EPA/600/R-92/035
March 1992
SEPARATION OF HAZARDOUS OMGANICS BY LOW PRESSURE
MEMBRANES: TREATMENT OF SOIL-WASH RINSE-WATER LEACHATES
by
D, Bhattacharyya and A. Kothari
Department of Chemical Engineering
University of Kentucky
Lexington, Kentucky 40506-0046
Cooperative Agreement No. CR814491
Project Officer
Richard P. Lauch
Water and Hazardous Waste Treatment Research Division
Risk Reduction Engineering Laboratory
Cincinnati, OMo 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268 .
-------
DISCLAIMER
The information in this document has been fended wholly or In pail by the United States
Environmental Protection Agency under assistance agreement number CR814491 to the University of
Kentucky Department of Chemical Engineering. It has been subject to the Agency's peer and
, administrative review and has been approved for publication as an EPA document Mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial products and practices
frequently cany with them the increased generation of materials that, if improperly dealt with, can
threaten both public health and the environment. The U.S. Environmental Protection Agency is chained
by Congress with protecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the agency strives to formulated and implement actions leading to a compatible
balance between human activities and the ability of natural systems to support and nurture life. These
laws direct the EPA to perform research to define our environmental problems, measure the impacts,
and search for solutions.
The Risk Reduction Engineering Laboratoiy is responsible for planning, implementing, and
managing research, development, and demonstration programs to provide an authoritative, defensible
engineering basis in support of the policies, programs, and regulations of the EPA with respect to
drinking water, wastewater, pesticides, toxic substances, solid and hazardous wastes, and
Superfund-related activities. This publication is one of the products of that research and provides a vital
communication Knk between the researcher and the user community.
This report describes the evaluation of low pressure composite membranes for the purification of
hazardous wastes in soil-wash rinse water leachates. SARM (Synthetic Analytical Reference Matrix)
soils were used for this work. The membrane performance was studied with leachate from distilled
water wash, surfactant wash and ehelant wash. In addition, feed pie-ozonation was'shown to provide
significant improvement in membrane water flux. The effect of turbidity constituents was also studied.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
-iii -
-------
ABSTRACT
Soil washing is a promising technology for treatment of contaminated soils. In the present work,
low pressure thin-film composite membranes were evaluated to treat the soil-wash leachates so that the
treated water can again be recycled back to the soil washing step. Experiments were done with SARM
(Synthetic Analytical Reference Matrix) soils. The FT30 membranes used in the study were stable over
extended operating periods. Membrane perfonnance was evaluated with leachates obtained from
different wash solutions. In a run after distilled water wash, for a recovery of 17-22 %, the permeate
flux drop was 24-33 %, TOC rejections were 80-85 %, conductivity rejections were 94-96 %, heavy
metal rejections were 92-98 %, and amount adsoibed on the membrane was 20-25 %. Hie use of EDTA
was found to enhance heavy metals removal from the soil and in the membrane run there was a 14 %
decrease in amount adsorbed on the membrane and no increase in flux drop. There was a 5 % decrease
in amount adsorbed on the membrane for leachates from surfactant wash. The presence of fine soil
suspensions also decreased the flux drop and unaccounted TOC. At high water recovery (80 %), there
was a 27 % flux drop due to the osmotic pressure effects of total dissolved solids which were highly
rejected by the membrane. There was some loss in TOC with the precipitates of heavy metals in the
leachate at ph 8 (28 % decrease in TOC). GC/MS analysis was done for specific organic compounds in
feed and permeate to investigate relative membrane separations. Membrane rejections were high for
pentacMorophenol (> 98%), 2,4-dinitrophenol (> 98 %), ethylbenzene (> 97 %). 4-aminobiphenyl (> 93
%), xylene (>81 %), and cWoroaniline (> 90%). Rejections were also high for Pb, Zn, Ni, Cu (94-98
%). Feed preozonation resulted in only 5 % flux drop for a recovery of 20 %. Solution-diffusion model
was modified to include an adsorption resistance term in water flux and this term was correlated with
bulk concentration using Freundlich isotherm. This correlation was then used to predict water flux drop
at different bulk concentrations or to predict water flux at different recoveries. This work was
conducted under EPA Cooperative Agreement No. CR814491.
- iv -
-------
CONTENTS
Foreword..... • ill
Abstract..... iv
Contents v
Figures viii
Tables.. . xi
Abbreviations and Symbols xiii
SECTION 1: INTRODUCTION ...1
SECTION 2: CONCLUSIONS 3
SECTION 3: LITERATURE REVIEW 6
Overview...... ...6
In Sim Treatment Technologies for Contaminated Soils 6
Soil Washing / Soil Hushing ..6
Solidification / Stabilization Techniques ..7
In Situ Vitrification ....7
Chemical Oxidation and Reduction........... ......7
Chemical Dechlorination J
Polymerization .........8
Biological Processes ....8
Thermal Treatment 8
Photolysis .8
Soil Vapor Extraction ,.8
Radio Frequency Heating . 9
Membrane Processes for Hazardous Waste Treatment 9
Reverse Osmosis 9
Pervaporation 10
Nanofiltration II
Ultrafiltration 11
Overview of SARM Soils and Soil Washing 11
Soil Washing Technology ...12
Characteristics of SARM Soils ......12
SECTION 4: REVERSE OSMOSIS THEORY 18
Overview ......18
Basic Concepts ...18
Osmosis and Osmotic Pressure 18
Reverse Osmosis and Effective Pressure Driving Force ..........18
Terminology in a RO Process.... .18
Concentration Polarization and Membrane Fouling ..20
v
-------
Thin-Film Composite Membranes 20
Membrane Transport Mechanisms and Models.. . 23
Solution-Diffusion Mechanism 23
Solution-Diffusion Model .23
Solution-Diffusion-Imperfection Model 24
Extended Solution-Diffusion-Model . ........24
Preferential Soiption-Capillary Flow Mechanism ......24
Preferential Sorption-Capillary Flow Model . 24
Surface Force-Pore Flow Model .........25
Irreversible Thermodynamics Models ..........25
SECTION 5: EXPERIMENTAL EQUIPMENT AND PROCEDURES 26
Overview .26
Soil Washing Experiments • ............26
Soils Used.. 26
Wash Solutions Used ....26
Distilled Water —26
Surfactant Solution 26
EDTA Solution 29
Soil Washing Procedure 29
Vacuum Filtration 29
pH Control 29
Ozonation Procedure.... ..29
Membrane Experiments 31
Membranes Used ......31
Apparatus and Procedure for Batch System ......31
Apparatus and Procedure for Continuous System 31
Membrane Cleaning .....34
SECTION 6: ANALYTICAL EQUIPMENT AND PROCEDURES 35
Overview 35
Apparatus and Procedure for TOC Analysis ...........35
Apparatus and Procedure for Metal Analysis 35
Apparatus and Procedure for GC/MS Analysis 36
SECTION 7: QUALITY ASSURANCE AND QUALITY CONTROL... .....38
Overview .....38
Sample Collection and Sample Custody ...38
Calibration 38
Data Reduction, Validation and Reporting... ....41
Data Quality Indicators .........41
Arithmetic Mean. 41
Standard Deviation 41
Precision 41
Accuracy 45
Linear Regression Statistics ......45
vi
-------
SECTION 8: RESULTS AND DISCUSSION 46
Soil Washing Studies ....46
Soil Washing with Surfactant Solution..... 46
Soil Washing with EDTA Solution . 48
Ozonation of SARM-Water Slurry 48
Membrane Studies 54
Membrane Stability... .....54
Studies with Leachate of Different SARM Soils. .....54
Studies with Two Stage Wash Leachate 61
Studies with Turbidity Constituents 70
Studies with Leachate of Different pH Wash 70
Studies with High Water Recovery ......77
Studies with Leachate after Surfactant Wash 83
Studies with Leachate' after EDTA Wash 83
Studies with Leachate Spiked with Model Compounds 89
Studies with Leachate at Different pH Values 98
Studies after Leachate Ozonation 98
Studies at Different Flow Rates 104
Summary ..110
SECTION 9: MEMBRANE WATER FLUX PREDICTION AND
CORRELATION ...112
Overview.... ...........112
Adsoiption Resistance Term 112
Correlation of Adsorption Term with Bulk Concentration 113
Prediction of Membrane Water Flux 116
REFERENCES .........120
vii
-------
FIGURES
3.1 Schematic of a typical washing system . 13
3.2 Soil washing effectiveness of water and surfactant solutions for
different size fractions of SARMIV ..16
3.3 Soil washing effectiveness of water and ehelant solutions for
different size fractions of SARM IV 17
4.1 Description of a reverse osmosis membrane process 19
4.2 Description of concentration polarization ' 21
4.3 FT 30 membrane: cross-section and chemical composition ...............22
5.1 Schematic of the overall experimental plan. ..........27
5.2 Experimental set-up for ozonation-membrane experiments ...........30
5.3 Experimental set-up for membrane experiments in batch system 32
5.4 Experimental set-up for membrane experiments in continuous system 33
6.1 Extraction procedure for GC/MS analysis 37
7.1 . Calibration curve for copper analysis .......39
7.2 Chromatographs from two injections of the same sample for
a typical GC/MS analysis 40
7.3 Reporting scheme for data collection, storage, validation and analysis 42
8.1 Effect of wash solution on teachability of Cu, Ni, Pb and Zn 50
8.2 TOC teachability from clean SARM and SARM IV with ozonation time 53
8.3 Distilled water flux with operating time for FT 30 (membrane 1) ......55
