WATER POLLUTION CONTROL RESEARCH SERIES •
17O2OEFA1O/7O
NEW AND ULTRATHIN MEMBRANES
FOR
MUNICIPAL WASTEWATER TREATMENT
BY
REVERSE OSMOSIS
ENVIRONMENTAL PROTECTION A6ENCY • WATER QUALITY OFFICE
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes
the results and progress in the control and abatement
of pollution in our Nation's waters. They provide a
central source of information on the research, develop-
ment, and demonstration activities in the Water Quality
Office, 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, Project Reports
System, Office of Research and Development, Water Quality
Office, Environmental Protection Agency, Room 1108,
Washington, D.C. 20242.
-------
NEW AND ULTRATHIN MEMBRANES FOR MUNICIPAL
WASTEWATER TREATMENT BY REVERSE OSMOSIS
by
North Star Research and Development Institute
Minneapolis, Minnesota 55406
for the
WATER QUALITY OFFICE
ENVIRONMENTAL PROTECTION AGENCY
Project #17020 EFA
Contract #14-12-587
October 1970
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price 70 cents
Stock Number 6601-0073
-------
EPA Review Notice
This report has been reviewed by the, Water
Quality Office, EPA, and approved for
publication. Approval does not signify that the
contents necessarily reflect the views and
policies of the Environmental Protection Agency,
nor does mention of trade names or commercial
products constitute endorsement or recommenda-
tion for use.
11
-------
ABSTRACT
A series of new and ultrathin membranes with thicknesses from 250 to 5000
angstroms (1 x 10~6 to 2 x 10~5 inch) and consisting of various polysaccharide
mixed esters and ethers were tested on microporous supporting films for
improved reverse osmosis treatment of municipal wastewaters. From the
screening studies with secondary effluent, ultrathin membranes prepared
from two polymers (out of a total of 44) looked very promising: cellulose
methyl sulfonate O-propyl sulfonic acid (CMSOPSA) and cellulose acetate
O-propyl sulfonic acid (CAOPSA). The CAOPSA membrane at a thickness
of 370 angstroms exhibited a water flux of 43 gallons per square foot per
day (gfd) and rejections of 96 percent for total dissolved solids, 93 percent
for nitrates, 96 percent for ammonia, 89 percent for total dissolved organic
carbon, and 99 percent for phosphate. The CMSOPSA membrane was found
to be significantly less degradable by hydrolysis, and thus more stable than
cellulose acetate.
Long-term (150 hours) testing of the CMSOPSA membrane with secondary
effluent resulted in an average water flux of 34 to 36 gfd over the last 100
hours, and rejections of 96 percent for total dissolved solids, 94 percent
for ammonia, and 83 percent for total organic carbon. The long-term flux
decline (not due to fouling) was virtually zero over the last 100 hours of the
test. A treatment with an enzyme-active laundry presoaking product was
found to be effective in cleaning the membranes and restoring the flux to
levels existing before fouling.
Preliminary membrane optimization studies showed that the reverse osmosis
properties of these new membranes could be improved by closely controlling
the polymer preparation procedures and the casting and annealing conditions.
The ultrathin nature of the membranes was an important factor in imparting
high flux properties to these membranes.
A preliminary economic analysis showed that, for a tubular configuration,
a cost of approximately ten cents could be expected for one square foot of
the composite of membrane and microporous support film. This cost was
insensitive to polymer raw material costs.
This report was submitted in fulfillment of Project Number 17020EFA,
Contract 14-12-587, under the sponsorship of the Water Quality Office,
Environmental Protection Agency.
111
-------
TABLE OF CONTENTS
Page
INTRODUCTION 1
EXPERIMENTAL 3
Polymer Preparation 3
Preparation of the Composite of Membrane and
Supporting Film 6
Reverse Osmosis Testing 7
Secondary Effluent Feed 7
RESULTS 13
Most Promising Membranes 14
Effects of Effluent Properties 16
Preliminary Membrane Optimization 19
Extended Time Testing 28
Preliminary Economic Analysis of Production Costs of
Tubular Ultrathin Membrane/Support Composites 35
ACKNOWLEDGEMENTS 37
REFERENCES 39
APPENDIX A
Preparation Procedures of Pertinent Polymers
APPENDIX B
Analytical Methods
APPENDIX C
Data and Results of Preliminary Economic Analysis
-------
LIST OF FIGURES
Number Title Paj
1 Structure pf Cellulose 4
2 Structure of 3 -Glucan 5
3 Dynamic Cell for Reverse Osmosis „ 8
4 Schematic Diagram of Reverse Osmosis Test Apparatus . 9
5 Polymer Structure 18
6 Water Flux as a Function of Membrane Thickness:
Effect of Support Film 23
7 Water Flux as a Function of Membrane Thickness:
Effect of Annealing 25
8 Water Flux Data from Secondary Effluent Test with
Cellulose Methyl Sulfonate O-Propyl Sulfonic Acid 32
9 Water Flux Data from Secondary Effluent Test with
Commercial Acetate Asymmetric Membrane 33
vii
-------
LIST OF TABLES
Number Title Page
1 Standard Values of System Variables for Reverse
Osmosis Testing 10
2 Minneapolis-St. Paul Sanitary District Wastewater
Data 11
3 Reverse Osmosis Performance of Representative
Membranes with Sodium Chloride Solution 15
4 Most Promising Candidate Polymers as Membranes for
Municipal Waste Treatment by Reverse Osmosis 17
5 Effect of Polymer Preparation Procedure (Designated
as Batch Number) on Water Flux and Salt Rejection for
Cellulose Methyl Sulfonate O-Propyl Sulfonic Acid 21
6 Effect of Casting Solvent on Reverse Osmosis -
Performance of Cellulose Acetate O-Propyl Sulfonic
Acid (211-63A) 22
7 Flux Declines for Ultrathin Membranes on Millipore
and Polysulfone Supports in an 0. 1-Percent Sodium
Chloride Solution 24
8 Effect of Thickness on Reverse Osmosis Properties
of Cellulose Acetate (E 398-10) in Salt and Secondary
Effluent 27
9 Membrane Characterization for Long-Term
Secondary Effluent Test 29
10 Long-Term Secondary Effluent Test Results 31
ix
-------
CONCLUSIONS
This program has resulted in a number of important conclusions on the
feasibility and ultimate use of new polymeric ultrathin membranes for
treatment of municipal wastewater by reverse osmosis. These conclusions
are listed below.
1. Ultrathin membranes of unusual new polymers are technically
feasible for treatment of municipal wastewaters by reverse
osmosis. New types of membrane materials have demon-
strated high potential for an optimum combination of water
flux and rejection at a low cost.
2. Ultrathin membranes of two new polymer materials were the
most promising of the 44 tested on secondary effluent during
the screening studies. These were cellulose methyl sulfonate
O-propyl sulfonic acid (CMSOPSA) and cellulose acetate
O-propyl sulfonic acid (C.AOPSA). Both membranes, at
thicknesses around 500 A, exhibited water fluxes above 40
gallons per square foot per day (gfd) with the following
rejections: phosphate, above 99 percent; total dissolved
solids, above 95 percent; and ammonia, above 93 percent. ,
For CMSOPSA and CAOPSA, the dissolved organic rejection
was 93 and 89 percent, respectively, and the nitrate r.ejection
was 81 and 93 percent, respectively.
3. The performance of an ultrathin CMSOPSA membrane during
a 150-hour, long-term test with a secondary effluent feed
was outstanding. During this test, in which the membrane
was cleaned with Biz* (a common household laundry pre-
soaking product) every twenty hours, the membrane
exhibited an average flux of 35 gfd (varying from 55 to 30
between cleanings) and the following rejections: total
dissolved solids — 96 percent; ammonia — 94 percent; and
dissolved organics — 83 percent. The average flux between
cleaning did not decline after the first forty hours.
4. Commercially available asymmetric cellulose acetate
membranes exhibited significantly lower water fluxes (less
than 15 gfd for 150 hours) than the ultrathin CMSOPSA
membranes, although the rejections were similar.
*Mention of proprietary equipment or products is for information purposes
only and does not constitute endorsement by the Water Quality Office,
Environmental Protection Agency.
xi
-------
5. Preliminary hydrolysis degradation studies (by pH effect)
showed that ultrathin CMSOPSA membranes were significantly
more resistant to basic hydrolysis than were the ultrathin
cellulose acetate membranes. Thus, in actual use, CMSOPSA
membranes may require fewer replacements than cellulose
acetate membranes.
6. Ultrathin cellulose acetate membranes exhibited water fluxes
and rejections similar to those of ultrathin CMSOPSA and
CAOPSA membranes. However, ultrathin membranes of
these new polymers have greater potential for more efficient
and low cost applications than do ultrathin cellulose membranes
because of (1) their greater resistance to hydrolysis, and-
(2) the expected improvement in reverse osmosis properties
of CMSOPSA and CAOPSA with structural and process
optimization.
7. Preliminary optimization studies of the new membranes
identified the variables that must be controlled to improve
the reverse osmosis performance. By (1) rigorously con-
trolling the polymer preparation; (2) optimizing the casting
procedures; (3) using a polysulfone support film;
(4) optimizing the annealing conditions; and (5) producing as
thin a membrane as practical, it should be possible to
produce a CMSOPSA membrane that has significantly
better reverse osmosis performance than any membrane
commercially available today.
8. A preliminary cost analysis showed that CMSOPSA or
other unusual polymer materials combined with a poly-
sulfone support film in a tube would cost about 10 cents
per square foot, exclusive of the tube.
xii
-------
RECOMMENDATIONS
This program has consisted primarily of screening new and ultrathin mem-
branes for reverse osmosis treatment of municipal wastewater. The scope
of this program did not include rigorous optimization, engineering con-
siderations, or field demonstrations. Because some important new candidate
membranes have resulted from this program, it is recommended that it be
continued in order to realize fully the practical advantages of these mem-
branes.
