WATER POLLUTION CONTROL RESEARCH SERIES • 12010 DRH 11/71
       Ultrathin Membranes for
       Treating Metal Finishing
    Effluents by Reverse Osmosis
ENVIRONMENTAL PROTECTION AGENCY • RESEARCH AND MONITORING

-------
         W&TJE.JQLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes
the results and progress in the control and abatement
of pollution in our Nation's waters.  They provide a
central source of information on the research, develop-
ment, and demonstration activities in the Environmental
Protection Agency through inhouse research and grants and
contracts with Federal, State, and local agencies, research
institutions, and industrial organizations.

Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Head, Publications Branch
(Water), Research Information Division, R&M, Environmental
Protection Agency, Washington, B.C.  20460

-------
       ULTRATHIN MEMBRANES FOR TREATING
           METAL FINISHING EFFLUENTS
              BY REVERSE OSMOSIS
                      by

  North Star Research and Development  institute
         3100 Thirty-Eighth Avenue South
          Minneapolis, Minnesota  55406
                    th.fiou.gk

The State of Minnesota Pollution Control Agency
                        the.

        Environmental Protection Agency
     Industrial Pollution  Control Section

               Project 12010 DRH


                 November  1971

-------
                      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 recommendation for
       use.
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price $1.00

-------
                                ABSTRACT
Reverse osmosis has been examined as a process for treating metal
finishing wastewaters.  Seventeen different membranes were evaluated
for the separations of heavy metal ions, acids, bases, and cyanides
from water.  They included commercially available asymmetric membranes
(approximately 0.002 inch in thickness) and ultrathin membranes (1 x 10 6
to 2 x 10 5 inch in thickness).  Experimental results showed that re-
verse osmosis is feasible and effective in treating these effluents for
both pollution control and metal ion and water recovery for possible
recycle.  Although no one membrane was found effective for all effluents,
membranes of five different polymers showed considerable promise.

Simulated acidic nickel, iron, zinc, and copper plating bath rinses
were effectively treated by ultrathin membranes of three polymers:
cellulose acetate, cellulose methyl sulfonate 0-propyl sulfonic acid,
and 8-glucan acetate dimethylaminoethyl ether.  Water fluxes were
generally above 30 gallons per square foot (of membrane) per day  (gfd)
at metal ion rejections up to 99.9 percent.

Simulated chromic acid rinses were effectively treated by ultrathin
cellulose methyl sulfonate 0-propyl sulfonic acid.  This membrane ex-
hibited a water flux of 27 gfd with 95-percent rejection of chromium
(at pH 2.5).

The sulfonated polyphenylene oxide was the only membrane to withstand
hydrolysis for over 170 hours by a highly alkaline (pH 11.4) copper
cyanide solution.  The rejection was 98 percent for copper and 93
percent for total cyanide, with a water flux of 45 gfd.

Preliminary engineering considerations on the application of reverse
osmosis to the treatment and recycle of rinse waters from an acidic
copper  (sulfate) plating bath are included.

This report was submitted in fulfillment of Project 12010 DRH under the
partial sponsorship of the Environmental Protection Agency.
                                   111

-------
                                CONTENTS

Section                                                            Page

    I     CONCLUSIONS 	    1
   II     RECOMMENDATIONS 	    3
  III     INTRODUCTION	    5
          Technical Background	    5
             Metal Finishing Operations .... 	    5
             Sources and Present Treatment of
             Wastewaters	    6
          Research Program	    7
   IV     EVALUATION PROCEDURES 	   11
    V     RESULTS	   17
          Acid Wastewaters Containing Single Metal Salts	   17
             Wastewaters Containing Nickel, Iron,
             Copper, and Zinc Salts	   17
             Wastewaters Containing Chromic Acid	   28
          Acid Wastewaters Containing Mixed Metal  Salts 	   33
          Alkaline Copper Cyanide Bath Rinse Waters 	   33
             Alkaline Hydrolysis of Membranes  	   33
             Tests with Copper Cyanide Feed	   36
             Conclusions	   41
          Acids and Bases	   42
          Test Variables Affecting Membrane Performance 	   44
             Operating Pressure 	   44
             Operating Temperature	   46
             pH Effects	   49
          Optimization of Ultrathin Membrane Systems
          for Better Reverse Osmosis Performance	   49
             Polymer Preparation and Degree of
             Substitution 	   50
             Annealing	   52

-------
                          CONTENTS  (Continued)
Section
                                                                   Paj
             Thickness	   5-
             Choice of Membrane Support	   5<
             Conclusions	   5i
  VI      POSSIBLE APPLICATIONS OF REVERSE OSMOSIS TO
          METAL RINSE SOLUTIONS 	   5
          Treatment of Rinse Waters from a Copper Sulfate
          Plating Bath	   5
             Bath Solution	   6
             Drag-Out Rate	   6
             Rinse Water	   6
             Reverse Osmosis System  	   6
             Reverse Osmosis Retentate	   6
             Evaporation	   6
             Recovered Water	   6
             Operating Mode — Reverse Osmosis	   6
             Results	   6
          Alternative Treatment System	   6
  VII     ACKNOWLEDGEMENTS	   6
 VIII     REFERENCES	   6
          Additional Literature Surveyed	   7
   IX     APPENDICES	   7
                                   vi

-------
                              FIGURES
                                                                Page

 1      Reverse Osmosis Test System	13

 2      Reverse Osmosis Test Cell	13

 3      Schematic of a Typical Reverse Osmosis Test	14

 4      Typical Water Flux Behavior Across Membranes
        During Reverse Osmosis Treatment of Metal Ion
        Solutions	15

 5      Effects of Increasing Feedwater Concentration on
        The Water Flux of Reverse Osmosis Membranes	22

 6      Effects of Increasing Feedwater Concentration on
        the Copper Ion Concentration of the Product Water. ... 26

 7      Water Flux Behavior During Reverse Osmosis
        Treatment of Copper Cyanide Rinse Waters 	 37

 8      Effect of Operating Pressure on the Water Flux
        Behavior of Reverse Osmosis Membranes	45

 9      Effect of Operating Temperature on the Water Flux
        Behavior of Reverse Osmosis Membranes	47

10      Effect of Operating Temperature on the Water Flux
        Behavior of an Ultrathin Cellulose Acetate Membrane.  .  . 48

11      Water Flux Behavior of Two Cellulose Acetate
        Membranes of Differing Acetyl Content During Testing
        with 0.1-Percent Sodium Chloride  	 51

12      The Effect of Annealing on the Water Flux and
        Rejection on the CMSOPSA Membranes	53

13      Effect of Membrane Thickness on Water Flux for
        Cellulose Acetate	55

14      Copper Plating Line	59

15      Application of Reverse Osmosis to a Typical Copper
        Plating Line	60

-------
                                TABLES

No.                                                              Page


  1      Origin and Present Treatment of Some Metal Finishing
         Waste Effluents	8

  2      Reverse Osmosis Membranes Screened 	 12

  3      Outline of Testing Program 	 16

  4      Reverse Osmosis Performance of Promising Membranes
         on Nickel, Iron, Zinc, and Copper Solutions	18

  5      Reverse Osmosis Performance of Less-Promising
         Membranes on Nickel, Iron, Zinc, and Copper Solutions. . 20

  6      Osmotic Pressure and Effective Driving Force at
         Various Feedwater Concentrations 	 23

  7      Water Fluxes and Copper Ion Concentrations Obtained
         at Pressures of 600 and 1000 psig	27

  8      Reverse Osmosis Performance of Membranes on
         Chromic Acid Solutions	29

  9      Reverse Osmosis Performance of Three Promising
         Membranes on Chromic Acid Solutions	31

 10      Metal Ion Rejections from Reverse Osmosis Tests on
         Membranes with Feedwaters Containing Single Metal Salts
         and with Those Containing Mixed Metal Salts	34

 11      Reverse Osmosis Performance of Precompressed Reverse
         Osmosis Membranes under Alkaline Conditions	35

 12      Rejection of Copper During Reverse Osmosis Treatment
         of a Copper Cyanide Waste Solution 	 39

 13      Rejection of Cyanide During Reverse Osmosis Treatment
         of a Copper Cyanide Waste Solution 	 40

 14      Initial and Final Rejections of 0.1-Percent NaCl
         Solutions to Determine Membrane Deterioration	41

 15      Rejection of Acids and Bases by Membranes During
         Reverse Osmosis Testing	43

 16      Flux Declines for Annealed and Nonannealed Cellulose
         Methyl Sulfonate 0-Propyl Sulfonic Acid Membranes  ... 52
                                 viii

-------
                         TABLES  (Continued)


No.                                                             Page

17      Thickness and Flux Data for Cellulose Methyl
        Sulfonate 0-Propyl Sulfonic Acid 	   54

18      Flux Decline Slopes for Ultrathin Membranes on
        Millipore and Polysulfone Supports in a 0.1-Percent
        Sodium Chloride Solution 	 55

19      Recommended Standards for Metal Finishing Waste
        Concentrations 	 58
                                 ix

-------
                               SECTION I

                              CONCLUSIONS

This study has led to the following conclusions regarding the feasibility
of treating metal finishing wastewaters by reverse osmosis:

1)   The treatment of metal finishing wastewaters by reverse osmosis
     has been shown to be feasible.  Membranes exhibiting water fluxes
     over 30 gfd have shown excellent rejections of metallic, acidic,
     or basic solutes.  Although no single membrane is capable of
     adequately treating all types of wastewaters, membranes are
     available which exhibit excellent rejections and high water fluxes
     for certain types of wastewaters.

2)   The membranes that show the most promise in treating acidic
     wastewaters containing nickel, iron, zinc, and copper salts are
     ultrathin cellulose acetate (E 360-60), ultrathin 3-glucan
     acetate dimethylaminoethyl ether, and ultrathin cellulose methyl
     sulfonate 0-propyl sulfonic acid.  These membranes are capable of
     concentrating an acidic copper sulfate rinse solution from one-
     tenth of one percent to between forty and eighty percent of the
     original plating bath concentration (51,500 mg per liter Cu) while
     still maintaining good water fluxes (13 to 25 gfd) and copper ion
     rejections (>99 percent).   In addition to their excellent per-
     formance during the treatment of wastewaters containing a single
     metal salt, these membranes performed similarly for wastewaters
     containing a mixture of the above metal salts.

3)   Three membranes show promise for treating wastewaters containing
     chromic acid:  ultrathin cellulose methyl sulfonate 0-propyl
     sulfonic acid, ultrathin cellulose acetate (E 398-10),  and
     asymmetric cellulose acetate (RO-97).  Chromic acid is more easily
     rejected at pH 5.0 than at 2.5 or less.  All three of the above
     membranes exhibit chromic acid rejections of between 90 and 98
     percent with fluxes up to 25 gfd at pH 5.0.  At pH 2.5, however,
     the rejections drop to between 83 and 90 percent.  In addition,
     both cellulose acetate membranes are slowly hydrolyzed at pH 2.5.
     The cellulose methyl sulfonate 0-propyl sulfonic acid membrane,
     which has exhibited some resistance to hydrolysis in short-term
     tests, appears to be the best membrane for use at pH 2.5 or lower.

4)   An experimental sulfonated polyphenylene oxide membrane (General
     Electric JEPDM-127) is the only membrane of all those tested that
     can withstand hydrolysis by a highly alkaline (pH 11.4) copper
     cyanide solution.  This membrane is able to reject over 98 percent
     of the copper and 93 percent of the total cyanide while maintaining
     a water flux of 45 gfd.

5)   Acids and bases in solution are more difficult to reject than are
     the metal salts.  The order of rejection for three acids and .one

-------
     base is l^SO^ > HC1 > HNOa > NaOH.  Asymmetric cellulose acetate
     (RO-97) exhibits the best overall rejection of these species, but
     other membranes also exhibit similarily high rejections for the
     sulfuric acid.

6)   Ultrathin 3-glucan acetate dimethylaminoethyl ether is unique
     among the membranes tested, in that it does not reject the acidic
     and basic species, and it is the only membrane which exhibits a
     higher water flux at pH 2.5 than at pH 5.0.

7)   The program has shown that reverse osmosis performance of ultra-
     thin membranes can be improved by  (a) modifying the chemical
     composition (degree of substitution), (b) optimizing the annealing
     conditions, (c) producing as thin a membrane as practical, and
     (d) using a polysulfone support film.  Proper manipulation of
     these variables can result in higher water fluxes and lower flux
     declines coupled with the required rejections.

8)   Preliminary engineering considerations on the application of
     reverse osmosis using ultrathin cellulose acetate to the treat-
     ment and recycle of waste rinse waters from an acidic copper  (as
     sulfate) plating bath show a 99.8-percent recovery of the copper
     sulfate and a 99.9-percent recovery of the water.

-------
                               SECTION II

                            RECOMMENDATIONS

The overall purpose of this program is to examine reverse osmosis as a
process for treating metal finishing wastewaters and to develop the
process for industrial application.  The major task of this first phase
was to evaluate membranes for their effectiveness in this application.
The scope of the first phase did not include optimization or extensive
engineering consideration.  The experimental studies conducted resulted
in five potentially effective membranes for reverse osmosis treatment
of metal finishing wastewaters, and it is recommended that this program
be continued to the field-demonstration phase.

The first task that should be carried out in a second phase of this
program is optimization of the reverse osmosis properties of the five
membranes found to work effectively on selected metal finishing effluents,
These membranes are (1) asymmetric cellulose acetate (RO-97), (2) ultra-
thin cellulose acetate (E 360-60), (3) sulfonated polyphenylene oxide
(JEPDM-127), (4) ultrathin g-glucan acetate dimethylaminoethyl ether,
and (5) ultrathin cellulose methyl sulfonate 0-propyl sulfonic acid.
The following properties must be adjusted for optimum performance.

     •    The synthesis of the 3-glucan dimethylaminoethyl
          ether and the cellulose methyl sulfonate 0-propyl
          sulfonic acid must be investigated and controlled to
          produce polymers of known purity and predictable degrees
          of substitution.  Thus, the polymer structure that would
          give optimum performance would be elucidated.

     •    Procedures for casting ultrathin membranes in the most
          applicable configuration (probably tubes) must be
          optimized to obtain high water flux.  Optimization
          would include investigation of the concentration of
          polymer casting solution, casting solvents, additives
          (such as surfactants) to the water or casting solution,
          and annealing to produce, consistently, a membrane as
          thin as possible.

     •    The conditions under which all these membranes will be
          tested must be optimized with respect to temperature,
          pressure, and pH.  The type of wastewater treated will
          largely determine these conditions.

The optimization testing would include long-term  (up to 500 hours)
reverse osmosis tests with simulated or actual metal finishing waste-
waters, during which the effects of oxidation and hydrolysis on the
membranes would be investigated.

-------
A second step in this proposed phase of the program would include
engineering studies, an economic analysis of membrane production,
and preparation for a field demonstration.  These considerations would
lead directly to a third phase, which would consist of small-scale,
on-site demonstration.

-------
                              SECTION  III

                              INTRODUCTION

North Star Research and Development Institute has completed the first
year of research on the program "The Use of Reverse Osmosis for Treating
Metal Finishing Effluents".  The Minnesota Pollution Control Agency, as
grantee from the Applied Science and Technology Branch, U.S. Environmental
Protection Agency, together with six private organizations, sponsored
the work as Program No. 12010 DRH.

This program was designed to serve the needs of the metal finishing
industry through improved pollution control and possible conservation of
valuable materials.  Reverse osmosis offers an efficient and potentially
economical method to meet these needs.  The first phase of the program,
described in this report, consisted primarily of determining which mem-
branes, from commercially available and experimental membranes, were
most promising for treatment of metal  finishing wastewaters.  The most
important membrane performance criteria were high rejections of specific
components, high water fluxes through  the membranes, and low declines in
flux during operation.


                          Technical Background

The metal finishing industry has an ever-growing problem in controlling
the effects of its wastewaters.  The wastes that cause the problems in-
clude rinse waters from metal electroplating solutions and from acid
and alkaline cleaning and pickling solutions.  These rinse waters, if
discharged into the environment without treatment, can pollute our
natural resources, inhibit or destroy natural biological activities,
and adversely affect materials of construction.  Specific examples of
these detrimental effects include the  toxicity of heavy metals and
cyanides to various forms of aquatic life,-1 the deleterious effect of
copper and chromium on biological sewage treatment processes (because of
their toxicity to the microflora),2 and the corrosive effects of acids
and bases on sewer lines and metal and concrete structures. ^ »**
Metal Finishing Operations

The following metal finishing operations are those most commonly found
in the electroplating and metal finishing industry and are the source
of the major portion of the waterborne wastes:

     (1)  Cleaning.  Cleaning processes are mainly designed to
          remove soil, oil, and grease.  They are usually of three
          general types:  organic solvent cleaners, acid pickling
          cleaners, and alkaline cleaners.  Organic solvent vapor
          cleaners are used mainly for oil and grease removal.
          Acid pickling cleaners are used for removal of oxides,
          sulfides, and other undesirable surface deposits.  They

-------
          are generally solutions of sulfuric, hydrochloric, or
          nitric acid, the strength of which depends upon the
          metal to be cleaned.

