United States
Environmental Protection
Agency
Hazardous Waste Engineering
Research Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S2-87/053 Sept. 1987
Project Summary
Separation of Dilute Hazardous
Organics by Low Pressure
Composite Membranes
D. Bhattacharyya, T. Barranger,
M. Jevtitch, and S. Greenleaf
The use of membrane processes for
waste purification and volume reduction
is gaining considerable attention in
many industries. For hazardous wastes
containing priority organics and salts,
reverse osmosis membranes can provide
simultaneous separation of both or-
ganics and inorganics. The industrial
development of non-cellulosic, (aro-
matic polyamide, sulfonated polysul-
fone, etc.) thin-film, composite mem-
branes has provided a means for reverse
osmosis treatment with high solute
separations and minimal compaction
problems.
The separation of dilute hazardous
organics was accomplished through this
project utilizing thin-film, composite,
aromatic polyamide membranes. This
technique offers advantages in terms of
high solute separation at low pressures
(1-2 MPa (145-290 psi)) and broad pH
operating ranges (pH 2 to 12). The
synthetic organic waste solutions used
in this study include polyaromatic
hydrocarbons (PAH), phenol, chloro-
phenols, nitrophenols, and chloroben-
zenes. The actual organic waste solution
was obtained from the site of a former
wood treatment processing plant in
Texas and was known to contain
chlorophenols. The membrane showed
excellent stability over long periods of
time. Standard NaCI rejections were
97-99% and the average pure water
flux at 2.068 MPa (300 psi) was about
14x10* cm3/cm2s (5.5x10 4 inVin2
S).
This Protect Summary was developed
by EPA's Hazardous Waste Engineering
Research Laboratory, Cincinnati, OH, to
announce key findings of the research
project that Is fully documented In a
separate report of the same title (see
Project Report ordering Information at
back).
Introduction
Industries engaged in organic chemical
manufacturing use contact process water,
and the spent aqueous wastes often con-
tain various hazardous priority pollutants.
Of the 272-363 million metric tons (300
to 400 million short tons) of industrial
solid wastes generated in the United
States each year, it is estimated that 38
million metric tons (42 million short tons)
are hazardous. Sixty-two percent of this
waste is generated by the chemical pro-
ducts industry. Another source of aqueous
hazardous waste is from unsecured in-
dustrial waste storage and from disposal
site leachate. These contaminants are
diverse in composition and concentration,
and contain a broad spectrum of priority
organics. Generally, the term hazardous
or toxic organics includes synthetic or-
ganic compounds such as pesticides,
herbicides, PCBs, and chlorinated hydro-
carbons. Many contaminant streams are
relatively dilute and thus a concentration
step prior to detoxification or disposal
may be necessary.
There are several techniques that have
been used to treat dilute aqueous wastes.
These include biological treatment,
chemical coagulation, carbon adsorption,
resin adsorption, stripping, and membrane
processes. Using membrane processes
for waste purification is gaining con-
-------
siderable attention in many industries.
Membrane transport mechanisms range
from molecular diffusion in solids that
are liquid-like to tortuous viscous flow
through microporous polymers. The three
major membrane processes for water and
wastewater treatment are electrodialysis,
reverse osmosis, and ultrafiltration.
Reverse osmosis can remove salts, or-
ganics and ionic species. Ultrafiltration
removes solute species primarily based
on molecule sizes of 10-200 A (3.9 x 108
- 7.9 x 107 inches). Electrodialysis
separation is based on ionic charge of the
solute. The reverse osmosis process
separates solutes based on relative dif-
fustvities of the solutes through the
membrane (molecule size of 2-10 A (7.9
x 10"9 - 3.9 x 108 inches)) at a given
pressure differential.
Membrane Separation Concepts
Membrane processes are generally
evaluated in terms of three parameters:
membrane rejection (R), permeate-water
flux (Jw), and extent of water recovery (r).
