios of water in each stage.
Flux rate and concentration of the concentrate in each stage.
Capital and operating costs.
The program is very flexible and has been written in such a way that
the effect of any of the variables can be studied without changing
its internal format. The program takes less than 2 ininutes for one
263
-------
complete calculation of the optimum number of modules, along with the
capital and operating costs of an RO unit.
Flow Chart and FortranListing of a Computer Program
The flow chart of the computer program is given in Fig. 60 for better
understanding of the sequence of operations involved in these optimiza-
tion studies. In this flow chart, a trapezoid symbol27 indicates an
input or output operation, whereas a rectangular box indicates any
processing operation except a decision. The "decision" is denoted
by a diamond symbol. READ stands for the input data of the program,
whereas WRITE gives the final output results of the program. The sym-
bolic notation of various parameters of the input and output data and
other important variables is also given along the flow chart.
There are three important decision checks made by the computer program.
First, it checks the velocity at each stage of concentration and, if
necessary, chooses a proper recycle ratio to maintain that velocity.
Second, it checks the final effluent concentration of the last stage
against the ultimate goal concentration and then changes the number
of parallel rows in one or all the stages, depending upon the deviation
between these two concentrations. Last, it compares each time new
values of capital and operating costs- with the previous corresponding
costs, and then, if necessary, changes the number of modules in series
in each parallel row until both capital and operating costs are minimum-
SYMBOLIC NOTATIONS OF VARIOUS INPUT DATA
USED IN THE COMPUTER PROGRAM
DGD = Design capacity of RO unit, gal./day
CIN = Initial concentration of the' feed, g/1
GOUT - Final concentration of the concentrate, g/1
D = Tube inside diameter, inch
SA = Module surface area, sq. ft
AP = Reference inlet pressure, psig
TEMP = Reference feed temperature, °C
FLNA = Reference NaCl solution (5000 ppm) flux rate at 600 psig
and 35°C, gfd
PINA = Reference osmotic pressure of NaCl solution, psia
AFL = Approximate average liquor flux rate at "AP" pressure
and "TEMP" temperature, gfd
VC1, VC2 = Velocity (ft/sec) — concentration (g/l) relationship
OPC1, OPC2 = Osmotic pressure (psia) —concentration (g/l) relationship
RRC1, RRC2 = Rejection ratios (percent) — concentration (g/l)
relationship
PDC1, PDC2 = Frictional pressure drop (psi) — concentration (g/l)
relationship
PDV1, PDV2 = Frictional pressure drop (psi) — velocity (ft/sec)
relationship
CRCAM = Membrane compaction rate
PINF = Percent increase in flux rate per degree centigrade
261;
-------
READ DGD, CIN, COOT, D, SA,
AP, TEMP, FLKA, PINA, AFL, VCl,
VC2, OPC1, OPC2, RRCl, RRC2,
PDC1, PDC2, PDV1, PDV2, CRCAM,
PINF, NB, RR(I), EFFE, ARGT,
NDAY, NYR, OPEN, CRGT, CMP1,
1CMP2, CMM1, &M2, CMN1, CMN2,
\REPC, REXP, CPEC, CEXP, QCOSl,,
'QCOS2, QSPRE, QMOD, CMODL
Calculate approximate num-
ber of modules in each stage
based on recovery ratios of
water
some minimum
number of modules in
series, say one, NM -
{same for all banks
Save the values of flow
rate and concentration of
the concentrate of Bank,
(NB-1)
I - NB
Save the previous value
of operating and capital
cost, TCI, TC2 by giving
another variable name
ZKC1, ZKC2_
£
Take flow rate and concen-
tration of the feed to Bank 1
same as those given In the
above READ statement, DGD, GIN
Compute"osmotic pressure,
PIF, and % rejection ratio,
TP, based on the concentra-
tion of the feed to tne next
bank or the concentration of
the concentrate from the pre-
vious bank
Yes
No
Calculate the number of
parallel rows, NR and the
inlet flow rate per paral-
lel row of modules, VIN
Compute the minimum velocity,
VMN, using velocity-concen-
tration relationship as given]
in the input data J
Figure 60. Flow Chart of Reverse Osmosis Optimization Studies
265
-------
No
[calculate recycle ratio,
IRYCL, and take the pressure
"Idrop, PD, zero
Take recycle ratio, RCYL,
equal to zero and the pres-
sure drop, PP, same as that
of previous bank
Take concentration, CM,
osmotic pressure, PIM,
and 1. rejection ratio,
TP, same as calculated
above
NI • 1
Modules
In series
NI (DO
.Loop),
NI> 1
Compute osmotic pressure, PIM,
and I rejection ratio, TP, of
the feed to the next module In
series based on the concentra-
tion of the concentrate from
the previous module In series
Compute total frlctlonal pres-
sure drop, PDi permeation re-
sistance, PR, flux rate. Ft,
concentration of the concentrate,
CM, and outlet velocity, VOUT
Compute nev minimum velocity,
VMIK, based on the concentra-
tion of the concentrate using
velocity-concentration rela-
tionship as given In the Input
data
Give a small Increment,
0.02 to the previous
recycle ratio, RCYL, and
take the pressure drop,
PP. sero
—&
Compute concentrate concentra-
tion, CCONC, concentrate flow
rate, FOOT, % recovery ratio
of water, RRW, and flux rate,
PU
Figure 60 (Cont'd). Flow*Chart of Reverse Osmosis
Optimization Studies
266
-------
Compute che absolute and
arithmetic difference, ACT,
Ctf, between the final con-
centrate concentration of
lait bank, CCONC, and the
final concentration to be
concentrated by RO, GOUT
Increaae or decrease the
number of parallel rove
of modules by on* If ACT
Z5
In the laat bank only de-
pending on whether CFP la
poiltiv* or negative
r~
5 < ACT < 50/i
\
1
1
CP\ ACF >/50
/
ACF£ 5
Increase or decrease the
number of parallel rows
of modules by one if ACF
<100, or by two If ACF>100
in all the banks depending
on whether CFP is positive
or negative
\ WRITE (a) Stage Number I, (b) No.
I of total modules, NUT, (e) No. of
\ parallel rova. NR, (d) No. of mod-
\ ules in series, MM, (e) Recycle ratio
\ RCYL, (f) inlet velocity, VIN, (g)
1 Outlet velocity. VOUT. (h) Pressure
\ drop, PDF, (1) Concentrate flow rate,
\ root, (j) Concentrate concentration,
\ CCONC, module
maintenance coat, CUCRR, (g) non-
membrane equipment depreciation
coat, CDCO
I
Figure 60 (Cont'd). Flow Chart of Reverse Osmosis
Optimization Studies
267
-------
Compute capital cost in dollars -
(a) total module cost, CTM, (b)
manifolding cost, CMAN, (c) cost
of main pressurizing reciprocat-
ing pump, CCRE, (d) cost of centri-
fugal booster pumps, CCBFP
Compute total operating cost in
cents per 1000 gallons of product
water, TCI, and total capital cost
in dollars, TC2
I
\ WRITE CTPP, CTGP, CMCP, CHRP, CDCO,
\TC1, CTM, CMkN, CCRE, CCBPP, TC2
Increase the number of
modules in series by one
in all the banks without
making any change in the
total number of modules
in any of the bank
Increase the number of
modules in series by one
in all the banks without
making any change in the
total number of modules
in. any of the banks
Figure 60 (Cont'd). Flow Chart of Reverse Osmosis
Optimization Studies
268
-------
NB = Number of banks or stages
RR(-I) = Approximate recovery ratios of product water in each stage
EFFE = Combined pimp-motor efficiency
ARGT = Grams of neutralizing reagent per 1000 gallons of feed
solution -
NYRRR = Module depreciation, years
NYR = Non-membrane equipment life, years
CPEN = Cost of pumping energy, cents/Kwh
CRGT = Cost of chemical reagent, cents/It
CMPl, CMP2 = Manpower cost ($)-unit size (gpd) relationship
CMML, CMM2 = Maintenance and materials cost ($)-unit size (gpd)
relationship
CMN1, CMN2 = Manifolding cost ($) —number of manifolds relationship
REPC, REXP = Main pressurizing reciprocating pump capital cost ($) —
capacity head (gmp x psi) relationship
CPEC, CEXP = Centrifugal "booster pump capital cost ($) — capacity head
gmp x psi) relationship
QCQSl, QCOS2 = Parameters of process instrumentation cost as a function
of RO unit size
QSPRE = Spare number of modules expressed as percentage of
minimum number of modules
QMOD ss Module cost, $ per sq ft of membrane
CMODL = Module maintenance cost, $ per module per year
Fortran listing of the program is also provided. There are about 300
Fortran statements describing the various operations and process calcu-
lations . Various comment statements, as indicated by a prefix "C,"
were written at various points of the computer program describing the
input and output data and other important process calculations. The
FORMAT statements of the output and input data were set up in such
a way as to print the important input variables and the output results
in a tabular form.
Sensitivity Analysis for the Mathematical Model
•ftie design and use of reverse osmosis as a method for concentrating
Pulping process effluents involves a complex, multistage, continuous
system. The logical development of an accurate description of such
a system, thus, becomes similarly quite complex. On dilute wastes
°f the pulp and paper industry, this complexity also arises from the
complex character of these dilute wastes. Most of these liquors, such
Q.S NSSC liquor, have significant amounts of colloidal and of fine par-
'ticulate suspended solids. These colloidal and suspended solids have
a tendency to "foul" the membrane surfaces, thus resulting in poorer
long-term flux rate and rejection characteristics of the membrane.
T° our knowledge, fouling has not been defined or explained mathemati-
°ally. In order to overcome this limitation, we made systematic experi-
^ntal studies for the determination of the required degrees of turbu-
lence and mixing necessary to minimize these fouling effects.
269
-------
FORTRAN PROGRAM LISTING
/JOB GO
/FTC LIST
BPS FORTRAN 0
C
c
S.0001
c
c
c
S.0002
S.0003
S.OOOA
S.0005
S.0006
S.0007
S.OOOB
S.0009
S.0010
S.0011
S.0012
S.0013
S.0014
S.0015
$.0016
S.0017
S.0018
S.0019
S.0020
S.0021
S.0022
C
C
C
C
C
C
C
C
C
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
COMPILER
CASE STUDY- OPTIMIZATION CRITERION OF LARGE SCALE
REVERSE OSMOSIS UNIT
DIMENSION NMT(IOO)tNR(lOO)»NMIIOO)•RRt100JtRCYLC100),VIN<100),
lVOUmOO),PDF(100).FOUT(lOO).CCONCUOO),OPtlOO).NMNI100),
2FLF(100),RRW(100»
GIVEN DESIGN CAPACITY OF REVERSE OSMOSIS UNIT.
INITIAL CONCENTRATION OF THE FEED, AND THE FINAL
CONCENTRATION TO BE CONCENTRATED BY REVERSE OSMOSIS
READ (5,1) DGO.CIN.COUT
GIVEN TUBE INSIDE DIAMETER AND MODULE SURFACE AREA
READ (5,2) 0,SA
GIVEN REFERENCE PRESSURE AND REFERENCE TEMPERATURE
READ IS,2) AP.TEMP
GIVEN REFERENCE FLUX RATE AND REFERENCE OSMOTIC PRESSURE OF
SODIUM CHLORIDE SOLUTION
READ (5,2) FLNA.PINA
GIVEN APPROXIMATE AVERAGE LIQUOR FLUX RATE AT
REFERENCE PRESSURE AND REFERENCE TEMPERATURE
READ IS,7) AFL
GIVEN EMPIRICAL CONSTANTS OF THE VELOCITY - CONCENTRATION EQUATION
READ (5.3) VC1,VC2
GIVEN EMPIRICAL CONSTANTS OF THE OSMOTIC PRESSURE - CONCENTRATION
EQUATION
READ (5,3) OPC1,OPC2
GIVEN EMPIRICAL CONSTANTS OF THE PERCENTAGE REJECTION RATIO -
CONCENTRATION EQUATION
READ (5,3) RRCt,RRC2
GIVEN EMPIRICAL CONSTANTS OF THE PRESSURE DROP - VELOCITY
EQUATION AS A FUNCTION OF CONCENTRATION
READ (5,4) PDC1,POC2,POV1,POV2
GIVEN LIMIT ON MAXIMUNM FRICTIONAL PRESSURE DROP
READ (5,7) PDMX
GIVEN MEMBRANE COMPACTION RATE EXPRESSED IN TERMS OF THE EFFECT
OF PRESSURE ON MEMBRANE CONSTANT
READ 15,1112) CRCAM
GIVEN PERCENTAGE INCREASE IN FLUX RATE PER DEGREE CENT.
READ (5,7) PINF
GIVEN NUMBER OF BANKS OR STAGES
READ (5,8) NB
GIVEN RECOVERY RATIOS OF MATER IN EACH STAGE
READ (5,1099) 4RRII),1-1,NB)
GIVEN INITIAL AND MAXIMUM NUMBER OF MODULES IN SERIES
READ 15,51 NZ1N.NMSM
GIVEN COMBINED PUMP / MOTOR EFFICIENCY
READ (5*7) EFFE
GIVEN PUMPING ENERGY FOR AUXILIARIES EXPRESSED AS PERCENTAGE
OF TOTAL PUMPING ENERGY
READ (5,7) PPNA
GIVEN PH OF THE FEED FOR REVERSE OSMOSIS PROCESSING AND THE
AMOUNT OF REAGENT USED IN CMS. PER 1000. GALLONS
READ (5,2) PHRO.ARGT
GIVEN MODULE DEPRECIATION IN NUMBER OF YEARS
READ (5,9) NYRRR
GIVEN NON-MEMBRANE EQUIPMENT LIFE IN NUMBER OF YEARS
READ (5,8) NYR
GIVEN NUMBER OF PARALLEL ROMS FROM EACH MANIFOLD
READ (5,8) NPRM
270
-------
C GIVEN COST OF PUMPING ENERGY IN CENTS PER KWHR.
S.0023 READ (5,7) CPEN
C GIVEN COST OF REAGENT IN CENTS PER IB.
S.0024 READ (5,71 CRGT
C GIVEN EMPIRICAL CONSTANTS OF THE COST OF MANPOWER IN
C $ / YEAR - UNIT SIZE IN GPO EQUATION
S.0025 READ 15,21 CMPUCMP2
C GIVEN EMPIRICAL CONSTANTS OF THE COST OF MAINTENANCE AND
C MATERIAL IN * / YEAR - UNIT SHE IN GPO EQUATION
S.0026 READ (5,2) CMMl,CMM2
C QIVEN EMPIRICAL CONSTANTS OF THE COST OF MANIFOLDING
C IN * / YEAR - NUMBER OF PAW FOLDS
S.0027 rtEAD (5,2) CMNUCMN2
C GIVEN EMPIRICAL CONSTANTS OF THE COST IN * - CAPACITY HEAD IN
C GPMXPSI EQUATION, ADJUSTING FACTOR, OPERATING LIMIT FACTOR,
C YEAR INDEX, AND NORMAL MODULE FACTOR FOR MAIN PRESSURIZING
C RECIPROCATING PUMP
S.0028 READ 15,101 RiPC,REXP,RADJ,ROLF,RYIX,RNMF
C GIVEN EMPIRICAL CONSTANTS OF THE COST IN $ - CAPACITY HEAD IN
C GPMXPSI EQUATION, ADJUSTING FACTOR, OPERATING LIMIT FACTOR,
C VEAR INDEX, AND NORMAL MODULE FACTOR FOR CENTRIFUGAL
C BOOSTER PUMP
S.0029 READ 15,10} CiPC,CEXP,CADJ,COLF,CYIX,CNMF
C GIVEN PARAMETERS OF PROCESS INSTRUMENTATION COST AS A
C FUNCTION OF R.O. UNIT SIZE
S.0030 READ (5,2) QCOS1.QCOS2
C GIVEN SPARE MODULES EXPRESSED AS PERCENTAGE OF
C MINIMUM NUMBER OF MODULES
S.0031 READ 15,7) QSPRG
C GIVEN MODULE COST PER SO. FT. OF MEMBRANE AND
C MODULE MAINTENANCE COST
5.0032 READ 15,21 QMOD,CHOOL
5.0033 1 FORMAT (3F1Q.2)
S.0Q34 2 FORMAT f2FI0.21
S.003S 3 FORMAT (2E16.7)
S.OOI6 4 FORMAT (4E16.TI
S.003T 5 FORMAT (2131
S.0038 7 FORMAT CIF10.2)
S.0039 8 FORMAT (113)
S.004Q 9 FORMAT (115)
S.0041 tO FORMAT (6F10.2)
5,0042 1099 FORMAT I4F10.2J
S.0043 1112 FORMAT (IE 16.7)
$.0044 CMOD*(QHOD)*(SA1
S.004I NDAY»(330.0Q)*(NYRRR1
S.0046 SUMM3-0.0
S.0047 NBM-NB-1
$.0048 RAP-600.00
S.0049 RTEMP-35.00
S.0050 PTF-ITEMP-RTEMP)*(PINF/100.00>
S.0051 G«(0.4087)/(D)**2
S.OOS2 TMR1-(1.0-CIN/COUT)
S.0053 TWR-DGD/TWRl
$.0054 F-TUR/1440.0
S.0055 AMCT-IFLNAI/IRAP-PINA)
S.0056 AMCTN.(AHCT)-(AP-RAP)*ICRCAM)
S.0057 FM»(AMCTNI*IRAPI
$.0058 WRITE (6,722) DGD.TWR.CIN,COUT,PHRO,AP,TEMP,FLNA,PINA
$.0059 722 FORMAT ( ' 1' ,20X, «OPTI MUATI ON CRITERION OF LARGE SCALE*/
130X,'REVERSE OSMOSIS UNIT'/////
271
-------
24X,'HAVENS IB TUBE TUBULAR MODULES*t
310X,-N.S.S.C. WHITE MATER'///
44X,'DESIGN CAPACITV OF REVERSE OSMOSIS UNIT*' f
56X,F10.1.2X,'GAL./DAY OF PRODUCT WATER'/55X,»OR'/
643X,'='.6X,FIC.1,2X,'GAL./DAY CF LIQUOR FEED RATE*//
74X,'INITIAL CONCENTRATION OF THE FEED-',12X,F10.I,2X,•GM./L.'