          Alkaline cleaners consist of organic emulsion cleaners,
          alkali soaking cleaners, and alkali electrocleaners.
          The organic emulsion cleaners are aromatic or aliphatic
          solvents coupled with an emulisifier.  Alkali soaking
          cleaners consist of sodium hydroxide, orthophosphates,
          polyphosphates, silicates, carbonates, organic emulsifiers,
          and synthetic wetting agents.  Alkali electrocleaners
          are similar to soaking cleaners except that they can
          be used as either anodic or cathodic electrocleaners.

          These cleaning methods are not mutually exclusive and
          are often used in combination.

     (2)  Plating.  In the plating process, the actual deposition
          of the metal takes place.  The material being plated is
          the cathode in an electrolytic cell.  Plating baths can
          be acidic or alkaline, depending on the metal to be
          plated, the type of plate desired, and the surface
          metal on which the plate is to be deposited.

     (3)  Stripping.  Stripping processes are necessary for the
          removal of undesirable metallic, metallic oxide, or
          other coatings which may be on the work piece to be
          plated or otherwise coated.  The type of stripping
          process depends on the film to be removed and the
          base metal.  Most stripping baths are acidic in nature
          and usually consist of solutions of sulfuric, nitric,
          hydrochloric, or to a lesser extent, hydrofluoric acids.
          In addition to the acidic baths, alkaline baths with
          sodium sulfide, sodium cyanide, and sodium hydroxide
          may also be used.  Electrolytic stripping has been
          used increasingly in the past few years.  In this method,
          the material to be plated is the anode, usually in baths
          equivalent to those mentioned above.
Sources and Present Treatment of Wastewaters

Waterborne wastes generated in the above operations include the
following:

     (1)  Rinse waters from plating, cleaning, and other surface
          finishing operations;

     (2)  Concentrated plating and finishing baths that are inten-
          tionally or accidentally discharged;

-------
     (3)  Wastes from plant or'equipment cleanup;

     (4)  Sludges, filter cakes, etc., produced by naturally
          occurring deposition in operating baths or by intentional
          precipitation in the purification of operating baths,
          chemical rinsing circuits, etc., when flushed down
          sewers.

From the viewpoint of the smaller plater, by far the most important
of these wastes is the rinse water.  It is a constantly flowing,
production-connected stream that is generally so large in volume that
it cannot be economically impounded for treatment before disposal.
This stream is usually concentrated enough to be toxic.

The inactivation or removal of the undesirable constituents from
metal finishing wastewaters prior to their disposal is therefore
necessary to minimize their detrimental effects on the environment.
Several methods exist which accomplish this task with varying degrees
of success.  The simplest method is the neutralization of an excessively
acid or alkaline waste.  Inactivation and removal of the metal and
cyanide species can be accomplished by oxidation or reduction to a less
contaminating state, precipitation to permit removal, or ion exchange
for removal or recovery.

Table 1 lists the origin and presently used treatments of the most
common types of metal finishing wastewater impurities.  Various
problems are encountered in the use of any one of these techniques, a
few of which include large space requirements, complicated operating
procedures, high cost, and insufficient removal of the contaminating
species.  In addition, the objective of most of the conventional methods
of treatment is ultimate disposal or destruction of undesirable con-
stituents, with relatively little attention being given to recovery of
the contaminating species or the water.
                            Research Program

The purpose of the present research program is the examination and
evaluation of the reverse osmosis process for treating metal finishing
wastewaters.  Treatment of these effluents by reverse osmosis could
not only render any waste effluents harmless to the environment, but
could also permit the reuse of a highly purified product water and the
recovery of metal and other values discarded in other treatment
processes.  Reverse osmosis is also advantageous because of its
relatively low space requirements (being modular in design), its
adaptability to various processes, its ease of operation, and the
possibility of lower operating cost than conventional processes.

Reverse osmosis can be used in combination with existing methods of
treatment to increase efficiency, or it can be used alone.  In com-
bination with existing methods, it can be used to treat water from a

-------
            Table 1.  Origin and Present Treatment of Some
                      Metal Finishing Waste Effluents5
 Type of Impurity
   in Effluent
           Origin
       Treatment
      Alternatives
Acids:
   sulfuric
   hydrochloric
   nitric

   hydrofluoric
   phosphoric
   perchloric
   acetic

Alkalies:
   caustic
   sodium carbonate
   silicates
   phosphates

Metal Ions:
   precious metals
   copper
   nickel
Cyanides:
   simple cyanides
   complex  cyanides
Chromic Acid:
metal pickling
metal pickling
pickling, phosphating,
polishing
cleaning solutions,
alkaline pickling
solutions
electroplating baths,
spent pickles
hardening,
cleaning solutions
electroplating
baths
electroplating, anodizing,
polishing,  chromating
solutions
neutralization
with alkali
neutralization
precipitation
neutralization and
precipitation
neutralization
with acid
sedimentation,
precipitation with
alkali, electrolysis,
ion exchange
destruction by
alkaline oxidation,
precipitation with
iron, ion exchange

oxidation,
precipitation,
sedimentation,
ion exchange

acid  reduction and
precipitation with
alkali, ion exchange

-------
continuous destruction process for recycling back to the plant opera-
tions , or it can increase the metal ion concentration prior to an ion
exchange treatment.  If used alone on an individual plating line, it
can provide rinse water for recycling and the reclamation of metal
salts or other chemicals for reuse.

Evaluation of reverse osmosis as a procedure for treating metal finishing
wastewaters must begin with the selection and evaluation of the
membranes which perform the actual separations.  Membranes were chosen
and evaluated with respect to their ability to reject a variety of
metal salts, acids, and bases while maintaining a high water flux
(>20 gfd).  Three types of membranes were evaluated in this program:
1) asymmetric, 2) ultrathin, and 3) homogeneous.

An asymmetric membrane has a thin, dense surface layer backed by a
thick porous layer.  The dense layer (MD.25 micron, 2500 angstroms,
or 1 x 10 5 inch thick) is in contact with the feed solution and serves
as the barrier to the passage of dissolved materials.  The porous
layer (vLOO microns, 106 angstroms, or 0.004-inch thick) provides
support for the dense layer and permits the diffusion of the purified
solvent into a collecting system.  Examples of asymmetric membranes
evaluated in this program are Eastman's cellulose acetate RO-97, RO-94,
and RO-89.

Ultrathin membranes are homogeneous polymeric films ranging in thickness
from 200 A to 6000 A (^ 1 x 10~6 to ^ 2 x 10~5 inch).  -They are similar
in physical structure and have the same function of rejecting dissolved
materials as does the dense surface layer of the asymmetric membrane.
Ultrathin membranes have no underlying support structures and can
therefore be supported on materials different from those of which they
are composed.  The thickness of the ultrathin membrane is easily con-
trolled by the rate at which it is cast.  Since the water flux is in-
versely proportional to membrane thickness, the ability to control  the
thickness provides one method of controlling the water flux.  In this
program, ultrathin membranes were made from commercially available
polymers, such as Eastman's cellulose acetate butyrate (EAB-171-5),
and from polymers synthesized at North Star, such as cellulose methyl
sulfonate 0-propyl sulfonic acid.

Homogeneous membranes are differentiated from ultrathin membranes in
that they range in thickness from 0.2 mil  (5 x 104 A) to 0.5 mil
(12.5 x 10^ A).They are supported on a separate polymeric film in the
same manner as the ultrathin membranes.  An example of this type of
membrane is General Electric's sulfonated polyphenylene oxide  (JEPDM-127)

In addition to screening a variety of different membranes, it was
important to determine  (on a preliminary basis):  1) the effects of
operating variables  (pressure, temperature, and pH) on the membrane
performance; and 2) the effects chemical and physical modifications of
the membrane would have on its performance.

-------
In order to screen a maximum number of membranes, short-term  (<20
hours) tests were carried out.  Longer tests  (>100 hours) were con-
ducted only 1) when it was suspected  that the composition of  the feed-
waters might adversely affect  the performance of the membrane, or 2) to
observe the effects of an increasing  feed concentration on membrane
performance.
                                    10

-------
                               SECTION IV

                         EVALUATION PROCEDURES

Tests conducted during this report period included screening various
types of membrane materials for their effectiveness in treating a
variety of simulated metal finishing waste solutions.  Also investigated
were the effects of operating variables on membrane performance and the
optimization of membrane performance by chemical or physical modifica-
tion.

Sixteen of the twenty-one membranes listed in Table 2 have been tested
by reverse osmosis using simulated rinse waters from representative
nickel, iron, copper, zinc, and chromium plating baths.  The other five
membranes had water fluxes less than one gfd or rejections less than
ten percent during preliminary tests on sodium chloride solutions and
were not tested further.  The membranes tested included commercially
available asymmetric membranes, ultrathin membranes cast from commercial
polymers, and ultrathin membranes cast from polymers prepared at North
Star.  Appendix A gives the chemical structures of the polymers;
Appendix B explains the ultrathin membrane casting procedures; and
Appendix C gives the procedures for synthesizing the polymers prepared
at North Star.

The simulated plating bath rinse solutions used in the membrane-testing
procedures contained the most common as well as the most troublesome
metal salts, acids, and bases.  Solutions containing nickel, iron,
copper, zinc, and chromium metal ions and cyanide ions represented the
acid and alkaline electroplating rinses.  Solutions containing sulfuric
acid, hydrochloric acid, nitric acid, or sodium hydroxide represented
the acid and alkaline cleaning and stripping solution rinses.  The
procedures for preparing the feed solutions are given in Appendix D.

The reverse osmosis tests were carried out in flat cells under 600
psig pressure at 25°C.  The test system and flat cell are pictured in
Figures 1 and 2, respectively.  Appendix E describes the reverse
osmosis test system.  For most of the tests, the metal ion concentra-
tions in the feed solution were maintained at 100 mg per liter by a
recirculating system.  The membranes were precompressed with a one-
tenth-percent sodium chloride solution followed by testing with solu-
tions containing nickel, zinc, iron, copper, and chromium, in that
order, all at pH 5.0.  This was followed by a chromium solution at pH
2.5 and a mixed feed solution at pH 2.5.  (Iron was omitted because
of the precipitation of ferric oxide caused by the oxidation of the
ferrous ion by the dichromate ion.)  The analytical procedures including
metal ion concentration, membrane thickness, conductivity, pH, and
water flux are given in Appendix D.  If it was suspected that the
membranes had been damaged during the testing procedure, a final retest
with the sodium chloride solution was made for comparison with the
initial test.  A schematic representation of one such test is shown in
Figure 3.


                                    11

-------
              Table 2.  Reverse Osmosis Membranes Screened
A.   Commercial membranes
     1.  Eastman asymmetric cellulose acetate, RO-97  (now RO-96)
     2.  Eastman asymmetric cellulose acetate, RO-94  (no longer available)
     3.  Eastman asymmetric cellulose acetate, RO-89  (replaced by RO-90)
     4.  Amicon Diaflo ultrafiltration membrane, UM-10
     5.  General Electric sulfonated polyphenylene oxide, JEPDM-127

B.   Membranes prepared from commercial polymers
     1.  Ultrathin cellulose acetate, (Eastman 398-10)
     2.  Ultrathin cellulose acetate, (Eastman 383-40)
     3.  Ultrathin cellulose acetate, (Eastman 360-60)
     4.  Ultrathin polyphenylene oxide, (General Electric)*t
     5.  Ultrathin polycarbonate,  (General Electric)*
     6.  Ultrathin nylon, (Milvex 4000 — General Mills)t
     7.  Ultrathin cellulose ether, (Ethocel — Dow)*
     8.  Ultrathin cellulose acetate butyrate, (Eastman 171-15)
     9.  Cross-linked asymmetric polyvinyl acetate, (#25-2813
         National Starch and Chemical)*t

C.   Ultrathin membranes prepared from North Star polymers
     1.  Cellulose acetate methyl sulfonate
     2.  Methyl cellulose methyl sulfonate acetate
     3.  3-glucan acetate dimethylaminoethyl ether
     4.  Cellulose methyl sulfonate 0-propyl sulfonic acid
     5.  Cellulose acetate melamine formaldehyde adduct
     6.  Methyl cellulose methyl sulfonate
     7.  Gold, vacuum deposited on polysulfone
*Membranes cast from these polymers exhibited water fluxes of less than
 one gfd in preliminary tests with sodium chloride solutions.
tMembranes cast from these polymers exhibited rejections of less than
 ten percent in preliminary tests with sodium chloride solutions.
                                    12

-------
Figure 1.  Reverse Osmosis Test System
Figure  2.  Reverse Osmosis  Test  Cell
                 13

-------
                     pHS.O
                              pH2.5
                                   pH5.0
                30   40
          Time (hours)
         50    60    70   80
                         90   100    110   120
 20 hrs  ?hr,s 20hrs   ? hr,s 20 hrs   ?hr,s20hrs  ? hr? 20 hrs
0.1% NaCI  iNi
Zn
Fe|     Cu     |Cr|     Cr    |Ni 10,1% NaCI
                             Zn
                             Cu
                             Cr
          Figure 3.   Schematic  of  a Typical  Reverse Osmosis Test
 Figure 4 shows a typical water flux  curve  for  two cellulose acetate
 membranes during the testing procedure.  After precompression, the
 water flux remained relatively steady  (+1  gfd) through the entire
 test, despite changes in feed solutions.

 Variables which would affect membrane performance were also investigated
 during this program.  Membrane variables included annealing, membrane
 thickness, and chemical composition.  Operating  variables included
 temperature, pressure, and pH.

 The outline in Table 3 summarizes the  test conditions, analytical
 determinations, and reverse osmosis  tests  conducted  during the first
 year of this program.
                                   14

-------
X
IT
UJ
                                                                                          CELLULJDSE
                                                                                          ACETATE
                                                                                          (E 398-10)
  CELLULOSE
  ACETATE
  (RO-97)
   10 -
   5  -
                                              40        50
                                                 TIME (hours)
80
90
                     Figure 4.  Typical Water Flux  Behavior Across Membranes During
                               Reverse Osmosis  Treatment of Metal Ion Solutions

-------
        Table 3.   Outline  of Testing Program
  I.   Test Conditions
      A.   Pressure	600  psig
      B.   Temperature	25°C
      C.   Feed flow	1650 ml/mln
      D.   Test cell	flat plate

 II.   Analytical Procedures (Appendix D)
      A.   Metal ion concentration -- atomic  absorption
      B.   Acid and base rejection -- pH measurement
      C.   Cyanide ion concentration -- total  carbon  determination

III.   Testing of Simulated Metal Finishing Waste  Effluents
      A.   Acidic rinses
          1.   Metal ions
              a.  Ni"14" (as sulfate)
              b.  Fe++ (as ammonium sulfate)
              c.  Cu"1^" (as sulfate)
              d.  Zn++ (as sulfate)
              e.  Cr   (as chromate ion)
          2.   pH of metal ion solutions
              a.  5.0 for Ki++,  Fe++, Cu++,  and Zn**
              b.  5.0 and 2.5 for Cr+6
          3.   Determinations made on feed and product water
              a.  Water flux
              b.  Metal ion concentration
              c.  Acid concentration
      B.   Basic copper cyanide rinses
          1.   pH = ~ 11
          2.   Determinations made on feed and product water
              a.  Water flux
              b.  Copper ion concentration
              c.  Cyanide concentration
      C.   Acid- and base-containing solutions
          1.   Acids and bases
              a.  Sulfuric acid
              b.  Hydrochloric acid
              c.  Nitric acid
              d.  Sodium hydroxide
          2.   Determinations made on feed and/or  product  water
              a.  Water flux
              b.  Acid and base concentrations

 IV.   Variables affecting Membrane Performance
      A.   Temperature
          1.   25° to 60°C
          2.   Copper sulfate feed at pH 5.0
      B.   Pressure
          1.   200 to 1000 psig
          2.   Copper sulfate feed at pH 5.0
      C.   pH
          1.   2.5 and 5.0
          2.   Copper sulfate feed
      D.   Membrane thickness

  V.   Optimization of Membrane System
      A.   Chemical composition
      B.   Annealing
      C.   Thickness
      D.   Membrane support


                                16

-------
                               SECTION V

                                RESULTS

The results of the reverse osmosis tests are presented in six sections
which are differentiated with respect to the type of wastewater or the
performance variable being studied. These six sections are:

     1)   Acid wastewaters containing single metal salts

     2)   Acid wastewaters containing mixed metal salts

     3)   Alkaline copper cyanide bath rinse waters

     4)   Acids and bases

     5)   Test variables affecting membrane performance

     6)   Optimization of ultrathin membrane systems for better
          performance.

             Acid Wastewaters Containing Single Metal Salts

Simulated wastewaters from plating baths containing nickel, iron, copper,
zinc, and chromium ions were chosen for testing because they are the most
commonly used and contain some of the more toxic metal ions.  The metal
ion concentration of these wastewaters is high enough to affect ad-
versely the waterway into which it is discharged.

The evaluation of membranes for the removal of these individual species
by reverse osmosis is presented in two separate sections.  The first
section presents the results obtained during evaluation of the nickel,
iron, copper, and zinc wastewaters, and the second presents the results
from tests with wastewaters containing chromic acid.  The results are
presented in these two sections because the chromic acid solutions
behaved differently from the solutions containing the other metal ions.

