EPA-660/2-73-019
DECEMBER 1973
                        Environmental Protection Technology Series
    Color Removal from
    Kraft  Mill Effluents
    by Ultrafiltration
                                  Office of Research and Development
                                  U.S. Environmental Protection Agency
                                  Washington, D.C. 20460

-------
            RESEARCH REPORTING SERIES
Research reports of the  Office  of  Research  and
Monitoring,  Environmental Protection Agency, have
been grouped into five series.  These  five  broad
categories  were established to facilitate further
development  and  application   of   environmental
technology.   Elimination  of traditional grouping
was  consciously  planned  to  foster   technology
transfer   and  a  maximum  interface  in  related
fields.  The five series are:

   1.  Environmental Health Effects Research
   2.  Environmental Protection Technology
   3.  Ecological Research
   4.  Environmental Monitoring
   5.  Socioeconomic Environmental Studies

This report has been assigned to the ENVIRONMENTAL
PROTECTION   TECHNOLOGY   series.    This   series
describes   research   performed  to  develop  and
demonstrate   instrumentation,    equipment    and
methodology  to  repair  or  prevent environmental
degradation from point and  non-point  sources  of
pollution.  This work provides the new or improved
technology  required for the control and treatment
of pollution sources to meet environmental quality
standards.

-------
                                    EPA-660/2-73-019
                                    December 1973
           COLOR REMOVAL FROM KRAFT MILL

           EFFLUENTS  BY ULTRAFILTRATION
                         by

                   H. A.  Fremont
                    D.  C.  Tate
                  R. L.  Goldsmith
                  Project S800261
              Program  Element 1B2037
                  Project Officer

              Mr. Edmond P.  Lomasney
          Environmental  Protection Agency
              Atlanta, Georgia 30309
                 prepared for the

        OFFICE OF RESEARCH AND DEVELOPMENT
       U.S.  ENVIRONMENTAL PROTECTION  AGENCY
              WASHINGTON, D.C. 20460
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price $i*i

-------
              EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents necessarily
reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or
commercial products constitute endorsement or recommenda-
tion for use.

-------
                        ABSTRACT
Reduction of color in pulp mill effluents by ultrafiltration
has been examined with a 10,000 gallon per day  (gpd) pilot
plant.  Treated streams included Decker effluents and pine
bleachery caustic extraction filtrate, which together com-
prise about 80% of the color from a bleached kraft mill.

High color removal (90-97%) was demonstrated when operating
at water recovery ratios of 98.5-9970.  Pilot plant capacity
(membrane flux) was 15-20 gal./day-ft^ when operation pro-
ceeded smoothly.  However, plugging of the membrane car-
tridges by residual particulates (even after precoat fil-
tration) was troublesome.

Several prefiltration, concentrate disposal, and water
reuse alternatives were evaluated.

Full-scale plant designs,  and approximate capital and
operating costs were estimated for systems of 1 and 2 MM
gpd capacity.   Capital costs are about $700,000 for a 1 MM
gpd plant,  and $1,200,000 for a 2 MM gpd plant. Corresponding
operating costs are about 45c/Mgal. (1 MM gpd) and 38c/Mgal.
(2 MM gpd).

Additional pilot plant tests are recommended to demonstrate
long-term solutions to the particulate problem and long-
term membrane cartridge life.

This report was submitted in fulfillment of Project Number
S800261,  by Champion International Corp. under the partial
sponsorship of the Environmental Protection Agency.  Work
was completed as of May, 1973.
                        111

-------
                     TABLE OF CONTENTS

                                                     Page
   I.   SUMMARY AND CONCLUSIONS                         1

        A.   NATURE OF PROBLEM                          1

        B.   NATURE OF STREAMS TREATED                  1

        C.   PROCESS CONCEPT                            2

        D.   PROGRAM RESULTS                            3

        E.   FULL-SCALE PLANT DESIGN AND PROCESS COSTS  5

  II.   RECOMMENDATIONS                                 7

IIII.   BACKGROUND                                      9

        A.   NATURE OF PROBLEM AND PROJECT GOALS        9

        B.   PROCESS DESCRIPTION                       12

  IV.   PILOT PLANT DESCRIPTION                        25

        A.   GENERAL                                   25

        B.   PRETREATMENT SEQUENCES                    28

        C.   OTHER SYSTEM MODIFICATIONS                30

        D.   DESCRIPTION OF MEMBRANE CARTRIDGES        33

        E.   SAMPLING AND ANALYTICAL PROCEDURES        50

   V.   RESULTS AND DISCUSSION                         53

        A.   FEED PRETREATMENT AND CHARACTERISTICS     53

        B.   REJECTION DATA AND EVALUATION             80

        C.   ULTRAFILTRATION RATE DATA AND
            EVALUATION                                88

        D.   PRESSURE DROP DATA                       115

-------
                     TABLE OF CONTENTS
                       (continued)


                                                     Page
        E.  IMPORTANT FACTORS CONTROLLING MEMBRANE
            FLUX                                     121

        F.  CLEANING PROCEDURES AND EFFICIENCY       125

        G.  MODULE MECHANICAL FAILURES               129

        H.  INCINERATOR STUDIES                      134

  VI.  FULL SCALE PLANT DESIGN AND COSTS             139

        A.  FIRST STAGE PINE CAUSTIC EXTRACTION
            FILTRATE                                 139

        B.  DECKER EFFLUENTS                         173

 VII.  WATER REUSE                                   189

        A.  PINE CAUSTIC EXTRACTION FILTRATE
            PERMEATE                                 189

        B.  DECKER EFFLUENT PERMEATES                189

VIII.  ACKNOWLEDGEMENTS                              191

  IX.  REFERENCES                                    193

        APPENDIX A - DETAILED PILOT PLANT PROCESS
                     DESCRIPTION                     197

        APPENDIX B - ULTRAFILTRATION PILOT PLANT
                     SPIRAL MEMBRANE CARTRIDGE
                     IDENTIFICATION                  217

        APPENDIX C - DETAILS OF FILTERS USED IN
                     PILOT PLANT PROGRAM             223

        APPENDIX D - MECHANICAL PROBLEMS OF DIFFERENT
                     SPIRAL MEMBRANE CARTRIDGES      231

        APPENDIX E - SLIME ANALYSIS                  237
                             VI

-------
                        FIGURES

                                                    Pa^e

 1.  SIMPLIFIED ULTRAFILTRATION FLOW  SCHEMATIC      13

 2.  CAUSTIC EXTRACTION FILTRATE—FLOW SCHEMATIC    18

 3.  TYPICAL MATERIAL BALANCE FOR TREATMENT OF
     CAUSTIC EXTRACTION FILTRATE  BY ULTRAFILTRATION 19

 4.  DECKER EFFLUENTS—FLOW SCHEMATIC               21

 5.  TYPICAL MATERIAL BALANCE FOR TREATMENT OF
     HARDWOOD DECKER EFFLUENT BY  ULTRAFILTRATION    22

 6.  SIMPLIFIED PILOT PLANT FLOW  SCHEMATIC          26

 7.  FEED PRETREATMENT FLOW SCHEMATIC               29

 8.  BAUER HYDRAS IEVE                                34

 9.  TOP OF 500 GAL. FEED TANK                       35

10.  ACID PUMP AND ACID DRUM                         36

11.  DETERGENT MIXING TANK AND PRECOAT MIXING
     TANK WITH MIXER                                 37

12.  PRETREATMENT SEQUENCE                           38

13.  DETAILS OF SPARKLER FILTER                      39

14.  SHRIVER FILTER PRESS                            40

15.  100 GAL. FILTRATE SURGE TANK                   41

16.  SINGLE CARTRIDGE AND FILTERS                   42

17.  SPARKLER VELMAC DISC FILTER                     43

18.  ULTRAFILTRATION UNIT                            44

19.  ULTRAFILTRATION UNIT  (TOP VIEW)                 45

20.  DETAILS OF CONTROL PANEL RIGHT SIDE            46

21.  DETAILS OF CONTROL PANEL LEFT SIDE             47
                          VII

-------
                        FIGURES
                        (continued)
                                                    Page
22.  SPIRAL WOUND MODULE DESIGN                      48

23.  SULFURIC ACID REQUIREMENT TO  NEUTRALIZE
     CAUSTIC EXTRACTION FILTRATE                     54

24.  SULFURIC ACID REQUIREMENT TO  NEUTRALIZE
     PINE DECKER EFFLUENT                            55

25.  SULFURIC ACID REQUIREMENT TO  NEUTRALIZE
     HARDWOOD DECKER EFFLUENT                        56

26.  EFFECT OF pH ON COLOR OF PINE CAUSTIC
     EXTRACTION FILTRATE                             58

27.  SUSPENDED SOLIDS OF CAUSTIC EXTRACTION
     FILTRATE                                        70

28.  SUSPENDED SOLIDS OF PINE DECKER EFFLUENT        71

29.  SUSPENDED SOLIDS OF HARDWOOD  DECKER EFFLUENT    72

30.  TOTAL SOLIDS OF PINE CAUSTIC  EXTRACTION
     FILTRATE AND PINE DECKER EFFLUENTS              73

31.  TOTAL SOLIDS OF HARDWOOD DECKER EFFLUENT        74

32.  COLOR OF PINE CAUSTIC EXTRACTION FILTRATE
     AND PINE DECKER EFFLUENT                        75

33.  COLOR OF HARDWOOD DECKER EFFLUENT               76

34.  EFFECT OF FEED PRETREATMENT ON COLOR OF
     PINE CAUSTIC EXTRACTION FILTRATE                78

35.  REJECTION AND CONVERSION DATA                  81

36.  MEMBRANE FLUX FOR STAGE la                      94

37.  MEMBRANE FLUX FOR STAGE Ib                      95

38.  MEMBRANE FLUX FOR STAGE Ic                      96

39.  MEMBRANE FLUX FOR STAGE 2                       97
                          Vlll

-------
                        FIGURES
                       (continued)
                                                    Page

40.  MEMBRANE FLUX FOR STAGE  3                       98

41.  MEMBRANE FLUX FOR STAGE  4                       99

42.  MEMBRANE FLUX FOR STAGE  5                      100

43.  EFFECT OF PREFILTRATION  EFFICIENCY  ON
     ULTRAFILTRATION RATE                           102

44.  ULTRAFILTRATION RATE DATA:  STAGE Ib           108

45.  ULTRAFILTRATION RATE DATA:  STAGE 2           110

46.  FLUX VS. TIME FOR STAGES la AND  Ib             114

47.  STAGE 1 PRESSURE  DROP                          116

48.  STAGE 2 PRESSURE  DROP                          117

49.  STAGE 3 PRESSURE  DROP                          118

50.  STAGE 4 PRESSURE  DROP                          119

51.  STAGE 5 PRESSURE  DROP                          120

52.  CLEANING EFFICIENCY OF STAGE  Ib  MEMBRANES     130

53.  CLEANING EFFICIENCY OF STAGE  3 MEMBRANES      131

54.  STAGE 3 COMPACTION CURVE                       132

55.  SIMPLIFIED FLOW SCHEMATIC:  TREATMENT OF
     PINE CAUSTIC EXTRACTION  FILTRATE              140

56.  FLOW SCHEMATIC FOR TREATMENT  OF  PINE  CAUSTIC
     EXTRACTION FILTRATE:  LOW FLOW CASE           141

57.  FLOW SCHEMATIC FOR TREATMENT  OF  PINE  CAUSTIC
     EXTRACTION FILTRATE:  HIGH FLOW  CASE           142

58.  SIMPLIFIED FLOW SCHEMATIC:  TREATMENT OF
     DEKCER EFFLUENTS                               174
                           ix

-------
                         TABLES
 1.   WASTE CHARACTERISTICS AT THE NORTH CAROLINA
     MILL                                            11

 2.   FILTERS USED                                    31

 3.   CHARACTERISTICS OF MEMBRANE  CARTRIDGES          49

 4.   ANALYTICAL PROCEDURES                           51

 5.   PARTICLE SIZE DISTRIBUTION OF  SUSPENDED
     SOLIDS IN PINE CAUSTIC EXTRACTION FILTRATE     61

 6.   PERFORMANCE - MAIN FILTERS                      63

 7.   TESTS WITH DIFFERENT FILTER  AIDS  ON  SPARKLER
     15.3 SQ FT LEAF FILTERS                         67

 8.   SUMMARY OF FEED CHARACTERISTICS                 69

 9.   WASTE CHARACTERISTICS AT THE NORTH CAROLINA
     MILL - DETAILED ANALYSES                        79

10.   REJECTION                                       82

11.   COMPOSITION DATA:  PINE DECKER EFFLUENT
     SAMPLES                                         85

12.   COMPOSITION DATA:  HARDWOOD  DECKER EFFLUENT
     SAMPLES                                         86

13.   COMPOSITION DATA:  PINE CAUSTIC EXTRACTION
     FILTRATE SAMPLES                                87

14.   MEMBRANE FLUX BY STAGE DURING  PILOT  PLANT
     PROGRAM                                         89

15.   EFFECT OF FEED PREFILTRATION ON  CLEANING      127

16.   CASES FOR CAPITAL COST ESTIMATES, PINE
     CAUSTIC EXTRACTION FILTRATE                    145

 7.   CHARACTERISTICS OF MEMBRANE  CARTRIDGES        147

-------
                        TABLES

                       (continued)
                                                    Page

18.  CHARACTERISTICS AND COSTS OF  MEMBRANE MODULES  148

19.  ULTRAFILTRATION SECTION DESIGN  DETAILS AND
     COSTS—CASE 1                                   153

20.  ULTRAFILTRATION SECTION DESIGN  DETAILS AND
     COSTS—CASE 2                                   158

21.  SUMMARY OF INSTALLED CAPITAL  COST ESTIMATES
     PLANT TO TREAT PINE CAUSTIC EXTRACTION
     FILTRATE                                        162

22.  OPERATING COSTS FOR TREATMENT OF  PINE CAUSTIC
     EXTRACTION FILTRATE, CASE 1                     163

23.  OPERATING COSTS FOR TREATMENT OF  PINE CAUSTIC
     EXTRACTION FILTRATE, CASE 2                     164

24.  OPERATING COSTS FOR TREATMENT OF  PINE CAUSTIC
     EXTRACTION FILTRATE, CASE 3                     165

25.  OPERATING COSTS FOR TREATMENT OF  PINE CAUSTIC
     EXTRACTION FILTRATE, CASE 4                     166

26.  OPERATING COSTS FOR TREATMENT OF  PINE CAUSTIC
     EXTRACTION FILTRATE, CASE 5                     167

27.  DAILY INCREMENTAL OPERATING COSTS,  TREATMENT
     OF PINE CAUSTIC EXTRACTION FILTRATE            168

28.  CASES FOR CAPITAL COST ESTIMATES, DECKER
     EFFLUENTS                                       175

29.  INSTALLED CAPITAL ESTIMATES—PLANT TO TREAT
     DECKER EFFLUENTS                                176

30.  OPERATING COSTS FOR TREATMENT OF  DECKER
     EFFLUENT, CASE 6                                178

31.  OPERATING COSTS FOR TREATMENT OF  DECKER
     EFFLUENT, CASE 7                                179
                           XI

-------
                        TABLES

                      (continued)
                                                    Page
32.  OPERATING COSTS FOR TREATMENT OF DECKER
     EFFLUENT, CASE 8                               180

33.  INCREMENTAL OPERATING COSTS  FOR TREATMENT
     OF DECKER EFFLUENT, CASE  9                     181

34.  INCREMENTAL OPERATING COSTS  FOR TREATMENT
     OF DECKER EFFLUENT, CASE  10                    182

35,  DAILY INCREMENTAL OPERATING  COSTS,  TREATMENT
     AND REUSE OF DECKER EFFLUENTS                 183
                           xi i

-------
                    SECTION I

             SUMMARY AND CONCLUSIONS
A.  NATURE OF PROBLEM

    At present, in kraft paper mills color pollutants are
    not substantially removed by conventional biological
    waste treatment methods.   The available means, such as
    lime precipitation and carbon or resin adsorption are
    highly expensive.

    In the Champion Papers' North Carolina mill, approx-
    imately 607o of the mill color discharge is contained in
    the bleach plant caustic extraction filtrate, and 20%
    in the Decker effluents.   These numbers are fairly typical
    for a bleached kraft mill.  Thus, reduction of color from
    these two streams can greatly reduce the overall color
    discharge from a kraft mill.

    The purpose of this program has been to examine ultra-
    filtration as a means of reducing color in kraft mill
    effluents more efficiently and/or more economically than
    the presently available methods.  The scope of the
    program included the six month operation of a 10,000 gpd
    pilot plant at the Champion Papers' North Carolina mill.

B.  NATURE OF STREAMS TREATED

    The major experimental effort dealt with treatment of
    pine caustic extraction filtrate, with lesser emphasis
    placed on pine and hardwood Decker effluents.  All
    three streams are highly alkaline (pH 10-12) and hot
    (120°F-135°F), and require neutralization (to pH7)
    and cooling (to 100°F-105°F) before treatment by
    ultrafiltration.  These limits are imposed by the
    characteristics of current cellulose acetate
    membranes, which will not exhibit long (economical)
    life if exposed to high pH and high temperature.

-------
       These streams also contain substantial quantities of
       particulates (e.g. 100-300 ppm),  of which 50% are
       smaller than 10/j.   To obtain acceptable membrane
       equipment performance it is essential to employ mem-
       brane equipment (configuration) which is not susceptible
       to plugging by particulates.  In addition, flocculation
       occurs in a freshly filtered feed on "aging".

   C.   PROCESS CONCEPT

       Ultrafiltration, can selectively concentrate color bodies
       and organics contained in the streams.  A simplified
       flow schematic is shown below:

                                          Organic
                               	O Concentrate
                                          5-20% organics
                                          0.5-2% ash
                                 ,:^!
    Feed to	^
Ultrafilter        \f^^T'      ^	f	Semipermeable
0.1-0.2% organics     "       TMembrane
0.4-0.6% ash
                                           Decolorized
                                           Permeate
                                           —0.01% organics
                                           0.3-0.5% ash
       Desirable features of the process (compared to reverse
       osmosis) are:

       o  Since only 5-30% of the dissolved solids are
          returned in the concentrate, very high water
          recovery (e.g. 99%) can be achieved;
       o  A low-volume, high-organic-solids concentrate
          is obtained, which has substantial heating value
          if ultimate disposal is by burning;
       o  Low-pressure systems (e.g. 100 psig) can be
          employed; and
       o  High capacity can be achieved since higher-flux
          membranes can be used.

-------
    Limitations of the process are:

    o  The effluent is not demineralized,  and the
       residual salt content (especially chlorine
       in caustic extraction filtrate) limits reuse
       potential; and
    o  Color removal is typically 90-97%,  somewhat
       lower than can be achieved by reverse osmosis.

    For all three streams examined,  pretreatment (neutra-
    lization, temperature reduction, and filtration) is
    required.

D.  PROGRAM RESULTS

    A pilot plant was operated from August, 1972 through
    February, 1973.  Operation was normally 24 hrs/day,
    7 days/week.  During this period four experimental
    aspects of the process were evaluated:

    o  Feed pretreatment (especially filtration);
    o  Ultrafiltration (separation efficiency and capacity);
    o  Concentrate disposal (incineration); and
    o  Water reuse potential.

    Feed pretreatment work focused primarily on means to
    remove particulates, by surface, packed-bed, and pre-
    coat filtration, to obtain satisfactory ultrafiltration
    performance.  Using filtration which removed particles
    of about 2p, and larger, 50-80% particulate removal was
    achieved.

    The ultrafiltration system membranes were in a spiral
    wound configuration, and were obtained from three
    vendors (T. J. Engineering, Gulf Environmental Systems,
    and Eastman Chemical Products).  Most membrane cartridges
    had the standard "mesh" flow-channel spacer; a few had
    "corrugated" spacers.  Operation was usually at  about
    100°F, 100 psig, and pH 6-7.

    Color removal efficiency was satisfactory; typical
    results are:

-------
                      7. Water       % Solids In      % Color
Influent              Recovery      Concentrate      Remova1

Pine caustic
extraction filtrate    98.5-99         15-20          90-92

Pine Decker
effluent               98.5-99         5-8            95-97

Hardwood Decker
effluent               98.5-99         5-8            95-97

    Capacity (membrane flux) was variable.  When operation
    proceeded smoothly, fluxes of 15-20 gallons/day-ft *
    of membrane were achieved.  At other times, fluxes were
    substantially lower.  Important factors which reduced
    flux were:

    o  Reversible membrane surface fouling by colloidal
       and macromolecular feed constituents.  This was
       reversed by water flushing and detergent cleaning;
    o  Irreversible membrane "compaction".  This was
       minor; and
    o  Reversible and irreversible particulate collection
       within the membrane cartridges.

    The latter limitation is the most serious problem encoun-
    tered in the test program.  Due to incomplete particulate
    removal in the pretreatment operation, particulates
    collected in the membrane cartridges.  This was especially
    severe for cartridges with "mesh" spacers, but not for
    the "hydrodynamically-clean" corrugated spacers.
    Manifestations of cartridge plugging were:

    o  Occlusion and inactivation of membrane surface,
       with reduced capacity;
    o  High pressure drop across the cartridges which
       resulted in:  (a) cartridge deformation and
       (b) seal failure;
    o  Flow maldistribution within cartridges and
       between cartridges in parallel, which aggravated
       cartridge plugging by particulates; and
    o  Various modes of module mechanical failures.

    Particulate plugging can be minimized substantially by:

-------
    o  Operation at high flow,  which prevents particulate
       collection;
    o  Regular and efficient cartridge cleaning; and
    o  Use of cartridge flow-channel spacers which are
       not susceptible to particulate collection (e.g.
       corrugated spacers).

    Concentrate disposal for pine caustic extraction filtrate
    by incineration or evaporation and admixture with primary
    sludge was demonstrated.  Return of the concentrates
    from Decker effluents to the weak black liquor system
    was chosen as an optimum means of disposal.

    Water reuse potential was examined.  Analyses of treated
    effluents and comparison to mill water standards led
    to the following conclusions; treated effluent from
    pine caustic extraction filtrate has limited or no
    reuse potential due to a high chloride content and
    some residual color; treated water from the Decker
    effluents may be beneficially reused in pulp washing.

E.  FULL-SCALE PLANT DESIGN AND PROCESS COSTS

    A series of design cases were examined.  Design para-
    meters included:

    o  Plant capacity (1 MM gpd and 2 MM gpd);
    o  Membrane type (mesh and corrugated spacers); and
    o  Different prefiltration alternatives.

    The key cost factor was plant capacity, as expected,
    since treatment costs are nearly proportional to the
    volume treated and not the contaminant loading.

    The installed capital costs are estimated to be:

                          Pine Caustic            Decker
                       Extraction Filtrate      Effluents

                            $770,000             $690,000

                          $1,250,000           $1,100,000

-------
The difference is due to concentrate disposal equipment
required for the pine caustic extraction filtrate.

The total operating costs (including amortization)
are estimated to be:

                    Pine Caustic            Decker
Flow             Extraction Filtrate       Effluents

1 MM gpd           46C/M gal. =            44(?/M gal. =
                   $0.58/ton bleached      $0.33/ton total
                   pulp
2 MM gpd           35C/M gal. -            37
-------
                    SECTION II

                 RECOMMENDATIONS
The results of the project studies using the nominal 10,000 gpd
ultrafiltration pilot plant demonstrate that the processing
system is technically sound as a means for reducing color
from the pine caustic extraction filtrate of a pulp bleachery
and from the Decker effluents from pulp washing.  Mechanical
difficulties, however, were encountered which should be
resolved.  These problems and unknown variables are highly
critical in developing reliable cost estimates for full scale
installations.

It is recommended that pilot plant studies be continued
for the purpose of confirming the present results over a
longer time span, and further examining the cost sensitive
areas of the system.  More specifically, these additional
studies could accomplish the following tasks;

1.  Examine membrane cartridges from several manufacturers
    for long-term reliability.

2.  Verify optimum operating conditions such as feed flow
    rates, pH, temperatures and operating pressures.

3.  Further specify prefiltration conditions for each
    influent.

4.  Test any new membranes that may become commercially
    available and which could afford the elimination of
    the neutralization and cooling treatments of the
    influent.

5.  Obtain more extensive operating experience using Decker
    effluents.

6.  The information from items 1 through 5 above could
    result in redesigning a full scale plant and assessing
    more reliably capital and operating cost estimates for
    such a plant.

-------
                           SECTION III

                           BACKGROUND



A.  NATURE OF PROBLEM AND PROJECT GOALS

    The 121 kraft pulp mills in the United States produce
    about 85% of the chemical wood pulp consumed.  In these
    pulping operations a substantial volume of wastewater
    is discharged, typically about 25,000 gallons per ton
    of pulp.  Of concern are the pH, temperature, BOD, and
    color loading of this effluent.  Conventional and
    generally inexpensive techniques are adequate for waste
    treatment except for color removal.  It may also be noted
    that conventional waste treatment does not provide for
    water reuse or chemical recovery and reuse, and, as such,
    is not conducive to eventual close- loop operations.

    Color bodies found in pulp mill wastes unfotunately are
    resistant to biological degradation.  The effluent color is
    due primarily to lignin and its degraded products which are
    chemically stable and are intractible to separation by
    presently proven commercial processes.  Consequently, new
    treatment techniques for color removal are undergoing
    active development and actual plant scale demonstration.
    Promising processes developed include chemical precipitation
    (1-10) , including lime precipitation  (stoichiometric and
    massive lime), adsorption (11-14), and reverse osmosis and
    ultraf iltration (15-25) .  For controlling color from
    bleaching effluents, it is also possible to. modify the
    bleaching sequence to CHE .... from CEH . . . ; where C=
    chlorination, H =hypochlorite bleaching, and E = caustic
    extraction.

    Segregation of mill wastes is often practiced and it is
    likely that segregation of waste streams by color will
    eventually be required for adequate waste treatment.
    Tertiary treatment systems which could not be considered
    cost-wise for treatment of the total  effluent, might be
    applicable if the bulk of the color is contained in a
    relatively small fraction of the mill effluent.
    For example, in Champion Papers' North Carolina m
    about 60% of the mill color is present in 2 x 10b gpd

-------
of bleachery first-stage pine caustic extraction fil-
trate.  This flow amounts to only about 4% of the
total wastewater.  Thus, if color could be removed
from this stream at total operating costs of $450 to
$800 per day, over 60% of the mill color could be re-
moved at a cost of about 0.9$ to 1.6C/1000 gal. of
total effluent or about 55$ to $1.00 per ton of
bleached pine pulp.

The second most important controllable source of color
in a kraft mill is the decker effluent.  This waste is
present in all kraft mills, while the pine caustic
extraction filtrate is found only in mills producing
bleached pulp.  At the North Carolina mill, approxi-
mately 2 x ID** gpd of mixed pine and hardwood decker
effluents are currently discharged.  This waste con-
tributes about 20% of the mill color.  Thus, decker
effluents are another case of interest for segregated
waste treatment.

Ranges of flow and composition for these streams at the
North Carolina mill are given in Table 1.

The program described in this report encompasses the
initial phase of a two-phase project for the develop-
ment and demonstration of ultrafiltration as a means of
color removal from kraft mill effluents. The site of the
project is Champion Papers' North Carolina mill which
is a representative bleached kraft mill.  Phase I of
the project  (the subject of this report) included on-
site operation of a nominal 10,000 gpd membrane demon-
stration plant to supplement data from previous pilot
tests and to specify more accurately design bases for a
full-scale demonstration plant.  Additional studies
determined requirements for feed pretreatment prior to
ultrafiltration, disposal of the concentrated wastes
produced by ultrafiltration, and the potential for water
reuse.

Briefly the project objectives have been threefold.

    1.  To demonstrate with commercially-available
equipment the effectiveness of ultrafiltration to re-
duce color in first-stage pine caustic extraction
filtrate and decker effluents to low levels;
                            10

-------
                           TABLE 1
       WASTE CHARACTERISTICS AT THE NORTH CAROLINA MILL
Flowrate* ,
Million
Waste
Pine Caustic
Extraction
Filtrate

Pine Decker
Effluent


Hardwood
Decker
Effluent

Gal/Day
1.3 to
2.1


0.5 to
2.7;
Avg. =
0.5
1.5 to
4.1;
Avg . =
1.5
pH
11.5
to
11.8

10.7
to
11.0

10.0
to
10.6

Color, ppm
12,000 to
50,000;
Avg. ^
28,000
5,000 to
10,000;
Avg. "*>
6,000
4,000 to
24,000;
Avg. ^
11,000
Total
Solids,
ppm
4,000 to
14,000;
Avg. ^
7,000
1,500 to
6,000;
Avg. i>
2,400
1,200 to
5,000;
Avg. ^
3,000
Suspended
Solids,
ppm
50 to
Avg. ^


50 to
Avg. 'v


100 to
Avg. ^



500;
80


250;
80


500;
200


* maximum flow is future projection
                               11

-------
        2.   To examine the  potential for reuse  of purified
    effluents  and means of  disposal of the concentrated
    wastes  produced by the  membrane process;  and

        3.  To demonstrate that process economics will
    be sufficiently attractive to lead to widespread adop-
    tion of this  pollution  abatement process  in the
    industry.

B.  PROCESS DESCRIPTION

1.  Principles of Ultrafiltration

    Ultrafiltration is a membrane process for concentration
    of dissolved  materials  in aqueous solution.  A semi-
    permeable  membrane is used as the separating agent and
    pressure as the driving force.  In an Ultrafiltration
    process (Figure 1), a feed solution is fed  into the mem-
    brane unit, where water and certain solutes pass through
    the membrane  under an applied hydrostatic pressure.
    The solutes whose sizes are larger than the pore size
    of the  membrane are retained and concentrated.  The
    pore structure of the membrane thus acts  as a molecular
    filter, passing some of the smaller solutes and re-
    taining the larger solutes.  The pore structure of this
    molecular  filter is such that it does not become plugged
    because the solutes are rejected at the surface and do
    not penetrate the membrane.  Furthermore, there is no
    continuous buildup of a filter cake which has to be
    removed periodically to restore flux through the mem-
    brane since concentrated solutes are removed in solu-
    tion.  Many Ultrafiltration applications  involve the
    retention  of  relatively high molecular weight solutes
    accompanied by the removal through the membrane of lower
    molecular  weight impurities.  Thus concentration of
    specific solution components can be achieved.

    Considerations important for determining  the technical
    and economic  feasibility of Ultrafiltration as applied
    to a specific process are the rate of solution trans-
    port through  the membrane  (flux) and the  separation
    efficiency (rejection).  Factors which control flux
    and rejection have been described elsewhere  (.?_§.r22).

    Membrane processes for  treating pulp mill wastes have
    been under development  for several years, focusing
    primarily  on  reverse osmosis.  In reverse osmosis all
                               12

-------
                                      PERMEATE

                                                '  '        .   .
                                             ' '  ' '      *       MEMBEANE
                                              • •  •  •   -   _   X
FEED                Q   O   .  * O *. ^  . o  9  • o' • O' O • Op O. O      CONCENTRATE

                              -     c  .   -o --
                             '                     '
                                   '  •     -
                                       .o
                                      '

                    .   •   • , '  '      .  *   I.   *    •  •  •   I-   .  ^MEMBRANE
                     •   *
                -  Dissolved  Salts


                O Color  Bodies






               FIGURE 1:  SIMPLIFIED ULTRAFILTRATION FLOW SCHEMATIC

-------
dissolved solutes are concentrated, and a demineralized
aqueous effluent is produced.  Ultrafiltration is, in
fact, a variation of reverse osmosis.  The fundamental
difference relates to the retention properties of the
membranes:  ultrafiltration membranes do not retain
salts and other low molecular weight solutes.

There are several potential advantages of ultrafiltra-
tion when compared to reverse osmosis, which are:

        a.  The use of ultrafiltration membranes leads
    to operation at lower pressures, typically 50 to
    300 psig.  The strength requirements of the mem-
    brane system are much less stringent than those for
    reverse osmosis syterns, which typically operate
    above 400 psig.  At the lower pressures used in
    ultrafiltration, power costs are less; membranes
    can last longer since the rate of "compaction" may
    be reduced;  and lower capital costs can be
    achieved due to less demanding requirements for
    pressure vessels, pumps, etc.  Reliability is also
    an important consideration, and system failure ob-
    served in reverse osmosis tests to date can be tied
    to the high operating pressures used.  Important
    factors have been membrane compaction  (resulting
    in reduced capacity), membrane catastrophic
    failure (membrane support rupture), and pump
    failure.

        b.  With reverse osmosis the feed pine caustic
    extraction filtrate or decker effluent can be con-
    centrated only to 8 to 10% total solids, or about
    20-fold.  This is due to limitations of both mem-
    brane fouling and a buildup in feed solution
    osmotic pressure.  Both problems are of substan-
    tially less importance for ultrafiltration, in which
    very high volumetric concentration ratios are
    obtainable, up to or exceeding 200-fold.  This is
    due primarily to a relatively slow increase in feed
    solids content with concentration by ultrafiltration,
    That is, only a small portion of the feed solids,
    specifically the higher molecular-weight organics,
    is retained by the ultrafiltration membrane.  Note
    that about 80% of the dissolved solids in the wastes
    of interest are low molecular-weight salts.  Since
    the retained solutes have relatively high molecular
    weights, osmotic pressure limitations are of minor
                            14

-------
    importance, and solids levels up to 20% can be
    achieved in low-pressure ultrafiltration.

        c.  Disposal cost of an ultrafiltration con-
    centrate by incineration or other means would be
    substantially less than that for a reverse osmosis
    concentrate, since a substantially smaller volume
    must be treated.  For example, in treating
    2 x 106 gpd, about 100,000 gpd of reverse osmosis
    concentrate would be generated, but only 10,000
    gpd of ultrafiltration concentrate.  Furthermore
    the high organic content of the latter provides
    substantial heating value, almost sufficient in
    itself to sustain combustion.  A major fuel cost
    would be required to burn a reverse osmosis con-
    centrate which contains primarily inorganics.  In
    addition a high chlorine content in a reverse
    osmosis concentrate of pine caustic extraction fil-
    trate could cause severe corrosion problems during
    incineration, and would require extensive off-gas
    scrubbing to remove volatile chloride particulates.

        d.  Membrane life data in the field in other
    applications has shown that ultrafiltration mem-
    branes are less susceptible than reverse osmosis
    membranes to deterioration in flux and rejection
    due to alkaline hydrolysis and/or compaction under
    pressure.

The major disadvantage of ultrafiltration vis-a-vis re-
verse osmosis lies in the quality of the aqueous efflu-
ent produced by the membrane system.  Two factors are of
importance.  First, since ultrafiltration does not
demineralize that waste treated, the effluent does con-
tain residual salts.  As discussed below, this should
not present any problem in the treatment of decker efflu-
ents.  However, for pine caustic extraction filtrate these
residual salts can limit the reuse potential of the
water.  Second, color rejection in ultrafiltration.is
somewhat lower than that in reverse osmosis. The greater
residual color in the effluent from an ultrafiltration
system can be a limiting factor.

A review of the costs of commercially-available membrane
equipment at the beginning of the project showed that
spiral wound membranes would be the lowest in cost.
Although some process limitations exist for membranes
                            15

-------
    in this  configuration,  it  was  felt that  this  system
    would be adequate  for processing both pine  caustic
    extraction filtrate and decker effluents.   The  advan-
    tages and disadvantages of other membrane configura-
    tions have been discussed  elsewhere (^6).   The  two
    alternatives,  hollow fine  fiber and tubular configura-
    tions, are not thought  to  be attractive.  The former is
    extremely susceptible to plugging by suspended  solids,
    and the  latter is  a relatively high cost system.   For
    these reasons  the  present  study has focused solely on
    the use  of membranes in the spiral wound configuration.

2-  Need for Pilot Plant Studies

    In a previous  program sponsored by Champion Papers
    (unpublished)  the  necessity of conducting in-field pilot
    studies  was apparent.  In  the preliminary program,
    several  hundred gallon samples were shipped to pilot
    plant facilities for ultrafiltration tests.  Data for
    membrane flux, membrane rejection efficiency, and
    material balances  were  obtained.  However,  factors which
    could not be evaluated  were:

            a)  differences in treating "fresh" (minutes-
        old) and "aged" (days-old) feed materials;
            b)  pretreatment requirements for particulate
        removal to obtain stable long-term operation;
            c)  operability of a staged membrane  system
        as a means of  continuously producing a high-solids
        concentrate;  that is, with feed "conversion" to
        permeate of 99+%;
            d)  effective techniques for membrane cleaning
        as a means to  sustain  high-flux in long-term
        operation;
            e)  membrane flux  and life in intermediate-term
        tests (six months). Note that long-term life tests
        were not a part of this program;
            f)  requirements for pH and temperature control,
        and  performance at different pH, temperature, pres-
        sure and feed  flow conditions;
            g)  examination of a statistically meaningful
        number of  membrane modules to determine failure
        modes; and
            h)  system operability in the field.

    Furthermore, feed  pretreatment  (filtration) data could
    only be  obtained with fresh samples since particulate
                                16

-------
    level and size were known to change with aging for
    pulp mill effluents.