84X,'FINAL CONCENTRATION CF THE CONCENTRATE*',
S7X.F10.1,2X,'GM./L.•//
T4X,'PH OF THE FEED =•,31X.F10.I//
94X, 'REFERENCE PRESSURE*' ,27X,FIO. 1 ,2X,« PS IG1//
X4X,'REFERENCE TEMPERATURE*',24X,F10.I,2X,•CENT. •//
Y4X,'REFERENCE 5000 PPM SODIUM CHLORIDE SOLUTION'/
Z4X,'FLUX RATE AT 600. PSI6 AND 35. CENT.«•,9X,F10.1,
D2X,*GAL./DAY/SO.FT.•//
E4X,'REFERENCE OSMOTIC PRESSURE OF'/
F4X,'SODIUM CHLORIDE SOLUTION"•.21X.FIO.1,2X,«PSIA«)
S.0060 WRITE (6.9722) QMOD,CMODL,CPEN
S.0061 9722 FORMAT (MX,'MODULE COST-' ,34X ,F10. 1 ,2X,
1'DOLLARS PER SO. FT. OF MEMBRANE'//
24X,'MODULE MAINTENANCE COST*•,22X,F10.1,2X,
3'DOLLARS PER MODULE PER YEAR'//
44X,'COST OF ELECTRIC POWER"',23X,F10.1,2Xt'CENTS PER KWHR.'J
S.0062 IF (ARGT-1.00) 9888,9688,6988
S.0063 6988 WRITE (6,7777) CRGT
S.0064 7777 FORMAT I MX,'COST OF CHEMICAL REAGENT*',21X,F10.1,
A2X,'CENTS PER LB. •)
S.0065 9888 WRITE (6,9777) NYRRR.NYR
S.0066 9777 FORMAT I MX,'MODULE DEPRECIATIGN««,2*X,110,2X,•YEARS'//
7AX,'NON-MEMBRANE EOUIPMENT LIFE-',16X,110,2X,'YEARS')
S.0067 DO 55 1*1,NB
S.0068 WR'RR(I)*DGD/100.0
S.0069 CMR^WR/AFL
S.0070 55 NMT(IXMR/SA
S.0071 SUM»0.00
S.0072 NMI*NZIN
S.0073 65 DO 50 1*1,NB
S.0074 IF (1-1) 21,21,22
S.0075 21 FF*F
S.0076 CF=CIN
S.0077 PIF*(OPCI)*(OPC2)*(CF)
S.0078 CS*(RRC1)+(RRC2)*(CF)
S.0079 PD-0.0
S.0080 GO TO 23
S.0081 22 NBB-I-1
S.0082 . FF«FOUT(NBB)
S.0083 PIF*OPINBB)
S.0084 956 CS»(RRCl»-t-(RRC2)*(CF»
S.0085 23 IF U.O-SUM) 871,172,172
S.OOB6 871 NI'NZXT
S.0087 SUM3*0.0
S.0088 GO TO 471
S.0089 172 NI-NHI
S.0090 GO TO 173
S.0091 171 SUH3«CM
S.0092 471 NI*NI+1
S.0093 173 NRIt)*NMT(I)/NI
S.0094 NMdl'NI
$.0095 TFR»FF/NR(I)
S.0096 GFR'TFR
S.0097 VR*G*TFR
272
-------
$.0098
$.0099
5.0100
S.0101
S.0102
S. 01(13
S.0104
S.0105
S.0106
$.0107
S.OlOfi
S.0109
S.0110
S.OUl
S.0112
S.0113
S.0114
S.0115
S.0116
S.0117
S.OU8
S.011
TFR=GFR
VR=G*FR
VINU)=VR
C CALCULATE PERMEATION RESISTANCE. FLUX RATE,
C CONCENTRATION OF CONCENTRATE IN EACH SERIES
00 90 If~l,NI
IF III-U 4
41 TFM«FR
CH«CF
AND FINAL
MODULE
TP»-tCS)/tOO.OC
GO TO 43
42 PtM*JOPClI»-«OPC2}*«CH>
Tl>— ICSI/100.0
TFM-TFM-P
43 V»G*TFH
IP (VI 361,361,362
362 AI*IPOC15*|PDC2>*»CM>
B1*IPOV1)»(POV2)*(CNI
TTI«V**B1
PDi«TTI»*l
PO-PD+PD1
PR.PIM*PD
TLL-IAP-PR1/RAP
TL«IFHJ*ITLL5
P»(FL*SAI/1440.0
TFR-TFR-P
FMM-TFRfP
Hl-TFR/FHH
If (Nil 361,361,382
382 H2»Hl**TP
90 CM»H2*CM
PDFI I)=PD
682
782
VOUTII)-IB1*ITFM-PI
IF (VOUTIII-VKIN) 681,661,682
If (1.0-SUHM3I 981,981,181
IF (CM-SUH3) 181,782*782
IF IPDMX-PDFtl » 181,181,171
181 NNII1-NI
NR(II«NHT«I>/NH«10I»CU*IOPC2»*
-------
S.0156
S.015?
S.0158
S.0159
S.0160
&.0161
S.0162
S.0163
S.0164
S.0165
S.0166
S.0167
S.0168
S.0169
S.0170
S.0171
S.0172
S.0173
S.0174
S.0175
S.0176
S.0177
S.0176
S.OIT9
S.0180
S.0181
S.0182
S.0183
5.018*.
S.0185
S.0186
S.0187
S.0188
S.0189
$.0190
$.0191
S.0192
S.0193
$-0194
S.0195
S.0196
5.0197
5.0198
$.0199
S.0200
$.0201
S.0202
S.0203
TLL=IAf»*PRI/RAI»
TL=FW*TLL
FLFtI»-
-------
37X, 'OUT'^X.'PRE-SSX.'FLOWSSXt'CONC. •t4Xf•OSM.',5X,•FLUX•,
A5X,«REC.«)
S.0204 WRITE (6,7261
S.0205 726 FORMAT («BER•,6X,'MOOU-',4X.'ALLEL*,4Xt*IES't6X,«RATIO*t4X,
BMFT./1,
54X, •( FT./',4X,»SSURE',4X,'RATE SSXtMGM./'^X.'PRE-SSX,'RATE* ,
B5X,'RATIO*)
S.0206 WRITE (6,727)
S.0207 727 FORMAT ( IOX,«LES«,6X. 'ROWS1,5X,«MODU-',13X,«SEC.) • ,4X,»SEC.) • ,
74X, 'DROP'.SX.MGPMJ'.AX.'L. >• ,6X, • SSURE',<»X, • «GPD/' ,<.X,'OF1 I
S.0208 WRITE (6.728)
S.0209 728 FORMAT (28X,«LES«,33Xf'(PSI)•,22X,«PSIA',5X,'SOFT)•.
C4X,'WATER1)
$.0210 00 655 1*1,N8
S.0211 655 WRITE (6,724) I ,NMT (I) ,NR( I» ,NH( t) ,RCYL( 1) ,VIN{ I) , VOUTU ),
IPDFm.FOUTII) .CCONCH) ,OPII) ,FLF(I),RRW(I)
S.0212 724 FORMAT (//I 6.3I9.9F9.2)
S.0213 I»l
S.02U NSTT=NM(I)
S.0215 NZ-0
S.0216 IMTT*0
S.0217 00 4442 1=1,NB
S.0218 GOBX=RCYL
-------
5.02*9 CHC1-IOGDD)**(CMH2)
S.0250 CMC*(CMCli*fCMMll
S.0251 CHCP=-(CMC/<330.0*OGDDI)*(UOE*5)
C CALCULATE MODULE REPLACEMENT COST
S.0252 DG0P»DGOD*NOAY
S.0253 CMRE=NHTT*CMOD
S.025*1
5*0259 CMAN«CCMNU*
S.026I CCRE2*CCREl*REPC
$.0262 CCRE3*CCRE2*RADJ*ROLF*RYIX
S.0263 CCRE*CCRE2*CNHF+CCRE3
C CALCULATE CAPITAL COST OF CENTRIFUGAL BOOSTER PUMPS
5.0264 CCBPP=0.00
S.0265 DO 9163 J-1,NBM
S.0266 CCBPX-FOUT(J)*RCVL(JI*PDF(JJ
S.0267 CCBP2*CCBPl**CEXP
S.0268 CCBP3»CCBP2*CEPC
$.0269 CCBP4«CCBP3*RADJ*ROLF*RYIX
S.0270 9163 CCBP-CCiP3*CNMF+CCiP4
$.0271 CCBPP«CC»PP»CCBI»
C CALCULATE PROCESS INSTRUMENTATION COST
S.0272 SURPC-TWR/10CO.OO
S.02T3 IF (SURPC-500.00) 666,667,667
S.0274 666 GLRPC*QCOS1
S.027S GO TO 670
S.0276 667 GLRPC*OCOS2
C CALCULATE DEPRECIATION COST
S.0277 670 TODC-CMAN+CCRE+CCBPP+GLRPC
S.0278 OCOS1-(TODC/DGOD)*«1.0E*5)
S.0279 CDCO-JDCOS11/(NYR*330.00)
C CALCULATE TOTAL CAPITAL AND OPERATING COST
S*0290 TCl«CTPP*CT6P*COPP*CMCP+CMRP+CDCO*CLKRR
$.0201 TC2«CTM*CHAN*CCRE*CCBPP*GLRPC
S.0282 NRITE <«,1777)
S.0283 1777 FORMAT l////18X*«CAPITAL COST IN DOLLARS'f34X,
1-OPERATING COST IN CENTS PER'/
275X,«IOOO.GALS. OF PRODUCT MATER*>
S.0284 WRITE (6,3888> CTM.CTPP.CHAN.CTGP
S.0285 3888 FORMAT (///4X,'TOTAL MODULES COST'*,22X(F14.i,
15X,'PUMPING ENERGY COST-*,24X,F10. 1Y/
34X,'MANIFOLDING COST-',24X,F14.1i
3SX. 'NEUTRALI ZATION COST- • ,24X(RO. 1)
$.0286 WRITE 16,9666) COPPtCCRE,CCBPP,CHCP,GLRPC,CMRP(CLKRRtCDCO
$.02*7 9666 FORMAT I/4X**MAIN PRESSURIZING RECIPROCATING*,
129X,«MANPOWER COST-* ,30X ,FIO.I/
24X,>PUMP COST«»«3IX»F14.1//
34X* (CENTRIFUGAL BOOSTER PUMPS COST-* ,10X,FM. 1, 5X,
4'MATERIAL AND MAINTENANCE COST"'tL4X,Fl0.1//
Z4X,'PROCESS INSTRUMENTATION COST*'»12X(F14.lt
5SX.'MODULE REPLACEMENT COST-',20X,F10.I//
664Xt'MODULE MAINTENANCE COST-'»20X,FIO.l//
276
-------
764X,'NON-MEMBRANE EQUIPMENT DEPRECIATION COST-*3X»FIO.I)
S.0288 WRITE <6»9555» TC2.TCI
S.02B9 9555 FORMAT C ^/MXt *TOTAL CAPITAL COST**,22X,F14.1t
15X,»TOTAt QPEIUTIN& COST-'»23X,Fi0.1l
S.029Q NDIFaNNSM-NSTT
S.029I IF (NOIF) 102,102,6569
S.0292 6569 SUH-2.00
S.0293 NZXT-NSTT
S.0294 GO TO 65
S.0295 102 STOP
S.0296 END
SIZE OF COMMON 00000 PROGRAM 14528
END OF COMPILATION MAIN
For properly identifying the most critical parts of a mathematical
model, we studied the effect of changes in the many parameters used
in the model, and thus determined the relative importance of each con-
tributing factor. The results obtained from these preliminary studies
indicated the model is most sensitive to:
1. Changes in the characteristic flux rates of reference
solutions of NaCl or other standard solutions at
standardized reference pressures and temperatures.
2. Osnotic pressure of the liquors.
3. Frictional pressure drop-velocity relationship.
k. Velocity-concentration relationship.
5, Percentage increase in flux rate per °C.
Sensitivity analysis also showed that the rejection ratio-concentration
Jlationship and membrane compaction rate are of secondary importance.
le depreciation, module cost per sq ft of membrane, nodule main-
tenance cost, and non-membrane equipment life are the most important
factors in determining the capital and operating costs of a reverse
osmosis unit, whereas the cost of pumping energy and neutralization
are of relatively less importance. A separate analysis is made to
study the effect of a number of stages, percentage recovery ratios
of water in each stage, and the number of modules in series on the
capital and operating costs, and is discussed in the latter part of
the report.
EXPERIMENTAL STUDIES ON THE CONCENTRATING RUNS OF
CALCIUM-BASE ACID SULFITE AND NSSC LIQUORS
Here we used our large-scale reverse osmosis trailer unit rated at
10,000 to 100,000 gallons daily for concentrating calcium-base acid
sulfite and NSSC liquors by k to 15 times the original concentration
°f the feed liquor. Thepje Studies could be readily carried out with
^a little as 500 gallon|f»|3f refeflUMquors by recycle of the liquor
277
-------
product fractions back to the feed tank. A schematic diagram of the
experimental set-up is shown in Fig. 6l. The feed liquor flows by
gravity from a 5000-gallon stainless steel tank to a 500-gallon plastic
tank from which the feed is pumped at about 20 psig to a Manton-Gaulin
main pump. The main pump discharges the feed to Stage I. The concen-
trate from Stage I is fed directly to Stage III in a tvo-stage concen-
trating run. For a three-stage concentrating system, the concentrate
from Stage I goes to Stage II first, and then to Stage III. The booster
Pumps A and B are used to overcome the pressure drop and to provide
a suitable recycle ratio, thus maintaining identical conditions of
inlet pressure and inlet velocity in all the stages. The final concen-
trate from the last stage is returned through a heat exchanger for
heating or cooling to maintain carefully controlled feed temperatures,
and then recycles to the 500-gallon feed tank. The flow rate of permeate
was measured from each stage. The total permeate of all the stages
is collected and mixed in a small tank, and then drained to a 50-gallon
plastic tank underneath the trailer. During the concentration run,
the permeate from the 50-gallon tank flows by gravity to a 300-gallon
plastic tank, and then is discarded to the sewer. Under constant con-
centrating conditions, the permeate mixes with the concentrate via
a small centrifugal pump, and then recombined concentrate and permeate
is returned through a heat exchanger.
A number of continuous concentrating runs were made with calcium-base
acid sulfite and NSSC liquors. The results of these studies are dis-
cussed in detail as follows:
Concentrating Run of Calcium-Base Acid Sulfite Liquor
Here the concentrating run of calcium-base liquor was made in two
I and III, using Type 310 R-Series Havens 18-tutoe nfcdules. These
have been in use intermittently for 1-1/2 years. Only the booster
Pump B was used for maintaining the inlet pressure and inlet velocity
in Stage III the same as those of Stage I. Three concentrating runs
of calcium-base liquor were made at an average velocity of h.2 feet
per second using three different modules configurations as follows:
(a) Twenty modules — 5 parallel rows of 2 modules in series
in Stage I and 5 parallel rows of 2 modules- in series
in Stage III.
(b) Twenty modules — k parallel rows of 2 modules in series
in Stage I and k parallel rows of 3 modules in series
in Stage III.
(c) Twenty-one modules — H parallel rows of 3 modules in
series in Stage I and 3 parallel rows of 3 modules in
series in Stage III.
278
-------
ro
}g
n
Stags
HI
Booster
pt
wr
IE
Back Pr»s«ur«
Regulator
ir
l2" S.S. RLg^
2-Psa
Exch
at
sif
ttL
Cooling
Water
tin)
Js
a" a.s. KP*
woo.
S-1
o
§
a
•t
5
5000 Gallon
S.S. Feed
Tank
F - Feed
E - Concentrate
t * Pferauto of
OZM
-Total
of «H th«
Stages
ye«dPuap
500 Gallon
•tic f
Tank.
Figure g!. Scheoatlo Diagraa of Experimental Setup
-------
Control flux rates on standard sodium chloride solution were measured
at k.2 feet per second "before each of the concentration runs. Each
flux run was made at 500 psig and 33-35°C. Before the start of each
liquor run, the modules were -washed with high-velocity water'and with
detergent BIZ solution of 15 g/1 concentration. Here we concentrated
calcium-base liquor only 3 times"in the solids concentrations range
of 10 to h2 grains solids per liter. ¥e found we could not concentrate
reliably to 10 percent solids because of greater probability of tube
failure with 360 tubes in 20 of the elderly modules during unattended
overnight operations.
Table 82 gives the flux rates and rejection ratios of calcium-base
liquor for each of these three different modules arrangements. It
is noted from Table 82 that the rejections of solids, optical density,
COD, and calcium are all above 98 percent, whereas BODg rejections
vary between 93 and 97 percent. And the flux rates are relatively
very high, on the order of 10 to 13 gfd at 500 psig and 33-35°C for
10 to if2 g/1 concentrations of calcium-base liquor. These high liquor
flux rates are probably due to the fact that the membrane becomes "open
after the detergent BIZ wash-up, and then calcium-base liquor forms
a dynamic membrane by which the rejections are improved significantly
at very little expense of flux water.
Concentrating Run of NSSO Liquor
Here dilute samples of NSSC liquors were concentrated using a recently
purchased bank of the 310 Ps-Series l8-tube Calgon-Havens modules manu-
factured and delivered during the first 6 months of 1971* Various
concentrating runs of ISSC liquors were made at an average velocity
of U.2 feet per second using two different modules configurations as
shown below:
Configuration of twenty-seven modules — k parallel rows
of 3 modules in series in Stage I and 3 parallel rows of
3 modules in series in Stage II and 2 parallel rows of
3 modules in series in Stage III.
Configuration of twenty-four modules — 5 parallel rows
of 2 modules in series in Stage I and h parallel rows
of 2 modules in Stage II and 3 parallel rows of 2 modules
in series in Stage III.
The booster Pump A and B were used prior to Stages II and III for
taining identical conditions of inlet pressure and inlet velocity in
all the three stages. For each module configurations, two eoncentrati&S
runs were made at two different inlet pressures of 600 and 800. psig-
The sodium chloride flux rates were measured at U.2 feet per second
before each concentrating run. Each flux rate run was made under con-
trolled conditions of inlet'pressures and at 33-35°C. Before the star
of each liquor run, all the' modules were cleaned either with high-
velocity water or with a detergent solution containing 15 grams BIZ
280
-------
TABLE 82
FLUX RATES AND REJECTION RATIOS DURING THE CONCENTRATION RUN OP
CALCIUM-BASE ACID SULFITE LIQUOR
{Type 310 R-Seriea Havens 18-Tube Old Modules)
Maximum No. of Modules » 21
No. of Stages ** 2
Inlet Pressure = 500 psig
Average Feed Liquor Temperature • 33-35°C
Average Flow Rate = 2.6 g.p.m. <* k,2 ft/sec
Average 5000 mg/1 NaCl Flux Rate at 500
pslg and 35°C » 16.0 gfd
Rejection Ratios of NaCl " 1»0-70 percent
Concentration Rejection Ratios, percent
of Liquor, Flux Rate, OD at
e/1 gi'd Solids 281 no COD BOD5 Calcium
(a) 5 Parallel Rows of 2 Modules in Series in Stage 1+5
Parallel Rows of 2 Modules in Series in .Stage III (Total
Modules = 20)
II*. 7
1*2.0
10.3
30.0
12.8
35.8
(b)
(c)
12.9
10.8
U Parallel Rovs of
Parallel Rovs of 3
Modules » 20)
12.3
11.0
k Parallel Rovs of
Parallel Rows of 3
Modules = 21)
11. U
10.0
99.1 99.6 99.1 97.7 99-5
98.9 99. ** 98.8 97.3 99.5
2 Modules in Series in Stage I + k
Modules in Series in Stage III (Total
99. b 99.9 99.0 95.2 99.6
99-8 99.6 98.5 93.8 99.2
3 Modules in Series in Stage 1+3
Modules in 'Series in Stage III (Total
99.2 99.7 98.8 97.1 99.7
98.2 99.0 97.7 96.1 98.7
281
-------
per liter. The modules during the first configuration were washed
with water alone, and no detergent was used during this configuration.
However, the modules during the second configuration were cleaned exten-
sively with water and with the BIZ detergent solution. In these tests to
verify the math model pulsing could not be used, yet we were able to
concentrate NSSC liquors by 10 to 15 times without pressure pulsing
or any kind of shutdown during any of these concentrating runs. High
velocities maintained flux rates at satisfactory levels even at high
solids concentrations.
Tables 83 and &k give the flux rates and rejection ratios of NSSC liquors
for two different module configurations. Figures 62A and 62B plot
the corresponding flux rate vs hour of operation for these two arrange-
ments of modules. It is noted from Table 83 that the liquor flux rate
decreased from 8.3 gfd at 8.6 g/1 to 3.3 gfd at 78 g/1 for 600 psig
inlet pressure, whereas for TOO psig, the flux rate decreased from
9.9 gfd at 6.7 g/1 to 2.8 gfd at 90 g/1 solids concentrations. The
liquor flux as well as NaCl flux rates given in Table Qk were measured
during the second module configuration and they both are relatively
higher than those of Table 83. This was because the modules during
the second configuration were given an extensive cleaning treatment
with detergent BIZ. The data in Table 81* show that the flux rate de-
creased from 8.8 gfd at 9.8 g/1 to 2.9 gfd at 89 g/1 for 600 psig pres-
sure, whereas for 700 psig, the flux rate decreased from 11.2 gfd at
10 g/1 to 2.6 gfd at 122 g/1 solids concentrations. The rejection
ratios of solids, optical density, and COD were all above 95 percent,
whereas sodium and BODs varied between 90-95 percent, except for two
permeate samples in which BODs rejections were only 86-88 percent.