Wastewaters Containing Nickel, Iron, Copper, and Zinc Salts

Short-term (< 20 hours) reverse osmosis tests were conducted on sixteen
membranes using low concentration feed solutions containing approximately
100 mg per liter of the nickel, iron (ferrous), copper, or zinc ion (see
Figure 3 for test procedure).  Of these sixteen, six membranes that had
shown either high rejections or high water fluxes were chosen for test-
ing with an acidic copper sulfate solution which was concentrated from
500 to approximately 40,000 mg per liter of copper.

Low Concentration Feeds.  Ultrathin membranes from three polymers showed
the most promise in rejecting the nickel, iron, copper, and zinc ions at
potentially high water flux.  These three polymers were cellulose acetate
                                  17

-------
   (E 360-60),  6-glucan acetate dimethylaminoethyl ether  (211-40A),  and
  cellulose methyl sulfonate 0-propyl sulfonic acid.

  The results  obtained for these membranes are presented  in  Table  4.   Of
  these four,  the  cellulose acetate membrane exhibited the highest  metal
  ion rejections  (> 98 percent) with a good water flux of 30 gfd.   The
  rejection of the g-glucan acetate dimethylaminoethyl ether was not  quite
  as high  (copper  was the least rejected at 95.1 percent), but its  water
  flux was an  exceptional 47 gfd.  Two different derivatives of the
  cellulose methyl sulfonate 0-propyl sulfonic acid were  tested.   The
  211-96B derivative had a high flux (42 gfd), but the rejections  were
  lower than those of the other membranes.  The 211-89A derivative, which
  contained a  smaller amount of 0-propyl sulfonic acid substituted on
  the cellulose unit than did the 211-96B derivative, exhibited a  reduced
  water flux of 19 gfd, but gave a high rejection for each metal species
   (above 97 percent).  It would be possible to vary the performance of the
  cellulose methyl sulfonate 0-propyl sulfonic acid membrane by varying
  the amount of the substituents on the cellulose unit.   A membrane could
  thus be synthesized to achieve a high rejection and a flux comparable to
  that of the  other two membranes.  (For a more complete  explanation  of
  the differences  in reverse osmosis properties caused by chemical changes
  in the polymer properties, see the section on optimization of ultrathin
  membrane systems, page 49.)  The water fluxes of the ultrathin membranes
  in Table 4 could be increased by casting them thinner.
      Table 4.  Reverse Osmosis Performance of Promising Membranes  on
                Nickel, Iron, Zinc, and Copper Solutions
Test Conditions:
   Temperature.
                                  25°C
   Pressure	600 psig
   Flow Rate	1650 ml/min
   Feed Metal Ion Concentration Range .... 100 + 10 mg/liter
   PH	
5.0



Cellulose Acetate
(E 360-60)*
S-Glucan Acetate Dimethyl-
aminoethyl Ether (211-40A)*
Cellulose Methyl Sulfonate
0-Propyl Sulfonic Acid
(211-96B)
Cellulose Methyl Sulfonate
0-Propyl Sulfonic Acid
(211-89A)


Thickness
1200A

900A

1050A


1100A



Water Flux (gfd)
At 20 hrs
35

•V47

66


21


At 70 hrs
30

•x-47

42


19


Metal Ion Rejection**
(percent)
Ni
>99.9

99.2

94.5


97.6


Fe
>99.9

96.6

91.2


97.7


Cu
>99.9

95.1

82.9


97.4



Zn
98.7

98.7

99.2


99.8


 *pH 2.5.
**Calculated from measured values.
                                      18

-------
The choice of membrane for the treatment of wastewater containing the
metal ions studied here would depend upon what flux was desired and
what concentration of metal ion could be tolerated in the product
water.  If high rejection was necessary, the cellulose acetate membrane
would be the choice.  If high flux was desired and the rejection could
be somewhat lower, the g-glucan acetate dimethylaminoethy1 ether or the
cellulose methyl sulfonate 0-propyl sulfonic acid could be used.

To understand the data presented in Table 4 completely, three points
must be explained.

     1)   The metal ion concentration is shown as 100 +_ 10 mg per
          liter.  This figure is given solely for the purpose of
          simplifying the table.  This number indicates the range
          of concentrations measured in the feedwaters and not the
          actual concentrations.  The rejection percentages were
          calculated from the actual measured (atomic absorption)
          values.

     2)   The water flux values are comparable because the membrane
          thicknesses were all approximately the same within the
          experimental error of + 100 A.

     3)   The pH of the feed solutions during the testing of the
          cellulose acetate (E 360-60) and 3-glucan acetate
          dimethylaminoethyl ether was 2.5 rather than the 5.0
          at which the other membranes were tested.  This is
          because they were tested early in the program when 2.5
          was the test pH.  This changed to 5.0 to fall within
          the pH range for minimum membrane hydrolysis.6

Table 5 presents the reverse osmosis performance data for the other
twelve membranes tested on the low-concentration wastewaters.  The
three points discussed above for Table 4 are also relevant for proper
interpretation of Table 5.

Several membranes in Table 5 exhibited very high metal ion rejections.
For example, the asymmetric membranes (>99.9 percent); ultrathin
cellulose acetate, E 398-10 (>98.7 percent); cellulose acetate melamine
formaldehyde adduct, 211-29B (>99.9 percent); and the methyl cellulose
methyl sulfonate acetate, 211-64A (>99.2 percent).  However, their
water fluxes were not above 20 gfd.  Two of the membranes, ultrathin
cellulose acetate (E 398-10) and ultrathin cellulose acetate melamine
formaldehyde adduct, could exhibit improved water fluxes if they were
cast thinner.  The ultrathin cellulose methyl sulfonate acetate could
also be cast thinner, but its water flux would still be expected to
be much lower than the desired 20 gfd.  The water fluxes of the asym-
metric membranes, of course, could not be improved.  The Diaflo
                                    19

-------
                  Table 5.   Reverse Osmosis  Performance  of Less-Promising Membranes
                              on Nickel,  Iron,  Zinc,  and Copper  Solutions
    Test  Conditions;

         Temperature	25°C
         Pressure	600 psig
         Flow Rate	1650 ml/min
         Feed Metal Ion Concentration Range	100 + 10  mg/llter
         pH	5.0

Asymmetric Cellulose Acetate
(RO-97)
Asymmetric Cellulose Acetate
(RO-9A)
Asymmetric Cellulose Acetate
(RO-89)
Diaflo (UM-10)
Sulfonated Polyphenylene
Oxide (JEPDM-127)
Ultrathin Cellulose Acetate
(E 398-10) t
Ultrathin Cellulose Acetate
(E 383-40)
Ultrathin Cellulose Acetate
Methyl Sulfonate (211- IOC) t
Ultrathin Cellulose Acetate
Melamine Formaldehyde
Adduct (211-29B)t
Ultrathin Methyl Cellulose
Methyl Sulfonate Acetate
(211-64A)t
Ultrathin Cellulose Acetate
Butyrate (EAB-171-15)
Ultrathin Gold on Polysulfone
Thickness
4 mils
4 mils
4 mils
—
—
1000A

1380&

1500A

1050A

1900A
650A

VLOOOA
Water Flux (gfd)
At 20 hrs
17
17
24
66
23
20

24

19

21

6
3

5
At 70 hcs
14
14
20
43
22
18

	

16

18

5
2

4.5
Metal Ion Rejection
(percent)*
Ni
>99.9
>99.9
>99.9
54.8
—
>99.9

—

99.5

>99.9

>99.9
90.1

53.0
Fe
>99.9
>99.9
>99.9
45.4
—
>99.9

__

98.5

>99.9

>99.5
94.7

50.5
Cu
>99.9
>99.9
>99.9
48.3
65.4
>99.9

99.7

97.5

>99.9

99.2
94.8

36.8
Zh**
>99.9
—
—
72.4
—
98.7

—

—

—

"
90.0

69.7
 ^Calculated from measured values.
**Zn rejection was not determined for membranes tested early in the program.
 tpH of 2.5.

-------
ultrafiltration membrane exhibited a good water flux  (42 gfd), but its
rejection was much too low, as would be expected.
High-Concentration Feed.  The importance of using membranes with high
water fluxes and high rejections became apparent during a test in which
an acidic copper sulfate feedwater  (pH 5.0) was allowed to concentrate
from 500 to approximately 40,000 mg per liter of copper.  (In all other
tests, the feed concentration was maintained at a constant value by
recirculating the product water.)  This test was conducted to simulate
conditions which would be expected during the actual reverse osmosis
treatment of a copper plating bath rinse water.

Six membranes chosen on the basis of their high flux-high rejection, or
moderate flux-high rejection properties (see Table 5), were tested with
this copper sulfate feed.  These included:

     1)   Asymmetric cellulose acetate  (RO-89)

     2)   Asymmetric cellulose acetate  (RO-97)

     3)   Ultrathin cellulose acetate  (E 398-10)

     4)   Ultrathin cellulose acetate  (E 383-40)

     5)   Ultrathin cellulose acetate  (E 360-60)

     6)   Cellulose methyl sulfonate 0-propyl sulfonic acid (211-89A)

Since water flux was found to be inversely proportional to membrane
thickness in the range 500 A to 1500 A  (page 54)» water fluxes at
1000 A-thickness were calculated for the ultrathin membranes.  This was
done to facilitate the comparison of their reverse osmosis properties.
(Rejection was not significantly affected by membrane thickness.)

Figure 5 shows these calculated water fluxes versus the copper con-
centration of the feedwater on semi-logarithmic plots for each of the
six membranes.  At a feed concentration of 2000 mg per liter copper,
the water fluxes of all six membranes began an approximately linear
decline to a concentration of 20,000 mg per liter copper.  Between
20,000 and 40,000 mg per liter copper,  the water fluxes declined more
rapidly.

The decline in water flux with increasing copper concentration in the
feedwater was attributed to two primary factors.  The first factor was
the increasing copper ion concentration of the feedwater, which caused
an increase in the osmotic pressure.  Table 6 gives the calculated
osmotic pressures at the tested copper  sulfate feed concentrations, and
the method of calculation.  As the  osmotic pressure increased, the
effective driving force  (gage pressure  minus osmotic pressure) decreased,
                                   21

-------
M    •*•
                               Cellulose Methyl Sulfonate 0-Propyl  Sulfonic  Acid  (2II-89A)
           Cellulose Acetate (E 360-60)
            Cellulose Acetate (E 383-40)
                                  Cellulose Acetate (RO-89)
            Cellulose Acetate (E 398-10)
            Cellulose Acetate (RO-97)
          1000
 2000


Figure 5.
       4000      6000   8000  10,000         20,000
   Copper Ion Concentration of Feed  Water (mg/l)

Effects of Increasing Feedwater Concentration on the
Water Flux of Reverse Osmosis Membranes
40,000

-------
                            Table 6.  Osmotic Pressure and  Effective  Driving Force
                                      at Various Feedwater  Concentrations
to
OJ
Concentration of Feedwater
(mg/liter Cu)
1,124
2,120
3,920
5,460
8,120
11,260
19,800
39,300
Calculated* Osmotic
Pressure (psi)
13
24
44
61
91
127
222
442
Effective Driving
(gage pressure -
587
576
554
539
509
473
378
158
Force at 600 psig
osmotic pressure)








       *0smotic  pressure  IT  =  14.7  cRT

       where:


            c =  molarity  (mole 1  )

            R =  0.082  (1-atm  deg"1 mole~ )
            T  =  temperature  (°K)

-------
This caused a decrease in the water flux of the membrane since water
flux varies directly with this effective driving force.  Therefore, at
a feedwater concentration of 20,000 mg per liter copper, the effective
driving force was 378 psig instead of 600 psig.  The increasing osmotic
pressure was probably the primary cause of the decline in water flux in
the linear portion  (2000 to 20,000 mg per liter copper) of the curves
in Figure 5.

The second factor which decreased the water flux was concentration
polarization.  The  following is a brief description of concentration
polarization.

     In reverse osmosis, the copper, salt, and water are brought
     to the membrane surface by convection.  The water selectively
     passes through the membrane, leaving most of the copper salt
     at the membrane surface.  Since the rate of transfer of the
     copper salt away from the membrane is diffusion controlled,
     an increase in the salt concentration at the membrane surface
     results (concentration polarization).  This polarization
     causes an increase in the osmotic pressure at the membrane
     surface, which lowers the product water flux.  Also, since
     the percentage salt rejection of the membrane is a constant
     (independent of salt concentration), the increase in the salt
     concentration  near the membrane surface increases the salt
     gradient across the membrane and, in turn, increases the salt
     concentration  in the product water.  At a salt concentration
     of 100 mg per  liter, a 90-percent-rejecting membrane would
     give product water containing ten ppm salt.  At a salt con-
     centration of  500 mg per liter, the same membrane would give
     product water  containing 50 ppm.  Thus, concentration
     polarization causes a lower water flux and an apparent lower
     salt rejection.

Concentration polarization not only can cause a decrease in flux, but
also can result in  a lower salt rejection.  The more rapid water flux
decline in Figure 5 between feedwater concentrations of 20,000 and
40,000 mg per liter copper was probably caused by concentration polariz-
ation.  This was evidenced by the fact that the high-flux membranes
exhibited a sharper drop in flux than did the low-flux membranes.
These sharper flux  decreases of the high-flux membranes were attributed
to greater concentration of polarization effects caused, in turn, by a
relatively faster buildup of copper sulfate at membrane surfaces.  It
should also be noted that as the copper concentration of the feedwater
increased, the difference in water flux between the lowest- and highest-
flux membranes steadily decreased.  In Figure 5, at 1000 mg per liter
copper, water fluxes of the membranes varied from 15 to 49 gfd—a 34
gfd difference.  At 40,000 mg per liter copper, they ranged from 9 to
15 gfd—a 6 gfd difference.
                                 24

-------
Figure 6 shows the curves of the copper concentration of the product
water versus that of the feedwater for each membrane (semi-logarithmic
plot).  The copper rejection for all six membranes was greater than
99 percent at all feed concentrations tested.  It was evident, and to
be expected, that as the feedwater concentration increased, the product
water concentration increased.

Of the three polymers giving the best reverse osmosis performance
using feedwaters at approximately 100 mg per liter of metal ions (Table
4), two were tested with the high~copper-containing feed; cellulose
acetate (E 360-60) and a cellulose methyl sulfonate 0-propyl sulfonic
acid  (211-89A).  Although complete data were not available on the per-
formance of the B-glucan acetate dimethylaminoethy1 ether at all the
copper feed concentrations investigated during this test, this membrane
exhibited a corrected water flux of 34 gfd and a 99.8-percent rejection
of copper ions when tested on a 5220 mg per liter copper (as sulfate)
feed solution.  This water flux is only slightly below that of the
cellulose acetate (E 360-60), when tested on a feed of similar copper
concentration (see Figure 5).  Assuming that the behavior of the
B-glucan acetate dimethylaminoethyl ether corresponded to that of the
other membranes tested, it would appear to be the third best performing
membrane.  Here again then, as in the tests with feeds containing lower
metal ion concentrations, the three best performing polymers were
ultrathin cellulose acetate (E 360-60), a cellulose methyl sulfonate
0-propyl sulfonic acid derivative, and 3-glucan acetate dimethylamino-
ethyl ether.

Pressure Effects.  It was stated previously that, as the copper con-
centration in the feedwater increased, the osmotic pressure increased,
thus reducing the effective driving force.  It would therefore be
expected that by increasing the pressure from 600 to 1000 psig, the
water flux and copper rejection of the membranes could be improved.

Table 7 gives the water fluxes and the copper ion concentration of the
product water at 600 psig and 1000 psig.  The feedwater concentrations
were 39,000 mg per liter copper and 43,000 mg per liter copper,
respectively.  Increasing the pressure increased the water flux and
decreased the copper ion concentration of the product water.  This
latter effect occurred because the flow of water through the membrane
is pressure-dependent, in contrast to the passage of the copper salt
through the membrane which is diffusion-dependent; i.e., the diffusion
of copper ion through the membrane remains constant while the flow of
water through the membrane increases with an increase in pressure.

This results in an effective decrease in the copper concentration of
the product water.