    Finally, variability in flow and composition of the
    wastes of interest, extremely  important parameters in
    process design, had to be determined at the mill.

    For these reasons, a pilot plant program was clearly
    warranted to realistically determine process technical
    and economic feasibilities.

3.   Pine Caustic Extraction Filtrate;  Flow Schematic
    and Typical Material jjalance

    Figure 2 shows a  flow schematic for the treatment of
    pine caustic extraction filtrate by ultrafiltration.  In
    the bleaching of  pine pulp, the pulp flows through a
    series of stages.  First is chlorination, with the
    aqueous waste discharged to drain.  In the second stage,
    the pulp is extracted with caustic, and it is the efflu-
    ent from this operation which  contains the bulk of the
    color discharged  from the bleaching system.  The  pulp is
    subsequently treated with calcium hypochlorite, and in.
    other bleaching and washing stages.

    One means of reducing the color discharge from a  bleach
    plant is to reverse the hypochlorite and caustic  extrac-
    tion stages.  Through this sequence a substantial frac-
    tion of the color is oxidized  by hypochlorite, but at a
    cost of increased chemical usage.

    In the proposed process, the traditional CEH  ...  bleach-
    ing sequence is used and the pine caustic extraction
    filtrate is processed by an ultrafiltration  system
    (Figure 2).  As described above, a membrane  is  selected
    which specifically concentrates the dissolved  organics
    and color bodies.  A purified, but not demineralized,
    water is obtained for reuse or disposal.  Reuse would be
    limited to non-pulping operations because of the  high
    chloride content.  The organic concentrate  is  disposed
    of either by evaporation to approximately  50%  solids,
    with subsequent ultimate disposal on  land, or  by  inciner-
    ation.  A typical material balance for  the  ultrafiltra-
    tion operation is shown  in Figure  3.

    By obtaining 90-96% color removal  from  the  pine  caustic
    extraction filtrate,  approximately 50-60% of the  total
                                17

-------
                    Chlorine
                     Water
 Caustic
   Calcium
Hypochlorite
     Pine
     Pulp
                 Chlorination
                    Stage
                     Drain
 Caustic
Extraction
  Stage
                                                 Other
                                                Bleaching
                                                 Stages
00
  Calcium
Hypochlorite
   Stage

     I
   Drain
                                                                              Organic  Concentrate
                                                                              Disposal by  Evaporation
                                                                                 or  Incineration
                                                           Ultrafiltration
                                                              Sy a tern	
                                                                               Purified Water for
                                                                               Reuse or Disposal
                     FIGURE 2:   CAUSTIC EXTRACTION FILTRATE—FLOW SCHEMATIC

-------
                                                                                 CONCENTRATE
                                                                                    1.67  Ibs
                                                                                1,300,000 ppm  color
                                                                                14.5%  total solids
                                                                              4,020  ppm  total chloride
                                                                              1,590  ppm  ionic chloride
            FEED -
          100 Ibs
VO
36,700 ppm color (unneutralized)
23,600 ppm color (neutralized)
0.91% total solids (neutralized)
785 ppm total chloride
705 ppm ionic chloride
                                                MEMBRANE
                                                                                PERMEATE
                                                                                98.33 Ibs
                                                                              1,920 ppm color
                                                                              0.68% total solids
                                                                            730 ppm total chloride
                                                                            690 ppm ionic chloride
                       FIGURE 3:   TYPICAL MATERIAL BALANCE FOR TREATMENT OF
                                CAUSTIC EXTRACTION FILTRATE BY ULTRAFILTRATION
                                       (Neutralization with H2SO.)

-------
    mill color will be removed.  In addition, about 2 x 10
    gpd of water will be available for potential reuse
    within the mill.  If the water cannot be reused and is
    instead sewered to the waste treatment system, the
    organic loading (as B.O.D.) of the waste treatment
    system will be reduced by approximately 207o.

4.  Decker Effluents:  Flow Schematic and Typical
    Material Balance

    The treatment of decker effluents is the other case of
    interest.   While bleach plant effluents are found only
    in bleached kraft mills, decker effluent is found in
    all kraft mills.  Thus,  the treatment of decker efflu-
    ent by ultrafiltration is an application of broader
    interest to the industry.  Referring to Figure 4, wood
    chips are digested and the resulting pulp is transferred
    to a blow tank.  The black liquor removed from the blow
    tank is processed in evaporators for chemical recovery.
    The pulp from the blow tank goes to a series of washers.
    At present, treated water is used in the washers, with
    the spent wash water processed in the black liquor
    recovery system.  The pulp is then washed in a final
    decker operation, again with the addition of treated
    water.  The decker effluent is too dilute to be processed
    with the black liquor,  and is discharged to the waste
    treatment plant.  At the North Carolina mill,  approxi-
    mately 2 x 106 gpd of treated water is added to the
    decker.

    In the proposed process, the decker effluent is concen-
    trated in an ultrafiltration system.  The purified water
    is recycled to the final washer,  or other pulp system uses,
    eliminating the need for adding treated water.  This
    purified decker effluent could also be used at other
    points in the mill.   The organic concentrate from the
    ultrafiltration system is processed in the black liquor
    recovery system.  A typical material balance for the
    ultrafiltration operation is shown in Figure 5.  Note
    that the color removal  efficiency is better than for
    the pine caustic extraction filtrate case.   This is
    because the color bodies are lower molecular weight for
    the latter, since substantial lignin fragmentation occurs
    in the bleach plant chlorination stage.

    This treatment of decker effluent by ultrafiltration has
    several desirable features.   These are
                               20

-------
    Wood
    Chips
f
            Digester
ro
H
    To
  Black
 Liquor ""^
Evaporators
       Chemicals
                            (other pulp
                            mill usage)
                        Blow
                        Tank
                        To
                      Black  <-s£
                     Liquor  *^
                   Evaporators
                                                                      Treated
                                                                       Water
                                  Pulp
                                                Water
                                                                       1
                                           Washers
                                                   Decker
                         (hardwood
                         and pine)
                                          Organic
                                         Concentrate

                                         To Black
                                          Liquor
                                       Evaporators
                                                  Ultrafilte:
                                                                            Purified
                                                                           —Watrex	
                            FIGURE 4:  DECKER EFFLUENTS—FLOW  SCHEMATIC

-------
   NEUTRALIZED
       FEED  	
     100 Ibs
0.26% total solids
11,000 ppm color
MEMBRANE
                                                                    CONCENTRATE
                                                                       2 Ibs
                                                                    6.8% total solids
                                                                    467,000 ppm color
                                                                 -  PERMEATE
                                                                     48 Ibs
                                                                  0.13% total solids
                                                                  500 ppm color
               FIGURE 5:  TYPICAL MATERIAL BALANCE  FOR TREATMENT OF
                        HARDWOOD DECKER EFFLUENT  BY ULTRAFILTRATION
                              (Neutralization with H-SO.)

-------
    --the water usage, and subsequent treatment require-
      ment in the waste treatment system, is reduced
      by 2 x 106 gpd;
    --approximately 20% of the color in the total mill
      effluent is removed;
    --salts ordinarily lost in the decker effluent are
      recovered and recycled to the black liquor
      evaporation system;
    --organics ordinarily lost in the decker effluent
      are returned to the black liquor system and have
      a corresponding heating value; and
    --the organic loading (as B.O.D.) to the existing
      waste treatment system is reduced by approximately
      20%.

As described in detail below, the technical feasibility
of treating both pine caustic extraction filtrate and
decker effluents has been demonstrated.   Projected costs
for both cases are thought to be attractive.  For the
treatment of pine caustic extraction filtrate, ultra-
filtration costs will be less than those for lime preci-
pitation, activated carbon adsorption, or utilization of
the alternative bleaching sequence.  For the treatment
of decker effluents, a positive return on investment
might be realized.
                            23

-------
                          SECTION IV

                   PILOT PLANT DESCRIPTION
A.  GENERAL

    A generalized flow schematic is shown in Figure 6.  A
    more detailed flow schematic and process description is
    contained in Appendix A.  Referring to Figure 6, feed
    material, either pine caustic extraction filtrate, pine
    decker effluent, or hardwood decker effluent was piped
    from the mill to a 500-gallon Fiberglas feed tank.
    Temperature, pH, and level were controlled at this point.
    Neutralized feed from the feed tank was pumped through
    a filter(s) for removal of suspended solids, and then to
    the ultrafiltration unit.  Details of the various filtra-
    tion sequences employed are given below.

    In the ultrafiltration system color bodies and organics
    were concentrated and removed as a low-volume concentrate,
    while water and salts were removed as permeate.  During
    operation, feed, concentrate, and permeate flows were
    measured, sampled, and analyzed, as well as other inter-
    mediate process flows.

    In the ultrafiltration part of the pilot plant, five
    recirculated "stages-in-series" were used.  This design
    was selected so that high conversion with high color
    removal efficiency could be achieved  (see Appendix A
    for details).  Spiral wound membrane cartridges were
    installed in seven housings  ("shells"); three shells were
    used in Stage 1; and one in each of Stages 2 through 5.  A
    single shell held either three 24-inch  long cartridges
     (T.J. Engineering, Downey, California;  or Eastman
    Chemical Products, Kingsport, Tennessee) or two 36-inch
    long cartridges  (Gulf Environmental Systems Co., San
    Diego, California).  Cartridges from all three companies
    were used in the pilot plant operation.  A detailed
    listing identifying cartridge use by stage and time is
    given in Appendix B.
                                            2
    Membrane areas were approximately 300 ft  for Stage 1,
    and 100 ft2 for each of Stages 2 through 5.  Total mem-
    brane area was about 700 ft^.  At a nominal membrane flux
    of 15 gal./day-ft   (gfd), the pilot plant capacity was
    10,500 gpd.
                               25

-------
                             Recirculation Flow
                                             (3 si ells)
               Booster
                Pump
Stage 1  Filter
 Pump
Stage 1
permeate
Sampling points:
1)  raw feed
2)  feed and concentrate for each
   stage (10 points)
3)  permeates from each shell
   (7 points).
4)  mixed permeate
                                                 Recirculation Flow
                                                                 Stage 2
                                                                  Pump
                                                                             (1 sh
                                                                all each)
                                                                 Stage 2,3/4
                                                                  permeate
Three stages  connected
      in series.
                                                                                                                     Concentrate
                                                                                                                          Flow
                                                                                Stage 5
                                                                                 Pump
                                                'Stage 5
                                                 permeate
                                                                                                                      Mixed
                                                                                                                     Permeate
                              FIGURE 6.  SIMPLIFIED PILOT PLANT FLOW SCHEMATIC

-------
After pilot plant start-up in mid-August, 1972, it became
apparent that the pre-filtration system provided with
the pilot plant was inadequate for the removal of sus-
pended solids.  The specific manifestations of inadequate
filtration were a rapid increase in pressure drop across
the membrane cartridges, and a drastically reduced ultra-
filtration rate with time.  In addition, it was suspected
that membrane fouling by colloidal and dissolved macro-
molecular materials was also a problem, although definitely
of lesser importance than the flow-channel plugging created
by particulates.  Several changes were made to the pilot
plant, including:

     1.  the installation of a single test cartridge
         which was used for tests to determine means of
         controlling the plugging and fouling problems;
     2.  the installation of more effective filters;
     3.  piping modifications to allow high-capacity,
         once-through water flushing of all stages; and
     4.  the installation of piping and tanks to allow
         detergent cleaning of membrane cartridges.

The sequence of changes can be summarized as follows:

     Mid-August startup period;  The originally installed
     Broughton lOy basket filters were used.

     Early September:  A single cartridge was installed
     for special tests to examine the plugging/fouling
     problem.  Also installed were a Bauer Hydrasieve
     filter for the removal of coarse fibers, and ly
     Cuno cartridge filters for more complete removal
     of suspended solids.

     Mid-September;  A Hydromation depth filter was put
     into use upstream of the Cuno polishing filters to
     prolong the polishing filter element life.

     Mid October:  A detergent cleaning  system was  inte-
     grated into the pilot plant.  The Hydromation  filter
     was replaced with a  Shriver filter  press  to obtain
     increased capacity  (prolonged filtration  cycles).

     End of November;  Flush valves and  the necessary
     piping for once-through flushing of all membrane
     shells was installed.

     December:  An automatic filter aid  feeder was  in-
     stalled  to permit unattended operation of the  pre-
                            27

-------
         coat filter over 24-hr periods.

         Early February:  The Shriver filter press was replaced
         by a Sparkler leaf filter to obtain more "representa-
         tive" filtration data.  A Sparkler Velmac disc filter
         was substituted for the Cuno polishing filters to
         examine a "backwashable" polishing filter.

    Changes in the flow diagram that resulted from the equip-
    ment modifications are shown in detail in Appendix A
    (Figure A. 2).

B.  PRETREATMENT SEQUENCES

    A more detailed flow schematic for the pretreatment opera-
    tion is shown in Figure 7.  Raw feed was piped from the
    mill to the 500-gallon feed tank.  During part of the test
    period a screen (Bauer Hydrasieve) was used to remove
    coarse fibers and particulates from the raw feed prior
    to introduction into the feed tank.

    As indicated, acid from a 55-gallon drum was pumped by
    an acid pump into the feed tank for neutralization.  For
    the bulk of the program, sulfuric acid was used, but for
    a two-week period, hydrochloric acid was added.

    When pre-coat filtration was employed, a continuous filter
    aid addition unit (BIF Industries screw-type adder) intro-
    duced filter aid directly into the feed tank, where
    vigorous mixing was achieved with a Lightnin mixer.

    Neutralized/ cooled feed was pumped by a filter pump
    through one of three filters, a Hydromation depth filter,
    a Shriver filter press/ or a Sparkler leaf filter.  The
    filtrate was held in a 100-gallon surge tank  (two 55-
    gallon drums).  In general/ flow into this surge was
    through a float valve which kept the surge tank full.
    Thus, the filtrate flow was controlled by the subsequent
    rate of consumption in the pilot plant.  At time,s ex-
    cess filtrate was drawn off and either sewered or re-
    turned to the 500 gal. feed tank.  When a precoat fil-
    ter was being used, a precoat suspension was mixed in
    a 55 gal. drum connected to the suction of the filter
    pump.  Filtrate was returned to the precoat mixing
    tank until an adequate precoat was developed.

    Filtrate from the filtrate surge was pumped into the
    suction of the Goulds multistage Stage 1 pump.  The
                                28

-------
               Raw
               Feed
        Feed Sampling Point

excess filtrate
                            Drain
                                           Broughton
                                               lOy
                                             Basket
                                             Filter
     55 Gal.
 Stage 1
Membranes
                                                   Polishing
                                                     Filters
                             Pretreated
                            Feed Sample
      FI6UBE 7.  FEED PRETKEATMENT FLOW SCHEMATIC

-------
    outlet from the pump passed through two Broughton nominal
    lOu basket filters.  Flow was then through polishing fil-
    ters (Cuno cartridges or Sparkler Velmac pot filter) to
    remove residual particulates.  To avoid the buildup of
    an excessive pressure drop across the polishing filters,
    a differential pressure switch was installed.  Excess
    pressure drop resulted in system shutdown.  The final
    polished filtrate was introduced into the three parallel
    membrane shells of Stage 1.

    A summary of the location in the process flow and the
    characteristics of the filters is given in Table 2.  Manu-
    facturer's information on the filters is contained in
    Appendix C.

C.  OTHER SYSTEM MODIFICATIONS

1.  Single Cartridge Tests

    The single spiral wound cartridge was used for tests to
    characterize prefiltration efficiency.  Raw feed was
    filtered by various means and processed in the single
    cartridge.  The buildup in pressure drop across the
    cartridge and decrease in ultrafiltration rate were con-
    sidered to reflect prefilter performance.

    The single cartridge was connected in parallel to the
    Stage 1 shells, and fed by the Stage 1 pump  (Appendix A,
    Figure 199).  With this arrangement it was possible to
    operate the single cartridge simultaneously, or inde-
    pendently of, the pilot plant.  The permeate and con-
    centrate from the single cartridge were either sewered
    or returned to the 100 gal. filtrate surge tank.

2.  Detergent Cleaning System

    A 55-gal. barrel served as a mixing tank for cleaning
    solutions.  The tank outlet was connected to the suction
    of the booster pump (Figure 7).  When cleaning, the
    concentrate(s) and permeate(s) from the pilot plant
    were returned by hoses to the detergent solution tank,
    allowing the indefinite recirculation of detergent solu-
    tion through the system.  This eliminated the need to
    prepare large volumes of cleaning solutions.

3.  Once-through Flushing

    The original piping of the pilot plant did not permit
                              30

-------
                                             TABLE'2

                             FILTERS USED  (IN ORDER OF FLOW SEQUENCE)
   Filter
Location
Description
                              Function
1. Bauer Hydrasieve
in raw feed line
to 500 gal. feed tank
Curved wire screen/ 6 in.
wide, 2 ft long/ screen
openings approx. 0.01 in.
(see Fig.  8 )•
                              to remove coarse
                              fibers and solids
2.  Hydromation
   Depth Filter
after filter pump/
before filtrate
surge tanks
Granular PVC depth filter,
with automatic backwashing.
1 ft2 cross-sectional area.
                              fine filtration to
                              remove particles of
                              about 1v and larger
3.  Shriver
   Filter
   Press
                        24 plate filter press;
                        16 in. x 16 in. plates;
                        total area of about 75 ft2.
                        Cloth used initially to hold
                        filter aid; subsequently
                        switched to paper sheets
                        (see Fig.  14) .
4. Sparkler
   Leaf
   Filter
                        Vertical leaf filter with
                        15.3 ft  area.  Leaves with
                        316 ss wire mesh, covered
                        with nylon bags, in carbon
                        steel vessel.   Contained
                        high-pressure spray nozzles
                        for cake removal (see Fig. 13).

-------
      Filter
               TABLE 2  (continued)

      FILTERS USED  (IN ORDER OF FLOW SEQUENCE)

Location                Description
                              Function
   5. Broughton
      Basket
      Filters
after Stage 1 pump/
before Stage 1
membranes
Two model 3000 basket fil-
ters with nominal lOy
stainless steel mesh
screens.  Total area
5.6 ft  .  Included an
automated differential-
pressure-controlled back-
washing sequence.  A third
basket  filter  (200 mesh
screen,  ) was used to filter
water for backwashing (see Fig. 16).
to remove suspended
solids of 10 y and
larger -(this was
the initial filter
provided with the
pilot plant)
   6. Cuno
K>     Cartridge
      Filters
after Broughton
filter, before
Stage 1 membranes
Two parallel filter
housings, each containing
two cartridges.  Disposible
1-5v cartridges in stain-
less steel housings.  Area
per filter of about 1.6 ft2
for Micro-klean II elements
and 1 ft2 for Micro-Wynd II
elements.  Parallel filters
permitted element change
without system shutdown.
final polishing
filtration before
membrane elements
   7. Sparkler Velmac
      Disc Filter
in place of
Cuno filters
Contained backwashable
polypropylene circular
discs* 5 v retention,
15 ft2 area.

-------
    once-through flushing of the membrane shells,  i.e.  the
    piping connected each stage to the next.   It was decided
    that cleanup of plugged membrane cartridges could be
    facilitated by once-through flushing with water. Conse-
    quently,  a water line was connected to the suction  of the
    pump for  each stage/  and flush valves were piped into the
    concentrate lines from each stage.  This  arrangement per-
    mitted once-through flushing of each shell in  a "forward-
    flow" direction.  A subsequent piping modification  allowed
    once-through "reverse-flow" flushing.

    Photographs of various system components  are contained
    in Figures 8-21.

D.  DESCRIPTION OF MEMBRANE CARTRIDGES

    The spiral wound cartridges consist of a  membrane envelope
    and mesh  spacer that have been rolled around a PVC  tube
    (see Figure 22).  One or more cartridges  in series  can
    be installed in a housing, which serves as a pressure
    vessel.  The process fluid passes through the  open  space
    provided  by the mesh spacer.  The permeate produced spirals
    through the porous backing inside the membrane envelope
    into the  PVC center tube.  A seal between the  OD of the
    cartridge and the ID of the shell prevents the process
    liquid from bypassing the cartridge.

    For the pilot plant program, 7-ft long, 4-in diameter
    epoxy-coated steel shells were used.  For the  special
    tests with a single cartridge, a 30-in long, 4-in diameter
    PVC shell was employed.

    Four different types of cartridges were purchased from
    three different suppliers.  All membranes were cellulose
    acetate.   More specific information on the cartridges
    can be found in Table 3.

    The Gulf  cartridges had the highest salt  rejection  and
    lowest flux.  These were therefore used primarily in
    Stages 4  and 5,  where the greatest loss of color into
    the permeate could occur and where the lowest  flux  was
    expected  due to high feed concentrations.  One set  of
    three cartridges with corrugated flow channel  spacers
    was purchased for the purpose of comparing this type of
    spacer with the standard mesh spacer.  The former is
    less susceptible to plugging by suspended solids.
                                33

-------
                         Fig. 8 - Bauer Hydrasleve
  Bauer
Hydrasieve

-------
                                     Feed
                                     Line
Solids
Adder
                                                 Cooling
                                                  Coil
                                                Connection
              Mixer
                              Acid
                              Line
Fig.  9 -
  Left:
of 500 Gal
Feed Tank
                                                             mixer and  body  feeder
                                                    Center:  acid and raw  feed  lines
                                                    Right:   connections to  pli  probe
                                                             and heating/cooling  coil

-------
Fig.  10 - Acid Pump and Acid Drum

-------
Detergent Mixing Tank (fore
ground) and Precoat Mixing
Tank with Mixer (rear).
                             Precoat
                              Mixing
                               Tank
                Detergent
                 Solution
                  Tank

-------
                                     Temperature
                                     Controller
Sparkler
 Filter
                           12  -Pre treatment  Sequence  (in
                            art) .
                           From left  to  right:   Sparkler
                           filter;  filter  pump;  bottom of
                           500-gal.  feed tank;  temperature
                           controller and  indicator; booste

-------
 Flush
Connections
        Fig.  13  -  Details of Sparkler Filter
            Pressure  gauges,  filtrate flow
            meter,  flush connections.

-------
Fig. 14 - £i               Press

-------
 Level
k.Switch
Float
Valve

                                                    ^^•^•^•^••^
                     Fig-  15  -  100 Gal.  Filtrate Surge Tanl
                        Two 55-gal.  drums  connected at the
                        bottom  to  provide  100-gal.  surge
                        volume;  level  switch  (left) and
                        level control  valve  (right; volume
                        ahead of the ultrafiltration  unit.

-------
             Broughton
              Basket
              Filters
  Single
Cartridge
                                                - Single Cartridge and Filters
                                           Single cartridge in PVC shell
                                           deft); Broughton basket filter
                                           (center); Cuno filter  (right).

-------
7 -  Sparkler Velmac  Disc  Filter
         Sparkler
          Velmac

-------
Fig. 18 - Ultrafiltration Unit:
   Center tier
Left:        control panel
Top tier:    circulation pumps,
             Stages 2 through 5.
             shells with membrane
             cartridges, Stages 2
             through 5
Bottom  tier: Stage  1 pump  and
             membrane cartridges
             (flush valves  in-
             stalled)
                                            irculation
                                             Pumps
                                           Stages  2-5
                                          Membrane
                                           Shells
                                         Stages 2-5
                                       Membrane
                                        Shells
                                       Stage  1

-------
.  19 -.UltraflUr'ablon  Unit:  Crop
       view")
Stages 2  through  5  pumps  and  shells
with membrane cartridges  in  fore-
ground

-------
Main switch,  pump switches, alarms
outlet pressure gauges and control
valves,  Stage 1-5

-------
                                                     ontroller
 Recirculation
 Flow Meters
                                                      pH, Temp.
                                                      Recorders
                                                 Inlet
                                               Pressure
                                                Gauges
                                              Sampling
                                               Valves
                                                           Permeate
                                                            Flow
                                                            Meters
concentrate
 Flowmeter
                      Mixea
                     Permeate
                     Flowmeter
                  Control Panel
                                47
       Left Side
Stages 1-5 recirculation flow
meters, inlet pressure gauges
(Stages 1-5), and control valves
(Stages 2-5); permeate flow meter,
one for each of the shells (7);
pH controller with indicator; pH
and temperature recorders.	

-------
                      Brine  Seal
              ROLL TO
              ASSEMBLE
 FEED SIDE
 SPACER
PERMEATE OUT
     PERMEATE SIDE BACKING
     MATERIAL WITH MEMBRANE ON
     EACH SIDE AND GLUED AROUND
     EDGES AND TO CENTER TUBE
                                                           FEED FLOW
                                   PERMEATE FLOW
                                   (AFTER PASSAGE
                                   THROUGH MEMBRANE)
                                                                                                                        SEE DETAIL A
                                                                          MESH SPACER
MEMBRANE
                                                         FIGURE  22.   SPIRAL
                                                       WOUND  MODULE  DESIGN
                                                                                                       BACKING MATERIAL
                                                                                                                      PERMEATE TUBE
                                                                                                                   GLUE LINE
                                                                                               DETAIL A

-------
               TABLE 3
CHARACTERISTICS OF MEMBRANE CARTRIDGES
SUPPLIER
model no.
membrane type
spacer material
length of
cartridge,
inches
number per shell
membrane area
per cartridge,
nominal salt
rejection, %
nominal flux,
gfd at
100 psig, 70° F.
number purchased
number of brine seals
brine sealing point
during normal
operation
Gulf Environ-
mental Systems
4000
UF
mesh
36
2
50
approx .
30"
7-10
4
1
upstream
TJ
Engineering
UF-H-32
Eastman
HT 00
mesh
24
3
32-35
0-10.
35-50
45
2
downstream
TJ
Engineering
UF
Eastman
HT 00
corrugated
24
3
12-15
0-10
35-50
3
2
downstream
Eastman
Chemical Products
MK 00 A
HT 00
mesh
24
3
32-35
0-10
35-50
3
1
upstream

-------
E.  SAMPLING AND ANALYTICAL PROCEDURES

    Referring to Figures 6 and 7, the following samples were
    collected:
         —raw  (untreated) feed;
         —neutralized and filtered feed  (feed to Stage 1);
         —final concentrate  (Stage 5 concentrate);
         —mixed permeate  (from all 5 stages);
         —feed samples for all stages  (5); and
         —permeate samples for all stages  (5).

    Analytical procedures are detailed  in  Table 4.  The analyses
    for pH, total solids, suspended solids  and color of the
    raw and pretreated feed, final concentrate and mixed per-
    meate were performed on a regular basis.  Occasionally,
    analyses of feed and permeate samples  from individual
    membrane stages were also carried out.  A few samples
    were analyzed for other constituents  (ash, various ions,
    and suspended solids composition and particle size dis-
    tribution) to better characterize performance of the sys-
    tem.

    Suspended solids analyses were performed by the standard
    gravimetric method.  However, it was found that only
    relatively small samples, 50-100 ml, could be filtered
    through the 0.45y Millipore membrane filters before
    blinding occurred.  Thus, the suspended solids analyses
    were only approximate, and poorly reproducible, since
    the total amounts of suspended solids present in the
    samples were small, and difficult to measure accurately.

    A turbidimeter and a nepholometer were also tried to
    assay for suspended solids, but both instruments were
    found to lack sensitivity because of strong light ab-
    sorption by the highly colored samples.
                                50

-------
Analysis For

PH

Color, Cobalt units
                      TABLE  4

               ANALYTICAL PROCEDURES

                              Procedure *
                            Standard  pH meter

                            Absorbance at  465 mja
                            compared  to absorbance of
                            standard  Pt/Co.  solution,
                            No.  118 at pH  7.6, "Field"
                            color measured at as-is pH.

                            Gravimetric, No. 148A

                            Gravimetric, No. 148B

                            Gravimetric, No. 148C

                            Gravimetric, No. 148D

                            Calibrated chloride ion
                            electrode, No. 203C

                            Flame Photometric Method,
                            No.  153A

                            EDTA Titrimetric, No. 67C

                            Gravimetric, No. 238A

                            Colorimetric,  No. 144B

Particle Size Distribution  Sizing on filters fllowed
of Suspended Solids         by analysis of fraction
                            passing through  44u filter
                            with Coulter counter.
Total Solids

Total Dissolved Solids

Suspended Solids

Ash

Ionic Chloride


Na+


Ca-H-

So =

Fe
^Referenced number is the method in Standard Methods
for Examination of Water and Wastewater, APHA, 13th
Edition,  1971.
                        51

-------
                          SECTION V

                    RESULTS AND DISCUSSION



A.  FEED PRETREATMENT AND CHARACTERISTICS

    The pilot plant was operated on pretreated pine caustic
    extraction filtrate for most of the test program, and
    on hardwood and pine decker effluents for a limited
    period during January , 1973.

    Before introduction into the ultrafiltration section of
    the pilot plant all three effluents were neutralized
    to a pH of 6 to 7 , cooled to a temperature of 95 to
    105 °F, and filtered to remove suspended solids.  A
    description of the pretreatment section of the pilot
    plant was given in Section III.B.  Results pertinent
    to the evaluation of the pretreatment operations are
    discussed below.

1.  Temperature

    Cellulose acetate ultrafiltration membranes show
    accelerated hydrolysis and compaction  {irreversible
    loss of flux due to collapse of the porous membrane
    structure) with increasing temperature/ and especially
    when exposed to temperatures in excess of 110 °F.  All
    three effluents treated in the pilot plant are generally
    discharged at a temperature of 120-135 °F/ depending on
    mill operating conditions and the season.  Therefore,
    it was necessary to cool the feed, which was done by
    passing cold water through the coil heat exchanger in
    the feed tank.  In the course of the experimental pro-
    gram, the operating temperature was not intentionally
    varied, and most data are for operation at a feed tempera-
    ture between 95 and 105 °F.

2.
    Cellulose acetate membranes show satisfactory life when
    operated within a pH range of 3 to 7.  All three efflu-
    ents, however, are highly alkaline, with pH's ranging
    from pH 10 to pH 12.  Therefore, the effluents were
    neutralized to a pH of 6 to 7, by addition of sulfuric
    acid.  Figures 23 to 25 contain pH curves for the neutral-
    ization of pine caustic extraction filtrate, pine decker
                              53

-------
          12J
          10
                                      BASIS:   100 ml of Caustic Extraction Filtrate
                        •\
                           \
u.
     H
           2 .
                                mm
                                   V
                                      V
                                        \
                                234

                                   Amount of  0.5N I^SC^ added,  ml

                        FIGURE  23:   SULFURIC ACID REQUIREMENT  TO NEUTRALIZE
                                            CAUSTIC EXTRACTION  FILTRATE

-------
     11
                                      BASIS:  200 ml of Pine Decker Effluent
PH
     10
                                                               3/22/73
                                                                                                 18
                          Amount of 0.1N H2SO4 added/ ml



               FIGURE 24.  SULFURIC ACID REQUIREMENT TO NEUTRALIZE PINE DECKER EFFLUENT

-------
  11.
                                             BASIS:   200 ml of Hardwood Decker Effluent
  10
PH
                                                    -rir
-rt-
-te
                               Amount of 0.1N H-SO, added, ral



               FIGURE 25.   SULFURIC ACID REQUIREMENT TO NEUTRALIZE HARDWOOD DECKER EFFLUENT

-------
    effluent,  and hardwood decker effluent respectively.   About
    4 Ibs  of  100% sulfuric acid are required to neutralize
    1000 gals,  of any one of the three effluents.

    The major  effects on feed characteristics of pH adjust-
    ment were:
        —a reduction in feed color;
        —an  increase in the total solids of the  feed by
          about  400-500 ppm, as a result of the addition
          of  sulfate ion (from H2S04);  and
        —some change in the colloidal nature of  the feed
          as  evidenced by a change in suspended solids
          content.

    Figure 26  shows  the effect of pH adjustment on color for
    a sample of pine caustic extraction filtrate.   The color,
    measured by sample absorbance at 625 my, gradually dropped
    with decreasing  pH until at pH 2 the sample became turbid
    due to heavy  flocculation of organics.

    It was observed  during several experiments with pine
    caustic extraction filtrate that the suspended solids
    level  increased  with reduced pH.  Typical data are given
    below.
         SUSPENDED  SOLIDS  OF PINE CAUSTIC EXTRACTION
                FILTRATE AS  A FUNCTION OF pH
        pH Adjusted With H^SO,        Suspended Solids,  ppm

                  8                              43
                  7                              39
                  6                              58
                  5                             113
                  2                            turbid (not measured)

   This behavior  of pine caustic extraction filtrate  is fur-
   ther discussed in Section V.E  :   Important Factors
   Controlling Membrane  Flux.

3.  Prefiltration

   In a previous  laboratory  test program,  the necessity of
   feed pretreatment for particulate removal to obtain
   satisfactory ultrafiltration rates was  apparent.   The
                               57

-------
Ul

00
p.
6

in
CM
vo

-P
(0


M
O
en
            1.0
              .8
              .6
              .4
                 0
                                                                                TT
                                                pH
                  FIGURE 26.   EFFECT OF pH ON COLOR OF PINE CAUSTIC EXTRACTION FILTRATE

-------
exact type of filtration and filter equipment could not
be determined under laboratory conditions with "aged"
feeds.  Hence, a substantial amount of time was devoted
during this pilot plant program to investigate filters
and their operating conditions in order to obtain satis-
factory removal of suspended solids.  The following
filters and filter combinations were examined in the pilot
plant program:
     8/14/72-9/27/72

     9/27/72-10/5/72
     10/17/72-2/1/73
                         Broughton nominal 10 y mesh filter;

                         1) Bauer Hydrasieve (removal of
                         fibers and gross suspended solids;

                         2) Hydromation depth filter (l-5y
                         solids removal); and

                         3) Cuno polishing filters (ly
                         cartridges;

                         1) Shriver filter press with wood
                         flour as precoat material and
                         filter aid (body feed, between
                         150 and 300 ppm); and

                         2) Cuno polishing filters (ly
                         cartridges;

                         1) Sparkler pressure leaf filter
                         with wood flour as the precoat
                         material and filter aid  (body
                         feed, between  150 and 300 ppm); and

                         2) Sparkler Velmac 5y  disc filter

Four different main filters were used:  the Broughton mesh
filter, the Hydromation depth filter, the Shriver filter
press, and the Sparkler leaf filter.  The installation
of each was carried out to improve pilot plant perform-
ance.  The Hydromation depth filter used a packed bed of
granular PVC as the filtration medium.  The Shriver  and
Sparkler filters were precoat filters and operated with
filter aid added to body the feed.  For most of the  experi-
mental program wood flour was used as the body feed; how-
ever, some tests were performed with other filter aids,
and the results are discussed below.
     2/2/73-3/1/73
                           59

-------
a.  Operational Experience

    Immediately after pilot plant startup, it became apparent
    that the Broughton filter was not providing adequate sus-
    pended solids removal.  Specifically, the membrane car-
    tridges were plugging rapidly with suspended solids, as
    evidenced by a rapid increase in pressure drop across
    the cartridges and a rapid decrease in ultrafiltration
    rate.  At that time a sample of the suspended solids in
    the pine caustic extraction filtrate was analyzed for
    particle size distribution.  The results of this analysis
    are contained in Table 5.  It is apparent that about 50%
    of the suspended solids are below lOy, and is not sur-
    prising that the nominal 10y Broughton filters proved  in-
    effective for particulate removal.  No actual data were
    obtained on the removal efficiency of suspended solids
    by the Broughton filter, but it was probably 50% or less.

    Samples of the pine caustic extraction filtrate were
    submitted to filter vendors to obtain recommendations for
    prefiltration.  Vendor tests showed that surface filtra-
    tion resulted in rapid filter blinding, and that depth
    filtration was required.  As an initial approach, a
    Hydromation depth filter was installed.  Excellent re-
    moval of suspended solids was obtained, as determined
    both by the reduction in suspended solids level and more
    stable ultrafiltration rates in the pilot plant  (data
    discussed in Section V.C.).  Unfortunately, the filter
    size was inadequate for operation of the entire pilot
    plant on a continuous basis.  The vendor was unable to
    deliver a larger filtration unit within the schedule and
    budget constraints of the program, and it was decided to
    switch to a precoat filter.
                                           2
    An existing Shriver filter press  (75 ft  area) was avail-
    able within the mill, and was installed in the pilot
    plant.  Based on a recommendation by the Sparkler Manu-
    facturing Co. , Wilner wood flour # 139 was chosen as the
    precoat and body feed material.  Substantially improved
    performance of the pilot plant was obtained with the
    Shriver filter press compared to the Broughton basket
    filters.  Pilot plant operation continued for about 3-1/2
    months using the Shriver filter.  However, it was real-
    ized in December that the filter was operating at~an
    unrealistically low filtration rate  (^ 0.1 gpm/ft ) and
    it was decided to install another filter which could
    treat the feed at more typical filtration rates
                               60

-------
                              TABLE 5

         PARTICLE SIZE  DISTRIBUTION OF SUSPENDED SOLIDS
       IN PINE CAUSTIC  EXTRACTION FILTRATE (UNNEUTRALIZED)
Particle Size  (microns)      Weight/  %     Cumulative Weight, %
       >149
        88-149
        44-88
        40-44
        32-40
        20-32
        10-20
         5-10
       2.5-5
       <2.5
9.9
2.8
3.9
1.03
1.88
13.76
17.68
19.84
20.94
8.26
99.99
90.09
87.29
83.39
82.36
80.48
66.72
49.04
29.2
8.26
                                 61

-------
           2
(^ 1 gpm/ft ).  Furthermore, it was suspected that the
Shriver press did not provide the quality of filtration
normally obtained with a precoat filter.  This conclusion
was based on often erratic performance of the ultrafiltra-
tion unit.  The following considerations may explain the
less than adequate performance of the Shriver filter:
     —it was an old filter and possibly contained
       defects;
     —the flow distribution between the different plates
       and frames  (24 plates) was poor and the precoat
       thickness was probably not uniform.  Furthermore
       it was not possible to form the precoat under high
       flow conditions which would have produced a more
       uniform precoat;
     —the filter press was operated at high pressure
       drop, between 30 to 60 psi, and this could have
       resulted in the "extrusion" of solids into the
       filtrate;
     —operational problems may have allowed the precoat
       to fall off when switching valves to change from
       precoat application to treating neutralized feed.