The first step that should be carried out in a second phase of this program
is optimization of the reverse osmosis properties of ultrathin membranes
of cellulose methyl sulfonate O-propyl sulfonic acid (CMSOPSA) and
cellulose acetate O-propyl sulfonic acid (CAOPSA). The following properties
must be adjusted for optimum performance:
• The polymer synthesis procedures must be investigated
and controlled to produce polymers of known purity and
predictable degrees of substitution. Thus, the exact
polymer structure that would give optimum performance
would be elucidated.
• Casting procedures in the most applicable configuration
(probably tubes) must be optimized. Optimization would
include investigation of polymer casting-solution concen-
trations, casting solvents, additives (such as surfactants)
to the water or casting solution, and environmental con-
ditions.
• The effect of annealing on water flux and rejections must
be determined.
• The production of as thin a membrane as practicable in
a consistent manner must be investigated.
The optimization testing would include long-term reverse osmosis tests
with secondary effluent (up to 500 hours) during which membrane cleaning
and feed pretreatment procedures would be investigated and optimized.
A second step in this proposed program would include engineering studies
and should yield a more comprehensive estimate of costs and serve as a
basis for preparation of a field demonstration of these new membranes.
These considerations would lead directly to a third phase which would
consist of a demonstration at a sewage treatment plant.
It is also recommended that a small effort, concurrent with the above, be
carried out to develop some of the more promising high-flux membranes,
such as the cellulose acetate adduct of methylol phenol, that could not be
explored in depth during this current program.
Xlll
-------
INTRODUCTION
Treatment of municipal wastewaters by reverse osmosis has great potential
as a simple, effective, and inexpensive method to obtain nearly complete
removal of both dissolved and particulate contaminants. To realize this
potential fully, new membranes with high water fluxes and high rejections
of the various contaminants are necessary. In the program described here,
a series of new and ultrathin membranes was developed and tested. These
new membranes ranged from 250 to 5000 angstroms in thickness (1 x 10~6 to
2 x 10~5 inch) and consisted of esters, ethers, and mixed esters and ethers
of polysaccharides (primarily cellulose). In previous research efforts in
water desalination, these membranes have provided water fluxes up to
35 gfd at 97-percent salt rejection (after 1200 hours of testing under 800-psi
pressure, at one percent sodium chloride feed and 25°C).'!' The ultrathin
membrane was supported by a microporous support film developed at North
Star Research and Development Institute, that had high chemical resistance
and strength. The composite system of membrane and-support has been
successfully applied to tubular configurations, and in situ replacement of
the ultrathin desalination film has been shown feasible. (1» 2) This approach
toward the treatment of municipal wastewaters by reverse osmosis offers
the possibility of high reverse osmosis performance and low cost.
The objective of the program reported here was to develop promising new
membrane candidates for municipal waste treatment by reverse osmosis.
This first phase has included the synthesis, fabrication, and screening of
new and ultrathin membranes for effective reverse osmosis performance.
A high-performing membrane would exhibit high water flux, adequate
rejection of undesirable species, and low flux decline. The approach toward
accomplishing this objective was two-fold:
• Initial screening of the membranes by running short
reverse osmosis tests using sodium chloride. Salt
rejections below 50 percent or water fluxes below 20
gfd were generally considered not acceptable.
• Reverse osmosis testing of the more promising mem-
branes from the salt-solution screening using secondary
effluent from a municipal sewage treatment plant. Along
with water flux, the rejections of the following species
were measured: total dissolved solids (TDS), ammonia,
organics after filtration as total organic carbon (TOO,
and soluble phosphates. Flux declines were observed
on the best membrane candidates for extended time
periods.
This report summarizes the results of the screening program, with emphasis
on the most successful new membrane candidates for reverse osmosis treat-
ment of municipal wastewaters.
-------
EXPERIMENTAL
The experimental effort for this program, included polymer preparation,
preparation of ultrathin membranes and support film, and testing reverse
osmosis performance. The general experimental procedures are summarized
below, with pertinent descriptions of equipment and apparatus.
Polymer Preparation
Structure
The basic repeating structure of cellulose is shown in Figure 1. The poly-
mers used in this program are formed by substituting esters or ethers for
the hydrogen atoms on the hydroxyl groups associated with the 2, 3, and 6
positions on the ring. The degree of substitution (DS) is therefore three, at
maximum substitution. For example, for the ester derivative cellulose
acetate, E 398-10, the acetate ester groups comprise an average of 2.4 of
the available sites (DS for acetate = 2.4), and the remaining 0.6 site is the
hydroxyl (-OH) group (DS for hydroxyl = 0.6). The structure of the basic
repeating unit of |3-glucan, another polysaccharide used in this work, is
given in Figure 2. The same 'substitutions at the 2, 3, and 6 positions as
cellulose are possible for p-glucan.
Transport properties of materials across polysaccharide membranes are
grossly affected by substitution of functional groups (e.g., esters and ethers
at the positions indicated in Figure 1). (2, 3, 4) Polysaccharide derivatives
with a high degree of substitution of ester groups, and, in addition, a minor
proportion of highly hydrophylic substituent, were considered desirable.
The type of hydrophylic substituent was varied both in charge (acidic, basic,
or neutral) and in the relative strength of the acid or basic groups.
Preparation Procedures
The procedure for the preparation of the polysaccharide derivatives followed
two basic techniques:
• The starting material was cellulose acetate with free
hydroxyl groups (that is, a DS of less than three for
acetyl). The cellulose acetate was then reacted with a
suitable hydrophylic reagent to produce the desired
derivative for film casting. Completion of the reaction
was determined in the product from the infrared absorption
of a 1-mil film at 3500 cm~l (absence of the hydroxyl
absorption peak indicates complete substitution). The
substitution of the new group was then estimated by
subtraction of the known acetyl content from 3. 0.
-------
The chain structure may be represented as:
G.
G.
G
3(1-4) P(l-*4)
where:
G = anhydroglucose unit
Cellulose ester:
H atom on the hydroxyl' groups associated with the 2, 3, or 6 position
substituted by ester group; for example, acetate /O - C -
Cellulose ether:
H atom on the hydroxyl group associated with the 2, 3, or 6 position
substituted by ether group; for example, methyl sulfonate:
(O
II
) - S - CH,
II
O
Figure 1. Structure of Cellulose
-------
CH2OH
All glucose units are linked beta (1-3) in the chain skeleton and beta (1-6) in
the appended glucose units. The chain structure may be represented as:
| 0(1-* 6)
G G
G-
G
I 0(1 - 6)
G G
0(1-3) 0(1-3) 0(1-3). 0(1-3)
where:
G = anhydroglucose unit
Ester and ether derivatives: substitution for some H atoms as in Figure 1.
Figure 2. Structure of 0-Glucan
-------
The starting material was pure cellulose (cotton linters).
A given amount of hydrophylic reagent was then reacted
with the cellulose for a given time period in order to
achieve a desired DS. The partially substituted cellulose
was then acetylated to full substitution. A variation of
this procedure included simultaneous substitution of a
hydrophylic agent and acetylation. Individual preparation
procedures of important derivatives are given in Appendix A.
Preparation of the Composite of Membrane
and Supporting Film
Membrane Casting
Two casting techniques, casting on a water surface and on a glass plate,
were used to form ultrathih membranes. The polymeric derivative was
first dissolved in an organic solvent. After the polymer was in solution,
the casting dope was filtered through a Seitz K-5 filter pad to remove gel
particles.
Casting on a Water Surface. The solution with up to five percent by weight
polymer was poured onto a water surface. The solution spread quickly and
gelled to form an ultrathin membrane. The thickness was controlled by
pulling the gelling membrane across the water surface at a constant rate.
The solvents included cyclohexanone or a 90:10 mixture of dichloromethane:
methanol. Most of the ultrathin membranes were formed by this technique.
Casting on a Glass Plate. A 2-mil Gardner knife was used to draw the
casting solution across a glass plate. The membrane was then formed by
evaporation of the solvent at ambient conditions in a clean-air hood. This
method was used for a few polymers that could not be cast on water. These
membranes, at about 10, 000 A in thickness, were always thicker than the
membranes cast on water.
Membrane Thickness
The thicknesses of the membranes were measured on a Reichert Metallo-
graph equipped with a Nomarski polarization interferometer. (6)
Membrane Support
Either a Millipore VFWP filter or a microporous polysulfone film was used
as a support for the ultrathin membranes in the reverse osmosis tests.
The preparation procedure for the polysulfone film is described else-
where. (2, 6) The support was laminated to the ultrathin membrane by
immersing the support in the water underneath the ultrathin membrane and
pulling it out of the water with the ultrathin membrane on its surface.
-------
Reverse Osmosis Testing
All the tests were performed with the membrane-support composite placed
in a flat circular test cell. The cell is shown in Figure 3, with the key
dimensions of channel height and diameter indicated. The membranes were
tested in a recirculating (constant feed concentration) system shown in
Figure 4, with provision for testing four membranes at one time. In addition
to the four reverse osmosis cells, the test loop contained a five-gallon brine
reservoir, a Model 241-144B Milton Roy pump, an accumulator, a heat
exchanger, a needle valve for system pressure control, a 100-mesh high-
pressure filter, a rotameter, and a five-micron Cuno filter.
The general conditions for the reverse osmosis measurements with sodium
chloride and secondary effluent are given in Table 1.. The feed flow rate of
3500 ml/min produces a velocity adequate to control concentration
polarization. C7) Biebrich scarlet dye was always added to the salt solution
to confirm that no pinholes were present.
Each membrane underwent twenty hours of reverse osmosis testing with the
sodium chloride feed. If this; initial screening test looked promising, the
membrane was then tested with secondary effluent.