Conclusions and Discussion.   Ultrathin membranes from three polymers,
cellulose acetate (E 360-60), cellulose methyl sulfonate 0-propyl
sulfonic acid, and 3-glucan acetate dimethylaminoethyl ether, have
shown water fluxes and metal ion rejections  superior to all other
membranes tested by reverse osmosis on feedwaters containing metal

                                 25

-------
  60
u>
E
|40
t>
•o
2
o_

° 30
o
|


820
ex
Q.
   10
                                                            I  ' I   I  I  I  I I I      '     T

                                                                 Cellulose Acetate (E 383-40)-
                                                            Cellulose Methyl  Sulfonate
                                                            0-Propyl Sulfonic Acid
                                                            (2II-89A)	
                                                           Cellulose Acetate (RO-89)
                                                    Cellulose Acetate (E360-(
                                                                                      Cellulose Acetate
                                                                                      -(E 398-10)
              200
                          400
10,000     20,000
           600    1000       2000      4000  6000
            Copper Ion Concentration of Feed Water (mg/l)
Figure 6.   Effects of Increasing Feedwater Concentration on the
            Copper Ion Concentration of the Product Water
40,000

-------
                  Table 7.  Water Fluxes and Copper Ion Concentrations Obtained at
                            Pressures of 600 and 1000 psig
Test Conditions:

     Temperature	25 °C
     Flow Rate	1650 ml/min
Membrane
Ultrathin Cellulose Acetate
(E 398-10)
Ultrathin Cellulose Acetate
(E 383-40)
Ultrathin Cellulose Acetate
(E 360-60)
Asymmetric Cellulose Acetate
(RO-97)
Asymmetric Cellulose Acetate
(RO-89)
Ultrathin Cellulose Methyl
Sulfonate 0-Propyl
Sulfonic Acid (211-89A)
Water Flux (gfd)*
600 psig
10.8

9.2

13.1

7.3
11.1
15.0

1000 psig
18.4

14.5

20.2

13.0
16.7
25.4

Copper Concentration
of Product Water
600 psig
(39,300 mg/1
Cu44" feed)
38.0

104.6

43.6

8.4
64.2
71.8

1000 psig
(43,400 mg/1
Cu++ feed)
15.5

48.5

33.7

4.8
27.2
41.0

*Water fluxes are corrected to 1000 A for the ultrathin membranes,

-------
ions.  Also, treatment of metal finishing wastewaters containing copper
sulfate by these same membranes in a reverse osmosis system operating
at a pressure of 600 psig, limits the practical concentration of the
copper sulfate solution to between 20,000 and 40,000 mg per liter
copper.  This concentration is approximately forty to eighty percent
of the copper concentration of a general acid copper plating bath
(51,500 mg per liter copper).  By operating the system at higher
pressures, it should be possible to obtain a concentration closer to
100 percent bath concentration, while maintaining good flux and
relatively low copper ion concentrations in the product water.

The effects of concentration polarization could also be minimized by
proper design of the reverse osmosis cells (or tubes) and the use of
turbulence promoters.  Reduction of these effects would result in
1) product waters containing lower copper concentrations, and 2) in-
creased water fluxes caused by the reduction of osmotic pressure
differences at the membrane surface (caused by the buildup of salts).
Such improvements would also enable the reverse osmosis system to
concentrate the feedwaters to a greater degree than was obtained in
the above test.

Wastewaters Containing Chromic Acid

Chromic acid was observed to behave differently from the other metal
species.  Short-term tests (<20 hours) of sixteen membranes were
conducted at pH 5.0 and 2.5 (Figure 3) using low concentration feedwaters
containing approximately 100 mg per liter of chromium.  These two pH's
were chosen because pH 2.5 is the pH of a 100 mg per liter chromic acid
(as chromium) solution, and pH 5.0 is the pH at which minimum hydrolysis
of cellulose acetate membranes occurs.6

With the knowledge that chromic acid is a powerful oxidizing agent and
that the pH of 2.5 of a 100 mg per liter chromium solution would
eventually cause membrane hydrolysis, it was decided to run additional
tests of longer duration in order to observe any effects these two
phenomena would have on membrane performance.
Short-Term Testing.  Table 8 gives the short-term test results for the
seven best membranes of the sixteen tested.  These seven were chosen on
the basis of their moderate to high water fluxes and their adequate to
good chromium rejections at pH 5.0.  Examination of the rejections
indicated that chromic acid was generally more difficult to reject
than were the other metal salts.  For example, ultrathin cellulose
acetate (E 360-60) rejected only 75.0 percent of the chromium (at pH
5.0) and >98 percent of the nickel, iron, copper, and zinc ions
(Table 4).

It was also observed that chromic acid was rejected better if the
solution was at pH 5.0, rather than 2.5.  This improved rejection
could be caused by a shift in the chromate-dichromate equilibrium.
The lower acidity at pH 5.0 would cause a decrease in the concentration
                                 28

-------
                      Table 8.   Reverse Osmosis Performance of Membranes on
                                Chromic Acid Solutions
Test Conditions:
     Temperature	25° C
     Pressure	600 psig
     Flow Rate	1650 ml/min
     Chromium Concentration Range. .  .   . 100 + 10 mg/liter
Membrane
Ultrathin Cellulose Methyl Sulfonate
0-Propyl Sulfonic Acid (211-89A)
Ultrathin Cellulose Acetate Melamine
Formaldehyde Adduct (211-29B)
Ultrathin Cellulose Acetate (E 398-10)
Ultrathin Cellulose Acetate (E 360-60)
Asymmetric Cellulose Acetate (RO-97)
Asymmetric Cellulose Acetate (RO-89)
Sulfonated Polyphenylene Oxide
(JEPDM-127)
Thickness
1100 A

1050 A

1000 A
1200 A
4 mils
4 mils
0.2 mil

Water Flux (gfd)
at 70 hours
19

18

18
30
14
20
23

Chromium Rejection
(percent)*
pH 2.5
90.8

83.3

88.3
47.8
90.0
85.1
56.5

pH 5.0
96.5

—

94.1
75.0
98.3
92.2
79.5

 *Calculated from measured values.

-------
of the dichromate ion, which may be difficult to reject, and a cor-
responding increase in the concentration of the chromate ion, which
may be easier to reject.  The testing of this hypothesis, however,
was not within the scope of this program.

Of the sixteen membranes tested, only one exhibited any change in water
flux with chromium, compared to that obtained with other metal ions.
That membrane was g-glucan acetate dimethylaminoethyl ether.  During
tests on solutions containing nickel, iron, copper, and zinc salts,
this membrane (900A) exhibited a water flux of ^47 gfd at pH 2.5.
However, when the feed solution was changed to the chromic acid at
pH 2.5, the water flux dropped to 29 gfd.


Long-Term Tests.  To determine what effects longer exposure to chromic
acid would have on membrane performance, three of the seven membranes
listed in Table 8 were selected for further testing.  These three
membranes, chosen because of their high chromium rejections, were
asymmetric cellulose acetate  (RO-97), ultrathin cellulose acetate
(E 398-10), and ultrathin cellulose methyl sulfonate 0-propyl sulfonic
acid (211-89A).  The two ultrathin membranes exhibited similar water
fluxes at the same thickness.

The three membranes were exposed to chromic acid feedwaters (100 mg
per liter chromium) for a total of 170 hours:  140 hours at pH 2.5,
followed by 30 hours at pH 5.0.  Table 9 gives the results of the reverse
osmosis tests.  A decrease in the chromium rejection for all three
membranes was observed in tests made at 65 and 140 hours at pH 2.5.
A test of the membranes with a copper sulfate solution at 140 hours
(pH 2.5) showed no evident degradation of the membranes attributable
to either oxidation or hydrolysis.  Therefore, no satisfactory explana-
tion for this decline can be given at this time.  It is expected,
however, that the membranes would show definite effects of oxidation or
hydrolysis during longer exposure to chromic acid feed solutions at
pH 2.5.

An improvement in the rejection of the chromic acid with a change in pH
from 2.5 to 5.0 was again observed.  The change in pH was not only
advantageous for better chromium rejection, but it also was advantageous
because at pH 5.0 the oxidizing power of the dichromate ion is lessened
and hydrolysis of the cellulose acetate is minimal.6  A major dis-
advantage to changing the pH is the need to add a neutralizing base to
the solution to maintain the pH at 5.0.


Conclusions and Discussion.  On the basis of water flux and rejection,
three membranes appear to be best in treating chromic acid solutions.
These are ultrathin cellulose methyl sulfonate 0-propyl sulfonic acid
(211-89A), ultrathin cellulose acetate (E 398-10), and asymmetric
cellulose acetate (RO-97).  The asymmetric cellulose acetate membrane
(RO-97) and the ultrathin cellulose methyl sulfonate 0-propyl sulfonic


                                  30

-------
                 Table 9.   Reverse Osmosis Performance of Three Promising Membranes
                           on Chromic Acid Solutions
Test Conditions:

     Temperature 	
     Pressure	
     Flow Rate 	
     Chromium Concentration Range.
25°C
600 psig
1650 ml/min
95+1 mg/liter



Membrane
Asymmetric Cellulose
Acetate (RO-97)
Ultrathin Cellulose
Methyl Sulfonate
0-Propyl Sulfonic
Acid (211-89A)
Ultrathin Cellulose
Acetate (E 398-10)


Membrane
Thickness
4 mils

530 X.



1200 &


Water Flux
Through Membrane (gfd)
65 hrs
17

25



17

140 hrs
18

27



18

170 hrs
17

28



18

Percent Rejection of Cr+6
in Product Water*
pH 2.5
65 hrs
93.8

95.4



84.8

140 hrs
89.6

90.0



82.8

pH 5.0
170 hrs
97.8

97.4



90.0

 ^Calculated from measured values.

-------
acid exhibited similar rejections at pH 5.0.  However,  the latter mem-
brane exhibited the higher water flux of the two.  In addition, because
of its ultrathin nature, the water flux of the ultrathin cellulose
methyl sulfonate 0-propyl sulfonic acid membrane could further be
improved by casting it thinner.  Such an improvement is evident by
comparing the flux of this membrane in Tables 8 and 9; a 530-A-thick
membrane exhibited a water flux of 25 gfd compared to the 19 gfd exhibited
by a 1000-A-thick membrane.

Of these three membranes only one, cellulose methyl sulfonate 0-propyl
sulfonic acid (211-89A), was among the four membranes found most
effective for treating the metal salt solutions, and this membrane was
the least effective of the four  (lowest flux).  This indicates that one
membrane cannot be used for the successful treatment of a variety of
waste effluents; i,.e.3 the membrane which can most effectively treat a
chromic acid waste solution is not necessarily the one which can most
effectively treat a copper sulfate waste solution.

It is evident from the above results that there are two possible
approaches to the reverse osmosis treatment of rinses containing chromic
acid.  The first approach is to use a membrane that is resistant to acid
hydrolysis and oxidation and has a high chromic acid rejection at low
pH.  Of the membranes tested, only the cellulose methyl sulfonate
0-propyl sulfonic acid membrane  (211-89A) has been shown to be somewhat
more acid (and base) resistant to hydrolysis than cellulose acetate
membranes.  This resistance was caused by the substituted methyl
sulfonate and ether groups, which are less susceptible to acid and
base hydrolysis than are acetate groups.  However, the resistance of
this membrane to chromic acid oxidation during long-term testing at pH
2.5 or lower has not yet been determined.

The second approach to the problem is adjusting the pH of the solution
to 5.0.  At pH 5.0, membrane hydrolysis  (and possibly oxidation) would
be minimal so that any one of the three high-rejecting membranes shown
in Table 9 could be used in the treatment of chromic acid wastewaters.
Such an approach would be possible, however, only if the chromium were
to be discarded or reclaimed by another method, since the addition of
the neutralizing base would contaminate the concentrate, making it unfit
for reuse.  If neutralization to pH 5.0 were  the approach taken, waste
solutions from alkaline cleaning or stripping baths could be used in the
neutralization process.
                                   32

-------
             Acid Wastewaters Containing Mixed Metal Salts

In addition to single metal salt solutions, mixed feedwaters containing
approximately 100 mg per liter of nickel, copper, zinc, and chromium
were also subjected to reverse osmosis.  Iron was excluded from these
mixtures because the ferrous ion was quickly oxidized to the ferric
ion by the dichromate and then precipitated as iron oxide.  The pH of
these solutions was 2.5, because of the chromic acid content.

Table 10 compares the rejections obtained during tests using solutions
containing single metal salts and those using a mixture of salts.  The
five membranes chosen for comparison were those found to be the most
promising for treating the various single metal rinse waters, as
discussed in the previous section.  In almost all of these tests the
differences in rejection values were less than two percent.  This
difference was not considered significant.

Only one of these membranes exhibited water flux changes when the feed
solutions were changed from single salt to mixed salt feedwaters.  That
was 3-glucan acetate dimethylaminoethyl ether.  As indicated previously,
this membrane dropped in flux from 47 to 29 gfd when switched from a
copper solution to a chromic acid solution.  When this membrane was
then switched to a mixed feed solution, the flux returned to 47 gfd.  It
is significant that there was no chromic acid in this particular mixed
feed solution, which would indicate that this membrane was sensitive
to the presence of chromic acid.
               Alkaline Copper Cyanide Bath Rinse Haters

In addition to the acidic electroplating solutions which have just been
discussed, the metal finishing industry uses alkaline plating baths
which consist mainly of heavy metal cyanide complexes and free cyanide
salts.  Because these complex cyanides are extremely toxic to aquatic
life, they must be removed from the waste rinse waters prior to dis-
charge into the environment.
Alkaline Hydrolysis of Membranes

Before reverse osmosis tests with a metal cyanide solution could be
undertaken, at least one membrane had to be found which could withstand
the alkaline hydrolysis at the high pH's of the metal cyanide rinse
solutions.  Several membranes which had been precompressed were therefore
tested for 48 hours with a one-tenth-percent sodium chloride solution
at pH 11.  Table 11 gives the water fluxes and conductivity rejections
of four representative membranes.

It was evident from this test that the only membrane to withstand
the alkaline conditions was the sulfonated polyphenylene  oxide
(General Electric JEPDM-127).  The water flux and rejection remained


                                  33

-------
                    Table 10.  Metal  Ion Rejections from Reverse Osmosis Tests  on Membranes
                               with Feedwaters Containing  Single.Metal Salts and  with Those
                               Containing Mixed Metal Salts
       Test Conditions:
            Temperature	25° C
            Pressure	600 psig
            Flow Rate	1650 ml/min
Membrane
Ultrathin Cellulose Methyl
Sulfonate 0-Propyl
Sulfonic Acid (211-89A)
Ultrathin Cellulose Acetate
(E 360-60)
Ultrathin Cellulose Acetate
(E 398-10)
Ultrathin B-glucan Acetate
Dimethylaminoethyl
Ether (211-40A)
Asymmetric Cellulose
Acetate (RO-97)
Single Metal
Ion Feedwater
Ni
97.6
>99.9
>99.9
99.2
>99.9
Zn
99.8
98.4
98.7
98.7
>99.9
Cu
97.4
>99.9
>99.9
95.1
>99.9
Cr**
90.8
31.3
88.3
26.3
90.0
Mixed Metal*
Ion Feedwater
Ni
98.4
>99.9
>99.9
97.7
>99.9
Zn
96.8
96.6
—
~"~
—
Cu
98.3
>99.9
99.9
97.3
>99.9
Cr**
92.2
30.8
86.3
-«• —
93.8
4s
        *Some of the early tests did not  include zinc or  chromium
       **At pH 2.5.

-------
                    Table 11.  Reverse Osmosis Performance of Precompressed Reverse Osmosis
                               Membranes under Alkaline Conditions
       Test Conditions:
            Temperature  .
            Pressure.  .  ,
            Flow Rate .  ,
            Feed	
            Time of Test,
25°C
600 psig
1650 ml/min
0.1-percent NaCl at pH 11.0 to 11.5
48 hours
Membrane
Asymmetric Cellulose
Acetate (RO-97)
Ultrathin Cellulose
Acetate (E 398-10)
Ultrathin Cellulose
Methyl Sulfonate
0-Propyl Sulfonic
Acid (211-89A)
Sulfonated Polyphenylene
Oxide (JEPDM-127)
Water Flux Through
the Membrane (gfd)
Initial
15
19

18

23
After 48 hours
9
40

31

22
Percent Rejection Based
on Conductivity
Initial
94
90

96

51
After 48 hours
9
22

86

52
l/l

-------
essentially constant during the entire test  (low sodium chloride
rejection is a property of this membrane).  The only other membrane
to exhibit any resistance was a cellulose methyl sulfonate 0-propyl
sulfonic acid derivative  (211-89A), which had a large flux change, but
a relatively small decrease in rejection.  The cellulose acetate
membranes were rapidly degraded by the alkaline conditions and were
obviously not candidates  for treating high pH feedwaters.


Tests with Copper Cyanide Feed

From the above, it was decided to  evaluate polyphenylene oxide and
the sulfonated cellulose  methyl sulfonate 0-propyl sulfonic acid
(211-89A) membranes on a  simulated copper cyanide rinse water, with
asymmetric cellulose acetate (RO-97) as a control.

The testing was divided into five phases in a total test period of 95
hours.  The three membranes were precompressed for 18 hours with a one-
tenth-percent solution of sodium chloride (first phase).  They were
then tested on a feed which was approximately a 1:100 dilution of a
standard alkaline copper  plating bath containing 26 g per liter cuprous
cyanide, 44 g per liter sodium cyanide, and 3.7 g per liter sodium
hydroxide  (second phase).7  The pH of the 1:100 dilution was 10.4.
After testing the membranes for 49 hours on the 1:100 dilution at pH
10.4, the pH was changed  to 11.4 by the addition of sodium hydroxide.
The membranes were then tested for 24 hours at this higher pH to
determine the effects of  a higher  alkaline concentration on the per-
formance of the membranes  (third phase).  This higher concentration of
sodium hydroxide would have the effect of accelerating  the alkaline
hydrolysis of the membranes, giving a faster indication of their
ability to withstand degradation.  The testing using the two 1:100
diluted feeds was followed by two  shorter tests.  The first of these
(fourth phase) was conducted on a  1:10 dilution of the  plating bath
(pH 11.4) to determine the rejection at higher solute concentrations.
The final test (fifth phase) was conducted using a one-tenth-percent
sodium chloride feed solution.  Significant  flux increases coupled with
rejection decreases would indicate if the membranes had been degraded
during the testing procedure.