For these reasons a Sparkler leaf filter was ordered
and installed at the beginning of February, 1973.  The
Sparkler filter provided excellent suspended solids re-
moval and filtration rates.  It was possible to main-
tain filtration rates of approximately 1 gpm/ft2 for
periods of up to 24 hrs, depending on the feed type and
characteristics.   (Note, filtration characteristics of
the caustic extraction filtrate and decker effluents
differed, as discussed below).

Table 6 contains a summary of the performance data of
the main filters in processing pine caustic extraction
filtrate.  The data for removal of suspended solids are
based on a suspended solids analysis by filtration on
0.45y Millipore filters.   As described before, this
method provided only approximate values for suspended
solids levels.  However,  it is possible to recognize
major trends in filter performance.  Based on the suspend-
ed solids removal data in Table 6, it was judged that
the Sparkler and Hydromation filters gave the best per-
formance; and performance of the Shriver filter press was
good at times, but erratic.  The filtration rates of
both the Sparkler and Hydromation filters are judged to
be in the range suitable for commercial application.
                           62

-------
                                                  TABLE 6

                                          PERFORMANCE - MAIN FILTERS
              Filter
               Area     Filter
    Filter    sq ft     Medium
 Max/Min
  Flows
  During
Operation/
   gpm
                                             Max/Min
                                           Filtration
                                          Rates During
                                           Operation,
                                            gpm/sq ft
 Min/Max
 Duration
Filtration
Cycle, hrs
                                                                          Filtration
                                                                        Efficiency as
                                                                         indicated by
                                                                       Subsequent Ultra-
                                                                       filtration Rates
Removal of
 Suspended
  Solids
 % Removed
Broughton   5.6       10u        8-3
                      mesh
                                                 1.5-0.5
                                           inadequate
    Hydromation 1    granular PVC    8-3
                 8-3
                                                          0.1-1.0
                acceptable
                                                                                             50-80
Shriver    75     wood flour    10-3
u>
                                                0.13-0.04
                              4-20
                                                                         sometimes
                                                                         acceptable
                                    20-80
    Sparkler   15.3   wood flour    25-5
               1.6-0.3
                                                            4-16
                acceptable
                                                                                             50-80

-------
Perhaps the most relevant information on suspended
solids removal by the different filters is obtain-
able from the ultrafiltration rates of the pilot plant.
In general, the ultrafiltration rate data show that the
Sparkler pressure leaf filter and the Hydromation depth
filter were equally effective in removal of suspended
solids.  At times the Shriver filter press performed
adequately.

Performance of Other Filters

The performance characteristics of the Bauer Hydrasieve,
the Cuno cartridges, and the Sparkler Velmac disc
filter were not closely monitored because of their rela-
tively limited influence on the performance of the
ultrafiltration unit.

The Bauer Hydrasieve adequately removed fibers and gross
suspended solids from the raw feed, which prevented
clogging of valves and pumps used before the main fil-
ter.  However, the performance of the Bauer Hydrasieve
was not a critical part of the filtration operation from
the point of view of ultrafiltration performance.

The Cuno cartridge filters were installed just before
the ultrafiltration unit for final polishing of the
feed.  These filters also served to protect the ultra-
filtration unit in case of gross failure of the main
filter. The Cuno filters were primarily used in combina-
tion with the Shriver filter press.  Some additional
suspended solids removal was achieved by the Cuno fil-
ters and it was necessary to change the disposable
cartridge elements every 5-10 hrs of operation (1000-
3000 gals.).  It cannot be ascertained whether the
solids removed by the Cuno filters are particulates
which would pass through any precoat filter, or are
suspended solids which passed through the Shriver filter
due to below-par performance.

In the final phase of program, the Cuno filters were re-
placed by a Sparkler Velmac disc filter because of the
larger filter area and the capability of cleaning the
filter discs.   In combination with the Sparkler leaf fil-
ter, the Velmac filter lasted for about 60 hrs (20,000
gal.) when it was used for the first time.  Most of the
solids removed by the Velmac filter could be removed by
                           64

-------
washing.  However, some reduction in filter capacity
was observed in subsequent runs due to an inability
to completely regenerate the filter medium.  The
ultimate useful life of the Velmac filter has not yet
been established, but it is anticipated that in opera-
tion of a full-scale plant, life will be economically
acceptable.

Tests with Different Filter Aids

Two sets of experiments were performed.  One with the
Shriver filter press, and another with the Sparkler leaf
filter.

The data from the Shriver filter tests were obtained 2
under conditions of low filtration rates  (^0.1 gpm/ft ),
and can only be used as a guide in evaluating perfor-
mance characteristics of different filter aids.  In these
tests, five different filter aid materials were used:
# 139 wood flour, (Wilner Wood Products, Norway, Maine),
Dicalite 436  (Grefco diatomaceous earth), Celite 545
(Johns-Manville diatomaceous earth), Hyflo Super Cel
(Johns-Manville), and Perlite 400F (Chemrock Corp.).
In the tests a precoat of about 6 to 10 Ibs was applied
to the filter, and the filter aid was added as body
feed at a level of approximately 150 ppm.. In these tests,
no observable difference in removal efficiency of
suspended solids was noted.  However, a substantially
longer filtration cycle was obtained with the wood flour
than with the other filter aids.  This suggests that
the filter cake formed with wood flour is less suscepti-
ble to plugging or blinding by the suspended solids-
filter aid suspension.

In the tests with the Sparkler filter, three different
types of filter aid were examined:  wood  flour, dia-
tomaceous earth  (two grades), and a Perlite.  The tests
were performed with two different effluents:  pine
caustic extraction filtrate and hardwood  decker efflu-
ent.  In these tests, a 3-lb precoat of each of the fil-
ter aid materials was applied to the 15.3 ft2 Sparkler
filter.  After the precoat was applied, wood flour was
ised as the bodying material, and added at a level of
 Approximately 150 ppm.
                          65

-------
      The effluent filtered was always fresh, and was with-
      drawn directly from the 500 gal. feed tank.  Operation
      was at a standard filter pressure drop of 20 psi
      (except for the one run with Hyflo Super Cel since this
      is a looser type of filter aid and could compress at a
      20 psi pressure differential.

      A summary of the test results is presented in Table 7.

      The data in Table 7 show that initial filtration rates
      for all of the precoat materials were excellent, in the
      range of 1.4 to 1.8 gpm/ft .  In the tests with the pine
      caustic extraction filtrate, wood flour and the Dicalites
      maintained their original  filtration rates over the full
      21 minute test period.  In the tests with the hardwood
      decker effluent, which contained a much higher level of
      suspended solids, an appreciable falloff in filtration
      rate was observed, as expected.

      The Hyflo Super Cel is a much coarser filter aid, with
      a greater porosity and higher initial filtration rate.
      However, the filtration rate dropped off sharply with
      both pine caustic extraction filtrate and hardwood
      decker effluent, indicating that a fairly dense filter
      aid is to be preferred as a precoat medium if long fil-
      tration cycles are to be achieved.

      Removal of suspended solids in all tests was about the
      same, approximately 50 to 75%.

      On the basis of the tests with both the Shriver and the
      Sparkler filters, it was concluded that the Wilner # 139
      wood flour was the preferred filter aid.  In addition,
      wood flour has the advantages that it is inexpensive
      and can be easily disposed of by incineration with the
      concentrate from the ultrafiltration unit.  For these
      reasons, wood flour was used for the experimental
      program.

4.    Feed Characteristics

      The detailed experimental data on feed characteristics
      for caustic extraction filtrate, pine wood decker
      effluent and hardwood decker effluent,  respectively,
      are presented in summarized form in Table 8.
                              66

-------
           TABLE 7
 TESTS WITH DIFFERENT FILTER AIDS
ON SPARKLER 15.3 sq ft LEAF FILTER
Filter Aid
Wilner |139

Dicalite 436
c>
"^ Dicalite 436
Hi call te 476
Hyflo Super
Cel
Wilner §139

Dicalite 476
Hyflo Super
Cel
Type of Filter Aid
wood flour

diatomaceous earth

"
H
Per lite

wood flour

diatomaceous earth
Perlite

Effluent
Pine caustic
extract
n

n
n
n

Hardwood deck-
er effluent
"
n

Pressure
Drop
psi
20

20

20
20
10

20

20
20

Precoat
Ibs
3

3

3
3
3

3

3
3

Filtration rate
gpm/ft2
initial
1.4

1.7

1.5
1.6
1.6

1.4

1.8
1.5

final
(21 min)
1.4

1.5

1.6
1.6
0.85

0.9

1.2
0.2

Suspended
Solids
in/out
ppm
132/47

20/10

75/18
44/20
50/19

300/160

260/180
280/100


-------
     The range of compositions shown in Table 8 are based on
     analyses of all samples collected throughout the pilot
     plant program.  The detailed data presented in Figures
     27 to 33 show that the feed composition varied widely
     over relatively short periods of time, often within
     the same day.  Sharp variations in feed composition
     were noticed particularly when start up or shutdown
     operations took place in the mill.  This alone suggests
     that improved effluent flow control could reduce the
     mill washwater volume.  Presumably operation at high
     solids-high color loadings did not impair pulp or paper
     quality.  It should be possible to maintain these
     loadings near the highest tolerable level, thus re-
     ducing water consumption.

a    Suspended Solids

     Table 8 and Figures 27 to 29 give the ranges of suspended
     solids that were determined for the three effluents,
     both untreated and pretreated.  As discussed in the
     section on Analytical Methods, the suspended solids data
     are only approximate, and this accounts, in part, for
     the scatter in the data.

     The suspended solids level of the hardwood decker efflu-
     ents , during the limited time when measurements were
     made, was found to be higher than those for the other
     two effluents.

,     Total Solids
D.    	
     Figures 30 and 31 show the total dissolved solids contents
     of pine caustic extraction filtrate, pine decker effluent,
     and hardwood decker effluent, respectively.  Pine caustic
     extraction filtrate had a much greater amount of total
     solids (avg. about 7000 ppm) compared to the decker ef-
     fluents which averaged about 2500-3000 ppm solids.

     Feed neutralization and prefiltration introduced about
     400-500 ppm of solids into the three effluents, due to
     the addition of sulfate ion.
                                 68

-------
                                                  TABLE  8

                                      SUMMARY  OF FEED CHARACTERISTICS
VO
                                         Pine  Caustic
                                      Extraction  Filtrate
Pine Decker Effluent
   Hardwood
Decker Effluent

1.


2.


3.


4.



pH
Untreated Feed
Treated Feed*
Color, ppm, Cobalt units
Untreated Feed
Treated Feed*
Total solids, ppm
Untreated Feed
Treated Feed*
Suspended Feed, ppm
Untreated Feed
Treated Feed*
Range Average

11.5-12
6.0-6.5

12,000- 28,000
45,000
10,000- 19,000
27,000

4,400- 7,000
11,400
** **

50-200 80
10-90 30
Range Average

10.7-10.9
6.0-6.5

4,000- 6,000
9,300
3,000- 4,000
7,000

1,700- 2,400
8,200
** **

50-150 80
20-40 (limited data)
Range

10.0-10
6.0-6.

8,000-
22,000
6,000
13,000

1,200-
4,600
**

100-350
50-180
Average

.6
5

11,000
8,000

3,000
* *

200
(limited dat
        *   pH  of  feed adjusted to 6.5-6.9,  followed by filtration  through a depth or precoat filter

       **   total  solids  in neutralized,  filtered feed were increased over untreated feed by about
           400-500  ppm due to sulfate  addition (H-SO.)

-------
  40C-
                                                      £•  unneutralized feed  (untreated)



                                                      •  neutralized and filtered  feed (treated)
  30C-
a

 *
to

•H
H
0                                                                                     A

•o

^20(*
0)
(X
CO
3
cn
                                                      &
        average for   average for      * *&  m         a*   A  A
         untreated     .treated     •    *
irtrj	   vuA
             J
                          /:
                          ".A
                                       A
                                        ®
                 1
. O  • BJI

  • ••  *
       August      September   October      November    December     January    February      March
                                                              1972  1973

              FIGURE 27.  SUSPENDED SOLIDS OF CAUSTIC EXTRACTION FILTRATE

-------
    400
    300
o
w
•S
to  200

CO
                                                 unneutrailzed feed
                                                 (untreated)

                                                 neutralized and
                                                 filtered feed
                                                 (treated)
   100
            average  for un-
            neutralized feed
average for
neutralized
   feed
                          L.
                          fSL
         November    December     January     February     March
                           1972  1973

          FIGURE 28.  SUSPENDED SOLIDS OF PINE DECKER EFFLUENT

-------
   400
    300 h-
o
CO
•o
I
8,  200
3
to
    100
average for unneutralized
                  X
                             a
                                         A
                  &
                                               unneutralized feed
                                               (untreated)

                                               neutralized and
                                               filtered feed
                                               (treated)
                                     £>
                                               February
                                                March
November    December     January
                  1972 1973
   FIGURE 29.  SUSPENDED SOLIDS OF HARDWOOD DECKER EFFLUENT

-------
             untreated Pine Caustic Extraction Filtrate
10,000
          •  neutralized and filtered Pine Caustic
             Extraction Filtrate

          o  untreated Pine Decker Effluent
5,8000
 •»
to

•H
rH
o
01
 i 6000
 «J
 4J
 O
  4000
            average for unneutralized
          Pine Caustic  Extraction Filtrate
                                                                      &
                &
                                                        &
             average for unneutralized
             Pine Decker Effluent
  2000
                                                                              a
                                                                                G> a
                                                                          a
                                                                          a
          August
                     September   October
                                      November    December     January     February   March
                                                      1972    1973
FIGURE 30.  TOTAL SOLIDS OF PINE CAUSTIC EXTRACTION FILTRATE AND PINE DECKER EFFLUENTS

-------
  10,000
    8000
p<
    6000
(0
4J
o
    4000
    2000
                    a
                   -fir
                                        unneutralized feed
a
A
  average for un-
neutralized feed
           November    December      January    February     March
                               1972  1973
          FIGURE 31.  TOTAL SOLIDS OF HARDWOOD  DECKER EFFLUENT

-------
   48  —
   40
   32
01
X  24
8.
cu
8  16
                G  Pine Caustic Extraction Filtrate  (untreated)

                                                               o
                A  Pine Caustic Extraction Filtrate
                   (neutralized and filtered)


                a  Fine Decker Effluent  (untreated)
                 Pine Decker Effluent
                 (neutralized and filtered)
            average for untreated Pine
            Caustic Extraction Filtrate
                                                    00
 average  for treated Pine
Caustic  Extraction
                                             o
                                             §
                                             o
            average for untreated Pine
                Decker Effluent
                                                                  o
                                                                o o
                                                         00
                                                              o  o  o

                                                                 o
                                                                 GO
                                                       O
                                                                   0
                                                         o
                                                        o
                                                    o
            average for_treated_ Pijie Decker Effluent
                _L
                     i
         J.
J_L
I  I  I  I  I
1
                                                             s
                                             ''  '  '  '  '  '  '  '  i
                                                                  ©     O
                                                       ___.,
                                                      "' ...... © o       o
                                                      ©   o         ©
                                                                     o
                                                                                   00
                                                                                     O  O  O©
                                                                                      o   o°©
                                                                             	QJ
                                                                                •  •
                                                                                        <  '  I  '  '  '  '
                                              November   December     January    February     March
                                                              1972  1973
            August      September    October


                  FIGURE 32.  COLOR OF PINE CAUSTIC EXTRACTION FILTRATE AND PINE DECKER EFFLUENT

-------
      48
                                                                          O unneutralized feed
                                                                          A neutralized and filtered feed
      40
      32
   o
   H

   X
Os
      24
   8
      16
              average for unneutralized feed
                                    O
                                    O
                                                                             o
                                                                      o
                                                                     o
                                                                              o
                                                                                      o
0

&
              average for neutralized and
                     filtered feed	
                                                               ©*~
i  i  t
                     I  i  i  i  i  i  I  i  i  i  i  i  I  i  i  i  i  i
                                                         I
           August     September   October      November    December     January
                                                                 1972 1973
                                                                       February    March
                               FIGURE  33.  COLOR OF HARDW6)OD  DECKER EFFLUENT

-------
c.   Color

     Figures 32 and 33 show the color contents of the three
     effluents.  It is seen from these figures that the
     pine caustic extraction filtrate was much more highly
     colored (avg 28,000 ppm),  than the other two effluents.
     Pine decker effluent had the least color, averaging
     about 6000 ppm.

     Figure 34 shows the effect of feed pretreatment (pH
     adjustment to 6.5 to 6.9 and precoat filtration) on the
     "field" color content of a series of samples of pine
     caustic extraction filtrate.  Feed pretreatment removed
     about 30-35% of the color from the raw feed; the major
     part was due to pH adjustment.

d.   Other Solutes

     Table 9 gives additional analytical data for a few feed
     samples of the three effluents.
     In summary, the following conclusions can be drawn about
     the characteristics of the three effluents:
          —Pine caustic extraction filtrate has higher
            levels of total dissolved solids, ash and color
            than either of the decker effluents;
          —Although the two decker effluents have similar
            levels of total dissolved solids, pinewood decker
            effluent has less color;
          —Hardwood decker effluent has the highest suspended
            solids level; pine caustic extraction filtrate
            and pine decker effluent levels are about the same;
          —Pine caustic extraction filtrate contains much more
            total and ionic chloride than  the decker effluents;
          --Decker effluents contain a substantial amount of
            sulfate ion, originating in the pulp digester;
          —The feed compositions  of all three  effluents vary
            widely within a relatively short period of time;
          --Pretreated effluents have about  30-35% less color
            than the untreated  (raw) effluents;
          --pH adjustment of the effluents introduces about
            400-500 ppm sulfate ion;
          ~-A major fraction of the suspended solids of the
            three effluents is smaller than  10^1, but is
            reasonably effectively removed by either depth
                                 77

-------


-0
c
(0
•o
0)
N
•H
id
M
9
oo 0)
4-t
0

^1
o
r-l
O
U

JU
CO
1
o
I—I
X
1
•o 20
^1
^s^
*s^
0 .^^^ O O
.t^^
^
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
FIGURE 34,
 20                        30                        40
       Color  of Unneutrailzed Feed, ppm x 10
EFFECT OF FEED PRETREATMENT  ON  COLOR OF PINE CAUSTIC EXTRACTION FILTRATE

-------
Components
                                               TABLE 9

                 WASTE CHARACTERISTICS AT THE NORTH CAROLINA MILL - DETAILED ANALYSES
                 Untreated Pine Caustic
                   Extraction Filtrate
Untreated Pinewood
  Decker Effluent
sampled on 1-3-73
Untreated Hardwood
  Decker Effluent
sampled on 2-1-73
PH
10-13-69
11.3
Color, ppm
S . S . , ppm
T.S. t ppm
Ash / ppm
ft a , ppm
Ca , ppm
Cl~, ppm
	 1
S04~~' PP™
Fe , ppm

1,180
78
1,010



10-16-70
11.5
14,000

6,800
3,800
2,000
5
1,600

1.6

2-10-73
11,7
28,000
80
7,000

1,200
60
1,275



3 pm
10.7
7,700
230
2,244
1,010
480
20
14
260
1.4
5 pm
10.7
7,300
70
2,284
784
425
16
14
190
1.2

8 pm
10.7
8,300
140
2,468
908
516
4
14
240
0.8

1 pm
10.0
18,300
400
3,848
976
58
16
30
350
2

3 pm
10.5
8,300
114
2,520
848
480
4
6
200
1.2

5 pm
10.7
7,700
214
2,244
1,464
441
12
35
200
0.1


-------
            filtration or precoat filtration.

B.   REJECTION DATA AND EVALUATION

     Comprehensive data were collected on detailed color and
     total solids rejections data for pine caustic extraction
     filtrate, pine decker effluent and hardwood decker efflu-
     ent, respectively.  Specifically, the following informa-
     tion was included:

          a.  Color of unneutralized feeds; neutralized and
     filtered feeds, final concentrates and mixed permeates
     from the ultrafiltration unit;

          b.  Total solids of unneutralized feeds, final
     concentrates and mixed permeates from the ultrafiltra-
     tion unit;

          c.  Concentration ratios, defined as the ratio of
     feed flow to the ultrafiltration unit to concentrate
     flow from the ultrafiltration unit;

          d.  Percent color removal, defined as

     color  of unneutralized feed-color of mixed permeate x 100;
              color of unneutralized feed
     and

          e.  Color removal efficiency of the ultrafiltration
     unit by stage for treatment of pine caustic  extraction
     filtrate.

     For presentation here these data have been condensed and
     displayed in Figure 35, which summarizes the separation
     efficiency  of the pilot plant during the operating period.
     For all three effluents, the concentration ratio achieved,
     the total solids level in  the concentrate, and the percent
     color  removal are shown.

 1.   Color  Rejection

     As  seen  in  Figure 35 excellent color removal was obtained.
     For example, for  about eighty-fold  concentration, color
     removal was approximately  90% for the pine caustic extrac-
     tion filtrate  and 95% for  both decker effluents.  The
     higher color removal efficiency  for decker effluents is
     hypothesized to be  related to the molecular  weights of
                                  80

-------
  20
I
Sio-
H C
S
   t,   £>
                                  7"
                                10  &
                                ft  sa
                                         »a ^
                        CEF      ,£
                     (-X.80  fold
                    concentration)
                                                                                        (-v-120 fold
                                                                                       concentration)
                                 a a
                                                                                 1
  100
   90
 I
 L-
«
n
    60
    40
                 A
              &  &
                 A
                &

            	  A
                             CEF
                          (i>80 fold
                          concentr ation)
 ^ Pine Caustic Extraction
   Filtrate (CEF)

-=-Pine Decker Effluent  (DE)
 a Hardwood Decker Effluent  (DE)
                                                                                       DE (^80  fold
                                                                                       concentration)
                                                  & *   £ —''3
                                                     'X «.
                                                                                     e>  &
                                                     ae>
                                                     a a
                                                             &  e>   &
                 i
                               1
         August     September   October    November    December     January
                               FIGURE 35.  REJECTION  AHD  CONVERSION DATA
                                                                      February      March

-------
                                                           Table  10


                                                           REJECTION

Effluent

Pine
Caustic
Extraction
Filtrate


Pine
Decker
Effluent
Hardwood
Decker
Effluent

Period

11/5-
11/8/72
11/28-
12/1/72
2/22-
2/27/72
1/30-
1/31/73

1/18-
1/23/73

Color Concentration , ppm

Raw
Feed

14,000
18,000
20,000
30,000
30,000-
40/000
6,000-
8,000

12,000-
16,000

Neutralized
Filtered
Feed
Average
11,000

17,000

23,000

4,500


9,000


Permeate

1200-1500

2000-3000

4000-5600

430-470


600-900


Final
Concentrate

500,000-
600,000
1,000,000-
1,500,000
1,600,000-
2,500,000
180,000-
220,000

420,000
500,000

Total
Solids in
Concentrate
%

12-16

12-15

18-22

5


6-7



Concen-
tration
Ratio

80-90

80-100

100-130

70-90



80-90


Color
Removal , %

92

90

86

94


95

--
00
to

-------
the color bodies in the effluents.  The presence of
lower molecular weight lignaceous materials in the pine
caustic extraction filtrate would be expected due to
lignin fragmentation in the bleach plant chlorination
stage.

The color removal was also dependent on the concentra-
tion ratio obtained:  the greater the concentration
ratio, the higher the color content of the mixed per-
meate.  This is to be expected since permeate color
increases with feed color, and at high conversion a
greater fraction of the membrane area is exposed to
highly colored feed.

This effect is further demonstrated by Table 10, which
contains rejection data for selected periods when
operation proceeded smoothly.  Also shown is the higher
color removal observed when processing decker effluents.

Appendix I contains some limited color rejection data
for individual membrane stages when treating pine caustic
extraction filtrate.  Color rejection increased with
feed concentration  (progressing through the stages of
the pilot plant).  This is due to the fact that with
increasing feed concentration a higher fraction of the
feed color bodies are high molecular weight species
since lower molecular weight solutes are preferentially
removed in the earlier stages.

In addition, color rejection was clearly dependent on
membrane type.  The T.J. Engineering modules showed a
wide variation in color rejection, between 87 and 96%,
which is attributed in part to poor quality control in
both membrane and cartridge manufacture, as well as
occasional leaks.  The color rejection of the four Gulf
Environmental Systems modules was between 98-99.9%, and
clearly superior to rejections for either the Eastman or
T.J. Engineering modules.  This was expected since the
HT-00 membrane (used in the latter two modules) intrin-
sically had a higher "molecular weight cutoff" than the
Gulf membranes.

Color rejection of the pilot plant was low on a few
occasions due to mechanical failures of several T.J.
Engineering cartridges:  Especially troublesome was o-
ring failures on the permeate collection tube  (see
comments column in Appendix D, and discussion in section
V.G., Module Mechanical Failures).
                          83

-------
2.  Total Solids Rejection

    Figure 35 also shows the effect of concentration ratio
    on total solids content of the final concentrate.  At
    eighty-fold concentration of pine caustic extraction
    filtrate, the final concentrate contained about 1570
    total solids.  The greater part of the solids were high
    molecular-weight organics since the ultrafiltration
    membranes used did not appreciably retain salts and low-
    molecular solutes.  The detailed pilot plant data demon-
    strate that the total solids rejections of the membranes
    were between 5 and 30%.

3.  Material Balances

    Material balance data for color and total solids was
    collected and evaluated at least every operating shift.
    Both field color tests (at as-is pH) and the standard
    color tests (at pH 7.6) methods were used.  Because of
    the pH sensitivity of the color values in the effluents,
    the standard test data was used for color balance
    calculations.

    Tables 11 and 12 contain composition data of samples
    from tests with pine and hardwood decker effluents.
    These data also show that retention of non-complexed
    salts and other low molecular solutes  (sodium ion,
    sulfate ion, ash and total solids) was real, but low.
    This can be seen best by comparison of permeate and
    concentrate assays.  This was probably due to the use
    of the Gulf modules in Stages 4 and 5.  These membranes
    had about 3070 NaCl rejection.

    Rejection of Fe444 and Ca44" was very high, and this is
    attributed to complex formation of these ions with re-
    tained organics.  The pH of concentrate samples was
    consistently lower than the filtered feed; and that for
    permeate samples was consistently higher.  Evidently,
    the retained organics were weakly acidic.

    The assays in Table 13 for pine caustic extraction
    filtrate samples also show similar behavior.  Note
    values for the contents of trivalent metal oxides
    (mainly Fe444 ) , Ca44 , 804"„ ash and chloride.  Un-
    fortunately, the samples were collected on different
    days and only qualitative conclusions can be drawn.
                              84

-------
 COMPOSITION DATA:
 TABLE 11
PINE DECKER EFFLUENT SAMPLES
Date Sampled  1-31-73  3:00
Assays

Total Solids
Ash
pH
Fe++
Na+
S04~
Ca++
Color

Assays

Total Solids
Ash
pH
Fe++
Na+
S04—
Ca++
Color

Assays

Total Solids
Ash
PH
Fe++
Na+
S04—
Ca++
Color
Permeate

1/376 ppm
948 ppm
7.0
0.8 ppm
425 ppm
468 ppm
0 ppm
430 ppm
Date Sampled
Permeate

1,296 ppm
868 ppm
6.6
0.6 ppm
325 ppm
470 ppm
0 ppm
430 ppm
Date Sampled
Permeate

1,392 ppm
716 ppm
6.5
0 . 4 ppm
350 ppm
420 ppm
0 ppm
470 ppm
Concentrate
( concentration
ratio = 50)
48,740 ppm
13,532 ppm
5.9
182 ppm
4,860 ppm
7,157 ppm
448 ppm
200,000 ppm
1-31-73 5:00 pm
Concentrate
( concentration
ratio = 50)
49,312 ppm
13,652 ppm
5.6
110 ppm
4,620 ppm
7,280 ppm
320 ppm
200,000 ppm
1-31-73 8:00 pm
Concentrate
(concentration
ratio = 50)
48,256 ppm
12,240 ppm
5.3
96 ppm
4,620 ppm
7,250 ppm
280 ppn
186,000 ppm
Neutralized &
Filtered Feed

3,500 ppm
1,700 ppm
6.5
13.97 ppm
600 ppm
839 ppm
16.05 ppm


Neutralized &
Filtered Feed

3,624 ppm
1,684 ppm
5.8
6.20 ppm
575 ppm
900 ppm
16.0 ppm


Neutralized &
Filtered Feed

3,688 ppm
1,648 ppm
5,8
5.2 ppm
600 ppm
970 ppm
8.0 ppm

Raw Feed

2,244 ppm
1,012 ppm
9.5
1.40 ppm
480 ppm
260 ppm
20 ppm
7,670 ppm

Raw Feed

2,284 ppm
784 ppm
8.5
1.2 ppm
425 ppm
190 ppm
16 . 0 ppm
7,330 ppm

Raw Feed

2,468 ppm
908 ppm
9.2
0 . 8 ppm
516 ppm
240 ppm
4.0 ppm
8,330 ppm
                          85

-------
                     TABLE 12




COMPOSITION DATA;  HARDWOOD DECKER EFFLUENT SAMPLES







Date Sampled  2-1-73  1:00 pm
Assays
Total Solids
Ash
PH
Fe++
Na+
S04—
Ca++
Permeate
1,344
572
6.8
1.2
325
200
0
ppm
ppm

ppm
ppm
ppm
ppm
Concentrate
(concentration
ratio = 62)
36
8


2
4

,512
,072
5.8
134
,900
,000
96.0
ppm
ppm

ppm
ppm
ppm
ppm
Neutralized &
Filtered Feed
5,184
1,636
6.0
5.2
880
350
24
ppm
ppm

ppm
ppm
ppm
ppm
Raw __ Feed
3,848
976
10.5
2.0
580
350
16.0
ppm
ppm

ppm
ppm
ppm
ppm
Date Sample  2-1-73  3:00 pm
Assays

Total Solids
Ash
PH
Fe++
Na+
SO4 —
Ca++
Color

Assays

Total Solids
Ash
pH
Fe++
rfa+
804 —
Ca-H-
Permeate

2,316 ppm
1,192 ppm
6.8
1 . 2 ppm
558 ppm
591 ppm
0 ppm
730 ppm
Date Sampled
Permeate

2,432 ppm
1,460 ppm.
6.9
1.0 ppm
625 ppm
660 ppm
0 ppm
Concentrate
( concentration
ratio = 60)
59,652 ppm
12,896 ppm
5.7
148 ppm
4,500 ppm
5,100 ppm
1,281 ppm
400,000 ppm
2-1-73 5:00 pm
Concentrate
(concentration
ratio = 60)
62,948 ppm
13,704 ppm
5.7
125 ppm
5,300 ppm
5,950 ppm
481 ppm
Neutralized &
Filtered Feed

6,296 ppm
2,280 ppm
5.6
12 ppm
920 ppm
1,040 ppm
48 ppm


Neutralized &
Filtered Feed

6,608 ppm
2,348 ppm
5.8
7 . 0 ppm
972 ppm
1,600 ppm
80 ppm
Raw Feed

2,520 ppm
848 ppm
10.5
1.2 ppm
480 ppm
200 ppm
4.0 ppm
15,000 ppm

Raw Feed

2,244 ppm
1,464 ppm
10.7
0.1 ppm
441 ppm
220 ppm
12 ppm
                        86

-------
                       TABLE 13

COMPOSITION DATA:  PINE CAUSTIC EXTRACTION FILTRATE SAMPLES
ASSAY
 PERMEATE
sampled on
  1-30-73
                                       CONCENTRATE
sampled on
  2-13-73
sampled on
  2-14-73
pH


Total Solids


Ash

Trivalent
Metal Oxides

Na+


S04—


Ca++


Cl-
    7.4


  4,560 ppm


  2,950 ppm


   3.32 ppm


  1,200 ppm


    580 ppm


      8 ppm


  1,275 ppm
     6.7


 142,850 ppm


  27,920 ppm


   4,120 ppm


  10,400 ppm


   1,809 ppm


   8,000 ppm


   5,86 8 ppm
 200,000 ppm


  35,000 ppm


   4,500 ppm


  12,000 ppm


   2,000 ppm


  10,000 ppm


   6,000 ppm
Specific
Gravity
  0.975
  1.0635
  1.0766
                          87

-------
    In summary, all the rejection data have confirmed the
    basic assumptions made at the beginning of the program.
    Color bodies and organics in the three effluents can be
    selectively concentrated by ultrafiltration.  Permeates
    had low color contents, especially the decker effluents,
    but still contained the greater part of the total dis-
    solved solids (ash) contained in the original effluents.
    High solids,levels in the concentrate were achieved,
    at times over 20% in the pine caustic extraction filtrate
    tests.

C.  ULTRAFILTRATION RATE DATA AND EVALUATION

    The detailed ultrafiltration rate (flux) data have been
    condensed for presentation and discussion in this section.

    The dependence of flux on time and operating conditions
    was complex.  Operating time was important for both
    short-term operation, i.e. periods up to 48 hours, and
    long-term, i.e. months.  Time effects can be attributed
    to four factors:
        --reversible membrane fouling/flow channel plugging
          (short term effect);
        --irreversible membrane fouling/flow channel plugging
          (long term effect);
        —membrane compaction (long term effect); and
        --membrane cartridge compression (long term effect).

    During the pilot plant experiments, various changes were
    made in operating conditions and pilot plant equipment,
    including several changes of membrane cartridges.  For
    this reason it is logical to discuss the operation of
    the pilot plant chronologically, and seven  distinct
    periods are discussed below.