8.4 NaCi rejections with operating time for FT-30 (membrane 1) 56
8.5 Distilled water flux with operating time for FT 30 (membrane 2) 57
8.6 NaCl rejections with operating time for FT-30 (membrane 1) 58
8.7 Permeate flux vs. time for SARM I leachate as membrane feed... 62
8.8 Permeate flux vs. time for SARM II leachate as membrane feed.... 63
viii
-------
8.9 Permeate flux vs. time for SARMIV leachate as membrane feed ....64
8.10 Permeate flux vs. time for SARM IV leachate from single
and double stage washing... 67
8.11 GC/MS ehromatograph of pentaehlorophenol and bisphthalate for
feed (double stage wash leachate of SARM IV) and permeate 68
8.12 Permeate flux vs. time for filtered and non-filtered leachate of SARM IV.... 73
- 8,13 Permeate flux vs. time for SARM IV leachate washed at different pH . 76
8.14 Permeate flux as a function of recovery for SARM IV leachate as
membrane feed ....80
8.15 TOC rejection as a function of recovery for SARM IV leachate as
membrane feed 81
8.16 Relative concentration of pentaehlorophenol in feed and permeate
of 20 and 80 % recovery run (SARM IV-distilled water wash) .82
8.17 Permeate flux vs. time for leachate after washing SARM IV
with 0.04 % surfactant solution 87
8.18 Relative concentration of pentaehlorophenol in feed and permeate
for distilled water and surfactant wash runs ....88
8.19 Permeate flux vs. time for leachate after washing SARM IV with
1.03 mol/1 of Na4F.DTA.4H20 (1000 mg/1 Versene 100EP) 93
8.20 Relative concentration of pentaehlorophenol in feed and permeate
for distilled water and EDTA wash runs... 94
8.21 Permeate flux vs. time for leachate spiked with 50 mg/14-ehloroaniline,
50 mg/14-aminobiphenyl, 75 mg/12,4-dinitrophenol
(SARM IV-distilled water wash) 96
8.22 GC/MS chromatographs of 4-chloroaniline, 4-aminobiphenyl,
pentaehlorophenol, and 2,4-dinitrophenol (spiked in SARM IV
distilled water wash) in feed and permeate.. 97
. 8.23 Permeate flux vs. time for leachate at different pH values
(SARM IV distilled water wash) 101
8.24 Permeate flux vs. time for leachate ozonated for different times
(SARM IV - distilled water wash) 105
8.25 GC/MS chromatographs of xylene, ethylbenzene, and styrene in
feed and permeate before ozonation (SARM IV -D.W. wash) 106
8.26 GC/MS chromatographs of xylene, ethylbenzene, and styrene in
feed after ozonation (SARM IV -D.W. wash) 107
ix
-------
8.27 Permeate flux vs. time for the continuous nm with suspended solids
at different flow rates (SARMIV - distilled water wash) 109
8.23 Summary of membrane performance with different leachates (SARM IV) Ill
9.1 Log ( RA
-------
TABLES
3.1 Physical characteristics of clean SARM ..... 14
3.2 Analytical profile of SARMs ...................15
5.1 List of chemicals used for synthetic solutions 28
7.1 A typical output from TOC Analyzer .43
7.2 Precision in calibration and analysis.............. 44
8.1 Results of SARM I washed with different concentration of non ionic
surfactant solution . 47
8.2 Results of SARM I washed with different concentration of ehelant solution .....49
8.3 Effect of ozonation on clean SARM slurry 51
8.4 Effect of ozonation on SARM IV water slurry. 52
8.5 Summary of batch experimental results with different SARM soils ....59
8.6 Effect of SARM type on membrane performance 60
8.7 Summary of batch experimental results with single and double stage washing..65
8.8 Effect of two stage washing on membrane performance 66
8.9 GC/MS analysis of pentachlorophenol and bisphthalate for feed (two stage
wash leachate of SARM IV) and permeate .......69
8.10 Summary of batch experimental results with suspended solids 71
8.11 Effect of suspended solids on membrane performance 72
8.12 Summary of batch experimental results with different pH of wash water 74
8.13 Effect of wash water pH on membrane performance 75
8.14 Summary of batch experimental metal results with different recovery. ....78
8.15 Batch experimental study of flux drop and TOC rejection with recovery 79
8.16 Summary of batch experimental metal results with surfactant washing. 84
xi
-------
8.17 Summary of batch experimental results with surfactant washing ....85
8.18 Effect of surfactant washing on membrane performance .....86
8.19 Summaiy of batch experimental metal results with chelant washing......... .90
8.20 Summary of batch experimental results with chelant washing ........91
8.21 Effect of chelant washing on membrane performance 92
8.22 Results of SARMIV leachate spiked with 50 mg/14-chloroaniline,
75 mgA 2,4-dinitrophenol and 50 mgfl 4-aminobiphenyl 95
8.23 Summary of batch experimental results with different pH of feed.... ............99
8.24 Effect of feed pH on membrane performance................ 100
8.25 Summaiy of batch experimental results with feed ozonation 102
8.26 Effect of feed ozonation on membrane performance ....103
8.27 GC/MS analysis of xylene, ethylbenzene and styrene in feed and
permeate of SARM IV leachate for ozonadon studies ....108
9.1 Adsorption resistance term calculated from experimental values'.. 114
xii
-------
ABBREVIATIONS AND SYMBOLS
A Solvent permeability constant, (cm/MPa/s)
B Solute permeability coefficient, (cm/s)
c Molar concentration of solute, (moI/L)
Cg Bulk feed concentration, (mg/L)
Cp Solute concentration in feed, (mg/L)
CF Solute concentration in permeate, (mg/L)
Cw Solute concentration at wall, (mg/L.)
cw Concentration of water in the membrane, (mol/1)
Cp Solute concentration in permeate, (mg/L)
Ds Diffusivity of solute, (cm2/s)
D= Diffusion coefficient of solute in membrane, (cm2/s)
Dw Diffusivity of water, (cm2/s)
DWF DistiHed water flux through membrane, (cm/s)
Fj Initial feed volume or feed flow rate, (ml or ml/s)
Fp ¦ Permeate volume collected or permeate flow rate, (ml or ml/s)
i number of ionic species in the solution
IC Inorganic carbon, (mg/I)
Js, Ns Solute flux through membrane, (mg/cm2-s)
Jw, Nw Water flux through membrane, (cm/s)
ks Solute distribution coefficient
K ¦ Stability constant
r Water recovery
R Rejection of the membrane
xiii
-------
Membrane resistance due to adsorption, (MPa-s/cm)
r8
Universal gas constant, (L-MPa/Mol-K)
K
Membrane resistance (1/A). (MPa-s/cm)
T
Temperature, (K)
TOC
Total organic carbon, (mg/L)
TC
Total carbon, (mg/L)
TDS
Total dissolved solids, (mg/1)
;VW
Molar volume of water, (L/mol)
Xs
Mole fraction of solute S
Greek Symbols
it Osmotic pressure of solute, (MPa)
Atz ' Osmotic pressure difference across the membrane, (MPa)
AP Pressure across the membrane, (MPa)
Ax Membrane thickness, (cm)
8 Membrane thickness, (cm)
xiv
-------
SECTION 1
INTRODUCTION
Soil contamination is one of the major environmental problems of today. Recently enacted
legislations and the high costs and high energy requirements associated with conventional
excavation and incineration, with subsequent disposal in a landfill, have created a need for
. innovative, cost-effective technologies for the cleanup. Soil washing is a promising technology
for treating contaminated sites, and it is one of the most successful fuE-scale technologies
developed in Europe for site remediation (Pheiffer, 1990). Cost-effective remediation by soil
washing would require, however, a simultaneous development of effective rinsate treatment
' techniques for separating and concentrating dissolved hazardous pollutants in the wash water
and recycling back the treated water for soil washing. These needs were the motivation for the
present work.
Membrane processes provide a very promising treatment technology for this particular
purpose of treating soil-wash rinse-water leachates. Membrane separation processes consume
less energy than conventional processes, and membrane systems are compact and modular.
Microfiltration, ultrafiltration, reverse osmosis and electrodialysis are fully developed
membrane technologies and pervaporation is a developing membrane technology. Reverse
osmosis (RO) in recent years has emerged as a fully developed mature technology and the
estimated worldwide sales of reverse osmosis membranes in 1988 were $118 million with a
projection of $335 million for 1998 (DOE, 1990). Considerable information is now available
for full-scale application of reverse osmosis technology in terms of membrane materials,
module design and, cost estimation. Extensive review can be found in recent books by Parekh
(1988), Rautenbach and Albrecht (1989), Cecille and Toussaint (1989), Drioli and Nagasaki
(1986), Sourirajan and Matsuura (1985) and Belfort (1984). High pressure RO membranes are
used for sea water desalination (osmotic pressure - 2.34 MPa) and low pressure RO membranes
are used traditionally for brackish water desalination (osmotic pressure -0.1 to 0.28 MPa). Low
pressure RO membranes have lower capital and operating cost and because of low pressure
¦ requirements they can be used in spiral element design, thus minimizing membrane fouling. As
compared to low pressure RO process (1,38-2.76 MPa), the energy requirements of high
pressure RO process is about 3-4 times and that of distillation is about 8-18 times for 3,785
; litres of purified water (DOE, 1990). Thin-film composite membranes provide high water flux
and higher rejection. The osmotic pressure of most hazardous waste streams is in the brackish
water range. In view of these attractive properties, it was decided to use low pressure thin-film
composite membranes in the treatment of soil wash rinse water leachates. The
ultra-low-pressure RO process or nanofiltration can be used in combination with low pressure
RO because nanofiltration membrane permeates monovalent ion but rejects divalent and
multivalent ions, as well as organic compounds having molecular weights greater than 200.