These low BODs rejections were probably due to anaerobic fermentation
reactions, since we observed gas bubbles being evolved and the odor
of hydrogen sulfide from these particular permeate samples.
It is to be noted that the NaCl solution flux rates for the new Ps-
Series Calgon-Havens modules were very low compared to the NaCl solution
flux rates of R-Series Havens old modules. These new modules had not
undergone hydrolysis reactions of the cellulose acetate membrane which
could increase their flux rates. Their rejections of NaCl were rela-
tively much higher than those of E-Series old modules. The difference
between the NaCl flux rates at 600 and 700 psig pressure was small and.
probably resulted from the membrane becoming more compacted at 700
psig. Consequently the flux rate decreased at the expense of rejections-
The rejections of NaCl were found to be about 7-10 percent higher at
700 psig than the corresponding rejections at 600. psig.
Model Verification
The objective of this study was to verify the results of the mathematics1
model against the experimental results. One way has been to compare
the flux rates of the mathematical model and experimental studies under
identical conditions of module configuration, reference NaCl solution
flux rate, pressure, temperature, and velocity.
282
-------
TABLE 83
FLUX RATE AND REJECTION RATIOS DURING CONCENTRATION RUN OF NSSC LIQUOR
(I Module Configuration)
(Type 310 PS-Series Havens l8-Tube New Modules)
Total No. of Modules * 27
No. of Stages = 3
Set up = *» Parallel Rovs of 3 Modules in
Series in Stage 1+3 Parallel Rovs of
3 Modules in Series in Stage II + 2
Parallel Rows of 3 Modules in Series
in Stage III
Average Feed Liquor Temperature •» 33-35°c
Average Velocity = 2.5 6Pm = ^.2 ft/sec
Average 5000 mg/1 NaC* Flux Rate at 600
psig and 35°C » 8.0 gfd
Average 5000 mg/1 NaCl Flux Rate at 700
psig and 35°C » 8.5 g^
Rejection Ratios of NaCl at 600 psig = 70 percent
Rejection Ratios of NaCl at 700 psig - 80 percent
Flux Rate,
gfd
Concentration
of Liquor,
8/1
(a) Inlet Pressure » 600
8.6 8.3
30.3 6.5
5U.7 U.8
78.0 3.3
(b) Inlet Pressure » 700 psig
6.7 9.9
19.5 7.0
30.h 6.0
57.9 !».!»
89.7 2.8
Percent Rejection Ratios
Solids
and NaCl
98
97
96
95
.2
.1
.8
.3
and NaCl
97
97
97
96
95
.6
.5
.U
.9
.9
OD
281
at
nm
Flux Rate
99.
99.
99.
99.
U
5
6
7
Flux Rate
99.
99.
99.
99.
99.
k
5
5
6
7
COD
« 8.
98
97
97
96
- 8.
97
97
97
97
96
0
.3
.6
.5
.0
5
.6
.3
.2
.3
.8
BOD 5
grd
97
95
9U
93
gfd
97
91*
86
95
91
.5
.2
.5
.1
.8
.0
.6
.0
.7
Sodium
97.
95.
9U.
92.
96.
96.
96.
95.
93.
0
5
7
7
3
5
0
3
8
283
-------
TABLE 84
FLUX RATE AND REJECTION RATIOS DURING CONCENTRATION RUN OF NSSC LIQUOR
(II Module Configuration)
(Type 310 PS-Series Havens 18-Tube New Modules)
Total No. of Modules - 24
No. of Stages - 3
Set up = 5 Parallel Rows of 2 Modules in
Series in Stage 1+4 Parallel Rows of
2 Modules in Series in Stage II + 3
Parallel Rows of 2 Modules in Series
in Stage III
Average Feed Liquor Temperature - 33-35°C
Average Velocity "2.5 gpm • 4.2 ft./sec
Average 5000 mg/1 NaCl Flux Rate at 600
psig and 35°C - 8.5 gfd
Average 5000 p.p.m. NaCl Flux Rate at 700
psig and 35°C = 1 0 gfd
Rejection Ratio of NaCl at 600 psig = ?2 percent psig
Rejection Ratio of NaCl at 700 psig = 79 percent psig
Concentration Rejection Ratios, percent
: Liquor,
8/1
(a) Inlet
9.8
24.9
34.9
52.7
69.9
89.4
(b) Inlet
10.2
27.3
34.4
47.8
75.7
121.9
Flux Rate,
gfd
Pretsure - 600
8.8
7.1
6.6
5.1
4.0
2.9
Pressure -700
11.2
9.0
8.5
7.3
5.0
2.6
Solids
psig and
97.4
97.3
97.4
97.1
97.0
96.2
psig and
97.6
97.9
97.9
97.8
97.3
95.1
OD at
281 ran
NaCl Flux
99.5
99.7
99.7
99.7
99.8
99.8
NaCl Flux
99.4
99.6
99.6
99.7
99.7
99.8
COD
Rate -
97.5
98.0
97.4
97.2
97.2
96.9
Rate -
98.1
98.1
98.3
98.2
96.0
96.7
BOD5
8.5 6»
87.8
95.9
94.2
93.9
93.6
96.5
10.0 gfd
95.8
95.3
98.7
95.5
97.3
93,4
Sodium
96.9
97.1
97.2
96.5
95.7
94.7
95.0
95.5
96.4
96.1
95.0
92.6
28U
-------
10.0 r-
9.0
8.0
7.0
» 6.0
t*
I
5.0
1..0
3.0
2.0
T.Og/1
Inlet Pressure • 700 psig
Inlet Preaaurs » 600 psig
' ' 1 L
28.0 g/1
30.0 g/1
J L
10,0 20.0 30.0 5*0.0 50.0
Hours of Operation
60.0
70.0
80.0
Figure 62A. Flux Rates During Concentration Runs of WSSC Liquor
(I Module Configuration)
P85
-------
10 r/1
ic
10 c/1
B-
s
w
Inlet Pressure «• 600 psig
Inlet Preesuro • 700 psig
I
J_
10
30 >»0 50
Hours of Operation
60
70
80
Figure 62B. Flux Rates During Concentrating Run of KSSC
Liquor (II Module Configuration)
286
-------
The mathematical model was run on the IBM 360, Model kh Computer. Table
85 gives a typical computer printout of the mathematical model.results
for one NSSC liquor concentrating run. It is to be noted here that
the mathematical model results were based on a straight-through concen-
trating run in which the feed was concentrated to the required solids
concentration in one single, throughput using a large number of modules.
For experimental studies, we used a relatively small number of modules
and made a concentrating run under recycling conditions. It should
also be noted that the capital and operating costs of a reverse osmosis
unit, as given in the computer printouts, are very approximate, and
their accuracy depends strongly upon the accuracy of the input data
used in the model.
Tables 86, 87, and 88 compare the experimental and mathematical model
flux rates for calcium-4>ase acid sulfite and NSSC liquors. The results
of the comparative study show that the agreement between the experimental
and mathematical model results was fairly good and within 10 percent.
There is more than 10 percent deviation between the experimental and
mathematical model flux rates for the concentrating run of NSSC liquors
at 700 psig. The mathematical model flux rates are higher than the
experimental flux rates . This was probably due to the fact that the
reference NaCl solution flux rate, as used in the model, was not the
same as was obtained experimentally. The model was very sensitive
to the initial NaCl flux rate, as has been shown in the sensitivity
analysis of the model.
EVALUATION OF IMPORTANT PARAMETERS FOR THE DESIGN OF
MANIFOLDING AND PIPING SYSTEM OF A LARGE-SCALE
REVERSE OSMOSIS UNIT
This section of the report deals with the study and evaluation of those
parameters which do not depend directly upon the operating conditions,
characteristics of the dilute waste streams to be processed, and the
performance of the reverse osmosis membranes. These parameters are
important in the design of a manifolding and piping system for the
best arrangement of modules, thus minimizing both the capital and the
operating costs of a reverse osmosis unit. In this study, three impor-
tant parameters were considered, as follows:
Number of stages
Percentage recovery ratios of water in each stage
Number of modules in series in each parallel row
A number of computer runs were made to determine the effect of the
above parameters on the total .number of modules, capital and operating
costs of a reverse osmosis unit. Table 89 gives the effect of the
first two parameters on the total number of Calgon-Havens modules,
capital and operating costs of a RO system for calcium-base acid sulfite
liquor, along with the important input data used, in the computer runs.
It is noted from Table 89 that there is no change in the total number
of modules for various combinations of recovery ratios in a 3- and
287
-------
TMLE 85
ro
CD
Co
OPTIMIZATION CRITERION OF LARGE SCALE REVERSE OSMOSIS UNU
(Havens 18 Tub* tubular Modules)
NSSC White Water
Design Capacity at Reverie Osmosis Unit - 45,000.0 gal./day of product water
" 48.632,5 gal,/day or liquor feed rate
Initial Concentration of the Feed -6.7 g/1
Final Concentration of the Concentrate « 89.7 g/1
pH of the Feed - 7,8
Referenc
Referene
Referene
flux Rat
Heferene
Module c
Pressure - 700.0 psig
Teaperatur* m 35°C
NaCl Solution " 500O ppm
at 600 pslg and 35°C - 8,5 (fd
Osmotic Pressure of SaCl Solution * 65.0 psia
S9.3 per sq ft o£ membrane
Module Maintenance Coat • #24 pel module per year
Colt of Electric Power - l.Zc per Kwh
Module Depreciation • 5 year*
ne Equipnieat Life w S years
Stage.
No.
1
2
3
Total
Modules
171
120
21
Ho. of
Parallel
Rows
57
40
7
Series
Modules
3
3
3
Recycle
Ratio
3.43
6.02
4.15
Vel. In,
ft/stc
4.30
4.47
4.53
Vel. Out,
ft/sec
3.77
3.9*
4.17
total
Pressure
Drop,
psi
101.96
113.41
131. «
Effluent
Flow
Kate,
gpm
15.58
3,77
2.25
Cone,,
g/l
14.19
56.26
92.67
Qsmot ic
Pressure,
psia
42.58
168. 79
278.01
Flux Kate,
gfd
8.13
6.11
4.2S
0/0
St.ec.
Ratio
of
Wittr
53 87
34.98
4 48
Minimum Number of Modules - 312
Spare Number of Modules - 31
f ot*l Kunber of Module* - 343
Overall Average Flux Kate » 8.4 gfd
Capital Coat in Dollars
total modules cost 55,185.2
Manifolding cost 3,500.0
Main pressurizing reciprocating pump coat 10,700.8
Centrifugal booster pumps cost 9,143.0
Process Instrumentation .cost 25,000.0
Total C4pit*l cost
101.528.9
Operating Cost in Cents per
1000 Gal. of Product Miter
Pimping energy cost 10.2
Neutralization cost 3^0
Manpower cost 15.4
Material and Maintenance 5.5
Module replacement cost 73.7
Module »»tnttn«iie* cost 55.0
Hosmembmne equipment, depreciation cost 44,5
Total operating cost 224.2
-------
k-stage.concentrating system. Also, the' total number of modules do
not change significantly and they Indicate a maximum increase'.of 16
when the number of concentrating stages are increased from 3 to k,
For both 3- and l*-stage concentrating systems, the' number of total
modules shows a relatively larger, increase of 69-85 -(from 305. to 37^-
390) with an increase in the number of modules from 2 in series to
5 in series.
TABLE 86
COMPARISON BITW1EI EXPERIMENTAL AHD COMPUTERIZED
MATHEMATICAL MODEL FLUX RATES — CALCIUM-BASE
ACID LIQUOR
Io. of Stages = 2
Computerized Mathematical
. Experimental Results Model Results
Concentration of Flux Rate, Concentration of Flux Rate,
Liquor, g/1 gfd Liquor, g/1 gfd
Two Modules in Series in Each Parallel How
1^.7 12.9 22.0 11.3
te.O 10.8 1*0.1 10.3
Three Modules in Series in Each Parallel Row
12.8 ll.lt 16,9 10.5
35.8 10.0 33.9 9-5
The results of the above studies seem to indicate that the series' modules
setup is a more important factor than the number of stages and in what
proportions the water is removed in each stage in determining the total
modules, and hence, the capital and operating coats of a reverse osmosis
concentrating system. It is also noted here that we need about 305
modules (setup of 2 modules in series In each stage), rather than 395
nodules with which the trailer was originally equipped, for concentrating
55,000. gallons per day of calcium-base liquor from 1.0 to 10 percent
solids concentrations at 600 psig and 35°C. This run was based upon
nodules hairing a flux rate of 13.5 gfd on & control teat with laCl
at 600 psig and 35°C. The bank of 305 modules was shown, in the sensi-
tivity analysis to be highly sensitive to reference NaCl flux rate.
A separate computer run was made to determine the' effect. of number
of modules in series on the capital and operating costs of a reverse
osmosis unit processing 750,000 gallons per day of NSSC liquor. The
results are given' in Table 90 in terms of individual items of important
289
-------
equipment, capital cost, and operating charges, along with the more
significant input data as used in the computer run. Figure 63 plots
these capital and operating costs vs^ number of modules in series. These
cost data are interesting "but they derive from a model which was developed
primarily for providing comparative data "but not necessarily true total
installed cost for a 750,000. gallons per day reverse osmosis plant.
The accuracy.strongly depends upon the accuracy and completeness of
the input data used in the model. Of course the relative variations
of these costs with number of modules in series are quite accurate.
It is noted from Fig. 63 that "both the capital and the operating costs
tend to become minimum for a setup of two modules in series. Therefore,
two modules in series can probably be considered best for optimizing
the design of a large system as based on present cost studies.
TABLE .8?
COMPARISON BETWEEN EXPERIMENTAL AND COMPUTERIZED
MATHEMATICAL MODEL FLUX RATES - NSSC LIQUOR
(I Module Configuration)
No. of Stages = 3
No. of Modules in Series in Each Parallel Row = 3
Computerized Mathematical
Experimental Results Model Results
Concentration of Flux Rate, Concentration of Flux Rate,
Liquor, g/1 gfd Liquor, g/1 gfd
Inlet Pressure = 600 psig
8.6 8.3 1^.7 6.8
30.3 6.5 31.9 6.0
5U.7 U.8 82.3 3-5
78.0 3.3
Inlet Pressure - 700 psig
6.7 9-9 13.2 8.2
19.5 7.0 lH.8 6.8
30.tj- 6.0 92.1 U.3
57.9 U.U
89.7 2.8
290
-------
TABLE 88
COMPARISON AMD COMPUTERIZED
MATHEMATICAL MODEL FLUX RATES - LIQUOR
(II MODULI CONFIGURATION)
No. of Stages « 3
No. of Modules In Series
in Each Parallel Row <* 2
Experiment: a 1 Results
Computerized Mathematical
Model Results
Concentration
of Liquor, g/1
9.8
34.9
69.9
89.4
10.2
27.3
47.8
121.9
Flux Rate Concentration
gfd of Liquort jj/1
Inlet Pressure * 600 t>sic
8.8 17.2
6.6 38.3
4.0 95.9
2.9
Inlet Pressure « 700 pslg"
11.2 18.0
9.0 41.5
7.3 128.0
2.6
Flux Rate
gfd.
7.7
6.6
3.6
9.1
7.9
3.5
291
-------
TABLE 89
EFFECT OF NUMBER OF STAGES AND RECOVERY RATIOS
OF VftTER ON THE CAPITAL AND OPERATING COSTS
(Calgon-Haveni 18-Tube Tubular Module!)
Feed Liquor - Calcium-Ban Acid Sulflte Liquor
Deilgn Capacity of Revern Oimoili Unit • 50,000 gal./day of Product Water
or
» 55,000 gal,/day of Liquor Feed Rate
Initial Concentration of the Feed * 10 g/l
Final Concentration of the Concentrate • 100 g/l
Inlet Prciaur* - 600 pllg
Reference NaCl - (5000 ppm)
Flux Rat* at 600 p>lg and 35°C • 13.5 gfd
Module Coit * $9.3 per iq ft of membrane
Module Maintenance Colt « $24 per module per year
Coit of Electric Power • 1.2e per Kehr
Module Depreciation " 5 yeara
Non-membrane Equipment Life • 5 yean
Percent ' Total Operating
No. of Modulea Recovery Ratloa Minimum Overall Average Coit, cent! per
Total No. In Serle* In of Water In No. of Module! Total No. Flux Rate Total Capital 1000 galloni of
of Stagei Each Stag* Each Stag* in Each Stage of Moduli!* gfd Coit, dollara product vater
3 2 22.6 66 305 10.5 98,112 197.5
22.3 66
43.7 146
2 22.6 66 305 10.5 99,363 199.1
44.2 132
23.8 80
2 45.1 132 305 10.5 97.556 196.9
19.8 60
25.8 86
3 21.7 66 9.9 101,060 202.7
21.5 66
47.3 162
3 21.7 66 323 9.9 101,642 204.2
42.A 132
26.5 96
3 43.2 132 323 9.9 99.592 201.8
19.0 60
28.4 102
4 22.7 72 343 9.3 103,504 210.1
22.3 72
45.3 168
5 20.2 70 374 8.6 108,597 220.0
20,7 70
49.8 200
4 2 31.3 92 305 10.5 97,212 196.6
12.8 38
17.2 52
29.0 96
3 29.6 90 330 9.7 100,334 204.1
9.5 36
16.3 51
35.0 123
4 31.5 100 356 9.0 104,747 212.9
9.8 44
18.2 60
31.1 120
3 31.5 105 390 8.2 110,617 224.6
7.0 45
18.8 65
33.4 140
Includes 10 percent apar* nodulea over minimum number required for effective proceaalng of feed loading.
292
-------
TA5LK 90
ro
\o
U)
inncx OF KMES OF MODULES in SERIES on TUB CAPITAL AMD
THE OPERATING COSTS OF REVERSE OSMOSIS (MIT
(C*lgoQ*Havena 18-Tube Tubular Hodulea)
Feed Liquor - NSSC White Vatec
Deaiga Capacity of Reverse Oaeioala Halt * 675,000 gel. /day ef Product
or
- 750.0OO gal. May af Liquor Feed Kate
Initial Concentration of the Feed • LO g/1
Fiftal Concentration of the Concentrate * 100 g/1
Inlit Preaaura - 600 P*lg
Feed Liquor Temperature " 35°C
Reference Sodiua Chloridt (500O ppn) and Hue Bate *e 600 p«lg end 35°C - IJ.5
Kodule Co»t - $9,3 p«r »q ft of Mr»br*n«
Kodule Halntcniace C««t - $14 per Module per Year
Coat of Electric Power * 1.2c per Kali
Nodule Deprtclaeloti * S Y<-ar«
Koo-«e«Bbrane IqnipBciit Life - S teara
Huaber of ConcentrMlDg Stage* - 3
Capital Coat, dollare
Operating Co«t inCeat« per LOOP Gallons of Frorfutt Utter
Ho. of
Hodale*
In Serie*
I
2
1
4
5
6
Intel
Me. of
Matale**
335T
358*
3818
'; 4OT+
4372
4745
OnzaU
Awrege
Flux Hate
8«d
12. »
w.a
11. S
10. «
9,8
8.1
Moduli-
Coat
540.107
576.951
614,27?
555,465
703.410
763,421
Me In
M*2l- Preaiurictng Booncer
folding Reciprocating PUBJ>
Cost Pu«p Coat Coet
101,800 40.96T
54,400
- 38, MO
30,»OQ
26.600
24,000
32,703
37,421
41,388
«4.SZ1
47,674
50,241
tTOCttl
InatruiBes-
tatioo Total Manpower
Cat la Co.to Cott
50,000 76^.517 3.7
759. 73»
785,232
621,853
868,651
928,630
Materiel & 8eutr*ll- Pumping
Maintenance tat Ion Energy
Coat* Colt Coat
2.4 0.0 *.0
».o
9.0
9.0 .