    Before detailing performance of the pilot plant during
    these periods, however, flux data for the membrane
    system during the entire program will be presented.
    In Table 14 membrane flux by shell is presented, from
    the beginning of pilot plant operation through the
    end of the program.  These fluxes were measured at
    the start of experiments; that is, after membrane clean-
    ing.  The efficiency of membrane cleaning was highly-
    variable, and this explains in part some of the varia-
    tion in day-to-day operation.  These data have been in-
    cluded in Figures 36 to 42, which show the flux behavior
    at both the beginning and end of experiments.  An
    evaluation of these data follows.
                              88

-------
                      TA8X.B X4




MEMBRANE FLUX BY STAGE DURING PILOT  PLANT PROGRAM
Cumulative
FEED* Date Hours of
Operation
PCEF 8-19-72






00
SO









8-21
8-22
8-25
8-30
8-31
10-25
10-28
10-31
11-1-72
11-2
11-2
11-3
11-6
11-7
11-10
f 11-11
28
32
38
50
66
80
110
135
190
200
204
220
240
280
300
336
356

la
22.5
24
13.5
12
20.2
10.5
36
30
22.5
34.5
30
28.5
24
18.9
10.5
12.9
16.5

Ib
.75
16.5
20.2
12
22.5
8.2
13.5
7.5

34.5
30
28.5
24
18.9
10.5
12
15
Permeate
1C
.4
6
15
7.5
18
7.5
15
12
13.5
33
27
28.5
22.5
21
18.7
22
18.7
Flux by Stage/ gfd
2
18
18
15
7.5
37.5
9
15
15

34.5
30
28.5
22.5
22.5
15
18.7
19.5
3
21
24
14.2
15
13.5
10.5
14.3
12
6.8
30
22.5
24
18
21
15
20.2
20.2
4
6
7.5
6.9
7.5
6
9.3
15
15
15
30
30
28.5
21
15
12.7
22.5
22.5
5
1.5
3.0
5.3
4.4
4.5
2.7
6.9
9
6.3
15
15
15
13.5
4.5
3.7
6.6
6.6
Average
flux,
gfd
10.3
14.4
10.3
10.3
14.4
8.2
14.4
14.4
10.3
29
25.3
25
19.9
16.6
11.8
15.8
16.2

-------
                     TABLE 14
                    (continued)

MEMBRANE FLUX BY STAGE DURING PILOT PLANT  PROGRAM
FEED* Date
PCEF 11-12-72





VO
o









'
11-16
11-17
11-18
11-28
11-29

12-1
12-2
12-3
12-5
12-6
12-7
12-8
12-9
12-10
r 12-11
Cumulative
Hours of
Operation
376

408
430
460
476

516
530
550
585
600
615
636
644
660
680

la
13.5
22.5
19.5
19.5
18
10.5

15
15
13.8
17.4
16.5
13.8
34.5
33.8
30
34.5

Ib
12.7
19.5
18
15
18
10.5

15
15
13
15
12.9
11.7
12
12
10.8
12
Permeate
Ic
17.2
20.2
18
18
18
9

14
21
12
15
14.4
11.9
21
18
18
19.5
Flux by Stage/ gfd
2
16.5
21
19.5
15
15
15

19.5
18
18.7
18
18
15
18
15
15
15
3
15
18.7
18
14
15
12.7

18
15
15
15
18
18
16.5
18
14
16.5
4
15
8.7
15
14
8.1
6.0

15
15
15
17
16.5
12.4
15
15
15
13.8
5
8.9
3.0
3.0
2.4
3.0
2.4

2.4
4.5
6.0
7.3
6
2
2
2
6.3
5.4
Average
flux,
gfd
13.6

15
13.4
13
8.8

13.6
14.2
12.9
14.3
13.9
11.5
16.3
15.6
15
16

-------
                      TABUS 14
                    (continued)
MEMBRANE FLUX  BY  STAGE DURING PILOT PLANT  PROGRAM
FEED* Date
PCEF 12-12-72





o
.j









i
12-13
12-14
12-18
12-19
12-20
12-21
12-22
12-23
12-26
12-27
12-28
1-4-73
1-5
1-6
1-7
' 1-8
Cumulative
Hours of
Operation
700
714
734
760
781
797
813
824
846
869
897
919
943
967
989
1010
1033
Permeate Flux by Stage, gfd Average
la Ib
33 12
34.5 10.5
34.5 10.5
34.5 10.5
13.5 36

34.5
34.5
34.5
34.5
34.5
36
34.5
34.5
37.5
34.5
34.5
Ic 2
17.2 13.2
16.5 16.5
15 13.5
13.5 15
13.5 12
11.7 10.5
12 12
12.8 10.7
11.3
10.5
12
10.5
17.2
17.2
18.5
18.5
20.4
3
15
16.5
16.5
15
15
14.3
13.8
13.8
12.7
15
16.5
15
15
15
16.5
16.5
16.5
flux,
4 5 gfd
13.2 3.4
6.7
6.0
4.5
3.0
4.3
3.8 3.5
6 6
3.6 3
4.4 4.5
5.7 4.5
3.3 2.3
3.9 5
4.5 4.8'
5 2.7
3.3 1.5
4.5 3.9

-------
                     TABLE 14
                    (continued)
MEMBRANE FLUX BY STAGE DURING PILOT PLANT PROGRAM
Cumulative
FEED* Date Hours of
Operation
PCEF 1-9-73
i 1-10
HDE 1-18



> '
5
1-19
1-22
1-23
1-24
•J
PDE 1-26


>
1-29
1-30
' 1-31
HDE 2-1-73
MDE 2-2
PCEF 2-12


^
2-13
2-19
, 2-20
1056
1078
1092
1098
1110
1120
1127
1136
1146
1158
1172
1184
1196
1000
1016
1025
1035

la Ib
31.
33
27.
22
29.
26.
25.
29.
29.
34.
27.
26.
24
24.
25.
22.5 26.
21 25.
Permeate
Ic
5

6

4
4
7
1
1
2
3
9

8
5 22.5
3 22.5
5 21
Flux by Stage/
2 3
18 17.
18.5 17.
10.8 13.
9.6 13.
11.5 13
9.9 11.
9.6 10.
10.4 11.
12.
13
11.
12.
10.
13.
13.
15
12
gfd

2
2
5
5

4
7
6
9

1
4
5
8
5


4
4.
4.
5.
5.
5.
4.
3.
5.
4.
4.
4.
5.
4.
15
15
15
13.

5
5
3
3
3
2
8
3
5
5
5
3
5



5
Average
flux,
5 gfd
3
3.8
6
6
6
3.8
4.5
6
5.3
6
4.5
4.5
4.5
16
13.5 16.2
5.4 16.2
9.6 15.8

-------
                                         TABLE 14
                                       (continued)
                    MEMBRANE FLUX BY STAGE DURING PILOT PLANT PROGRAM
FEED* Date
PCEF 2-21






S
i
2-22
2-23
2-26
2-27
2-28
3-1

' 3-2
Cumulative
Hours of
Operation
1052
1072
1090
1106
1128
1140
1154

1171

la
19.
18.
18
17.
16.
16.
19.

17.


5
8

3
5
5
5

1

Ib
24
24.
24
24
22.
23.
19.

24.



8


5
3
5

6
Permeate Flux
Ic 2
21
21
19.5
18.7
16.5
18
19.5

17.1
by Stage/ gfd
3
12
13.5
12
13.8
12
12
14.9

12.2
4 5
13.5 9.6
15 10
15 8.3
15 4.5
14.2 10
13.5 9
11.8

13.2
Average
flux/
gfd
15.
15.
14.
14.
13.
13.
13

12.
6
3
8
1
7
7


6
*FEED Abbreviations Key:
      PCEF - Pine Caustic Extraction Filtrate
       HDE - Hardwood Decker Effluent
       PDE - Pine Decker Effluent
       MDE - Mixed Decker Effluents

-------
36
32
28
24
20
12
                        A
i_i   at beginning of test

O   at end of  test
    A
   A
     A
      A
                A
    	      Eastman Cartridges
      •si	
                I
                                                        A.
                                         A
                                                         A
                                                   OO
                                      ,low circulation rate
                                        A
                                                             circulation
                                                               rate
        low  circulation
              rate
                                  T.J.  Cartridges
                                                          T.J. Cartridges  (old  set)
                                                           (1948,1970,3749)
                                                        T.J.  Cartridges from
                                                        Stage Ib (low flux)     T.J.  Cartridges
                                         (4713,4723,4724)
                                 I
                                                          (4709,4711,4719)

                                                           I             I
     (1950,4714,4722)

          I	
August
          September    October
                                           November    December     January
                                                             1972   1973
                               FIGURE 36 :  MEMBRANE FLUX. FOR STAGE la
February
                                                                                      March

-------
   32
   28
   24
  •O
  M-l

  ^20
vo
    16
    12
     8
           A  at beginning of test


               at end of test
                   A
        A



        A




\—      A
                   A
        	  T.J.  Cartridges (old set)
                                                                         A
                                      T.J.  Cartridges


                                        0 574 71 i~'
   T.J.  Cartridges (old set)


    (2026, 4283, 4395)


t             I           I
          August
               September    October    November    December     January

                                                          1972 1973

                            FIGURE  37:   MEMBRANE FLUX FOR STAGE lb
                                                                                   February
                                                                                         March

-------
    28
    24
    20

  •o
VO CP
**  .16
    12
            A  at beginning of  test

            O  at end of test
                A
               A
        	    T.J.  cartridges (old set)
              A
1
             A
             A
             A
                  T.J.  Cartridges (4722,4725,3747)
                                                                                           ' low  circulatio
                                                                                                   rate
                                                                                               \
 T.J.  Cartridges
    (old set)
("3749, 3782, 3751BT'
         August      September    October    November     December     January     February
                                                                 1972   1973
                                                                   March
                                  FIGURE  38:   MEMBRANE FLUX  FOR STAGE  Ic

-------
32
28
         A   at  beginning  of  test

         O   at  end  of  test
A
                                         A
24
20
16
12
   A
 A


T.J.  Cartridges (old set)


    I	I
                                         I
                                              T.J. Cartridges
(4714,4716,4717)    "^

       I             I
                                      T.J.  Cartridges
                                    (CorrugatedT^Spacer)

                                      I           1
      August     September     October    November     December     January
                                                             1972  1973
                               FIGURE 39:  MEMBRANE FLUX FOR STAGE 2
                                       February
                                                                                 March

-------
  TJ
00
    28
    24
    20
    16
    12
        	   T.J. Cartridges  (old set)
                at beginning of test

                at end of test
 A

A
    A

     A
                                               T.J.  Cartridges (4708,4712,4736)
A
                                             I
                                          I
                         i
I
          August      September    October    November    December      January    February
                                                                 1972   1973
                                                                                   March
                                   FIGURE 40:  MEMBRANE FLUX FOR  STAGE 3

-------
          Gulf
       	^	..
       Cartridges
       _(#!, #2)
             T.J. Cartridges

                 (old set)
           A   at beginning of test

                at end of test
 T.J. Cartridges        Gulf Cartridges   T.J. Cartridges

14737','4735,4731V  ^   ":""(f 1,  #2)   ^ ""("4737", 4 731 f 4721JT
                                  A

                                  A
  24
  20
                                            A
(n
  12
                 A
                                            o
                                            °
                                             A
August     SeptemberOctober    November     December       January     February
                                                      1972  1973
                                                                                                March
                                  FIGURE  41:   MEMBRANE FLUX FOR STAGE 4

-------
                     Gulf Cartridges
                         (13, #4)


             <—^ at beginning of
                test
             O at end  of  test
              P.J.  can
                         T.J. Cartridges

                        (4718,4720,4721)
                                     Gulf Cartridges (#3, #4)
                                                                              T.J. Cartridges
                                                                          >- -33	—	
                                                                              (4718,4720,4735)
  •O
§X
    12
August
September
                                   October     November     December      January
                                                                  1972 1973
February
March
                                   FIGURE  42:  MEMBRANE  FLUX FOR STAGE 5

-------
1.  Operating period:  8/19/72-8/31/72  (Start up).
Pilot plant operation was initiated with pine caustic
extraction filtrate on 8/19/72.  At this time, the mem-
brane system was equipped with Eastman cartridges in
Stage la; T.J. Engineering cartridges in Stages Ib, lc,
2 and 3; and Gulf cartridges in Stages 4 and 5.

It was immediately apparent that two related problems
were present.  First was that the ultrafiltration rate
declined very rapidly during the course of an experi-
ment due to membrane fouling.  Second/ restoring the
flux by flushing or cleaning the membranes proved to be
difficult.  Subsequent runs showed the identical pattern
of rapidly declining flux after start up, and great
difficulty in membrane cleaning.

This behavior was evaluated, and the following con-
clusions were drawn:
     —the filtration system, consisting solely of the
       Broughton lOia basket filters, was inadequate
        (as discussed in Section V.A.); and
     —the problem was possibly aggravated by the fact
       that under certain conditions it was possible
       for unfiltered feed to bypass the Broughton
       filters and enter the membrane system.  This
       possibility was eliminated by installing a check
       valve in the appropriate place  (Stage 1 recir-
       culation loop) on 8/31/72.

On the basis of these results it was decided to evaluate
other filtration alternatives to avoid or minimize the
membrane flux decline.  In order not to expose the en-
tire membrane system to the possibility of inadequate
filtration by the filtration sequences to be examined,
it was decided to conduct these tests with a single
membrane cartridge.

2.  Operating period:  9/1/72-10/24/72  (single cartridge
tests).  A series of 50 to 100-hr tests were conducted
with various filtration sequences and a single membrane
cartridge.  The first test was with the Broughton filters.
Data for short term operation are contained in Figure 43.
As is evident, there was a very rapid flux decline, which
confirmed the conclusion that the lOy Broughton filter
was inadequate.

A more retentive filtration system was then installed.
The new filtration sequence consisted of:
                          101

-------
o
10
       60
                             Hydromation Filter
                     stopped for
                     pump repair
                                         depressurization
                                                  -£>- —-
            Filter


           	I	
0
                 20
40
60      80      100     120
     Operating Time,  hours
140
160
180
              FIGURE 43;
                EFFECT OF PREFILTRATION  EFFICIENCY ON ULTRAFILTRATION RATE
                (T.J.  Eng.  Co.   Cartridge,  110  psig; lOa'F,  5 gpm circulation rate)

-------
     —a Bauer-hydrasieve filter (to remove fibers and
       gross solids);
     —a Hydromation depth filter (as the main filter);
       and
     —ly Cuno cartridge filters (for final polishing).

The suspended solids removal efficiency of this filtra-
tion sequence was substantially better than that of the
Broughton filters.  This was reflected in a substantial
improvement in membrane flux stability, as seen in
Figure 43.  It was possible to maintain a membrane flux
exceeding 20 gfd with this filtration sequence during
a 130-hr test.  The short term flux decline (i.e. over
24 hrs) was about 25 to 35%, and it was possible to
largely recovery flux by intermittent shut down  (system
depressurization).  Complete flux recovery was achieved
by detergent cleaning.

The cleaning procedure consisted of a low-pressure  (20-
40 psig) and high flow  (5-6 gpm) once-through water
flushing, and detergent cleaning (described in Section
V. F.) .

Unfortunately the filtration capacity of the Hydromation
filter was  less than that required to treat the  feed
needed for  operation of the entire pilot plant.  The
filter vendor was unable to deliver a larger unit within
the budget  and schedule constraints of the program.
Therefore it was decided to install a precoat  filter,
which was expected to give particulate removal equivalent
to that of  the Hydromation depth filter.

As described in the  section on  pretreatment, a Shriver
filter press was available within the mill.  This was
installed in place of the Hydromation  depth filter,
keeping both the Bauer hydrasieve and  the  Cuno polishing
filters in  the system.

The single  membrane  cartridge was operated for about  50
hrs with  the Shriver filter press,  using wood  flour as
a precoat material  and body feed.   This  filtration  se-
quence was  also  found to  effectively remove suspended
solids  as shown by  the data in  Figure  43.   Its filtra-
tion  efficiency was  somewhat  less than that for  the
Hydromation filter,  since the  rate  of  membrane flux
decline  within  relatively  short periods,  e.g. 24 hrs,
was  somewhat more rapid.  However,  it  appeared that
                         103

-------
acceptable ultrafiltration rates could be maintained,
and membrane cleaning efficiency was greatly improved.
Moreover, the Shriver filter press had sufficient
capacity for operation of the entire pilot plant.  Thus
at the end of these tests with the single cartridge it
was decided to resume operation of the entire pilot plant.

3.  Operating period:  10/25/72-10/31/72 (renewed pilot
plant tests).   The pilot plant operation was renewed
on 10/25/72 using the Shriver filter press.  The first
step was to clean the pilot plant according to the
detergent cleaning method developed for the single car-
tridge.  The result was a substantial improvement in
flux for all stages  (see Table 14).  However, during
operation on subsequent days, the flux decline was un-
satisfactory.  This was particularly true for Stages Ib,
2 and 3.  At the time it was concluded that the initial
set of membrane cartridges was irreversibly fouled or
plugged by inadequately filtered feed during August,
and was  not suitable for obtaining meaningful results.
A new set of membrane cartridges was ordered, in order
to conduct the remainder of the experimental program
under representative conditions, that is, with mem-
brane cartridges which had not been subjected to ex-
treme upset conditions.  It now appears likely that
the unsatisfactory flux decline observed at the end of
October was due to a combination of this hypothesis
and inadequate filtration with the Shriver filter press.

4.  Operating period;  11/1/72-11/27/72.  The new set
of T.J. Engineering cartridges was received and installed
in the pilot plant on 11/1/72.  Operation was resumed
with pine caustic extraction filtrate.  Initial fluxes
were high, in the range of 25-35 gfd.  By reference
to Figures 36 to 42 and Table 14, it .can be seen that
performance was substantially improved over that of
any earlier period.  Nevertheless, a continuous decline
in flux and increase in pressure drop across the mem-
brane stages was observed. Complete flux recovery by
the cleaning procedure developed for the single spiral
was not obtained.  However, the system appeared to be
stabilizing at a flux rate level of approximately
15 gfd, but pressure drops in Stages 1, 2 and 3 were
high.

One membrane change was made during this period.  When
a leak occurred in Stage 5 on 11/6/72, the cartridges
were replaced with two of the original Gulf membrane
                         104

-------
cartridges.  These cartridges remained in the pilot
plant until 2/12/73.

During this start up period with the new membranes
several observations were made on the condition of the
cartridges.  These were:

    Date                  Observation

    11/14/72 to           Brine seals of almost all
    11/17/72              cartridges, except the two
                          Gulf cartridges in Stage 5,
                          were found to have "flipped",
                          potentially allowing feed to
                          short-circuit the membranes.
                          Membrane cartridge compres-
                          sion effects were also noticed
                          (see Section V.G.:  Module
                          Mechanical Problems).  The
                          brine seals were reinforced
                          and the cartridges returned
                          to their respective shells.

    12/7/72               The brine seals of the car-
                          tridges in shells la, Ib, and
                          Ic had flipped.  The cartridges
                          were reinstalled with the
                          brine seals properly positioned.

In addition, other limitations of the system became
apparent during this period.  Specifically, the pilot
plant piping did not permit once-through flushing of the
membrane cartridges at low pressure and high flow.  Also,
it was suspected that the Shriver filter press was not
providing a high efficiency for suspended solids
removal.

Thus, the following system modifications were undertaken
and completed by 12/7/72.
     —A continuous wood flour body feed addition
       system was installed.
     —A regular procedure was instituted to apply a
       layer of fine paper on top of the filter cloth
       of the Shriver filter press.  This aided in
       the formation of the precoat layer, as well as
       facilitating filter cleanup.
                       105

-------
     —Piping modifications were made such that each
       membrane shell could be flushed at high flow and
       low pressure on a once-through basis.

5.   Operating period; 11/28/72-1/10/73.  After the changes
to the filtration system and piping had been made, oper-
ation continued on a regular basis.  Additional membrane
changes were made, as summarized below.

                           Change

                           Stage la cartridges leaked;
                           replaced with three TJ
                           Engineering cartridges from
                           the initial set.

     12/19/72              Switched position of Stage
                           la and Ib cartridges.

     12/21/72              Stage Ib(formerly la) car-
                           tridges were replaced by three
                           TJ cartridges from the initial
                           set.

     12/21/72              Installed two Gulf cartridges
                           from the initial set in
                           Stage 4.

     1/4/73                Installed three TJ Engineering
                           wide-channel corrugated-
                           spacer cartridges in Stage 2.
                           Observed that two of the three
                           cartridges removed had torn
                           brine seals.

Stage 1

On 12/8/72, Shell la was observed to have a leak.  The
membrane cartridges were replaced with three of the
original set of TJ Engineering cartridges.  Surprisingly,
these three older cartridges showed a very high flux,
about 30 gfd.

To determine if the position of the la Shell influenced
                         106

-------
performance, the three cartridges in Stage la were
shifted to the Ib position on 12/19/72.  For the two
days during which these cartridges were tested in this
position, membrane flux continued to be high.

On 12/21/72, the three cartridges in the Ib position
were replaced by another three TJ Engineering cartridges
from the initial set, in order to see if they also had
high flux.  As seen from the data in the Table 14,
these three cartridges did indeed exhibit high, stable
flux, and continued to do so for the rest of the test
program.  Specifically, flux remained more or less
constant until 1/10/73, when operation with decker
effluents was initiated.

It was immediately suspected that unequal flow distribu-
tion between the parallel shells in Stage 1 caused this
behavior.  Just before beginning the tests with the
decker effluent, flows through the shells in Stage 1
were measured.  It was observed that approximately 65%
of the total flow was passing through Stage Ib, 30%
through Stage la, and only 5% through Stage Ic.

It was concluded from these results that it is neces-
sary to maintain a high flow through spiral-wound car-
tridges with standard mesh spacers to maintain a satis-
factory flux.  High flow also results in facilitated
cleaning of the membrane cartridges and flux recovery.
Furthermore, in an actual plant installation, flow
distributors between parallel shells will be required.
Finally, it appears that a minimum circulation rate of
about 8 gpm will be required to maintain high flux.

Data showing performance of these membrane cartridges
operating under acceptable conditions are contained in
Figure 44.  Shown is flux as a function of operating
time.  During any period between cleaning, flux de-
creased due to reversible membrane fouling.  However,
membrane flux was completely recovered during cleaning.

Stage 2

Stage 2 fluxes ranged between 15 and 22 gfd until mid-
December.  After this time, flux began to decline slowly.
Pressure drop for Stage 2 increased to a level of about
70 psi  (at 4.5 gpm circulation rate) on 11/29/72.  After-
                         107

-------
•a

 *.
x
rH
t,
40
30
20
10
   1/4/73
                     Pine Caustic Extraction Filtrate
               Decker Effluents
              - -•>-1 -^	 1:-	-3-
                                                                                                        -f*-

                                                                   _L
                         1000
?
                                                                   1100
                                                              1/18/73
                                      Cumulative  Operating Time, hours
1200
                                               FIGURE  44

                                  ULTRAFILTRATION RATE DATA:   STAGE Ib

-------
wards/ pressure drop decreased and this was due to brine
seal reversal.  On 1/4/73, when the Stage 2 cartridges
were removed to replace them with three new TJ Engineer-
ing cartridges  (wide-channel, corrugated spacers), it
was observed that two out of the three Stage 2 cartridges
had brine seal failures.  It is likely that after brine
seals failed, feed partially bypassed the cartridges
which then slowly fouled due to low flow (see Table
14).

Three corrugated spacer cartridges were installed in
Stage 2 on 1/4/73 to determine if this "hydrodynamically
clean" flow channel spacer would permit operation with-.
out cartridge plugging.  It is generally necessary to
operate this type of cartridge at a higher feed flow
rate than in mesh spacer cartridges.  The open cross-
sectional area in corrugated space cartridges is greater,
and unless one operates at the same or higher linear
velocity than in a mesh-spacer cartridge, concentration
polarization can reduce flux below the water flux of the
cartridges.  It was anticipated that the flow rate re-
quired to overcome this effect of concentration polari-
zation would exceed the 5-6 gpm which the Stage 2 circu-
lation pump could deliver.  Thus, the purpose of the
test was not to collect meaningful ultrafiltration rate
data but to demonstrate that the cartridges could oper-
ate without flow channel blockage and the corresponding
irreversible buildup in pressure drop and flux decline.

An examination of the data in Figure 45  shows that no
long-term degradation in flux was obtained in the tests
with the pine caustic extraction filtrate.   (The data
for the decker effluents are explained below). The
clean module rates, were observed to be  about 50% of the
water flux of the cartridges, and this demonstrates that
concentration polarization was relatively severe.  By
contrast, flux of pine caustic extraction filtrate in
all other stages was generally observed  to be identical
to the water flux.

Since the initial flux of the corrugated spacer  car-
tridges could be recovered repeatedly, it is  concluded
that the use of this flow channel spacer can  result in
complete membrane cartridge cleanup.
                         109

-------
  30
g
8 20
tr
x
  10
   0
  1/4/73
                    Pine Caustic Extraction Filtrate
                                    Decker Effluents
                          1000
                                  / 1100
                               1/18/73
Cumulative Operating Time, hours
                      FIGURE 45.  ULTRAFILTRATION RATE DATA:  STAGE 2

-------
Stage 3

The membrane flux of the Stage 3 membrane cartridges
remained between 12 and 18 gfd througout this test peri-
od.  It is concluded that neither brine seal failure
nor substantial irreversible flow channel plugging
occurred.  The gradual decline in flux is attributed to
irreversible membrane compaction (see Section V.F. for
the effect of time on water flux of these membranes).

Stages 4 and 5

The flux of Stage 4 also remained reasonably high, until
12/12/72, at which point a leak occurred and flux was
no longer measured.

Membrane flux for the Gulf cartridges in Stages 4 and 5
generally ranged between 2 and 6 gfd, and this is
attributed to the lower flux characteristics of the Gulf
membranes.  It is also possible that these membrane
cartridges may have partially dried out during the
period that they were not used.

6.   Operating period;  1/18/73-2/2/73  (Operation with
Decker Effluents). During the end ot January, 1973, the
pilot plant was operated for periods of approximately
40 hrs with hardwood and pine decker effluents.  Un-
fortunately, these tests were conducted when the pilot
plant system was not operating with acceptable flux in
most of the stages.  Thus, the tests with the decker
effluents provide preliminary information on the separa-
tion efficiency of the membranes and, to a limited ex-
tent, values of membrane flux.

Referring to Table 14, meaningful flux can be obtained
from the data for Stages Ib, 2 (corrugated spacer car-
tridges) , and 3.  The data for Stage Ib and 3 indicate
that the flux for the two decker effluents was some-
what less than flux for the pine caustic extraction
filtrate.  This was more pronounced for Stage 2.  This
behavior is explained by the following hypothesis.  The
organics contained in the decker effluents have higher
molecular weights than those in the pine caustic
extraction filtrate, since fragmentation of the organics
in the latter waste occurs in the bleach plant chlorina-
tion stage.  Thus, at a given retained organics loading,
                         111

-------
concentration polarization should be more severe for
the decker effluents since the diffusivity of the dis-
solved solutes is lower.  For this reason, flux in
Stage 2 was reduced by almost a factor of two below
that for pine caustic extraction filtrate.

The lower flux in Stage 1 may also be due to a gradual
failure of the Stage 1 pump.  It was discovered that
the capacity of the Stage 1 pump was low and this could
have resulted in a flux reduction due to concentration
polarization.  Also, near the end of the tests with the
decker effluents, the Stage 2 pump began to fail, and
this may account in part for the reduced Stage 2 flux.

7.   Operating period;  2/3/73-3/1/73. At the end of
the tests with the decker effluents it was apparent that
additional modifications to the pilot plant were desir-
able.  The following changes were made:
     —a Sparkler pressure leaf filter replaced the
       Shriver-filter press;
     —a Sparkler-Velmac 5y disc filter (backwashable)
       replaced the Cuno cartridge filters;
     —the Stage 1 and Stage 2 pumps were repaired,
       and their flow capacity was restored.

During the first week in February, 1973, all the TJ
Engineering cartridges were tested in the single car-
tridge test shell.  On the basis of the results of these
tests, the best membranes were selected and installed
in the pilot plant.  Details are given in Appendix B.
Briefly, the cartridges in Stages Ib, 2, and 3 were not
changed;  the four Gulf cartridges in Stages 4 and  5
were replaced by six TJ Engineering cartridges;  two of
the three cartridges in each staqe had been previously
used in the same stage in November, 1972.  Stages la and
Ic were filled with other good remaining TJ Engineering
cartridges.

Initially some persistent problems with O-ring leaks
were encountered, but eventually these were solved.
Operation began on 2/12/73 with pine caustic extraction
filtrate.  In general, membrane fluxes were similar to
those observed in November and December.
                         112

-------
The membranes in Stages Ib and 3 had not been changed.
In treating pine caustic extraction filtrate, the
flux for Stage 3 returned almost to the original level.
But in the case of Ib there was a certain amount of
irreversible flux loss.  The reason for this is not
clear.

In Table 14 flux levels are shown for both the begin-
ning of an experiment  (approx. 1 hr) and the end
(usually 8 to 17 hrs).  The flux decline for Shell Ib
usually did not exceed 20%, compared to declines of
40 to 50% for Shells la and Ic.  This is presented
graphically in Figure 46, which contains data for
Shells la and Ib.  It is seen that the initial water
flux for the two shells were identical.  However, the
initial flux with pine caustic extraction filtrate was
higher for Shell Ib and did not fall off as rapidly as
that of Shell la.  Furthermore, membrane cleaning
recovered the initial flux for Shell Ib, while for
Shell la the initial flux decreased with time.  In
addition, the water fluxes of the two stages diverged,
with water flux for Shell la declining relative to that
for Shell Ib.

At this point, circulation flows in the different shells
were measured and it was found that the flow through
Shell la was approximately 1.3 gpm;  Shell b, 9.5 gpm;
and Shell Ic, 4 gpm.  So again it was apparent that the
shell with the highest circulation, flow exhibited
superior performance.

In general, flux for the pine caustic extraction filtrate
was equal to the water flux of the membranes.  Notable
exceptions were the behavior in Stage 2, with corrugated
spacer spirals, and in Stage 5.  As can be  seen in Figure
42, the Stage 5 flux for the TJ Engineering  cartridges
was more or less equal to the Stage 5 flux  for the Gulf
cartridges, even though the TJ Engineering  cartridges
had a substantially higher water flux.  This demonstrates
that at least in Stage 5 the feed concentration at-
tained a level sufficiently high to reduce  flux, that
is, flux in this stage was limited by concentration
polarization, and not by the membrane water  flux.
                         113

-------
fr.
   12
    8
                                                   Stage  la,~'  pine  Caustic.
                                                             '  Extraction
                            G  Stage Ib,
                            4  Stage la,^
                            I  Stage Ib,
                                                               Filtrate  Flux
                                                              Water
                                                             ' 72°F,  90 psi
   °J
2/20/73
            10
20
30      40      50
 Operating Time, hours
60
70
80
90
                 FIGURE 46:  FLUX VS TIME FOR STAGES lA and IB

-------
     Since the color rejection of the TJ Engineering
     cartridges was lower than that of the Gulf cartridges
     (more color was observed in the Stage 5 permeate)
     it would be desirable in a full-scale installation
     to use lower flux, more selective membranes in the
     final stages).

D.   PRESSURE DROP DATA

     Pressure drop  (and circulation rate) data for the five
     stages are shown in Figure 47 to 51.  Pressure drop was
     strongly dependent on circulation rate, as can be seen
     from inspection of the data.  For example, when the
     Stage 1 pump lost capacity  (flow), pressure drop across
     this stage decreased sharply.   (Note:  one would expect
     turbulent-flow pressure drop to increase with the 1.8
     power of circulation rate.)

     During November, pressure drops for all stages became
     quite high, exceeding 50 psig in the early stages.  In
     some stages  (e.g. Stage 2), brine seal failure led to
     a reduction in pressure drop since feed flow bypassed
     the membrane cartridges.

     The high pressure drop was due to inadequate prefiltra-
     tion and an inability to flush and clean the cartridges
     thoroughly.

     The changes made in the pilot plant in early December,
     piping modifications to permit once-through flushing and
     the use of paper on the plates of the Shriver filter
     press, resulted in improved membrane cleaning.  This
     was generally reflected in lower pressure drops in the
     pilot plant after mid-December.  The high pressure drop
     in Stage 1 in February, 1973 was due to operation at
     increased flow.

     Pressure drop across Stage 2 when corrugated spacer car-
     tridges were used was negligible, and no increase with
     time was observed.

     Thus, it is concluded that negligible flow-channel
     plugging occurred with the corrugated-space cartridges.
                              115

-------
  20
  10
o;
             • •

            •   tt
                                       *••  •

                      M   •
                                                                     .V
  60
  50
V)
a 40

M
Q

0)
M

« 30
0)
0)
M
  20
  10
                                                                                     A  A
4  44
  4

4    4
                                                       4 4

                                                        4
           4  44
                     **
                                           4

                                          4
                                 4 4 4t
                            4    4  A
       Auqust      September   October      November     December    January     February       March
                                                               1972 1973
                                    FIGURE 47:  STAGE  1  PRESSURE DROP

-------
fa
•a

-------
& 6

8 «
1  2
0)
                                        • t    ••
                                                     V


                                                      •*••  •%*%•  *—
  60
§40

0)
n
9
CO
n
o*
  20-
  1C-
                4

                4
4     44

 4   4

   4
                                                                                      44
                                                                          44*  44




                                                                            4 4M
4



4  4
                                        44


                                         4
      August      September    October    November     December     January     February      March
        y                                                     1972 1973

                                    FIGURE 49:  STAGE 3 PRESSURE DROP

-------
 •O   2
 
-------
i
                                               •        •    •      /•   -            % •
                                         •»  •       »•»•»•••  r •    •«•«•    *•
             V


H   4
T3
0>
0)
0>   9
    2
                                                                                    • w  •
                                         .      •        ft    •      * •
             •
                •
   60
   50
o

                                         • •**


-H  4°'
to
   30

 0)
 M



 S  20
 0)
 M
  10-
                                              *
                                                       •     .
                                       *             A     *A     A4   AM*«*      /  *
                                                            *       *                  A
                                         *             A
                                                      *  A*         *             *


            •*•
       August      September    October    November     December     January     February     March

                                                             1972 1973

                                   FIGURE 51:   STAGE  5  PRESSURE DROP

-------
E.   IMPORTANT FACTORS CONTROLLING MEMBRANE FLUX

1.   Slime Formation on the Membrane Surfaces

     The membrane flux .decline observed was controlled by
     several interrelated phenomena.  It is instructive first
     to consider a "normal" case, in which a continuous and
     uniform feed flow is maintained across the membrane
     surface.  In this case, a "gel" layer of slightly solu-
     ble colloidal and macromolecular solutes can build up
     on the membrane surface with time, reducing flux.  This
     behavior has been reported previously for a variety of
     feed solutions (27)f and undoubtedly occurred in the
     ultrafiltration of the three effluents processed in
     this program.

     During the program "slimes" removed from membrane
     spirals were analyzed, and other tests were performed
     in an attempt to characterize the troublesome "foul-
     ants".

     Analyses of the slimes indicated the presence of large
     amounts of ash in the slime  (60-70%).  The ash con-
     sisted primarily of finely divided clay, calcium com-
     pounds and anti-foam agents such as silicones.  These
     ash materials were traced to paper machine white water
     which was used for a time as make-up for the caustic
     extraction stage.  Beginning in January white water was
     not used for this function.

     Materials similar in composition to the slime are found
     in the effluents of every pulp and paper mill process
     using water which has been in contact with cellulose
     pulp.  In all cases, these effluents can be filtered
     (or even ultrafiltered), and the filtrate when refiltered
     24-72 hours later will have 10-100+ppm of a grayish floe
     which appears similar to the material removed from the
     membrane surfaces.  This suggests that the minimal
     membrane fouling and slime formation observed in earlier
     laboratory experiments may have been a result of the
     age of the material studied.  The slime-forming material
     may have flocculated during the time of sample shipping
     and most of it may have been removed in the laboratory
     prefiltration.
                              121

-------
In order to develop a better understanding of the
mechanism of slime formation, a series of tests and
analyses were conducted. The slimy solids were examined
bacteriologically in several laboratories.  No evidence
was developed to indicate the presence of any micro-
organism.

An extensive experiment was conducted on fresh pine
caustic extraction filtrate.  The fresh sample was
adjusted to a given pH.  Suspended solids were deter-
mined by filtration of 500 cc samples through Whatman
No. 40 filter paper.  The filtrates were allowed to
stand for 24 hrs at room temperature and at 120°F.  The
suspended solids in these filtrates were then deter-
mined as before.

A typical set of data from these experiments is given
below:

          Pine Caustic Extract Sample 12/7/72
Adjusted
  PH
 (original)
12.5

 8

 7

 6
                  Suspended Solids in
                 Filtrate after 24 hrs,ppm
Suspended        at Room
Solids,  ppm   Temperature     at 120°F
  53
  43

  39

  58

 113
 4

29

41
 1

28

43
From experiments of this type it appears that part of
the problem of prefiltration and membrane fouling may be
due to pH adjustment to pH 7 or below.  The mechanism
which causes the slimy material to appear acts as
though it is a micellar system with an isoelectric point
between pH 6 and 7.
                         122

-------
A series of detailed chemical analyses were conducted
on the solids formed in the parent liquor, the retentate
and the concentrate as well as the solids flushed out
of the pilot unit during the washing operation.

One of the samples most fully investigated was Sample
#3, January 19, 1973, which was a water flushing from
the pilot plant.  Details are given in Appendix E.

From such analyses, it appears that the basic slime
forming material is polysaccharidic in nature and there
is some component of long chain hydrocarbon materials.
Materials derived from the cellulosic fibers are
believed to be the basic contributors to this slime.

One other set of observations of merit in this discus-
sion is that during ultrafiltration of both caustic
extract filtrates and decker effluents with bulk pH of
about 7, the concentrates displayed pH's of about 5-6
and the permeates had pH's 1-2 points higher.

From considerations such as these, it is hypothesized
that the slime forming and membrane fouling constituents
are present in the effluents predominantly as small
miceliar structures.  These constituents are derived
from the cellulose fibers and may be stabilized  in the
liquid by small amounts of tall oil soaps.

It is hypothesized further that the concentration ef-
fects at the membrane surface, and the concommittant pH
depression, cause the micellar structure to agglomerate,
resulting in the slime formation on the membrane sur-
face.

The results obtained thus far in developing some under-
standing of this solid formation problem suggest that
the formation of these materials, and membrane  flux
decline, can be minimized by:
     —operating the system at as high a pH as
       feasible consistent with the use of cellulose
       acetate membranes;
     —operating the membrane modules at high  feed flow
       to minimize concentration polarization.
                          123

-------
2.   Plugging of Membrane Cartridge Flow Channels

     The formation of slimes on the membrane surface can be
     greatly aggravated by blockage of the cartridge flow
     channels.  Specifically, if the flow channels become
     partly blocked, flow channeling through only part of the
     cartridge results in a part of the membrane area being
     exposed to stagnant, or nearly stagnant, feed.  Concen-
     tration polarization and slime formation become severe,
     and part of the membrane area becomes ineffective.  Also,
     when the brine seals "flipped" flow bypassed the car-
     tridges, at least in part, and the reduced net feed
     flow led to greater concentration polarization and
     membrane fouling.