This project deals with the use of thin-film, low pressure composite membranes for
! concentrating and separating hazardous pollutants in the soil-wash rinse-water leachates. Soil
washing experiments were done with different wash solutions. The separation characteristics of
the membranes were evaluated in terms of membrane feed total organic carbon' (TOC), heavy
metal concentration, dissolved solids, suspended solids, pH, presence of specific compounds in
1
-------
the feed, feed pre-ozonation, extent of recovery, and water flux. The specific objectives of the
study were:
1. Determine the stability of thin-film composite membranes used for the separation and
concentration of hazardous pollutants.
2. Study membrane performance with leachates obtained from washing SARM (Synthetic
Analytical Reference Matrix) soil with distilled water, with two stage wash of distilled
water, with distilled water at different pH, with surfactant solution, and with chelant
solution.
3. Study the effect of turbidity (fine soil suspensions) constituents on membrane separation.
: 4. Study the effect of feed concentration and feed pH on membrane performance.
5. Establish membrane separation and concentration for selected spiked model compounds. •
6. Establish the effect of water recovery on the permeate flux drop and membrane
rejections,
7. Establish the effect of feed pre-ozonation on permeate flux and over-all rejections in the
Gzonation-membrane process,
8. Study the effect of different flow rates on the permeate flux in a continuous system.
9. To predict membrane flux drop behavior by using a solution-diffusion model containing
an organic adsorption term.
2
-------
SECTION 2
CONCLUSIONS
This study has shown that thin-film composite membranes can be used effectively for the
treatment of soil-wash rinse water leachatcs to produce permeates for reuse. The
permeate can be recycled back to the soil washing step. If the permeate needs to be
discharged, further treatment may be required. The treatment of the concentrated stream
would be much easier and cost-effective due to reduction in the volume to be treated.
This membrane process offers distinct advantages in terms of high solute separations at
low pressures ( < 2 MPa ), high water flux, low energy and capital costs, broad pH
operating range, and by virtue of being compact and modular. If EDTA recovery is also
one of the objectives, then a loose RO membrane like a nanofiltration membrane may be
used to recover EDTA with further treatment of the permeate. In addition, the
ozonation-membrane process would be effective in reducing the flux drop and increasing
the over-all rejections. Specific conclusions are:
1. For non-ionic surfactant concentrations higher than critical micelle concentration, there
was an increase in dissolved leachate TOC. This could be due to increased solubilization
of the hydrophobic organics in the surfactant's micelles and the surfactant's ability to act
as a detergent and its associated low interfacial tension.
2. EDTA washing showed enhanced teachability of Cu and Pb. This could be explained by
the higher stability of metal-EDTA complexes.
3. Ozonation of S ARM-water slurry indicated some oxidation of the soil matter itself. This
suggests that ozonation times should be kept small.
4. Two stage washing enhanced the TOC teachability from the soil.
5. The FT30 membranes used in the study were stable as demonstrated by the consistency
of distilled water fluxes and sodium chloride rejections over an extended operating period
(> 200 days).
6. In the case of distilled water wash, the permeate flux drop for a recovery of 17-22 % were
in the range of 24-33 % and % unaccounted TOC from the mass balance on the
membrane system was around 20 to 25 %. This suggests that there was some adsorption
of organic pollutants on the membrane surface.
7. In the case of surfactant wash, the permeate flux for a recovery of 17 % was about 10 %
higher, but % unaccounted TOC was about 5 % lower than the distilled water wash run.
This suggests that because of surfactant's ability to form micelles and bind hydrophobic
. organics there is less adsorption of organics on the membrane surface. .
8. In the case of EDTA wash, there was no significant increase in flux drop for a recovery
3
-------
of 18 % and there was a 14 % decrease in % unaccounted TOC as compared to distilled
water wash run. This shows that the EDTA wash actually enhanced membrane
performance. This suggests that the metal-EDTA complexes were able to reduce
membrane fouling.
9. In the case of different pH wash, there was no significant difference in the membrane
' performance for all the cases. This was because of the high buffer capacity of the soil
which resulted in the same pH of the leaehate after washing.
10. In the case of run with suspended solids for a recovery of 19 %, the permeate flux and %
unaccounted TOC was about 10-11 % lower as compared to the ran with no suspended
solids. This suggests that presence of fine suspensions for a low recovery may actually
reduce adsorption of organics on the membrane surface.
11. Raising the pH of the leaehate resulted in precipitation of some heavy metals and some
¦ TOC was also removed with the precipitate. Lowering the pH of the leaehate resulted in
formation of slight white cloudiness, probably due to formation of calcium carbonates
(not verified by analysis) from the bicarbonates that might have leached from the soil at
low pH. Hence, the pH adjustment of the leaehate should be done carefully because
during the membrane runs at high and low pH, there could be a precipitate formation on
the membrane surface resulting in membrane fouling.
12. Feed pre-ozonation significantly reduced feed TOC indicating formation of carbon
dioxide. The permeate flux drop for a recovery of 20 % was only 5 % and there was a
reduction in % unaccounted TOC also. This suggests that ozonated products do not
interact strongly with the membrane and hence there is reduced adsorption on the
membrane surface.
13. The TOC rejections for almost all runs for 18 - 22 % recovery were in the range of 82
86 % indicating good membrane performance.
14. For almost all the runs, the heavy metal rejections were high (92-98 %) and the'
conductivity rejections were also high (94-96 %). This shows excellent membrane
performance.
15. Membrane rejections were also found to be high in terms of selected compounds:
pentachlorophenol ( > 98 %), 4-aminobiphenyl (>93 %), ethylbenzene (>97 %), xylene
(>81 %), 4-chIoroaniline (>90 %) and 2,4^dinitrophenol (>98 %).
16. The flux drop (with suspended solids) in the continuous run was found to be less than in
the batch run. Further, the permeate flux drop was found to increase with the decrease in
feed flow rates (Reynolds number < 1000). This suggests some relation of wall
concentration to adsorption.
17. At high water recovery (80 % ), there was a large flux drop. This could be explained in
. terms of significant osmotic pressure contribution from total dissolved solids (primarily
inorganics) which are highly rejected by the FT30 membrane and due to the solubility
limit of some sparingly soluble organics being exceeded at high concentrations resulting
in increased membrane surface hydrophobicity and hence more adsorption' on the
membrane surface. However, the TOC rejections were still found to be high (80 %).
4
-------
18. Solution-diffusion model was modified to include an adsorption resistance term. This
term was correlated with bulk concentration using Freundlich isotherm. The predicted
flux drop at different bulk concentrations and water flux at different recoveries was found
to agree with the experimental values.
5
-------
SECTION 3
LITERATURE REVIEW
OVERVIEW
The purpose of this section is to review briefly (1) the state-of-the-art in situ treatment
technologies for contaminated soils, and (2) membrane processes for wastewater and hazardous
waste treatment.
Early remedial actions for contaminated soils consisted mainly of disposal at a landfill.
Subsequent findings of leaking landfills, heightened public concern and recent legislations have
prompted the development of technologies for the in situ treatment of contaminated soils. In
situ treatment means treating fee waste materials where they are located and is capable of
reducing the risk posed by these wastes to an acceptable level or completely eliminating that
risk.
Membrane processes are being used for the separation and concentration of inorganics and
organics from wastewaters; The advantage of membrane processes is that they can be
simultaneously used to concentrate and purify wastewater containing both organics and
inorganics and produce a 20-50 fold decrease in waste volume that can now be treated with
other processes like incineration. In addition, some membrane processes allow for selective
separation of hazardous pollutants. Reverse osmosis, ultrafiltration and nanofiitrarion have
been used to separate and concentrate organic "and non organic wastes, Pervaporation is an
emerging technology for the treatment of organic contaminated wastewaters.
IN SITU TREATMENT TECHNOLOGIES FOR CONTAMINATED SOILS
Soil Washing / Soil Flushing
Soil flushing is an in situ extraction of inorganic and organic compounds from soils.
Extractant solvents are passed through the soils using an injection/recirculation process. These
solvents may include water, water surfactant mixtures, acids or bases, chelating agents,
oxidizing or reducing agents. Nash (1988) found that in situ soil washing with aqueous
surfactants would be ineffective at soil sites that have contaminants with relatively high
sorption values. Andrews et al (1990) have reported extraction of aromatic contaminants from
a sandy loam soil using supercritical carbon dioxide. Truett et al. (1982) have reported the use
of water flushing for full scale remediation at a former herbicide factory site in Sweden.
Soil washing consists of similar treatments, but the soil is excavated and treated at the
surface in a soil washer. U.S. EPA Edison, New Jersey, has mobile soil washer and other
systems are under development. Literature review of soil washing and results'of EPA project
reports (989) on soil washing with, synthetic soils (SARM) is discussed later in this section.
6
-------
Solidification / Stabilization Techniques
The goal of these technologies is to immobilize the toxic and hazardous constituents in the
waste. This is done by changing the constituents into immobile forms, binding them in an
immobile, insoluble matrix and/or binding them in a matrix which -minimizes the material
surface across which transfer or loss of contaminants may occur. CulEnane et al. (1986) have
given an extensive discussion on these techniques. The technology was evaluated on PCB
contaminated soils at a site (USEPA, 1989) and the results showed the treated soil to have a
dense, homogeneous structure of low porosity, the treatment increased volume by 8.5 percent
and freeze/thaw tests showed unsatisfactorily large weight losses. Achieving uniform mixing
may be difficult and loss of volatiles may occur during mixing procedures.