The membrane rejection parameter, R, is
a measure of the extent of solute
separation,
R =
(D
in which Cp and C, are the permeate and
feed-steam solute concentrations, respec-
tively. Primary separation of solutes
occurs at the thin-film (skin) barrier layer.
Thin-film membranes result in higher
flux at pressures considerably less than
asymmetric cellulose-acetate membranes.
A number of models have been devel-
oped to describe the transport of solute
and solvent through membranes. A
commonly used model is the solution-
diffusion model. The water and solute
fluxes (under a chemical-potential driving
force) are given by:
water, JW = A(AP- ATT) and; (2)
solute, Js = B(AC); (3)
in which (AP - ATT) is the net trans-
membrane pressure, and "A" is the
membrane permeability (function of
temperature) constant. In Equation 3, "B"
and "C" are solute permeability (function
of solute-distribution coefficient between
solution phase and membrane phase)
and concentration gradient between the
membrane surface and the permeate,
respectively. In the case of negligible
concentration polarization, AC and ATT
become the concentration and osmotic
pressure difference between the bulk
solution and the permeate, respectively.
Surface Force-Pore Flow
The model developed by Sourirajan
gives a better understanding of the rejec-
tion and flux phenomena. The negative
and positive adsorption of solute at the
membrane-solution interface arises from
net repulsive or attractive forces acting
on the solute from the adjacent membrane
surface. The model admits that a layer of
water is preferentially adsorbed at the
pore wall. In some cases this layer of
water can be displaced by some molecules
of solute exerting stronger adsorption
forces toward the pore wall. For example,
the Stoke's radius of water and phenol
are 0.87A and 2.1 A (3 4 x 10 9 - 8.3 x
109 inches), respectively. A layer of
phenolics displacing the water layer will
definitely cause a water flux drop by
reducing the available path of the fluid.
This phenomenon occurs at low pH under
no ionization of the solutes.
In order to predict the separation of
organics by a membrane, one needs to
know the pore distribution of the mem-
brane. The skin pore distribution of the
membrane used in this study was mea-
sured by C02 (217°K) and by N2 (77°K)
gas adsorption technique. Utilizing the
pore distribution, and a new calculation
technique for the simultaneous solution
of the radial velocity profile of the solvent
through the pores, the solute concentra-
tion in the product water and the inter-
action parameters were applied to
compute rejection of phenolics. Figure 1
shows excellent agreement between cal-
culated and experimental results for a
typical run involving nitrophenols.
Objectives
The development of a low pressure
membrane (noncellulosic composite
membranes) process to concentrate
selected priority pollutants from hazardous
wastes will substantially improve con-
ventional destruction techniques. This
work deals with the use of thin-film,
composite membranes for concentration
and separation of pollutants from aqueous
waste streams. Spiked waste streams
utilized included PAHs, phenolics, chloro-
and nitrophenols, chlorinated benzenes,
phthalates and heavy metals. An actual
waste stream was also collected for treat-
ment in the membrane system. Separation
of pollutants and operation of the mem-
brane system were evaluated as a func-
tion of system pressure, flow rate, and
input waste concentration.
Experimental Procedure
Membrane studies were conducted in
batch, continuous thin channel, and con-
tinuous spiral-wound modules (Figures 2
and 3). The batch operating condition;
were 1400-1800 ml of feed solution, t.
system pressure (AP) of 1.38 - 2.07 MPe
(200-300 psi) and pH = 4.5-11.8. The
continuous operating conditions were
AP = 0.69-2.07 MPa (100-300 psi)
Reynold's number (Re) = 4000-9000, anc
pH = 3.3 - 11.8. For both the batch anc
continuous systems standard distillec
water flux and salt (NaCI) rejection were
obtained prior to experiments with an\
hazardous organic compounds. The
membranes used in this study were made
of aromatic polyamide. For each experi-
mental run, samples of feed, concentrate,
and permeate were collected, properly
stored, and analyzed.