For all the reverse osmosis tests, water flux was obtained by timing the
upward movement of the water meniscus in a 10-ml pipette. The product-
water flow rate in ml/min was multiplied by the conversion factor for these
cells of 21.2 to obtain gallons per square foot per day (gfd). In the case of
the salt solution screening tests, the percent rejection of salt was calculated
from the salt concentrations in the feed and the product as determined
from conductivity measurements.
For secondary effluent the constituents for which the feed and product water
were analyzed included total organic carbon (TOO, ammonia nitrogen,
nitrate nitrogen, orthophosphate, suspended solids, and dissolved solids.
Except for, suspended solids, the feed was filtered with a 0.45-micron filter
before analysis. The analytical procedures and any procedural modifications
are described briefly and referenced in Appendix B.
Secondary Effluent Feed
The feed for the secondary effluent"tests was obtained from the Minneapolis-
St. Paul Sanitary District plant. -This plant has a high-rate activated sludge
process designed for 75-percent removal of biochemical oxygen demand
(BOD). Average analytical data orivthe raw, primary, and secondary
effluents for a typical month are given-in. Table 2. To prevent excessive
biological activity before reverse osmosis testing, the secondary effluent
feed was stored at 4°C.
-------
00
ULTRATHIN
MEMBRANE
POROUS SUPPORT
FILM
FEED SOLUTION FEED SOLUTION
•INLET /-OUTLET
FEED SOLUTION
CHAMBER
PRODUCT
WATER OUTLET
^\W^\\\\\\\\\\V I
1.75 IN. EFFECTIVE DIA.
KRAFT PAPER
POROUS PLATE
Figure 3. Dynamic Cell for Reverse Osmosis
-------
N2 REG
CO
r i —
SUPPLY — i
TANK
(20 1)
4*
CIIRftP
SUPPLY DRAIN TANK L—
DIAPHRAGM ®
pi 1 MB _ 1
1 1
£ (S ^
SURGE GAGE '
TANK SNUBBER^
FLUSH COIL «=
/ N2 TANK
-®— i
xy Jt
DRAIN
jSTARTERl
III
3*. 220 VAC
ROTAMETER 100-MESh
RETURN
SUPPLY
C
F
:3I6SS
TANlf" ~"~~"|
JEPTH 1
'IITFR I
1 PRESSURE 1 1
1 CONTROL IJ
/ VALVE JT
1 ®
^^1 STRAINEf
O^ FLUSH
^jT- ' VALVE
)/ (0-1000 PSD
f^
< SNUBBER
FLUSH
DRAIN
y
REFRIGERATED
WATER BATH
TEST CELL UNIT (4)
r^
*
, L,
czz
TREATED
i
6b- i
xy
r?
.::. 71
TEST
CELL
|
Figure 4. Schematic Diagram of Reverse Osmosis Test Apparatus
-------
Table 1. Standard Values of System Variables for Reverse
Osmosis Testing
Screening Tests with Salt Solution
Feed 0.1 weight percent sodium chloride
pH 7
Temperature 25°C
Pressure 600 psig
Feed flow rate 3500 ml/min
Time of test 20 hr
Screening Tests with Secondary Effluent
Feed Secondary effluent filtered through
a 1/u cartridge
pH* As received
Temperature 25°C
Pressure 600 psig
Feed flow rate 3500 ml/min
Time of test 6 hr
Precompressed 20 hr, per conditions for salt .
solution
*In sbme screening tests of secondary effluent, the effluent was
acidified.
10
-------
Table 2. Minneapolis-St. Paul Sanitary District
Wastewater Data*
Chemical Oxygen Demand
(COD)(mg/l)
Biochemical Oxygen Demand
(BOD)(mg/l)
Ammonia — N (mg/1)
Kjeldahl N (mg/1)
Phosphate (mg/1)
Total Suspended Solids
(mg/1)
Total Dissolved Solids
(mg/1)
Chloride (mg/1)
Dissolved Oxygen (mg/1)
Temperature (°F)
PH
Soluble Organic s (mg/1)
Raw**
370-520
222
4-12
19-24
7-19
291
604
83-99
0.40
70
7.3
—
Primary
290-430
167
4-12
18-23
7-18
117
606 '
75-98
0.42
70
7.3
—
Secondary
190-220
56
.4-7
9-15
6-15
49
566
59-98
2.6
—
7.4
58t
*Permission for use granted by the District, April 1970.
**Single figures = average for July 1969.
Range figures = data from 1967-1968.
tMeasured at North Star.
11
-------
RESULTS
The experimental work included the four following major efforts:
1. Membrane screening to determine the most effective
polymer.
2. A brief look at major effluent variables and their effect
on reverse osmosis performance.
3. Preliminary optimization of the best membrane candidates
to define the important optimization variables and improve
performance.
4. Some extended-time testing of the best membrane
candidates to evaluate their long-term performance.
In addition, a preliminary economic analysis was carried out on some of
the new polymer materials. Important data, their interpretation, and
conclusions are summarized and presented below.
Forty-four polymers were tested in the reverse osmosis cell for water flux
and rejection using a 0.1-percent sodium chloride feed (in some early cases,
a one-percent feed). The following membranes were tested for reverse
osmosis performance using the secondary effluent feed (cellulose acetate is
referred to as CA).
Anionic
CA sulfate
Cellulose methyl sulfonate O-propyl sulfonic acid
CA O-propyl sulfonic-acid
CA adduct of methylol phenol
CA phosphate
CA phthalic acid half ester
Cationic
CA methylol morpholine adduct
CA pyrrolidone adduct
3-glucan acetate diamethylaminoethyl ether
13
-------
Cationic -Anionic
CA pyridine dicarboxylic acid half ester
Nonionic
CA
CA methyl sulfonate
CA dimethyl sulfonate
Xylan acetate
Methyl CA
Methyl cellulose methyl sulfonate
Cellulose trimethyl sulfonate
CA propane sulfonate
P-glucan diacetate
The most promising new membranes were of the anionic type. The reason
is not known at this time.
The salt-screening procedures'saved a considerable amount of time in this
membrane screening program. The membranes-with salt rejections less
than 50 percent or water fluxes less than 20 gfd were eliminated. In
general, the salt rejections were found to be proportional to the secondary
effluent rejections. One apparent exception, however, was the rejection
of the dissolved organics; higher TOC rejections than expected were'obtained
using low-salt-rejecting membranes. Some typical salt-screening results
are given in Table 3. Most of these membranes did not pass to the secondary
effluent testing stage.
Most Promising Membranes
Of the 44 membranes of different polysaccharide derivatives screened, the
following new polymers were found to exhibit outstanding reverse osmosis
properties:
• Cellulose methyl sulfonate O-propyl sulfonic acid (CMSOPSA)
• Cellulose acetate O-propyl sulfonic acid (CAPOSA)
• Cellulose acetate phosphate
• Cellulose acetate adduct of methylol phenol
14
-------
Table 3. Reverse Osmosis Performance of Representative Membranes with
Sodium Chloride Solution
Conditions:
NaCl 0.1 perc
ent
; VFWP Filter
Feed pH 7
Polymer and Batch Number
Anionic
Cellulose Acetate Sulfate (211-52A)
Cellulose Acetate Phosphate (211 -58B)
Cationic
p-glucan Acetate Dimethylaminoethyl
Ether (211-40A)
Cellulose Acetate Adduct of Dimethylol
Piperazine (211-82B)
Cellulose Acetate Diethanolamine Adduct
(211-45A)
Cellulose Acetate Ethyl Trimethyl
Ammonium Chloride (211-71A)
Nonionic
3-glucan Diacetate (87-26A)
Methyl Cellulose Methyl Sulfonate
Acetate (211-64A)
Cellulose Acetate Methyl Sulfonate
(211-10C)
Thickness
(A)
1600
1400
2000
1600
5900
1800
1800
1700
1300
Flux after
~20 hr (gfd)
13
31
10
66
48
5
14
7
14
Salt
Rejection (%)
84
72
99
52
28
94
94
98
97
-------
Performance of these membranes is shown in Table 4. For purposes of com-
parison, a commercially available asymmetric cellulose acetate membrane
is included in the table.
Two membranes looked particularly promising for both high flux and high
rejection: cellulose acetate O-propyl sulfonic acid (CAOPSA) and cellulose
methyl sulfonate O-propyl sulfonic acid (CMSOPSA). Both membranes
exhibited high ammonia rejections (96 and 93 percent, respectively), adequate
TOG rejections (89 and 93 percent, respectively), and high TDS rejections
(96 and 95 percent, respectively). For practical purposes, the phosphate
rejections were 100 percent. The rejection of nitrates was high for
CAOPSA (93 percent), but lower for CMSOPSA (81 percent). The water
flux through these membranes was above 40 gfd. Presumably, if the
CMSOPSA membrane had been as thin as the CAOPSA. membrane, CMSOPSA
would have surpassed CAOPSA substantially in water flux.
t;
The DS of both of these membranes for the propyl sulfonic acid substituent
was estimated at less than 0.1. Increasing the DS of this functional group
would increase the water flux of the membrane but decrease the rejection.
The structures of these two polymers are given in Figure 5.
The cellulose acetate adduct of methylol phenol was included because of its
high flux (64 gfd) and high TOC rejection (89 percent). Chemical and
process optimization should improve its performance considerably.
The commercial cellulose acetate asymmetric membrane (RO-97) that was
used for comparison in Table 4 exhibited generally adequate rejections but
low water flux. Data on the reverse osmosis performance of ultrathin
cellulose acetate is presented and discussed later in this report.
Effects of Effluent Properties
The water flux and salt rejection of membranes for municipal wastewater
treatment are affected by operating variables such as temperature, pressure,
feed pH, feed concentration, and effluent filtration. Temperature and
pressure effects are well known, and were not studied in this program.
The remaining variables were given brief consideration and, except for
membrane cleaning (discussed later in this report), their effects on reverse
osmosis performance are presented in this section.
Feed pH and Resistance of Membrane to Degradation
The pH of the feed affects the structural integrity of a polymeric membrane.