In Situ Vitrification
In situ vitrification is a process whereby hazardous wastes are converted into a glassy
' substance utilizing very high temperatures (Fitzpatrick et al., 1987). The process is carried out
by Inserting large electrodes into contaminated soils containing significant level of silicates.
Heat is produced by passing high currents of electricity and a melt is produced. A vacuum
collection system is needed to collect any volatilized organics. Contaminants are trapped in the
melt, which, as it cools forms a form of glass. It was originally used for stabilizing radioactive
wastes. The process is energy intensive and has the potential to cause contaminants'to migrate.
Chemical Oxidation And Reduction
Chemical oxidation is accomplished by removal of electrons from the atom and chemical
reduction is accomplished by addition of electrons to the atom. These reactions can change the
form of hazardous materials in the soil to less toxic or change their solubility, stability or
' separability and/or improve their handling/disposaL Iron, aluminum, trace metals, and
adsorbed oxygen has been identified as catalysts that promote soil-catalyzed oxidations
(Furukawa et al.. 1973). Organic wastes that are water soluble and have half-eeU potentials
below the redox potential of a well oxidized soil are amenable to treatment by soil-catalyzed
reactions. Ozone has been used for treatment of soils and ground water at the Karlsruhe site in
Germany (Rice, 1984). A formaldehyde spill at UMah, California was treated by alkaline
hydrogen peroxide oxidation and biological treatment techniques. The applicability of strong
oxidizing agents is usually limited because of their non-selectivity and reaction with any
organic material present in the soil. Sodium borohydride and zinc have been successfully used
in field experiments with soils to chemically reduce organic contaminants (Staiff et al., 1981).
: Ferrous sulfate has been used to treat soils that have been contaminated with hexavalent
' chromium (Sanning and Black, 1987). The disadvantage of adding reducing agents is that if
' ' ' ' isceptible to reduction, treatment have
Chemical Dechlorination
The purpose of this technology is to displace chlorine from chlorinated organic compounds.
The KPEG dechlorination process involves the application of a potassium
hydroxide-polyethylene glycol reagent. The reagent strips one or more chlorine atoms from the
PCB or dioxin molecule. The KPEG process was evaluated with contaminated soils (Esposito
et al., 1989) and was found to be effective in removing volatiles but had only a limiting effect
on removing the inorganic contaminants, alkali polyethylene glycolate type reagents have been
demonstrated to reduce chlorinated dioxin levels in soil (Arienti et al, 1986). Reaction
byproducts of these processes are not well understood and the need to deliver, mix and heat the
7
-------
reagent and the soil is a limiting factor.
Polymerization
Polymerization is the conversion of a monomer or low order polymer to a larger chemical
multiple of itself. Often, such large polymers have greater stability and are less mobile. It is
most effective for immobilization of organic constituents like aromatics, aliphatics and
oxygenated monomers. Treatment solutions containing sulfate related constituents have been
successfully used in polymerization reactions in the soil (Williams, 1982). This technology can
be difficult to apply because of the difficulty in bringing sufficient contact between the catalyst
and the monomers. This treatment can be expensive.
Biological Processes
These processes cause the breakdown of organic wastes into biomass and or other less toxic
products utilizing biological processes. In situ biological treatment of contaminated soils is
being evaluated by enhancing microbial activity in the soils or introducing specialized
microorganisms (mutants) or using enzymes produced by microorganisms or even using seed
producing plants to remove and accumulate compounds from the soil (EPA, 1990). Some of
the important parameters affecting biodegradation include pH, temperature, soil moisture
content, soil oxygen content, and nutrient concentration. Soil extraction/excavation followed
by onsite biological treatment is one of the most frequently used technologies for treatment of
contaminated soils.
Thermal Treatment
Thermal Treatment removes organic contaminants by indirectly heating the soils to
temperatures sufficient to vaporize the hazardous contaminants (Fox et al., 1991). Treatment of
organic vapors is also required. Indirect heating avoids contact between soil, flame and
combustion products. This technology has been demonstrated for treatment of volatile organic
compounds from the soil (Nielson et al, 1989).
Photolysis
Photolysis is a process that breaks down a chemical by light energy. The light energy in the
Ultraviolet (UV) region is generally sufficient to break down compounds and/or transform
molecular structures. Photolysis may be enhanced by addition of proton donor materials. UV
light in presence of organic solvents and aqueous surfactant emulsions have been found to be
effective for treatment of dioxin contaminated soils. This technology is limited by lack of
penetration by UV light
Soil Vapor Extraction
This technology is mostly' used to reduce volatilization or volatile materials from a
contaminated site using a vapor extraction system. This can be done by injecting air into
• contaminated soils and using a vacuum system to extract the vapor filled air from recovery
wells (Hutzler et al., 1989). This technology is limited by the fact that the soils must be
permeable and fairly homogeneous and additional technologies may be required to treat
nonvolatile compounds not amenable to this process. The design considerations for soil
. cleanup by soil vapor extraction has been studied by Ball and Wolf (1990).
8
-------
Radio Frequency Heating
Radio frequency (RF) heating heats the soil uniformly to the point where volatile and
semivolatile organies are vaporized into the soil matrix. Vented electrodes are then used to
recover gas streams that can be subjected to other treatment technologies (Dev et al., 1988).
This technology also could not be applied for soils with nonvolatile or inorganic contaminants.
MEMBRANE PROCESSES FOR HAZARDOUS WASTE TREATMENT
Reverse Osmosis
In this technology, solutes in an aqueous waste increase in concentration as the water flows
from the inlet to the outlet of a membrane unit, due to the passage of water at a significantly
lower solute concentration (permeate) through the membrane. This technology was initially
used for treatment of hazardous waste with cellulose acetate membranes. The rejections of
cellulose acetate membranes were: 95 % for TOC. over 90 % for COD, 59 % for benzene, 67 %
for bis(2~ethylhexyl)phthalate, 25 % for phenol, 10 % for methylene chloride, over 42 % for
cyanide, over 92 % for arsenic and 67 % for chromium (Shukrow et al., 1981). Edwards and
Schubert (1974) had earlier used cellulose acetate membranes for the separation of salts of the
herbicide 2,4-D and separations as high as 50 % were obtained.
Factors affecting reverse osmosis separation of aqueous wastes have been reported to be
type of solutes, feed concentration, extent of water recovery, feed pH, temperature, operating
pressure and type of membrane (Kurihara et al., 1981). Membrane processes are generally used
in combination with other processes like filtration, coagulation, pH control, or adsorption. The
application of reverse osmosis has gained considerable importance after the development of low
pressure thin-film composite membranes. These membranes offer high solute separations at
! low pressures, wide temperatures and broad pH ranges at low energy consumption and capital
cost Chian et al. (1975) compared ceEulose acetate and composite (NS-1GG) membranes for
the separation of pesticides which included several chlorinated hydrocarbons and four
organophosphorus compounds. Rejections of over 99 % were obtained for both membranes but
NS-10G had higher flux. Dickson et al. (1975) have shown that rejections of alcohols, ethers,
ketones, and aldehydes were better for aromatic polyamide than cellulose acetate membranes.
Cadotte et aL(1980) found that the FT30 polyamide membrane was resistant to compaction, had
excellent chemical stability, and a wide pH operating range; on the other hand, the cellulose
acetate membranes were found to suffer from compaction, chemical and biological attack, had
excellent chemical stability, and was sensitive to low and high pH.
Many of the waste streams in the electroplating and metal-finishing industries contain heavy
metals. Reverse osmosis provides an attractive alternative to conventional technologies to
remove, recover or recycle these metals. Robison (1983) describes the use of reverse osmosis
for recovery of nickel 6om plating baths. Imasu (1985) discusses the use of reverse osmosis to
: recover plating chemicals using cellulose and FT30 membranes at three plating shops. The RO
process was found to be cost effective in both treating the plating wastes and producing water
for use in the plating process. Slater, Ferrari, and WisniewsM (1987) studied reverse osmosis to
remove cadmium from metal finishing wastes. They found rejections in excess of 98 % with
thin film composite membranes.
The reverse osmosis process has been applied to the treatment of municipal wastewater also
with the focus on the removal of dissolved solids since these are not removed by conventional
: processes. The process can also be used for removal of organies, color, and nitrates. Nusbaum
9
-------
¦and Aigo (1984) have reported the successful use of reverse osmosis for municipal wastewater
treatment in Orange County, California. The use of reverse osmosis have also been studied for
treatment of wastewaters in pulp and paper industry. In most cases, piate-and-frame or tubular
modules am used because they are easy to clean and resist fouling (Chakravorty and Srivastava,
1987; lonsson and Wiramerstedt, 1985). Basu and Sapkal (1987) used reverse osmosis to treat
black liquor from an alkaline pulping process. Color rejections of over 95 % and chloride
rejections of over 80 % were obtained for feed concentrations from 2-12 % of IDS.
Pre treatment consisted of primary sedimentation, pH adjustment, clarification, multi-media
sand filtration, and cartridge filtration. Reverse osmosis has been found to be an excellent
option for treating wastewater from textile industry because reverse osmosis could be used to
reuse water, recover and recycle dyestuffs and other valuable components and/or recover the
thermal energy in hot wastewaters (Buckley et al., 1985; Slater, Ahlert, and Uchrin, 1987).
Reverse osmosis has also been used to treat many waste streams in the petroleum industry.
However, extensive pretreatment is often required to maintain high performance. Studies have
been done to treat oily waste waters at a facility that manufactures diesel engines (Spatz 1981)
and to treat oily wastewaters generated in the metal-cutting and metal-forming ¦ industry
(Cartwright, 1989). Reverse osmosis has also found applications in power-generation industry
to produce boiler-quality water for steam generation (Patra et al., 1987) and to reduce
wastewater effluents (Hess et al., 1988).