     The flow-channel plugging problem was encountered for
     all membrane cartridges with standard mesh spacers.
     The cartridges in shells la and Ib which operated at
     very high feed flow (^7-10 gpm) showed substantially
     less collection of solids - as concluded from flux
     stability and nearly complete recovery of water flux on
     cleaning.

     Other cartridges which became plugged did so irreversibly,
     Solids continued to accumulate with time, until the
     brine seals flipped.  Autopsies of a few cartridges with
     high pressure drop showed extensive amounts of a grayish
     gelatinous material which filled the mesh spacer, and
     presumably, occluded membrane area.

     For the TJ Engineering cartridges another problem arose.
     The brine seal was on the downstream end of the cartridge
     Thus, the entire cartridge exterior was at the cartridge
     inlet pressure, while the interior pressure was at the
     inlet pressure, less the pressure drop to that point  in
     the cartridge.  If a 20 psig pressure differential
     existed across a cartridge (inlet to outlet), the static
     pressure difference between the exterior and interior,
     at the outlet end, would be 20 psig.  This resulted in
     a net compressive force being exerted on the cartridges.

     Since the TJ Engineering cartridges were loosely wound,
     extensive deformation occurred.  This further aggravated
     the internal flow maldistribution problem.
                              124

-------
     Finally, in addition to flow channeling within a car-
     tridge, channeling occurred among the Stage 1 shells.
     This appears to have been an unstable situation.  When
     a flow maldistribution occurred, the shell with high
     flow remained clean, while the low flow shells continued
     to collect solids and plug since the low flow was not
     able to sweep the accumulated particulates out of the
     modules.

3.   Prefiltration Efficiency

     The efficiency of suspended solids removal greatly in-
     fluenced the degree of membrane cartridge plugging by
     suspended solids, and hence, the membrane fouling
     characteristics as discussed above.  In addition,
     depth and precoat filtration could have removed colloidal
     material which contributed, in part, to the slime
     formation.

4.   Feed Circulation Rate

     As discussed above, increasing the feed circulation rate
     reduced the flux decline for two reasons.  First, it
     reduced the rate of particulate collection within the
     modules.  Second, it reduced concentration polarization
     and, hence, the rate of slime formation.

F.   GLEANING PROCEDURES AND EFFICIENCY

1.   Procedures

     The cleaning procedure that was found to be the most
     effective consisted of the following three steps:

          Step 1: Low-pressure and high-flow water flush of
                  the pilot plant in a once-through manner.
          Step 2: Low-pressure and high-flow detergent
                  cleaning for about 30 minutes with recircu-
                  lation of the detergent solution.
          Step 3: Repetition of Step 1 at the end of the
                  detergent cleaning.

     The water flushing was the most important part of the
     cleaning procedure and was done at 20-50 psig and 2-3
     gpm per shell.  Reverse-flow flushing was found to clean
     the system faster than forward-flow flushing.
                              125

-------
     Detergent cleaning was not always required and sometimes
     was performed on alternate days.   However, on most
     occasions, detergent cleaning helped in removal of
     the gel-type foulant at a faster  rate.   After 12/5/72,
     the pilot plant was cleaned only  at the end of every
     15-22 hrs of continuous operation.  The detergent
     cleaning part consisted of circulating about 40 gallons
     of cleaning solution, containing  about 1% neutral or
     slightly alkaline detergent, for  about 30 minutes.

     It was found that either one of the following sets of
     operating conditions for detergent cleaning was
     effective:

          Procedure used during    10-40 psig pressure
          December and January
                                   2-3 gpm circulation rate
                                   per shell

                                   100°F

          Procedure used during    40-70 psig pressure
          October, February and
          March                    4-5 gpm circulation rate
                                   per module

                                   100°F

     The following detergents were used for cleaning:
          a)  Ultraclean, Part A.; Abcor, Inc. (a phosphate
              type low-alkaline detergent).
          b)  Tide (pH adjusted to 7).
          c)  1% EDTA + 1% Triton X-100 (Rohm & Haas Co.).
          d)  1% citric acid + 1% Tergitol 15-S-7 (Union
              Carbide).
          e)  Ultraclean  (enzyme detergent,  Abcor, Inc.).

     It was found that cleaning efficiency was not noticeably
     dependent on detergent type.  Ultraclean Part A, which
     does not require pH adjustment, was used most frequently,
     Tide was also used on several occasions.

2.   Effect of Feed Prefiltration on Cleaning

     Table 15 summarizes the effect of feed prefiltration and
     different cleaning procedures on  pilot plant performance
                              126

-------
                                            TABLE  15

                           EFFECT  OF  FEED PREFILTRATION ON  CLEANING
Operating
 Period
   Feed
Filtration
 Systems
                                    Operating
                                    Time , hrs
                                                  Cleaning Procedures
                                     Water
                                    Flushing
                  Detergent
                   Cleaning
     Cleaning
    Efficiency
8-19-72 to

    9-26-72
Broughton 10y
  Filter
                                   4-10
Inadequate;
needed modi-
fications in
the pipelines
                                                                   es
Poor.  Required about 4 hrs
(2-3 times repetition of
cleaning cycles)  to recover
flux.
9-27-72 to

    10-5-72
a. Hydromation,
   depth filter
      and
b. 1v Cuno
   cartridge
   filter
                                    20
                                                                  Yes
                              Improved.  Required about
                              2 hrs to clean the system.
10-17-72 to

     11-27-72
   Plate & Frame
   Press with
   wood flour
   precoat & body
   feed; and
   1v Cuno
   cartridge
   filter
                                    20
                                                                  Yes
                              Required about 2-2*$ hrs
                              to clean the system.
11-28-73 to

     2-2-73
                                    20
                                 Modified pipina
                                 improved flushing operation.
                                 See details in text
                                 for operating conditions.
                              Required about l-l>s hrs
                              to clean the system.
2-11-73 to

     4-1-73
   Sparkler leaf
   filter with
   wood flour
   precoat & body
   feed; and
   1 v Cuno
   cartridge
   filters
                                    20
                              Required about 1  hr to
                              clean the system.

-------
     and cleaning efficiency.   The pilot plant ultrafiltration
     rate and cleaning efficiency were very poor at startup
     of the plant (8/19/72 to 8/30/72), at which time the
     filtration system was grossly inadequate.  By contrast,
     the pilot plant could be cleaned within an hour after
     the installation of efficient filtration and flushing
     systems.

3.   Effect of Intermittent Shutdown on Ultrafiltration Rate

     During a study with the single cartridge it was observed
     that substantial flux recovery was possible with a
     brief shutdown of the unit.  Figure 43 shows the effect
     of shutdown on this recovery.  However, during the pilot
     plant operation similar flux recovery was not obtainable.
     The probable reason for this is that, in the single
     cartridge case, the particulates and membrane foulants
     were released during shutdown and  on restarting were
     flushed out of the system.  There was negligible recircu-
     lation of foulants in the single cartridge case. How-
     ever, during the pilot plant operation, foulants released
     during shutdown were picked up again either by the same
     membrane stage or a subsequent one because of the high
     degree of internal recirculation.  This was the major
     reason for poor cleaning efficiency of the pilot plant
     during the period when it was not possible to flush on
     a once-through basis.

     Intermittent shutdown of the system during cleaning was
     also found to improve the cleaning operation.  The same
     held for flow interruption during the water flush cycle.

4.   Cleaning Efficiency in Different Membrane Stages

     The cleaning efficiency of the membrane cartridges is,
     for very obvious reasons, dependent on the flow distri-
     bution within the cartridges.  If the brine seal of a
     cartridge fails and creates a short circuit, flow through
     the cartridge will be reduced, and hence the effective-
     ness of a flush or cleaning cycle.  Also, if the flow
     distribution within the different parts of the brine
     channel is not uniform, those parts where there is lit-
     tle flow will remain fouled.  Many of the TJ Engineering
     cartridges became deformed under the compressive force
     discussed above, thereby affecting the flow distribu-
     tion within their brine channels.  These cartridges,
                              128

-------
     compared in particular to the Gulf cartridges,  showed
     high pressure drop and inadequate cleaning efficiency
     as measured by water and effluent fluxes.

     Membrane cartridges with poor ultrafiltration performance
     were not readily cleaned.  Both characteristics can be
     traced to unsatisfactory flow distribution within the
     cartridges.

     Stage 1 membrane modules were arranged in three parallel
     shells.  The cleaning efficiency and the ultrafiltration
     rate of these shells were dependent on flow distri-
     bution among the shells, in addition to the flow distri-
     bution within each cartridge.  The variation in flow
     distribution in the three shells depended upon  the type
     and history of individual cartridges and their  varying
     resistance, as discussed before.  Figure 52 shows the
     effect of cleaning on Stage Ib.  It is seen that clean-
     ing was reasonably effective, and that the ultrafiltra-
     tion rate remained relatively high.  During the same
     period, Stages la and Ic however showed lower flux.
     These cartridges were inspected and found to have under-
     gone severe deformation  (compression).  They had, as a
     result, low flow during cleaning and ineffective
     cleaning.

     The three wide channel corrugated spacer cartridges
     installed in Stage 2 were cleaned most easily.

     Figure 53 shows the effect of cleaning on the Stage 3
     membrane cartridges.  These cartridges had the longest
     operating life, from 11/1/72 to 4/1/73.  Figure 53
     shows that the water flux of these cartridges decreased
     with time.  The main reason for this decline appears to
     be membrane compaction.  The flux decline parameter,
     measured by the slope of a log-log plot of water flux
     vs. time was about -0.078  (Figure 54).

G.   MODULE MECHANICAL FAILURES

     During the operation of the pilot plant a variety of
     failures with several of the 45 TJ Engineering car-
     tridges used occurred.  Mechanical problems with the
     three Eastman and four Gulf cartridges were negligible.
     Appendix D details the different problems experienced
     during the pilot plant operation.  By and large, these
     problems appeared to be four fold:
                              129

-------
                 O  Water  flux before cleaning




                 A  Water  flux after cleaning
  18
                        A
                      A
                                       A
                             A
o
CM
  16
  14
                            (§>
   O
              O O
                      A
-H 12
(0
G
o
r-
  10
   8
0)
              I
     I
                December      January    February       March

                       1972 1973


          FIGURE 52:  CLEANING EFFICIENCY OF  STAGE Ib MEMBRANES
                          130

-------
              0   Water flux before cleaning {60  psig,  72°F)
3
rH
6-,
              A   Water flux after cleaning (60 psig,  72°F)
              O   Pine Caustic Extraction Filtrate Flux (100 psig,
                                                    	100°F)
    18
    16
    14
    12
    10
     6
	  O
                  O     O   CO
           A

        	     A
     A
I—   A
      o
            i
                               O
                O
                       A    A
                               A  A
                                A
                                       O
                                O
                                    O
                        O
                                                 COO
                                                           A
                                                            o
                                                      i
                                                    February



           FIGURE 53:   CLEANING EFFICIENCY  OF STAGE 3  MEMBRANES
       November     December      January
                          1972  1973
                          131

-------
   iocr
   so r-
                            Slope = -.078
M
0)
-P
it)
                             1
1
                        10                100

                      Cumulative Operating Time, hours
                        1000
                   FIGURE 54:   STAGE 3 COMPACTION CURVE
                                132

-------
     —O-ring failures;
     ~-membrane cartridge compression, resulting in
       deformation of the cartridges;
     —brine seal failures; and
     —glue seam or membrane failures.

1. O-ring Failures.  Membrane leaks in most cases were
due to inadequate O-ring seals on the permeate collec-
tion tubes, and poor color rejection was corrected
simply by replacing the O-ring seals.  On at least five
different dates involving stages la, Ic, 4 and 5, the
poor color rejections were definitely identified as
O-ring failures.

2. Membrane Cartridge Compression.  TJ Engineering
cartridges have brine seals on the downstream end, thus
the exterior pressure on the cartridges is higher than
the interior pressure due to pressure drop of feed
flowing through the cartridge.  Several cartridges were
observed to "deform" and "shrink" from their round
shape.  These cartridges were primarily in stages where
severe pressure drop across the cartridges was present.

It is suspected that module compression created dead-ends
which could not be cleaned, thus resulting in low ultra-
filtration rate and accumulation of fouling materials.
The latter phenomenon created additional pressure drop,
consequently more compression, and further aggravated
the problem.

Gulf cartridges, however, did not show this problem.
Gulf cartridges have brine seals on the upstream end,
and the pressure within the cartridge is higher than
the exterior pressure.  The net result is that the Gulf
cartridges received an "expansive" force, which maintained
the brine channel width intact and prevented dead-ends.

3. Brine Seal Failures. As the membrane cartridges
collected solids and/or underwent compression, pressure
drop across the cartridges increased.  This resulted in
brine seal reversal, permitting feed flow to bypass the
membrane cartridges.  It is possible that this effect
accounted for gradually declining flux in several of the
membrane she11s.
                         133

-------
     The  four Gulf  cartridges which were  in  the  system for
     a period of  about  seven months did not  evidence  brine
     seal reversal.  At the end  of  the  test  period, how-
     ever,  three  of the four brine  seals  were  observed to
     have weak  tapes which needed reinforcement.   Several
     TJ Engineering>cartridges required replacement of brine
     seals (mainly  foam gaskets  and tapes).

     The  outer  wraps of almost all  the  cartridges  remained
     in good shape  despite the compression or  expansion
     forces (10 to  20 psi per cartridge).

     The  pressure drop  data show that the brine  seals of the
     spiral cartridges  probably  failed  to maintain a  proper
     seal when  the  pressure drop across the  seals  reached
     about 20 psi.

     4. Glue  Seam or Membrane Failures.  In  some instances,
     leaks can  only be  attributed  to membrane  cartridge
     failures.   There have been five failures  of this kind
     involving  stages  lbr  lcr  4 and 5.   All  cartridges w.exe
     from TJ  Engineering Co.

     Membrane cartridge failures may have happened because
     of
          —leakage at  the glue seam;
          —leakage at  the material fold adjacent to the
           permeate collection tube;
          —membrane and/or backing material wrinkles
           causing direct leakage;
          —pinholes in membrane.

     One membrane cartridge with poor color rejection was
     autopsied  and was  found to have glue seam failures in
     several  places, especially at points where longitudinal
     wrinkling  was present.   The Gulf cartridges had mem-
     branes directly cast on membrane-support sheets.  Gulf
     has reported that  spiral modules made in this manner
     show more  uniform brine channel thickness  (28_) .

H.   INCINERATOR STUDIES

     Disposal of the concentrate from the ultrafiltration of
     pulp mill  effluents is a problem only if the concentrate
                             134

-------
contains substantial amounts of chlorides.   In the
studies discussed here it was found that disposal of
the concentrate from the pine caustic extraction fil-
trate does require special processing, but that the
concentrate from the decker effluents can be bene-
ficially used in the weak black liquor system.  The
following discussion applied, then, only to the concen-
trate from the pine caustic extraction filtrate.

The concentrate produced in ultrafiltration of the
caustic extraction filtrate would contain about 20%
total solids.  These solids would be predominately
chlorinated lignins with some amount of sodium salts
such as sodium chloride.

The content of organic and inorganic chlorides in the
concentrate makes disposition a problem if it is to be
injected into any of the present pulp mill streams.
For example, if the concentrate were processed in the
black liquor recovery system, about 500 pounds per day
of hydrogen chloride would be liberated in the recovery
boiler from combustion of the chlorinated lignins and
also about 500 pounds per day of sodium chloride would
be introduced into the recovery boiler from the inor-
ganic solids in the concentrate.  These chloride-con-
taining materials could lead to corrosion problems in the
boilers, dusting problems in the boiler stacks due to the
liberation of finely divided sodium chloride, and also to
raising the chloride level in the pulping system, as a
function of the material recycle system.

Injecting the concentrate into the lime mud kilns would
lead to similar problems.

As a result of the engineering evaluations it was con-
cluded that the disposition of the concentrate from the
pine caustic extraction filtrate would require either an
incinerator specifically designed to burn organic
chlorides, or, evaporation to high solids and subsequent
admixture with primary sludge for disposal as land fill.

Laboratory studies have been conducted on evaporation of
concentrate to various levels of solids.  In all of the
tests the material dried with no apparent increase in
viscosity to about 80% solids.  The material does not
scale nor foam when concentrated by evaporation.  When
                         135

-------
a 507. solids concentrate was mixed with primary sludge,
the admixed sludge was judged to be satisfactory for
land fill.  As a result of the laboratory tests it was
concluded that evaporation of the concentrate to about
50% solids could be done in commercially available equip-
ment using waste hydrogen available on the plant site
as the heat source.  The 50% concentrate would be added
to the dewatered primary sludge and carried to land fill.
This method of disposal would add 10% to the present
daily primary sludge weight.

Concentrate from the ultrafiltration unit was incinerated
in February at the John Zink Company, Tulsa, Oklahoma
pilot facility.  The purpose of the tests was to
establish parameters on which to base a budgetary
estimate for a full scale incineration system to dispose
of the concentrate.

The tests demonstrated that the Zink incinerator will
provide a method of disposing of the concentrate.

The operation of the incinerator requires the addition
of supplementary fuel.  To minimize operating costs,
hydrogen from the electrochemical cells, which is
presently unused, would be burned.

The tests demonstrated the need for the use of a venturi
scrubber and a vent scrubber to dispose of the poten-
tially plume-forming, finely-divided inorganic salts
formed in the combustion of the concentrate.  In a full
scale installation, the pine caustic extraction filtrate,
prior to neutralization, would be used as the scrubbing
fluid for the venturi scrubber and the stack scrubber.
The partially neutralized pine caustic extraction fil-
trate would then be treated in the ultrafiltration
system.

The economic feasibility of the use of an incinerator
for the disposal of the concentrate is heavily dependent
on the availability of an inexpensive fuel source. ' If
hydrogen is used as the fuel source, the fuel costs
will be those associated with the capital and operating
costs of the system required to deliver it to the
incinerator.
                          136

-------
The partial evaporation system for disposal of the
concentrate requires less energy input but does require
increased costs for land fill operations with the
primary sludge.

The capital and operating costs of both systems are
improved, the more concentrated the solids level pro-
duced by the ultrafiltration system.  These costs are
also strongly influenced by the salt rejection charac-
teristics of the specific membrane system used;  the
incinerator system is more cost sensitive than the
evaporation system.  If low salt rejection membrane
systems are used, the ratio of the organic content to
inorganic content in the concentrate will be increased
and for a given solids level the heating value of the
concentrate will be increased.  Increasing the organic
solids content of the concentrate will yield a material
of both higher heating value per unit concentrate
weight and also less water to be evaporated in the
combustor.

The choice of the system for disposal of the concentrate
depends on several design parameters of an ultrafiltra-
tion plant as well as the specific plant site for which
it is designed.  At the present there are at least two
technically feasible methods for disposing of the
ultrafiltration concentrate from pine caustic extrac-
tion filtrate.  In the capital estimates for these
systems  (discussed in Section VI.A.) compromise cost
figures are used.
                         137

-------
                          SECTION VI

               FULL SCALE PLANT DESIGN AND COSTS



A.  FIRST STAGE PINE CAUSTIC EXTRACTION FILTRATE SYSTEM

1.  Flow Schematic

    A general flow schematic has been presented in the
    Introduction as Figure  2.  A second generalized pro-
    cess flow schematic, more specific for the ultrafiltra-
    tion system, is given in Figure  55.  First-stage pine
    caustic extraction filtrate flows from the pine pulp
    bleachery through a pretreatment system.  Pretreatment
    consists of neutralization and filtration.  The treated
    feed is cooled to 100°F.  Although future membranes
    would process pine caustic extraction filtrate at its
    normal temperature  (120-130°F), current membranes prob-
    ably cannot withstand a temperature in excess of 100°F
    and exhibit long life.  It is to be noted that no
    long-term life data are available at temperatures
    higher than 100°F for membrane systems treating pine
    caustic extraction filtrate.

    The pretreated and cooled feed is concentrated in the
    ultrafiltration system.  The permeate (treated effluent)
    is sewered and flows to the mill's waste treatment sys-
    tem.  The concentrate from the ultrafiltration system
    is used to sluice the filter cake, and this suspension
    is pumped to an evaporator and evaporated to 50% solids.
    Excess hydrogen currently available in the mill from
    the electrochemical cells will be used as fuel.  The
    evaporator discharge will be mixed with primary sludge
    and disposed of as land fill.

    A more detailed process flow schematic is given in
    Figures 56 and 57.  These two figures differ in that
    Figure 56  is for a low-flow membrane system (spiral
    wound cartridges with standard mesh spacers), and
    Figure 57 is for high-flow membranes (spiral wound
    cartridges with corrugated spacers).  The reasons for
    examining these two cases are discussed below.

    Sulfuric acid is mixed with pine caustic extraction
                              139

-------
Hydrogen
Evaporator
                                  To
                                               Caustic
                                              Extraction
                                               Filtrate
                                                  Pine Pulp
                                                   Bleachery
                                                                              Water
                                                     Pretreatment
                                                            Filter
                                                             Cake
                                                                               UF

                                                                            Treatment
Permeate
  to
Sewer
                                                                Available
                                                                2 MM gpd
                                                                at 100°F
                                                                                                Concentrate
                 FIGURE  55:   SIMPLIFIED  FLOW  SCHEMATIC:  TREATMENT OF PINE CAUSTIC EXTRACTION FILTRATE

-------
    Feed
Solids
Filter Aid
                                Bodying Pump
                                              Filter Pump
                                                                Dry Dumped
                                                             & Sluiced away in
                                                              UF Concentrate
                                                    To Evaporator *— -7-^
                                                                                                          Back Washed
                                                                                                            into Feed
                   Concentrate
                   used to Slurry     •»—[Xj-
                   Dumped Filter Cake
           Permeate
          from Stage 1
             Water
              for
             Back.
           Flushing
Overflow
to Sump
                    HX—;£**
             Connect
             :o Points A
                              Flushing
                                Pump
                                                Stages
                                              &  3  Required
                                              but  not Shown
  Stage
Circulation
i.   £
                                     Flush to Feed

                                    Permeate Lines (.typical 1
                                                  Flush
                                                 to Feed
                                                                            Stages 2, 3, 4
                                                                     (Stage 1 Lines to Water Store)
                                                                  Sump
                                                                          Permeate Pump
                                                                           (if reqd.)
                                                                                                        Feed
                                                                                                        Pump
                            FIGURE  56:  FLOW SCHEMATIC FOR TREATMENT OF PINE CAUSTIC EXTRACTION FILTRATE:

                                                             LOW FLOV7 CASE

-------
reed
                                                          Dry Dumped and
                                                          Sluiced Away in
                                                          UP Concentrate
                                               To Evaporator •*— -^_.
                Concentrate Used
                  9 Slurry Dun1
                  Filter Cake
        to Slurry Dump«d^__jsxL
       Permeate
   From Stage 1
Sta

 *
     Water
     Store for
     Back-
     flushing
                   Overflow
                   to SumP
                        Flushing
                          Pump
                                        -pad):
                                             ranei
                                            a-
                                          Stage 4
                                         Circulation
                                                                   Stages 2 t 3
                                                                   Required but
                                                                   Not Shown
Flush to Feed

    Permeate Lines (tyfical)
                                                                                 Membranes
                                                      Heat
                                                    Exchanger
                                   Stage 1
                                 Circulation
                                                              Feed
                                                              Pump
                                                                                             Flush to
                                                                                               Faed
                                                         I  Sump   \
                                                                        Stages 2,3,4
                                                                     CStage 1 lines to water store)
                                                                       Permeate
                                                                         Pump
                                                                   (if required)

                         FIGURE 57t   FLOW SCHEMATIC FOR TREATMENT OF PINE CAUSTIC  EXTRACTION  FILTRAT3

                                                        HIGH PLOW CASZ

-------
filtrate in an in-line mixing section.  The rate of acid
addition is controlled by a downstream pH probe and con-
trol system.

A small portion of the neutralized feed flows to a bodying
tank.  Filter aid (tentatively wood flour) is added by a
solids feeder.  The rate of addition of filter aid is
such as to maintain a body feed level of 50-100 ppm.  The
body feed slurry is pumped with the bodying pump back
into the feed line, where it mixes with the neutralized
pine caustic extract.  This stream is pumped through leaf
filters to remove most of the suspended solids contained
in the feed.  Instrumentation for the filter system will
control operating pressure, pressure drop, and mechanical
operation.  Dry cake will be dumped, sluiced away with
the ultrafiltration concentrate, and sent to the evaporator,

A bypass line to the bodying tank on the downstream side
of the leaf filters is used to apply the precoat.  Specif-
ically, after a cake has been dumped, pine caustic ex-
traction filtrate, with filter aid added, is circulated
through the leaf filters and returned to the bodying tank.
When an adequate precoat has built up, normal flow into
the membrane system is resumed.

The feed from the leaf filters,is further treated in back-
washable scavenging filters.  These polishing filters re-
move any filter aid or other large particles which pass
through the leaf filters.  The solids' discharge from the
scavenger filters (backwash) is mixed with the feed pine
caustic extraction filtrate for additional solids removal
in the leaf filters.

The filtered, neutralized feed flows to a surge tank
prior to treatment in the ultrafiltration system.  This
surge capacity allows for short-term interruption in
the feed flow or filter operation, without causing an
immediate membrane system shutdown.  Feed from the surge
tank is pumped through a heat exchanger and cooled
to 100°F.  Flow is then through a series of four mem-
brane stages.  For the low-flow membranes, Stage 1 can
be operated on a once-through basis; that is, it is
not necessary to recirculate concentrate within the
Stage 1 flow loop.  The concentrate from Stage 4 is
used to sluice the filter cake, and flows to the evap-
orator for further concentration.  The permeate from
Stage 1, used for backflushing the membranes, flows
to a water storage tank.  The permeates from Stages 2,
3, and 4 are collected in a sump, and pumped to the sewer
                         143

-------
    by a permeate pump.   The membrane stages will be set up
    for reverse-flow flushing,  an operation which has been
    found to be effective in membrane cleanup.   The Stage 1
    permeate will be used for this purpose.  In addition,
    if detergent cleaning of the membranes is required,  the
    Stage 1 permeate storage tank will be used to mix the
    detergent solution.

    The high-flow membrane system shown in Figure 57 is
    identical to the low-flow system, except that recircula-
    tion in Stage 1 is required to maintain high velocity
    through the membranes.

    Details of the equipment for these process flow schematics
    are given below with cost estimates.

2.  Equipment Description and Capital Cost Estimates

    Five capital cost estimates for full-scale, battery-
    limits plants are presented.  These cases have been
    prepared to provide some insight into the relative cost
    sensitivity of the various parts of the process system.
    The significant process parameters used for each case
    are shown in Table 16.

    The equipment cost estimates presented are based on
    budgetary estimates obtained from potential vendors.
    In the following. Case 1 will be described in detail;
    the other cases will be presented as modifications of
    Case 1.

    Membrane Module Alternatives—Different Flow Channel
    Spacers"  Spiral wound membranes will be used.  These
    membrane cartridges are currently available from Gulf
    Environmental Systems with either the standard mesh
    spacers  (60 ft2/cartridge @ $1.50/ft2) or corrugated
    spacers  (40 ft2/cartridge @ $2.25/ft2).  In the pilot
    plant operation, as described above, most of the mem-
    brane cartridges had the standard mesh spacers.  In
    addition to the mechanical failures encountered with.
    the T. J. Engineering cartridges, all cartridges with
    mesh spacers were difficult to clean.  In the early
    part of the pilot plant operation, an inability to
    thoroughly clean the membrane cartridges resulted in
    increased pressure drop across the cartridges with time,
    as well as reduced capacity.  It was subsequently ob-
                               144

-------
                                                        TABLE 16

                                             CASES FOR CAPITAL COST ESTIMATES

                                             PINE CAUSTIC EXTRACTION FILTRATE
       Pine Caustic
        Extraction
       Filtrate Flow
Case    (x 106 gpd)
   Membrane Modules                     circulation
 Low Flow    High Flow      Membrane        Rate
             (corrugated        Flux,     gpm/membrane
(mesh spacer)   spacer)    aal/dav-£t2     cartridge .
Membrane          Prefiltration           operating
  Cost     Precoat Backwashable Screen    Manpower ,
$/sq ft    Filter  Depth Filter Filter   No. of Men
                                                        18
                                                                                  $1.50
                                                        18
                                                                      15
                                                         $2.25
                                                        18
                                                                      15
                                                          $2.25
                                                        18
                                                                                  $1.50
                                                        18
                                                                                  $1.50

-------
served that improved prefiltration and strict adherence
to prescribed cleaning procedures alleviated the prob-
lems associated with mesh-spacer cartridges.  That is,
a "steady state" operation was achieved.  Specifically,
over long periods of operation it was possible to re-
turn repeatedly to a base-line membrane flux and car-
tridge pressure drop after cleanup.

Experiments with membrane cartridges with corrugated
spacers suggest that it will be substantially easier to
clean these cartridges, possibly by water flushing alone,
In particular, reverse-flow flushing is anticipated to
be substantially more effective for corrugated-spacer
cartridges that for mesh-spacer cartridges.  Thus, cor-
rugated-spacer cartridges offer several possible advan-
tages.  These are:

     —maintenance of a higher average flux  (reduced
       membrane area requirement);
     —facilitated cleanup, reducing down-time for
       cleanup, and possibly eliminating the need for
       detergent cleaning;
     —longer membrane cartridge life; and
     --possible operation without precoat filtration.

These advantages can be of substantial importance in
terms of process costs.  It is to be noted, however,
that these factors have not been demonstrated to date.

The cost figures for the two different membrane car-
tridges given above are identical, i.e. $90/cartridge.
These are the costs currently quoted by Gulf Environ-
mental Systems.  It is not clear why costs should be
independent of the square footage of membrane in each
cartridge.  On the contrary, for large-scale production
costs should be proportional more to membrane area, and
not to the number of cartridges.  For present estimates,
however, the Gulf price structure will be used, al-
though it will tend to give a conservative picture for
corrugated-spacer membrane cartridges.  The two cases
are summarized below.
                 2
     1.  $1.50/ft , for membrane cartridges with mesh
         spacers.  A circulation rate through these
         cartridges of 8 gpm is assumed to be suffi-
         cient to minimize membrane fouling and pre-
                          146

-------
         vent extensive solids collection within the
         cartridges.  These membranes will be assumed
         to have a three-year life.  Precoat filtra-
         tion is required.
                 2
     2.  $2.25/ft , for membrane cartridges with cor-
         rugated spacers.  A circulation rate of 15 gpm
         through the membrane cartridges is assumed to
         be sufficient to maintain cleanliness and
         avoid cartridge plugging.  Membrane cartridge
         life is assumed to be three years.  Precoat
         filtration may not be required.

The relatively high flows through the membrane cartridges
have been selected on the basis of the pilot plant opera-
tion.  For shells (with mesh-spacer cartridges) which
were operated at high feed flow, membrane fouling and
particulate collection within the cartridges was greatly
reduced.  The three-year life assumed is an important,
but not critical, factor.  For example, a two-year mem-
brane life would add about 4C/1000 gal. to the plant
operating costs  (detailed below).

For design purposes, it has been assumed that corrugated-
spacer cartridges can be obtained with 60 ft2 of membrane
area.  The desirability of having a high square footage
per cartridge relates to mechanical problems encountered
in the pilot plant operation.  Specifically, it is
highly desirable to minimize the total number of seals,
gaskets, and glue seams in the system since these are
the points of potential operating problems.

The flow characteristics of the two different membrane
cartridges are given in Table 17.

                       TABLE 17

        CHARACTERISTICS OF MEMBRANE CARTRIDGES

Spacer                   Mesh                Corrugated
Feed Pressure            120 psi             120 psi
Outlet Pressure          80-90 psi           80-90 psi
Circulation Rate         8 gpm               15 gpm
Pressure Drop            8 psi/cartridge     8 psi/cartridge
Maximum Number of
  Cartridges in Series   4                   4
                          147

-------
The operating pressures chosen are 120 psig at the inlet
to the cartridges and 80-90 psig at the outlet from the
cartridges.  At the circulation rates given, pressure
drop per cartridge will be about 8 psi.  Correspondingly,
the maximum number of cartridges which can be assembled
in a series configuration is four.  Thus, four cartridges
will be installed in a single housing, and this will be
denoted as a "membrane module".

In any single stage, capacity will be fixed by the total
membrane area.  Since more than four membrane cartridges
will be required, it is necessary to pipe membrane
modules in parallel.  The number of parallel modules is
determined by the total area requirement.

The costs of membrane modules for cartridges with mesh
and corrugated spacers are given in Table 18.

                       TABLE 18

      CHARACTERISTICS AND COSTS OF MEMBRANE MODULES

Spacers                  Mesh               Corrugated
Circulation Rate         8 gpm ,.,            15 gpm
Area                     240 ft             240 ft2
Bare Membrane Cost       $360               $540
Housing and Installa-
  tion                   $390               $390
TOTAL COST               $750               $930

The modules with mesh spacers will cost $750 each; and
those with corrugated spacers, $930 each.  The cost
difference is due solely to the greater bare membrane
area cost for corrugated-spacer spirals.

Installation includes all racks, internal piping,
shipping, and on-site direct labor.  Excluded are costs
for pumps, the building and foundations, engineering
design, materials procurement, construction supervision,
and startup.

Membrane cartridges with corrugated spacers will prob-
ably be somewhat larger physically than membrane car-
                         148

-------
     •bridges with mesh spacers,  and the housings could cost
     more.   However, for present purposes,  the housing and
     installation costs have been assumed to be the same.

a.   Details of Case__!_ Design

     (i)  Design Bases.  The Case 1 design bases are as follows:

          —2 x 10  gpd pine caustic extraction filtrate
            processed;
          —mesh-spacer membrane cartridges used;
          —precoat filtration needed; and ^
          —membrane flux of 18 gal./day-ft .

     (ii)  Feed Rate and Its Effect on Plant Size.  It has been
     assumed that the pine caustic extraction filtrate flow
     is constant at 2 x 10° gpd (1,390 gal./min).  However,
     membrane flux depends on membrane cleanliness and
     changes with time.  In any case, the plant has been
     sized for a flux of 18 gal./day-ft2, which is an average
     flux  over the operating period between cleaning cycles.
     Flux  varies from 25-30 gal./day-ft2 for clean membranes
     to 10-15 gal./day-ft2 for fouled membranes.

     The simplest way to operate a system is to process feed
     at a  rate greater than 2 x 106 gpd immediately after
     membrane cleaning, and then to reduce the feed flow in-
     to the membrane plant as the membranes become fouled.
     During the latter part of this operation it would be
     necessary to accumulate surplus feed for use during
     the next run immediately after cleaning.  However
     this  requires supplying a large surge capacity for the
     feed,  which would be prohibitively expensive.  There-
     fore,  operation will be such that a small fraction
     (one-quarter or less) of the membrane plant will be
     flushed and cleaned at any single time.  That is, the
     flushing/cleaning sequence will be cycled through
     isolatable sections of the ultrafiltration plant.  This
     will  increase piping and operating costs somewhat, but
     a  net benefit will accrue due to the elimination of a
     requirement for a large feed surge capacity.

     It is  further assumed that each segment of the plant
     can be cleaned in one hour per day.  Consequently, the
     operating time available is 23 hrs/day.
                             149

-------
 Excess membrane  area will  be  required to  produce  the
 permeate  used for membrane flushing.   Since  the plant
 volume is about  10,000  gals.,  conservatively not  more
 than  100,000  gpd will be required for once-a-day  flush-
 ing and cleaning.  This will  increase the membrane  plant
 area  requirement (capacity) by 5% or  less.

 Finally,  some surge capacity  must be  available, and a
 one-hour  surge (80,000  gals.)  will be provided.

• (iii) Plant Size.  The  ultrafiltration system must  process
 2.1 x 106 gals,  in 23 hrs,  at an average  flux of  18 gal./
 day-ft2.   The permeate  rate is essentially the same,
 since more than  98.5% of the  feed must pass  through
 the membranes.   Thus, the  calculation of  the membrane
 area  required is:

 Membrane  area =2.1 x 106  x 24/23 x i  =  122,000  ft2.
                                    J.C
 The minimum size of the pretreatment  section of the
 process,  including the  filters,  is 2.1 x  106 gals,  within
 23 hrs, or 1,500 gpm.   A slightly larger  filtration sys-
 tem will  be chosen to allow for cleaning  the filtration
 section and emergencies.   Specifically, the  filtration
 section will  be  designed to handle 2.1 x  106 gals,  with-
 in 22 hrs, or 1,600 gpm.

 (iv)  Neutralization and Filtration Sections.  The feed
 will  be neutralized with sulfuric acid before filtering.
 Based on  pilot plant experience,  the  feed will be
 filtered  first in a precoat filter, adding filter aid
 to body the liquid at 50-100 ppm.   This will be fol-
 lowed by  a scavenging filter.   The filter aid tenta-
 tively will be wood flour.  Costs  for equipment in  the
 neutralization and filtration  sections are given  below.