For cellulose acetate, the optimum feed pH is between 5 and 6, where
degradation by hydrolysis is minimal. (8) In addition to hydrolytic damage
to the membranes, the feed pH can affect the molecular structure of the
solution constituents as well as the ionic form of an anionic or cationic
membrane. These latter effects can result in significant changes in reverse
osmosis perf9rmance.
16
-------
Table 4. Most Promising Candidate Polymers as Membranes for Municipal Waste
Treatment by Reverse Osmosis
Conditions :
Temperature 25°C
Effluent Filtered Secondary Effluent
Feed pH ~ 7 for Salt Solution and 7-8 for Effluent
Feed Flow Rate 3500 ml/min
Support Film Millipore VFWP Filter
Each membrane was tested initially for 20 hours in 0. 1 percent NaCl at pH 7
Polymer and Batch Number
Cellulose Methyl Sulfonate
O-Propyl Sulfonic
Acid (211-96A)
Cellulose Acetate O-Propyl
Sulfonic Acid (211-95A)
Cellulose Acetate Phosphate
(211-94A)
Cellulose Acetate Adduct
of Methylol Phenol
(211-83B)*
Cellulose Acetate Eastman
(RO-97)*
Thickness
(A)
570
370
260
740
Asymmetric
Water Flux After
~20 hr with Salt
Solution (gfd)
44
43
49
64
20
Percentage Rejection
NaCl (with
Salt Solution)
90
96
74
50
.
96
TDS
95
96
89
73
96
N03
81
93
64
43
74
NHs
93
96
81
70
96
TOC
93
89
89
89
94
PO4
>99
>99
96
95
>99
*Effluent feed rate, 1000 ml/min.
-------
0
II
0-S-CH3
6
0
II
Cellulose* 0-S-CHs
ii
0
H H H 0
I I I II
0-C-C-C-S-OH (DS
-------
The effect of high pH on the reverse osmosis performance of CMSOPSA
membranes was shown to be significantly less than on that of cellulose
acetate. For cellulose acetate, hydrolysis in acidic or basic solution
results in removal of the acetate groups and subsequent poor rejection.
An experiment with CMSOPSA (211-89A) and cellulose acetate (E 398-10)
using a sodium chloride solution at pH = 11, showed that the reverse
osmosis performance of the CMSOPSA membrane deteriorated much less
than that of the cellulose acetate. The salt rejection of the cellulose acetate
membrane was 18 percent, and that of CMSOPSA was 90 percent, after
twenty hours of testing. CMSOPSA may also deteriorate slower than
cellulose acetate under the less extreme pH conditions for municipal waste
effluents.
Feed Concentration
A test was conducted to determine the effect of feed concentration on the
rejection performance of a membrane. In the test, the sodium chloride
concentration of a salt solution feed was increased from 500 mg/1 to
5000 mg/1 to simulate the concentration range occurring in a reverse
osmosis unit during treatment of an effluent. As expected, there was no
effect on the rejection performance of cellulose acetate (RO-89). The salt
rejection of a CMSOPSA (211-89A) membrane decreased only slightly, from
96 to 95 percent. Some decrease could be expected because of the existence
of some ion exchange character in this derivative. (9)
Preliminary Membrane Optimization
The reverse osmosis performance of a membrane can be improved to a
considerable extent by modifications in its fabrication process. The mem-
brane variables that can be modified for obtaining high water"fluxes, high
rejections, and low flux declines include:
• Polymer preparation process and DS
• Casting conditions
• Type of support film for ultrathin membranes
• Membrane annealing conditions
• Membrane thickness
Controlling the effects of the above variables would improve the performance
of the membrane/support composite for optimum treatment of municipal
wastewaters. A preliminary optimization study was carried out on the more
promising membranes, CMSOPSA, CAOPSA, and cellulose acetate. Since
the flux and rejection of membranes with a sodium chloride feed has been
shown to be roughly the same as the unfouled flux and TDS rejection of a
secondary effluent feed, and because a greater number of experiments can
be carried out in a given time period with sodium chloride, the 0.1-percent
sodium chloride feed was used under the conditions in Table 1.
19
-------
Polymer Preparation and Degree of Substitution
Table 5 illustrates the results obtained from different batches of CMSOPSA.
The membranes listed were prepared from three different batches: (1) the
original batch (211-89A); (2) a duplicate batch <211-96A); and a third batch
(211-96B), that was prepared with a higher DS for O-propyl sulfonic acid
to obtain higher water flux. The results are summarized below.
1. Batch 211-89A was the original preparation. Membrane 3
was cast from the same polymer solution as membranes 1
and 2, but one month later. It was cast very thin (280A)
to achieve a high water flux. This membrane exhibited the
same salt rejection as membranes 1 and 2, but with only
slightly greater water flux. The unexpectedly low water
flux could be due to polymer solution changes during the
one-month period, producing a tighter membrane. Thus,
the storage of polymer solutions must be considered in an
optimization study.
2. Batch 211-96A was prepared with approximately the same
reaction conditions used previously for. Batch 211-89A
cited above. Membrane 4 cast from Batch 211-96A did
not exactly reproduce any results of membranes 1, 2,
and 3.
3. Batch 211-96B was prepared with reaction conditions adjusted
to obtain a product with a higher DS of O-propyl sulfonic acid.
Membrane 5 formed from this batch had a higher water flux
and a lower salt rejection than membranes 1, 2, 3, and 4.
This is the expected behavior for a polymer more highly
substituted with flux-promoting substituents.
These results indicate that, at present, it is difficult to prepare a polymer
with predictable properties, and thus the reaction conditions must be
rigorously controlled in an optimization study. Investigation and study of
the reaction conditions were not within the goals of this screening program.
Casting Conditions
Among the many casting conditions affecting the performance of the membrane,
an interesting observation was made regarding the type of solvent used. It
was found that some polymer-sol vent solutions gave membranes that were
cloudy in appearance, even though the solution appearance indicated good
polymer solubility. The cloudy-appear ing membranes exhibited higher flux,
but lower salt rejection, than the normal transparent membranes. These
cloudy membranes probably contained voids or gel particles; however, the
exact explanation for this behavior is not known. An example is given in
Table 6.
20
-------
Table 5. Effect of Polymer Preparation Procedure (Designated as Batch Number) on
Water Flux and Salt Rejection for Cellulose Methyl Sulfonate O-Propyl
Sulfonic Acid
Conditions:
Pressure 600 psig
Temperature 25°C
Salt Solution 0. 1 -percent NaC
Feed pH , 7
Feed Flow Rate
Support Film . .
3500 ml/min
Millinore VFWP
Membrane
Number
1
2
3
4
5
Batch
Number
211-89A
211-89A
211-89A
211-96A
211-96B
1
Filter
Thickness
(A)
1300
1200
280
570
1200
Water Flux After
20 hr (gfd)
37
30
40
44
66
Percentage
Rejection
94
95
95
90
75
-------
Table 6. Effect of Casting Solvent on Reverse Osmosis Performance
of Cellulose Acetate O-Propyl Sulfonic Acid (211-63A)
Solvent(s)
Cyclohexanone
Dichloromethane /
methanol
Membrane
Appearance
Cloudy
Transparent
Thickness
(A)
2000
1800
Flux
(gfd)
37
26
Percentage
Salt Rejection
54
96
Type of Support Film
Millipore VFWP filters were used for supporting ultrathin membranes
during the early part of this program. More suitable microporous support
films made from polysulfone have been developed at North Star Research
and Development Institute for supporting ultrathin membranes and are
presently used for the supports for tubular ultrathin membranes. (&• 2) To
determine optimum performance of the composite in long-term tests, the
two supports were compared for their effects on water flux and flux decline.
Effect on Water Flux at Low Membrane Thicknesses. High fluxes can be
obtained by use of the polysulfone support film rather than a Millipore
support film. Figure 6 shows the effect of the support on the water flux/
thickness relationship of two ultrathin membranes.
The use of Millipore or polysulfone supports caused no difference in flux
at high thicknesses for both membranes. However, thinner membranes,
which are necessary to achieve higher fluxes, require the use of poly-
sulfone supports to attain these higher fluxes. Polysulfone enhanced the
flux of cellulose acetate (E 398-10) membranes to below 500 to 700 angstroms
in thickness. Polysulfone supports were necessary to attain high flux for
CMSOPSA (211-96B) membranes below 2000 angstroms in thickness. The
polysulfone sheets^have smaller, more uniformally sized and a larger
number of pores than Millipore filters. (D We believe that these pore
properties contribute toward the superior support behavior of-the poly-
sulfone. Further research will be necessary, 'however, to determine the
exact reasons for the improved support properties of polysulfone.
Effect on Flux Decline. Lower flux declines can be realized by using the
polysulfone support film. Flux-decline values from the slopes of
logarithmic plots of flux versus time for cellulose acetate and CMSOPSA
membranes on both supports are given in Table 7.
22
-------
to
oo
100
80
60
«40
20
X
O Cellulose Acetate (E 398-10) on Millipore VFWP Support
• Cellulose Acetate (E 398-10) on Polysulfone Support
& CMSOPSA (2II-96B) on Millipore VFWP Support
A CMSOPSA (2II-96B) on Polysulfone Support
X
X Polysulfone Support
X
Support
Polysulfone Support
Millipore Support
-
Reciprocal Membrane Thickness x 10 (A )
10,000 2000
1000 800 700 500 400 0
Membrane Thickness (A)
300
Figure 6. Water Flux as a Function of Membrane Thickness: Effect of
Support Film.
-------
Table 7. Flux Declines for Ultrathin Membranes on Millipore and
Polysulfone Supports in an 0.1-Percent Sodium Chloride
Solution
Polymer
Cellulose acetate (E 398-10)
CMSOPSA (211-96B)
CMSOPSA (211-96B)
Annealing
None
None
Air Dry
Flux Decline*
With
Millipore
Support
-0.08
-0.14
With
Polysulfone
Support
-0.05
-0.11
0
*Slopes from logarithmic plots of flux versus time from 1 hour to 260
hours of testing.