9.0
9.0
Replace -
n«nt
Coat
43.2
5».7
55 0
58.7
63.0
68.4
Hei»e«braoe
Module Equtyveat
Maintenance Depreciation Total
Coat Coat Coat
16. 0
38.5
41.0
43. 6
47.0
il.O
20.1
16.21
15.3
14.9
14.8
1.4.8
119.4
121.7
1Z6.5
132.6
139.9
149.3
Include* Id p€*cest space itofele* over
itonber required for effective ^roceealag of f«e4 lo«41ftg«
-------
1V1
I
4'
1139
-------
CONCLUSIONS
The mathematical model developed in this study proves to be most sensi-
tive to the reference NaCl flux rate, the osmotic pressures of the
liquors .being processed, module membrane life, and module and membrane
replacement costs.
There is a fairly good agreement between the liquor flux rates of the
Mathematical model and that obtained from the experimental studies.
The agreement vas within 10 percent.
The number of modules in series appears to be a more important considera-
tion than the number of stages and the recovery ratios of water in
each stage in the design of the manifolding and piping system. Probably
two modules, 18 tube (288 linear ft) in series is the best number based
on the present capital and operating costs of an RO concentrating system.
295
-------
SECTION IX
MEMBRANE MODULE LIFE STUDIES
The economics of reverse osmosis processing are critically dependent
upon the life expectancy for the membrane system., as well as upon the
membrane performance in terms of flux rates and solute rejections.
In order to develop data for the economic study to be discussed in
Section X, detailed records were maintained throughout this project
for the purpose of establishing the life history of the membrane equip-
ment used. Additional experimental programs specifically designed
to develop membrane life experience were conducted where possible within
the 3-1/2 year program. Limitations on the number of life study runs
which could be conducted on a sustained basis prevailed in terms of
availability of reliable, continuously operated, high pressure, pumping
equipment and also of manpower for detailing such studies. Nevertheless,
substantial experience was gained toward this objective.
Most suppliers have indicated a 2-year life to be a major goal of mem-
brane research and development programs and of membrane manufacturing
quality control in the 1966-1971 period of development. Unless sub-
stantial values can be recovered, the economics of waste processing
by RO appear questionable if membrane module replacement is required
More frequently than once each two years. Laboratory studies and some
isolated instances of field experience may lend confidence to projec-
tions for life expectancies at that level and may even extend to three
years or more. However, actual experience in this research and demon-
stration project and also that of analogous field trials by others,
of which the authors are aware, indicate the average life expectancy
in terms of a statistically significant number of modular units of
membrane equipment manufactured in 1970 or before has yet to be ade-
quately proven out at the one-year level. The time lag in developing
well-controlled and reliable data is an obvious obstacle to drawing
conclusions as to the life expectancy of newly designed or improved
equipment recently introduced to the market.
The 387 modules with which the large trailer unit was originally
equipped were of 1967 design and were manufactured early in 1968. Two
hundred sixty of these were rebuilt at the factory a year later in
September 1969 after somewhat less than 2000 hours of actual service
under full pressure. Important improvements have since been made,
and several new types or modifications have issued from this manufac-
turer in 1970 and 1971. Similar model changes and improvements are
apparent for equipment being offered by all suppliers active in the
field throughout the period of conducting this research and demonstra-
tion project.
The laboratory and small pilot scale life expectancy runs were planned
and conducted to supplement the experience on the large trailer-mounted
field unit. Specific problem areas affecting life performance of mem-
brane modules exposed to service conditions in concentration processing
297
-------
of pulp and paper industry effluents could best be evaluated in this
manner. Those conditions for waste treatment proved to be a severe
test and the results were substantially less*than required for meeting
minimum standards of economic feasibility in large-scale commercial
operations.
Data being accumulated in studies continuing after terminating this
specific Research and Demonstration Project which is subject for this
report may be expected to show continuing progress being achieved toward
extending membrane and module life beyond the minimum 12-month goal
and on toward the more favorable 2- and 3-year life levels.
Equipment
Most of the various laboratory and pilot test stands, and also the
large trailer-mounted field unit as described in Sections V, VI, and
VII have at times been employed individually and collectively in these
membrane module life testing studies. Reliability of high pressure
pumps was a first problem. It took some patience and much "doing"
to maintain the several rotary piston pumps available for early studies
in around-the-clock, 7-day week service, but the pilot plant crew,
with the help of the laboratory staff and group leaders, maintained
a night and day "on call" program with remarkable loyalty and persever-
ance. Later, the larger Milton Roy variable stroke, positive displace-
ment , piston pumps became available especially for sustained and con-
tinuous research service and less out of hours attention was required.
At least one such life study has been under way at all times within
this project in the ^-year period, 1966 through 1971 > and as many as
four units have been operating simultaneously at times on the various
•types of equipment supplied for testing by cooperating manufacturers.
Much of the exploratory testing involved a range of the types and kinds
of membrane modules becoming available and was made possible by loan
of equipment without charge by a number of manufacturers developing
modified designs for waste treatment as well as for salt water conver-
sion service. Such equipment is described by "type but the cooperating
suppliers are not specifically identified. Cooperating concerns are
listed in the appendix.
Methods
The pattern for life testing utilized pumping test stands described
in detail in prior sections of this report. Each unit was complete
with automated instrumentation and controls employing a system of relay8
and timers to maintain system temperatures, pH and pressure levels.. ' >
These sensors were integrated with automated shutoffs to the pump power
supply in case of failures. The timers operated pulsing systems (tem-
porary programmed release of pressure for short periods of 30 seconds
to several minutes every hour or so) to control fouling and also were
used at times for control of periodic changes in feed flows and the
like.
298
-------
The feed systems usually could be maintained on a continuous, straight-
through basis in field tests, at the mills where adequate supplies of
fresh liquor were available continuously or where feed tanks could
"be filled periodically during the day. The Appleton laboratory and
pilot plants have tanks available for up to 10,000 gallons feed storage.
This capacity was used for several test runs of limited duration, uti-
lizing truck load quantities "but usually the life testing on the
Institute campus was carried out with recycle of 50 to 500-gallon quanti-
ties of test solution made up in batches and replaced at suitable inter-
vals. Both permeate and concentrate product flows were returned and
mixed to provide a continuous supply of feed liquor.
Most of the principal pulping and bleaching effluents meriting specific
studies were subjected to testing in the life test studies. But the
standard test solution for routine studies was made up at a solids
concentration of 1 percent by diluting 50 percent Ca-base spent sulfite
liquor concentrate (commercially available as "lignin liquor" under
various trade names). This "standard" feed liquor, approximating a
principal type of pulp wash water, has been the principal comparative
test solution used for these life test evaluation studies. Other
specific pulp wash waters and bleach liquors have been subject for
extended studies when the logistics of supply permitted.
Analytical control for the life tests necessarily had to be based on •
the minimum expenditure, of man hours for measuring, sampling, and con-
ducting .control assays necessary in maintaining a continuous record
on performance of membrane units under life te.sts. For most of the
runs reported, feed flows were routinely metered by use of the positive
displacement piston pumps,-with adjustable stroke length. Six individual
Milton Roy pumps, three, of which were ..purchased especially for this
project, provided infinitely variable flows with reproducible stroke
length adjustment, and were selected for this purpose in flow capacities
up to 3 to 5 gallons per minute each. Permeate flows were taken period-
ically during each day from several individual modules, and wherever
possible from banks of two or more modules to obtain statistically
sound, averaged data. Temperature and pH could be recorded automati-
cally at times, and where necessary, but availability of such facilities
was necessarily limited. More usually the temperature was controlled
within desired limits without recording, and the pH in feed tanks was
manually adjusted at suitable periods as needed. Pressure control
was usually maintained with manually adjusted back pressure valves
(Cash Acme K-20 valves on lab and small pilot units). This was not
an ideal system of pressure regulation for unattended, around-the-
clock operations, but worked with fair reliability at reasonable cost
if the valve-seats were maintained in good condition and free of
plugging by suspended solids.
life testing for the 38f modules oa the large trailer-mounted unit
was a special objective during each of the first three field trials.
Later this large unit was moved to Appleton for special studies which
helped to advance the program for evaluating life expectancy." A card
299
-------
file with records of performance for individual modules was maintained
in as much detail as could be accomplished with available manpower.
Life 'Test 'Data for various Membrane "Mo&nL_e Systems
Capillary 'Fiber
Several capillary fiber test modules were submitted for research evalua-
tion, and the data and results were interesting in the limited areas
of life study actually conducted on these units. However, no extended
life testing could "be achieved with the early types of test units avail-
able. Susceptibility to plugging by suspended solids contained in
the feed liquor, by precipitates and crystalline deposits which devel-
oped during concentration, or by sliming as a result of microbiological
cell growth, comprised problems which were not satisfactorily solved
in these early tests. Sustained life studies could not be gotten under
way. Several major chemical concerns are understood to be actively
advancing the technology for use of the hollow fiber systems since
the time these initial tests were conducted. Although plugging was
a serious problem, no failures of the membrane structural support system
occurred and no leakages of connections were experienced.
Sheet_Membrane Pack Systems — Spiral Wound
Problems with plugging by suspended solids contained in the pulp mill
effluent feed liquors or precipitates and the like which formed during
operation were also apparent with the sheet membrane pack, systems,
such as the spiral wound module as discussed in Section V. Oils
adversely affected the development of sustained operation, and therefore^
made life studies dependent upon selecting feed liquors free of suspends"
solids or use of test solutions which could feasibly be clarified by
such methods as filtration at the 3 to 10 nm level. With these pre-
requisites taken care of, it was possible to conduct limited, small-
scale life study tests on spiral wound modules.
Three modifications of spiral wound assemblies were tested:
1, A Schedule 80 PYC tube containing two spiral wound
modules with a total membrane area of 6.8 square feet
and a maximum operating pressure of UOO psig.
2, Two 2-1/2 inch Schedule 80 stainless steel tubes containing
ten spiral wound modules, each tube with a membrane area
of 69 square feet and capable of operating at 600 psig.
3. Succeeding studies were conducted utilizing spiral wound
nodules of a mad! fled type with larger than standard
openings in the mesh separator to ndnimize plugging of
tie flow channel.
300
-------
The PVC tube "with two spiral units was intended for use in several
preliminary trials to evaluate plugging problems prior to exposing
the larger, more expensive stainless steel units to the process stream,
The PYC unit was installed on a test stand with a variable stroke pump
and auxiliary equipment to maintain an operating pressure of 350 psig
while processing a Ca-base sulfite pulping wash water that had been
defibered through a 100-mesh screen.
unit was operated five days a week for a total of eight weeks (6l8
hours), with weekly backwash to maintain the open flow channels. During
this period there was no indication of plugging. This was a substantial
improvement over earlier laboratory experience with standard modules
having less space between membrane flow channels.
After eight weeks of operation at 350 psig, the PVC unit was replaced.
The two larger stainless steel units were installed for operation at
600 psig. These were operated successfully for six weeks. At that
time the high pressure pump had to be diverted to other test programs,
so it was necessary to terminate studies on these spiral modules until
a pump could again be made available.
The data tabulated in Table 91 for these two tests show the flux rate
declined to some extent, but this appeared to be stabilizing at approxi-
mately 8.9 gfd during the last four weeks of operation. Rejections
averaging on the order of 96 percent solids, 99 percent color, 88 percent
biochemical oxygen demand (5 day), 95 percent for chemical oxygen demand,
and 99 percent for calcium were achieved while processing this waste
stream.
At the time of the unavoidable shutdown, unrelated to performance of
the spiral equipment , a note was made in the records that "these modules
should be put back into use as soon as practical, since the results
look too good to warrant premature termination of the study." Unfor-
tunately, due to pressure of other work, the processing of this stream
with spiral wound modules could not again be undertaken for several
months. When the modules were tried again (in late 1969), they were
found to be fouled and plugged by growth of microbiological slime during
storage , Incomplete washing and' improper storage conditions were
apparent, and replacement of the spiral modules was required.
The rebuilt modules were put back in service for a study of processing
an evaporator condensate, and performed satisfactorily for more than
2000 hours . The system for control of pH failed at 2131 hours , allowing
a rise to above 10, which damaged the membranes and ended the. tests
on these units .
A third system was provided by a different supplier of spiral wound
equipment. This consisted of a newly designed wide-channel spiral
^ound module, a multistage centrifugal pressure pump and a preset ori-
fice for k50 psig pressure operation. This uafet was used-, to concentrate
301
-------
TABLE 91
FLUX RATES AND PERMEATE QUALITY WITH SPIRAL WOUND MODULES ON LIFE STUDY
.easing an Acid Sulflte Wash Water
(2 Gallon/Minute Flow Rate)
Flux,
reek gfd at 25 C
1
2
3
1*
5
6
7
8
6
6
6
6
6
k
5
5
.2
.5
.8
.3
.2
.8
.1
.0
Feed
Solids,
6/1
Two Modules
13-7
10.0
14.0
12.9
12. k
15.0
12.7
12.7
Feed
PH
Rejection, percent
permeate Solids
at 350 psig
*.3
k.k
3.3
4.7
3.fc-
4.3
4.8
3.6
Twenty Modules at
9
10
11
12
13
Ik
12
10
9
10
9
8
•9
.7
.0
.U
.6
.2a
12.2
10,8
12.5
13-7
11.7
11.2
4.2
4.3
3.1
3.6
3.9
M
3
3
2
3
3
Operating
• 1
.6
.9
.7
A
3.8
k
2
.3
.8
Pressure
98
96
95
97
96
97
98
95
OD
98
99
99
99
99
99
99
99
BODc
90
92
90
91
90
89
87
89
COD
97
97
96
97
96
96
95
96
Calcium
98
99
99
99
99
99
99
99
600 psig Operating Pressure
3
4
2
3
3
4
.0
.0
.9
.2
.3
.2
97
97
95
96
96
98
99
99
99
99
99
99
87
86
88
88
86
84
95
95
95
96
96
9*
99
99
99
99
99
99
At 1.2 gpm flow rate.
NSSC "white water," The liquor was pretreated by passage through a
50 ym filter. The waste stream, containing a high content of suspended
solids in a near colloidal state, had been tried unsuccessfully in
several narrow-channel modules, The newest design, however, was
thought to be less apt to plug with particulate matter, and the pulse-
less operation of the centrifugal pumping system warranted investigation
in comparison with the pulsing normally used on this feed in tubular
systems.
A flux loss of U5 percent (from 12.7 to 7-0 gfd) was experienced in
the first three days of operation, followed by complete plugging of
the 50-nm feed filter. Subsequently, the spiral module plugged
302
-------
irreversibly after the filter was removed upon the suggestion of the
manufacturer.
The rejection of solids, BODs, COD, and color were on the order of
92-99 percent (Table 92). Hie flux rate deteriorated to less than
one gfd at 6 percent solids and could not "be restored. The unit was
then returned to the manufacturer.
The laboratory and small field-scale runs on spiral wound modules demon-
strated that:
1. Early designs of the spiral wound module tended to plug with
suspended particulate matter, by precipitates which formed during con-
centration of some pulping effluents, and also by microbiological sliming.
2. Tests on clear feed streams, such as evaporator condensates,
indicated the spiral units could be used satisfactorily for concentration
in the range of 1 percent to 10 percent solids, and with good (90 percent
or better) rejections if problems with formation of suspended solids
and microbiological slime could be avoided,
3. The newest "open path" spiral modules presented no mechanical
plugging problems during a 2131-hour concentrating study with an acid
sulfite evaporator condensate. However, as with all units equipped
with cellulose acetate membranes, they could not tolerate pH levels
much above 7«5» and failure occurred when the pH control allowed the
pH of the feed to rise to higher levels.
k. There was indication of microbiological fouling (slime forma-
tion) in the study with recycling feed, but properly controlled straight-
through feeding of fresh substrates with periodic cleanups at intervals
of about 1 week would be expected to minimize or eliminate this problem.
5. Although module failure by plugging and fouling was a serious
problem with the spiral-wound sheet membrane system, there was at no
time any evidence of structural failure in the membrane support system
nor were leakages experienced in the connections within or outside
of the well-designed module.
TUBUMR MODULES
The tubular modules evaluated in these laboratory studies were of three
basic classes:
1. Replaceable modules, where failure of individual membrane
or support tube components required the replacement of
the entire module as a unit.
2. Replaceable 'tubes within a module in which the individual
support tubes, complete with membranes could be replaced
if membrane or tube failures occur.
303
-------
TABLE 92
RATES AND PERMEATE QUALITY WITH SPIRAL WOUND MODULES OH LIFE STUDY
Processing NSSC White Water at lj-50 psig Operating Pressure
Hours
3
72
168
306
Flux,
gfd at 25 C
12.7
7.0
3.8
2.0
Feed
Solids,
g/1
Liquor Fed to
9-5
9.6
10.3
22,5
solids
Rejection
BOD5
, percent
COD
Optical
Density
Slowly Concentrate
J
99
99
99
92
93
93
95
96
96
97
96
99
99
99
99
326
tea
^39
Liquor Fed to Rapidly Concentrate
After Adjusting the Volume to 50
Ijiters ¥ith a 1 Percent Solids Liquor
2.3
1.6
2.2
1.1
o.i*
0.3
0.1
12.0
30.2
^0 ^
65.9
86.6
103.8
127. k
98
99
99
99
98
98
95
..
92
91
—
9k
92
• 89
—
97
98
98
98
98
96
..
98
99
99
99
99
99
3. Replaceable membranes, based upon membrane inserts
within individual tubes which could be replaced if
the rejection or flux characteristics became un-
acceptable.
These basie tubular designs could be further differentiated into porous
and non-porous support tube structure, and the porous support styles
into spiral wrapped, braided, or woven fiberglass and resin-bonded
sand core types.
Identification according to the number of tubes per module could also
be used as a means for classification. In order to simplify presenta-
tion of the data, discussion, and conclusions of the experimental work
in this section, the headings used are in terms of the number of tubes
per mnflule for ^ach design. Additional descriptive information has
SOU
-------
been included under each heading where required to classify the type
of module under study.
Each tubular system (spiral wrapped, braided, woven fiberglass and the
sand core) was evaluated, as in the case of the spiral wound module
study for the total operating hours during which acceptable rates and
quality of product water were maintained, as evidenced by the rejection
of solids, biochemical oxygen demand, chemical oxygen demand, and color
constituents from a recycled 0.5-1 percent solids pulp wash water re-
constituted from a 50 percent concentrate of calcium-base acid sulfite
pulp liquor. Special operating conditions of pressure, temperature,
fluid velocity, and other process modifications are also discussed under
the individual headings as they apply.
Seven-Tube Modules
Studies on the life of tubular units first began in December 1966 with
ten of the early prototype designs available at that time. These were
7-tube modules of the spiral wrapped design and had the tubes cast per-
manently in resin heads. They were evaluated on a test stand comprising
a rotary type, four piston pump, a 50-gallon plastic drum reservoir
for the recycling liquor, and a pressure regulating valve set at 600
psig.
As the study progressed it became apparent that some modifications to
the pumping equipment and controls would be necessary to provide accept-
able process control. These included:
1. The installation of a stainless steel cooling coil (sixth
day) to control rising temperatures due to conversion of
mechanical energy into heat in the recycle system.
2. A vertical one-inch diameter 6-foot long section of
stainless steel pipe as an improvised accumulator chamber
was installed immediately after the pump outlet (56th
day) to reduce the sharp pressure pulsations from a
range of about 125 psig per stroke to about 25 psig.
3. Still another change occurred at the IJ^th day with place-
ment of a forepump (Centrifugal, 18 psig maximum 20 gpm)
ahead of the main pressurizing pump in an attempt to
further reduce the "hammer" of sharp pressure fluctuations
from the reciprocating piston pump available for these
early stage evaluations.