Reverse osmosis has also been used for other applications in wastewater treatment. Hsuie
et al. (1989) have studied the separation of uranium conversion process effluent compounds
from radioactive wastewaters. Film Tec's FT3G membranes provided rejections greater than 98
% and it was found that the radioactivity of the permeate water could be reduced to the lower
detection value. Bhattaeharyya et al. (1984) . studied the FT3Q aromatic polyamide membrane
(in spiral wound module) "and asymmetric aliphatic-polyamide membrane (hollow fiber
module) for the separation of biotreated coal-liquefaction water. At 90 % recovery, both
membranes removed 94-98 % of the organics and 100 % of the color. The IDS and
conductivity rejections were 77 % and chloride rejection was 80 % for hollow fiber module; the
spiral wound module rejected 84 % of TDS and conductivity and 91 % of chloride. Spiral
wound module had higher fluxes also. Siler and Bhattaeharyya (1985) have also shown good
membrane rejections and performance with the FT30 membranes for treating oil shale retorting
wastewaters containing organics (phenolics, aliphatic acids), inorganics, color, odor along with
oils and fine suspended solids. Bhattaeharyya et al. (1987) and Bhattaeharyya and Madadi
(1988) have showed that the FT30 membranes have rejections of over 98 % for PAH
compounds with little drop in permeate flux. For ionizable organics such as phenol,
chlorophenols, and nitrophenols rejections and fluxes were found to be highly dependent on pH
(higher at high pH). Williams (1989) showed that preozonation of feed significantly reduced
flux drop for chlorophenol and chloroethane-chlorophenol mixtures.
Pervarsoration
Pervaporation is a membrane process in which a organic mixture contacts one side of a
permselective membrane and the organic solutes are removed as a vapor from the other side.
The vapor pressure on the permeate side of the membrane is maintained lower than the vapor
pressure of the feed liquid. This process has been used for the dehydration of 90 %
ethanol/water solutions to yield 99.5 % pure ethane! (Bruschke, 1988). Wijmans et al. (1990)
have discussed the removal of organic solvents from dilute aqueous streams using
pervaporation. Membranes more permeable to organic vapor than water were used. Lipski and
Cote (1990) have formulated a model to predict performance, capital cost and operation cost of
pervaporation for removal of organic contamipants from water and to optimize design variables
such as membrane thickness and the module hydrodynamics.
10
-------
Nar.ofiltration
Nanoflltration membranes have separation properties between reverse osmosis membranes
(rejection of most solutes) and ultrafiltration (rejection of only large molecules) membranes.
These membranes are often negatively charged so it is the anion repulsion which mainly
determines rejections. In addition, these membranes also reject compounds with molecular
weights above 200 to 500 (Eriksson, 1988). This process therefore allows for selective
separations, including separation of solutes with charge differences and separation' of high
molecular weight organics from high concentration monovalent salt solutions. Buidoff et al.
(1987) used nanoflltration membranes to remove color causing compounds from the effluent of
the caustic extraction stage of a wood pulping process. Dyke and Bartels (1990) studied
removal of organics from offshore produced waters using nanofiltration membranes. They
found that a produced water of up to 176 mg/L freon extractable organics, will yield a permeate
for discharge of less than 48 mg/1 fieon extractable organics at 50 % recovery, while rejecting
less than 20 % of the NaCl.
Bhattacharyya et al. (1989) have investigated the separation of selected inorganic and
organic compounds using the NF40 membranes. It was shown that, mixtures of cadmium and
nickel could be selectively separated by adding sodium chloride to the solution. Experiments
with the phenol system showed that rejection was less than 5 % when the phenol was not
ionized but increased to 71 % when the solutes were ionized. Williams (1989) found that feed
preozonation of chlorophenol and chlorophenol-chloroethane mixtures reduced flux drop and
increased rejections.
Ultrafiltration
Separation by this process is based on solute size; typically solutes with molecular weights
> 1000 are rejected by the membranes. Bhattacharyya et al (1977, 197S, 1979) have used
charged ultrafiltration membranes for the treatment of TNT manufacturing wastes, heavy
metals, and treatment of acid mine water. Dorcia (1986) found color removals as high as 88 %
and COD rejections of 60 % using ultrafiltration' membranes for treatment of paper mill bleach
plant effluent Chakravorty (1989) used ultrafiltration techniques to produce recyclable water
and concentrated pollutants in a pulping industry. Rautenbach and Albrechet (1989) provides a
review of different applications of ultrafiltration membranes in wastewater treatment
OVERVIEW OF SARM SOILS AND SOIL WASHING
This chapter consists of two main sections. The first one gives an overview and status of soil
washing as a technology for treating contaminated soils. The second section is mainly based on
the EPA reports (Traver, 1989) on development and use of EPA's synthetic analytical reference
' matrix (SARM) and are discussed here mainly to give a brief description of the S ARM soils as
. well as to summarize the results obtained by Traver et al. in soil washing experiments with
these soils. This overview would be useful because S ARM soils were also used in this work.
11
-------
SOIL WASHING TECHNOLOGY
Soil washing technology involves excavation of contaminated soils and then contacting
them with a liquid, extractive agent (usually water). Soil washing is a physical process which
works via two mechanisms- particle separation and dissolution of the contaminants into the
extractive phase. The contaminants in the extractive phase are both in dissolved form and in
dispersed form (suspended or colloidal particles). A significant portion of the contaminants are
attached to the fines (silt, humus, and clay) and hence the separation of the fines from the bulk
of the coarse (sand and gravel) material by soil washing results in significant volume reduction.
Figure 3.1 gives a schematic of a typical soil washing system. It essentially consists of soil
washing, particle sizing and a water treatment system. Clean soil is often clean enough or it can
be used as a feedstock for other treatment technologies. The small volume of fines is either
further treated or disposed as hazardous waste. The treated water is either discharged or
recycled.
First soil washing experiments were done for treatment of crude oils and PCBs
contaminated soil by Ellis et al. (1985). They used variety of aqueous surfactant solutions and
found that 86 % of the crude oil and 98 % of the PCBs were washed from the soil. Despite the
repeated success in the laboratory, soil washing by aqueous surfactants of a sandy soil
contaminated with waste oils and solvents at Volk Field Air National Guard Base did not help
in cleaning within statistical limits (Downey and Elliott, 1990). Soil washing is one of the most
successful full-scale technologies for in situ remediation of contaminated soils (Pheiffer et al..
1990). Five high-throughput soil washing technologies in the Netherlands and the ERG were
reviewed. Hie results of the review show that soil washing generated 10-20 % of the initial
volume as sludge and typical cleaning efficiencies for soil washers ranged from 75-95 %
removal. Glynn et al. (1990) evaluated soil washing technology for remediation of soils
contaminated by the release of petroleum products from leaking underground storage tanks.
Their studies show that removal greater than 90 % of petroleum products could be achieved
with synthetic sols (lower removal for actual soils) but the primary mechanism for contaminant
removal was particle separation. Therefore, enhancement of solubilization mechanism would
be required to effectively remove contaminants from soils containing small fraction of fine
materials. Soil washing was also evaluated as one of the technologies for treatment of
contaminated soils by Esposito et al. (1990). Their studies were done using SARM soils and
their results are discussed in the later sections of this chapter.
CHARACTERISTICS OF SARM SOILS
SARM, an acronym for synthetic analytical reference matrix, is a synthetic soil which was
prepared for use in bench-scale and pilot-scale performance evaluation of treatment
technologies by EPA's Superfund research program (Traver et al., 1989). The physical
characteristics of clean SARM is given in Table 3.1 and Table 3.2 gives the contaminant levels
in four different SARM formulations. Traver et al. (1990) did soil washing studies with SARM
soils using table shaker and then separating the soil into > 2 mm, 250 Jim to 2 mm, and < 250
pm size fractions using a wet-sieve. Soil washing was done with water, chelant solution
(tetrasodium. salt of EDTA) and an anionic surfactant solution (phosphated formulation).
Figure 3.2 and Figure 3.3 show the results of soil washing of SARM IV. The results indicate
that water alone can efficiently remove a significant portion of both the inorganic and organic
contamination from the > 2 mm soil fraction, and the addition of chelant can enhance metals
removal from the middle and fine soil fractions.
12
-------
Water
Recycle
Contaminated Soil
Wash with
EDTA / Surfactant
Wet & Dry
Steve System
Soil Slurry
Coarse Material
Medium
Size Particles
Contaminated
Water r
Water Treatment
System
Further Treatment
or Disposal
Dewatering
Clean Soil
Figure 3.1 Schematic of a typical soil washing system (adapted from Pheiffer.
1990).