As shown in. Table 7, the flux decline was -O.'OS for nonannealed cellulose
acetate using the polysulfone support film, compared to -0. 08 with the
Millipore support. The results with nonannealed CMSOPSA (211-96B)
showed the same effect.
Effect on Rejection. The type of support had no effect on salt rejection.
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 7 shows the effect of annealing on the water
flux/membrane thickness relationship for cellulose acetate and CMSOPSA.
The membranes were annealed by air drying after being placed on poly-
sulfone supports. The water fluxes for both polymers were reduced.
Effect on Flux Decline. The effect of annealing by air drying a CMSOPSA
(211-96B) membrane on a polysulfone support was significant and is
indicated in Table 7. The flux-decline slope for the annealed membrane
on a polysulfone support was zero, but the nonannealed membrane on a
polysulfone support has a flux decline slope of -0.11.
24
-------
DC
100
80
X
3
40
0
*-
O
20-
Nonannealed
CMSOPSA
Annealed
CA
2 3
Reciprocal Membrane Thickness x I03
10,000 2000
1000800700600 500 400
Membrane Thickness (A)
300
Figure 7. Water Flux as a Function of Membrane Thickness: Effect of Annealing
-------
Effect on Salt Rejection. Annealing increased the salt rejection of the
CMSOPSA membranes from approximately 70 percent (for nonannealed) to
94 percent. For cellulose acetate, the salt rejection was increased from
approximately 92 to 95 percent by annealing.
Effect of Membrane Thickness
The solution-diffusion model forms the basis for the plot of flux versus
reciprocal thickness. The final result of this proposed mechanism of
transport states 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 pro-
portional to the constant term.
Figures 6 and 7 illustrate the effect of thickness on water flux for both
polymers CMSOPSA (211-96B) and cellulose acetate (E 398-10) annealed
and nonannealed. With a polysulfone support, the relationship for cellulose
acetate was linear with inverse thickness over the thickness ranges studied.
Although the annealed CMSOPSA membrane exhibited linearity over the
thickness range> the nonannealed CMSOPSA did not. The effect of the
Millipore support on nonlinearity has been discussed earlier.
It is apparent that these membranes, regardless of polymer, must be ultra-
thin to achieve high water flux while still maintaining high rejections. A
cellulose acetate (E 398-10) membrane was cast at a thickness of 260 A by
casting on a polysulfone support in a 1-1/4-inch tube. The support/membrane
composite was air dried, cut from the tube, and measured for reverse
osmosis performance in a flat cell. Its performance compared with a flat-
cast 2100-A-thick cellulose acetate membrane is given in Table 8. This
thin membrane exhibited a water flux generally predicted from the inverse
thickness relationship while maintaining high rejections. It also compares
favorably with the CAOPSA membrane prepared in the same manner
(Table 3).
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
CMSOPSA and CAOPSA make this a necessary step. By (1) rigorously
controlling the polymer preparation, (2) optimizing the casting pro-
cedures, (3) using a polysulfone support film, (4) optimizing the annealing
conditions, and (5) producing as thin a membrane as practical, a CMSOPSA
membrane, that, at present, surpasses cellulose acetate in stability, should
be produced with significantly better reverse osmosis performance than any
membrane available today.
26
-------
CO
Table 8. Effect of Thickness on Reverse Osmosis Properties of Cellulose Acetate
(E 398-10) in Salt and Secondary Effluent
Conditions:
Pressure
Temperat
Salt Soluti
Effluent .
Feed pH
Feed Flov
Support F
Thickness
(A)
2100
260*
600 psig
ure 25°C
Filtered
~7 for Sa
/ Rate 3 500 ml/
Water Flux
After~20 hr
with Salt
Solution (gfd)
14
47
ent NaCl
Secondary Effluent
It Solution and 7-8 for Effluent
tnin
VFWP Filter
Percent Rejections
NaCl
(with
Salt
Solution)
95
93
TDS
>99
95
N03
82
89
NH3
90
88
TOC
92
93
P04
>99
>99
*Cast on polysulfone support in tube and air dried.
-------
Extended Time Testing
In the screening studies, all the reverse osmosis tests with secondary
effluent were carried out for six hours after a twenty-hour reverse osmosis
test (precompression) with salt solution. To determine applicability to an
actual municipal waste situation, it was necessary to determine the
behavior of the membranes during testing of a longer duration. Reverse
osmosis tests of approximately one week's duration were carried out on
selected membranes, and preliminary information on long-term flux, flux
decline, and rejections was obtained.
Experimental Conditions and Procedures
The experimental procedure consisted of circulating a fresh, filtered
secondary effluent feed at 600 psig through the four flat test cells containing
the membranes. After twenty hours, the membranes in the test cells were
cleaned, deionized water was circulated through the cells briefly under 600
psig, and fresh, filtered secondary effluent was circulated through the cells
under 600 psig again for twenty hours. These twenty-hour test cycles, with
the cleaning and deionized water procedures, were repeated seven times
(150 hours). Each cycle was designated as a "test section" and given a
number.
A preliminary long-term test was run to establish the procedures for
secondary effluent filtration and membrane cleaning. The most effective
filtration medium was glass fiber mats. '^O* H' The total suspended solids
in the secondary effluent was reduced from 52 to 15 mg/1. The cleaning
procedure used after each twenty-hour test section in this test series con-
sisted of circulating a solution of 20 g/gal. Biz in the cells for 1. 5 hours.
The long-term tests provided information on average water flux, fouling
flux decline, long-term flux decline, and rejection. The average water
flux of a test section was found by dividing the area under the test section
curve (flux versus time) by the time period (~20 hours). The fouling flux
decline was the decrease in flux that occurred during each twenty-hour test
section. The long-term average flux decline was the decrease of the
average water flux values from the first to last test sections. The rejections
measured were total dissolved solids, ammonia, and dissolved organics
(measured as total organic carbon).
Observed Reverse Osmosis Properties
The four membranes tested are listed in Table 9, along with their initial
salt rejection data. One of the most promising high-flux, high-rejecting,
new ultrathin membranes, CMSOPSA, was used. Another new ultrathin
membrane, cellulose acetate methyl sulfonate, was used because of its
high salt rejection in the initial screening study (at a thickness of 1300 A —
Table 3). A thinner membrane would retain the high rejection, but increase
28
-------
Table 9. Membrane Characterization for Long-Term Secondary Effluent Test
CO
CO
Conditions:
Cell Flat
Solution 0.1 -percent NaCl
Pressure 600 psig
Feed Flow Rate 3 500 ml/min
Temperature 25°C
pH.... 7
Support for ultrathin membranes — polysulfone
All ultrathin membranes annealed by air drying
Polymer and Batch Number
CMSOPSA (211-89A)
Cellulose Acetate Methyl
Sulfonate (211-IOC)
Cellulose Acetate (E 398-10)
Cellulose Acetate (RO-97)
Thickness
(A)
210
440
290
Asymmetric
~20-hr
Flux (gfd)
55
42
50
16
Salt
Rejection (%)
94
94
94
96
-------
the water flux compared to the membranes tested earlier. For comparison
purposes, the asymmetric cellulose acetate membrane (RO-97) was chosen
because it is the tightest standard commercial membrane. As a further
control, an ultrathin cellulose acetate membrane (E 398-10), which had
salt rejections comparable to the other membranes, was used. During
this test, all the ultrathin membranes were annealed by air drying to
achieve higher rejection. The control membrane of cellulose acetate
(RO-97) was used as received.
It was convenient to characterize the membranes by initial twenty-hour reverse
osmosis tests with sodium chloride. These tests also served as a "quality
control" for the membranes to be used. Also, the twenty-hour flux com-
pacted the membrane and eliminated further flux decline due to membrane
compression during the long-term test. In Table 9, the water fluxes at
the end of twenty hours varied between 42 and 55 for the ultrathin membranes
and under 20 for the asymmetric membrane. Salt rejections for all the mem-
branes were in the 95-percent range.
The experimental results are summarized in Table 10. The ultrathin
cellulose acetate membrane apparently developed a defect at the
beginning of the fourth test section (after 60 hours of testing). At
the beginning of the fourth test section, the initial water flux was
abnormally high (67 gfd) and the rejections had decreased significantly
(e.g., conductivity rejection from 96 percent to 85 percent). The
data for this membrane are thus given only for the first sixty hours
of testing.
The specific results for important reverse osmosis parameters are
summarized below.
• Average Water Fluxes
The highest average water flux — 43 gfd — was observed
for the CMSOPSA membrane in the first test section.
After forty hours, the average water flux for this mem-
brane remained at about 36 gfd for the remainder of the .
test (150 hours). The cellulose acetate ultrathin mem-
brane exhibited an average flux of 34 gfd in the third test
section. For both of these membranes, the actual water
fluxes from the effluent varied from 54 gfd to 30 gfd.
Much lower average fluxes were observed for the cellulose
acetate methyl sulfonate and the asymmetric cellulose
acetate control membrane (20 gfd and 13 gfd, respectively,
over the last two-thirds of the test).