Membrane feed, concentrate, anc
i permeate samples were analyzed in terms
of total organic carbon (TOC) (direct in-
jection), high performance liquid chro-
matography (HPLC) (direct injection 01
after solvent extraction if below detection
limit), and gas chromatography (GC) (after
solvent extraction-concentration). The
reproducibility and recovery of the solvent
extraction-concentration step were
checked with known synthetic solutions
and spiked samples. The objectives of the
HPLC analysis by reverse phase columns
were two-fold: establishment of mem-
brane output concentrations and correla-
tion of membrane rejection behavior with
HPLC elution times. Previous studies have
indicated an increase in rejection with
HPLC elution time.
Results
Membrane separation (batch cell and
continuous unit) of selected classes ol
priority pollutant mixtures was studied at
0.69 - 2.07 MPa (100 to 300 psi). The
batch cell study focused on the mixtures
of selected sparingly soluble PAH com-
pounds (naphthalene, anthracene,
phenanthrene), phenols, phthalates,
chlorobenzenes, and a field-collected,
contaminated ground-water sample. The
continuous unit study focused on the
mixtures of selected chlorophenols and
nitrophenols. One run involving phenol
and a salt mixture was also conducted.
For chloro- and nitro-phenols, a wide
range of pH values was selected to estab-
lish the rejection behavior of nonionized
and ionized species. Membrane stability
for both the continuous and batch units
was checked with standard NaCI runs
and with distilled water. The membrane
shows a 15% drop in distilled water flu>
with standard NaCI rejections remaining
constant at 97-98%, indicating gooc
membrane stability.
-------
7000
900
800
700
600
= - 500
400
300
200
10,0
00
Calculated data
Symbols, experimental data
2,4-DNP — pH - 3 0
Rett = 10.000
_L
000 250 500 750 1000 1250 1500 1750 20.00 22.50 2500
Pressure, N/rrf x 10~3
Figure 1. Effect of pressure on experimental and calculated re/ections of single solutions
of 2-NP, 2,4-DNP
Feed Solution
Inlet
• Cell Volume 1900cm
Permeate
Figure 2 Batch membrane unit
Separation of naphthalene (solubility
20-22 mg/l), anthracene (solubility 0 12
mg/l), phenanthrene (solubility 0.58
mg/l), and dimethylphthalate (solubility
278 mg/l) was carried out in the batch
cell. Rejection of dimethylphthalate was
Membrane Area
Tank
about 97%, and rejection of naphthalene
was 98.0% For the higher molecular
weight anthracene and phenanthrene,
which are chemically similar to naphtha-
lene, the rejections were 98-99%. The
flux drops with these compounds were
only 3-5%. The material balance analysis
of PAH compounds showed significant
loss of these compounds through the
test, probably due to adsorption on the
membrane.
Additional experimental runs were
performed on a mixture of chlorophenols
(phenol, 2-chlorophenol (2-CP, 2,4-
dichlorophenol (2,4-DCP), 2,4,6-trichloro-
phenol (2,4,6-TCP), and 4-chlorocresol
(4-CCR)), and nitrophenols (2-nitrophenol
(2-NP) and 4-nitrophenol (4-IMP)). It is
important to emphasize at this point that
phenolics are ionizable species. As such,
the pH values of various mixtures were
varied from pH = 4.5 to pH = 11.8 to
observe the separation characteristics at
various ionization conditions. At solution
pH > 11 essentially all species are 100%
ionized.
Experimental runs involving the chloro-
phenol mixture were performed at pH
values of 5, 9, and 11. In each chloro-
phenol experimental run involving high
pH, there was a smaller drop in permeate
flux than in low-pH runs. The rejection of
the individual species increases with pH
value and with each chloro substitution
on the phenol molecule. Similar runs
with nitrophenol mixtures at high pH (pH
= 11.5) showed very high rejections of
phenol (98.3%), 2-NP (99.3%), and 4-NP
(99.1%).
