Separation of Hazardous Organics by Low Pressure Reverse Osmosis Membranes: Phase II, Final Report. Project Summary


               United States
               Environmental Protection
               Agency
Risk Reduction
Engineering Laboratory
Cincinnati, OH 45268
               Research and Development
EPA/600/S2-91/045  Jan. 1992
EPA       Project  Summary
                Separation of Hazardous
                Organics  by  Low  Pressure
                Reverse Osmosis  Membranes
                Phase  II,  Final  Report
               D. Bhattacharyya and M.E. Williams
                 Extensive experimental  studies
               showed that thin-film, composite mem-
               branes can be used effectively for the
               separation of selected hazardous or-
               ganic compounds. This waste treatment
               technique offers definite advantages in
               terms of. high solute separations at low
               pressures (<2 MPa) and broad pH op-
               erating range, and the use of charged
               membrane would allow the selective
               separation of some organlcs from feeds
               containing high salt concentrations. In
               addition, feed pre-ozonation of selected
               organlcs provided significant Improve-
               ment in flux and  rejection characteris-
               tics for both charged and uncharged
               membranes because of the formation
               of ionizable organic acid intermediates
               during the ozonation that did not inter-
               act as strongly with the membrane. The
               overall ozonation/membrane process ef-
               fectively produced permeate water of
               high quality while it minimized the vol-
               ume of waste that must be further
               treated.
                 This Project Summary was developed
               by EPA's Risk Reduction Engineering
               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
                 Chemical manufacturers generate mil-
               lions of tons of wastes containing various
               hazardous priority pollutants each  year.
               The development of alternative technolo-
               gies for the treatment of various hazard-
               ous wastes is becoming increasingly im-
               portant as concerns grow over their dis-
posal. These and other wastes, such as
leachate from unsecured disposal sites,
contain a wide variety of priority pollutants
such as pesticides, herbicides, RGB's,
chlorinated hydrocarbons, and heavy met-
als. Much of this waste is relatively dilute
and so must be concentrated  before fur-
ther treatment.
  Several  methods such as biological
treatment, stripping, and carbon adsorp-
tion have been used to treat dilute waste-
water. Ozonation has also been found to
be effective in oxidizing some hazardous
organics to less toxic compounds. Mem-
brane processes can  be used to purify
wastewater and produce a 20- to 50-fold
decrease in waste volumes that must be
treated with other processes such as in-
cineration or wet air oxidation.

Membrane Concepts and
Applications
  Membrane processes are gaining con-
siderable attention for the purification and
volume  reduction  of dilute hazardous
wastes. The development of low pressure
reverse osmosis (RO) membranes (such
as aromatic polyamide  and  sulfonated
polysulfone)  has resulted in  membrane
processes that give high water flux at low
pressures {< 2 MPa). These thin-film, com-
posite membranes can provide high sol-
ute separations and have definite advan-
tages in terms of energy savings, capital
cost, and broad pH (2 to 12) operating
range. The low pressure membranes can
simultaneously separate hazardous organ-
ics and inorganics. Membranes containing
charged groups can also selectively sepa-
rate ionizable compounds.
                                                              Printed on Recycled Paper

-------
Membrane performance is measured in
terms of membrane rejection (R), perme-
ate water flux (Jw), and extent of water
recovery (r), The rejection is a measure of
solute separation by the membrane and is
defined as
whore C and C, are the solute concentra-
tions in the permeate and feed streams,
respectively.
Thin-film, composite membranes have
been used to remove major pesticides,
carcinogenic substances, priority organic
pollutants, etc. However, although low
pressure polyamids membranes have suc-
cessfully separated chlorinated phenolics,
there was a water flux drop due to mem-
brane/solute interactions.