• Fouling Flux Decline
The flux-versus-time data for two of the membranes are
given in Figures 8 and 9. In Figure 8 for the CMSOPSA
membrane, the initial water flux from the effluent in each
30
-------
Table 10. Long-Term Secondary Effluent Test Results
CO
Conditions:
Cell Flat
Secondary effluent Filtered through glass fiber mat and
Periodic cleaning procedure . 20 gm Biz /gal. water at 0 psig and
25°C circulated for 1. 5 hr every 20
hr during. test
Membrane
CMSOPSA (211-89A)
Cellulose Acetate
Methyl Sulfonate
(211-10C)
Cellulose Acetate
(E 398-10)*
Cellulose Acetate
(RO-97)
Deionized
Water Flux
(gfd)
Initial Final
54 42
37 22
50
16 13
Average Water Flux
Within Test Section (gfd)
1
43
30
39
14.5
2
40
24
37
13.5
3
35
20
34
. 13.5
4
36
21
_
13.2
5
36
21
_
13.2
6
34
18
-
13.2
7
36
19
_
13.2
Average
Percentage
Rejection
TDS
96
97
93
96
*An apparent membrane failure was observed after Test Section 3; the average rejections include only Test Sections 1,
Feed Data
[Average Concentration in Filtered Feed (mg/1)]
TSS** , ,TDS NH3(N) TOC
15 668 8.1 25.1
**TSS = 52 for unfiltered feed.
NH3
94
96
92
92
TOC
83
85
83
85
Final
Percent
Salt
Rejection
95
96
-
96
2, and 3.
-------
60
50
40
•o
H-
o>
* 30
u_
o>
o
20
10
1
DIW
(1)
Effluent
DIW
1
(2)
Effluent
1 1
(3)
D'W
1 1 1 1 1 1 1 1
(4) (5) (6) (7)
DIW
DIW
Effluent
(DIW = Deionized water)
* Membrane cleaned at this point
1
1
1
D1W
DIW
Effluent
1
1
1
10 20 30 40 50 60
70 80
Time (hr)
90 100 110 120 130 140 150
Figure 8. Water Flux Data from Secondary Effluent Test with Cellulose Methyl
Sulfonate O-Propyl Sulfonic Acid
-------
60
50
40
«^
-21
X
if 30
co B
00 5
20
1 I0
m
o
r
5> o
cy c
K
c
(0
jn
i i I i 1 1 I 1 I i i i i i
(DIW =Deionized water)
*Membrane cleaned at this point
"
-
— -
,DIW
^ (1) (2) (3) (4) (5) (6) (7)
Effluent DIW Effluent
* , i * * *
1 1 1 1 1 I.I II 1 III 1
) 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Time (hr)
Figure 9. Water Flux Data from Secondary Effluent Test with Commercial Acetate
A G-wmmo-f v»i r» l\/ToTV»Hr*Qno
-------
test section was always two or more gfd less than the
deionized water flux. The initial flux from the effluents
varied from 55 gfd to about 40 gfd. The water flux at the
end of the twenty-hour test sections varied from about
38 gfd to about 30 gfd. All the membranes had some
deposits on their surfaces when they were inspected at the
end of the test. The fouling flux decline can probably be
improved by further optimization of any or all of the
following variables: cell design, feed flow rate, and total
suspended solids removal.
The asymmetric cellulose acetate membrane control
exhibited virtually no fouling flux decline (Figure 9). It
is reasonable to assume that under the same conditions,
high-flux membranes will generally experience greater
fouling flux decline than low-flux membranes. Thus, with
high-flux membranes, there is the necessity of some
degree of suspended solids removal from a feed high in
total suspended solids.
• Long-Term Average Flux Decline
As shown in Table 10, all the membranes experienced some
long-term average flux decline. This long-term flux decline
was virtually nonexistent for all the membranes after the
second test section (about forty hours).
• Rejection Performance
The average percentage rejection is listed in Table 10.
As indicated earlier, the rejections for ultrathin cellulose
acetate were measured in Test Sections 1, 2, and 3, only.
The ammonia rejections were high, up to 96 percent. The
TDS rejections were more than adequate, above 95 percent
in most cases. The TOG rejections were observed at
84 ± 1 percent. This rejection performance was consistent
throughout the 150-hour test.
Conclusions
Ultrathin membranes exhibited considerably higher flux than the thicker,
commercially available, asymmetric membranes. For ultrathin membranes,
CMSOPSA was highly promising as a practical reverse osmosis membrane
for municipal wastewater treatment. Optimization of the structure of this
polymer for treatment of secondary effluent should result in higher fluxes
without decreasing its high rejection or increasing its estimated low
degradability.
34
-------
Preliminary Economic Analysis of Production Costs
Of Tubular Ultrathin Membrane/Support Composites
An economic analysis was made of process costs relating to production of
commercial quantities of ultrathin membrane/support composites for
reverse osmosis water desalination. This analysis did not include the costs
for the mechanical support tube, such as a fiber glass epoxy tube. A
modified Zevnik-Buchanan approach was used, which is based on process
capacity, process complexity, functional units of the process, and the
Nelson Refinery Construction Index. The final estimated selling price
included a 25-percent return on investment after taxes. The procedures
and data are given in Appendix C.
It is convenient to consider the tubular configuration because of its high
potential practicality in this application. The system of polysulfone support
liner and ultrathin membrane looks promising from a production cost view-
point. At a production capacity of 10 million tubes (one square foot of mem-
brane per tube) per year, the estimated price for a polysulfone support liner
and an ultrathin cellulose acetate membrane was 9.5 cents per tube. At a
production capacity of one million tubes per year, the price was estimated
to be 40 cents per tube.
One advantage of ultrathin membranes for reverse osmosis treatment of
wastewater is their (cellulose acetate and possibly others) ability to be
regenerated in situ. (2) Thus, the tube would not be discarded for membrane
replacement. The replacement costs of the membrane, including all process
and labor costs, would be somewhat less than the cost of the original mem-
brane and liner, 8.3 cents per tube. If the membrane could be replaced
five times the average cost for original membrane plus polysulfone support
and the five replacement membranes would be 8. 6 cents per tube.
Tubes with other ultrathin membranes, such as CMSOPSA, would have
virtually the same cost as the combination of cellulose acetate and poly-
sulfone given above. This would be true even if the cost of the new
polymer were as high as $50 per pound (Appendix C). The reason why
polymer cost has so little effect on total cost is the extremely small
weight of polymer required for a square foot of membrane. Thus, the
selling price is unchanged for all practical purposes when polymers with
an approximately 100-fold raw material cost increase are used.
35
-------
ACKNOWLEDGMENTS
This research was directed and carried out by Dr. L. T. Rozelle, Principal
Investigator; Mrs. B. R. Nelson, Research Chemist, who was responsible for
the analytical work; Dr. E. M. Scattergood, Research Engineer, who was
responsible for the reverse osmosis studies; Mr. J. E. Cadotte, Research
Chemist, who was responsible for the new polymers and Miss Paulette
Johnston, Technical Assistant.
The support of this study by the Water Quality Office Contract No. 14-12-587,
and the advice and assistance of Dr. Carl A. Brunner of the Robert A. Taft
Water Research Center, the Project Officer, are gratefully acknowledged.
37
-------
REFERENCES
1. Rozelle, L. T., Cadotte, J. E., King, W. L., Senechal, A. J., and
Nelson, B. R., Development of New Reverse Osmosis Membranes for
Desalination, ' Office of Saline Water Research and Development
Progress Report, in press.
2. 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 Report No. 531, U. S. Government
Printing Office, Washington, D. C. (June 1970).
3. Francis, P. S., and Cadotte, J. E., Second Report on the Fabri-
cation and Evaluation of New Ultrathin Reverse Osmosis Membranes, Office
of Saline Water Research and Development Progress Report No. 247,
U.S. Government Printing Office, Washington, D. C. (April 1967).
4. Cadotte, J. E., Rozelle, L. T., Petersen, R. J., and Francis, P. S.,
"Water Transport Across Ultrathin Membranes of Mixed Cellulose
Ester and Ether Derivatives", Journal of Applied Polymer Science, Applied
Polymer Symposia No. 13, 73 (1970).
5. Rozelle, L. T., Cadotte, J. E., and McClure, D. J. , "Ultrathin
Cellulose Acetate Membranes for Water Desalination",
Journal of Applied Polymer Science, Applied Polymer Symposia No. 13, 61 (1970).
6. 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.
(June 1968).
7. Sherwood, T.K., Brian, P. L. T., Fischer, R.E., and Dresner, L.,
"Salt Concentration at Phase Boundaries in Desalination by Reverse
Osmosis," I&EC Fundamentals, 4, 113 (1965).
8. Vos, K. D. , Burris, F.O., Jr., 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).
9. Kraus, K.S., Phillips, H. O., Marcinkowsky, A.E., Johnson, J.S.,
and Shor, A. J. , "Hyperfiltration Studies, VI, Salt Rejection by
Dynamically-Formed Polyelectrolyte Membranes", Desalination, 1,
225 (1966).
39
-------
10. Dean, R.B., Claesson, S. , Gellerstedt, N. , and Boman, N. , "An
Electron Microscope Study of Colloids in Waste Water", Environmental
Sci. and Tech., 1, No. 2, 147 (February 1967).
11. Meiners, A. F. , etal.. An Investigation of Light Catalyzed
Chlorine Oxidation for Treatment of Waste Water, TWRC-3, U.S. Department
of Interior, Federal Water Pollution Control Administration (December
1968).
40
-------
APPENDIX A
Preparation Procedures of Pertinent Polymers
-------
APPENDIX A
PREPARATION PROCEDURES OF PERTINENT POLYMERS
Cellulose Methyl Sulfonate O-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
occasionally for six hours, then allowed to stand at room temperature for
four days. The etherified fiber was washed with methanol, 2-percent sul-
furic acid, and glacial acetic acid.
The fiber was then dispersed in 250 ml of glacial acetic acid. Twenty grams
of methyl sulfonic acid followed by 50 ml of acetic anhydride were added to
produce a clear, thin solution within twenty 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 Blender, washed, and dried at 100°C. The product
(17 g) was soluble in hot cyclohexanone.