13
-------
TABLE 3.1: PHYSICAL CHARACTERISTICS OF CLEAN SARM
Characteristic
Average
Cation exchange capacity, meq/100 g
133
TOC, %
3.2
' PH
8.5
Gravel size distribution, %
Gravel
3
Sand
56
Silt
28
Clay
12
14
-------
TABLE 3.2:
ANALYTICAL PROFILE OF S ARMS
Contaminants Concentration, mg/kg
SARMI SARMTI SARMffl SARMIV
Volatiles
Acetone
4353
356
358
8030
Chlorobenzene
316
13
11
330
1,2-Dichloroethane
354
7
5
490
Ethylbenzene
3329
123
144
2708
Styiene
707
42
32
630
Tetrachloroethylene
408
19
20
902
Xylene
5555
210
325
5576
Scmivolatiies
Anthracene
5361
353
181
1920
Bisphthalate
1958
117
114
646
Pentachlorophenol •
254
22
30
80
Metals
Arsenic
18
17
652
500
Cadmium
22
29
2260
3631
Chromium
24
28
1207
1314
Copper
231
257
9082
10503
Lead
236
303
14318
14748
Nickel
32
38
1489
1479
Zinc
484
642
31871
27060
15
-------
100
100
0 40
Surfactant
Wash
Volatiles Semivolatiles Inorganics
0 40
Water Wash
> 2mm size fraction
0.25 mm to 2 mm size fraction
H < 0.25 mm size fraction
Volatiles Semivolatiles Inorganics
Figure 3.2 ¦ Soil washing effectiveness of water and surfactant solutions for
different size fractions of SARMIV (adapted from Traver et al, 1990).
16
-------
Cheiant Wash
Volatiles Semivolatiles Inorganics
Water Wash
g§ > 2mm size fraction
0 025 mm to 2 mm size fraction
[j < 0.25 mm size fraction
Volatiles Semivolatiles Inorganics
Figure 3.3 Soil washing effectiveness of water and cheiant solutions for different
size fractions of SARMIV (adapted from Traver et al., 1990).
17
-------
SECTION 4
REVEKSE OSMOSIS THEORY
OVERVIEW
This chapter consists of two main sections. The first one deals' with basic definitions and
concepts used frequently in reverse osmosis literature and in this thesis. The second section
deals with a brief review of different models and mechanisms that have been proposed to
describe the transport of solute and solvent through reverse osmosis membranes.
BASIC CONCEPTS
Osmosis And Osmotic Pressure
A semi-permeable membrane is one that is permeable to solvent but not to solute. The
chemical potential of a pure solvent is higher than that of the solvent in a solution, hence, when
a semi-permeable membrane separates the two, the solvent permeates through the membrane to
the solution side to maintain thermodynamic equilibrium. This process is called osmosis. The
pressure head developed on the solution side in the above process at equilibrium (no net flow of
solvent) is called osmotic pressure. For dilute solutions, it can be calculated using Van't Hoffs
equation (Weber, 1972):
7C=icRT (1)
Reverse Osmosis And Effective Pressure Driving Force
Reverse Osmosis or RO refers to the phenomena when a pressure greater than osmotic
pressure is appied to the solution side and the solvent flows from the solution side to the pure
solvent side. In reality, even some solute is transported through the membrane and hence the
need to consider osmotic pressure for both sides. Therefore, effective' pressure driving force is
defined as (AP - Air),1 where AP is applied pressure difference and Are is the osmotic pressure
difference.
Terminology In A RO process
Figure 4.1 illustrates the basic terminology used in describing a reverse osmosis membrane
process. Initial solution is the Feed, the solution passed through the .membrane is called
permeate and the concentrated feed solution is called retentate. Flux is measured in terms of
volume flow through the membrane per unit time per unit area. Membrane performance is
evaluated in terms of recovery, rejection and flux drop.
The recovery r which is a measure of the amount of feed water that has passed through the
membrane is given by:
18
-------
Feed
Feed Volume, F
Solute Concentration C
High Pressure Side
Ap - An
= net or effective transmembrane
pressure difference
Ap
= transmembrane pressure
difference
= Bulk Concentration
Low
¦' ¦. '
Pressure Permeate Side
Retentate
or Concentrate
Solute Concentration, C
R
Retentate Volume, F
J = Water Flux
W
C = Solute Permeate Concentration
P
J = Solute Flux
S
F = Permeate Volume
Figure 4.1
Description of a reverse osmosis membrane process.
-------
The rejection R which is a measure of solute separation by the membrane is given by (for dilute
solutions with no concentration polarization):
R= 1- 7f (3)
Flux Drop which depends on the nature of the solute and membrane is given by:
DW - Jw
Flux Drop = (4)
F DWF w
Concentration Polarization And Membrane Fouling
This phenomena is illustrated in Figure 4.2 and. refers to the formation of a boundary layer
near the membrane on the high pressure side of the membrane, and the concentration in this
boundary layer Cw is higher than the feed bulk concentration C£. This may happen because of
solvent-membrane affinity or solute-membrane affinity. In case of solvent-membrane affinity,
solute is highly rejected by the membrane and its concentration increases near the membrane
leading to concentration polarization. In case of solute-membrane affinity, solute is attracted by
the membrane and gets adsorbed on the membrane surface again leading to concentration
polarization. The wall concentration can be described by film theory (Bird et al, 1960) and this
description shows that as mixing on the high pressure side of the membrane is increased, the
mass transfer coefficient in the boundary layer increases leading to a decrease in concentration
polarization (Rautenbach and Albreeht, 1989). Concentration polarization is not desirable. It
can result in a decrease in water flux (because osmotic pressure increases at the membrane
surface), increase in solute flux, changes in separation properties of the membrane, and fouling.
' Fouling is the accumulation of material at the membrane surface due to solute adsorption.
Thin-Film Composite Membranes
Thin-film composite membranes are asymmetric and have high permeate flux and high
¦ separation characteristics. Initial reverse osmosis membranes had homogeneous thickness of
the membrane resulting in large flow resistance and hence low permeate flux. Thin-film
composite membranes has three layers: (1) a relatively dense skin layer, (2) a porous support
layer, and (3) a backing for mechanical strength and. support. The porous support is fabricated
first, from one polymer, and a thin film of different polymer is coated on the porous support
(Cadotte and Peterson, 1980). The membrane performance is primarily dependent on the
chemical nature and physical structure of skin layer since most of the mass transfer resistance is
in the skin layer. FT30 is a successful commercial thin-film composite membrane
manufactured by Film Tec Corporation. Figure 4.3 shows the cross-section of a FT30
membrane and the chemical composition of the top skin layer. The chemical composition of
the skin layer is crosslinked polyamide with some negative charges. The amine derivative
gives the best performance and durability. However, oxidizing agents like chlorine or ozone
can attack the amine groups and selectivity of the membrane would be lost.
20
-------
w
Feed
Membrane
Permeate
Jw
C| = initial concentration
= bulk concentration
= wall concentration
¦Figure 4,2 Description of concentration polarization.
21
-------
Membrane Feed
Porous Pofysulfone Support {50 Microns)
Polyester Support Backing (125 Microns)
(a)
(b>
Figure 4.3 FT 30 Membrane: cross-section and chemical composition.'
22
-------
MEMBRANE TRANSPORT MECHANISMS AND MODELS
Several models have been developed to describe the transport of solute and solvent through
reverse osmosis membranes. The mechanism of transport in RQ is still a matter of controversy
and several mechanisms have been suggested. Review of these mechanisms and models could
be found in literature (Soltaniah and Gill, 1981), Some of the models and mechanisms are
briefly reviewed here.
Solution Diffusion Mechanism .
Several models have been derived that are based on solution diffusion mechanism. This
mechanism assumes that the membrane skin is non-porous. According to this mechanism,
suggested by Lonsdale et al. (1965), the solute and solvent dissolve in the nonporous skin layer
of the membrane and then diffuse across it in an uncoupled manner due to chemical potential
gradient (because of differences in concentration and pressure across the membrane); therefore,
the solubilities and diffusivitles if the solution components are of highest importance. Pore
flow (i.e convection) is ruled out in this mechanism. Some of the models based on this
mechanism are discussed below.
Solution-Diffusion Model
This model was originally developed on solution diffusion mechanism by Lonsdale et aL
(1965). The water flux is proportional to the solvent chemical potential difference and is given
by:
JW=A(AP-Ait) (5)
D CryV
where A= J*"' w. The solute flux is proportional to the solute chemical potential difference
R TAx
g
and is given by:
J =B(Cb-Cp) (6)
where Equations (5) and (6) may be combined with the definition of rejection
(Equation 3) and overall material balance to give:
i Dsks i
_L_1+—s_s _L (7)
R Ax Js
These equations predict that the solute flow through the membrane is independent of water
flow; rejection by the membrane depends on the solute distribution coefficient between the
solution and the membrane phase and the solute diffusivity in the membrane phase. The model
also predicts that feeds with higher osmotic pressures result in less water flux. The primary
• advantage of this model simplicity as only two parameters are required. However, the
solution-diffusion model is good only for systems with high rejections.
23
-------
Solution-Diffusion-Imperfection Model
This model (Sherwood et al.5 1967) is also based on solution diffusion mechanism but it
assumes some Imperfections, or pores, on the membrane surface layer which contribute to the
pore flow of solute and solvent. This model accounts for systems that exhibit lower separation
than the separation predicted by solution-diffusion model The water flux through the
membrane is given by
where K. is a coupling coefficient and ew is concentration of water. The solute flux is given by
where K-is a solute permeability coefficient equivalent to B in equation (6).
Extended Solution Diffusion Model
This model (Burghoff et al., 1980 and Jonsson, 1980) is also based on solution diffusion
mechanism but includes an extra pressure term in the solute chemical potential equation. This
model was found to be good when the solute partial molar volume is large and the rejections are
low.
Preferential Sorption-Caoillary Flow Mechanism
This mechanism first suggested by Sourirajan (1970) assumes that the membrane is highly
microporous. According to this mechanism, the transport is determined by surface phenomena
and partly by fluid transport through the pores. Depending on the physicochemical nature of
the solute-solvent-membrane system, one of the component of the solution is preferentially
sorbed by the membrane and then forced through the membrane capillary pores under pressure.
Some of the models based on this mechanism are briefly discussed below.