Objectives
This work deals with the use of low
pressure thin-film, composite membranes
for the concentration and separation of
selected chlorophenols (CP's) and
chtoroethanes (CE's) with and without feed
pre-ozonation. Two types of membranes
were used: aromatic polyamide (FilmTec
FT30*) and carboxylated (negatively-
charged) polyamide (FilmTec NF40) mem-
branes. Membrane feed pre-ozonation will
result in the formation of organic acid in-
termediates; this should result in less flux
drop and, since these compounds are ion-
izable, in high separations by charged as
well as uncharged membranes. Separa-
tion and flux characteristics of both charged
and uncharged membranes were studied
w'rth both non-ozonated and ozonated (se-
lected CP's and CE's) solutions to evalu-
ate the effectiveness of the ozonation/
membrane process for the enhanced re-
moval of priority organic pollutants.

Experimental
Membrane experiments were conducted
with a system (Figure 1) containing both a
batch and a continuous flow cell so that
the performance of two different mem-
branes could be studied at the same time.
The thin-film, composite membranes used
In the experiments were the NF40 and the
FT30-BW; the NF40 was placed in the
batch cell and the FT30 in the flow cell.
The pump shown in the system provided
flow of solution through the continuous
cell and mixing in the batch cell; com-
pressed nitrogen supplied the pressure
driving force for the system. The operat-
ing conditions were: system pressure (P™)
of 0.3 to 1.4 MPa, feed pH's of 3.0 to 9.4,
and system temperature of 24°C.
The procedure used for experiments in-
volving ozonation/membrane process is
outlined in Figure 2. Membrane feed solu-
tions were ozonated in a 1.8-L stirred re-
actor with a flow of 0.20 standard L/min
(SLPM) O? containing 2% ozone. Pre-
ozonation times ranged from 0 to 60 min.
After ozonation, solutions were mixed for
several hours to allow decomposition of
residual ozone. Membrane experiments
were then carried out with the ozonated
solutions.
Membrane performance was measured
in terms of flux drop from that of distilled
water flux (DWF) and solute rejection.
Membrane feed, concentrate, and perme-
ate samples were analyzed by Total Or-
ganic Carbon (TOC) and High Pressure
Liquid Chromatography (HPLC) direct in-
jection. TOC was measured with the use
of a Beckman Model 915-B Carbon Ana-
lyzer. HPLC analysis (phenolics) were per-
formed with a Varian 5000 liquid chro-
matograph with a MCH-5 column (reverse
phase octyldecylsilane on silica) and a
UV-50 variable detector at 220 and 280
nm.

Results and Discussion
Membrane separation of selected haz-
ardous organics with and without feed pre-
ozonation were investigated with the NF40
and FT30 membranes. Single component
studies were conducted with trichlorophe-
nol (TCP); multicomponent systems ex-
amined consisted of TCP/humic acid (HA)
mixtures and mixtures of CP,
dichlorophenol (DCP), TCP, trichloro-
ethane (TCE), and tetrachloroethane
(TTCE). A wide range of ozonation times
and pressures were studied, and since
many of the compounds studied were ion-
izable, several membrane feed pH values
were investigated. Membrane performance
results are presented in terms of solute
rejection and flux drop. The percent of
flux drop (at a particular pressure) is de-
fined as:
% Flux Drop = Distilled Water Flux - Flux with Watewater
Distilled Water Flux
spectively. The DWF of the FT30 (an aro-
matic polyamide membrane) was also 11
x 10-4 cm/s at 1.38 MPa; NaCI and Na SO4
rejections for this membrane were 96%
and 97%. The low NaCI rejection of the
NF40 illustrates the principal advantage
of the charged membrane over RO mem-
branes such as the FT30. The FT30 mem-
brane gives high rejections of most sol-
utes, whereas the NF40 membrane can
be used to selectively separate compounds
with different charges.

Studies with Trlchlorophenol
Membrane feed solutions of 50 mg/L
TCP were ozonated from 0 to 60 min. The
pH of the solutions decreased after
ozonation; this drop in pH indicated the
degradation of the TCP to intermediate
organic acid compounds. HPLC analysis
indicated that after 5 min of ozonation the
TCP concentration was reduced to 24.7
mg/L, and after 15 min, was
-------
-e
Pump
Continuous Flow Cell
Membrane Area = 140 crtf
Cooling
Coil
Permeate
gj _ valve

/^N — Pressure Gauge

j 7 I — Thermocouple

Figure 1. Schematic of system for membrane experiments.
Flowmeter
Batch Cell
Membrane Area = 55.0 crtf
Permeate
Compressed
Nitrogen
Ozonation of Solution
Ozonator
Stirred
Reactor
Mixing of
Solution to
Remove
Residual
Ozone
pH
Adjustment
(if necessary)
Membrane
Experiment
J~~l
Figure 2. Schematic of ozonation process forozonation/membrane experiments.