                                                Cost
      Acid store  - 500 gals., assumed  to be
      available.                                 n/c

      Acid pump and piping  - The  probable  acid
      use  rate is 8000 Ib/day,  about 0.4 gpm.
      On-off control through a  low pressure
      diaphragm metering pump  (^1/2 hp)  and
      iron piping.                              $2,000
                          150

-------
                                           Cost

pH measurement and control - In line pH
electrodes with a recorder, on-off control
and alarm switches, installed.            $6,000

Filter aid solids handling - Filter aid
will be received in bags.  Daily usage
about 1,000 Ib/day (about 70 ft3/day).
Hopper and controlled rate conveyor.      $3,000

Bodying tank - 600 gal. carbon steel.
5 ft diameter x 4 ft high, with 1/4 hp
agitator.                                 $1,600

Bodying pump - 1 gpm, 1/2 hp.             $  500

Precoat filter - Tests at Sparkler Manu-
facturing indicate that an achievable
filtration rate is 1.5 gpm/ft2.  For con-
servative design purposes, 1 gpm/ft2 is
assumed and 1,600 ft-2 filter area will be
installed.  Two units, Sparkler MCRO-
1000-3, each having 1000 ft2, vertical
leaf, horizontal tank filters for dry
case discharge cost $35,100.  Scaled to
1,600 ft2.                               $30,000

Piping and valving - for the precoat
filters.$ 7,000

Scavenging filters - 10 units Velmac
backflushable felt-type filters, each
13" diameter and 62" high, rated 100-
200 gpm.                                 $10,000

Piping and valving - for scavenging
filters.                                 $10,000

In-line mixing section -                 $'1,500

Instruments and controls - other than
for pH, for measuring pressure and
differential pressure and automating
filter flushing.                         $16,000

Filter pump - Size 8 x 10-13 to pump
1,600 gpm into 94 feet  (40 psi), 60 hp.
Operated at 62% efficiency.  In iron.
Driven and started, not installed or
connected.                               $ 3,200
                    151

-------
                                                Cost

     Back washing and other filter cake re-
     moval - The scavenging filters will be
     back washed into the feed.  The precoat
     filters will be dumped (dry), probably
     2 or 3 times each day.  About 100 ftV
     day of cake will be discharged.  This cake
     will be mixed with 20,000-30,000 gpd of
     concentrate and conveyed either to the
     evaporator (for pine caustic extraction
     filtrate) or to the black liquor plant
     (for decker effluent).                   $   800

     Slurry pump - to remove the mixture of
     filter cake and concentrate, 14 gpm.     $   400
     SUBTOTAL FOR EQUIPMENT                   $85,000

     Transportation, based on non-instrument,
     heavy pieces, totalling about  $63,000
     and coming from 800 to 1,000 miles.      $ 2,500

     Installation, direct labor only, in-
     cluding rigging, pipe fitting  and
     electrical connections of (approx):
         2 pieces filter hydraulic motors, 1/2 hp
         acid pump, 1/2 hp
         dry feeder, 1/2 hp
         tank stirrer, 1/4 hp
         filter pump, 60 hp
         instruments and controls,  20 amp max.
         filter cake mixer, 1/2 hp
         slurry pump, 1 hp
                                              $12,600

     TOTAL PRETREATMENT SECTION               $100,100
     (Engineering design, equipment pro-
      curement, supervision and startup
      not included)
                                         T
 (v) Ultrafiltration Section.  122,000  ft"" of  membrane
area will be used.  This will be incorporated into  four
stages  (Figure 56); details are given  in Table 19.

Each membrane module requires 8 gpm inlet feed flow.
Two-hundred parallel modules have been provided in
                          152

-------
Stage
Number
                           TABLE  19

                ULTRAFILTRATION SECTION  DESIGN

                  DETAILS  AND COSTS—CASE  1
Number of
Modules
             200
 160
                                                  90
          60
Circulation
Rates  (gpm)
Pump Differ-
ential  (psi)
Utilized
Horse Power
            1600
           (feed)
             120
           (feed)
             175
1280
  40
  50
720
 40
 25
480
 40
 20
Efficiency for
Estimating h.p,
              64%
  60%
 68%
 56%
Approximate
Pump Size
            6x8-16   6x8-13   6x8-11
                    4x6
Cost Estimate
(Driven, Started and
Installed
Electrical Connection
           $5,200   $3,000   $2,600   $1,400
Flush Pump:
   Addi tion a1 Equipment

1600 gpm, 40 psi, 60 HP
Permeate Pump:    same as  Flush Pump
Water Store:
20,000 gallons, delivered and
  installed
Flush and drain piping, installed
               $ 3,200,

               $ 3,200


               $10,000,

               $ 9,000,
                            153

-------
Stage 1, requiring 1,600 gpm.  The feed rate of 1,520 gpm
is adequate for once-through operation and circulation
in Stage 1 is not required.  The feed pump delivers 120
psig, and a Stage 1 circulation pump is not provided.

The feed to Stage 2 is 920 gpm, which can feed 120 parallel
membrane modules on a once-through basis.  Since a greater
number of modules is used, recirculation is necessary.
Perhaps of greater importance is the difficulty of op-
erating two, once-through stages in series.  Specifically,
an imbalance would occur since the second pump would up-
set the flow and pressures in the first stage.

Stages 3 and 4 would require circulation in any event
since the net feed flow to each of the stage will be
relatively low.
 (vi) Summary of Ultrafiltration Section  Costs
                                                Cost
Membranes (510 x $750)                        $382,500
   (includes $183,600 for membrane
   cartridges)
Pumps                                           18,600
Tanks and piping                                19,000
Electrical connections for a total of
     350 hp to 6 pumps (transformer ex-          9,000
     eluded)	

TOTAL                                         $429,100
(Engineering design, equipment procure-
 ment, supervision of construction, and
 startup not included)

(vii) Other Equipment.  Other equipment  costs are  given
belovr.

                                                Cost

     Clean feed surge tank - will have about
     one hour's surge capacity (80,000 gals.),
     installed.                               $ 30,000

     Heat exchanger  (counter current) - to
     cool 1,600 gpm from 130°F to 100°F using
     1,600 gpm of water which is heated from
     70°F to 100°F.  Installed cost.
                          154

-------
                                           Cost

    24 million BTU/hr
    150 BTU/(hr)(ft2)(°F)
    30°F AT ,,
    5,350 ft                             $ 75,000

Control panel -                          $ 10,000

Building - about 40 ft x 50 ft con-
crete pad with sumps and pump pads of
cast concrete.  A prefabricated type
of building 30 ft to cross beams.
Very limited heating as the system
releases substantial heat  (electrical
substation, if required, is not in-
cluded) .                                 $ 48,000

Evaporator - based on direct contact
between flame and solution to con-
centrate 20,000-30,000 gpd of 20%
solids to 50% solids.  Assumed to
be fueled by burning hydrogen avail-
able from Electrochemical Plant.  The
50% solids stream will be disposed of
with the primary sludge from the
waste treatment plant.  Incineration
is preferable to evaporation and may
be possible.  However, preliminary
capital costs provided by the John
Zink Company are high compared to
previous estimates.  Furthermore,
poor burning efficiency will re-
quire additional fuel consumption.
Therefore, for present purposes,
incineration is questionable.
Evaporator, installed, including
hydrogen handling system.                $125,000

Total Other Equipment -                  $288,000

Engineering charges  - including all
design and drafting, equipment procure-
ment, construction supervision, and
shift engineers for  startup and super-
vision for the  first twelve months of
operation.                               $300,000
                      155

-------
     (viii) Capital Cost Summary, Case  1.
                                                    Cost
    Filtration and neutralization section         $100,000
    Ultrafiltration section  (includes             $429,000
         membranes @$183,000)
    Other equipment and building                  $288,000
    Design, Administrative, and Supervision       $300,000
                                                $1,117,200

    Contingency, @ 10%                             111,700
    TOTAL                                       $1,228,900

    Total utilized energy is 374 hp excluding the flushing
    pump which is run only when a circulation pump is stopped.

b.  Details of Case 2 Design

    (i) Design Bases.  The design bases for Case 2 are given
    below:

         —2 x 10  gpd pine caustic extraction filtrate
           processed;
         —corrugated-spacer membrane cartridges used;
         —precoat filtration needed; and 2
         —membrane flux of 18 gal./day-ft .

    (ii) Ultrafiltration System Costs.  For this case costs
    increase due to (1)  increased membrane costs, and {2} in-
    creased feed circulation rate through the membrane
    spirals.  The costs  for the neutralization and filtra-
    tion section, other  equipment and building, and design
    and administrative costs remain unchanged.  Only the
    Ultrafiltration section costs must be changed.

    As before, 510 membrane modules are required, but
    each is more expensive.  The net feed flow to the mem-
    brane plant is 1,520 gpm.   Since each module requires
    15 gpm, the first stage can be operated only once-
    through if it contains 102 modules.  Instead of limiting
    the number of modules in Stage 1 to this number,  re-
    circulation will be  employed.   This requires an addi-
    tional circulation pump for Stage 1.
                             156

-------
    Ultrafiltration design factors and modified costs are
    given in Table 20.  The appropriate flow schematic is
    given in Figure 57.

    Capital Cost Summary, Case 2.
    Filtration and neutralization section
         (from Case 1)
    Ultrafiltration section (includes
         membranes §$275,400)
    Other equipment and building
         (from Case 1)
    Design, Administration, and Supervision
         (from Case 1)
                                                $1,217,100

    Contingency, @ 10%                             121,700
    TOTAL                                       $1,338,800

    Total utilized energy is 509 hp, excluding the flushing
    pump  which is run only when a circulation pump is stopped.

c.  Case 3 Design

    Design Bases.  The design bases for Case 3 are given
    below:

         —2 x 10  gpd of pine caustic extraction filtrate
           processed;
         —corrugated-spacer membrane cartridges used;
         —no precoat filtration required;2and
         —membrane flux of 18 gal./day-ft .

    This case is similar to Case 2, but since corrugated-
    spacer membrane cartridges are used, it has been assumed
    that precoat filtration is not required.  Instead, a
    simple screen, such as a Bauer Hydrosieve, will be
    employed in place of the precoat and scavenger filters.
    Feed neutralization and mixing is still required.
    Cost savings are achieved from (1)  elimination of  ex-
    pensive filtration equipment; and (2)  the building
    cost is reduced by 15% ($7200).  Thus,  the capital costs
    for Case 3 will be $80,000 less than Case 2.  The  total
                              157

-------
                           TABLE 20

                ULTRAFILTRATION SECTION DESIGN

                  DETAILS AND COSTS—CASE 2
 Stage
 Number
                               4
Number of
Modules
  200      120
        120
         70
Circulation
Rates  (gpm)
 3000     1800     1800     1050
Pump Differ-
ential  (psi)
   40
40
40
40
Utilized
Horse Power
  100
70
70
40
Efficiency for
Estimating h.p,
   70%      60%
         60%      61%
Approximate
Pump Size
 8x10-13  6x8-13   6x8-13   6x8-11
Pump Cost  (Driven,
Started and Installed,
but without Electrical
Connections)
$4000    $3,500   $3,500   $2,800
                             158

-------
                      TABLE  20
                     (continued)

           ULTRAFILTRATION SECTION DESIGN
             DETAILS AND COSTS—CASE  2


               Additional Equipment

Flush Pump:     1800 gpm, 40 psi, 60  HP.        $ 3,500,

                (note that Stage  1 must be
                 flushed in  two halves to
                 obtain adequate  flow)

Permeate Pump:  1600 gpm, 40 psi, 60  HP.        $ 3,200,

Feed Pump:      1520 gpm, to 90 psi,  125 HP.    $ 4,700,
                (64% efficiency)

Water Store:    20,000 gallons, installed       $10,000,

Flush and drain piping installed                $ 9,000,
      SUMMARY OF ULTRAFILTRATION SECTION COSTS


Membranes,  (510 x $930)
  (includes $275,400 for
   membrane cartridges)                         $474,300

Pumps                                            25,200

Tanks and Piping                                 19,000

Electrical  connections for  a total of
  485 HP to 7 pumps  (transformer excluded)       10,500


TOTAL                                           $529,000
                           159

-------
    capital cost is then $1,258,000.   This cost includes
    membranes at $275,400.   The total energy used is 447 hp.

d.  Case 4 Design

    Design Bases.  The design bases for Case 4 are the same
    as those presented for Case 1 with the exception that
    a depth filter system is used in place of the precoat
    filter system.  No changes in capital requirements
    have been included.  The capital estimates from this
    case are used to display the operating cost changes.
    The capital cost for Case 4 is $1,228,900.  Total
    utilized energy is 374 hp.

e.  Case 5 Design

    Design Bases.  Design bases for Case 5 are:

         —1 x 10  gpd pine caustic extraction filtrate
           processed;
         —mesh-spacer membrane cartridges used;
         —depth filtration system used; and
         —membrane flux of 18 gal./day-ft2.

    This case is similar to Case 1, but the system is designed
    to handle 1 x 106 gpd  rather than Case 1 design of
    2 x 106 gpd.  In addition, the system design uses a depth
    filter in place of the precoat filter system used in
    Case 1.  The capital cost estimate for Case 5 has been
    developed from the Case 1 capital estimate by scaling
    as indicated below:
                     CAPITAL COST SUMMARY

                          Case 1        Scale        Case 5
                       Capital Cost    Factor    Capital Cost

    Filtration and
    neutralization                       1
    section            $ 100,000                   $ 70,800
    Ultrafiltration                      ,
    section              429,000           -         225,800
    Other Equipment      240,000         1^          120,000
                               160

-------
            CAPITAL COST SUMMARY (continued)

                      Case 1        Scale       Case 5
                   Capital Cost    Factor    Capital Cost

Building              48,000        -^         30,000
                                    J.. b

Design, Administra-
tion and Supervision 300,000                   250,000
Total                                         $700,600

Contingency, @ 10%                              70,000
TOTAL                                         $770,600

Total utilized energy is 187 hp, exluding the flushing
pump which is run only when a circulation pump is stopped.

Summary of Capital Cost Estimates

A summary of the capital cost projections for the five
case studies for battery-limit plants to treat pine
caustic extraction filtrate is presented in Table 21.
Because a plant for this service would be the first of
its kind, conservative estimates have been used in pre-
paration of the projections.  This is especially so in
the estimates for design, administration and supervision.
It is felt that the startup and first year supervision
should be amply provided for.  As will be noted, in
Case 5 this accounts for about one-third of the capital
costs.

These capital values are used in the following sections
in the development of the projected plant operating
costs.

Operating Cost Summary

Projected operating costs for Cases 1 through 5 are
presented in Tables 22,  23, 24, 25, and 26 and are
summarized in Table 27.

Bases for Estimates.  The estimates have been developed
on the basis of a 365-day operating year.  The costs
are incremental operating costs for treating pine caustic
extraction filtrate in an existing pulp mill complex.
                           161

-------
                                 TABLE 21

               SUMMARY OF INSTALLED CAPITAL COST ESTIMATES

             PLANT TO TREAT PINE CAUSTIC EXTRACTION FILTRATE


Case No.              11211


Filtration &
Neutralization   $  100,000   $  100,000   $   35,000   $  100,000   $   70,800
Section


                    429,000      529,000      529,000      429,000      225,800
Other Equipment     240,000      240,000      240,000      240,000      120,000



Building             48,000       48,000       40,000       48,000       30,000


Design,
Administration      300,000      300,000      300,000      300,000      250,000
& Supervision


Subtotal          1,117,200    1,217,100    1,114,000    1,117,200      700,600



Contingency         111,700      121,700      114,000      111,700       70,000
Total            $1,228,900   $1,338,800   $1,258,800   $1,228,900   $770,600
Membrane Costs
Included in      $  183,600   $ 275,400    $  275,400   $  183,600   $ 91,800
Capital
Utilized HP             374         509           447          374        187
                                    162

-------
OJ
TABLE 22
Operating Costs for Treatment of Pine Caustic Extraction Filtrate, Case :
Quantity Unit Cost $/Day $/Year 
-------
                                TABLE 23
Operating Costs for Treatment of Pine Caustic Extraction  Filtrate; Case  2
                       Quantity         Unit Cost   $/Day     $/Year  C/M-gal   Total
Statistical
                                                                                     of

Material
Acid. . . 	 	 8/000 Ib/day IC/lb
Filter Aid 	 1/000 Ib/day 3£/lb
Total Material 	

Conversion Exoense
Labor (including benefits) 5 Man-yrs $ll,800/yr
Repair and Maintenance
Material. . . . .$ 763.400 i na ATV

Electric Power .... 509 HP 0. 746
-------
Ui
                               TABLE 24
Operating Costs for Treatment of Pine Caustic Extraction Filtrate;  Case  3
                      Quantity        Unit Cost   $/Day     $/Year  C/M-gal   Total
         Statistical
                                                                                              of
Material
Acid 	 8,(

Filter Aid
Total Material . . ,
Conversion Expense

300



Labor (including benefits)
Repair and Maintenance
Material $
Labor
Electric Power
Insurance and Taxes 1/2

682
0.5
447

Ib/day



4 Man-yrs

,400
Man-yrs
HP
x Maintenance

l$/lb



$ll/800/yr

1.5%/yr
$20,000/yr
0.746
-------
                                 TABLE  25
Operating Costs for Treatment of Pin<
Quantity
Material
Acid 	 8,0
Filter Aid
Conversion Expense
Labor (including benefits)
Repair and Maintenance


Insurance and Taxes I/2
Total, excluding Deprec
Depreciation
Other Facilities . . $1,
Total Incremental Cost
00. Ib/day
3 Caustic Ext:
Unit Cost
l«/lb
4 Man-yrs $ll,800/yr
744,300 1.5%/yr
0.5 Man-yrs $20,000/yr
374 HP 0.746$/HP-hr
x Maintenance Material
i at ion
183,600 3-year life
044,300 15-year life



ractio:
$/Da
$ 80.
$ 80.
$129.
30.
27.
66.
15.
$269.
$167.
190.
.$627.
$707.
n ri
£
00
00
32
58
40
96
29
55
67
74
9f>
96
.itrate: c<
$/Year
$ 29
$ 47
11
10
24
5
$ 98
61
69
$229
$258
,200
,200
,165
,000
,400
,580
,386
,200
,620
r2p5
,405
ase 4
<=/M-gal
4
6
1
1
3
0
13
8
3;
35
.0
.47
.52
.37
.35
.76
.47
.38
,39
.39
% of
Total

11.
18.
4.
3.
9.
2.
38.
23.
27.
	 	 8?,
100.

3
3
3
9
5
1
1
7
0
7
0
Statistical
Effluent Treated (x 106 gallons)                           2        730

-------
                                  TABLE 26
   Operating Costs for Treatment of Pine Caustic Extraction Filtrate:  Case 5       %   -^
                         Quantity         Unit Cost    $/Day      $Aear   C/M-gal    Total
Material
Acid.  .  .  .  ....... 4,000  Ib/day        l£/lb   $ 40.00
Filter Aid                                   .
     Total Material	$ 40.00  $ 14,600    4.00     8.7
Conversion Expense
Labor  (including benefits)    4 Man-yrs  $ll,800/yr   $129.32  $ 47,200   12.93    28.2
Repair and Maintenance
     Material	$  364,800         1.5%/yr     15.00    54,720    1.5      3.3
     Labor  	  .   0.5 Man-yrs  $20,000/yr     27.40    10,000    2.7      6.0
Electric Power             187 HP     0.746$/HP-hr     33.48    12,220    3.35     7.3
Insurance and Taxes     1/2 x Maintenance Material      7.50    27,360    0.75	1.6
     Total, excluding Depreciation                   $212.70  $ 77,636   21.27    46.5
Depreciation
     Membranes          $   91,880     3-year life     83.84    30,600    8.38    18.3
     Other Facilities   $  664,800    15-year life    121.42    44,320   12.14    26.5
Total Conversion Expense                             $417.94  $152,548   41.79    91.3
Total Incremental Cost                               $457.94  $167,100   45.79   100-.0
Statistical
Effluent Treated  (x 106 gallons)                        1          365

-------
                                 TABLE 27

                    DAILY INCREMENTAL OPERATING COSTS

               TREATMENT OF PINE CAUSTIC EXTRACTION FILTRATE
Case No.
Pine Caustic
Extraction
Filtrate              22221
Treated
x 106 gpd


Capital
Investment       $1,228/900   $1,338,800   $1,258,800   $1,228,900   $770,600
Daily Operating Costs - Dollars/day

  Materials      $110         $110         $ 80         $ 80         $ 40

  Conversion
  Expense less    301.87       327.23       278.81       269.55       212.70
  Depreciation

  Membrane        167.67       251.50       251.50       167.67        83.84
  Depreciation

  Other
  Facilities      190.74       194.23       179.43       190.74       121.42
  Depreciation    	       	       	       	       	


  Total          $770.28      $882.96      $789.74      $707.96      $457.94
  Costs
  $/1000 gal.       38.50$      44.14*        39.48$       35.39$       45.79$
  Treated
  Costs
  $/Ton          $  0.962     $  1.103     $  0.987     $  0.085     $  0.572
  Pine Pulp


                                   168

-------
These costs are presented as representing steady-state
operating costs for an ultrafiltration plant using
cellulose acetate membranes from the second year of
operation and on.  As indicated previously, in prepara-
tion of the capital costs, since this plant would be
the first of its kind and scale, due allowance has been
made for special startup and supervisory costs for the
first year of operation.

The bases for estimating costs were as follows:

     Materials;  Acid costs and filter aid costs are
     present plant or vendor estimates.

     Labor;  Labor costs have been estimated on the
     basis of 1973 estimated base salaries plus 30%
     added to cover benefits.  It is assumed the
     plant can be operated with 4 or 5 man-years per
     year of operator time for the first year or so,
     and that this manpower level could be substantially
     reduced as operating experience is gained.  For
     these estimates, however, 4 or 5 operators are
     used.

     Electric power;  Electric power costs are estimated
     on the basis of power costs of IC/kw-hr.

     Depreciation;  The membrane life is taken as 3 years
     and the costs are depreciated over this time period
     on a straight-line basis.  The remainder of the
     plant is depreciated on a straight-line basis over
     a 15-year period.

A summary of the daily operating costs in presented in
Table 27.

A comparison of Cases 1 and 2 shows that an operating
cost increase of approximately 6C/1000 gals, is in-
curred if corrugated-spacer membrane cartridges are
used instead of the mesh-spacer cartridges.  Examination
of Cases 1 and 3 shows that there is only about a 1
-------
Case 4 considers the use of mesh-spacer cartridges but
installing backwashing filters with automatic cleaning
cycles.  For this case, filter aid would not be re-
quired and four men would operate the system.  The
capital cost of these filters has been assumed to be
equal to the capital cost of precoat filters, but
operating costs are reduced by elimination of filter aid
and a labor component.  Some pilot plant experience
has been obtained with such a filter—the Hydromation
granuar PVC backwashing filter.  The model tested fil-
tered adquately but was not used extensively in the
pilot plant since its capacity was too small.  Case 4
shows operating costs for a system with a Hydromation-
type filter and with mesh-spacer cartridges.  It is
seen that a savings of approximately 3C/1000 gals, over
Case 1 can be obtained.

Both the capital and operating costs for an ultrafiltra-
tion system are strongly dependent on the volume pro-
cessed.  It has been observed in the pilot plant program
that the concentration of color bodies and organics in
the pine caustic extraction filtrate varies substantially
from day-to-day, and even hour-to-hour.  In principle,
it should be possible to control the pine caustic
extraction filtrate flow rate such that the concentra-
tion of contaminants is at the maximum allowable level
in terms of obtaining bleached pulp of acceptable
quality.  Through this means it is thought that the flow
of pine caustic extraction filtrate can be substantially
reduced from 2 x 106 gpd, possibly to as low a volume
as 1 x 10° gpd.  By this means, the cost of waste treat-
ment per unit of pulp produced can be substantially
reduced, even though the treatment cost per unit of
effluent may increase.  In Case 5, it has been assumed
that the total flow of pine caustic extraction filtrate
can be reduced to 1 x 106 gpd.  Mesh spacers are used,
and a backwashing filter is employed.  Capital costs
for this plant have been scaled from the 2 x 10^ gpd
plant using the scaling factors presented on page 148.
Case 5 is to be compared to Case 4.  Although the cost
per unit effluent is higher in Case 5, the cost per
unit pulp produced is substantially lower.

Some Economic Evaluations

The capital cost estimatesgindicate that  an ultrafiltra-
tion plant to treat 2 x 10  gpd of pine caustic extraction
filtrate would have an installed cost of  $1.2-$1.4 million
                          170

-------
based on present equipment and labor costs.  A similar
plant to treat 1 x 10° gpd would cost almost 750 to
800 thousand dollars.

The operating cost estimates indicate that the most
significant cost items are membrane depreciation, other
facilities depreciation, operating labor and materials —
which account for 70% to 90% of the daily costs in the
cases presented.

The most sensitive single cost factor displayed is the
total flow of caustic extraction filtrate to be treated.
If the flow of this material can be limited by judi-
cious bleachery process flow control to 1 x 10^ gpd,
the capital cost of the membranes would be reduced to
one-half, the total plant cost would be almost 60% and
the daily operating costs would decrease from a level
of $700 - $800 per day to about $450 per day.

Membrane flux is a second important parameter in process
costs.  The cases examined above have not considered
variation in membrane flux, but have assumed a rate of
18 gal./day-ft2 .  For a given capacity, plant cost will
vary almost inversely with membrane flux.  Operating
costs will be greatly affected due to both membrane
replacement cost and capital depreciation.  The value
chosen in the design calculations  (18 gal./day-ft2) is
higher than that observed in most of the pilot plant
tests.  However, since some of the membrane cartridges
exhibited this ultrafiltration rate, it is assumed to
be an achievable value.  In addition, future advances
in membrane technology will undoubtedly provide
higher-flux membranes.

Membrane life is very important since membrane replace-
ment is a substantial operating cost factor.  The  3-yr
life which has been chosen for design calculations is
longer  than  any  life demonstrated to  date  with pine
caustic extraction filtrate.  This life, however,  is
not unreasonable for brackish water desalination
applications, and is considered realistic.  It may be
noted that a 2-yr membrane life  (mesh-spacer cartridges)
would increase operating costs by approximately
4C/1000 gals.  Again, with new developments in mem-
brane technology, longer life should be obtainable.
                          171

-------
Other anticipated membrane improvements would permit
operation at alkaline pH's and at high temperatures.
Three such membranes are in an advanced stage of
development.  These are the NS-1 membrane, developed
by the North Star Research and Development Institute,
Minneapolis, Minnesota  (29) ,the polybenzimidazole
membrane, developed by the Celanese Research Company,
Summit, New Jersey (3P) ,and the dynamic membranes
developed by the Oak Ridge Natural Laboratory  (16).Not
only can acid costs and cooling costs be eliminated,
but membrane flux should also increase.  This is due to
two factors.  First, membrane flux increases with
temperature;  second, in the treatment of pulp mill
effluents membrane flux generally decreases when the
feed pH is changed.  This is primarily a fouling
phenomenon, with non-neutralized feeds exhibiting a
substantially reduced fouling rate.  Thus, new mem-
branes will not only reduce pre-treatment costs, but
can also reduce capital and operating cost since a
higher membrane flux can be realized.

When using cellulose acetate membranes, for which pre-
treatment is required, some additional cost savings may
be realized.  First, the acid cost for neutralization
may be reduced or eliminated by neutralizing with waste
acid available within the mill.  A second materials
cost which may be eliminated is that for filter aid.
If corrugated-spacer cartridges can be used without
precoat filtration, or if a backwashable depth filter
is adequate, filter aid would not be required.

Another major operating cost is labor.  It has been
assumed that four or five man-years/year would be re-
quired to run the combined pre-treatment and ultra-
filtration systems.  If the system is simplified by
modification or elimination of the pre-treatment system,
this operating labor load could be reduced substan-
tially.  Furthermore, as operating experience is gained,
the labor requirement should decrease.

Thus, several potential savings in both capital and
operating costs can be obtained through new develop-
ments in membrane technology, as well as by obtaining
operating experience for a full-scale plant.
                          172

-------
B.  DECKER EFFLUENTS

1.  Flow Schematic

    Figure  4 (page 21)  has presented a generalized flow
    schematic showing integration of an ultrafiltration
    system in the decker and black liquor systems.   A
    generalized flow schematic, more specific to the ultra-
    filtration process,  is shown in Figure 58.  The
    effluent from the decker undergoes pre-treatment
    (neutralization and filtration), is cooled by exchange
    with incoming fresh water, and flows into the ultra-
    filtration system.  The permeate from the ultrafiltra-
    tion system is recycled to the final stage of pulp
    washing, or is used for other pulp mill water require-
    ments.  The concentrate from the ultrafiltration section
    is used to sluice the filter cake, and this suspension
    is processed in the black liquor recovery system.

    More detailed flow schematics for the decker effluent
    treatment system are given in Figures 56 and 57
    (pages  129 and  130).These drawings, previously presented
    for the pine caustic extraction filtrate system, are
    applicable to the treatment system for decker effluents.
    Only the means of disposal of the permeate and concen-
    trate fractions from the ultrafiltration system are
    different.

    In preparation of the five cases of projected economics
    presented, the process parameters have been varied as
    shown in Table  28.

2.  Capital Cost Projections

    The five capital cost projections for installed battery-
    limit plants are presented in Table  29.  The vendor
    equipment budgetary estimates used in preparation of the
    pine caustic extraction filtrate capital projections
    have been used as the basis for the estimates and are
    not repeated here.

    Cases 6 and 9 differ from the capital estimate for
    Case 1  (pine caustic extraction filtrate) only by the
    cost of the evaporator system used in Case 1.  Details
    for the costs are given under the Case 1 discussion.
                              173

-------
                Other Pulp
                Mill Water .
               Requirements
                                                                      Fresh Water
                                     1
                     Pulp Washers
Black
Liquor
Plant
                            Decker
                   Dilute
                   Black
                   Liquor
                                 Present
                                 Effluent
T
               Filter
Cake
                                                 Fresh Water
                                            Concentrate
                               UF

                            Treatment
                                                                               Permeate
           FIGUP^ 58:  SIMPLIFIED FLOW SCHEMATIC:  TREATMENT OF DECKER EFFLUENTS

-------
                                                           TABLE 28

                                              CASES FOR  CAPITAL COST  ESTIMATES

                                                       DECKER EFFLUEHTS


           Decker            Membrane Modules                      Circulation                       Prefiltration
          Effluent         Low Flow    High Flow     Membrane          Rate        Membrane                                  Operating
          Flow Rate                    (corrugated      Flux        gpm/roembrane      Cost     Precoat Backwashable Screen    Manpower
 Case     (x  106 gpd)      (mesh spacer)    spacer)    gal./day-ft^    cartridge      S/sq ft    Filter  Depth Filter Filter   No. of Men


  6           2               x                            18             8          $1.50        x                              5



  72                            X               18            15          $2.25                             x         «
M
«J
Ui

  8           1               x                            18             8          $1.50        x                              5



  9           2               x                            18             8          $1.50                  x                    4



  10           1               X                            18             8          $1.50                  x                    4

-------
                                TABLE 29

     INSTALLED CAPITAL ESTIMATES -  PLANT TO TREAT DECKER EFFLUENTS
Case No.            6a           7_b           Bc           £a          10_


Prefiltration &
Neutralization   $100,000     $100,000     $100,000     $100,000
Section                        -65,000
                               =35,000      =70,800                  $ 70,800

Ultrafiltration   429,100      529,000      429,100      429,100      225,800
Membrane                                      1.9

                                           =225,800
Other Equipment   115,000      115,000       115,000/2    115,000       53,000
                                             =5., 000

Building           48,000        40,800        48,000       48,000       30,000
                                               1.6

Design,                                      =30<000
Administration    300,000      300,000       250,000      300,000      250,000
& Supervision '    -      -       -      -      -
Subtotal          992,200    1,019,800      629,600      992,200      629,600



Contingency        98,200      101,000       60,000       98,200       60,000




Total          $1,090,400   $1,120,800     $689,600   $1,090,400     $689,600




Membrane Costs
Included Tn183,600      275,400      183,600      183,600       91,800
Capital                                     =91,800



Utilized HP           374          447          187          374          187


NOTES;

a - costs  taken from Case 1 and adjusted
b - costs  taken from Case 3 and adjusted
c - costs  obtained by adjustment and scaling from Case 6



                                176

-------
    Case 7 differs from Case 3 (pine caustic extraction
    filtrate) only by the cost of the evaporator system.
    Details are given under Case 3.

    Case 8 and Case 10 are derived from Case 6 by the
    scaling procedures used under the pine caustic extrac-
    tion filtrate Case 5.

3.   Estimated Operating Costs

    Projected operating costs for each of the 5 cases are
    presented in Tables  30, 31, 32, 33 and 34 .  The
    estimates are based on a 365-day operating year.

    The projected operating costs as presented are incre-
    mental operating costs for a plant for treating the
    decker effluents and their reuse.  These costs are
    presented for steady state operations from the second
    year of operation and on.  In preparation of the
    capital costs, since this plant would be the first of
    its kind and scale, due allowance has been made for
    special startup and supervisory costs for the first
    year operations.

    The same estimating bases have been used for the decker
    effluent cases as were used in the pine caustic extrac-
    tion filtrate cases.  The life of the membranes is
    taken as 3 years and the life of the other facilities
    as 15 years for purposes of depreciation estimates.
    Labor costs as  presented include a 30% adder for
    benefits.

    The daily incremental operating costs for the five cases
    are summarized in Table 35.  These cases are evalua-
    ted in greater detail below,  which makes allowance for
    process credits.