Preferential Sorption-CapUlary Flow Model
This model developed by Sourirajan et al (1970) is based on the above mechanism. The
water flux according to this model is given by
Nw =Jw+K3APcw =A(AP-A~)+K3APcw
(8) •
NS =JS+K3^CB =K2+K3^CB
(9)
¦ NW=A(AP—(7CF-Ttp))
The solute flux is given by
(10)
(11)
For dilute solutions, these equations can be combined to give
(12)
24
-------
In contrast to the solution-diffusion model, these equations predict that rejection depends on
diffusivity, partition coefficient and effective length based on pore rather than membrane
material.
Surface Force-Pore Flow Model
The surface-pore flow model ( Sourirajan and Matsuura, 1985) is based on the above
mechanism and applies specifically to thin-fflm composite membranes. The transport occurs
through cylindrical shaped pores through the barrier layer of the composite membranes. A
layer of water molecules is preferentially sorted on the walls of the cylindrical pores. For this
model, differential equations axe solved to obtain radial velocity profiles in the pores, which are
then related to rejection.
Irreversible Thermodynamics Models
These are mechanism independent transport models and have been derived using
irreversible thermodynamics. These are based on one of the principle of irreversible
thermodynamics that the process (system) can be divided into smaller systems'at equilibrium
and determining the thermodynamic equations around each system. The Kedem Katchalsky
(1958) is the most well known. Pusch (1986) and Soltanieh and Gill (1981) have provided
excellent reviews of these models. The main drawback of these models is that they do not give
any insight into the transport mechanism.
25
-------
SECTION 5
EXPERIMENTAL EQUIPMENT AND PROCEDURES
OVERVIEW
Figure 5.1 is an overview of the different experiments and studies done. Essentially, there
were two basic experimental set ups - one for soil washing and one for membrane experiments.
Membrane studies were mostly done in a batch system but one final study was done in a
continuous system. In addition to this, a vacuum filtration set up was used for most of the
experiments and an ozonation system was used for'ozonation studies. Table 5.1 gives a list of
names, abbreviations, ¦ purities and brand names of chemicals used. Detailed experimental
procedures for various project objectives are provided under the appropriate sections on Results
¦and Discussions.
SOE, WASHING EXPERIMENTS
Soils Used
Experiments were done with SARM (Synthetic Analytical Reference Matrix) supplied by
U.S.EPA. The decision to use SARM was based on several factors. First, RCRA permit
regulations restrict off-site treatment of contaminated Superfund site soils. Second, SARM
soils broadly represent a wide range of soils and contaminants (volatiles, semivolatiles and
meals). Third, soil washing technology has been evaluated by EPA using the same SARM
soils. The description of SARM soils has already been discussed in a previous chapter. Most
of the experiments were done with SARM IV (high organics, high metals content).
Wash Solutions Used
Distilled Water
•Distilled water was used for soil washing or for making other wash solutions. Distilled
water was used instead of tap water because tap water characteristics could vary from day to
; day and it was decided to minimize variables. pH of distilled water was varied for some
experiments.
Surfactant Solution
Triton X-100, a non-ionic surfactant was used for surfactant washing. Triton is a trademark
of Rohm and Haas Company. It is an anhydrous liquid consisting of water soluble
octylphenoxypolyethoxyethanol with an average of 10 moles of ethylene oxide. The selection
of Triton X-100 was based on its ability to act as a highly effective hard surface detergent, to
act as an efficient emulsifier for aromatic liquids and on its ability to disperse or solubilize
¦ many water insoluble materials.
26
-------
WASHING SOLUTION
o.w.
Tman
3)TA
Two
fH
Wasn
Wash
Wash
Slags
Wasn
3,8,8
Wssn
SARM
Retentate
MIXER
Soil:Water::1:10
1 hr. mixing
Suparnatani
Washed
SARM
FT 30 Membrane
System
P = 1.72 MPs
Well-Mixed
W/////////7777A
Vacuum
Filtration
Fiitarad Laacftat*
suspended
Solids
pH
Control
4
Permeate
Ozonation
Membrane Feed
TOC « 80 ta 160 mg/l
ANALYSIS
TOC
GC/MS
AA
PH
Conductivity
Suspended' Solids
Dissolved Solids
Figure 5.1 Schematic of the overall experimental plan.
27
-------
TABLE 5.1: LIST OF CHEMICALS USED FOR SYNTHETIC SOLUTIONS
Chemical
Purity
Brand
2,4-DinitrophenoI
>99%
Fluka
4-chIoroaniline
>98%
Fluka
4-aminobiphenyl
98 %
Aldrich
Methylene Chloride
HPLC-GC/MS Grade
Fisher .
Triton X-100
High purity
Aldrich
Versene 100 EP
Technical grade
Dow Chemical Co.
Sodium Chloride
Reagent grade
Fisher
28
-------
EDTA Solution
Versene 100EP solution was used for EDTA washing. Versene is a trademark of Dow
Chemical Company, It is an aqueous solution of the tetrasodium salt of
ethylenediaminetetraacetic acid (Na4EDTA). It has 39 % of Na4EDTA.4H20 as active
ingredient. The selection of Versene 100EP was based on its demonstrated ability as a widely
used chelant for controlling metal ions over a broad pH range. Versene has been shown to be
effective in soil washing (Traver et aL, 1989).
Soil Washing Procedure
The variables important in soil washing are temperature, pH, reaction time, surfactant type
and concentration, chelant type and concentration, solution-to-soil ratio and number of rinses.
However, because of time constraints, it was decided to minimize evaluation of some variables
based on the results in literature. A 10:1 solution-to-soil ratio was used for all experiments. All
experiments were done at ambient temperature (22-25 °C). Reaction time (mixing time) was
fixed for one hour and for most of the experiments single rinse was done. Other variables were
varied and their effects observed. Mixing was done using a bench shaker but to improve
mixing, after initial few experiments, mixing was done using a top agitator. After mixing, the
soil-water slurry was allowed to settle in a cylindrical jar. Settling time of one hour was found
to be good to obtain efficient settling. The top supernatant was labeled as leachate and removed
for further experiments. The washed soil was stored in a closed glass container. '
VACUUM FILTRATION
For most of the experiments, suspended solids from the leachate were removed by vacuum
filtration at 58 cm Hg using a MilHpore GS type filter of 0.22 jam pore size. The filtered
leachate was used for further experiments while the suspended solids removed were stored as
waste.
pH CONTROL
pH of all samples were measured using a pH meter. For experiments involving pH control,
the pH of the samples was raised using 1 N NaOH and lowered using 1N I^S04_
OZONATION PROCEDURE
Solutions were ozonated (0 - 30 min.) in a 500 ml stirred reactor (a 500 ml cylinder). The
ozone was produced by a Welsbach Model T-816 ozonator which uses an electric discharge to
generate ozone from an ultrapure oxygen feed. The gas stream was passed through the reactor
using a fine pore diffuser at a rate of 0.2 Standard litres per minute (at 22 C and 156 kPa). The
ozone concentration was determined as outlined in Standard Methods (1985) and was found to
be 2 vol % (62 mg/1). For ozonation-membrane experiments, the procedure is outlined in
Figure 5.2. The ozonated solutions were stirred for sixteen hours to remove any residual ozone
in the solution since ozone could damage the membrane. The actual stirring time is not critical;
the objective is to remove residual ozone prior to membrane experiments. This could also be
accomplished by adding NaHSOs. After mixing, any precipitate, if formed, was removed by
vacuum filtration and pH of the filtrate was adjusted as needed for membrane run.
29
-------
OZONATION OF SOLUTION
r-^
( wtnomi j
&
CUM RUMf tAUftt
gi#
ma lit van
pawn
LO
O
Ozonator
Stirred
Reactor
To Exhaust
MIXING OF
SOLUTION TO
REMOVE
RESIDUAL
OZONE
MEMBRANE
EXPERIMENT
X
pH
Adjustment
(if necessary)
I
r
A
Figure 5.2
Experimental set-up for ozonation-membrane experiments.
-------
MEMBRANE EXPERIMENTS
Membranes Used.
Thin-film composite polyamide membranes (Film Tec's FT .30) were used for the
membrane experiments. The structure and composition of FT-30 membrane has already been
discussed. The FT 30 membranes have a pH tolerance of 2-11 and high water flux of 10-13 x
1Q"4 cm/s. The stability of these membranes was continuously monitored in terms of distilled
water flux and standard NaCl (1000 mg/1) rejections. These techniques are routinely used to
determine changes in membrane performance after runs with various pollutant systems.
Apparatus And Procedure For Batch System
The experimental set-up is shown in Figure 5.3. The set-up consists of a 2.0 L stainless
steel cylindrical pressure cell. The membrane was placed at the bottom of the cylinder and had
an area of 39 cm2* A variable speed agitator was used to maintain mixing conditions to prevent
concentration polarization. The feed solution was introduced into the batch cell and
compressed nitrogen (AP=1.75MPa) was used to provide the pressure driving force for the
experiments. Permeate water fluxes were monitored throughout the run. Samples of feed,
retentate and permeate were taken for analysis.
Apparatus And Procedure For Continuous System
The experimental set-up is shown in Figure 5.4. This system contained a feed cell and a
continuous flow cell containing the membrane in series. The feed cell consisted of a cylindrical
2.0 L stainless steel pressure vessel. The continuous flow cell consisted of two machined
stainless steel plates that when fastened together contained a thin, coiled channel that was 1.4
cm wide, 100 cm long and 0.084 cm in height. This means the equivalent dia was 0.158 cm.