In Figure 7, the water fluxes of each membrane during the first three
phases of the test are plotted versus time.  All three  membranes
exhibited an initial flux decline  due to membrane compaction.  When
the feed was changed from the sodium chloride solution  to the 1:100
bath dilution  at  pH 10.4  (from first  to  second  phase),  the  sulfonated
polyphenylene oxide and the cellulose methyl sulfonate  0-propyl sul-
fonic acid exhibited an initial increase in water flux  which was
attributed to membrane swelling caused by the alkaline  conditions.

After this initial flux increase in the second phase, the sulfonated
polyphenylene oxide exhibited the  expected compressive  flux decline,
while the cellulose methyl sulfonate 0-propyl sulfonic  acid and the


                                   36

-------
OJ
         60
         50
         40
20
          10
                  0.1% NaCI
                       10
                                   Sulfonated  Polyphenylene Oxide (JEPDM-127)
                                       Ultrathin Cellulose Methyl Sulfonate
                                       0-Propyl Sulfonic Acid (2II-89A)
                                       Asymmetric Cellulose Acetate (RO-97)

                                                                _Q	
                                       1/100  Bath Concentration at pH  10.4
                                                                                         1/100 Bath Concentration at pH 11.4
                        20
30
40          50
  Time (hours)
60
70
80
90
                          Figure  7.   Water  Flux Behavior During  Reverse Osmosis Treatment
                                      of Copper Cyanide  Rinse Waters

-------
cellulose acetate both exhibited a gradual increase in water flux
during the remainder of the testing phase.  Although such an increase
in water flux might be indicative of some membrane degradation, this
evidence alone was regarded as insufficient to conclude any degradation
had occurred.  No significant changes in copper or cyanide rejection
were observed during this phase of the test.

When the pH was raised to 11.4 (third phase), however, definite flux
changes were evident.  The water flux of the cellulose methyl sulfonate
0-propyl sulfonic acid increased rapidly, and that of the asymmetric
cellulose acetate (RO-97) decreased  (probably caused by the collapse
of the porous substrate).  The sulfonated polyphenylene oxide membrane
exhibited no apparent change in water flux.

Tables 12 and 13 give the copper and cyanide rejections of all three
membranes for the first three phases of testing using copper cyanide
feed solutions.  For both the cellulose methyl sulfonate 0-propyl
sulfonic acid and the cellulose acetate (RO-97) membranes, the copper
and cyanide rejections decreased when the pH of the feed solution was
increased from 10.4 to 11.4.  For the cellulose methyl sulfonate
0-propyl sulfonic acid, the copper rejection dropped from 99.7 to 56.2
percent and the cyanide rejection dropped from 95.8 to 49.3 percent.
The cellulose acetate (RO-97) exhibited smaller increases of 98.1- to
75.9-percent rejection for cyanide.  The sulfonated polyphenylene
oxide membrane exhibited no change in the rejection of either of these
two solute species.

From the above data it appeared that the sulfonated polyphenylene
oxide was the only membrane to withstand hydrolysis by the highly
alkaline feed solutions.  This membrane would therefore be the only one
for which solute rejections at a higher feed concentration could be
quantitatively determined  (although others are given).  The test with a
feed at a 1:10 dilution of the plating bath  (pH 11.4) showed a decrease
of about three percent in the copper rejection (98.5 to 95.4 percent)
and a similar decrease of about two percent in the cyanide rejection
(93.0 to 90.9 percent).  This decrease in rejection was most probably
caused by concentration polarization.

A final test with a one-tenth-percent sodium chloride solution was made
to determine to what extent the membranes had deteriorated during the
testing procedure.  Table 14 gives the initial and the final salt re-
jections for the three membranes.

The large decrease in the salt rejections of the cellulose methyl
sulfonate 0-propyl sulfonic acid and the cellulose acetate (RO-97)
membranes indicated (and substantiated earlier observations) that these
two membranes had been extensively damaged in the pH 11.4 solution.
The sodium chloride rejection of the sulfonated polyphenylene oxide
membrane, on the other hand, increased substantially.  Such an increase
could be caused by membrane compaction and the deposition of solid
                                   38

-------
                 Table  12.   Rejection of Copper During Reverse Osmosis Treatment of a Copper
                            Cyanide Waste Solution
     Test Conditions:
          Temperature	25°C
          Pressure	600 psig
          Flow Rate	1650 ml/min
          pH	10.4 and 11.4
          Feed	1/100 bath concentration — 218 + 2ppm Cu
                                    1/10 bath concentration — 2400 + 15 ppm Cu
          Time of Tests  	 50;hours on 1/100 bath concentration
                                    23 hours on 1/10 bath concentration
Membrane
Sulfonated Polyphenylene
Oxide (JEPDM-127)
Ultrathin Cellulose Methyl
Sulfonate 0-Propyl
Sulfonic Acid (211-89A)
Asymmetric Cellulose
Acetate (RO-97)
Percent Copper Rejection
1/100 bath cone.
pH 10.4
98.7
99.7
98.1
pH 11.4
98.5
56.2
75.9
1/10 bath
cone.
pH 11.4
95.4
44.5
59.7
Copper in Product Water (mg/1)
1/100 bath cone.
pH 10.4
2.8 + 0.5
0.58 + 0.05
4.06 + 0.04
pH 11.4
3.2 + 0.5
95.6 + 1.0
52.6 + 0.5
1/10 bath
cone.
pH 11.4
110 + 1
1330 + 15
968 + 15
CO
VO

-------
                     Table 13.  Rejection  of Cyanide During  Reverse Osmosis Treatment of a
                                Copper  Cyanide* Waste Solution

         Test  Conditions:

              Temperature	25 °C
              Pressure	600 psig
              Flow Rate	1650 ml/min
              pH	10.4 and 11.4
              Feed	1/100 bath concentration — approximately 325 + 4 mg/1 of Carbon
                                         1/10 bath concentration — 2905 + 50 mg/1 of Carbon
              Time of Test	50 hours on 1/100 bath concentration
                                         23 hours on 1/10 bath concentration
Membrane
Sulfonated Polyphenylene
Oxide (JEPDM-127)
Ultrathin Methyl Sulfonate
0-Propyl Sulfonic
Acid (211-89A)
Asymmetric Cellulose
Acetate (RO-97)
Percent Rejection of Cyanide
1/100 bath cone.
pH 10.4
93.3
95.8
94.5
pH 11.4
93.0
49.3
64.8
1/10 bath
cone.
pH 11.4
90.9
33.3
48.1
Total Carbon in Product Water (mg/1)
1/100 bath cone.
pH 10.4
20.1 + 0.5
12.8 + 0.5
16.8 + 0.5
pH 11.4
24.8 + 0.5
180 + 4
125 + 4
1/10 bath
cone.
pH 11.4
266 + 4
1940 + 50
1515 + 50
•t-
o
         *Cyanide was determined  using  a Beckman Total Organic Carbon Analyzer (TOC).   Results  are  reported
          as mg/1 of carbon.   To  convert to mg/1 cyanide, multiply by 2.17.

-------
            Table 14.   Initial and Final Rejections of
                       0.1-Percent NaCl Solutions to
                       Determine Membrane Deterioration
Membrane
Sulfonated polyphenylene oxide
Ultrathin cellulose methyl
sulfonate 0-propyl sulfonic
acid (211-89A)
Asymmetric cellulose acetate
(RO-97)
NaCl Rejections (percent)
Initial
71.2
91.4
96.3
Final
84.8
19.6
0
oxidation products on the membrane.  Thus, the membrane had not been
damaged.  The water flux of the sulfonated polyphenylene oxide membrane
after 95 hours of testing was 45 gfd.
Conclusions

Three conclusions may be drawn from the above data.

     1)   The sulfonated polyphenylene oxide membrane is an excellent
          candidate for treating alkaline cyanide waste solutions.
          Under acid conditions this membrane exhibited low rejections
          toward metal ions (Tables 5 and 8).  Under alkaline condi-
          tions it exhibited:

          a)   No changes in water flux or copper and cyanide
               rejection when the pH was increased from 10.4
               to 11.4 (i.e.3 no hydrolysis);

          b)   An increase in the sodium chloride rejection
               over the entire test time;

          c)   A water flux of 45 gfd after 95 hours of
               testing; and

          d)   High rejection of the solvent species  (98.5
               percent for copper and 93 percent for cyanide).

     2)   The cellulose methyl sulfonate 0-propyl sulfonic acid
          cannot be used for the reverse osmosis treatment of alkaline
          cyanide waste solutions.  The rapid increase in water  flux
          at pH 11.4, the decreased copper and cyanide rejections at
                                   41

-------
          pH 11.4, and the decreased salt rejection indicate that the
          membrane was very susceptible to alkaline hydrolysis.

     3)   Concentration polarization is probably responsible for
          small decreases in the copper and cyanide rejections of the
          sulfonated polyphenylene oxide membrane when the feed con-
          centration is increased by a factor of ten.  This problem
          could be lessened by the introduction of turbulence promoters
          into the reverse osmosis system.


                            Acids and Bases

Most metal finishing operations use acid or alkaline cleaning solu-
tions to remove stains and surface contaminants.  These solutions most
often contain sulfuric, hydrochloric, or nitric acids or sodium
hydroxide.  These acids and the base are also found in stripping solu-
tions which are used to refinish previously plated work, to rework
damaged plating, and to remove undesirable metallic or inorganic
surface coatings.

Reverse osmosis tests were conducted to determine the general extent
to which each of the above acidic or basic species was rejected by
promising membranes for treating wastes containing metal ions.  The test
procedure included precompressing the membranes for over twenty hours
with a solution containing zinc sulfate and sulfuric acid, to determine
the effect of a metal salt on acid rejection, and then testing them on
each individual acid and base for a period of one hour.  The order of
testing was sulfuric, hydrochloric, and nitric acids followed by
sodium hydroxide.  Table 15 gives the membranes and their rejections
for each acidic or basic species.

Several conclusions can be drawn from the data in Table 15.

     1.   Sulfuric acid was rejected better from a solution containing
          a metal ion — in this case, zinc — than from a solution
          which contained only the acid.

     2.   The order of rejection of the acidic and alkaline species
          was H2S04 > HC1 > HNOa > NaOH. The only membranes that
          rejected nitric acid to any extent were asymmetric cellulose
          acetate (RO-97) and cellulose methyl sulfonate 0-propyl
          sulfonic acid (211-96B).  None rejected sodium hydroxide.

     3.   The 3-glucan acetate dimethylaminoethyl ether did not
          reject any of the acids or base.

During the testing with the acidic solutions, the water fluxes of the
seven membranes remained stable at their 20-hour values.  However, when
the sodium hydroxide solution was tested, the fluxes of the five ultra-
thin membranes increased three to eight times, and the flux of the


                                  42

-------
       Test Conditions:
            Temperature.
            Pressure .  .
            Flow Rate.  .
            Feeds.  . .  .
Table  15.   Rejection of Acids and  Bases by Membranes During
            Reverse  Osmosis Testing
	25°C
	 600 psig
	 1650 ml/min
	 H2S04 and 100 mg/1 Zn** at pH  2.4
         0.1N H2S04 — pH 1.5
         0.1N HC1   — pH 1.2
         0.1 N HN03 — pH 1.2
         0.1N NaOH     pH 11.8
Membrane
Cellulose Acetate
(E 398-10)
Cellulose Acetate
(E 360-60)
Asymmetric Cellu-
lose Acetate
(RO-97)
Cellulose Methyl
Sulfonate
0-Propyl Sul-
fonic Acid
(211-89A)
Cellulose Methyl
Sulfonate
0-Propyl
Sulfonic Acid
(211-96B)
3-glucan Acetate
Dime thy lamino-
ethyl Ether
(211-40A)
Percent Rejection Based on pH
H2S04
& Zn++
96 + 3

84+3

99+1


98 + 1




98+1




0



H2S04
84 + 3

37 + 3

94 + 1


90 + 2




92 + 2




0



HC1
37 + 17

20 + 17

75 + 5


60 + 5




75 + 5




0



HN03
0

0

37 + 17


0




20 + 17




0



NaOH
0

0

0


0




• 0




0



pH of Product Water
H2S04
& Zn++
3.8

3.2

5.5


4.2




4.1




2.4



H2S04
2.3

1.7

2.8


2.5




2.6




1.4



HC1
1.4

1.3

1.8


1.6




1.8




1.2



HN03
1.2

1.1

1.4


1.2




1.3




1.1



NaOH
11.8

11.8

11.7


11.8




11.8




11.8



CO

-------
asymmetric membrane decreased by half.  This performance would indicate
that the membranes were not degraded by any of the acidic solutions
during the testing period and substantiates earlier evidence that these
membranes are rapidly hydrolyzed by alkaline conditions.

The poor rejection performance of these membranes toward several of the
acidic species was not necessarily a detrimental aspect of their overall
performance.  Consider a situation in which a waste zinc sulfate solu-
tion containing sulfuric acid was to be treated by reverse osmosis to
recover the zinc and not the sulfuric acid.  The 6-glucan acetate
dimethylaminoethyl ether would be the ideal membrane to perform this
task, since it could retain over 98 percent of the zinc salt and would
allow passage of the acid.  In similar situations where all, part, or
none of the acid was to be recovered, a suitable membrane could be
chosen to do the job.
             Test Variables Affecting Membrane Performance

The reverse osmosis performance of a membrane is not only dependent on
the chemical and physical nature of the membrane itself, but also on
the conditions under which the testing is performed.  These conditions
include operating pressure, operating temperature, and pH of the feed-
water.
Operating Pressure

Studies of the effects of operating pressure on the water flux of a
membrane showed that one distinct advantage to be gained by increasing
the pressure was an increased water flux.  The results of these
studies are plotted in Figure 8.

The ratios of the water flux at 1000 psig to that at 200 psig were
calculated and are given below.

     1)   Asymmetric cellulose acetate  (RO-89)	3.5

     2)   Ultrathin cellulose acetate  (E 360-60)	4.5

     3)   Ultrathin cellulose methyl sulfonate 0-propyl
          sulfonic acid (211-89A)  	  4.3

     4)   Ultrathin cellulose acetate  (E 383-40)	4.8

     5)   Ultrathin 3-glucan acetate dimethylaminoethyl
          ether  (211-40A)	4.9

     6)   Ultrathin cellulose methyl sulfonate 0-propyl
          sulfonic acid (211-96B)  	  4.8
                                   44

-------
  40
                    Asymmetric Cellulose  Acetate
                               (RO-89)
                              Ultrathln
                              Cellulose Acetate
                              (E 360-60)
  30
.2
"5
   20
   10
                                           Ultrathin
                                           Cellulose Acetate
                                           -(E 383-40)
                                                         Ultrathin
                                                       Cellulose Methyl
                                                       Sulfonate 0-Propyl
                                                       Sulfonic Acid
                                                        (2II-89A)
                                                 Ultrathin
                                                 fj-Glucan  Acetate
                                                 Dimethylaminoethyl
                                                 Ether  (2II-40A)
                                     Ultrathin
                                     Cellulose Methyl  Sulfonate 0-Propyl
                                     Sulfonic Acid (2II-96B)
                                      600
"400        600       800
 Operating  Pressure  (psig)
                                                              I
200
IOOO
      Figure  8.   Effect of Operating  Pressure on  the Water Flux
                   Behavior  of Reverse  Osmosis  Membranes
                                         45

-------
For the ultrathin membranes, the water flux at 1000 psig averages 4.7
times that obtained at 200 psig.  However, the ratio of fluxes for the
asymmetric membranes was lower  (3.5).  This lower ratio could be a
result of the compaction of the porous substrate of the asymmetric
membrane.  The increase in the water flux over this pressure range was
not linear.

Previous tests with highly concentrated copper sulfate solutions (20,000
to 40,000 mg per liter, copper) indicated that to obtain adequate water
fluxes, the operating pressure should be 600 psig or greater.  These
higher pressures were needed to offset the flux-diminishing effects
of osmotic pressure while still maintaining an acceptable water flux
across the membrane.  Of course, any advantages to be gained by in-
creasing the pressure must be balanced against the disadvantages of
such an increase.  These disadvantages include higher equipment and
operating costs and water flux declines caused by membrane compression
and concentration polarization.

Increases in the pressure had little effect on the metal ion rejection
of the membrane.

Operating Temperature

Studies of the effects of operating temperature on the water flux of
six membranes showed an initial increase in water flux with increasing
temperature.8  The results of these studies are plotted in Figure 9.
A temperature increase from 25 °C to 45°C increased the water flux by
a factor of 1.4  (average).  However, after the temperature exceeded 45°C,
a drop in the water flux occurred for four of the six membranes.  This
was probably caused by an annealing of the membranes by the high
temperature.  Annealing, like air drying, reduces the water flux of a
membrane.  It is also an irreversible process.  If the temperature were
to be returned to a previously tested lower temperature, the water flux
of the membrane would be lower than that previously obtained at that
temperature.