3
-------
20-

10
System: NF40 Membrane
SOmg/LTCP
DWF=10.8X10'4,cm/s
Membrano
FeedpH
03.3-3.6
08.8-9.4
O

O
i i Hii ii 11111 11111 1111 i |" " I' ' " I " " I " ' ' I' ' " I' " ' I " ' ' I ' "
0 5 10 15 20 25 30 35 40 45 50 55 60 65

Ozonation Time (mm)

Flgur* 3. Effect of ozonation on flux drop for the NF40 membrane with trichlorophenol.
Membrane
FeedpH
3.3-3.6
O 8.8-9.4
System: NF40 Membrane
SOmg/LTCP
P =1.38MPa
Temp.=24°C
0 5 10 15 20 25 30 35 40 45 SO 55 60 65

Ozonation Time (mm)
brane. Ozonation improved rejection since
the organic acids that were formed ion-
ized at lower pH's than does TCP and so
these were rejected better than TCP at
lower feed pH's.
Figure 5 shows the flux drop for the
FT30 membrane. Under non-ionized con-
ditions (feed pH 3.3 to 3.6), the TCP inter-
acted strongly with the membrane, caus-
ing a large flux drop. The drop in water
flux was greater than that for the NF40;
the charge on the NF40 weakened the
interactions between the TCP and the
membrane. As with the NF40, ozonation
reduced flux drop for the FT30 membrane
by reducing solute interactions with the
membrane. Under ionized conditions (feed
pH 8.8 to 9.4), the TCP and organic acids
formed during ozonation do not interact
with the FT30 membrane as strongly as
under non-ionized conditions and so the
flux drop was smaller. TOG rejections were
in the range 80% to 96%, and feed pH
and ozonation did not greatly affect TOC
rejection for the FT30 membrane; this
membrane did not depend on charge for
separation as does the NF40 membrane.

Studies with Trichlorophenol/
Humic Acids
Experiments to determine the effect of
water recovery on water flux and solute
rejection were conducted with mixtures of
50 mg/L TCP and 10 mg/L HA. HA's are
high molecular weight compounds that are
present in soils and so can be found in
groundwater containing hazardous organic
leachates. These compounds are highly
rejected by the membrane but can cause
large water flux drops resulting from ad-
sorption.
For both non-ozonated and ozonated
TCP/HA, no drop in flux was observed
even at high recoveries (80%); the TCP/
HA and the organic acids formed during
ozonation were ionized at the high feed
pH (9.1 to 9.3) and so little change in
water flux occurred. Rejection of the non-
ozonated TCP/HA decreased from 80% at
10% recovery to 64% at a recovery of
80%; however, for the ozonated solution
84% TOC rejection was maintained even
at 75% water recovery. The decrease in
TOC rejection for the non-ozonated TCP/
HA could have been the result of en-
hanced concentration of the HA at the
membrane surface. Since degradation of
the HA is expected during ozonation, HA
concentration on the membrane surface
did not increase for the ozonated solution,
and TOC rejection remained high.
F70w» 4. Effect of ozonation on TOC rejection for the NF40 membrane with trichlorophenol.