The chlorinated benzenes were run in
the batch cell as a mixture of chloro-
benzene (CB), 1,4-dichlorobenzene (1,4-
DB, and 1,2,4-trichlorobenzene (1,2,4-
TCB). Similar mixtures were tested under
stirred and quiescent conditions. Both
tests were taken to at least 80% recovery
of the feed solution and both showed
rejections ranging from 67.6 to 99.7%.
Experiments with contaminated water
from a wood processing plant were con-
ducted in the batch reactor at pH 7 and at
pH 11. The various phenol and PAH com-
pounds in the wastewater performed quite
similarly to the compounds tested in the
synthetic waste mixtures. Total rejection
of organics in the waste stream ranged
from 87% to 97%.
Continuous runs were made with
chloro- and nitro-phenol mixtures in the
turbulent flow regime. High flux drop
was observed under non-ionization condi-
tions (pH = 3.3) of the various solutes.
These results were obtained even under
a high Reynolds number (9000). The
behavior was similar to that observed in
the batch cell.
To establish the flux drop and rejection
phenomena, various types of mixtures
were run using thin channel cells. Runs
-------
Low
Pressure
Pump
1
Low Pressure Pump
55 Gal Drum
Sand Filter
Receiving Tank
Cartridge
Filter
. F/owmeter
By-Pass
Stream
Sample Flushing
Tank Tank
'3
High Pressure Pump
Concentrate
H.
Flowmeter
Permeate
Figure 3. Membrane flow diagram of the continuous unit operation
were conducted with mixtures of phenol,
2-CP, 2,4-DCP and 4-CCR containing a
total molar concentration of 1.8 mM. At
300 psi the flux drop was 31%. With
2,4,6-TCP in the mixture the flux drop
was 48%. A solution of 2,4,6-TCP (1.6
mM) indicated a flux drop of 35%. The
flux behavior of the chlorophenol mixture
(phenol, 20CP, 2,4DCP, 4-CCR) over the
range 0.69 - 2.07 mPa (100 psi - 300 psi)
was linear with P, thus indicating the
absence of surface polarization
phenomena.
To understand the effect of multi-
component systems on the flux behavior,
mixtures of phenol and 2-CP (46.6 ppm
and 54.8 ppm), phenol and 2,4-DCP (45.5
ppm and 62.8 ppm) and 2-CP and 2,4-
DCP (55.0 ppm and 71.9 ppm) were used.
The order of flux drop compared with
double distilled water (DDW) flux was
13%, 24%, and 25%, respectively. It can
be concluded that 2,4,6-TCP and 4-CCR
are the two compounds causing the most
flux decrease. For any type of mixture
under non-ionization conditions, the re-
jections of the various chlorophenols were
always of the order: Rphen0| < R2.CP < R2.4-
DCP < R2.4.6 TCP- Tnis sequence was also
the same as the HPLC elution time pattern
in a reverse phase (C18) column.
Conclusions
This study of thin-film composite mem-
branes for the separation of selected
classes of hazardous organic compounds
has proven quite effectively the benefits
of such a process. This particular waste
treatment technique offers definite ad-
vantages in terms of high solute separa-
tion at low pressures, insignificant
compaction problems, and broad pH
operating ranges (pH 2 to 12) The
aromatic polyamide membrane showed
excellent stability over long periods of
operating time with respect to permeate
water flux and rejection quality A surface
force-pore flow model was utilized to
predict rejections of various phenolic
compounds.
-------
D. Bhattacharyya, Theresa Barranger, Milan Jevtitch, and Suzanne Greenleaf
are with Department of Chemical Engineering at the University of Kentucky,
Lexington, KY 40506-0046.
John F. Martin is the EPA Project Officer (see below)
The complete report, entitled "Separation of Dilute Hazardous Organics by Low
Pressure Composite Membranes," (Order No. PB 87-214 870/AS; Cost:
$18.95, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Officer can be contacted at:
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
";°i t 0 .3 I z
Official Business
Penalty for Private Use $300
EPA/600/S2-87/053
0000329 PS
u s
*
CHICAGO
------- |