Cellulose Acetate O-Propyl Sulfonic Acid (211-95A)
Cotton linters (18 g of Hercules A-2000) were swelled in 100 g of 20-percent
NaOH at -10°C for 24 hours. Propane sultone (3. 0 g) was blended with the
wet fiber, using a spatula. The mixture was stirred occasionally for six
hours, then allowed to stand at room temperature for four days. The etheri-
fied fiber was washed with methanol, 2-percent sulfuric acid, and glacial
acetic acid. The acetic acid wet fiber was diluted to 400 ml with acetic acid,
and 20 g of 10-percent H2SO4 in acetic acid were added. The reaction pro-
ceeded rapidly, after addition of 100 ml of acetic anhydride, to give a clear
solution in one hour. A mixture of 20 ml of water and 30 ml of acetic acid
was added to decompose the excess acetic anhydride and sulfate ester, and
mixing was continued 10 minutes. The solution was poured into two liters of
deionized water containing 20 g of sodium acetate. The bulky precipitate was
chopped in a Waring Blendor, filtered, washed, and dried at 80°C. The yield
was 34 g, and the product was soluble in cyclohexanone.
Cellulose Acetate Phosphate (211-94A)
Eastman cellulose acetate 398-10 (10 g) was dried at 105°C and dissolved in
100 ml of reagent-grade pyridine. The solution was cooled in an ice bath,
and 0. 45 g of phosphorus oxychloride was added dropwise. After stirring
20 minutes, the viscosity of the solution began to increase, indicating cross-
linking by the POClo. The reaction mixture was poured into 600 ml of
deionized water at IO°C and stirred for 30 minutes. The product was dis-
persed in a Waring Blendor, filtered, washed, and air dried. The yield was
10. 0 g.
A-l
-------
Cellulose Acetate Adduct of Methylol Phenol (211-83B)
A ten-percent solution of methylol phenol in acetic acid was made by reacting
18. 8 g of phenol (0. 2 mole) with 6.0 g of paraformaldehyde (0. 2 mole) in
163.2 g of acetic acid. Cotton linters (Hercules A-2000) were preactivated
by soaking in water for 24 hours. The ;fiber was filtered and washed three
times with glacial acetic acid. The acetic acid wet fiber was mixed with
200 g of glacial acetic acid and 10 g of 10-percent H2SO4 in acetic acid. The
acetylation reaction was carried, out by the addition of 50 ml of acetic ;
anhydride. After one hour of reaction time, excess acetic anhydride and
sulfate ester in the product were hydrolyzed by addition of 10 ml of water and
40 ml of acetic acid and mixing for 15 minutes. Fifty grams of-the 10-per cent
methylol phenol solution was then added, and the product was precipitated in
water, filtered, washed, and dried. The product (17 g) was soluble in
c yc lohexanone.
Cellulose Acetate Methyl Sulfonate (211-10C)
Ten grams of Eastman cellulose acetate 360-60 was 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 was 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. 0 g) was
soluble in cyclohexanone and was estimated to have an acetyl DS of 2. 1 and
a methane sulfonate DS of 0. 9.
A-2
-------
APPENDIX B
Analytical Methods
-------
APPENDIX B
ANALYTICAL METHODS
Total Organic Carbon
Combustion-Infrared Analysis
American Society for Testing and Materials, D2579, Par t 2 3 (1969)
This method covers the determination of organic carbon in the range from 2
to 150 mg/1 of carbon in water. The instrument used for the analysis was a
Beckman Model 915 total organic carbon analyzer. The analysis required
acid blowing of the samples prior to the organic carbon determination to
remove all of the inorganic carbon. Microsamples of the acid-blown solu-
tions were then injected into a heated catalytic packed tube in a stream of
purified air. Organic matter was oxidized to carbon dioxide, which was
measured by means of a nondispersive infrared stream analyzer.
Ammonia Nitrogen
Distillation and Titration
Association of Official Agricultural Chemists, 1 Oth ed. (1965)
This method covers the determination of ammonia nitrogen in the range from
0. 1 mg/1 to 20. 0 mg/1 as nitrogen. The sample was transferred to a Wagner
distillation apparatus and made1 alkaline with sodium hydroxide. The alkaline
solution was distilled, and the freed ammonia was absorbed in boric acid and
titrated.
Nitrate Nitrogen
Brucine Colorimetric Determination
Standard Methods for the Examination of Water and Wastewater,
12th ed. (1965)
This method covers the determination of nitrate nitrogen in the range of 0 to
11 mg/1 as nitrate. Brucine and nitrate react under acid conditions to form
a sulfur-yellow color of intensity proportional to the original nitrate concen-
tration. Photometric measurements were made on a Beckman Ratio Record-
ing Spectrophotometer, DK-2A.
B-l
-------
Orthophosphate
Photometric Amino Reduction Method
American Society for Testing and Materials, D - 515, Part 23 (1969)
This method covers the determination of orthophosphate in the ranges from
0. 5 to 20 mg/1 and from 0. 1 to 1. 0 mg/1 as'phosphate. Orthophosphate
reacts with ammonium molybdate in an acid medium to form a molybdo-
phosphate, which, in turn, is reduced'to a molybdenum-blue complex with
aminonaphthol sulfonic acid. The color is proportional in intensity to the
phosphate concentration. When a bizmuth salt is added to the sulfuric acid
reagent, the intensity of the blue color increases fourfold. Photometric
measurements were made on a Beckman Ratio Recording Spectrophotometer,
DK-2A.
Total Suspended Solids
Filtration
Standard Methods for the Examination of Water and Wastewater,
12th ed. (1965)
Suspended matter is determined by the filtration of the sample through a
0. 45-micron Millipore filter. The difference in the weight of the dried
filter before and after filtration is the total weight of suspended solids in
the sample.
Total Dissolved Solids
Evaporation
Standard Methods for the Examination of Water and Wastewater,
12th ed. (1965)
The sample is filtered through a 0.45-micron filter, and the filtrate
evaporated in a weighed dish. The difference in the weight J.s the
total dissolved solids in the sample.
B-2
-------
APPENDIX C
Data and Results of Preliminary Economic Analysis
-------
APPENDIX C
DATA AND RESULTS OF PRELIMINARY ECONOMIC ANALYSIS
Analysis Procedure
The procedure used for these preliminary cost analyses on the production of
ultrathin membrane reverse osmosis tubes was based on a modified* method
of Zevnik and Buchanan. (1)
The Zevnik-Buchanan approach required the input of four types of. informa-
tion:
1. The capacity of the proposed process;
2. An assessment of the process complexity;
3. A process diagram for determining the number x>f functional
units; and
4. A construction cost index.
From these, a total plant investment cost can be estimated. The above pro-
cedure was modified to give a cost per unit item of production based on
operation 365 calendar days per year. These modifications included assess-
ment of labor costs, raw material costs, utilities consumption, depreciation
of capital equipment, and general overhead expenses.
In this method of cost analysis, the capacity of the proposed process is
chosen, e.g., ten million square feet of composite membrane tubular units
per year.
The process complexity is related to the cost of capital equipment needed for
the process. A "complexity factor" is derived from process parameters
covering temperature, pressure, and need for corrosion-resistant metals and
is calculated from Equation C-l below:
(F + F + F )
CF = 2 x 10 T P (C-l)
where:
CF = complexity factor
F, = temperature factor
F = pressure factor
F_ = alloy factor
ci
^Private communication to North Star Research and Development Institute.
C-l
-------
The maximum temperature to be anticipated in this process is 200°F (660°
Rankine). From Zevnik and Buchanan, F, = 0. 02. This process will operate
at atmospheric pressure. From Zevnik and Buchanan, F '='0.
The alloy factor would also be zero for this process because plastic-lined
tanks would be used, and corrosion-resistant metals would not be required.
Then,
(F, + F + FJ
CF = 2 x 10 l p a
. 0. 02
= 2 x 10
- 2.09
For each process a schematic diagram was drawn, and the number of
functional units was determined. The investment cost per functional unit
was determined from Figure C-l. . The appropriate-complexity factor was
used at the given production capacity to determine the cost per functional unit
in the process. Figure C-l is based on a 1963 Nelson Refinery Construction
Index (NRC) of 202. Construction costs since then have risen to the extent
that the NRC Index now stands at 350. (2)
An assumption was used in applying Figure C-l from the original Zevnik
and Buchanan plot to membrane application. The production capacity of the
original plot was in units of pounds per year. In the case of membrane-tube
combination, the cost to produce one square foot was considered to be'equiv-
alent to one pound of commodity chemical. Although this assumption is
admittedly inexact, it is 'considered'sufficiently good for an order-of-
magnitude estimate.
The process investment was calculated by multiplying the number of func-
tional units by the cost per functional unit and by the NRC index ratio (based
on 202). The process investment was further multiplied by a factor of 1. 33
to make an allowance (•"-' for indirect investment costs for utilities and
general facilities (land, building—"utilities and off-sites"). The resulting
number then constituted the total investment.
By assuming a ten-year life for equipment, the investment cost was con-
verted to an operating cost. To this was added other usual operating costs
to obtain a total cost per unit of membrane area. The price for a 25-percent
return on investment was then calculated.
C-2
-------
^ 0.7
.2
u
c
u.
0>
Q.
2
JO
"o
o
<£>
o
0.4
O.I
-0.07
§
to
0.04
Vt
w
0.02
0.01
Complexity Factor
_L
i i i i i i i
20
I 2 4 7 10
Production Capacity, 10 Ft /Yr
Figiife C-l. Process Investment versus Capacity (NRG = 202)*
40
#From Zevnik and Buchanan
(1)
C-3
-------
Results
Tubes with Ultrathin Cellulose Acetate Membranes
All the reverse osmosis tubes discussed here have been considered to be
eight feet long with an inside diameter of one-half inch. This configuration
contains approximately one square foot of membrane area.
A schematic process diagram with seven functional units is shown in
Figure C-2. The process includes fabricating polysulfone support films and
ultrathin membranes. It also includes stripping of old membrane and recast-
ing new membranes in used tubes. Both operations would be carried out in a
tube manufacturing plant." Although only six functional units are required for
each operation, the plant would consist of seven units total.