4.   Potential Process Credits from Decker Effluent Recycle

    The design of the ultrafiltration plant to treat decker
    effluents is developed on the basis of splitting the
    stream into a concentrate containing organic materials,
    including color bodies, and a permeate stream of low
    color containing the bulk of the dissolved materials,
    primarily sodium sulfate.  The low-volume concentrate
    would contain 10-20% organic material, and would be
                              177

-------
-J
00
                                            TABLE 30
                   Operating Costs for Treatment of Decker Effluent;   Case 6                 % of
                                   Quantity         Unit Cost   $/Day     $/Year  C/M-gal   Total
          Material
          Acid	8,000 Ib/day        l«/lb   $ 80.00
          Filter Aid	1,000 Ib/day        3C/lb     30.00   	
               Total Material	$110.00  $ 40,150     5.5       14.9
          Conversion Expense
          Labor (including benefits)    5 Man-yrs  $ll,800/yr   $161.64  $ 59,000     8.08     21.9
          Repair and Maintenance
Material 	 $ 606,800 1.5%/vr

Electric Power 375 HP 0. 746£/HP-hr
Insurance and Taxes 1/2 Maintenance Material
Depreciation
Membranes $ 183,600 3-year life
Other Facilities $ 906,800 15-year life
Total Conversion Expense
Total Incremental Cost
24.94
27.40
66.96
12.26
.$293.20

$167.67
165.63
$626.50
$736.50
9,102
10,000
24,400
4,476
$107,017
61,200
60.453
$228,673
$268,823
1.25
1.37
3.35
.62
14.67
8.38
8.28
31.33
36.83
3.3
3.9
9.1
1.7
39.8
22.8
22.5
85.1
100.0
          Statistical
          Effluent Treated (x 106 gallons)                            2       730.0

-------
                                            TABLE  31
                   Operating Costs for Treatment of Decker Effluent:  Case 7                  % of
                                   Quantity         Unit  Cost   $/Day     $/Year  $/M-gal   Total
          Material
          Acid ........... 8,000  Ib/day       l$/lb    $ 80.00
          Filter Aid
               Total Material	$80.00  $ 29,200     4.0     10.6
          Conversion Expense
          Labor  (including benefits)    4 Man-yrs   $ll,800/yr   $120.32  $ 47,200     6.47    17.1
          Repair  and Maintenance
               Material	  $   545,400          1.5%/yr     22.41     8,810     1.12     3.0
^>              Labor	0.5 Man-yrs   $20,000/yr     27.40    10,000     1.37     3.6
vo
          Electric Power             447 HP     0.746C/HP-hr     80.03    29,210     4.00    10.6
          Insurance and  Taxes      1/2  x Maintenance  Material    11.20	4,090	0.56	1.5
               Total, excluding Depreciation                   $270.76    98,827     1.35    35.8
          Depreciation
               Membranes          $   275,000      3-year  life   $251.50    91,800    12.57    33.2
               Other Facilities    $   845,400    ISryear  life    154.41    56,360     7.72    20.4
          Total  Conversion Expense                            $676.67  $246,985   $33.81    89.4
          Total  Incremental Cost                               $756.67  $276,185   $37.81   100.0
          Statistical
          Effluent  Treated  (x 106  gallons)                        2          730

-------
00
o
TABLE 32
Operating Costs for Treatment of Decker Effluent: Case 8
Quantity Unit Cost $/Day $/Year <=/M-gal
Material
Acid 	
Filter Aid
Total Material .
Conversion Expense
. .4.000 Ib/dav 16/lH
500 Ib/day 3£/lb

Labor (including benefits) 5 Man-yrs $ll,800/yr
Repair and Maintenance
Material 	 s 2Q7."*nn i <;» An-
Labor . . . . . .
Electric Power
Insurance and Taxes
Total, excluding
Depreciation
Membranes ....
0* 5 Man— vrs $20 .000 A/r
187 HP 0.746C/HP-hr
1/2 x Maintenance Material
Depreciation
. . s qi . ann ^-v^ar i i f &
Other Facilities . .$ 597,300 15-year life
Total Conversion Expense 	
Total Incremental Cost


$ 40.00
15
$ 55 00

$161.64
12.22
27.40
33.48
6.11
$240.85
$ 83.84
109.10
.$433.38
.$488 38


$ 20,070
$ 59,000
4,450
10,000
12,220
2,230
$ 87,910
30,600
39,820
$158,333
$178,259

5.5
16.15
1.22
2.74
3.35
0.61
24.08
8.38
10.91
43.47
48.87
% of
Total

11.3
33.1
2.5
5.6
6.9
1.2
49.3
17.1
22.3
88.7
100.0
          Statistical
          Effluent Treated (x 10 6 gallons)                         1          365

-------
                                             TABLE 33
oo
jLiiwj.diidiv.aa. w^co. a. i, j.iivj \* wo 1.0 J.WJL j. j.cai,uicii u \JJ- L/cwrt
Quantity Unit Cost
Material
Acid 	 8,000 Ib/day l£/day
Filter Aid
Total Material 	

Conversion Expense
Labor (including benefits) 4 Man-yrs $ll,800/yr
Repair and Maintenance
Material 	 & 606/800 1 5%/yr
Labor • ....... 0.5 Man— yrs $20,000/yr
Electric Power 	 374 HP 0. 746C/HP-hr
Insurance and Taxes 1/2 x Maintenance Material
Total, excluding Depreciation 	

Depreciation
Meirbranes 	 $ 183,600 3-year life
Other Facilities . $ 906,800 15-year life,




C J. OJi. JL .LUC
$/Day

$ 80.00

.$ 80.00


$129.32
9494
27 40
66 96
12.26
.$260. 88

$167.67
165.63
.$594. 18

.$674. 18

a<- , v-aac y
$/Year



$ 29 .200


$ 47,200
9 102
10 000
24 400
4,476
$ 95,178

$ 61,200
60,453
$216 ,831

$246 ,031

C/M-gal



4 0


6.47
1 25
1 37
3 35
0.62
13.06

8 38
8.28
29. 72

33. 72

% of
Total



11 9


19.2
3 7
4 1
9 9
1.8
38.7

24 9
24.6
88. 1

100.0

           Statistical
           Effluent Treated  (x 106 gallons)                        2          730

-------
                                            TABLE 34
             Incremental Operating Costs for Treatment of Decker Effluent;  Case 10          % of
                                   Quantity         Unit Cost   $/Day     $/Year  C/M-gal   Total
          Material
          Acid ........... 4,000 Ib/day       l$/lb   $ 40.00
          Filter Aid
               Total Material	$ 40.00   $ 14,600    4.0       9.1
          Conversion Expense
          Labor (including benefits)   4 Man-yrs  $ll,800/yr  $129.32   $ 47,200   12.93     29.3
          Repair and Maintenance
£              Material	$  297,,300         1.5%/yr    12.22      4,460    1.22      2.8
10              Labor.	    0.5 Man-yrs  $20,000/yr    27.40     10,000    2.74      6.2
          Electric Power	    187 HP     0.746«/HP-hr    33.48     12,220    3.35      7.6
          Insurance and Taxes  .  . 1/2 x Maintenance Material     6.11	2 ,230	0.61	1. 4
               Total, excluding Depreciation	$208.53   $ 76,110   20.85     47.2
          Depreciation                                             ~~~~~
               Membranes	$   91,800     3-year life  $ 83.84   $ 30,600    8.38     19.0
               Other Facilities  . $  597,300    15-year life  $109.10     39,820   10.91     24.7
          Total Conversion Expense                            $401.47   $146,530   40.14     90.9
          Total Incremental Cost                              $441.45   $161,130   44.14    100.0
          Statistical
          Effluent Treated (x 106 gallons)                       1           365

-------
                               TABLE  35


                   DAILY INCREMENTAL OPERATING COSTS

                TREATMENT AND REUSE OF DECKER EFFLUENTS
Case No.
Decker Effluent
Feed Rate
FTo"6 gpd
Capital
Investment
2 2
$1,090,400 $1,120,800
121
$689,600 $1,090,400 $689,600
Daily Incremental Operating Costs
Materials
Conversion
Expense less
$110 $ 80
293.20 270.76
$ 55 $ 80 $ 40
240.85 260.88 208.53
  Depreciation

  Membrane
  Depreciation
 167.67
 Other
 Facilities      165.63
 Depreciation    	
 251.50
              154.40
83.84
              109.10
167.67
            165.63
83.84
             109.10
 Total
$736.50
$756.67      $488.38      $674.18      $441.45
 Costs
 C/1000 gal.
 Treated
  36.83
  37.81
48.87
 33.72
                                                      44.14
                                   183

-------
    sent to the dilute black liquor system.  The permeate
    would be reused, either in the pulp washers or in other
    fresh water pulping uses and hence would eventually be
    returned to the dilute black liquor stream for chemical
    recovery.

    Operation with recycle of permeate and concentrate
    should result in process credits which could offset the
    costs associated with installation and operation of the
    ultrafiltration plant.  Potential credits could be in
    the following:

        —reduction of the fresh water requirements of
          the pulp mill (^3-5%);
        —reduction of the total mill effluent and waste
          processing operations (^3-5%);
        —retention of the salts presently being dis-
          charged in the effluent  (^16-24 tons/day);
        —increasing the organic content of the dilute
          black liquor without increase in liquor
          volume  (^8-10 tons/day);
        —reduction of the mill effluent color by about
          20%; and
        —reduction of organic solids treated in the
          secondary treatment plant by about 20%.

    The potential credits to be derived are a function of
    the specific pulp mill in which such an ultrafiltration
    plant would be installed.  The potential credits dis-
    cussed below are those that might be applicable at the
    pulp mill  in which these pilot studies were conducted.
    At other pulp mills the washing procedures/ water costs,
    waste disposal costs and nature of the total mill
    effluent may well differ.  However, it is felt that the
    conservative figures presented below are typical "ball-
    park" credits which would be available at pulp mills
    where the decker effluents are presently sewered.

a.  Water Credits

    At present, water at the Canton Mill costs c.bout
    $110/106 gal. for fresh water feed and treatment of the
    water in the waste treatment operation.  The potential
    water credit, then, is a function of the actual amount
    of water used.  In this study, two cases have been
                             184

-------
    examined.  At 2 x 10* gpd recycle flow, the credit
    would be 2 x 110 = $220 per day.  At 1 x 106 gpd
    recycle flow the credit would be $110 per day.   (It
    should be noted those water costs are low compared to
    many mills because of geographical location and well
    planned waste disposal facilities.)

    In the mill the water recycle would reduce the fresh
    water requirements and also the effluent handling
    requirements by 3-5%.  No direct credit is taken for
    this physical reduction of volume flow in this
    presentation.

b.  Retention of Salts

    Recycling both the concentrate and the permeate frac-
    tions of the decker effluents will retain all the con-
    tained materials presently going to the waste treatment
    plant.  Depending on the operating conditions, the
    decker effluents contain 1200-3300 ppm of sodium sulfate
    For the calculation below, a value of 2500 ppm at
    2 x 106 gpd flow is taken.  A value of !$/# is assumed
    for the sodium sulfate value.
        Potential credit for retained salts =


        (2500 x 10~6) (2 x 106) (8.33) (0.01) =  $416


    Recycle of these materials, of course, will reduce the
    dissolved salt content of the total effluent.  No credit
    is taken for this result.

    Organic Content

    About 98% of the organic content of the decker effluent
    is in the concentrate stream.  This stream will contain
    >10% organic material.  The heating value of this
    material is assumed to be 8000 BTU/f.  Because no total
    water increase in the weak black liquor stream is antici-
    pated as a result of the recycle and reuse of the water,
    the total heat value of the retained organics is taken
    as a potential credit.
                             185

-------
    The organic content is taken as 1000 ppm at a
    2 x 10° gpd flow.  The heat input is assumed to be
    worth 50C/106 BTU.  Then,

        Potential heat value credit =


        (1000 x 10"6)(2 x 106) (8.33) (8000) (0.50 x 10"6) =

        $66.6

    The decker effluents exit temperatures are 120-125°F.
    Returning this stream to the process has two effects.
    The thermal balance of the pulping process is improved
    and the thermal load on the mill effluent is decreased.
    No credit is taken for either effect in this calculation,

d.  Total Potential Credits
    The potential credits from installation of the process
    are totaled below.  In the calculations it has been
    assumed that there is a basic load of material to be
    removed in the washing operations and that the effect
    of varying the washing volumes will be to change the
    concentration of these materials, but that the total
    amount removed will remain the same.

    For 2 x 10^ gpd flow the total potential process credits
    are $220 + 416 + 67 = $703 per day.

    For 1 x 10  gpd flow the total potential process credits
    are $110 + 416 + 67 = $593/day.

5.  An Evaluation of Process Economics

    The capital costs of the ultrafiltration plants for
    treating the decker effluents are those for treating the
    pine caustic extraction filtrate, modified by elimina-
    tion of the concentrate disposal systems required for
    the latter.  Consequently the capital costs and the
    daily operating costs displayed in Tables 29 and 35 are
    lower than for similar cases in the pine caustic extrac-
    tion filtrate treatment.

    Again the most sensitive parameter in the study is the
    flow of the decker effluent stream.  For example,
                              186

-------
Case 9 (2 x 10  gpd) has a capital cost of $1,090,400
and a daily operating cost of $674.18; Case 10,
which presumably represents the same level of treat-
ment but on a more concentrated stream of 1 x 106 gpd,
has a projected capital cost of $689,600 and a daily
operating cost of $441.45.

As in the previous study the other parameters which
are highly sensitive are membrane flux and life, capi-
tal depreciation, operating manpower, and materials.

The decker effluent cases differ from the pine caustic
extraction filtrate cases in that for the former the
potential credits defray the operating costs of the
treatment plant.  On the assumption that the projected
capital and operating costs and potential credits are
realistic, a decker effluent treatment plant, depending
on size and process configuration, would have either
a small net operating cost  (comparison of $736.50/day,
operating cost, Case 6, and $703/day potential credit),
or a small net credit  (comparison of $441.45/day, opera-
ting cost, Case 10, and $593/day potential credit).
Consistent with the precision of the cost estimates
and knowledge of the plant flow volumes, it is likely
that an ultrafiltration plant to treat decker effluents
could be a low-cost or net-credit operation.  The
process could produce color and BOD reduction and plant
flow reduction values which have not been treated as
credits in this evaluation.

It must be emphasized that the attractive steady state
operating economics presented here are projections from
the pilot operation.  As in the previous set of projec-
tions for the pine caustic extraction filtrate these
economics require additional experimental engineering
verification, especially in the area of:

    —quantifying the minimum controllable flow of
      decker effluents which can be used while pro-
      ducing a pulp of acceptable quality for subse-
      quent mill operations;
    —demonstrating that reliable membranes can be ob-
      tained commercially;
    —evaluating alternative plant designs, including
      additional experimental work, to assure that the
      plant is capable of continuous 365-day operation
      with minimum capital and operating costs / especial-
      ly labor.
                          187

-------
The technical feasibility of treating the decker
effluents has been demonstrated using the feed treat-
ment and ultrafiltration processes described.  It
would be premature, at present/ however, to proceed
with a full-scale plant design until the information
indicated above is obtained.
                          188

-------
                          SECTION VII

                          WATER REUSE
    One potential value to be obtained from the use of an
    ultrafiltration process to treat pulp mill process
    effluents is to provide a permeate which can be re-
    used within the mill.  This would reduce the fresh
    water requirements, as well as the total plant efflu-
    ent volume.

    During the course of the program, as permeate samples
    were produced and analyzed, evaluations of water reuse
    potential were performed.

A.  PINE CAUSTIC EXTRACTION FILTRATE PERMEATE

    The pilot plant produced permeates which had 90-96% of
    the color removed and which represented 99-99.5% of
    the flow stream fed to the system.  The permeate con-
    tained most of the non-organic dissolved materials (salts)
    and a residual color of usually greater than 1000 ppm—
    even with the high color removal effectiveness  (see
    Appendix for representative values).

    Because of the stream color and high dissolved material
    content (usually ^ 6000 ppm) it is felt that this per-
    meate would have limited usefulness in the pulping and
    bleaching process areas.  At best, it is felt, this
    permeate could be used to augment the volumes of other
    "dirty" water presently used in operations such as wood
    washing and cleaning of fly ash collectors.  No high
    value reuse capability has been developed for this per-
    meate .

B.  DECKER EFFLUENT PERMEATES

    The pilot plant produced permeates from both the pine
    and hardwood decker effluents which had about 98-99.5%
    of the color removed, and which represented about 99%
    of the stream feed to the system.  As discussed in
    other parts of this report, the concentrate from the
    system would be returned to the weak black liquor stream.
    The permeate with its low color  (^ 100 ppm color) and
    dissolved salts, primarily sodium and sulfate, is a
    stream of high potential value for direct multiple re-
    use in the washers of the pulping system, or as make up
    water in the pulping operations.  As such, the reuse of
                              189

-------
the decker effluents at a mill such as the Canton mill
would:

     a) reduce the treated water demand by about 4%;
     b) reduce the plant effluent by about 4%;
     c) recycle 15-25 tons/day of chemicals normally
        lost in the decker effluents; and
     d) reduce the organic loading of the biological
        waste treatment system by about 20%.
                            190

-------
                       SECTION VIII


                     ACKNOWLEDGEMENTS
The support of the project by the Environmental
Protection Agency and the help provided by Mr. George
Webster, Mr. Edmond P. Lomasney, the Federal Grant
Project Officer, Dr. Karl Suva and Ralph Scott is
acknowledged with sincere thanks.

The project was directed by Dr. Henry A. Fremont of
Champion International.  The experimental program was
supervised by Mr. Dan C. Tate of Champion International.

The membrane equipment was provided by Abcor, Inc.
Drs.  Robert L. Goldsmith, James R. Ryan, and David J.
Goldstein, and Messrs. Sohrab Hossain and A.C.F.
Ammerlaan, all of Abcor, Inc., contributed to the
technical support of the project and report preparation.

The progress of the project would not have been possible
without the cooperation and counsel of Champion Inter-
national personnel, including John 0. Parrott, Plant
manager, Champion International, Canton Mill, and his
very competent staff; Richard Wigger; Austin Moore;
Robert Townsend; C. E. Fredericks; and, Dr. A. E.
Vassiliades.
                           191

-------
                      SECTION IX

                      REFERENCES
1.  Interstate Paper Corporation for the Environmental
    Protection Agency, Program #12040 ENC, Grant #WPRD
    183-01-68, "Color Removal from Kraft Pulping Effluent
    by Lime Addition," December 1, 1971.

2.  Spruill, Edgar L., "Paper Mill Waste:  Treatment for
    Color Removal," IW, March/April 1971, pp. 15, 21-23.

3.  Gould, Matthew, "Lime-Based Process Helps Decolor
    Kraft Wastewater," Chem. Eng. , January 25, 1971,
    pp. 55-57.

4.  Tejera, N.E.  and Davis, M.W. , Jr., "Removal of Color
    and Organic Matter from Kraft Mill Caustic Extraction
    Waste by Coagulation," TAPPI, 53, No. 10, October
    1970, pp. 1931-1934.

5.  National Council for Air and Stream Improvement, Inc.,
    "The Mechanisms of Color Removal in the Treatment of
    Pulping and Bleaching Effluents with Lime.  I.  Treat-
    ment cf Caustic Extraction Stage Bleaching Effluent,"
    Technical Bulletin No. 239, July 1970.

6.  National Council for Air and Stream Improvement, Inc.,
    "The Mechanisms of Color Removal in the Treatment of
    Pulping and Bleaching Effluents with Lime.  II.  Treat-
    ment of Chlorination Stage Bleaching Effluents,"
    Technical Bulletin No. 242, December 1970.

7.  Davis, C.L.,  Jr., "Tertiary Treatment of Kraft Mill
    Effluent Including Chemical Coagulation for Color
    Removal," TAPPI, 52, No. 11, November 1969, pp. 2132-
    2134.

8.  Middlebrooks, E. J., Phillips, W.E., Jr. and Coogan,
    F.J., "Chemical Coagulation of Kraft Mill Wastewater,"
    IW Water and Sewage Works Supplement, March 1969,
    pp. 7-9.

9.  Smith, S.E. and Christman, R.F., "Coagulation of
    Pulping Wastes for the Removal of Color," Journal
    WPCF, 41, No. 2, Part 1, February 1969, pp. 222-231.
                             193

-------
                      REFERENCES
                      (continued)
10.  "Projects of the Industrial Pollution Control Branch,
     July 1971," Water Pollution Control Research Series
     # 12000-07/71, pp. 5-22, 5-23, 5-24, 5-25.

11.  Ibid., pp. 5-13, 5-26.

12.  Proceedings of the TAPPI 8th Water and Air Conference,
     Boston, Mass., 1971, paper entitled "Activated Carbon
     System for Treatment of Paper Mill Washwaters."

13.  Ibid., "Color Removal from Kraft Bleach Wastes by Ion
     Exchangers."

14.  "Carbon Treatment of Kraft Condensate Wastes," TAPPI,
     5^, 241, 1968.

15.  Lacey, R.E. and Loeb, S., eds., Industrial Processing
     with Membranes,  Wiley-Interscience, New York, 1972,
     pp. 223+.

16.  Moore, G.E. Minturn, R.E., et. al., "Hyperfiltration
     and Cross-Flow Filtration of Kraft Pulp Mill and
     Bleach Plant Wastes," ORNL-NSF-EP-14, May, 1972.

17.  Wiley, A.J., Dubey, G.A. and Bansal, I.K., "Reverse
     Osmosis Concentration of Dilute Pulp & Paper Effluents,"
     for the Environmental Protection Agency, Project
     #12040 EEL, February, 1972.

18.  Morris, D.C., Nelson, W.R. and Walraven, G.O., "Recycle
     of Papermill Waste Waters and Application of Reverse
     Osmosis," for the Environmental Protection Agency,
     Program #12040,  January, 1972.

19.  Bansal, I.K., Dubey, G.A. and Wiley, A.J., "Develop-
     ment of Design Factors for Reverse Osmosis Concen-
     tration of Pulping Process Effluents," presented at
     Membrane Symposium, National Meeting of American
     Chemical Society, Chicago, Illinois, September 14-
     18, 1970.

20.  Beder, H. and Gillespie, W.J., "Removal of Solutes
     from Mill Effluents by Reverse Osmosis," TAPPI, 53,
     No. 5, May 1970 , pp.  883-887.
                            194

-------
                      REFERENCES
                      (continued)
21.  Bregman, Jacob I., "Membrane Processes Gain Favor
     for Water Reuse/' Environmental Science £ Technology,
     4_, No. 4, April 1970, pp. 296-302.

22.  Wiley, A.J., Dubey, G.A., Holderby, J.M. and Ammerlaan,
     A.C.F., "Concentration of Dilute Pulping Wastes by
     Reverse Osmosis and Ultrafiltration," Journal WPCF,
     42, No. 8, Part 2, August 1970, pp. R279-R289.

23.  Ammerlaan, A.C.F. and Wiley, A.J., "Pulp Manufacturers
     Research League Demonstrates Reverse Osmosis Process,"
     TAPPI, 52, 1969.

24.  Ammerlaan, A.C.F. and Wiley, A.J., "The Engineering
     Evaluation of Reverse Osmosis  as a Method of Pro-
     cessing Spent Liquors of the Pulp and Paper Industry,"
     prepared for the New Orleans Meeting of A.I.Ch.E.,
     March 17-20, 1969.

25.  Ammerlaan, A.C.F., Lueck, B.F. and Wiley, A.J.,
     "Membrane Processing of  Dilute Pulping Wastes by
     Reverse Osmosis,"  TAPPI, 52,  No. 1, January 1969,
     pp. 118-122.

26.  "Industrial Ultrafiltration,"  in  Membrane Processes
     in Industry and Biomedicine, Plenum Press,  1971.

27.  Flinn, J.E., ed., Membrane  Science and  Technology,
     Plenum Press, New York,  1970,  pp.  33+,  47+.

28.  Foreman, G.E., et al.,  "The Improvement of  Spiral-
     Wound Reverse Osmosis Membrane Modules," Research
     and Development Progress Report,  No.  675, Office of
     Saline Water, April,  1971,  pp.13.

29.  Cadotte, J.E., and Rozelle, L.T.  of North Star
     Research and Development Institute,  "In Situ-Formed
     Condensation Polymers  for Reverse Osmosis Membranes"
     for Office  of  Saline Water, U.S.  Department of  the
     Interior, Third Quarterly Report,  January 9,  1972,
     through  April  8,  1972,  Contract  No.  14-30-2883.
                           195

-------
                       REFERENCES
                       (continued)
30.    Hossain, S.,  Goldsmith, R.L.,  Tan, M., Wydeven, T.,
      and Levan, M.I., "Evaluation of 165°F Reverse
      Osmosis Modules for Washwater Purification", for
      Intersociety Conference on Environmental Systems,
      July 1973, San Diego.
                          196

-------
                APPENDIX A
DETAILED PILOT  PLANT PROCESS DESCRIPTION




          (as  initially installed)
                   197

-------
                      APPENDIX A

      DETAILED PILOT PLANT PROCESS DESCRIPTION

               (as initially installed)
A generalized description of the process was given in
Section iv.A. The following information describes the
pilot plant, as initially installed, in detail.

A detailed process flow is shown in Figure A-l.  The
process description which follows is based on normal
operation.  Components which were used at other times
are described in Table A-I, a complete listing of all
system components by intended function.

Referring to Figure A-l, the numbers contained in the
diamond symbols refer to the process streams listed at
the bottom of the Figure.  This table gives average,
maximum, and minimum flow rates at different points in
the system.

Feed flow from the plant was to the 500 gal. fiberglas
feed tank (T-l).  Flow was through a float valve  (V-7),
which kept T-l filled so long as feed was available from
the mill.  If feed flow were interrupted, a low level
shutdown switch in the tank (LS-F) shut down the pilot
plant and sounded an alarm.  The unit was fitted with
a Lightnin mixer  (M-l) which kept the tank contents well
mixed, enabling pH and temperature to be controlled ac-
curately.

Temperature was measured and controlled by a probe in-
stalled in the tank and control unit (TIS-1), which
controlled flow of cold water or steam through a heat
transfer coil installed in the tank (HE-1).  Although
feed left the mill hot (approximately 120-135° F) some
cooling occurred before the feed flow reached the feed
tank.  Consequently, a control system xvas provided to
allow either heating or cooling by manual selection.
An automatic shutdown switch and audible alarm was in-
cluded in the temperature controller, which would shut
down the system should a maximum preset temperature be
exceeded.  A separate temperature probe in T-l was con-
nected to a recorder giving a continuous record of feed
temperature  (TR-1).
                           198

-------
                                        jtn*

                                        n«s«
«r
fr*

Vi
«



M-7
^i..__.

1
c
i^

.,„ .,. , ••>— ' '
JMUMUIUUli 	 	 Lc 	
MM MM>»tt« 	 1*7^ 	
«• »»;iT '•':.!* '
*^^T^ "•« ^A»*« •J*4_s/K«k * ! O*«S* 4Ljd
_=)_._. _l

*r£



                                                                                             —T    fruTB* lAfrWAjwl
                                                                                             ti    )   <".*.*.  |

J
-7~y«rw
ti««J
^°3
	 —ffitfj-
er
•m
	 A 	
;

NO

VO .
                                                                                                                                lUUt*
                                                                            I
                                                                     &1MC 4- 100 s«. PT. MMMAMi: AIC4
      (12f&'-t)
                                                                    in  »w

                                                            ±SM,^-©(2>       »Mt~,


                                                           -tfT^^Hfe^EHI
                                                           fca  ^/T^T""T^r'n
                                                           'Sf,**1^   *           ft
r«i».-o      ^
      "• WAIVC





     [Xj YAlVt OMN O««IN* I/' OH"»»TI<»H




     ^•4 IMIVt CLOSED PUHINC ur OCKKATItfM


    »>A •»HHtrOKDl»

     T»r  Tfftt.*. ftKCOftOtM

     f S • T(.M^. SWITCM

     Ll • civil SWITCH

am
s
<

1>
k
t
t

OK
i
»,**
j«
*
!•
•TC i
IT
i
1
1

£!£
!•
!•*
••
«,•
*n i

4.1
L*
1

«T*
ns
'»
>
.»^
"? f

.>
i

p-i
UEl
11
it
^^^
ml

i.ii
i

a
11
*
nu
t



s
im
tB.
i
«•
1

41
>
t

i
a
i
i
H
n


f
wr
•
m
»?
n
1 1

I
n
•»i
•
JL

W
H
ff


aa
v~
MM
et

i*
™i

.i
f
V
Jw
t>
£
-
PB
1*

1
ID
• JKh
IV
PI
IB
i
nf
f t

I
r
«*
r.
M
_l
1
I
!
n


•* t-
K
"
nrc
|4

1

E
10
>0
•
tM
*



«

MM
1

I
k

M
L
u
•
i
•
r



i
i

at
»f

s
1

i
a
i
W
m
fT



M
to.
1
k
,•
• .
r?
ii
*

i
1

f

L

a
i

9,1
11
1

ff
4
1
•K
tti

s..

•DC
11
1
IK
Lft
»
,-

t , -
*T*
T
>
* v
BE
11
•"
•

1

tSS
M
H
I
                                    • "now a i*rn MV
                                                                                                                                23S£.3r.1*
                                                                                                                                Mtinmtjn HM
                                                                                                                                     m*jH
                                                                                                                                                =aa
                                                                                                                                                     «*    IVIt-lfnvMUlIT

                                                                                                                                                                CMMHDCt, MJW

                                                                                                                                                                «W
                                                                                                                                                                *r*irf
                                                                                                                                                                           5136-5003

-------
                                                  TABLE A'I

                                              COMPONENT LIST
  Code
Name
Identification
  CS-1
Composite
Sampler
o
o
  F-1A \
  F-1B J
  P-2
  FM-1
  FI-1C
Process
Filters
Water Filter
Integral Feed
Flow Meter
First Stage
Concentrate
Flow Meter
Location
(assumes operator is
 facing control panel)
                                                                                Function
Sigmamotor Pump, Model
AJ>4, 4-channel finger
pump
Enclosed in dog house
on top of module tray
Two Broughton Corp.
Model 3000 'filters
with 10 y stainless
steel baskets

Broughton Corp.
Model 990 filter with
200 mesh stainless
steel basket

Water meter
Brooks Instrument
25 gpm rotameter
Model 1305
On module, left hand
side
On module, left hand
side
Mounted on top of
barrel of P-l; on
module

Control panel
Continuously pumps small
quantitites of four process
streams to composite
sample collector*.  The
four boxes labelled CS-1A,
CS-1B, CS-1C, and CS-1D
in Figure 2 represent
channels for raw feed,
neutralized feed, final
concentrate, and mixed
permeate, respectively.

Filters neutralize the.feed
prior to introduction into
the membrane assemblies.
Filters water used, for
backwash of F-1A and
F-1B.
Records the total amount
of process fluid handled
by the membrane system.

Measures Stage 1- concen-
trate flow.

-------
                                                  TABLE A-I
                                                  Page 2
  Code
Name
Identification
                                                         Location
                                                                 Function
ts>
  PI-C
FI-1PA
FI-1PB
FI-1PC
FI-2P
FI-3P
FI-4P
FI-5P .

FI-P
   HE-1
   LS-F
               Concentrate
               Flow Metera
Concentrate
Flow Regulator
and Flow Meter
               Permeate Flow
               Meters
               Mixed Permeate
               Flow Meter
Heat Transfer
coil
Feed Tank Level
Switch
                 Brooks Instrument
                 10 4pm rotameters,
                 Model 1305
Brooks Instrument
Self-contained flow
controller, Model
1350-8802-5-65C

Dwyer Instruments
polycarbonate rota-
meters, Models RMC
141,142,143
Brooks Instrument
25 gpm rotameter,
Model 1305

25 sq ft, 1/2"
diameter titanium
coil

Gems level switch,
Model LS-1900
                         Control panel
Control panel
                                          Control panel
                       Measures concentrate
                       flows for Stages 2-5.
Regulates and measures
concentrate flow from
Stage 5.
                       Measures permeate flow
                       rates from each of the
                       seven membrane assemblies.
                                          Control panel
Feed tank  (T-l)
Feed tank  (T-l)
Measures flow rates of
mixed permeate from all
membrane assemblies.

Heat transfer surface
for heating or cooling of
process fluid,

Shuts down membrane system
and sounds alarm on low
level of process fluid

-------
                                                  TABLE A-1
                                                  Page 3
  Code
Name
Identification
  LS-P
  M-l
  P-l
M
O
N9
Permeate Sump
Level Switch
Gems level switch.
Model LS-1900
Location
Function
Feed Tank Mixer  Lightnin  mixer, Model
                 ND-2A
Stage 1 Feed     Goulds multistage
Pump             centrifugal pump, Model
                 MB-13400
               Circulation      Goulds multistage
               Pumps for Stages centrifugal pumps/
               2,3,4, and 5     Model MB5100
In permeate sump tank
(T-2)
                         Feed tank (T-l)
                         On bottom of module
                                          On top right hand
                                          side of: module
Shuts down permeate
pressurization pump (P-8)
on low level in permeate
sump. tank.

Mixes feed tank contents
allowing efficient
neutralization and temp-
erature control.

Pumps process fluid
through the automatic
backwash filter unit and
Stage 1 membrane assemblies
This pump also provides
feed pressurization in
subsequent membrane stages.

Boosts pressure of feed
sequentially to Stages 2,
3,4, and 5.
  P-6
Feed Booster
Pump
Corcoran 1 hp
centrifugal pump,
Mounted on feed tank
panel
Transfers feed from feed
tank to suction of P-l.
               Acid Pump
                 Precision Control
                 diaphragm pump, Model
                 11321-71
                         Mounted by acid drum
                         on mezzanine
                       Pumps acid from 55 gal.
                       drum to feed tank for
                       feed neutralize*"'on.

-------
                                                 TABLE A-1
                                                 Page 4
 Code

 P-8
Name
Identification
 PHIS-1
Permeate
Pressurization
Pump
pH Indicator/
Controller
Location
Little Giant plastic
head pump, Model
1P732 (Grainger)
Leeds and Northrop
monitor, Model 7070-
02-3-107-6-04, with
pH electrodes and probe
Mounted on underside
of module tray
Monitor mounted in
in control panel; pH
probe assembly in
feed tank
Function
CO
o
u>
 PHR-1
pH Recorder
Rustrak Instrument
millivolt recorder
Control panel
Transfers pressurized
permeate into membrane
modules to keep them
pressurized and wet during
system shutdown.

Measures and controls
pH of the; process fluid.
An electrode probe
assembly is mounted in the
feed tank and input is
transmitted to the control
panel.  The low alarm on
the monitor is used to turn
the acid pump on and off
with a dead band range of
0.5 pH unit.  The high
alarm of the monitor is
used to shutdown the system
and sound an alarm on
rising pH.
Records pH. The
output from the Leeds and
Northrop controller, serves
as the input to the
recorder. Full-scale on
the recorder (10 divisions)
corresponds to the pH
range 2-12.

-------
Code
PI-
PI
-Flj
-F2J
PI-6
                                                TABLE A-I
                                                Page 5
           Name
                 Identification
Water Filter
Pressure Gauges
Broughton Pressure
Gauges
                         Location
           Process Filter
           Pressure Gauge
                 0-400 psi Pressure
                 Gauge
                       Function
On water filter
                         On bottom of module,
                         in the piping running
                         from P-l outlet to
                         process filter inlet
Measures pressure drop
across water filter (F-2).
High pressure drop during
the backwash cycle in-
dicates the need to "blow-
down" the water filter.

Measures pressure,, at the
filter inlet.
PI-1I
PI-IE
PI-2I
PI-2E
PI-3I
PI-3E
PI-4I
PI-4E
PI-5I
PI-5E

PS-F
           Pressure Gauges
PS-2
                 0-300 psi Pressure
                 Gauges
                         Control panel
           Filter
           Differential
           Pressure Switch-
                 Static 0-Ring
                 Pressure Switch,
           Low Pressure
           Switch
                 Static O-Ring
                 Pressure Switch,
                         Mounted in module
                         between inlet and
                         outlet piping of
                         process filter
                         Mounted in piping
                         manifold on suction
                         side of P2-P5; top
                         right-hand side of
                         module
                       Measures inlet and outlet
                       pressures to each of the
                       five membrane stages.
                       Senses pressure differen-
                       tial across the process
                       filter, triggering an
                       automatic backwash cycle
                       when the pressure drop
                       builds up to a specified
                       level  (nominally 75 psi)

                       Shuts down pumps P2, P3
                       P4, and PS when their
                       suction pressure drops
                       below a preset value
                       (nominally 35 psi}

-------
                                                  TABLE A-I
                                                  Page 6
 Code
Name
Identification
Location
Function
 S-l
 T-l
 T-2
to
s
 TIS-1
Strainer




Steam Trap


Feed Tank
Permeate Sump
Tank
Temperature
Indicator/
Controller
Common Y Type
Pipeline Strainer
Next!to upper left-
hand corner of con-
trol panel
                                                         On exit line of HE-1,
                                                         located on Feed Tank
500 gallon fiberglas
tank
5 gallon polyethylene
tank
United Electric, Model
1202 Temperature
Indicating/Controller
On top of module
tray
Mounted on Feed Tank
control panel
 TR-1
Temperature
Recorder

Temperature
Indicators
Rustrak Instruments
Temperature Recorder

Temperature Gauges
Control panel
                                                         On module; in inlet
                                                         piping to membrane
                                                         stages 1,2,3/4, and 5
Removes suspended solids
before concentrate flow
regulator to prevent
plugging

Removes condensed steam
Surge for process fluid,
and mixing vessel for
control of pH and temp-
erature.

Serves as reservoir for
permeate for system pres-
surization during shutdown

Measures and controls temp-
erature of process fluid
in feed tank.  Electronics
control operation of V-9
allowing either steam or
cooling water to pass
through HE-1 when called
for by the sensing element.

Continuously records feed
temperature in Feed Tank.

Indicates temperature of
feed to each- of the five
membrane stages

-------
                                                   TABLE A-1
                                                   Page 7
  Code
  V-1I
  V-2I
  V-2E
  V-3I
  V-3E
  V-4I
  V-4E
  V-5I
  V-SE;
             Name
                 Identification
             First Stage
             Control Valve
                          Location
                 1" globe valve
             Control Valves   1/2" globe valves
                          Lower left-hand
                          front of module

                          Control panel
                        Function
                        Throttles feed flow from
                        P-l.

                        To control flows and pres-
                        sures through membrane
                        stages 2,3,4, and 5; to
                        isolate any of stages 2,3,
                        4 and/or 5, as desired.
  V-1SI
  V-1SE (
  V-2S1 7
  V-2SE
  V-3SI
NV-3SE
.*V-4SI
  V-4SE
  V-5SI
  V-5SE/
        \
V-1SPA
V-1SPB
V-1SPC
V-2SP.
V-3SP
V-4SP
V-5SP  J

V-1R
V-2R
V-3R
V-4R
V-5R
             Feed Sample
             Valves
                 1/4" Petcocka
                         Control Panel
Permeate Sample
Valves
1/4" petcocks
In permeate piping on
right hand side of
Module
                        To sample inlet and outlet
                        feed flows to each of the
                        five menbrane stages.
Valves for collection of
permeate samples from each
of the seven membrane
assemblies.
               Pressure Relief
               Valves
                 Brass Pressure
                 Relief Valves
                         In piping to inlet of   Pressure relief valves
                         each of membrane stages preset at 175 psig to
                         1 through 5             prevent over-pressurization
                                                 of any of the five membrane
                                                 stages should system mal-
                                                 function occur

-------
                                                TABLE A-I
                                                Page 8
Code
Name
Identification
Location
                                                                               Function
V-PR
V-6


V-7




V-8


V-9
Permeate Pressure Brass Relief Valve
Relief Valve
             Check valves
             Sample Valve


             Float Valve




             Drain Valve
             Steam
             Solenoid
             Valve
                 Check-all 1/2"
                 Union Check valves
                 1/2" globe valve


                 Fisher Controls
                 float operated valve,
                 Type 17IF


                 2" ball valve
                 Automatic Switch
                 Co. Steam Solenoid
                 Valve, Model 8222B24
                         In permeate line from
                         Stage 5
                         In piping on right
                         hand side of module
                         On feed inlet piping
                         to Feed Tank

                         Feed tank
                         On tank outlet
                         In feed tank on inlet
                         to HE-1
                        Pressure relief valve
                        preset approximately at
                        50 psig,+*> relieve prensur
                        if overpressurization
                        should occur in any of
                        the pern.eate piping.

                        To allow pressurized
                        permeate to pass into the
                        membrane stages upon shut-
                        down, and to prevent
                        process fluid from enterin
                        the permeate sump tank whe
                        the system is operative.