During the 60-week test period the product water quality from intact
(leak-free) modules was excellent (Table 93). The flux rates varied
with the age of the module, the temperature of%operation, and the
condition of the membrane surface.
305
-------
TABLE 93
FLUX RATES AND REJECTIONS FOR SEVEN-TUBE MODULES ON LIFE STUDY
Processing 1 Percent Acid Sulfite Wash Water
Week
1
5
10
15
20
25
30
35
to
45
50
55
60
aA
Stream
Feed
Permeate
Feed
Permeate
Feed
Permeate
Feed
Permeate
Feed
Permeate
Feed
Permeate
Feed
Permeate
Feed
Permeate
Feed
Permeate
Feed
Permeate
Feed
Permeate
Feed
Permeate
Feed
Permeate
measure of the
Flux,
gfd at 25 C
13.8
11.7
7.8
7.8
8.1*
9.6
9.7
8.8
8.4
7.6
7.6
8.0
7.4
lignin content.
PH
5.6
5-3
4.9
4.9
5-2
4.8
4.9
4.6
5.1
4.0
5.0
4.2
5.8
4.0
4.8
4.2
4.8
4.3
4.6
4.5
4.8
4.4
5.4
4.4
5.2
5.0
Solids,
ms/l
5644
115
4668
108
6513
108
4829
257
5018
258
5055
195
5190
237
4960
200
5635
178
5430
164
5615
179
5332
211
5620
2to
BOD ,
mg/15
1608
845
1315
442
1275
427
1361
512
1237
650
1100
378
1211
347
1195
U25
1420
551
1530
142
1438
305
Itol
418
1282
359
Opticala
Density
at 281 ran
42.5
0.56
to. 6
0.26
57.8
0.27
39.7
0.33
42.6
0.39
36.0
0.22
44.9
0.32
38.0
0.33
38.0
o.to
42.0
0.43
41.3
0.42
35.0
0.35
4o.9
0.38
Color
CoPt
_._
700
5
450
5
too
5
650
5
700
5
530
5
500
5
480
5
380
5
321
5
306
-------
At the start of the trial the flux rates were on the order of 16* gallons
per square foot per day (gfd) at 600 psig. The flux rapidly decreased
to 11 gfd and then held at this level for 35 days*
After 35 days there was evidence that heavy sline formation in the feed
reservoir and. in the modules was affecting the performance of the system.
The liquor in the reservoir, which was. on a twice weekly change schedule,
was odoriferous within eight hours after makeup, and there were long
stringlike slime formations floating in the fluid. Periodically,, pads
of slime sloughing off the tubes would block the regulating valve and
the pressure would become erratic.
The flux rate decline, due to slime formations on the membranes, could
"be stopped and the rate restored to 8-9 gfd by ball-flushing the modules
periodically with a 3A-inch soft polyurethane ball (a procedure fully-
described in Section V). Later the addition of a biocide (an alkyl
dimethylbenzyl ammonium chloride product) to the feed stream helped
to reduce the rate of slime formation and the frequency of ball-flushing
cleaning.
This carefully controlled life test run started with ten modules and
was continued around-the-clock and seven days a week, with replacement
of modules as they failed over a period of lM months (^33 days 5•
Two of the original modules, Hos. 3-19 (6552 hours) and 3-27 (6680 hours)
continued in active service for nine months (Table 9*0- One ran to
failure without replacement, (also at 9 months) three were replaced
once, at shorter intervals and the others on several occasions. Eight
modules remained in operating condition at the end of the run, and at
that time had as a group demonstrated an average life of 4812 (200 days),
or nearly seven months with the oldest at 5808 hours (2k2 days) and
the newest at 3329 hours (139 days). Overall, the 25 modules entering
upon the entire 1^ months' test had averaged more than six months' oper-
ating life. Seven of these remaining 7-tube modules entered a subse-
quent comparative test with 18-tube modules, and their average life
was extended by another month (578 hours) (see Table 96). Two more
individual modules reached the 9-month life level, lo. 3-68 (668k hours)
and No. 3-70 (6215 hours).
Elsewhere, the Bureau of Reclamation, U.S. Department of Interior, was
reported to have had a similar bank of these same 7-tube modules in
continuous service for a period of 18 months in a study conducted for
the Office of Saline Water.
These results were concluded to be promising evidence that, with con-
tinuing improvement in manufacture and design, the life of tubular
modules could be expected to eventually extend beyond a minimum one-
year practical life and on toward a 2-year objective. The larger-scale
demonstrations were undertaken with tubular design equipment on this
premise and on the basis of comparative freedom from serious plugging
problems experienced with capillary fiber and sheet membrane designs.
307
-------
g
00
TABLK 94
MODULE REPLACEMENT SCHEDULE FOR THE SEVEN-TUBE LIFE STUDY
While Processing a 1 Percent Solids and Sulfite Wash Water
#3-19 #3-20
Out on Failed
3-2-67 6.17^67
for 1*128 hr
cleaning
1813 hr #3-60
#3-31 In on
6-17-67
In on 5753 hra
3-2-67
1/2 day
#3-19
In on
3-2-67
Failed
9-27-67
6552 hr
#3-73
In on
9-27-67n
3329 hr
in use at end of run.
#3-21
Failed
29^2 hr
#3-32
In on
5-8-67
Failed
5-20-67
288 hr
#3-31*
In on
5-20-67
Failed
12-5-67
2036 hr
#3-75
In on
12-5-67
3972 hr
#3-22
Out as
mate to
#3-21 on
U-19-67
291*2 hr
Failed
6-28-67
3720 hr
#3-75
In on
6-28-67
Out as
mate to
#3-19
2288 hr
#3-23 #3-2U
Failed Failed
6-22-67 12-12-66
k2bk hr 96 hr
#3-68 #3-18
In on In on
6-22-67 12-12-66
5637 hra
Out for
cleaning
3-2-67
damaged
1818 fhr
#3-30
ID on
3-2-67
Failed
8-20-67
38UO hr
#3-72
In on
8-21-67
U221 hr
#3-25
Failed
5-11-67
3311 hr
#U-17
In on
5-11-67
Out as
mate to
#3-26
5-11-67
#3-33
In on
5-15-67
Failed
6-5-67
1*18 hr
#3-1*3
la on
6-15-6?
Failed
6-30-67
311 hr
#3-69
In on
6-30-6T
5^52 hra
#3-26 #3-27
Out as Failed
mate to 10-2-67
#3-25 6680 hr
5-11-67
3311 hr
#3-26
In on
5-15-67
Failed
5-21-67
35l»0 hr
#5-10
In on
5-22-67
Replaced
no fail
6-5-67
#3-1*2
In on
6-15-67
5808 hra
#3-28
Failed
8-16-67
5558 hr
#3-71
In on
8-16-67
U323 hra
-------
A field trial was set up on a calcium-base acid sulfite pulp wash water
with 2k. of the 7-tube modules which continued for a 23-week period.
Two module failures were encountered during this period. Both modules
had been used extensively before this test and both failed due to tube
rupture after 1^50 and 1550 hours.
A problem of much concern in the use of later designs of tubular modules
with replaceable tubes was not apparent in these first life studies
with the T^tube cast-in-head modules. The rigid casting of the tubes
in resin heads was an effective method of eliminating problems of seal-
ing the connections for tubes in the end cap structures. Leakages in
the end caps were a comparatively minor problem of proper gasket instal-
lation and maintenance of external module manifolding connections.
However, the structural failure of a single tube in the 7-tube module
with a rigid cast-in-place head structure resulted in total failure
and forced;discarding of the entire 7-tube module. This manufacturer,
in common with other suppliers, was researching and designing new and
improved modules. Various systems were under development for replacing
individual tubes or relining rigid modular structures.
One of the tubular type equipment suppliers came forth at this time
(1967-1968) with new l8-tube modules having a compact head structure
based upon replaceable tubes for the combined purpose of:
1. Increasing the ratio of membrane surface area to a unit
of module volume.
2. Reducing the pressure drop in the tube connecting, turn-
around structure of the end cap and of the manifolding
connections between modules.
3. Reducing initial capital costs and the major operating
charges for module replacement an'd maintenance.
Similar design changes were apparent in new equipment becoming available
from other manufacturers.
18-Tube Modules
Upon completion of the lU-month life study with 7-tube modules, a second
trial was started in February, 1968, which was initially equipped for
comparative purposes, with a series of eight of the 7-tube modules re-
maining from the first study, in parallel with two banks of the new
Havens 18-tube modules each with 3 modules in series. This arrangement
provided a three bank system in which the 7-tube and the 18-tube modules
could be compared while processing the same dilute calcium-base feed
liquor under the same conditions of operation.
Pumping equipment was improved by converting from the rotary pumps (h
piston) to heavier reciprocating single piston pumps with variable
309
-------
stroke length more suitable for continuous research service. Standard
bladder-type accumulators were also employed from this point on to dampen
the pressure pulsations of the piston pumps at 600 psig.
The initial product water flux rates of the 18-tube modules were found
to be almost 50 percent higher than for the 7-tube modules (Table 95)•
Quality of the product water was comparably high for both types of equip-
ment. This indicated an improvement in the modules, and did in fact
indicate an improvement in membrane performance. However, there was
a serious increase in the number of module failures. Leakages in the
18-tube module end cap seal structures became a critical problem to
contend with.
With six l8-tube and eight 7-tube modules on line we experienced 20
and 6 failures, respectively, during a 5H-day test period.
This rapid module replacement resulted in widely scattered flux and
water quality data. Observation of poor rejection was usually soon
followed by module failure. The 7-tube modules failed due to tube
ruptures and most of the failures in the l8-tube units were due to leaks
in the seals between the individual tubes at the end cap of the module.
While these leaks could be repaired on site, it was not an economically
satisfactory situation in terms of time and expense for repair and re-
placement. All experiments with the 7-tube modules were terminated
at this time when manufacture ceased.
A second trial with new 18-tube modules having components redesigned
to reduce the cold flow in some of the plastic parts was inaugurated
April 15, 1968. Three parallel banks of three 18-tube modules in series
was set up for this test.
In this study at 600 psig, the flux could "be maintained at 8 to 1^ gfd
(corrected to hO°C) for a recycle feed actually operating in the range
of 30-35°C. A combination of pressure pulsing and "ball flushing of
the modules was employed during the run to develop the data shown in
Table 96. Although only the optical density rejections are summarized
in this table, other rejections were on the order of 96-98 percent for
solids, 88-95 percent for COD, and 99+ percent for color when processing
with intact modules. Losses in product water quality were traced to
"module failures" due to either tube rupture or leaks at the intertube
seals at the end caps .
While the first study with the 18-tube modules was plagued with leaks,
this type of module failure was somewhat reduced in modules of later
design. However, the number of tube ruptures markedly increased.
Between the start of the trial on April 15, 1968 and August 1, 1968
(109th day), one module failed due to tube rupture and twelve leaked
heavily at the tube ends. A second tube ruptured on August 1, 1968
and by the end of the run on January 3, 1969 (266 days) twenty-eight
310
-------
TABLE 95
FLUX RATES AND MODULE REPLACEMENT FOR SEVEN- AND EIGHTEEN-TUBE MODULES ON LIFE STUDY
Processing 1 Percent Acid Sulfite Wash Water
Flux Rate, gfd at 25 C
Mediae Replacement,
total hours at time of failure
Week
0
1
2
3
1*
5
7-Tubea l8-Tubeb
lU.5 25.2
9.7 17-8
7.1 10.7
7.1* 15.2
6.2 11.3
6.2 9.6
l8-Tubec
17.5
16.1
10.9
11.0
8.9
7.4
7-Tubea
5915
6215
1*500
1*729
1*991*
6684
l8-Tubeb
95
1*6
30
21*
16
1*38
899
6
l8-Tubec
95
318
16
2
1*38
96
966
805
1
1
5.6
5.7
9-*
8.0
10.7
9-7
Out of
system
Out of
system
a Bank of eight modules in series.
Bank of three modules in series.
c Bank of three modules in series.
21*
311
-------
modules had failed due to tube rupture and twenty-two by developing
leaks (Table 96).
Examination of several of the tubes in consultation with representatives
of the manufacturer led to the conclusion that some of the tubes had
failed due to a malfunction of the tube winding machine during manufac-
ture of these modules, as evidenced by the insufficient bonding between
the fiberglass and the resin binder.
Since several of the tubes that had failures had been in modules which
had been repaired several times as "leakers," there was also the possi-
bility that the tubes had been damaged during these repair operations.
To repair the 18-tube modules it was necessary to remove the tube bundle
from the tight fitting product water jacket. Such a step subjected
the outer tubes of the bundle to considerable friction, (abrasion) and
strain, as well as the possibility of wall damage from resting against
or striking a sharp object while unprotected by the module casing.
Approximately 80 percent of the tube failures were observed to occur
in the outer tubes around the perimeter of the 18-tube bundle within
a module.
Pilot-Scale Mais with Small 1000-5000 Gallon
per Day PiLlot_Unit_s__at..Mll_l Sites •
An on-site field trial with a neutral sulfite semichemieal wash water
with nine of the 18-tube modules "was operated at a maximum of 550 psig.
When the operating pressure was raised above this level, leaks appeared
and disappeared, with resultant changes in product water quality without
known correlation to operating conditions other than the pressure.
Each of the five field demonstrations for this project were preceded
with preliminary runs with smaller 1000-5000 gallon per day pilot units
over periods of three months or more to establish preliminary parameters
of operating on each new substrate and to obtain data for design instal-
lation and operation of the larger field units. Diese preliminary runs
provided some additional information on life history of modules. In
some cases the modules were used on several different mill wastes during
their test life.
Most of the small field trial experiences paralleled the laboratory-
module life studies for the spiral wrapped tubular designs discussed
in the first part of this section. However, many variables were under
study'and life data were difficult to tabulate.
These initial tests of the newly developed 18-tube design constructed
with replaceable tubes demonstrated much need for further improvement
in details of module design and need for close control'of quality
during manufacture. Evidence for such improvement was apparent in suc-
cessive shipments from the factory but much further improvement was
critically needed.
312
-------
U)
TABLE 96
FLUX RATES AMD MODULE REPLACEMENT FOR EIGHTEEN-TUEE SPIRAL WOUND MODULES ON LIFE STUDY
Processing 1 Percent Solids Acid Sulfite Wash Water at 600 psig
Flux Rates, gfd g Uo°C
Optical Density
Rejection, percent
Total Operating Hours Prior to Failure
Weeks Bank A Bank B Bank C
1-5 8.8 9.0 9.2
5-10 8.0 8.0 8.2
10-15 17.0 11.6 15.0
15-20 11.2 10.7 11.*
20-25 10.6 10.5 11.7
25-30 9.* 13.2 13.1
30-35 l*.l 11.6 9.*
35-End 13.* 15.* 13.7
Bank A
98
98
99+
99
99
96
97
99
Bank B
97
99
99
99
98
99
77
99
Bank C
97
95
98
99
98
99
99
99
Bank'A
190
1178
66
2520R *8
2*R 333R
3U22R 3587R
177R
1798
Bank
182
712
1027
2126R
3167R
896R
269
H55R
39R
2
*5*
673R
100
B
182
201R
1023
30*0
362R
2R
1R
233R
1*5R
*1*
1**R
Bank
1*06
1386
113
3111R
315R
60R
525
1689R
2586R
C
no*
157*R
12R
^ = modules vith ruptured tubes, other values "leakers."
Uote: Banks A, B, and C = Three parallel rovs of three modules in series.
-------
SMALL-SCALE LAB AND PILOT EVALUATIONS OF OTHER TUBULAR DESIGNS
'Ill-Tube'Modules
One of the early modules of the replaceable membrane type, produced
by another' supplier' consisted of 1^ polyester reinforced "braided fiber-
glass'tubes k feet'long and one-half inch inside diameter. A strip
of cellulose acetate membrane was inserted as a lining to form a tube
with a loose seam resting against the woven fiberglass wall. A slight
overlap of the lengthwise edges of the cellulose acetate strip provided
a compaction seal under'operating pressures above about lOO.psig. This
module was also equipped with turbulence promoters inside of the tubes
to permit lower operating velocities without concentration polarization.
Operation of this module at velocities one-fourth those for an open
one-half inch tube permitted the' installation of more modules (tubes)
in parallel for a pump of given flow rate output. Recycle to reach
a given' concentration level could be reduced. These appeared as sub-
stantial advantages of this design.
Our early trials with one of these modules 1, 5, 10, and 20 percent
solids concentration of calcium acid sulfite wash water were conducted
at 1*00, 600, and 800 psig.
Excellent flux rates, even at 20 percent solids, could be achieved at
800. psig (Table 97) and at flows as low as 0.25-0.5 gpm (12-25 cm/sec
velocity in empty tubes). At the same time rejections on the order
of 96-98 percent solids, 93-98 percent COD, and 98-99 percent optical
density were possible under these operating conditions (Table 98).
Shortly after this preliminary trial was completed, however, the module
failed with massive leakage, apparently due to the opening of one or
more of the membrane-overlap seams during the time the module was at
atmospheric pressure. The support tube did not rupture.
In July 1969 a second module was received, along with notification that
a slight positive pressure should be maintained at all times on the
high pressure (liquor) side of the tubes.
This five foot module was placed on life study with the 1 percent solids
calcium-base acid sulfite liquor. The module produced colored product
water samples occasionally. The' color indicated continuing problems
with leaks at the seam of the membrane. The flux rate gradually de-
clined (Table 99). After 330 hours (2 weeks) of operation, the flux
rate had declined from an initial 2k gfd to 13 gfd (a 1*0 percent loss)
and there was evidence of membrane (or module) fouling.
At this time a motorized bypass ball valve with less rapid action was
installed in the pressure line between the pump and the module to permit
slower and less violent reduction of the pressure from 600 psig to 25
psig for two minutes every four hours (pressure pulsing). During the
-------
TABLE 97
FLUX RATES WITH BRAIDED FIBERGLASS TUBULAR MODULES
Tap Water And Ca-Base Wash Water
Feed
Solution
Tap water
1 Percent solids
wash water
5 Percent solids
wash water
10 Percent solids
wash water
20 Percent solids
wash water
Inlet
Pressure,
psig
600
800
800
800
800
600
400
800
800
800
600
400
800
800
800
600
400
800
400
800
Temperature,
°C
34.5
34.5
33
34
33
35
34.5
35
34
33.5
34
33
34
34
34
34
34
34.5
34.5
36
Average Linear
Velocity,
cm/ sec
48
46
29
16
6
30
21
30
16
7
31
32
31
17
7
31
32
31
32
65
Flux Rate,
gfd
19.7
24.1
18.8
18.3
16.6
13.1
9.9
14.4
13.9
12.3
11.0
6.8
11.4
10.3
9.0
8.0
6.9
9.9
4.6
9.7
315
-------
TABLE 98
PERMEATE QUALITY FROM BRAIDED FIBERGLASS TUBULAR MODULES
Ca-Base Wash Water at Various Concentration Levels
Feed Concentrations
Solids Concentration,
percent
1
5
10
20
Solids,
mg/1
11300
55200
81400
•. *. *H ** *.