4
-------
30-
System: FT30 Membrane
SOmg/LTCP
DWF = 10.5 X1O4 cm/s
P =1.43 MPa
Temp. = 24?C
Membrane
FeedPH
O 3.3 - 3.6
08.8-9.4
0 5 10 15 20 25 30 35 40 45 50 55 60 65
Ozonation Time (min)

Figure 5. Effect ofozonation on flux drop for the FT30 membrane with trichlorophenol.
40

30 H

I"
58
10
System: NF40 Membrane
50mg/LCP, DCP, TCP
100mg/LTCE,TTCE
NoOzon.
60 Min. Ozon.
234 5 6 78 Q 10

FeedpH
Figure 6. Effect of feed pH andozonation on flux drop for the NF4O membrane with hazardous organic
mixtures.
Studies with Mixtures
Studies were performed with mixtures
of 50 mg/L CP, DCP, and TCP with 100
mg/L TCE and TTCE to determine the
effect of multicomponent systems on flux
and TOC rejections for ozonated and non-
ozonated solutions. The flux behavior of
the mixture with the NF40 was found to
be linear over the pressure range 0.34 to
1.38 MPa (50 to 200 psi), indicating the
absence of surface polarization phenom-
ena. Flux drops as a function of pH for the
NF40 membrane at 1.38 MPa are shown
in Figure 6. Flux improved substantially
for increase in feed pH for both the
ozonated and non-ozonated mixtures;
ozonation did not greatly improve flux
drops found for the mixture at a fixed feed
pH. However, Figure 7 shows that TOC
rejection was enhanced significantly for
the mixture after ozonation for 60 min.
While the non-ozonated mixture rejection
was almost constant for the different feed
pH's, it was increased to as high as 80%
for the ozonated mixture. The increase
was because of the formation of organic
acids that ionized and were rejected by
the membrane. Although the phenolics in
the mixture were ionizable, the pKa of
these compounds are much higher than
those of the intermediates, and also, the
CE's present were not ionizable. As a
result, the rejection of the non-ozonated
mixture did not increase over the pH range
studied. The organic acids formed after
ozonation, however, had much lower pKa's
and so are rejected by the charged mem-
brane.
The flux behavior of the mixture with
the FT30 membrane was also found to be
linear over the pressure range studied.
FT30 membrane flux drops are shown in
Figure 8 for an operating pressure of 1.43
MPa. Flux drop was also a strong func-
tion of feed pH for this membrane, and
feed pre-ozonation did improve flux drop
over the range of feed pH's studied. As
with the single component TCP system,
ozonation produced intermediates that did
not interact with the membrane to the
same extent as the mixture and so flux
was enhanced. TOC rejection slightly in-
creased with ozonation and feed pH, rang-
ing from 81% to 95%.

Overall Removal of
Trichlorophenol
Figure 9 illustrates an example calcula-
tion of an ozonation/membrane process
for the separation of hazardous organics;
TCP was used as the model compound.
After ozonation for 30 min the TCP con-
centration would be reduced to <0.2 mg/
L; the TOC of the feed solution would be
66 mg/L due to the formation of CO2 dur-
-------
o No Ozon.
A 60 Mm. Ozon..
System: NF40 Membrane
50 mg/LCP, DCP, TCP
1000 mg/LTCE, TTCE
P =1.38MPa
Temp. = 24°C
D-
-n-
i
3
6

FeedpH
10
FJgun 7. Effect offaedpH and ozonation on TOG rejection for the NF40 membrane with hazardous
organic mixtures.
SO-

45-

40-

35-

30,

25'

20-

75-

0
System: FT30 Membrane
50 mg/LCP, DCP, TCP
100 mg/LTCE, TTCE
DWF=10.9X10-*cm/s
P =1.43MPa
Temp. 24°C
O No Ozon.
A 60 Min. Ozon.
r
5
T~
6
T"
7
T

8
10
FeedpH
FlgunS. EffactoffaodpHandozonationonfluxdropfortheFTSOmembranewithhazardousorganic
mixtures.
ing the ozonation. For a membrane that
rejects 90% (either the NF40 or FT30),
the permeate contains only 6.6 mg/L TOO
(due primarily to organic acids); the TCP
concentration in the permeate would be
less than 0.2 mg/L. For 90% water recov-
ery, the concentrate TOC, also mostly due
to organic acids formed during ozonation,
would be 1194 mg/L. This greatly reduced
volume could be disposed of by incinera-
tion. The overall ozonation/membrane pro-
cess would produce permeate water of
high quality (overall TCP removal of
>99.8%) and greatly reduce the volume
and risk of waste that must be further
treated. Also, since the process has been
shown to be effective when using charged
membranes, selective separation of haz-
ardous organics from feeds containing high
salt concentrations would be possible and
allow high fluxes at low operating pres-
sures even for feeds with high osmotic
pressures.