                        To draw off raw feed for
                        sampling or drainage.

                        To maintain feed tank
                        at the desired level so
                        long as process feed is
                        available.

                        Drain valve to drain tank
                        contents

                        To control flow of steam,
                        or cooling water to HE-1,
                           actuated by TIS-1.

-------
                                                  TABLE A-I
                                                  Page 9
  Code

  V-10
Name
Identification
Location
Function
Check Valve
  V-ll
First Stage
Pressure Relief
Valve
8
  V-12
  V-13
Check Valve
First stage
bypass valve
  V-14
Ball Valve
1" Union Checkall
Check Valve
D'Este Type V
Direct-acting back-
pressure regulator
with external pressure
sensing
On suction of P-l
Behind process filters.
1" Checkall Union
Check Valve
1" globe valve
In piping between
filter outlet and
Stage 1 inlet
At top of control
panel
Brass ball valve
On suction side of
booster pump (P-6);
on feed tank
Prevents flow reversal
into the feed tank when
the system is shutdown
and operating under
permeate pressurization.

Maintains constant pressure
at filter outlet.  If '
pressure drops at the
filter outlet due to cake
build-up on the filter,
V-ll closes thus reducing
the amount of bypass arounc
the filter, and increasing
the inlet pressure to
the filter.

Prevents flow backing-up
into filters during system
shutdown and pressurizatior
with permeate.

To control recirculation
rate around stage 1.  In
normal operation it is
anticipated that this
valve will be closed.  This
is to be confirmed in
actual operation.

Isolation valve, for service
of P-6.

-------
                                                TABLE A-1
                                                Page 10
Code

V-15
Name
Identification
Location
Function
Concentrate
Solenoid
Valve
1/4" Asco solenoid valve,On module,in concentrate This normally-closed
                              Model 8262A233
                         line just before FI-C
V-16
Solenoid Valve
1" normally-closed
Asco solenoid valve,
Model 8211B27
In piping on moduli
just before FM-1
V-17
Bypass valve
                         On outlet line of
                         HE-1 (on tank)next to
                         ST-1
V-18
Transfer valve
1" brass globe
valve
Feed Tank, on booster
pump outlet
solenoid valve is open in
operation, allowing
concentrate to be removed
continuously from the
system.  When the system
is shutdown the solenoid
valve closes, thus
preventing pressurized
permeate removal from
the system through the
concentrate line.

To prevent continued
drainage of feed tank
contents into the membrane
system during automatic
shutdown.  This solenoid
valve will open only when
the system is operational.

This valve is to be opened
when cooling water is
passed through HE-1.   This
permits water flow to pass
to drain without having
to pass through the steam-
trap (ST-1).

This valve allows removal
of feed tank contents
after neutralization and
temperature control to
another system,if desired.
Feed material can be
pumped out this line for
pilot filtration tests.

-------
                                                  TABLE A-I
                                                  Page 11
 Code
Name
Identification
 V-19
 V-20
 V-21
to
M
o
 V-23
 V-24
 V-25
 V-26
 V-27
 V-28
 V-29
 V-30
 V-31
 V-FIA")
 V-F1B (
 V-F2A >
 V-F2BJ
Location
Function
Isolation Valve  1" brass ball valve
Flush valve
1/2" brass globe valve
Permeate by-
pass valve
1/4" needle valve
Isolation
Valves
                               1/4" needle valves
Solenoid Valves
1" Asco 2-way normally-
closed brass solenoid
valves, Model 8211B27
On outlet line of
booster pump (P-6),
on feed tank

In concentrate line
from stage 5; on
module behind water
filter (F-2)
On bottom side of
module tray
                          At inlets to pressure
                          indicators; on rear
                          side of control panel
On piping to process
fluid filters
                                                 System isolation, as
                                                 needed.
When open, this permits
rapid flushing of all five
membrane stages since the
flow regulator (FI-C) can
be bypassed

To allow continuous bleed
from the permeate pump
(P-8) back to the permeate
sump tank  (T-2), to avoid
overheating P-8 when
pressurized permeate is
not being transferred to
the membrane system.

To isolate pressure'gauges
should removal be required
during opera*-ion.
To isolate process filters
from feed during the
backwash cycle.

-------
                                                TABLE A-1
                                                Page 12
Code
Name
Identification
Location
                                                                               Function
V-F3A)
V-F3B (
V-F4A f
V-F4B )
V-F5A")
V-F5B /
V^F6A \
V-F6B .)

V-CSA ^
V-CSB y
V-CSC {
V-CSD }
Solenoid Valves
Check Valves
Petcocks
2" Asco 2-way
normally-closed brass
solenoid valves,
Model 8211B82
1" and 2" brass check
valves
1/4" Petcocks
On piping to process
fluid filters
Installed in piping to
process fluid filters
Controls flow of backwash
water through the filter
during backwash cycle.
These four solenoid valves
are on the water lines to
the process filters.

Insures positive seating
of solenoid valves under
reverse pressurization.
V-CSA on feed tank in-  To isolate the composite
let line, V-CSB on      sampler from the four
inlet to P-l above      sampling points as
P-l barrel, V-CSC on    needed.
concentrate line
behind process filters,
V-CSD on mixed per-
meate line behind
filters

-------
A similar system was used to control and record pH.  A
probe was installed in the tank and connected to a con-
trol system (PHIS-1).   On rising pH the "low" contact
actuated an acid pump (P-7)  which transferred sulfuric
acid from a drum to the surge tank.  This same contact,
utilizing a "dead band" of 0.1 pH unit, stopped the acid
pump on falling pH.  A "high" alarm, set at a pH above
the normal operating range, would shut down the system
and sound an audible alarm if pH in the Feed Tank rose
above a preset limit.

Feed from the Feed Tank was pumped through a five-stage
ultrafiltration system which contained spiral-wound mem-
brane modules.  Feed from the feed tank was pumped by a
small booster pump  (P-6) through a cumulative flow meter
(FM-1) to the suction of the Stage 1 pump  (P-l).  Flow
from the pump was through a throttle valve  (V-II) and a
back-flushing filter system  (F-1A and F-1B) for removal
of residual suspended solids, prior to introduction in-
to the ultrafiltration unit.  Filter inlet and outlet
pressures were measured  (PI-6 and PI-II).  An automatic
backwashing cycle was triggered by a differential pressure
switch  (PS-F) installed between the filter inlet and out-
let lines.  A pressure relief valve (V-ll) installed on
the upstream side of the filter, but controlled by a sen-
sing element on the downstream side, maintained a constant
pressure at the filter outlet.  Recirculation of feed
through the back pressure regulator was to the Stage 1
pump suction.  After the filter, flow was through a check
valve  (V-12) into three parallel passes of membrane cart-
ridges.  Membrane Assemble 1-A initially contained East-
man Kodak spirals  (3)  and assemblies 1-B and 1-C contained
TJ Engineering spirals  (3 each).  Permeates from each pass
were collected separately and their flows were measured
and sampled individually.

On the inlet side of Stage 1, pressure  (PI-1I) and tem-
perature (TI-1I) were measured, and a sample valve  (V-1SI)
was used for collection of Stage 1 feed.  A pressure re-
lief valve  (V-1R) was provided as  a safety  device.  The
concentrate was sampled through a sample valve  (V-lSE)
and its flow rate  (FI-1C) and pressure  (PI-IE) measured.
A line was provided to recirculate part  of the concen-
trate  through a "bypass valve"  (V-13).   Stage 1  could
operate with or without recirculation through V-13.

Operation of the subsequent membrane stages was  always
with recirculation, since this was required to maintain
                           212

-------
a feed flow rate through the membrane cartridges at or
above about 4 gpm.  Flow on a "once-through" basis would
fall below this level, and recirculation was required
to achieve satisfactory operation.

Flow to Stages 2 through 5 was introduced into the suc-
tion side of the circulation pump for each stage.  A low
pressure switch (PS-2) was installed to shut down these
stages should an upset in flow occur.

For each stage (Stages 2 through  5 are identical) flow
from the booster pump passed through a throttle valve
into the membrane cartridges.  Pressure and temperature
were measured on the inlet to the membranes and pressure
and flow rate on the outlet.  Pressure relief valves were
provided as safety devices should overpressurization oc-
cur.  Sample valves allowed collection of feed and con-
centrate samples for each stage.  Permeate flows were
individually sampled and collected,  and their flow rates
measured.  Each stage had an internal recirculation loop
which permitted maintenance of the desired flow rate
through the membrane cartridges.  Initially, Stages
2 and 3 contained TJ Engineering  cartridges  (3 each) and
Stages 4 and 5 contained Gulf Environmental Systems cart-
ridges  (2 each) .

The final concentrate  from Stage  5 was passed through  a
flow control valve and  flow meter (FI-C) prior to collec-
tion for incineration  tests or discharge.

Each stage also had a  means of introducing pressurized
premeate to the feed  side of the  membrane cartridges  for
wet storage during system shutdown.   The system  for sup-
plying pressurized permeate to the membrane  modules was
quite simple.  Permeate  from the  five stages was  manifol-
ded and introduced into  a permeate surge tank  (T-2) .
A permeate pump  (P-8)  automatically  pumped permeate to
the membranes  for each stage.  Flow  to  the membranes was
through check  valves,  so that when the  system was opera-
ting high-pressure  feed could not be pumped  into the per-
meate system.  During shutdown the permeate  surge tank
remained  filled since all permeate  fed  to  the membrane
system was returned  as permeate.   During operation, per-
meate passed  through  the  surge tank  and overflowed  to
drain.
                                         o
Membrane  areas were  approximately 300 ft   for  Stage  1
and 100  ft2  for each  of Staqes  2  through  5.  Total
                            213

-------
membrane area was approximately 700 ft2.  At a nominal
membrane flux of 15 gal./day/ft2 (gfd) the plant had a
capacity of 10,500 gpd.

A revised pretreatment system flow schematic is given
in Figure A-2.  The changes were made to permit instal-
lation of various filters, a cleaning system, and once-
through flushing of the membrane shells.
                            214

-------
     •••
IV

-------
               APPENDIX B
      ULTRAFILTRATION PILOT PLANT
SPIRAL MEMBRANE CARTRIDGE IDENTIFICATION
                    217

-------
                              APPENDIX  B                Page 1

ULTRAFILTRATION PILOT PLANT SPIRAL MEMBRANE  CARTRIDGE  IDENTIFICATION
UF"^-^^^ Date
Cartridge^^^^^
Location 	 ..
Stage
la


Ib


Ic


2


3


4


5



Cartridge
Position
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
8-19-72
Type
E
E
E
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
Gulf
Gulf

Gulf
Guli

«
1
2
3
1
2
3
4
5
6
7
8
9
10
11
12
1
2

3
4

8-22-72
Type
E
E
E
T.J.
.T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
Gulf
None
Gulf
Gulf
None
Gulf
*
1
2
3
16
17
18
13
14
15
7
8
9
10
11
12
1

2
3

4
8-30-72
Type
T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
E
E
E
T.J.
T.J.
T.J.









*
4
5
6
1
2
3



19
20
21









8-31-72
Type
E
E
E
T.J.
T.J.
T.J.
Gulf

Gulf
T.J.
T.J.
T.J.
.T.J.
T.J.
T.J.
T.J.
T.J.
T.J.
Gulf

Gulf
f
1
21
3
1
21
3
1

2
19
20
21
10
11
12
13
14
15
3

4
218
                                                  Eastman
Lot numbers  of  the  first  set  of membrane
cartridges were not recorded,  and they
are therefore numbered starting from
No. 1.  Prom 11/1/72 all  cartridges are
identified by manufacturer's lot number'.
Entries denote when cartridges were changed;
if no entry  is given, then the cartridge was»not changed.

-------
                              APPENDIX B•




ULTRAFILTRATION PILOT PLANT SPIRAL MEMBRANE CARTRIDGE  IDENTIFICATION
UF -^^^ Date
Cartridge"*— 	
Location --^
Stage
la


Ib


Ic


2


3


4


5


Cartridge
Position
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
11-1-72
Type
T.J.
n
it
n
n
n
n
ti
n
n
n
n
n
n
n
n
n
H
n
it
n
*
4713
4723
4724
4715
4709
4711
4725
4722
4719
4717
4716
4714
4708
4736
4712
4737
4735
4731
4718
4720
4721
11-3-72
Type


















T.J.
T.J.
T.J.
*


















4718
4720
4523
11-5-72
Type






T.J.
T.J.
T.J.









Gulf

Gulf
ft






3747
4725
4722









3

4
11-18-72
Type















T.J.
T.J.
T.J.



#















3752
4731
4735



                                              E
Eastman
                                  219

-------
                              APPENDIX  B                   Page 3

ULTRAFILTRATION PILOT PLANT SPIRAL MEMBRANE CARTRIDGE IDENTIFICATION
UF"^^^ Date
Cartridge "~-^^^
Location ^^--^...
Stage
la


Ib


Ic


2


3


4


5


Cartridge
Position
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
12-8-72
Type
T.J.
n
ii
it
n
n
n
n
n
n
n
n
n
n
n
n
n
n
Gulf

Gulf
. #
1948
1970
3749
4715
4709
4711
3747
4725
4722
4717
4716
4714
4708
4736
4712
3752
4731
4735
3

4
12-19-72
Type
T.J.
n
n
n
' n
n















#
4715
4709
4711
1948
1970
3749















12-21-72
Type
T.J.
n
it
it
n
n
ii
it
it
n
n
n
it
n
it
Gulf

Gulf
Gulf

Gulf
#
4715
4709
4711
2026
4283
4395
3747
4725
4722
4717
4716
4714
4708
4736
4712
1

2
3

4
1-4-72
Type









T.J.*
T.J.*
T.J.*









f









4829
4828
4827









          *Wide  channel
           corrugated  spacer
           cartridges
E = Eastman
                                 220

-------
                               APPENDIX  B                 Page 4




ULTRAFILTRATION PILOT PLANT SPIRAL MEMBRANE  CARTRIDGE IDENTIFICATION
UF*" — "---^ Date
Cartridge "--v^^^
Location — ,.
Stage
la


Ib


Ic


2


3


4


5


Cartridge
Position "
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet
2-8-73
Type
T.J.
it
it
ti
ii
ti
n
ii
n
n
n
n
n
n
ii
H
n
n
n
n
it
* .
1970
3748
2209
2026
4283
4395
3751B
3782
3751A
4829
4828
4827
4708
4736
4712
4731
4721
4737
4735
4720
4718
2-10-73
Type






T.J.
T.J.
T.J.












*






3749
3782
3751B












2-14-73
Type
T.J.
T.J.
T.J.


















*
1950
4722
4714



















Type





















*





















                                               E = Eastman
                                 221

-------
                  APPENDIX C
DETAILS OF FILTERS USED  IN PILOT PLANT PROGRAM
                        223

-------
       Can save by recovering solids
       formerly lost in effluents
       Reduces operating costs and
       improves efficiency in process
       separations and in pol Iution
       control
       Minimal maintenance and no
       power costs—there are no
       pumps, motors, nozzles or
       moving parts
       Provides fast return on invest-
       ment  (often within months)
       Unique screen is self-cleaning,
       non-clogging and puncture-
       proof . Needs I itt le or no
       attention
       Easy to install. Saves space
                                   The patented C-E Bauer
                                   Hydrasieve™ is a simple, highly
                                   efficient screening device for
                                   removing solids from low con-
                                   sistency slurries.
                                     Exclusive design features and
                                   the  unit's ability to operate con-
                                   tinuously for extended  periods
                                   without attention make the rugged
                                   Hydrasieve a reliable profit builder
                                   in a broad range of fluid removal
                                   applications. Hundreds of units
                                   are now in use.
                                     Typical liquids/solids separa-
                                   tion operations include dewater-
                                   ing, thickening, recovering usable
                                   solids, classifying, and
                                   fractionating.
Installations. Waste water treat-
ment and pollution control sys-
tems. Municipal  sewage plants.
Chemical, plastic, and ceramic
classification. Synthetic and
natural fiber recovery. Salvaging
rubber fines. Processing soup in-
gredients, fish, citrus fruits, etc.
Recovering  hog hair and other val-
uable solids for meat and hide
processors. and similar operations.
   C-E Bauer developed the Hydra-
sieve originally for the pulp and
paper industry. In high density
pulping systems, it pre-thickens
pulp ahead of the Bauer Helipress"
     HOW IT WORKS
                Gravity feed
                of liquids/solids I
Self cleaning.
non clogging stainless
steel screen for
continuous dewatering
                                   ,
                             Rugged all stainless
                             steel or fiber
                             glass construction
                             minimizes maintenance
                              _. Headbox


                               Alternate
                               feed inlet
Removed or
recovered
solids
                                              Note how "Coanda" effect strips liquid from
                                              bottom of the stream flowing down  the
                                              screen. This "wall attachment" effect
                                              accelerates fluid removal action.
                                                       224
                     •The Hydrasieve is manufactured under U.S. Letters Patents No. 3451.555 and 3.452.876. and patents pending
                     Canadian patent No  835737; Great Britain 1.196.303/1.196.304. Menico 99.076. France 1.529.448: Belgium
                     729 813. Other patents are pending in the United States, and foreign countries.
"Cleaned" effluent improves plant
pollution control efficiency,
lowers B.O.D.. reduces sewage rates
     Downward curve of Bauer screen bars
     divides the flow of slurry into separate
     streams between the vertical supports thus
     preventing clogging or blinding.

-------
         the revolutionary
HYDROMATION IN-DEPTH  FILTER
             225

-------
           THE  SPARKLER  MCRO  FILTER
         IS  A  REVOLUTIONARY DESIGN
             WITH  THE  EMPHASIS  ON
           SIMPLICITY AND  EFFICIENCY
            OVERHEAD  SUSPENSION  DESIGN
Engineered for heavy duty service
Rugged structural steel frame
Rigid support for the tank and cover
Stationary cover - - unnecessary to break
inlet and outlet piping to open filter
Retractable tank with 4 point suspension
Perfect meshing of tank and cover — posi-
tive seal
•  No cake disturbance in opening - - plates
  remain stationary
•  Internal self-sluicing
•  Dry cake removal either by retracting tank
  or internal conveyor screw
•  While in operation Vz of frame space free
  for traffic
•  Simple to completely automate
•  Capacity 10 to 300C sq. ft. filter area
                              226

-------
F.O.
ITEM
NO.
1
2
3
4
5
6
7
8
9
to
II

PAR Jo. PER CONSTRUCT. MAT'L.
DWG . NO .
3002-000-
3001-001 -
3000-01 1 -
3001 -014-
-015-
-016-
-017-
-018-
3001 -019 -
250?
£75-
BRONZE
-001
-001
-00^
-004
-001
001
-004
SEE

	
0-RING MATERIAL
NITRILE BUNA N
TEFLON COATED
300
MA.
MA)
CUJ
ADC
CU!
BRC
BA!
DA
-019-001
X.OPERATIN
<.OPR.TEMF
;TOMFR
STN.STL STEEL
-002 -003
-00 -001
-005 -006
-004 -004
-002 -003
002 003
-005 -006
•uuo uuc
TABULATION
< c
A &
AVAILABLE
VITON TEFLON
TEFLON COAT
3001-01
G PRES
'ERATUR
)RFSS
5T. ORDER
3UGHTON OF
5KET FILTEf
TE SHIPPED
Nn
ID. No,
R MEDIA
•

9-002 3001-0
SURE
F






QUAN DESCRIPTION

FILTER ASSEMBLY
BASKET ASSEMBLY
•'OP HOUSING
SAFETY LOCK
"OP CAP
CASING
INLET CONNEC"ION
RPlTTAM
0-RING
'2. ££/)AJ?
z /ita*/1 A/-/"/*^"

TEFLON
ENCAPSULATED
19-003 3001-019-004
PSI
•F


BAJ
473 W
WELLE


U—
^X
0ZZ&
fA


5


v^/^v. C^
SEC!
=7.! 25 — -i

f)! 1
f .? /"
L t
jfaj
ft

\

••"

^
sf

	 >*•
— 1
|

43.
fltdt



1
52
94
a ' — ' 3.375
'ION VIEW

19
©
©
\
Q FILTER ASSEMBLY
}T WILSON
^lUilion Com/jam
/ j
ASHINGTON STREET
ISLEY, MASS. 021S1
017)237-1755
BROUGHTON CORPORATION
GLENS FALLS NEW YORK, 12801
3000 FILTER ASM-2" PLAIN END
DWG. BY: ^u,^
CK'D. BY; «r>/
SCALE:
^ DA
ow

TE: i
G. NO
?oo
£-*<>.*
2-000-00_
/r^

-------
li
li
14
li
   !««• XH.T
   COYK »JH»tT
V«LVC
rcc
                 IITI4-C
                                                  NOTE
                                                  WHtN ORDERING PARTS. LIST SERIAL
                                                  NUMBER AND DRAWING NUMBE R F-2916
                                                    SPARKLER  MFG. COMPANY
                                                           CONROE   TEXAS
                                                  STANDARD L-6  PARTS  LtST
                                                  "GUARDIAN" TRAP  FILTER
DATE

SCALE:
                                                     If 5 '64
ORN BV

CKD BY
                                                                 WE J

                                                                 BCD
F29I6
                                       228

-------
                     TYPE  IB
                  175 P.S.I. @ 250°F
                  3/4* NPT Connections
                        TYPE CT
                     300  P.S.I, (n 200°F
                   3/t" or 1" NPT Connections
    1B1
                                           1B2
Three-piece housing; centerpost construction. Available in  one
and two-high cartridge models.

STANDARD — cast iron and steel  housing, asbestos shell gaskets,
fiber cap nut gasket, drain plug, steel internals, or, 304 stainless
steel housing and 304 stainless steel internals. Mounting pads on
all models are drilled and tapped for mounting bracket.
              DIMENSIONS-WEIGHTS
                                                                 CT-101
                                                                                                         CT-102
1B1
jrvj
3/«"
3/4"
12V
22%"
4l/4 "
4l/»"
9lbs.
12 Ibs.
Three-piece housing; ring nut construction holds sump to head.
Available in one and two-high cartridge models.

STANDARD — cast brass head and ring nut, drawn 304 stainless
steel sump, Buna N head gasket, 302 stainless internals. Heads with
H" connections  have  mounting pads drilled  and tapped for
mounting bracket. Also available  with a  304 stainless steel cast
head (ring nut nickel plated.)
                                                                           DIMENSIONS-WEIGHTS
           HOUSING CATALOG NUMBERS
 Mounting bracket
                                               #35581-01
             Filter utilizes:
               Micro-Klean Series G78 cartridge(s)
               all Micro-Wynd cartridge(s)
               Micro-Screen Series 52243 cartridge
               Micro-Screen Series 52043 cartridge
               Poro-Klean Series 50387 cartridge
           HOUSING  CATALOG  NUMBERS
  Mounting bracket
                                               135581-03
              Filter utilizes:
                Micro-Klean Series G78
                all Micro-Wynd carmdge(s)
                                                    229
                                       ORDER CARTRIDGES SEPARATELY

-------
      THE   RIGHT   FILTER   FOR   ANY   FLUID
                                      '-&
MICRO-KLEAN D
                                        ^4
                                         4
                                        3
                                          •«
                                MICRO-WYND
Liquid  and gas filters of AMF Cuno  design are available from
your distributor in several  types. Each filter tvpe  is described
below with detailed information further in the catalog.

1   MICRO-KLEAN  II... depth type,  disposable  fiber  car-
    tridge in micronic ratings. Ideal  for  removing  contaminant
    which is fibrous, abrasive or gelatinous in character. Typical
    applications  include gas, alcohols, glycols, coolants,  fuels,
    oils, lubricants, cosmetics, paints and  varnishes,  syrups, com-
    pressed air, water. See  pages 4 and 5. For Micro-Klean air
    line filters see pages 14  and 15.
    MICRO-WYND II . . . depth type,  disposable  wound  car-
    tridge in ratings from 1 to 350 microns. Intended for use where
    Micro-Klean fibers or resin binders are not compatible with
    the  fluid, or, where a wound  type cartridge is preferred.
    Typical  applications include plating solutions,  organic sol-
    vents, waxes, detergents, plastics, resins, deodorants, animal
    and vegetable  fats  and  oils, metal cleaning solutions.  See
    pages 6 and 7.
                4
                                MICRO-SCREEN
                                                            MICRO-SCREEN . . . surface type, metal screen  cartridge
                                                            —cleanable  and  reusable—in micronic  and screen mesh
                                                            ratings. Ideal filter media for final protection in ultra-clean
                                                            fluid  systems . . .  offering complete  freedom from media
                                                            migration. Also  intended for applications  involving  high
                                                            temperature  and corrosive conditions.  Typical applications
                                                            include steam,  strong acids,  concentrated  alkalis, strong
                                                            reducing agents and other chemicals which react with  fiber
                                                            type cartridges. See pages 8 and 9.
   PORO-KLEAN  . . . depth type, sintered metal cartridge—
   cleanable and reusable in 5, 10, 20 and 40 micron ratings.
   Used in place of Micro-Screen when contaminant is gelat-
   inous or fibrous in character. Also recommended  for heavy
   viscous fluids. Typical applications include steam, air, water,
   solvents, polymers,  acids, cellulose solutions,  process gases.
   See pages 8 and 9.
   PORO-KLEAN
        AUTO-KLEAN
                                                           CUNO-PORE .  . . depth type, disposable media in cartridge
                                                           and disc form.  Will remove contaminant which  is fibrous,
                                                           abrasive or gelatinous in character. Possessing natural dessi-
                                                           cating properties, it is ideal for removing trace quantities of
                                                           water from oil. Recommended for free-flowing liquids where
                                                           optical  clarity  is  of  primary  importance. Cartridge  type
                                                           filters are used  on dielectric insulating oils, coolants, deter-
                                                           gents,  cosmetics,  light  lubricants,  fuels,  degreasers  and
                                                           chemicals.
                                                           See pages 16 -19.
                                                           AUTO-KLEAN   . . edge  type, all metal filter. One  turn of
                                                           the handle cleans cartridge and restores full flow. Ideal for
                                                           removing  particles as small as 36 microns (.0015"). Typical
                                                           applications include paint, adhesives, resins, greases,  inks,
                                                           tar, cellulose solutions, waxes, soaps, hydrocarbon oils, fuels
                                                           and lube oils. See pages 20-23
                                                   230

-------
           APPENDIX D






MECHANICAL PROBLEMS OF DIFFERENT




   SPIRAL MEMBRANE CARTRIDGES
              231

-------
                         APPENDIX  D



MECHANICAL PROBLEMS OF DIFFERENT SPIRAL MEMBRANE CARTRIDGES
Date
11-5-72
12-7-72
2-7-73
tv>
U>
hi
2-13-73
8-21-72
11-15-72
12-7-72
Operating Information
Stage
la
la
la
la
Ib
Ib
Ib
Cartridge
Number
T.J. 4713
4723
4724
T.J. 4713
T.J. 4715
4709
4711
T.J. 1970
3748
2209
T.J.
T.J. 4715
T.J. 4715
Cartridge
Position
Inlet
Middle
Outlet
Inlet
Inlet
Middle
Outlet
Inlet
Middle
Outlet

Inlet
Inlet
Membrane Cartridge
Characteristics
(AP) , psi
50-60




50-60
50-60
Rejec-
tion
O.K.

O.K.
Low
Low
O.K.
O.K.
Problems
Brine
seals
loose fit
Brine seal
flipped
Brine
seals
flipped
O-ring
leak
Cartridge
failure*
Brine
seals
weak
Brine
seal
flipped
Comments
Reinforced brine seal
May have happened around
12-2 when AP was high
May have occurred around
1-4-73 when AP suddenly
dropped from a high value
to a low one
All three cartridges were
tested individually and
found to give good color
rejection
Membrane replaced on
8-22-72
Reinforced brine seal with
additional tapes


-------
                          APPENDIX D
                        (continued)
MECHANICAL PROBLEMS  OF DIFFERENT SPIRAL MEMBRANE  CARTRIDGES
Page 2
Date
12-20-72
8-21-72
9-13-72
N> 11-15-72
u>
u>
1-4-73
12-6-72
Operating Information
Stage
Ib
Ic
Ic
Ic
2
3
Cartridge
Number
T.J. 1948
T.J.
T.J.
T.J. 4722
T.J. 4717
4714
T.J. 4708
Cartridge
Position
Inlet


Inlet
Inlet
Outlet
Inlet
Membrane Cartridge
Characteristics
(AP) , psi






Rejec-
tion
Low
Low
O.K.
LOW
O.K.
O.K.
Problems
Cartridge
failure*
Cartridge
failure*
Brine seal
weak
0-ring
failure
Brine
seals
squeezed
out, flow
leaking
underneath
brine
seals
Brine seal
flipped;
other two
cartridges
were O.K.
Comments
Membrane replaced on
12-21-72
Membrane replaced on
8-21-72
Reinforced brine seal
with tapes
Replaced O-ring and
rejection became normal
May have happened
between 12-3-72 and
12-5-72 when the
(AP) was high


-------
                        APPENDIX D
                       (continued)                           Page  3

MECHANICAL PROBLEMS OF DIFFERENT SPIRAL MEMBRANE CARTRIDGES
Date
10-26-72
10-31-72
11-16-72-
NJ
*
12-19-72
12-21-72
2-7-73
11-1-72
Operating Information
Stage
4
4
4
4
4
4
5
Cartridge
Number
T.J.
T.J.
T.J. 4737
4735
4731
T.J. 3746
T.J. 4724
Gulf
T.J. 4721
Cartridge
Position


Inlet
Middle
Outlet
Inlet
Middle
InJ.et
Outlet
Membrane Cartridge
Characteristics
(AP), psi







Rejec-
tion
Low
Low
Low


O.K.
Low

Problems
Unknown
reasons
Brine seals
weak
O-ring
failure
Brine seal
flipped
O-ring
failure
Brine seal
weak
O-ring
leak
Comments


All three cartridges we're
O.K. when tested indivi-
dually later on. Replaced
cartridges.
(AP) rose to 50 psi before
it happened
Replaced all cartridges
by two Gulf cartridges
Brine seal of the second
Gulf cartridge was O.K.
Slot too wide.
Replaced cartridge.

-------
                        APPENDIX D
                        (continued)
Page 4
MECHANICAL PROBLEMS OF DIFFERENT SPIRAL MEMBRANE CARTRIDGES
Date
11-3-72




2-7-73

to
Ul







Operating Information
Stage
5




5










Cartridge
Number
T.J. 4718
4720
4523


Gulf
Gulf









Cartridge
Position
Inlet
Middle
Outlet


Inlet
Outlet









Membrane Cartridge
Characteristics
(AP), psi
















Rejec-
tion
Low




O.K.










Problems
Cartridge
failure*



Brine seals
had loose
tapes

Continents
Cartridges 4718 and 4720
were found to be O.K.
during individual cartridge
screening tests. Replaced
all three T.J. cartridges
by Gulf cartridges


.
*Any one of the following reasons:
1. Leakage at the glue searn;
2. Leakage at the material fold
adjacent to the permeate
collection tube;
3. Membrane and/or backing material
wrinkles causing direct leakage;
4. Pinholes in membrane.

-------
  APPENDIX E
SLIME ANALYSIS
        237

-------
                      APPENDIX E

                    SLIME ANALYSIS
A sample of the material flushed from a fouled membrane
cartridge was analyzed.  The material was de-cationized
using Amberlite IR-120 resin.  The flow sheet showing the
make-up of the sample and the fractions obtained is as
follows:

The six fractions from the cation column were examined.
The results leave some strange unanswered questions.

Fractions "A" and "C" should be the same.  When pyro-
lyzed, the chromatograms closely matched.  However, the
TGA showed degradations at 270° and 520° with a 10%
white ash for "A" and degradations at 270° and 440° with
1.3% white ash for "C".  The difference in ash content
is unexplainable; especially since neither sample should
contain any inorganic salts.

The chromatograms from the pyrolysis of these fractions
closely match the pyrolysis chromatogram for glucose.
The pyrolysis by-products can be grouped into three
groups.  This is a 600° pyrolysis.
     1.  Very volatile - retention times 0.6 to 2.2 min.

     2.  Intermediate - rentention times 6.0 to 12.3 min.
     3.  High boiling - retention times 14.0 to 30.0 min.
Groups 1 and 2 closely match the pyrolysis of glucose;
however group 2 is definintely stronger than glucose.

When pyrolized at 950°, the chromatograms match glucose
even closer.  Apparently some of this material is poly-
saccharides in nature.  It should be noted that even
though a similarity exists to a pyrolysis of cellulose,
several peaks are different.

Fraction D.  The 600° pyrolysis of "D" match as far as
groups 1 and 2 are concerned; however the high boiling
group of compounds are completely different.  A sample
of the crystals from the concentrate was dissolved in
water and acidified.  The precipitate was filtered.
The pyrolysis of this solid matches exactly fraction "D".
TGA showed degradation at 300, 500, and 650°.  Ash was
25% white.
                            238

-------
                                       Sample I 3 Membrane wash - pH 6.6
                                  Added 20 mg  NaOH AH-10.4 then added 400 ml MeOH
                                    Passed solution through cation column as 1:1 MeOH
Effluent
  Evap. to remove MeOH
  Fpt. separated - filtered
Filtration very slow - no wash

About 75% of ppt.
  Sample "A"
46% of 13
                                                                                 Cation resin column
                                                                               Backwashed with water
                      Remainder of effluent
                        air-dried
                        Black granular crystals
                        washed with water - filtered
Covered with
 acetone
  1
                           I
  Filtrate
  1
Acetone extract
Sample "B"
Filtrate
                                  Solid
                                  Sample "C"
                                  10% of 13
                                                                               cation resin column
                                                                            washed with 72.94 mg HC1
                                                                              (not enough HC1
                                               Effluent neutralized with
                                                   56 mg NaOH
                                                   Volume 190 ml
                                              	1
                                             ppt.
   Collected backwash
  milky in appearance
separated on standing
             filtered
                                                             1
                                                       40 ml Evap.  to dryness
                                                       Solids calculated as Na
                                                              1.8%  of 13
                                                                                                                  1
                                                                                                                150  ml
                                                                                                                separated
                                                                                                           PPt.
                                                                                                    Extracted with ethur
                                                 1
                                          Solid Sample
                                            11% of 13
                                                I
                                            Filtrate
                                            discarded
                                                                                          I
                                                                                     Ether extract
                                                                                             "E"
                                                                    I
                                                              Ag.  sol.
                                                      acidified '
                                                                                                      extracted with ether
                                                                                                     	  I
                                                                                         I
                                                                                     Ether extract
                                                                                      Sample "F"
                                                                                             Ag.  sol.

-------
Fraction B.  The chromatogram of this fraction was very
poor.  A group of 12 to 15 materials with an estimated
boiling range 230 to 300°.  When reacted to form the tri-
methylsilyl derivative, some differences are noted.  The
IR showed strong aliphatic and carbonyl.  The extract
also contained some hydroxyl absorption; but spectra
quality is poor.


Fraction E.  IR showed strong aliphatic/ especially CH2.
Also minor indications of hydroxyl  (extremely weak),
carbonyl (one only), and aromatic  (possibility of a
little phenylphthalein - used as an indicator).  The GC
showed several distinct peaks and a group of lesser peaks.
Slight changes occur when the fraction is derivatized.

Fraction F.  This fraction chromatographs similar to
fraction "B", except that a very large distinct peak is
found which matches one of the peaks in "E".  This
material does not appear to form any derivative.  These
materials would all be very high boiling.  The IR showed
strong aliphatic (CH? and CH-,) .  Also minor indications
of hydroxyl, carbonyl (2 bands) and aromatic.  "E" appeared
to be more contaminated than "F", although the single
carbonyl indicates that at least one component is absent
in "E" .  Other than the carbonyl difference, both "E"
and "F" present very similar spectra.  The strong ali-
phatic indicates a hydrocarbon possibility such as kero-
sene.  Disregarding the fact that such a material should
not be in "F", the GC would indicate that if hydrocar-
bons were present, it would have to be C-15 to 18.
                        140

-------
                        ~ . Subjcrt Ftvld &,
                                             SELECTED WATER RESOURCES ABSTRACTS
                                                    INPUT TRANSACTION  FORM
 5
    Organization
        Champion International Corporation
        Knightsbridge
        Hamilton, Ohio   45020
    Tifle
      COLOR REMOVAL FROM KRAFT MILL EFFLUENTS BY ULTRAFILTRATION
i Q |\ Authorfs)
      H  A. Fremont
      D.  C. Tate
      R.  L. Goldsmith (Abcor, Inc
           Cambridge,  Mass.)
                              16
                                  Project Designation
                                                 S 800 261
                                  Note
 22
    Citation
          Environmental Protection. Agency report
          number, JEPA-660/2-73-019, December 1973.
    Descriptors (starred First) #puip anl^J**  	
                                         SEND TO: WATER RESOURCES SCIENTIFIC INFOBMATI
                                                U.S DEPARTMENT OF THE INTERIOR
                                                WASHINGTON. D. C 502 40
  WR:I02
  WRS1C
       (REV. JULY
                                                                                   CEN
  U.S. GOVERNMENT PRINTING OFFICE: !»*-S46-314:179

-------