COD,
mg/1
14700
37200
111900
281000
Calcium,
550
2500
2700
7500
Optical Density
at 281 nm
86.6
450
640
1450
Rejection Ratios, percent
Solids Concentration,
percent
1
5
10
20
Solids,
mg/1
98
98
96
COD,
mg/1
98
97
95
93
Calcium,
ing/1
97
98
92
Optical Density,
at 281 nm
99
99
98
98
next 17U hourr. (7 days) "by this modified method, of operation, the flux
rate remained relatively constant or increased slightly, "but the quality
of the product water continued to deteriorate below acceptable standards
and the study was discontinued,
For other tubular designs, pressure pulsing seemed to be one of the
better methods of reducing severe fouling in the processing of pulping
effluent containing suspended solids and colloidal suspensoids. The
unsealed membrane seams of this modular design were indicated to be
impractical for operating by that method. The manufacturer then sup-
plied two newly redesigned braided fiberglass modules, complete with
"seamless" cellulose acetate membrane-tube replacements. These were
put on line for an extended module life study on March 2k, 1970. A
third module of the same design was added on June 22 after the unit
had been in operation (2070 hours, 88th day). Pressure pulsing (600
psig to 2^4 psig) was practiced on an 80-second per hour cycle throughout
this run,
The physical stability of these vertically mounted modules, the flux
rates, and the product water quality (Table 100) were excellent for
approximately 5800 hours (35 weeks). Occasionally color appeared
316
-------
TABLE 99
FLUX RATES AND PERMEATE QUALITY WITH FOURTEEN-TUBE MODULES ON LIPB STUDI
Processing 1 Percent Solids Acid Sulfite Wash Water
Flux Rate, Optical Density of
Hours gfd at to C Permeate, 281 nra
1 2^.5 0.54
kQ 19.9 kA
16? 15. ^ 2.1
209 15-8 2.7
237 19.8 2.8
330 13.0 2.1
379 12.0 3.3
te5 11.5 5.8
502 16.1 35-0
in the permeate water for short periods of a day or so in samples from
Module 1, but at no tine did the rejections fall below 90 percent for
solids, 85 percent for BODs, 90 percent for COD, and 90 percent for
optical density (color). In most product water samples taken weekly,
these rejections were on the order of 98 percent, 97 percent, 98 per-
cent, and 99 percent, respectively.
During the first 30 days of operation, the flux rates declined from
a starting value of 11.2-12.5 gfd to levels of 6-7 gfd for the two
modules first placed in operation. Within 20 days after the third
module with a tighter membrane had been installed, its flux rate had
also levelled off at 7 gfd from an initial 9 gfd.
After approximately a month of operation we noted a buildup of a gray-
black "furry" slime coating on the outside of the individual tube of
Modules 1 and 2. Washing the outsides of the modules with tap water
and an air-water mixture did little to remove the deposit, but peri-
odic washing with an enzymatic laundry detergent followed by a rinse
with pH 3 water did keep the formation from spreading or developing
into layers, which would slough off into the product water. Several
trials with one-half hour soakings of the entire outside of the module
in 100 ppm copper sulfate had little effect.
317
-------
TABLE 100
FLUX AND PERMEATE QUALITY WITH FOURTEEN-TU8E MODULES ON LIFE STUDY
Processing I Percent Solids Acid Sulflte Kash Water
(600 psig Operating Pressure)
COD^
Week
1
U)
M
CO
10
15
20
25
30
35
40
Stream
Flux R*te,
gfd at 40°C
Optical
Density.
Rejection,
percent
at 281 nm
Rejection,
percent
Feed
P-l
P-2
Feed
P-l
P-2
Feed
P-l
P-2
Feed
P-l
P-2
P-3
Feed
P-l
P-2
P-3
Feed
P-l
P-2
P-3
Feed
P-l
P-2
P-3
feed
P-l
P-2
P-3
Feed
P-2
9.9
9,4
8.7
8.2
8.5
7.9
8.6
8.2
7,7
9,4
7.4
6.7
8.2
7.2
6.7
11.9
6.5
7.4
9.9
5.1
7.5
3.8
9820
398
98
9692
111
112
11544
134
106
11026
267
124
119
10670
678
135
146
11244
690
227
454
10882
1091
136
502
11296
310
139
197
11228
1122
..
96
99
__
99
99
99
99
„
98
99
99
94
99
99
»_
94
98
96
__
90
99
95
__
97
99
98
__
90
2270
154
66
2543
77
93
3185
93
84
5110
136
78
82
3465
492
171
152
2690
229
62
1S4
2630
246
42
115
2280
92
38
58
2750
340
..
93
97
«-
97
96
...
97
97
..
96
98
97
~~
86
95
96
^_
92
98
94
„ —
91
98
96
~.
96
98
97
— _
88
12020
520
133
12320
114
117
14460
168
135
13460
321
103
112
13330
914
207
226
13900
879
119
578
13820
1386
133
618
13460
326
112
206
13740
1359
..
96
99
--
99
99
._
99
99
..
98
99
99
•«
93
98
98
„_
94
99
96
90
99
96
-.*
98
99
98
.,
90
73.2
3.1
0.68
71.8
0.54
0.48
86.2
0.77
0.66
81.1
1.58
0.51
0.54
82.1
4.98
0.75
0.84
85.0
5.23
0.63
3.55
80.8
7.78
0.63
3.50
85.6
1.62
0.58
1 22
81 2
7.78
--
96
99
99
99
_„
99
99
__
98
99
99
94
99
99
_.
94
99
96
90
99
96
98
99
99
90
-------
The third module, which was washed with the detergent solution from
the start of its inclusion in the series, did not develop the slime
coating and the tubes remained clean.
Problems arising from extended operation of these 1^-tube modules were
mainly concerned with fouling problems which were difficult to control
at low velocities . Some leakage was apparent but rupture of the support
structure was experienced on only one occasion.
16-Tube Woven Fiberglass Modules
This modular design incorporated the idea of a replaceable tube mounted
on heavy stainless steel end plates fabricated with tubular U bend
interconnections. These modules were vertically mounted and the product
water drained down the tube walls into collecting pans.
The cellulose acetate membranes were cast inside of the woven fiberglass
tube which had an 0-ring coupling on each end for attachment to the
end plates. This 0-ring coupling was excellent, with trouble-free,
fast action during replacement and this design appeared in other ways
advantageous in terms of mechanical design features as compared with
other tubular systems. The top end plate was mounted rigidly and the
bottom allowed to float free; thereby providing a slight tensile stress
on the tubes, eliminating the compressive stress which apparently had
caused difficulties in some of the other tubular designs.
The first two of these modules contained 16 tubes in a modular configur-
ation, U x 16 inches (end plate size) by 106 inches long. The tubes
were 9/l6 inches inside diameter and contained 1.2 square foot of mem-
brane per tube.
Data in Table 101 summarizes test results in terms of flux rate (gfd),
product water quality (percent rejection) and tube failures. The flux
rates and product water quality were both excellent, but the tube fail-
ures, six in a period of eight days of operation (average tube life
of lUo hours), caused discontinuation of the study. Of the six failed
tubes, three had pinholes in the tube walls and three tore loose at
the tube coupling joint.
The second generation of tubes arrived in our laboratories in April
1970 and were mounted on the module end plates used in the first trial.
The end plates were modified with tie rods between the end plates in
an attempt to relieve tension on the tubes without causing the tubes
to be under compression.
Within a few hours tube failures were apparent, which were similar
to those in the first study; namely, ferrule separation at the tube
ends. The slightly unmatched lengths of the tubes under different
pressures resulted in the tubes "bowing and whipping" with each stroke
of the pump. Eventually this seemed to result in fatigue rupture of
the tube, at the rigid joint with the tube ferrule.
319
-------
TABLE 101
FLUX RATES AMD MODULE REPLACEMENT FOR SIXTEEN-TUBE BRAIDED MODULES ON LIFE STUDY
Processing 1 Percent Solids Acid Sulfite Wash Water at 600 psig
Flux rate, gfd at to C
Optical Density
Rejection, percent
Operating Hours
Prior to Tube Failure
Hours Module A Module B
^5 n.l 10.9
UJ
0 139 12 .k 9.9
k 2^.2 22.6
117 17.5 16.8'
253 l^.S 9.1
Module A Module B Module A Module B
First Trial
99
99
Second Trial
98
93
99
99
99 126,130,138 131*-
158,158
98 k
96
99 266,260,253 205,267,2*
267
-------
Flux rates varied widely during this test period, from 26 to J gfd
depending upon the age of the tubes in the module. The quality of
the permeate was quite high (Table 101), even at the higher flux rates
of 18--22 gallons per square foot per day.
Within a l6-day period, ten tube failures were experienced and the
study was again discontinued due to lack of tube replacements. Struc-
tural failure of the membrane support tubes was a major problem in
testing this type of tubular equipment.
36-Tube Modules withHonporous Tubes
During the latter part of December 1969j two 36-tube modules were re-
ceived from still another manufacturer for evaluation in the life study
program. These modules were arranged with six banks of six tubes,
each in a h x 11 x 106 inch space. The inner tube connections were
an integral part of the end plates and provided for series flow through
all 36 tubes. Optional end plates of a different design• could also
be used to provide parallel flow through the 36 tubes for use in modi-
fied methods of processing at high flow rates.
The membranes were cast on porous paper support tubes and could be
replaced by removing the end plates from either end of the module.
Data on flux rates and product water quality for the two modules are
given in Table 102. Data are also presented from operation of a third
module which was received on January 17» 1970 near the end of the run,
While the product water quality was generally excellent, there were
periods of highly colored permeates immediately after pressure-pulsing
cycles, and the modules failed "by massive leakage after 6ll, 822, and
276 hours (Modules 1, 2, and 3, respectively).
Shortly after the trials with these three modules were discontinued,
a company representative attempted to modify all three by installing
newly designed end plate gaskets and new tubes. Difficulties were
experienced in removing the inner membrane tubes, and the nodules were
returned to the factory where special tools were available for removing
the tubes without damage to the inner support structure. Swelling
of the replaceable paper tubes may have "been related to the pressure
pulsing system with back flows, for which the nodules had not been
designed. However, the swelling of the paper tubes did not appear
to be the cause of tube failures. Cracks were apparent in the plastic
sleeves at the tube ends where they mated with the 0-rings of the end
plates.
Three newly redesigned modules were installed on February 11, 1970 .
for series operation through the three in series. Due to the" difficulty
encountered in the first trial, care was taken to provide atmospheric
discharge of "both the product water and the concentrate to avoid siphon-
ing and "back flow problems which may have contributed to the' tube
swelling and seizure.
321
-------
TABLE 102
FLUX RATES AND REJECTIONS FOR THIRTY-SIX TUB! MODULES ON LIFE STUDY
1 Percent Acid Sulfite Wash Water at 600 psig
Rejection, percent
KLux. . Optical Density
Hours Stream gfd at TO C Solids BOD- COD at 28l na
1 P-i 111-.2 — __ 99 99
P-2 15.0 — — 99 99
88 p-i 11.5 99 97 99 99
P-2 11.8 99 97 99 99
257 P-l 9.0 99 9k 98 99
P-2 10.5 99 9k 98 99
^25 P-l 8.1 99 9k 98 99
P-2 10.6 99 96 99 99
56? P-l 6.8 99 96 99 99+
P-2 9-6 99 96 99 99+
728 ' p-2 9.6 99 98 99 99
^ P-3 13-2 99 98 99+ 99
21k p-3 UL.JI. 99 99 99 99
'.Che flux rates steadily declined in spite of pressure pulsing (30 seconds
per hour); weekly treatment vith a 1000 ppm copper sulfate solution;
and an increase in fluid velocity from 3.^, 3.0, and 2,8 feet per second
in the three modules to U.I, 3.8, and 3-5 feet per second, respectively.
Product water quality remained high during the first part of the study.
But by the end of the twelfth week (19U6 hours) all were producing
colored permeates» although only Module 3*s rejections were markedly
low.
Since the manufacturer had developed new seal rings, the modules were
rebuilt with all new seal rings and six new tubes in Module 2 and two
in Module 3. "Zhe difficulty was encountered in removing the tubes
in this last replacement, in spite of the fact that back flow and siphon-
ing had been eliminated.
The third trial terminated after seven days (l60 hours) due to the
reproduction of colored permeates in all modules of the line. All
three were returned to the manufacturer for refitting.
Ihe three completely new modules in the fourth trial developed colored
permeates in Modules 1 and 2 after 275 hours and 280 hours, respec-
tively, with gradually decreasing quality until both were removed from
322
-------
the system at the 851st hour. On the other hand, Module 3 produced
high quality permeates until the 1118th hour, when the permsate suddenly
colored and the trial was discontinued.
The series of tests with the 36 tube modules, with replaceable membranes
mounted in paper tubes, showed evidence of leakage of the sleeve sealing
connections within the membrane support structure, "but no evidence
of failure of the membrane support structure. These modules provided
much evidence of highly developed engineering design of the support
structure and of the excellent manifolding system. The leakage problem
within the membrane support structure was being intensively studied
and advances were indicated at the time of terminating the tests.
Life Performance in Large-Scale'Field Trials -
'With the Trailer-Mounted' Unit -
The advancing technology and growing experience with large laboratory
and small pilot RO units having membrane areas in the 10 to 100 sq
ft category provided the base for designing the larger trailer-mounted
unit which could more adequately evaluate the practical problems and
economics of field-scale commercial operation. Actual experience with
the life performance of a large BO system was therefore an important
objective in planning the experimental program for the Held test unit
designed and constructed in 1968. This unit as delivered in October
1968 had 387 modules with 6580 sq ft of working membrane area, plus
20 spare nodules for replacement. Special planning included setting
up a card catalog for following the individual life history of all
modules on this unit.
A first test of performance capabilities of the membrane system, in
terms of resistance to damage during shipment.for long distances,
occurred in transporting the trailer unit complete with modules mounted
in place some 2kOO miles from the factory in San Diego to the site
of the first demonstration at the sulfite pulp mill in Appleton, Wis-
consin. The five banks of modules were mounted on adjustable channel
steel racks with flexible metal hose connections to the multi-valved
manifolding system at the completion of the factory testing program
prior to delivery. Only a minimum of taping and support blocking was
felt to be required, and this appeared to have been fairly adequate,
although a few modules did shift substantially out of line without
apparent damage in transit. All but a few of the topmost modules high
in the stacks were still filled with water on arrival after 5 days.
enroute.
The operation and performance of the trailer-mounted unit in the three
field demonstrations have been reviewed in detail in Section VII of
this report. The reader should refer especially to the analysis of
downtime reported in Tables 32, 37, and kf for each of the demonstrations
The following paragraphs summarize the life expectancy performance '
of the 387 modules originally supplied with the unit in October
323
-------
and also of the performance of 238 modules which had been returned
to the factory for rebuilding in the period of field demonstration
Ho. 3.
Table 103 summarizes the accumulated data.
TABLE 103
LIFE PERFORMANCE OF 387 MODULES ON TRAILER-
MOUNTED FIELD UNIT
No. modules in service
Hours available on-site'
operation
Total trailer operating
hours
No. modules failed and
replaced
Type of module failure
Loss of membrane
Tube rupture
Leak in tube seals
scale
Field
Failure rate per 100
operating hours per
1000 sq. ft of membrane
Field
Field
Demonstration
No. 3
in strati or
No. 1
387
1386 i
690 .
8
i Demonstration
No. 2
387
985
59^
^6a
1st
Period
387
Bkl
lt66
62
2nd
Period
238b
226
211
58
0
6
1
1
6
30
2
8
0
55
k
3
0
Ik
uu
0
0.176
1.178
1.996 6.19k
\h Modules partially plugged with loose membrane (from other modules)
vere cleaned and returned to service.
238 Modules rebuilt at factor;'-.
32U
-------
The three 3-month trials extended over a period of nearly 12 months .
Records .show a total of 3^38 hours were available; ,. for operation at
the three field sites, of which 1967 hours, or about one-fourth of
the' one-year period resulted in actual operation of the unit under
pressure with mill effluent feed. During the first field demonstration,
8 module failures occurred during 690 hours of operation. Since the
manifolding system was equipped with shut-off valves on each set of
from 3 to 5 modules, the group of modules in which a failure had occurred
could "be closed out of the system without need for shutting down the
whole unit. Repairs or replacement could be conducted when time per-
mitted during the course of sustained operation, and no shutdown time
was experienced in this run "because of module failures.
In the second demonstration , available operating time totalled
hours, "but in this case k6 nodules failed or, were replaced during the
run. An additional kk modules were found to have partial or complete
plugging with loose membrane. Nine and one-half hours of operation
time were lost when it was found necessary to shut the entire unit
down to locate, to backwash, and to replace those kk modules. Appar-
ently, a rare case of membrane stripping occurred, and the loose mem-
"brane traveled through the remaining affected modules downstream. Most
of the Modules were found to "be easily back washed to remove the loose
membrane. Those were cleaned, tested, and returned to service. Exami^
nation showed a total of 6 modules from which the membrane had been
stripped in one or more tubes, "but it could not be determined whether
more than one module of the six was the original cause of membrane
loss since the loose membrane from one module could be cause for strip-
ping or plugging in modules downstream.
A much more serious effect on life expectancy of the total number of
modules arose from the 30 modules having structural failures of the
membrane support system, with resultant rupture during the course of
this second demonstration.
In the third demonstration, 62 module failures and replacements occurred
during the first period of k66 operating hours. Fifty-five of these
failures were due to tube rupture, k to apparent leaks in tube seals,
and 3 were found to have calcium sulfate scale. Module failures
accounted for 23 hours of downtime.
After rebuilding 238 of the modules at the factory, another 211 hours
of operation occurred in the second, period for this third field demon-
stration. Module failures and replacements totalled 58 and were cause
for 23 hours of downtime. Of these failures, lU were due to tube rup-
ture. In the second period most 'of th'e failures were due to hk tube
seal leaks , tut no downtime resulted from that problem. It was Apparent
that the small plastic tube seal inserts used during the rebuilding
program at the factory were faulty, since leakage had been a minor
problem theretofore.
325
-------
An attempt was made to evaluate the' rate of module failure In terms
of downtime for each hundred' operating hours per thousand square feet
of membrane. For the first demonstration,, this downtime factor was
calculated to "be 0.176. In the' second demonstration it rose to 1.178,
while the first half of the third field demonstration showed 1-996
and the second period 6,79^- This evaluation factor, based on failure
of module units, does not include allowance for variation in size and
cost of the wide variety of modules being developed and marketed. The
module is the failure unit under consideration. It will "be of interest
to follow development and use of this or a similar factor for congpara-
tive evaluation of failure as cost evaluation of membrane systems de-
velops .
The rate of module failures at 0.176 during the first 3-inonth demon-
stration appeared promising of "becoming & supportable item of operating
cost in a large installation. Hie failures occurring later in the
life of the module support structure during the second 3-month field
demonstration were found to be far too high. Sie rupture failures
at the 2.0 factor level which occurred as aging advanced in the first
period of operation during the third demonstration confirmed the rising
rate of failure at levels far beyond economic feasibility. Continued
operation of the trailer unit could not "be justified until more reliable
membrane modules could be proven out.
Membrane life expectancy tests were subsequently continued with equip-
ment supplied "by a number of principal supply firms introducing new
and improved equipment during the period 1968 through 1971-
• Discussion of Life Expectancy of RQ Equipment
The overall picture for life performance of membrane equipment of these
studies was disappointing because the critically high costs of membrane
module maintenance and replacement substantially reduced commercial
feasibility of the HO concentrating process at the stage of RO engineer-
ing and development prevailing during the 3-1/2 to k years of study
on this research and demonstration project.
High levels of technical performance for the RO process in terms of
capabilities for concentrating dissolved substances from dilute solution
has appeared excellent throughout this project. But commercial feasi-
bility for large-scale operations plainly must wait for proven success
in accomplishing the next step in design and production of membrane
equipment. A much higher level of reliability for sustained operation
at life expectancy levels greater than 12 months must be attained.
The' capillary fiber and the spiral wound sheet membrane systems showed
good evidence of having structural stability and freedom from internal
leakages , but the experience gained during the course of this 3-1/2.
year demonstration showed few' of the industry effluents which were. .
subject for study to be free of problems of fouling of the membrane or
alternately plugging of the passageways in these types of membrane
equipment•
326
-------
On the other hand, tubular-type module configurations were found capable
of handling the fouling problem at high velocities. Failures due to
plugging of the large passageways, 1/2 inch in diameter or greater
seldom occurred. But the provision of structural stability and freedom
from serious leakages for the tubular system has been an engineering
design problem not completely solved and proven during the period of
this research and demonstration grant study.