Conclusions
Dilute hazardous organics were effec-
tively separated by two types of thin-film,
composite polyamide membranes. Also,
the separation characteristics of the or-
ganics by the membranes were improved
by feed pre-ozonatioh. The two mem-
branes studied, a charged nanofiltration
membrane (NF40) and a low pressure
(RO) membrane (FT30), had water fluxes
of 10 to 13 X 10" cm/s at 1.4 MPa; NaCI
and Na?SO4 rejections were 30% and 97%,
respectively, for the NF40 membrane and
96% and 97%, respectively, for the FT30
membrane.
For the ionizable compounds (CP's)
studied, flux drops were highly dependent
upon feed pH. The NF40 membrane flux
drop was 17.4% at feed pH 3 but <6% at
pH 9 for TCP and over 30% at pH 3 but
only 3.7% at pH 7.9 for a CP-CE mixture.
The FT30 membrane flux drop with TCP
decreased from 27.9% at pH 3 to 4.3% at
pH 9.4 and from 44.8% at pH 3.0 to <12%
at pH >7.9 for the CP-CE mixture. Feed
pre-ozonation reduced flux drop of both
membranes for TCP and the CP-CE mix-
ture below feed pH 6.
Rejections of the NF40 membrane in-
creased from 29.8% to over 70% for TCP
as feed pH was increased from 3.4 to >9;
TOC rejections of the CP-CE mixture by
the NF40 membrane were <15%. For the
same mixture, the FT30 membrane TOC
rejections were in the range 80% to 96%.
Ozonation improved the NF40 membrane
rejection to as high as 87.6% for TCP and
to over 80% TOC rejection for the CP-CE
mixture. HA's present in TCP solutions
did not affect separation characteristics of
-------
CO TOG mg = 34
Feed
100mg/LTOC
Volume m 1L
Solute = TCP
TOCmg = 100
Ozonation
(30 min)
66mg/LTOC
Volume = 1L
TOO mg = 66
TCP<0.2mg/L
w pH adjusted to ~ 8
Membrane
v>.y>.v;-v>-v^
5)5% Water Recovery
Concentrate
(primarily organic acids)
1194mg/LTOC
Volume = 0.05 L
TOC mg = 59.7 mg
20
gal
ft2day

at200psi
Permeate (90% rejection)
(primarily organic acids)
6.6 mg/L TOC
Volume = 0.95 L
TOC mg = 6.3
Overall TCP removal = > 99.8%
Overall TOC removal - [ CO2 formation] + [Membrane Separation]

- 93.7%
Figure 9. Example calculation for ozonation/mombrane process.
the two membranes. It was shown that
the ozonation/membrane process could re-
move >99.8% of the hazardous organic
TCP and 93.7% of the TOC.
The full report was submitted in fulfill-
ment of Cooperative Agreement No.
CR814491 by the University of Kentucky
Department of Chemical Engineering un-
der the sponsorship of the U.S. Environ-
mental Protection Agency.
•&U.S. GOVERNMENT PRINTING OFFICE: 1992 - 648-080/40133
-------
D. Bhattacharyya andM.E. Williams are with the University of Kentucky Department
of Chemical Engineering, Lexington, KY 40506-0046.
Richard P. Lauch is the EPA Project Officer (see below).
The complete report, entitled "Separation of Hazardous Organics by Low Pressure
Reverse Osmosis Membranes -Phase II, Final Report," (Order No. PB91 -234
625/AS; Cost: $26.00; 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 Project Officer can be contacted at:
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
United States
Environmental Protection
Agancy
Center for Environmental Research
Information
Cincinnati, OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA PERMIT NO. G-35
Official Business
Penalty for Private Use $300
EPA/600/S2-91/045
-------