Figure 10 shows the hysteresis effect on water flux for an ultrathin
cellulose acetate membrane  (E 360-60) as a function of temperature.  The
water flux of the membrane was 28 gfd at the initial testing temperature
of 25°C.  After the membrane was tested at 60°C, it was retested at
25°C and had a water flux of 14.5 gfd.  All four membranes whose water
flux declined after reaching 45°C behaved similarily.  Such behavior
would support the conclusion that these membranes had been annealed.

The two membranes for which the water fluxes were not affected by the
temperature 1) could have been annealed at a higher temperature prior
to testing [as is suspected in the case of the cellulose acetate, (RO-89)],
or 2) needed higher temperatures before any significant effects of
annealing became evident  (as was probably the case with g-glucan
                                   46

-------
   50
   40
X
=  30

   20
                 20


            Figure 9.
                           Asymmetric Cellulose Acetate
                                   (RO-89)-
                                                  Ultrathin Cellulose Acetate
                                                 \N<«	(E 360-60)
                                                   \
                                                    \
                                                      \
                                                       \
                                                        \
                                                 \        x
                                                  \        \
                                                             Ultrathin Cellulose
                                                              Methyl Sulfonate
                               Ultrathin Cellulose Acetate
                                     E 383-40)
                          Ultrathin p-6lucan Acetate
                          Dimethylamimoethyl Ether
                                 (2II-40A)
                                    rUltrathin Cellulose Methyl
                                    Sulfonate 0-Propyl Sulfonic Acid
                                    	-(2II-96B)
     30          40          50
       Operating Temperature  (°C)
60
70
Effect  of  Operating  Temperature on the Water Flux
Behavior of Reverse  Osmosis Membranes
                                            47

-------
   40
   30
X
iZ

  20
    0  10
20
                             I
                        1
                                            	 Increasing Temperature
                                            	Decreasing Temperature
                         I
30         40          50
  Operating Temperature (°C)
60
            T
70
           Figure 10.  Effect of Operating Temperature on the Water Flux
                       Behavior of an Ultrathin Cellulose Acetate Membrane
                       (E 360-60)
                                         48

-------
acetate dimethylaminoethyl ether).  Ultrathin 3-glucan acetate di-
methylaminoethyl ether membranes may exhibit higher flux and more
stability at higher temperatures compared to the cellulose-derivative
membranes.

Increases in the operating temperature caused no significant changes in
the metal ion rejection of the membranes.

The above data indicate that if temperatures greater than 25°C are used,
membranes (cellulosic) annealed at a temperature higher than the work-
ing temperature should be used.  Such a situation would occur if it were
not economically or physically feasible to reduce the temperature of a
warm rinse from a hot plating bath prior to reverse osmosis treatment.
pH Effects

Of all sixteen membranes tested at pH's of 2.5 and 5.0, only one
showed any change in reverse osmosis performance with a change in pH.
That membrane was g-glucan acetate dimethylaminoethyl ether.  A 900-
angstrom-thick membrane exhibited a water flux of ^ 47 gfd at pH 2.5
and a water flux of ^ 15 gfd at pH 5.0.  It was suspected that this
pH dependency was a result of the acid-accepting properties of the
dimethylamino group.
             Optimization of Ultrathin Membrane Systems for
                   Better Reverse Osmosis Performance

The polymers from which the ultrathin membranes are cast can be
chemically modified, or the resulting membranes can be physically
changed, to obtain optimum reverse osmosis performance.  The com-
mercial asymmetric membranes cannot be manipulated in this manner,
since they have already been optimized and are ready for use as-is,
when purchased.  The following tests and explanations, therefore, are
concerned only with ultrathin membranes.

The reverse osmosis performance of a membrane can be improved to a
considerable extent by modifications in its fabrication process.  The
membrane variables that can be modified for the highest water fluxes,
highest rejections, and lowest flux declines include:

     •    Polymer preparation process and degree of substitution

     •    Membrane-annealing conditions

     •    Membrane thickness

     •    Type of support film
                                   49

-------
Controlling the effects of the above variables would improve the
performance of the membrane-support composite for optimum treatment of
metal finishing wastewaters.
Polymer Preparation and Degree of Substitution

Small differences in the chemical composition of similar membranes
can result in large differences in their reverse osmosis properties.
In the Eastman cellulose acetate polymer series  (which have already
been optimized), the E 398-10 polymer, which is  39.8-percent acetate,
produces a lower flux and a higher rejection membrane than the E-360-60,
which is 36.0-percent acetate.  Figure 11 gives  the water flux curves
for these two membranes at similar thicknesses.  The water flux of
the E 360-60 membrane with the lower acetyl content  (or higher hydroxide
content—see Appendix A) had a higher flux  (<40  gfd) and lower salt
rejection (78 percent) than did the E 398-10 membrane with the higher
acetyl content (or lower hydroxide content) which had a water flux of
^28 gfd and a rejection of 93 percent.

Similar differences for these two membranes were also evident from the
reverse osmosis testing of chromic acid.  The E  360-60 membrane ex-
hibited a water flux of 30 gfd and a chromium rejection of 75 percent
(pH 5.0), and the E 398-10 membrane exhibited a water flux of 18 gfd
and a chromium rejection of 94 percent (pH 5.0).

In the previous tests two different ultrathin cellulose methyl
sulfonate 0-propyl sulfonic acid membranes were  evaluated (Table 4).
The 211-96B derivative exhibited a water flux of 42  gfd and metal ion
rejections of 83 to 99 percent, and the 211-89A  derivative exhibited a
water flux of 19 gfd and metal ion rejections of 97  to >99 percent.
These water flux and rejection differences were  caused by the dif-
ference in the amount of propyl  sulfonic acid substituted on the
cellulose unit (degree of substitution — DS).

The 211-96B formulation for the ultrathin cellulose  methyl sulfonate
0-propyl sulfonic acid membranes had a slightly  higher degree of
substitution (as indicated by reaction conditions) than did the 211-89A
formulation.  This is the basic reason that the  211-96B membrane
exhibited higher water flux and lower metal ion  rejections than did the
211-89A membrane.  By rigorously controlling the reaction conditions
during the synthesis of this membrane polymer, it should be possible
to prepare a polymer with certain desirable and  predictable properties.
The investigation and study of the reaction conditions were not within
the goals of this screening program.
                                   50

-------
  60
  50
                                                     T
                                             T
                                 T
                                  Test Conditions9
                                     Temperature-25°C
                                     Pressure—600 psig
                                     Flow Rate-l650ml/min
                                     Support- Polysulfone
                                     Feed - 0.1% NaCI
                                                                            78% Salt Rejection
•S.40
X

-------
Annealing

It is well known that heat annealing decreases the flux decline of
asymmetric cellulose acetate membranes.   It also decreases the water
flux and increases the salt rejection.  The same effects have been
observed with ultrathin membranes when annealed by heat or air drying.


Effect on Water Flux.  Figure 12 shows the effect of annealing on
the water flux of cellulose methyl sulfonate 0-propyl sulfonic acid
(211-96B).  One membrane was annealed by  air drying after being placed
on a polysulfone support; the other membrane was tested wet.  The
annealed membrane exhibited a lower water flux.

Effect on Flux Decline.  The effect of annealing (by air drying) on
the flux decline of a cellulose methyl sulfonate 0-propyl sulfonic acid
(211-96B) membrane was significant and is indicated in Table 16.  The
flux-decline slope for the annealed membrane was zero, but the non-
annealed membrane had a flux decline slope of -0.11.
            Table  16.   Flux Declines  for Annealed  (air dried)
                        and Nonannealed  Cellulose Methyl
                        Sulfonate 0-Propyl  Sulfonic Acid Membranes
Annealing
None
Air dried
Flux Decline*
-0.11
0.00
                       *Slopes  from logarithmic plots  of
                        flux  in gfd versus  time in hours
                        from  1  hour to  260  hours  of
                        testing.
Effect on  Salt Rejection.  Annealing  increased  the  salt  rejection  of
the cellulose methyl  sulfonate  0-propyl  sulfonic  acid  (211-96B)
membrane from approximately  62  percent  (for  nonannealed)  to  96 percent
(Figure 12).

It is possible to  control  annealing so  that  the three  membrane prop-
erties of  water  flux,  flux decline, and  rejection are  optimum for
treating specified types of  effluents.
                                   52

-------
    60
    50
    40
   X
   3
-; £ 30
   V.
   9>
  "5




    20
     10
                               -000-
         _o-o-
    -O-O	OOO	OOO-
                                                                -O-
                                                    I      I     I     I
                                                 Test Conditions5
                                                    Temperature-25°C
                                                    Pressure-600 psig
                                                    Flow Rate-1650mI/min
                                                    Feed-0.1% NaCI
                                                    Support-Polysulfone
                                                                                 62% Salt Rejection
                                                                                    Cellulose Methyl  Sulfonate
                                                                                    0-Propyl Sulfonic Acid
                                                                                    (211-968)1850 A
                           96% Salt Rejection
                      -o—<**> Cellulose Methyl Sulfonate
                               0-Propyl Sulfonic Acid
                               (2II-96B) 2000A (air dried)
                 20
60
                         Figure 12.
100
140        180
 Time (hours)
220
260
300
          The Effect of Annealing on the  Water Flux  and
          Rejection on the CMSOPSA Membranes

-------
Thickness

The solution-diffusion model9 forms the basis for the flux-versus-
reciprocal-thickness plot.  A conclusion from this proposed mechanism
of transport is that the water flux is equal to a constant term
multiplied by the reciprocal thickness and the net pressure driving
force.  Since the pressure was kept constant for all the tests, a plot
of flux versus reciprocal thickness should yield a straight line
through the origin with a slope proportional to the constant term.

Figure 13 illustrates the effect of thickness on the water flux of
an ultrathin cellulose acetate  (E  398-10) membrane.  The relationship
was linear with inverse thickness  over the thickness range studied
(400A to 2000 A).

Similar thickness-flux relationships were exhibited by the ultrathin
cellulose methyl sulfonate 0-propyl sulfonic acid (211-89A) during
tests with metal salt solutions.   Table 17 gives the data from these
tests.
                 Table 17.  Thickness and Flux Data for
                            Cellulose Methyl Sulfonate
                            0-Propyl Sulfonic Acid
                            (211-89A)
               Membrane Thickness
                       (A)
70-Hour Water Flux

      (gfd)
                     1100

                      530

                      380
       19

       25

       27
A plot of the flux versus  reciprocal  thickness was not  linear.  This
nonlinearity, however, was probably caused by the experimental nature
of the membrane, and further  development of membrane  formation procedures
would be expected to linearize  the thickness-flux relationship.
Choice of Membrane Support

Because ultrathin membranes are rather  fragile,  they must be mounted
on a support to be used for reverse osmosis.  Two membrane supports
have been used to date—Millipore VFW filters10  and polysulfone .films.
The microporous polysulfone films were  developed at North Star and are
presently used for the support of ultrathin  tubular membranes.11' 12
                                  54

-------
Ul
Ln
        50
        40
      TJ
      5
      X
      
-------
Membranes supported on the polysulfone exihibited lower flux declines
than membranes supported on the Millipore filters during a 260-hour
test.  Flux-decline values from the slopes of logarithmic plots of
flux versus time for the cellulose acetate (E 398-10) and cellulose
methyl sulfonate 0-propyl sulfonic acid  (211-96B) membranes on both
supports are given in Table 18.
       Table 18.  Flux Decline  Slopes  for Ultrathin Membranes on
                  Millipore and Polysulfone Supports in a 0.1-
                  Percent  Sodium  Chloride Solution
Ultrathin Membrane
Cellulose acetate (E 398-10)
Cellulose methyl sulfonate
0-propyl sulfonic acid
(211-96B)
Flux Decline*
Millipore
Support
-0.08
-0.14
Polysulfone
Support
-0.05
-0.11
     *Slopes from logarithmic plots of  flux  in  gfd versus time
      in hours from 1 hour  to 260 hours of testing.
Polysulfone was chosen  for use as  the  support material  for the above-
stated reasons as well  as for its  inertness  to  acids  and bases.13
The highly alkaline and acidic nature  of many metal finishing waste-
waters would require the use of a  support having  this property.
Conclusions

A large number of variables must be manipulated  for  any membrane
optimization.  The possibilities of attaining  the high water  fluxes and
rejections and low flux  declines that  appear achievable with  ultrathin
membranes make this  step necessary.  By  (1) rigorously controlling
the polymer preparation,  (2)  optimizing  the annealing conditions,
(3) producing as thin a  membrane as practical, and  (4) using  a poly-
sulfone support film, ultrathin membranes  should result which have
significantly better reverse  osmosis performance than they now exhibit.
                                   56

-------
                               SECTION VI

                    POSSIBLE APPLICATIONS OF REVERSE
                    OSMOSIS TO METAL RINSE SOLUTIONS
The treatment of metal finishing rinse solutions is ideally suited to
reverse osmosis technology.  These waste rinses are generally (1) toxic,
(2) moderate, but steady, in volume, and (3) contain low concentrations
of the metal salts and other bath solutes.  The present methods for
disposing of these wastewaters include lagooning, direct dumping into
waterways, and chemical treatment prior to disposal.  In many instances,
none of these methods of disposal can adequately handle the job.  The
direct discharge of the toxic wastewaters into rivers and streams can
harm aquatic life.  Discharge of slugs of highly concentrated wastes
into sewers can upset the biological treatment processes.  (Table 19
gives some recommended standards for sewer discharges and Appendix F
gives the metal concentration in the rinse waters of some common
plating baths.)  The large volume and rather long reaction times re-
quired for chemical treatment require large holding tanks for the
wastewaters.  The low concentration of the reacting species in the
rinse wastewaters also prolongs the time for effective chemical
treatment.  Even after these wastewaters have been treated, there still
remains the problems of disposal of a still-toxic sludge, and water
that contains a high concentration of nontoxic salts.

The major advantages of reverse osmosis in treating these rinse waters
are that reverse osmosis could not only render any waste effluent
harmless to the environment or reusable as purified process water,
but also it could permit the recovery of metal and other values dis-
carded in conventional treatment processes.
      Treatment of Rinse Waters from a Copper Sulfate Plating Bath

Figure 14 is a diagram of a representative copper plating line.  The
circled section on this figure encloses the copper sulfate plating bath
and the countercurrent rinses.  The following discussion considers the
application of reverse osmosis to the treatment of these countercurrent
rinse waters.  Figure 15 is a diagram of this application.  The data on
bath makeup,drag-out rate, rinse water concentration and rinse rate
were taken from Chapter 13 of Industrial Wastes^ Their Disposal and
Treatment.^-1*  The data on water flux, copper rejection, and the
membranes to be used were taken from the experimental work reported in
the previous section (Acidic Wastewaters Containing Single Metal Salts),
An explanation of each part of this system follows.
                                   57

-------
                Table 19.  Recommended Standards for Metal
                           Finishing Waste Concentrations
Sewer Discharge
Component
pH
Ni (mg/1)
Cu (mg/1)
Cr+6 (mg/1)
CN (mg/1)
Typical Limits15
6-10
2.0
0.3
0.05
0
Typical Range16
5-10
2-10
0-4
0-10
0.2
New York (1963)
--
3.0
5.0
0.5
0.2
In Final Electroplating Rinse
Component
Ni
Cu
Cr+6
CN
Reasonable Concentration
Limits17 (mg/1)
39
39
16
39
                                   58

-------
Work Piece

Zn Diecast
        Alkaline
        Cleaner
On
                    Rinse
                           Waste
                 Finished
                  Piece
                        Drying
                         Oven
                                                       Tap or Softened Water
gSane?
(Cath.)
                                                                      £
       Rinse
Acid
   5%HCL
                                                         Waste
                                                                  Waste
     Counter
     Current
                                                   I
..„,
R|nse
 Copper
Sulfate
  Bath
                                                                    Deionized  Water
                                                                                         .
                                   Rinse
                                                          l

                                                       Waste
                                                                                                  Copper
                                                                                                  Cyanide
                                                                                                  "Strike"
                                                                                                    Waste
                    Rinse
                                        Figure 14.  Copper Plating Line

-------
ON
O
                     r
                             Drag-out - 0.008 gpm
          0.008 gpm
         0.008 gpm
         I
              Bath
           50,OOOmg/ICu|
                                                                          sk.
         I
  Rinse
I20mg/l Cu
 Rinse
20mg/l Cu|
to Oven

0.001 gpm
              50,000 mg/l Cu|
                            Evaporation
                                       u  20.000mg/ICu
                                                              4.0 gpm
    R.Q
                                  0.014 gpm
      3.976gpm     20mg/l Cu
   Makeup
    Water
              3.99 gpm
              Recovered
              Water
               (99.8%)
           Plating Both
           CuS04- 27oz/gal (51,000 mg/ICu)
 Membrane-1000& Cellulose Acetate E 360-60
 Water Flux-25gfd atSOOpsig
 Copper Rejection-99.9%
 Membrane Area-230 ft2
 Rinse pH-4.5
                Figure 15.  Application of Reverse Osmosis to a Typical Copper Plating Line

-------
Bath Solution

Before a reverse osmosis system can be applied to the treatment of
plating rinse waters, the various properties of the plating bath must
be considered.  These include (1) chemical composition and concentra-
tion, (2) acidity or alkalinity, (3) temperature, and (4) rate of
evaporation.  The first three bath properties will determine the prop-
erties of the rinse solution to be treated.  The last property will
determine the concentration of the reverse osmosis treated feed needed
for return to the plating bath.  If no evaporation takes place, the
concentrated feed solution must be of the same concentration as the
bath.  If evaporation does occur, the reverse osmosis treated feed can
be of a lower concentration than the bath in order to replace water
lost during evaporation.