Several causes might have been responsible for structural failure in
the tubular membrane equipment. Support structures based upon high
strength, corrosion-resistant material such as 316 stainless steel
or with other types of less expensive metal structures, protected by
coatings or resin linings , were proven quite adequate in eliminating
tubular rupture. But these materials of construction had to bear a
serious cost handicap, and all manufacturers producing prototype equip-
ment with expensive metal support structures or housing were indicated
to be advancing design of commercial equipment toward use of lighter
weight, less expensive composite structures, such as resin-bonded fiber-
glass and the like.
The results of the studies reported for this project strongly indicate
that the resin-bonded fiberglass structures available for these studies
were subject to stress fatigue in sustained operation after a period
of several months. The stresses of high pressure operation at 600
psig or higher were greatly accentuated by programmed pulsations and
by fluctuations during the course of shutdowns and startups during
six months and more of operations. Serious leakages were probably
also associated with stress fatigue in the connecting seal structures
in a number of types of equipment tested. Use of 0-ring seals has
appeared to be an excellent answer to the problem of sealing internal
connections within the membrane support structure. It was plain that
further improvement in design could logically follow intensive engineer-
ing studies known to be in progress.
Basically, it may be concluded that the life expectancy of RO equipment
is a problem of improvement in terms of engineering design and quality
control during manufacturing operations. The structural failure problem
seems solved by high levels of engineering design for the capillary
fiber and spiral wound systems. However, these excellent qualities
are counterbalanced by the problems inherent in design of equipment
which can provide high velocities and turbulence across the surface
of the membrane to keep the membrane clean and free from fouling
problems. The studies in this grant project have therefore trended
toward preference for tubular systems because of the proven capability
to maintain high levels of flux through the membrane at high rates
of rejection with greater degrees of freedom from fouling or plugging
problems.
Emphasis in these studies has been heavily directed toward the search
for reliable designs and configurations for the membrane support struc-
tures. Membrane failures as such have been surprisingly few, and
327
-------
apparently are within the range of commercial feasibility at the present
stage of development. Cellulose acetate membranes, if operated at
moderate temperatures below ^5°C, and within pH ranges of about 3.0
to 7-5 j have appeared adequate in terms of life expectancy. Preliminary
tests with nylon and other types of membrane formulations also appear
promising in terms of membrane life expectancy if operated within recom-
mended limits for temperature, pH, and the like, and apparently have
substantial advantage over cellulose acetate in operating over a wider
range of temperature and pH. The problem then appears to be a matter
of improving the structural design and the quality of manufacturing
operations for the membrane support structures.
Conclusions as to life Expectancy and Sustained
Performance of RQ Equipment
life expectancy of RO membrane equipment must be adequate as measured
on two basic criteria:
1. RO membrane equipment must be capable of sustained operation
free of problems of being plugged or of becoming irreversibly
fouled by suspended particulates, colloidal suspensoids, or by
large molecular weight solutes contained in the feed liquor
or which develop during concentration.
The plugging and fouling problem is of particular concern in
processing industrial effluents as discharged in pulp and paper
manufacturing operations based upon wood-derived organics and
inorganic pulping chemicals. The extended evaluatory studies
conducted under this research and demonstration grant point to
the self-cleaning capabilities provided by high velocity and
highly turbulent flows within tubular RO systems as being best
adapted to processing such effluents. Capillary fiber and sheet
membrane pack systems at their present stage of development
apparently will require high levels of pretreatment for most
of these industrial effluents. Such pretreatment requirements
might be provided by ultrafiltration.
2. The life expectancy of RO membrane equipment must be adequately
proven in terms of long-term stability of the membrane support
structure, and of sustained leak-free performance of internal
connections based upon the support structure. The studies
under this R & D Grant project have failed to demonstrate
adequate leak-free, life expectancy of membrane support struc-
tures of the tubular conformation as designed, manufactured,
and available for testing within the U-year program period.
The requirements of establishing the feasibility of RO membrane
processing systems seem best fulfilled by tubular conformations
in terms of sustained performance free of plugging and fouling
problems when treating dilute pulp and paper processing ef-
fluents . On the other hand, the capillary fiber and sheet
328
-------
membrane systems have been indicated as most advanced in terms
of freedom from failure of the membrane support structure and
associated leakage of internal connections.
None of these systems were found capable of meeting both of
these requirements basic to establishing large-scale feasi-
bility for processing of these dilute industrial processing
effluents.
329
-------
SECTION X
PROCESS ECONOMICS
Justification for the extensive application studies undertaken on this
Research and Demonstration Grant Project for evaluation of reverse
osmosis as a method of concentration processing of dilute pulp and
paper manufacturing effluents has necessarily required frequent assess-
ment and reassessment of economic feasibility. Process concepts were
new and novel when undertaking the project in 1966. Reverse osmosis
is still under intensive research and development, and is not yet in
large-scale commercial practice as of January 1972. As such, it has
not been possible to have the benefit of actual experience in developing
capital costs for this membrane equipment in large-scale commercial
production, nor can conclusions be drawn on the firm base of actual
costs of operation from large-scale operations of the RO concentration
process in other fields, such as saline water conversion.
Nevertheless, extended engineering studies have been reported28'.29
upon which fairly adequate estimates of the present and future economic
outlook could be established. Personal interviews and letter surveys
with responsible executives and engineers in the RO field have tended
to confirm the estimates in the above-cited cost studies for the Office
of Saline Water. These surveys have indicated capital charges might
be expected to be in the range of $1.00 per installed gallon of permeate
water production for plants producing one million gallons per day,
and also have forecast a range of 50^ to $1.50 per 1000 gallons for
cost of the permeate water produced as a measure of operating charges.
These estimated and preliminary ranging figures have been in use with
qualifications for the purpose of designing, directing, and evaluating
progress of developing the RO process.
However, progress of membrane equipment suppliers has been slow in
achieving projected minimum levels of performance and long-term life
expectancy of the modular membrane structures during the four-year
period of these project studies. Although technical developments have
in other ways been excellent and important, it has therefore been neces-
sary to substantially increase those preliminary estimates in the cost
evaluations for this summary report. Capital costs for equipment remain
high, but of more concern are the operating charges for maintenance
and replacement of short-lived RO equipment, which in some instances
of pilot-scale experience have been estimated to range to 60 or even
80 percent of total operating charges at 1971 levels of performance.
These high charges should be greatly reduced when large-scale production
of membrane equipment is achieved. Evidence does point to substantial
advances by the equipment suppliers.
The objectives for this stannary of process economics are directed to
evaluating the experience gained in the laboratory, pilot, and field
demonstrations in terms of existing conditions. Needed areas for im-
proving the economics for putting RO to active use in concentrating
331
-------
dilute pulping industry effluents can be "better pinpointed and progress
accelerated in achieving those goals.
Since membrane equipment costs will in all probability remain high
for some time, ve can expect the first industrial applications of RO
will be on a modest scale for concentration of wastes containing market-
able values, and in this manner partially balance the large expense
of early stage RO process developmental charges. Larger scale, low-
cost concentration processing of dilute wastes can then be expected
to follow later on when membrane equipment is marketed in sufficient
volume to benefit from the reduced costs of large-scale commercial
production.
In order to better evaluate the costs which might be applicable to
conditions prevailing at the five sites of the individual field demon-
strations conducted for this project, a computer program was developed
which could provide a comparable picture based on the actual waste
flow processed, and upon operating conditions, power costs, and other
significant variables. No attempt was made to conduct a detailed engi-
neering cost survey, and the data presented in Table 10^ should be
so understood and evaluated as a preliminary study appropriate to the
present state of the art for manufacture and operation of RO equipment.
Waste flows studied at these mills ranged in volume from 125,000 gal.
per day to 1,000,000 gal. per day, and the solids concentrations of
the effluent to be processed varied from 1/2 percent solids to 3.0
percent solids. Two sets of cost data were compared. The first was
based on good RO processing experience obtained in tests on these wastes
with 1968-1970 Havens modules, and the second on high performance mem-
branes which became available from several suppliers for field testing
in 1971. Flux rates on test salt solutions with use of the older mem-
brane equipment were at a level of 12.5 gfd, while the new membranes
delivered at twice that flux rate. Several of the factors used require
further definition:
All data were based on operations at 600 psig feed pressure
and 35°C.
Module cost complete with membranes was quoted at July 1972
expectations of $9-30/sq ft of effective surface.
Module maintenance charges at $2^/module/year were quoted
to cover cost of servicing, back washing, repairs, and
labor for replacement with 2-year membrane life.
Power cost varied with the mill between 0.6 and 1.20/Kwh.
Module depreciation was based upon 5-year life of the
module support structure.
332
-------
TABLE 104
OIDICATED CAPITAL ADD OPERATISG COSTS OF REVERSE OSMOSIS- UNITS BASED UPON
FIVE FIELD DEMOJISTBATIOHS - CALGOH-HAVEHS 18-TUBE TUBULAR MODULES
Reference Pressure » 600 psig
Reference Feed Liquor Temperature = 35°C
Module Cost = $9.30 per si. ft. of Membrane
Module Maintenance Cost = $2U per Module per Year (2 year life of membrane)
Cost of Electric Power » 0.6-1,2* per Rwhr
Module Depreciation = 5-Year Life of Module Support Structure
Non-membrane Equipment Life = 5 Years
Number of Concentrating Stages = 3
Dumber of Modules in Series in Each Stage * 2
LO
Demonstration Modules (Reference
5000 mgA NaCl Flux Rate at
600 psig and 35°C » 12.5 gfd)
Latest Modules (Reference
5000 mgA Had Flux Rate at
600 psig and 35°C - 25 gfd)
Feed Liquor
Calcium-base acid
s.ulfite liquor
HSSC white water
Ammonia-base acid
salfite liquor
Design Capacity
of RO Unit ,
thousand gpd
of feed liquor
500
1W
125
Concentration
Range,
8/1
12-100
20-100
30-100
Overall
Average
Flux Ratec,
gfd
9.6
9.8
7.9
Total
Dumber
of
Modules
2923
813
70U
Total
Capital
Costd,
thousand
dollars
602. Q
207.0
179.0
Total
Operating
Cost,
$/1000 gal.
product water
1.5^
1.91*
2.21
Overall
Average
Flux Hate,
gfd
18.9
19.2
15.5
Total
Dumber
of
Modules
1U85
115
358
Total
Capital
Cost,
thousand
dollars
3U3.0
133.0
116.0
Total
Operating
Cost,
*/1000 gal.
product water
0.95
1.20
l.W
Caustic-stage kraft
bleach effluent liquor 1000
Chemimechanical pulp
•-rash vater 550
5-50 11.2 5207 1050.0 1.32
6-100 11.0 3025 658,0 1,35
22.2 2580 582.0 0.82
21.5 151*1* 387.0 0.83
*Used throughout the 5 demonstration sites.
Recent laboratory and pilot plant experience.
Overall average flux rate
where n = number of concentrating stages
F. = average value of flux rate in the i-th stage
» percentage recovery of water in the i-th stage
= summation over n stages
The calculations of true overall average flux rate were made separately in the computer program,
using a sufficiently large number of concentrating stages.
T'otal capital cost includes: $100 for each manifolding unit of 6 modules
10? spare modules for all waste flows
10% additional for pulsing of NSSC liquor
$25,000 automated instrumentation/each 0.5 mgd.
-------
Non-membrane equipment as pumps, instruments, piping, etc.,
were based on a 5-year life.
The number of concentrating stages were based on a 3-stage
optimum system.
Each set of modules contained two in series .
Manifolding costs were estimated at $100 per 3 sets of 2
modules in series (6 modules).
Ten percent spare modules were allowed overall, and 10 percent
additional modules were provided as an allowance for downtime
during periodic pulsing of the system to control fouling
problems .
Instrumentation capital costs were estimated at $25,000 for
each half million gallons of permeate production per day.
The computer programming of flux rates was conducted separately
for each substrate.
Results of Computerized Cost Evaluation
The data on capital costs for RO installations at the five mill sites
tended to confirm the earlier estimates of $1.00 per gallon of daily
permeate water production on the basis of experience, with membrane
equipment available and used during the course of these studies in
the period 1967-1970. Substantial capital cost reductions of 10 to
^0 percent were indicated for use of the new high performance membranes
released by several suppliers for field testing in 1971.
The capital cost data in Table 10U would seem to indicate that mills
might build plants to concentrate dilute effluents of from 100,000
gpd to one million gpd at costs in the range of $1.00 per daily gallon
of permeate water production.
The critical charges for sustained operation and maintenance of an
RO concentrating system were, however, much higher than predicted in
the preliminary estimates, ranging from $1.32 to $2.21 per 1000 gal.
of permeate flow, with membrane equipment used on this project to a
range of $0.82 to $1.U8 per 1000 gallons for the newly available high
performance membrane equipment.
Reverse osmosis is not a complete process of waste treatment, and addi-
tional charges must be added for final disposal or utilization pro-
cessing of the concentrate.
However, some perspective as to how the costs provided in Table
can be evaluated as part of a waste treatment system can be gained
by comparison with the costs of bio-oxidation processing for conventional
33U
-------
disposal type treatment to reduce the BQD5 of such wastes. A rule
of the thumb cost estimate for activated sludge or trickling filter
treatment at U-l/2 cents per pound of BOD5 at the 90 percent removal
basis has been quoted at times ,in recent cost surveys. The BODs values
for the five effluents listed as feed liquors in Table IQh. have been
provided in the various individual field study reviews in Section VII
and may be summarized as follows:
Pounds Biological BODs Removal
per 1000 $/1000 Gallons
Gallons at l4-l/2$/Pound
Ca-base wash water 25.87 $1.16
NSSC white water 19-52 0.88
NH3 wash water 62.5 2.82
Alk. ext. KBE 1.6 0.07
CM wash water 58.7 2.6k
The RO charges provided in Table 10k can be interpreted as marginally
competitive against proven disposal processes as long as BODs removal
is the principal standard of environmental quality to be complied with
in treating these wastes. Bio-oxidation in aerated oxidation lagoons
may accomplish satisfactory levels of BODs removal at substantially
less cost than by the activated sludge or trickling filter systems.
Aerated lagoons, whenever feasibly installed, would be the choice for
disposal treatment of these wastes. Other factors do affect the choice,
however.
Although RO is not in itself a complete treatment process, its possi-
bilities for achieving complete treatment in an integrated system are
distinctly advantageous, and comprise the justification for the studies
reported and for continuing research and development.
Solutes are recovered in the concentrate at high levels of
rejection for final disposal or for recovery and utilization
of any values.
Permeate water is of high quality for recovery and reuse in
the mill system.
Closure of the mill water system is the goal of complete waste
treatment.
Recovery of values from the concentrates and in the form of
reusable permeate water can substantially reduce the cost of
complete treatment.
335
-------
The evaluation of RO and of the costs of such processing has been advanced
in a supplementary RO project conducted by the staff of Green Bay Packag-
ing Inc. at the site of the second field demonstration. This organization
has specialized processing conditions peculiar to that mill, and goals
have been established for long-range closure of their mill system.
These conditions have encouraged further studies for application of
RO to the concentration of their on-machine wash water or "white water"
effluent.
The mill staff has conducted a substantial program of pilot-scale evalua-
tions independent of the studies covered in "this report during 1970-
71. Basic results of the original project findings are being confirmed
and interest continues at that mill, but adverse experience with module
maintenance and replacement problems has been such that they have as-
signed far higher charges in this category. The previous table (lOU)
listed the module maintenance charge at $2^ per module per year. The
Green Bay experience with the equipment they have tested points to
$100 per year. Modified cost data based on information supplied by
the mill staff are provided in Table 105-
The computer run has evaluated and proven a substantial optimization
results from the design of the module manifolding system in terms of
employing a single module, two, three and four modules in series. The
white water is indicated to be much more advantageously processed,
with only one or at most two modules in series in terms of both the
capital and the operating cost categories.
Capital costs for the mill study are somewhat higher than provided
in the original cost comparison provided in Table 10^, but the operating
charges range from 2 to k times higher. This is largely due to the
much greater charge assigned to the module maintenance and membrane
replacement costs at the $100 per module per year level.
There is much indication in recent studies of advantage arising from
use of fewer modules in series, newer developments in membranes becoming
available on the market, and use of higher velocities of flow rather
than pulsing for fouling control. These factors have a principal effect
on the increased operating charge.
Pressure pulsing of the entire module system by depressurizing at fre-
quent intervals of each hour or so around the clock for months and
years has been recognized as a severe test of module life performance.
Much work directed to eliminating need for pulsing was reported in
Section VIII of this report. Use of fewer modules and higher velocities
greatly reduced or even eliminated need for pulsing. The optimization
studies further confirmed that finding against use of pulsing. Confir-
mation trials in the mill are needed to prove these findings under
practical plant operating conditions and are being planned.
The need for extending proven life performance of membrane modules
in terms of freedom from irreversibly plugging by suspended matter
336
-------
TABLE 105
SUPPLEMENTARY CAPITAL AND OPERATING CHARGE ESTIMATES
Mill Pilot Experience on NSSC White Water
LO
U>
—J
Overall Average
No. of Flux Rate at Total
Modules 600 psig and 35°C, No. of
in Series gfd Modules
1 8.3 887
2 7.4 985
3 6.6 1102
4 5.7 1284
Total Capital
Cost,
thousand
dollars
231.3
235.0
250.0
277.8
Design capacity of RO unit - 144,000 gal./day of liquor
Initial concentration of the feed = 20 g/1 feed rate
Final concentration of the feed = 100 g/1
Reference 5000 ppm sodium chloride solution
Flux rate at 600 psig and 35°C =10.0 gfd
Module cost = $9.3 per sq ft of membrane
Module maintenance cost = $100 per module per year
Cost of electric power = 1.2 cents per Kwhr
Module depreciation = 5 years
Non-membrane equipment life = 5 years
Total Operating Cost,
$/1000 gal.
product water
3.79
4.06
4.46
5,08
-------
and other materials during concentration processing, and even nore
importantly the extension of the life performance in terms of freedom
from failures in the membrane support and in connecting seals is a
primary goal in cost reduction for the RO process".
Practical applications for RO in the' concentration processing of dilute
pulp and paper manufacturing effluent streams avait proving out of
these substantial life performance criteria for the membrane nodules.
338
-------
SECTION XI
ACKNOWLEDGMENTS
The development of design factors for construction and installation of
pilot and field demonstration units, the conduct of laboratory, pilot,
and field studies, together with analytical work, engineering evaluations
and report writing were performed "by a single team. The project was
initiated by the Pulp Manufacturers Research League and cooperating
member pulp manufacturing corporations. The League merged into The
Institute of Paper Chemistry, April 1, 1970, and its staff "became the
Effluent Processes Group within the Institute's Division of Industrial
and Environmental Systems. Averill J. Wiley directed the project
throughout.
Mr. J. M. Holderby, Consultant to the Research League, contributed ac-
tively to the supervision of field demonstrations as conducted by the
staff field engineers; first by Mr. A. C. P. Ammerlaan and subsequently
by Messrs. Kenneth Scharpf and I. K. Bansal. Mr. Bansal has contributed
especially to the optimization studies and the computer program evalua-
tions of process economics in Sections VIII, IX, and X. Mr. George
Dubey, Research Associate, has supervised laboratory studies conducted
throughout the project, and especially those described in Section V.
The conduct of a research and development project extending to five field
demonstrations at different pulp and paper manufacturing facilities, has
necessarily involved extensive and expert participation and cooperation
of responsible individuals from each of those organizations.