'*
Original Tubes. The cost per tube (one square foot) of ultrathin cellulose
acetate /polysulfone tube composite was estimated to be 9.5 cents based on a
capacity of 10 million square feet per year. The data and calculations are
given in Table C-l. The estimates were made as follows:
Using Figure C-2 and a complexity factor of 2, investment costs per
functional unit were estimated at $50,,000. Based on a process con-
taining seven functional units and adjusting for an NRC of 350, the
process investment costs were calculated to be $810,000.
The raw material costs for Table C-l were obtained from. Tables C-2
and C-3. These tables show the costs for the porous polysulfone liner
(Table C-2), and ultrathin cellulose acetate membranes (Table C-3).
The chemical costs were obtained either from the Oil, Paint, and Drug
Reporter \^) or the manufacturers.
Tubes with Regenerated Cellulose Acetate Membranes. The price of regen-
erating an ultrathin membrane was'estimated to be 8.3 cents per square foot
with a 25-percent return on investment. This figure was obtained by sub-
tracting the cost of the polysulfone solution from the list of costs in
Table C-l and recalculating a new selling price. All costs except that of
polysulfone would be essentially the same for this case as for the case
presented in Table C-l when the work is done in the same plant.
If the ultrathin membranes in polysulfone tubes were regenerated five times,
the average cost for each of the six tubes (original plus five regenerations)
would be 8. 6 cents per tube. '••••
Tubes with Other Polysaccharide.Derivatives. • The effect on the production
costs of ultrathin membranes from more expensive polymers (such as
cellulose methyl sulfonate O-propyl sulfonic acid) are considered here.
C-4
-------
Support
Manufacturing
i
Store
and
Mix
o
i
01
Tube
Assembly
Membrane
Stripping
H
Membrane
Casting
Annealing
Assembly
- and
Packing
Figure C-2.
Process Diagram Showing Functional Units of
Reverse Osmosis Tube Manufacturing Plant
-------
Table C-l. Cost Estimate for Producing Ten Million Square Feet
per Year of Composite Polysulfone Support, Ultrathin
Cellulose Acetate Membrane
Plant Size:
Temperature:
Pressure:
Alloy:
10 million ft2/yr, 27, 400 ft2/day
200°F or 660°R
1 atm
None
Temperature Factor:
Pressure Factor:
Alloy Factor:
Complexity Factor:
$50,000 (NRC = 202)
Total Factor, F: 0. 02
Functional Units: 7
Cost per Functional Unit:
NRC Index: 350
3^n
Process Investment: 7 x 50,000 x-^jx 1.33 = $810,000
Operating Labor: 7 men at $4/hr
$270/day from Tables C-2 and C-3
20% of other production costs
5% of process investment per year
10% of process investment per year
100% of labor
6% of process investment per year
0.02
0
0
2
Raw Materials:
Utilities + Misc:
Maintenance:
Depreciation:
Overhead:
Interest:
Daily Costs:
Item
Labor + Overhead
Raw Materials
Maintenance + Depreciation + Interest
Utilities + Misc.
Contingency (5% of above)
Price per Unit Membrane Area:
Break-even price =
4^0 °° = 5.4
-------
Table C-2. Estimated Raw Material Costs of Polysulfone Tube Liners
Assume:
Film is cast from 15% by weight polysulfone and 85% dimethyl
formamide solution
Cost of polysulfone is $l/lb
Cost of DMF is 24$/lb
Film thickness is 2 mils
Specific gravity is 0. 4
Then cost of materials is:
(0. 15 x 100 t 0. 85 x 24) x Q. 002 x 62. 4 x 0. 4 = Q< /ft2 Qr $268/
U • J- 0 X J. ^
Table C-3. Estimated Raw Material Costs for Ultrathin Cellulose
Acetate Membrane
Assume:
Film is cast from 5% by weight CA and 95% ethyl acetate solution
Cost of CA is 70
-------
The raw material costs for these membranes are given in Tables C-4 and
C-5. It is expected that the cost of the new polymer would be between $5 and
$50 per pound.
From Tables C-4 and C-5, the total raw material cost for the tubes is
virtually the same when they contain either cellulose acetate or other poly-
saccharide derivative membranes. The raw material cost comprises the
only difference in the selling price of the tubes containing the two mem-
branes. Thus, the selling price is unchanged for all practical purposes
when polymers with an approximately 100-fold raw material cost increase
are used. Due to their ultrathin nature, the extremely small amounts of
the polymers that are necessary for the membranes in the tubes account
for their insignificant effect on the total selling price of the tubes.
Effect of Plant Capacity on Tube Cost. The effect of reducing tube production
to one million^units per year is shown in Table C-6. The greatly increased
cost of 40£/ft compared with 9. 5£/ft2 for ten million units per year is due
largely to the higher proportional cost of labor.
C-8
-------
Table C-4. Estimated Raw Material Cost of Ultrathin Membrane
of a Polysaccharide Derivative with a Cost of
Five Dollars per Pound
Assume;
Film is cast from 5% by weight derivative and 95% ethyl acetate
solution
Cost of derivative is $5/lb
Cost of ethyl acetate is 13£/lb
Membrane thickness is 1000 A
Specific gravity is 1. 3
Then cost of materials is:
(0. 05 x 500 + 0. 95 x 13) x 1000 x 10"8 x 62. 4 x 1. 3 n non^/^
0,05 x 2.54 x 12 0.(WO<5/it
Table C-5. Estimated Raw Material Cost of an Ultrathin Membrane
of a Polysaccharide Derivative with a Cost of Fifty
Dollars per Pound
Assume:
Film is cast from 5% by weight derivative and 95% ethyl acetate
solution
Cost of derivative is $50/lb
Cost of ethyl acetate is 13
-------
Table C-6. Cost Estimate for Producing One Million Square Feet
per Year of Composite Polysulfone Support, Ultrathin
Cellulose Acetate Membrane
Plant Size: 1.0 million ft2/yr, 2, 740 ft2 /day
Temperature: 200°F or 660°R Temperature Factor: 0.02
Pressure: 1 atm Pressure Factor: 0
Alloy: None Alloy Factor: 0
Total Factor, F: 0.02 Complexity Factor: 2
Functional Units: 7
Cost per Functional Unit: $15,000 (NRC = 202)
NRC Index: 350
Q Cf)
Process Investment: 7 x 15,000 X-T-X 1.33 = $240,000
Operating Labor: 7 men at $4/hr
Raw Materials: $2 7 /day
Utilities + Misc: 20% of other production costs
Maintenance: 5% of process investment per year
Depreciation: 10% of process investment per year
Overhead: 100% of labor
Interest: 6% of process investment per year
Daily Costs: Item
Labor + Overhead
Raw Materials
Maintenance + Depreciation + Interest
Utilities + Misc.
Contingency (5% of above)
Total _ 772
Price per Unit Membrane Area:
Break-even price = 77| *4Q°° = 28£/ft2
Increase in cost for 25% return after taxes = 240' 00° x 0> g X 10° = 12$ /ft2
(Taxes assumed at 50% of gross profit) 1. 0 x 10
c\
Selling price for 25% return after taxes = 40$ /ft
C-10
-------
REFERENCES
1. Zevnik, F., and Buchanan, R., "Generalized Correlation of Process
Investment", Chem. Eng. Prog., 59, No. 2, p. 70(1963).
2. Oil and Gas Journal (September 1970).
3. Oil, Paint and Drug Reporter (February 1970).
C-ll
-------
1
Accession Number
w
5
2
Subject Field & Group
05D
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
NORTH STAR RESEARCH AND DEVELOPMENT INSTITUTE
3100 Thirty-Eighth Avenue,South
Minneapolis. Minnesota 55406
Title
New and Ultrathin Membranes for Municipal Wastewater
Treatment by Reverse Osmosis
10
Authors)
Rozelle, L.T.
Scattergood, E.M.
Nelson, B.R.
Cadotte, J.E.
16
Project Designation
Program # 17020 EFA
21
Note
22
Citation
23
Descriptors (Starred First)
Wastewater Treatment*, Reverse Osmosis*, Semipermeable Membranes*,
Separation Techniques, Tertiary Treatment, Water Pollution Treatment*,
Municipal Water, Organic Compounds, Ammonia, Phosphates, Membranes*
25
Identifiers (Starred First)
Ultrathin Membranes*, Polysaccharide Derivatives*
27
Abstract
A series of new and ultrathin membranes with thicknesses from 250 to 5000 angstroms
(1 x 10~6 to 2 x 10~5 inch) and consisting of various polysaccharide mixed esters and
ethers were tested on microporous supporting films for improved reverse osmosis treatment
of municipal wastewaters. From the screening studies with secondary effluents, ultrathin
membranes prepared from two polymers (out of a total of 44) looked very promising:
cellulose methyl sulfonate 0-propyl sulfonic acid (CMSOPSA) and cellulose acetate
0-propyl sulfonic acid (CAOPSA). The CMSOPSA membrane was found to be significantly less
degradable by hydrolysis, and thus more stable than cellulose acetate.
Long-term (150 hours) testing of the CMSOPSA membrane with secondary effluent
resulted in an average water flux of 34 to 36 gfd over the last 100 hours, and rejections
of 96 percent for total dissolved solids, 94 percent for ammonia, and 83 percent for
total organic carbon. The long-term flux decline (not due to fouling) was virtually
zero over the last 100 hours of the test. A treatment with an enzyme-active laundry pre-
soaking product was found to be effective in cleaning the membranes and restoring the
flux to levels existing before fouling.
A preliminary economic analysis showed that, for a tubular configuration, a cost of
approximately 10 cents could be expected for one square foot of the composite of membrane
and TiHrrnpnrnng support- f-ilin. _ TrH fi mst was insensitive to polymer raw material costs.
Abstractor
L.T. Rozelle
Institution
North Star Research and Development Institute
WR:I02 (REV. JULY 1969)
WRSIC
SEND. WITH COPY OF DOCUMENT. TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR s-r —
WASHINGTON. D. C. 20240
* GPO: 1970-389-930
------- |