In Figure 15, the bath concentration is approximately 50,000 mg per
liter of copper, the solution is acidic, the temperature is 60°F, and
the rate of evaporation is negligible.  The copper lost from the bath
during the plating process is replenished by using copper anodes.
Drag-Out Rate

The drag-out rate is the amount of bath solution being carried by the
work piece (s) from the plating bath to the rinse.  In Figure 15 the
drag-out rate is 0.008 gallons per minute.  This is a function of the
rate of work-piece throughput, surface area, shape, orientation of the
work piece, and drain time.  The drag-out rate will determine the metal
salt and acid concentration of the rinse waters.
Rinse Water

The concentration of metal salts and other solutes in the rinse water
is an important factor in determining the product appearance.  Spotting
of the piece will result if the concentration of a species is higher
than a certain limit.  Such an occurrence would require  the addition
of a polishing step after the piece had been dried.

Table 19 gives a recommended limit for the copper concentration in
the rinse of 39 mg per liter.  Figure 15 shows that  the  first rinse
solution substantially exceeds this value; thus, the necessity for a
second rinse which has a value below the recommended limit.

The countercurrent rinse system used in Figure 15 not only reduces
the copper concentration in the final rinse water, but also conserves
on the amount of water needed to produce a high-quality  product.
                                   61

-------
Reverse Osmosis System

The overflow of 4.0 gallons per minute from the first rinse in Figure 15
is that which would be treated by reverse osmosis.  Figure 15 shows the
reverse osmosis system concentrating the 120 mg per liter rinse to
20,000 mg per liter (a 166.5-fold increase) and recovering 3.98 gallons
of water per minute.  To accomplish this recovery, the entire reverse
osmosis system must have the following performance parameters:

     •    Pressure — 600 psig

     •    Membrane water flux — 25 gfd

     •    Rejection of copper — 99.9 percent
                                 2
     •    Membrane area —  230  ft

     •    Membrane — ultrathin cellulose acetate (E  360-60) at
          1000 A
Reverse Osmosis  Retentate

To determine  to  what  degree  the  rinse water  should be  concentrated,  two
factors have  to  be  considered:

      1.    The effect  of  osmotic  pressure  in  reducing the water  flux;

      2.    The effect  of  concentration polarization in  reducing  the
           water  flux  and the copper rejection.

In the case in Figure 15,  the rinse water was  concentrated to only
20,000 mg  per liter of copper because higher concentrations would
reduce the flux  of  the membrane  (Factors  1 and 2), requiring more
membrane area to treat the waters.   Higher concentrations  would also
cause the  copper concentration of the recovered water  to  increase
since the  copper rejection of the membrane is  constant.
 Evaporation

 For  recycle,  the retentate must be of the same concentration as the
 plating bath.   In Figure 15 the reverse osmosis retentate is too
 low  for recycle and must be concentrated further.   This is done by
 evaporation to a concentration of 50,000 mg per liter.   The resulting
 retentate can be returned to the plating bath, and the  evaporated
 and  condensed water can be combined with the water from the reverse
 osmosis system for recycle to the final rinse solution.
                                    62

-------
Recovered Water

The concentration of the metal salts in the recovered water depends
mainly on the rejection of the reverse osmosis membrane and the
concentration of the reverse osmosis retentate.  In Figure 15 the
concentration of copper is 20 mg per liter  (or a 99.9-percent rejection
of the retentate) and the flow rate is 3.98 gpm for the reverse osmosis
unit and 0.01 gpm from the evaporation process.  This water, 3.98 gpm,
is returned to the final rinse.  Additional deionized water 0.01 gpm,
must be added to maintain material balance.
Operating Mode — Reverse Osmosis

The mode of operation used in Figure 15 was a single-pass operation
using tubular units to support the membrane.  To maintain the volumetric
flow rate and velocity in the tubes during reverse osmosis treatment,
the cross-sectional area or the number of tubes would have to be
reduced in downstream sections of the reverse osmosis system.

The minimum amount of space that would be required by the reverse
osmosis tubes alone would be 19.5 cubic feet.  This volume would contain
approximately 450 1/2-inch (ID) tubes with a total membrane surface
of 230 square feet.  The dimensions of the tubular unit would be 36  in.
by 18 in. by 51 in.  This calculated volume was based on the use of
American Standard TM 4-14 tubular modules (14 tubes per module, 51-
inches module length, 4-1/2-inch module diameter and 7.28 square feet
of membrane area).


Results

The recovery of the copper salts and water is given below.

                                              With RO        Without RO

     Makeup deionized rinse water             0.01 gpm        4.0  gpm

     Water recovered for reuse                99.8%             0%

     Copper sulfate recovered for reuse       99.9%             0%
                                   63

-------
                     Alternative Treatment Systems

Several other potential applications for reverse osmosis in treating
metal finishing wastewaters are given below.  In Systems I and II
reverse osmosis is used exclusively,  and in Systems III and IV,  a
secondary method of treatment is used in conjunction with reverse
osmosis.
                                System I
               Bath
     Destroy*—
 Rinse
                       Concentrate
R.O.
                                                          Recovered
                                                          Water
This can be operated as an open or closed system.  If the concentrate
is contaminated with foreign materials which were brought into the
plating bath by the work piece, it cannot be recycled back to the
bath (the contaminant would buildup).   It should, therefore, be
destroyed by a conventional treatment process.  If no foreign materials
harmful to the bath are present, the concentrate can be recycled.
                                   64

-------
                               System II
             Bath
            Makeup
Deionized
  Water
                                             1W*%I|



               	I                       w   w
             Bath
  Rinse
            Destroy
Concentrate

R.O.
Reuse In
Other Processes
            To Sewer
In this system, reverse osmosis  is  used  to  treat  the wastewater to
obtain a concentrate for conventional  disposal  and recovered waste for
reuse in other plant processes.   It would be necessary  to  divert the
water to other processes if the  metal  and other solutes were too high
for the final rinse,  but low enough for use in cleaning or pickling
or other rinse systems in the industrial plant.
                                  65

-------
                               System III
              Bath
                                                     r —
                       Rinse
                       Impurity
                        Removal
                        R.O.
                                                             Recovered
                                                             Water
A purification unit is added to the  system to remove any impurities
from the concentrate which might harm the  plating bath if recycled.
This can be appreciably more practical when treating the higher con-
centration achieved by reverse osmosis.
                               System IV
                Bath
               Makeup
               ,  1  f
                Bath
           Metal
          Recovery
                       Rinse
   Ion
Exchanger
                                               R.O.
                  Stripped Concentrate
Makeup
Water
Recovered
 Water
This system is used to recover the water  and precious metals in the
concentrate.  All other solutes are drained into the sewer, or, if
toxic, chemically destroyed.   One such application could be the recovery
of gold from the plating of electronic components.
                                  66

-------
                              SECTION VII

                            ACKNOWLEDGEMENTS


This research was carried out by Dr. L.T. Rozelle, Principal Investigator;
Mrs. B.R. Nelson, Research Chemist, who was responsible for the reverse
osmosis studies, analytical work, and report writing; Dr. E.M. Scattergood,
Research Engineer, who assisted in the reverse osmosis studies;
Mr. J.E. Cadotte, Research Chemist, who was responsible for the new
polymers; and Miss Paulette Johnston, Technical Assistant.  The authors
are appreciative of the advice and assistance of Messrs . Myron
Browning and Dale Bergstedt on the metal-finishing applications.  The
authors are also grateful for the cooperation of Honeywell Inc., and
Mr. Tom Zenk, in particular, for inspection and discussion of their
electroplating operation.

This program was sponsored by the Minnesota Pollution Control Agency
as grantee from the Environmental Protection Agency with help provided
by Mr. C. Risley, Project Officer, and Mr. E.L. Dulaney, Hqs. Program
Manager.  Co-sponsors for this program were American Standard, Inc.,
Aqua-Chem, Inc., Honeywell Inc., Lindsay Company, Marvel Engineering
Company, and North Star Research and Development Institute.  The support
and assistance of these organizations and individuals is hereby
acknowledged.

This report was submitted in fulfillment of Project No. 12010 DRH,
under the partial sponsorship of the Environmental Protection Agency.
                                    67

-------
                              SECTION VIII

                               REFERENCES

 1.  Pickering, Q.H., and Henderson, C., "The Acute Toxicity of Some
     Heavy Metals to Different Species of Warmwater Fishes", Air and
     Water Poll. Int. J., 10,  453 (1966).

 2.  Interaction of Heavy Metals and Biological Sewage Treatment
     Processes, Public Health Service Publication No. 999-WP-22
     (1965).

 3.  Shea, J.F., Reed, A.K., Tewksbury, T.L., and Smithson, G.R., Jr.,
     A State of the Art Review on Metal Finishing Waste Treatment,
     Federal Water Quality Administration, U.S. Department of the
     Interior, Program No. 12010 EIE 11/68, Grant No. WPRD 201-01-68
     (November 1968).

 4.  Dobb, E.H., "Metal Wastes, Contribution and Effect", Tech. Proc.
     Amer. Electroplaters Soc., 53 (1958).

 5.  Weiner, R., Waste Treatment in the Metal Finishing Industry,
     Chapter 2, Robt. Draper,  Ltd., Teddington, England  (1963).

 6.  Vos, K.D., Burris, F.O.,  and Riley, R.L., "Kinetic Study of the
     Hydrolysis of Cellulose Acetate in the pH Range of 2-10",
     Journal of Applied Polymer Science, 10, 825 (1966).

 7.  Metal Finishing Guidebook, 37th Ed., Metal and Plastics Publica-
     tions, Inc., Westwood, New Jersey  (1969).

 8.  Loeb, S., Desalination by Reverse Osmosis, Chapter 3, Merten, U.,
     Ed., M.I.T. Press, Cambridge, Massachusetts (1966).

 9.  Merten, U., Desalination by Reverse Osmosis, Chapter 3, Merten, U.,
     Ed., M.I.T. Press, Cambridge, Massachusetts (1966).

10.  "Ultracleaning of Fluids and Systems", Millipore Corporation
     (1963).

11.  Rozelle, L.T., Cadotte, J.E., Corneliussen, R.D., and Erickson,
     E.E., Development of New Reverse Osmosis Membranes for Desalina-
     tion, Office of Saline Water, Research and Development Progress
     Report No. 359, U.S. Government Printing Office, Washington,
     D.C. (October 1968).

12.  Rozelle, L.T., Cadotte, J.E., and McClure, D.J., Development of
     New Reverse Osmosis Membranes for Desalination, Office of Saline
     Water, Research and Development Progress Report No. 531,
     U.S. Government Printing Office, Washington, D.C.  (June 1970).

13.  "Chemicals and Plastics,  Physical Properties", Union Carbide (1970)

                                    69

-------
14.  Buford, M.G., and Masselli, J.W., Industrial Wastes, Their
     Disposal and Treatment3 Chapter 13, Rudolphs, N., Ed., Reinhold
     Publ. Co. (1953).

15.  Day, R.V., "Disposal of Plating Room Wastes", Plating, 46, 929
     (1959).

16.  Hendel, F.J., and Stewart, V.T., "Flow Through Treatment of Metal
     Industry Wastes", Sewage and Ind. Wastes, 25, 1323  (1953).

17.  Graham, A.K., Electroplating Eng. Handbook, 2nd Ed., Reinhold
     Publ. Co., New York  (1962).
                     Additional  Literature Surveyed

Golomb, A., "Application  of Reverse  Osmosis  to  Electroplating Waste
Treatment", Plating, 1001 (October 1960).

Hauck, A.R., and  Sourirajan,  S.,  "Performance of  Porous  Cellulose
Acetate Membranes  for  the Reverse Osmosis Treatment  of Hard  and Waste
Waters", Environmental Science and Technology,  3,  No. 12,  1269
(December 1969).

Flinn, J.E., Ed.,  Membrane Science and Technology, Plenum  Press,
New York  (1970).

Reverse Osmosis of Single Salt Solutions, Havens  Industries,  San Diego,
California  (August 1965).

Sourirajan, S., "Separation of Some  Inorganic Salts  in Aqueous  Solution
by Flow, Under Pressure,  Through Porous  Cellulose Acetate  Membranes",
I and EC Fundamentals, 3,  206 (1964).

Sourirajan, S., "Characteristics of  Porous Cellulose Acetate Membranes
for the Separation of  Some Inorganic Salts in Aqueous Solution",
J. Applied  Chem.,  14,  506 (1964).

Zievers, J.R., Grain,  R.W., and  Barclay, F.G.,  "Waste Treatment in
Metal Finishing:   U.S. and European  Practices", Plating, 55, No. 11
(1965).
                                    70

-------
                               SECTION  IX


                               APPENDICES


                                                                 Page

A.   Chemical Structures	73

B.   Casting Procedures 	   79

C.   Polymer Synthesis	81

D.   Analytical Procedures	83

E.   Reverse Osmosis System  	   87

          Figure 1.  Schematic Diagram  of  Reverse
                     Osmosis Test Apparatus  	   88
          Figure 2.  Dynamic Cell for Reverse  Osmosis  	   89

F.   Common Plating Baths 	   91

     References for Appendices	92
                                    71

-------
                        A.  Chemical Structures
Cellulose
The chain structure may be represented as
   Where G = anhydroglucose unit

Cellulose Acetate

In cellulose acetate polymers the hydrogen atoms on the hydroxyl groups
at positions 2, 3, and 6 of the anhydroglucose unit are partially or
completely (cellulose triacetate) substituted by acetate groups.
The total degree of substitution (DS) is therefore three, and the
various membranes differ with respect to the DS.1  The proportion of
acetyl or hydroxyl groups on the three commercial cellulose acetate
polymers is  given at the top of the next page.
*This is a modified Haworth formula.  The hydrogen atoms attached  to
 the ring carbon atoms are not labeled.  The pyran rings are not
 planar as shown, but are deformed into a "chair" type  of  structure.
                                  73

-------
        Polymer Designation
         Substituted Groups
          D.S,
Cellulose Acetate (E 398-10)
Cellulose Acetate (E 383-40)
Cellulose Acetate (E 360-60)
                OH
                OAc*

                OH
                OAc

                OH
                OAc
          0.6
          2.4

          0.7
          2.3

          0.9
          2.1
*Ac = -C-CH,
Other Cellulose Derivatives

     The cellulosic polymers which contain functional groups other
than ethyl acetate or hydroxide are depicted below.  The DS of each
substituted group, where ..known, is given.   The groups are substituted
on the 2, 3, and 6 positions of the anhydroglucose  unit.

     Cellulose Acetate Butyrate (EAB 171-15)2
         Cellulose
                            0
                          0-fi-CH3
OH
(0.75)
(0.25)
                          0-C-CH2-CH2-CH3      (2.0)
                                74

-------
Cellulose  Acetate Methyl Sulfonate  (211-10C)
                                     (2.1)
           /        r.
 Cellulose
           \
                                     (0.9)
Methyl Cellulose Methyl Sulfonate Acetate
 Cellulose ^- - OCH3
                    0           0

                  0-f-CH3 or 0-
-------
Cellulose Methyl Sulfonate 0-Propyl Sulfonic Acid
                    9
                 •0-S-CH3
                    6
      Cel Iulose - 0(CH2)3-S03
                    0
                  0-S-CH3
                    6
V
Methyl Cellulose Methyl Sulfonate
      Cellulose
         OCH3

         OCH3

         0-|-CH3
           6
                        76

-------
Non-Cellulose Polymers
     Structure of  g-Glucan
          CH2OH
All glucose units are  linked beta  (1+3) in the chain skeleton and
beta (l-*6) in the appended glucose units.  The chain structure may
be represented as:
     |p(H6)
     where
       G = anhydroglucose  unit
                                   77

-------
     The chemical structure of 3-glucan  acetate dimethylaminoethyl
ether is represented as:
        p-glucon
   0
0-C-CH3

   0
0-C-CH3
     Sulfonated Polyphenylene Oxide (JEPDM-127)3
                            (S03H)
                                         N
            where x is  0.3  to 0.5.
                             78

-------
                         B.  Casting Procedures


Casting Ultrathin Membranes'4

Approximately 0.5 cc of a five-percent (W/W) solution of the polymer in
cyclohexanone was placed on the sloped side of a pan partly filled with
water, as shown in the sketch below.