The administration of the project was greatly facilitated by assistance
of the Reverse Osmosis Subcommittee of the Research League under the
Chairmanship of Mr. G. K. Dickerman, Technical Assistant to the President
of Consolidated Papers, Inc., until his retirement in 1969* and subse-
quently of Mr. William R. Nelson, Director of Corporate Development for
Green Bay Packaging Inc. The Subcommittee membership has also included
the following:
Dr. Jack Jayne, Environmental Research Team, Corporate
Research and Engineering, Kimberly-Clark Corporation,
Neenah, Wisconsin
Mr. Francis C. Schroeder, Director, Environmental Control,
Potlatch Forests, Inc., Northwest Paper Company, Cloquet,
Minnesota
Mr. Donald Pryor, Manager, Environmental Control,
Consolidated Papers, Inc., Wisconsin Rapids, Wisconsin
Mr. Milton A. Lefevre, Manager, Forest Chemical Products,
Scott Paper Company, Oconto Falls, Wisconsin
339
-------
ACKNOWLEDGMENTS (Cont'd.)
Mr. Donald Arps, Supervisor of Environmental Control,
Appleton' Papers, Inc., Combined Locks, Wisconsin
The completeness of these studies has been made possible through the
excellent cooperation of membrane and process equipment suppliers and
of associated concerns developing specialized pumps and instrumentation
in this new field of concentration processing of dilute industrial ef-
fluents. The following suppliers especially are to be commended for
their cooperation, often at very great expense to their concerns and
without charge to the project:
Membrane Equipment
Havens International and later Calgon Havens Division
of Calgon Corporation
Gulf General Atomic Company and later Gulf Environmental
Systems Company
Aerojet-General Corporation and later Envirogenics Company
Aqua-Chem, Inc.
American Standard Inc., Conseps Div., now Division of
Abcor, Inc.
Westinghouse Electric Corporation
Osmonics, Inc.
Eastman Chemical Products, Inc., Subsidiary of Eastman
Kodak Company
"Permasep" Products Division, E. I. du Pont de Nemours
and Company
Supporting Equipment
Goulds Pumps, Inc. — Centrifugal pumps
A. W. Cash Valve Manufacturing Corporation — Back
pressure control valves
Sweco, Inc. — Vibrating screens
Manton-Gaulin Manufacturing Company — Reciprocating
pumps — Field unit
Milton Roy Company — Reciprocating pumps — Pilot scale
•3UO
-------
ACKNOWLEDGMENTS (Cont'd.)
Mr. Ralph H. Scott, Chief of Paper & Forest. Industries, Environmental
Protection Agency, as Federal Project Officer, has had responsibilities
for immediate supervision of this project in accordance vith objectives
of the Federal Research and Demonstration Grant Program. Mr. George R.
Webster, Water Quality Officer, and Mr. William J. Lacy, Chief, Indus-
trial Pollution Control Branch, have also contributed substantially to
the development and expediting of the Federal participation in this
project. Their help and also that of their associates at EPA is
acknowledged with sincere thanks.
-------
XII
REFERENCES
1, Wiley, A. j.» Jtamerlaan, A. C, F,» and Dubey, G. A. Tappi, 50,
no. 9:'fc55 (196?). ~^ ~~
2. Findlay, A. ' Osmotic^Pressure. Longmans Green and Co., New York,
(1913).
3. Merten, U. Desalination by Reverse Osmosis. M.I.T. Press,
Cambridge, Mass,, p, 1-10, 55-100, 161-200, 2lU-l8 (1966).
k. Spiegler, K. S. Principles of Desalination. Academic Press, lev
fork, p," 350-^00 (1966). "
5- Ammerlaan, A. C, F., Lueck, B. F., and Wiley, A. J. Tappi, 5_2_»
no, 1: 118 (1969).
6, Bansal, I. K., Dubey, G. A., and Wiley, A. J. Membrane Processes
in Industry and Biomedi c i ne, Edited by Milan Bier, Plenum Press
(1971).
7. Kopecek, J. and Sourirajan, S. Jour. Appl. Polymer Sci., 13.; 637
(1969).
8. Amnerlaan, A. C. F. and Wiley, A. J. Chem. Eng. Progr. Syarp.
Series., 65, (97): lU8 (1969).
9. Spent Sulphite lAquors — A Bibliography. The' Institute of Paper
Chemistry, Appleton, Wisconsin, 19^0, plus supplements to 1969.
10., Nelson, W. R. and Graves, J. f. A Fluosolids Spent liquor Recovery
' System for a'NSSC Pulp Mill. Presented before the June 15, 1967
meeting of the Central States Water Pollution Control Associa-
tion, St. Charles, 111.
11. Morris, D. C., Kelson, W. S., and Walraven, G. 0. The'Use of
'Reverse Osmosis in Treatment of Weak Semi chemical W_aste Waters.
Presented at the 26th PurdueIndustrial Waste Conference,May
U-6, 1971-
12., Gehm, I.'W. and Gove, G. W. ' Kraft Mill "Waste Treatment "in^the
'U.S. -'A Status Report, NCASI Tech. Bull. No. 221, p. U6, Dec.,
1968.'
13. Berger, H. F. ' 'Techn6lpj;Lcal'Trends, in Mill Effluent Color Reduction.
'Devatered 'Sludge 'Disposal and Kraft Mill Atmospheric 'Tjirrfssion
- Control, NCASI Tech. Bull. No. 228, p. 2-13, June. 1967.
3l»3
-------
(Continued)
lU. Massive Color Removal -Systems--Being Constructed by International
Paper Company. 'Southern'Pulp'Paper Mfr. 32, no. k: 26 (April
10, 1969).
15. Clarke, J. and Davis, M. W., Jr. 'Tappi, 52, no. 10: 1923 (Oct.,
1969) •
16. Tejera, N. E. and Davis, M. W., Jr. 'Tappi, 53_, no- 10:, 1931-4
(Oct., 1970).
17. Fuchs, R. I. Decolorization 'of'Pulp Mill Bleaching Effluents 'Using
Activated'Carbon. ICASI Tech. Bull. lo. 181, May. 1965.
18. Smith, D. R. and Berger, H. F. Tappi, 'gl, no. 10:. 37A (Oct., 1968).
19. Berger, H. F. Tappi, ]f£, no. 8: 80A (Aug., 1966).
20.. Diffusion Washers will Permit Complete lecycle. Can.'Chem.
Processing, 51, no. 12: ^9 (Dec., 1967),
21. Wiley, A. J. 'Progress in Developing Reverse Osmosis for Concentra-
tion of Pulp and Paper Effluents. Beport to International
Congress on Industrial ¥aste Waters, Stockholm, Sweden Nov. 2-6,
.1970.
22. Proceedings of 125th Meeting, Technical Advisory Committee of the
Environmental Research Group (Unpublished), The Institute of
Paper Chemistry, p. 95» Jan. 5 and 6, 1971.
23. Perona, J. J., Butt, P. H., Fleming, S. M., Mayr, S. T., Spitz, R.
A. 5 Brown, M. K., Cochran, H. D., Kraus, K. A., and Johnson, J.
S. , Jr., 'Environmental Science and Technol. ^L, no. 12: 991
(Dec., 1967).
2k. Aggarwal, J. P. and Sourirajan, S. Ind. Eng. Chem., 6l, no. 11: 62
(NOV., 1969).
25. Vos, K. D., Burris, F. D., and Riley, R. L. Jour. Appl. Polymer
Sci., 10: 825 (1966).
26. Wiley, A. J. , Dubey, G. A., Holderby, J. M., and Ammerlaan, A. C. F.
£• Water Pollution Control Federation, Part 2: R279 (Aug., 1970).
27. McCracken, D. D. A Guide to Fortran IV Programming, 5th printing,
John Wiley & Sons, Inc., New York, p. 23-7, 67-80,. Aug., 1967-
-------
REPEREICES (Continued)
28. Harris, P. L., et_ al_, Kaiser Engineers and Wong, C. M., et 'al,
Office of Saline Water, Office Saline Water Besearch and
Development Progress Beport No. 509., U.S. Government Printing
Office (Dec. 1969), Engineering 'and 'Economic, Eval;uation Study
of Reverse 'Osmosis.
29. Hittman Associates Inc. for Office of Saline Water, Office of
Saline Water Research and Development Progress Report No. 6ll,
U.S. Government Printing Office (Oct. 1970), Reverse Osmosis
Desalting"State-of-the Art (1969).
-------
SECTION XEII
PUBLICATIONS
The first paper in the following listing describes data obtained in an
experimental program preliminary to designing the program for this
Research and Demonstration Grant 120^0 EEL. The following four papers
were prepared and published on the basis of the data generated in the
grant.
At least two additional technical papers are requested for future prepa-
ration and delivery based on the fourth field demonstration for RO pro-
cessing of alkaline extraction kraft bleach effluents (see Section VII of
this report) and on optimization of RO plant design (see Section IX).
1. Wiley, A. J., Ammerlaan, A. C. F., and Dubey, G. A.
Application of Reverse Osmosis to Processing of Spent
Liquors from the Pulp and Paper Industry. TappjL, 50.,
no. 9: 1+55-60 (Sept., 1967).
2. Ammerlaan, A. C. F. , Lueck, B. F., and Wiley, Averill J.
Membrane Processing of Dilute Pulping Wastes by Reverse
Osmosis. Tappi, 52, no. 1: 118-22 (Jan.. 1969).
3. 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. Chem.
Eng. Progr. Symp. Series, 65, no. 97; 1^8-55» Water —
1969.
h. Wiley, Averill J., Dubey, George A., Holderby, J. M.,
and Ammerlaan, A. C. F. Concentration of Dilute
Pulping Wastes by Reverse Osmosis and Ultra Filtration.
Jour. Water Pollution Control Federation, ^2, no. 8,
Part 2: R279-89 (Aug., 1970),
5. Bansal, I. K. , Dubey, George A., and Wiley, Averill J.
Development of Design Factors for Reverse Osmosis
Concentration of Pulping Process Effluents. Reprinted
from Membrane Processes in Industry and Biomedicine,
Plenum Press, 1971-
-------
SECTION JCCV
GLOSSABY
1. Alkaline extraction stage KBE — The caustic extraction effluents from
the second and fourth stages of
bleaching in the example of the CEDED
sequence.
2. Bank/Stage — A group of nodules connected externally in series,
parallel or series/parallel arrange-
ments comprising a separate concen-
tration process.
3. BQDg — Biochemical oxygen demand based on the oxygen requirements of
living organisms over a 5-day period
while utilizing components of a
waste stream for growth and/or re-
production .
h. Ca-base sulfite pulp — Produced "by the calcium-base acid sulfite
process.
5, Chemimeehanical (CM) pulping process — A high yield process "based on
a short chemical cook followed by
mechanical refining to separate the
fibers in the softened chips.
6. Chlorination stage KBE — Some bleach sequences in multistage bleach-
ing may chlorinate two or more times,
but in this demonstration study it
usually refers to the first chlori-
nation stage as in CEDED (chlorina-
tion, caustic extraction, chlorine
dioxide, caustic extraction, and a
final stage of chlorine dioxide
bleach).
7. COD — Chemical oxygen demand is the measurement of the oxygen equiv-
alent of that portion of the organic
matter in a sample that is suscepti-
ble to oxidation by strong chemical
oxidants (chromic acid).
8. Compaction — Decrease of water permeation rate with time at a fixed
pressure.
9. Concentrate — The solution existing from the RO unit after removal
of a portion of the' water through
the membrane.
-------
GLOSSARY (Cont'd.)
10. Condensates — The condensible products from the evaporative.concen-
tration of pulping liquors of the
acid sulfite, bisulfite, NSSC and
alkaline sulfate (kraft) processes.
11. I)ynamic membrane — The formation of a layer on the' surface of an
intact membrane or other relatively-
impervious, but porous, surface
which decreases the transport (in-
creases the rejection) of solutes
without markedly affecting the trans-
port (flux rate) of the solvent.
12. Electrical resistance —A quantitative measure of the electrolytes
present in a solution.
13. Fluo-solids — Fluid solids process of burning to obtain dry ash
pellets.
lU. Flux rate — Rate of permeation or transport of the solvent (water)
through the membrane measured as
gallons per square foot per day
GF2/D or simplified to gfd.
15. Fouling — Stoppage of fluid flow (permeation) through the membrane
by a foreign material on the membrane
surface or within the membrane matrix,
as opposed to plugging which is a
blockage of the flow of the process
liquor along the module flow path.
16. Hydrolysis — Deacetylation of cellulose acetate membrane in a strong
acidic or alkaline medium.
IT- Inside diameter — Tubular RO systems in this study were of 0.5 inch
I.D.
18. KBE — Kraft bleach effluents, an effluent of the various single
stage or multistage methods of
bleaching kraft pulp.
19. Kraft pulp —Pulp produced in alkaline sulfate (kraft) process of
pulping.
20.. Lignin —An amorphous polymeric substance related to cellulose that
together with the cellulose forms the
woody cell walls of plants and the
cementing material between them.
350
-------
GLOSSARY (Cont'd.)
21. LLgnosulfonates — Compounds formed by the reaction of the bisulfite
(HSOa) ion and sulfurous acid in the
cooking liquor with the lignin in
the wood; present in the liquors as
salts of the base used in pulping.
22. Membrane constant —Flux rate/effective driving pressure.
23. NHa-base pulp — Produced by the ammonia-base acid sulfite process.
2i*. NSSC — Neutral sulfite-semichemical pulping, usually with sodium
bisulfite but other bases such as
ammonia are also used.
25. NSSC white water — Pulp wash water produced in "on-machine" washing
of linerboard pulp.
26. Optical density —Measurement of light absorption of an appropriate
dilution of the sample at 28l nm in
a 1 cm square silica cell, and re-
ported on basis of original, un-
diluted material.
27. Osmosis — Diffusion through a semipermeable membrane separating a
solvent and a solution that tend to
equalize their concentrations.
28. Osmotic pressure — The hydrostatic pressure required to stop the
osmotic diffusion across a semi-
permeable membrane between two solu-
tions of dissimilar concentrations.
29. Permeation resistance — (Osmotic pressure of the liquor) + (osmotic
pressure increase due to concentra-
tion polarization and fouling).
30. Plugging — Stoppage of the flow of process liquor through the
modules due to some foreign material,
suspended solids, scale, precipitate,
loose membrane in the flow path; as
opposed to fouling which is a block-
age of the flow of permeate through
the membrane.
31. Pressate — Fluid from the Zenith screw press in the chemimechanical
pulping process.
351
-------
GLOSSARY (Cont'd.)
32. Pressure drop — The loss of pressure through a system due to fric-
tion losses in piping and modules,
process stream velocity and Yiseosity,
33. Pressure pulse, hard — Rapid periodic and sharp reduction of the
pressure from an operating level to
atmospheric for the purpose of re-
.storing the flux rate lost lay the
deposition of a fouling material on
the membrane surface,
3l». Pressure pulse, soft — Periodic slow reduction of the pressure from
an operating level to atmospheric
for the purpose of restoring the flux
rate lost "by the deposition of a
fouling material on the menibrane
surface.
35. Pressure regulator — A spring or loaded, adjustable valve -which
can be used to set the pressure on
the upstream solution under process.
36. Becovery ratio — The ratio of the cmantity of a component (solids)
in the concentrate to the quantity
of the same component in the original
feed.
37• Recycle ratio — The ratio of the quantity of the feed liquor re-
cycled to the quantity of the liquor
entering a recycling operation as
fresh feed,
38. Reynolds number — A dimensionless number equalling (diameter of
pipe x average linear velocity of
the fluid x fluid density) •=- fluid
•viscosity.
39, RO — Osmosis in reverse flow through a semipermeable membrane when
external pressure in excess of the
osmotic pressure is applied.
l+O.. Semipermeable membranes — A menibrane which, is selective in that
certain components in a solution
(ordinarily the solvent) can pass
through the menibrane while one or
more components cannot.
352
-------
GLOSSARY (Cont'd.)
Ul. Spent liquor — Liquor separated from the pulp containing the resid-
ual cooking chemicals and dissolved
constituents of the wood.
1*2.. Stage/Bank -See Bank/Stage.
it3. Sulfite pulp — Pulp produced by one of the various modifications of
chemical and semichemical pulping
with solutions of sulfite and bi-
sulfites.
kk. Temperature coefficient — The rate of change in the flux rate
through a semipermeable membrane
with changes in the temperature of
the process stream.
*45. Turbulent flow — A flow in a fluid in which the velocity at a given
point varies erratically in magnitude
and direction.
1*6. Velocity — Linear velocity of flow across the membrane surface.
353
-------
SECTiOI XV
APPENDICES
Photographs, Fig. 6kt 65, and 66.
355
-------
Figure 6k. Pretreatment Section in Building Adjacent to Trailer-Mounted
RO Field Test Unit. Ca-Base Pulping Wash Water Screened in Column
at Left and pH Adjustment Carried out in Agitated Day Tank in
Center, with Flow Stabilized in Second Day Tank at Right.
Flow Rate About kO Gal./Min. (Refer to Page 95)
356
-------
*
Figure 65. RO Pilot Unit Operating in Pulp Mill to Process
Alkaline Extraction Stage Kraft Bleach Effluent.
Northwest Paper Company, Cloquet, Minnesota
357
-------
Figure 66. Photo of RO Pilot Installation Operating to Concentrate
Dilute Chemimechanical Pulp Wash Water. Large 5000 Gallon Feed
Tank in Center Background with Sweco Vibrating Screen and
Base of Zenith Press Above. Pretreatment System in Foreground,
Cooling in Tube-Type Heat Exchanger at Right» pH Adjustment
in Day Tanks Center Foreground, and Pump Test Stand
on Wheels in Left Rear
358
•US. GOVERNMENT PRINTING OFFICE: 1972 484-486/243 l-J
-------
Accession Nuniher
Subject Field &•. Group
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
Pulp Manufacturers Research League and The Institute of Paper Chemistry, Effluent
Processes Group, Division of Industrial and Environmental Systems
Title
Reverse Osmosis Concentration of Dilute Pulp and Paper Effluents
1 0 1 Authcr(s)
Wiley,
Averill J., Project
Director
i>t Project Designation
EPA WQ Contract No.
12040 EEL
21 Note
Dubey, George A.
Bansal, I. K.
22
Citation
23
Descriptors (Starred First)
*Reverse Osmosis, *Pulp Wastes, *Waste Treatment, Water Reuse, Chemical
Recovery, Economic Feasibility, Technical Feasibility, Membrane Fouling.
25
Identifiers (Starred First)
*Waste Water Treatment, ^Industrial Wastes, *Semipermeable Membranes, Separation
Techniques, Desalination, Water Costs, Membrane Cleaning.
27
Abstract
Adaptation of reverse osmosis as a method of concentration for dilute effluents
of pulping, bleaching, and paper manufacture was conducted in laboratory, pilot
scale, and in large 50,000 gallon per day field demonstrations at pulp mills.
Most of these dilute wastes at 1 percent solids contained suspended particles,
colloidal suspensoids, large molecular-weight wood derived organics, and/or
scale-forming inorganic chemical residues. Tubular membrane systems capable
of being operated at self-cleaning velocities increasing beyond 2.0 feet per
second, as concentration advanced to 10 percent solids, were apparently best
adapted to processing these effluents at sustained high flux rates and relatively
free of. fouling problems. Capillary fiber and spiral wound sheet membrane
systems required expensive clarification treatment before and during concentra-
tion. Tubular systems studied were subject: to excessive failure rates in terms
of life of membrane support structures or to leakage of internal connections
based on the support structure. Feasibility of employing RO for concentration
of dilute pulping and bleaching effluents depends on developing routes to sub-
stantial improvement in life expectancy of RO equipment to maintain high flux
rates and rejections at much lower membrane maintenance and replacement costs
than prevailed with equipment available for these studies conducted from 1967
through 1971.
Abstractor
Institution
WR:102 (REV. JULV 18S9)
WRSIC
SEND. WITH COPV OF DOCUMENT. TOt WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OP THE INTERIOR
WASHINGTON. D. C. 20240
* OPS: IB70-383-J30
------- |