                  Polymer
                  Solution
                                   SISSJJJ'/f
As the solution reached the water it spread spontaneously across the
surface.  When this spontaneous spreading stopped, the membrane was
drawn by hand to obtain the desired thickness.  The thickness was con-
trolled by the rate of spreading.  The cast film remained on the
water surface for several minutes to allow the solvent to diffuse into
the water and air.  The membrane was cut to size with scissors and a
wetted support was brought up underneath the floating membrane.  The
membrane-support composite was removed from the water and placed
directly in the reverse osmosis cell.


Casting Polysulfone Supports5»6

Polysulfone resin (30 g Union Carbide Bakelite P-3500) was added slowly
to 170 g of dimethylformamide, with rapid stirring, and heated to 100°C
to give a clear solution.  The casting was performed in a room maintained
at below 20-percent relative humidity.  The polysulfone casting solution
was spread on a glass plate using a Gardner casting knife with a four-mil
clearance.  The plate was quickly immersed in water to gel the film
which then floated free from the glass plate.  The air surface (glossy
side) of the film was used to support the membrane.
                                    79

-------
                         C.   Polymer Synthesis


Cellulose Acetate Methyl Sulfonate  (211-10C)

Ten grams of Eastman cellulose acetate E 360-60 were dried at 105°C
for two hours and dissolved in 140 ml of pyridine.  After cooling the
solution to -10°C, 10 g of methane sulfonyl chloride in pyridine were
added in small portions.  The solution was stirred for 24 hours at room
temperature and precipitated into cold water.  The precipitate was
dispersed in the water in a Waring Blender, filtered, and dried at 60°C.
The product (13 g) was soluble in cyclohexanone and was estimated to
have an acetyl degree of substitution (DS) of 2.1 and a methane sulfonate
DS of 0.9.
Methyl Cellulose Methyl Sulfonate Acetate (211-69A)

Ten grams of 15 cps viscosity grade methyl cellulose were added to
200 ml of glacial acetic acid.  To this solution was added 10 g of
methyl sulfonic acid, followed by 50 ml of acetic anhydride.  The
solution was mixed for two hours.  The resulting yellow solution was
precipitated in water containing sodium acetate.  The precipitate
was dispersed in a Waring Blendor, filtered, washed with water, and
air dried.  The product (12 g) was soluble in cyclohexanone.
g-Glucan Acetate Dimethylaminoethyl Ether (211-40A)

Ten grams of g-glucan  (Pillsbury purified grade) was dispersed in 100
ml of 20-percent NaOH.  Five grams of 2-dimethylaminoethyl chloride
hydrochloride were added, and the solution was mixed for 24 hours at
room temperature.  The etherified product was precipitated in a liter of
methanol, filtered, and washed with methanol and glacial acetic acid.
This product was dispersed in 180 ml of glacial acetic acid.  Two ml of
70-percent perchloric acid were then added, followed by 50 ml of acetic
anhydride.  After a few minutes, the resulting hazy solution was
precipitated in water containing sodium acetate, filtered, washed with
water, and air dried.  The product (13. g) was soluble in cyclohexanone.


Cellulose Methyl Sulfonate 0-Propyl Sulfonic Acid  (211-89A)

Ten grams of cotton linters  (Hercules A-2000) were swelled in 100 g of
20-percent NaOH at -10°C for 24 hours.  Three grams of propane sultone
were blended with the wet fiber using a spatula.  The mixture was
stirred occassionally for six hours, then allowed to stand at room
temperature for four days.  The etherified fiber was washed with
methanol, 2-percent sulfuric acid, and glacial acetic acid.


                                   81

-------
The fiber was then dispersed in 250 ml of glacial acetic acid.  Twenty
grams of methane sulfonic acid followed by 50 ml of acetic anhydride
were added to produce a clear, thin solution within 20 minutes.  The
product was precipitated in 1.5 liters of deionized water containing
20 g of sodium acetate.  The precipitate was mixed for 1/2 hour to
decompose excess acetic anhydride, chopped in a Waring Blendor,
washed, and dried at 100°C.  The product  (17 g) was soluble in hot
cy clohexanone.
Cellulose Methyl  Sulfonate 0-Propyl  Sulfonic Acid  (211-96B)

The reaction was  performed as above  except  that:   1) 18  g of cotton
linters were reacted with 6  g of propane  sultone;  2) the amount of
methane sulfonic  acid was increased  to  25 g; and 3) the  reaction time
with the methyl sulfonic acid was  two hours rather than  twenty minutes.
Cellulose Acetate Melamine  Formaldehyde Adduct  (211-29B)

Ten grams of  cellulose  acetate  (Eastman 398-10) were  dissolved  in  100 g
of dimethylformamide.   Twenty grams  of trimethylol melamine  (American
Cyanamide Parez  607) were added,  resulting  in a clear solution.  The
solution was  heated  for 30  minutes at 130°C, evolving formaldehyde
vapors.  The  product was isolated by precipitation in water,  filtered,
washed, dried at 60°C.   It  gave a milky-appearing solution in cyclo-
hexanone.
Methyl  Cellulose Methyl  Sulfonate (211-19A)

Five  grams  of  100  cps  viscosity grade  methyl cellulose  were  dissolved  in
100 ml  of glacial  acetic acid.   Fifty  milliliters  of  acetic  anhydride
and 0.1 gram of 70-percent perchloric  acid were  added and  mixed  for
five  minutes to give a clear solution.  The  solution  was poured  into a
0.5-percent sodium acetate solution.   The precipitate (6.0 g)  of methyl
cellulose acetate  was  dispersed in a Waring  Blendor,  filtered, washed,
and air dried.

Three grams of the dried product were  stirred in 100  ml of dioxane and
20 g  of methane sulfonyl chloride were added. The product was pre-
cipitated in a 0.5-percent sodium acetate solution, filtered,  washed,
and dried.
                                 82

-------
                       D. Analytical Procedures
Metal Ion Concentration

The feed water was made up to the desired concentration of metal ion by
dissolution of the following salts in approximately 18 liters of de-
ionized water.  Higher concentrations were prepared by multiplying
these amounts by the appropriate factor.
                                Amount  (g)       Metal Cone,  (mg/1)

                                   8.06                   100

     Fe(NH4)2(S04)2  ' 6H20        12.69                   100

     CuS04 ' 5H20                  7.09                   100

     Cr03                          3.47                   100

     ZnS04 ' 7H20                  7.90                   100

     CuCN                         26.00                18,450
To accurately determine the metal ion concentration in the  feed water
and product water, analyses for the appropriate ion were  made using
atomic absorption spectrophotometry.7    Each time samples  of the
product water (20 ml) were collected, a sample of the feed  water  (20  ml)
was also-collected.  The average deviations in measuring  the  metal  ion
concentration are listed below.
     Range of Measurement  (mg/1)      Average Deviation  (mg/1)

             0 to 2                              +0.02

             0 to 10                             +0.05

             0 to 100                            +0.5

             0 to 1000                           +5.0
                                   83

-------
The analyses were made with a Varian-Techtron AA-120 Atomic Absorption
Spectrophotometer using an air-acetylene flame.  Measurements were
recorded on a Beckman ten-inch linear recorder.
Ultrathin Membrane Thickness

After the ultrathin membrane had been cast  (see Appendix B), a one-inch-
square portion adjacent to the portion used for the reverse osmosis
studies was removed and placed on a black opaque glass plate and air
dried.  The thickness was then measured by interferometry techniques
using a Reichert Metallograph equipped with a Normaskii polarization
interferometer. **
Salt Rejection by Conductivity

The rejections of sodium chloride  feed  solutions were determined by
conductivity measurements.  The  instrument  used was a Leeds and Northrup
Co. multipurpose electrolytic conductance/resistance conductivity
bridge equipped with an Industrial Instruments  (Beckman) conductivity
cell (G 10Y116).
Acid-Base Rejection by  pH Measurements

The rejection of  acidic or basic  species was  calculated  from pH
measurements according  to the  following equations:
                                      antilog  pHf
              % Rejection (H )  -   1 - antilog  pH
                                                "P


                                       antilog (14  - pHf)
              % Rejection (OlT) =  1 - antilog (14  - pHp)



where:

              f = feed water

              p = product water

These measurements  were made on  a Beckman pH  meter, Model 76.
                                  84

-------
Water Flux

The rate of water passage through the membrane was measured using a
graduated pipette which was placed at the product water outlet in the
reverse osmosis cell.  The rate of flow in ml per minute was determined
and multiplied by the factor 21.2 gal-day/ml-ft2 to obtain the
membrane water flux in gal/ft2 • day.
                                   85

-------
                       E.  Reverse Osmosis System

The reverse osmosis test loop pictured in Figure E-l contained a 20-
liter reservoir, a Model 241-144B Milton Roy Pump, an accumulator
(surge tank), a heat exchanger, six stainless steel test cells, a 100-
mesh, high-pressure filter, a needle valve for system pressure control,
and a Rotameter-type flow meter.

The design of the test cell is shown in Figure E-2.  The cell was
machined from AISI 316 stainless steel.  Mechanical support for the
membrane was supplied by a two-inch-diameter sintered stainless steel
plate, 1/4-inch thick.  The porous support film was placed between the
porous plate and the membrane to protect the membrane from the rough
surface of the plate.  A seal was obtained with a 1.75-inch (I.D.) "0"
ring.  The two end plates were held together by four nuts.
                                   87

-------
                                                     N2 Reg
                                                                         Pressure Gage
oo
CO

fT ^V
Supply
Tank
(ZOIiters)
T Sll
Supply Drain T«
C


liaphragm
Pump
i i i
Tstarter 1
:*<
1 I 1 IA
> oonWAf* -, 1

vy- — i_j — HI
— ®" /J
	 it
\Surge
/Tank |
/Flush gi
- / M,-i Tnnl/ ^
rge ^ N2Tank
® Drain
otameter 100 1\
	 1 ^ Slruif
1 /C>- | 	
I 09 tr—
I Pressure
i I Control I i
\ Valve V
®
UN Straii
O 	 1 Flush N
eU
(( ))(°-|00°P$i)
F^"
JR J \Snubber R
Sl*f / Flush ^
— ® 	 j
Drain
Test Cell
rtesh ' 	 	
4r

ner
/alve r~"
Treated
Water « —
Outlet p
1 	
efrlgtrated
iVat«r Bath
WL
Unit (6)
	 1
3 !
3 i
® !
™™™"j
Test |
Cell 1
•j i
1
	 1
                          Figure 1.  Schematic Diagram of Reverse Osmosis Test Apparatus

-------
oo
vo
                                      FEED SOLUTION
                                     -INLET
                                                            FEED SOLUTION
                                                            OUTLET
                                                               FEED SOLUTION
                                                               CHAMBER
     ULTRATHIN
     MEMBRANE

POROUS SUPPORT
           FILM
       PRODUCT
       WATER OUTLET
                                     k\\\\\\\\\\\\\\\\
                                                            1.75 IN. EFFECTIVE  DIA.


                                                                    KRAFT PAPER
                                                                        POROUS PLATE
                           Figure 2,  Dynamic Cell for Reverse Osmosis

-------
                           F.  Common Plating Baths
Bath Formula and
Plating Conditions (8)
Nickel
44 oz/gal nickel sulfate (hexahydrate)
6 oz/gal nickel chloride (hexahydrate)
5 oz/gal boric acid
pH 1.5 to 4.5 2
current density 25 to 100 amp/ft
temperature 115 to 140° F
Copper (acidic)
27 oz/gal copper sulfate (.pentahydrate)
6.5 oz/gal sulfuric acid
current density 20 to 50 amp/ft2
temperature 60 to 120° F
Copper (alkaline)
3.5 oz/gal copper cyanide
5.9 oz/gal sodium cyanide
0.5 oz/gal sodium hydroxide
pH 12 2
current density 20 to 50 amp/ft
Zinc
24.0 oz/gal zinc sulfate (heptahydrate)
6.0 oz/gal sodium acetate
pH 3.0 to 4.5 2
current density 100 to 300 amp/ft
temperature 80 to 120° F
Chromium
33.0 oz/gal chromic acid
0.33 oz/gal sulfuric acid
Iron
48 oz/gal ferrous ammonium sulfate
(hexahydrate)
pH 2.8 to 5.0 2
current density 20 amp/ft
temperature 75 °F
Metallic and
Cyanide
Concentration
(mg/D

75,000 Ni






51,500 Cu




19,000 Cu
31,000 CN




41,000 Zn





130,000 Cr


50,500 Fe




Rinse
Concentration
(mg/1)*

156 Ni






108 Cu




40 Cu
65 CN




85 Zn





270 Cr


106 Fe




*Based on 0.5 gph drag-out and 4 gpm rinse rate.
                                       91

-------
                       References for Appendices

1.   "Cellulose Acetate", Technical Brochure, Eastman Chemical Products,
     Inc. (1959).

2.   "Cellulose Acetate Butyrate", Technical Brochure, Eastman Chemical
     Products, Inc.  (1962).

3.   Plummer, C.W., Kimura, G., and LaConti, A.B., Development of
     Sulfonated Polyphenylene Oxide Membranes for Reverse Osmosis,
     Office of Saline Water, Research and Development Progress Report
     No. 551, U.S. Government Printing Office, Washington, B.C.
     (January 1970).

4.   Francis, P.S., Fabrication and Evaluation of New Ultrathin Reverse
     Osmosis Membranes3 Office of Saline Water, Research and Development
     Progress Report No. 177, U.S. Government Printing Office,
     Washington, D.C.  (February 1966).

5.   Rozelle, L.T., Cadotte, J.E., Corneliussen, R.D., and Erickson,
     E.E., Development of New Reverse Osmosis Membranes for Desalination,
     Office of Saline Water, Research and Development Progress Report
     No. 359, U.S. Government Printing Office, Washington, D.C.
     (October 1968).

6.   Rozelle, L.T., Cadotte, J.E., and McClure, D.J., Development of
     New Reverse Osmosis Membranes for Desalination, Office of Saline
     Water, Research and Development Progress Report No. 531  (June 1970).

7.   Slavin, W., Atomic Absorption Speotroscopyt Interscience
     Publishers, New York  (1968).

8.   Metal Finishing Guidebook, 37th Ed., Metals and Plastics Publica-
     rions, Inc., Westwood, New Jersey  (1969).
                                     92

-------
    Accession Number
 w
                            Subject Field & Group
                                05D
SELECTED WATER  RESOURCES  ABSTRACTS
       INPUT TRANSACTION FORM
    Organization
      Minnesota Pollution Control
      717 Delaware  St.,  S.E.
      Minneapolis.  Minnesota   55440
    Title
      Ultrathin Membranes for Treating Metal Finishing Effluents by Reverse Osmosis
10

Authors)
Dr. Lee T. Rozelle
North Star Research &
Development Institute
3100 Thirty-Eighth Avenue Soutl
Minneapolis, Minnesota 55406
16

21

Project Designation
12010 DRH
Note
22
    Citation
23
     Descriptors (Starred First)
      •^Reverse Osmosis
25
     Identifiers (Starred First)

      treatment of metal finishing wastewaters, ^Ultrathin polymer membranes,
      ^Laboratory study, Metal salts, Electroplating wastes,  Cleaning  and Pickling
        wastes                                                                	
    Abstract
    Reverse osmosis has been examined as a process for treating metal finishing
 wastewaters.  Seventeen  different membranes were evaluated for the separations of
 heavy metal ions, acids, bases,  and cyanides from water.  They included commercially
 available asymmetric membranes (approximately 0.002 inch in thickness) and ultrathin
 membranes (1 x 10   to 2 x 10~5  inch in thickness).  Experimental results showed that
 reverse osmosis is feasible and  effective in treating these effluents for both pollution
 control and metal ion and water  recovery for possible recycle.  Although no one membrane
 was found effective for  all effluents, membranes of five different polymers showed
 considerable promise.
   Simulated acidic nickel, iron, zinc, and copper plating bath rinses were effectively
 treated by ultrathin membranes of three polymers:  cellulose acetate, cellulose methyl
 sulfonate Q-propyl sulfonic acid, and B-glucan acetate dimethylaminoethyl ether.
 Simulated chromic acid rinses  were effectively treated by ultrathin cellulose methyl
 sulfonate 0-propyl sulfonic acid.  The sulfonated polyphenylene oxide was the only
 membrane to withstand hydrolysis for over 1?0 hours by a highly alkaline (pH 11-4)
 copper cyanide solution.   Preliminary engineering considerations on the application of
 reverse osmosis to the treatment and recycle of rinse waters from an acidic copper
 (sulfate) plating bath are included.
   This report was submitted in fulfillment of Project 12010 DRH under the partial sponsor-
  ihip of the Environmental Protection Agency
  5tracto^ ,	.  .  	        '"St'Jyt'O" _ JT	
Abstra
     ' °Robert A. Wasson
                                  ice of Research  and Monitoring, EPA
 WR:I02 (REV. JULY 1969)
 WRSIC
                             SEND WITH COPY OF DOCUMENT. TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
                                                      U.S. DEPARTMENT OF THE INTERIOR
                                                      WASHINGTON, D. C. 20240
ft U. S. GOVERNMENT PRINTING OFFICE : 1972— 484-482/17
                                                                               « GPO: 1970-389-930

-------