Reynolds number was calculated as a function of flow rate:
Re = 2,246.3 Q
where Re is the Reynolds number and Q is the flow rate in litres per minute. The membrane
was placed between the two plates on a polysulfone support and had an area of 140 cm2. The
channel of the flow cell provided mixing of solution flowing through the cell to prevent
concentration polarization. The pump shown in the system was used to provide flow of feed
through the system, flow rates were varied using the valve just before the continuous cell. The
system had a dead volume of 120 ml (i.e. after a ran, with all the water drained out including
from the flow meter, 120 ml of liquid still remained in the system). The feed solution was
introduced into the feed cell and after it had circulated for several minutes, a feed sample was
withdrawn from the valve after the cooling cell. The temperature was maintained constant at 24
CC by the cooling coil and the temperature was monitored by a thermocouple. Compressed
nitrogen was used to provide the pressure driving force and the transmembrane pressure of the
membrane cell AP =i.75MPa) was monitored by the gauges as shown in Figure 5.4. Permeate
water flux was monitored throughout the ran. Feed, permeate and retentate samples were taken
for analysis.:
31
-------
Mixer
Membrane Area
39.0 cm
Compressed
Nitrogen
Permeate
Figure 5.3 Experimental set-up for membrane experiments in batch system.
32
-------
Continuous Flow Thin Channel Cell
2
Membrane Area = 140 cm
Flowmeter
Cooling
Coil
Permeate
Feed Solution Reservoir
IBJ — Valve
(p) — Pressure Gauge
~ - Thermocouple
Compressed
Nitrogen
Figure 5.4
Hxperimental set-up for membrane experiments in continuous system.
-------
Membrane Cleaning
Membrane cleaning was done by distilled water after each run. Distilled water permeation
was done at low pressure (0.5 MPa) at high mixing. If the distilled water flush did not help in
recovering 90 % of the initial distilled water flux, 10-30 % methanol-water solution was used
for cleaning. Methanol solution helps in the distilled water flux recovery because it acts as a
solvent to remove most of the organics that might be adsorbed on the membrane surface.
34
-------
SECTION 6
ANALYTICAL EQUIPMENT AND PROCEDURES
OVERVIEW
A variety of analytical procedures were used for this work. Samples of leachates,
membrane feeds, retentates, and permeates as well as ozonation samples were analyzed by
Total Organic Carbon (TOC), GC/MS (after extraction) and Atomic Absorption
Spectrophotometer. Conductivity and pH measurements were also taken. The conductivity
meter was calibrated by using standard NaCl solutions. For some studies, total dissolved solids
and suspended solids in the leachate were also calculated using methods described in Standard
Methods (1985). The reproducibility of all analysis was checked by multiple analysis of the
samples, the precision, accuracy and the data quality indicators are explained in the section on
Quality Assurance And Quality Control Some analysis of heavy metals and of some
semivolatile organics were done by Technology Applications, Inc. (EPA analytical support lab)
by using EPA approved methods.
Conductivity and pH measurements were made with standard conductivity meter (Fisher)
and a digital pH meter (Orion Research). Permeate flow measurements were made by
measuring the collected volume in a measuring cylinder and using a stopwatch to measure the
elapsed time.
APPARATUS AND PROCEDURE FOR TOC ANALYSIS
The TOC of samples was determined using EPA QA procedures (EPA Method
600/4-79-020). The analysis was performed using a Beckman Model 915B Total Organic
Carbon Analyzer Computational System with teleprinter. The Model 915B incorporates
channels for determination of both total carbon (TC) and inorganic carbon (IC); the TOC is the
difference between the TC and IC. The sample is directly injected into one of the channels
where the organic is converted to carbon dioxide in the TC channel or carbonate is converted to
carbon dioxide in the IC channel; the carbon dioxide is conveyed by the carrier gas to an
integral infrared analyzer. The operating conditions for the instrument were: TC furnace
temperature - 950 C IC furnace temperature -155 C carrier gas (ultrapure air) flow rate - 300
-- ml/min; sample injection volume - 50|il — 400Lil.
APPARATUS AND PROCEDURE FOR METAL ANALYSIS
Selected metal analysis was performed using a Varian AA575 Atomic Absorption
Spectrophotometer. The samples were treated with concentrated nitric acid to prevent
formation of hydroxides in the dilute solutions. The samples had to be diluted so that the metal
concentrations fell into a linear range analyzable by the AA. Calibration was performed
between each samples. The absorbanee of the unknown sample was plotted against the
calibration points to determine the concentration of the diluted sample. This value was
35
-------
multiplied by the concentration factor to get the concentration of the original sample. The
operating conditions for analysis of Cu were: Lamp Current - 3.5 mA, Fuel - acetylene,
Support - air, Flame stoichiometry - oxidizing, Wavelength - 324.7 nm, Spectral band pass - 0.5
nm, Working range - 2 to 8 mg/1
APPARATUS AND PROCEDURE FORGC/MS ANALYSIS
¦ Some organic compounds in membrane feeds and permeates were investigated by GC/MS
to establish relative membrane separation behavior. The objective of these analysis was to find
relative concentration of specific organic compounds in membrane feed and permeate as well as
ozonation samples. Percentage area count rejections in the linear calibration range were used to
study the membrane performance and effect of ozonation on specific compounds. The analysis
were done using the HP 5S90A GC/ HP 5971A MSD. The MSD includes a GC/M3D interface,
an electron impact ion source, a hyperbolic quadrupole mass filter and electron multiplier
detector. Real time control of the MSD is from HP G1030A MS Chemstation (DOS Series)
: analytical workstation which includes program to tune (calibrate) the MSD, acquire data and
process data. The MSD was tuned every time to meet the tuning criteria and samples were
introduced at the GC inlet port by direct injection. Identification of the spectra was done using
library search and confirmation by injecting known samples obtained from Supelco. Library
searches were done with NIST/EPA/MSDC 49K, mass spectra database which has a database of
spectra of about 40,000 compounds (mostly environmental). GC columns, data acquisition
parameters, temperature program for the GC and other operating conditions were dependent on
the compounds being analyzed and are given along with the discussion on the analysis (Results
and Discussion section) of these compounds. All the aqueous samples were extracted with
methylene chloride according to the EPA method 604. The procedure is shown in Figure 6.1.
The pH of samples was adjusted below 3 so that all organic solutes may be in non-ionized
condition.
36
-------
~ Adjust pH of Solution
/ Yes \
Third Extraction
Aqueous Solution
Aqueous
Soiutian
CH2Ci2' Solution
in Eflenmeyer Flask
Shake in Separatory
Funnel; Separate
Discard
For Analysis
Figure 6.1
Extraction procedure for GC/MS analysis.
37
-------
SECTION 7
QUALITY ASSUKANCE AND QUALITY CONTROL
OVERVIEW
The main purpose of this section is to do data quality assessment in terms of precision,
accuracy, and data reduction, validation and reporting. Quality assurance of experimental and
analytical procedure is also assessed in terms of sample collection and custody, and calibration
of analytical instruments.
SAMPLE COLLECTION AND SAMPLE CUSTODY
Samples of soil wash leachates, membrane feeds, permeates and retentates as well as
ozonation samples were collected. They were collected using a measuring cylinder and beakers
that were washed with distilled water after cleaning with concentrated sulfuric acid. About 100
ml of each samples were collected. pH and conductivity measurements of the samples were
done as soon as they were collected. All the samples that were taken for later analysis were
stored at 4 °C in colored borosiicate glass vials with Teflon-lined screw caps. For GC/MS
analysis, the samples were' extracted in methylene chloride and analyzed within 1 days. Each
sample was labeled with an identification number, date of experiment and sample type.
CALIBRATION
Multilevel calibrations were used for each instrument, with at least one level of caEbration
performed daily. The TOC instrument was calibrated with potassium hydrogen phthalate for
TC analysis and sodium carbonate-sodium bicarbonate for IC analysis. The standards were
injected in between the samples to check calibration. Hie microprocessor controlled instrument
directly provides, after calibration with known standards, concentration (as carbon), mean
value, standard deviation and % variance. Table 7.1 show a typical output of the TOC analysis.
The standards for calibration of Cu for analysis by Atomic Absorption spectrophotometer was
done using copper nitrate dissolved in distilled water and treated with nitric aqid. Figure 7.1
show the calibration curve (after linear regression with least squares method) for copper
analysis. The absarbance of standards were plotted against concentration and concentration of
unknown samples were determined using this calibration curve. Known standards were used
before each sample. GC/MS was tuned routinely to prepare MS for data acquisition. The MS
tuning compound was perfluorotributylamine (PFTBA) and was used to tune MS in terms of
mass peaks (69, 219, 512), peak width (0.4-0.6), relative abundance (69 peak = 100 %, 219
peak > 30 % and 502 peak > 1 % ), and isotope ratio (isotope mass 70, 220 and 503 should be
in the range 0.5-1.5 %, 2-8 % and 5-15 % respectively). For % Area Count analysis of samples,
multiple samples were injected to check reproducibility. Figure 7.2 show chromatographs from
two injections of the same sample for a typical GC/MS analysis.
38
-------
100
90
. 80
'«o
* 70
o
« 60
G
^ O U
o ¦¦
«
.< 40
30
20
10
0
0 12 3 4 5 6 7 8 9 10
Concentration (mg/1)
Figure 7.1 Calibration curve for copper analysis.
Wavelength : 324.8 nm
Lamp Current . 3.5 rxxA.
Fuel : Acetylene
Support : Air
Flame : Oxidizing
Slit
Range : 0.02-10.0 pprn
Aspiration Rate:2.1 ml/miri
Air/Fuel
39
-------
^fiunaance
120000 4 9.,5.8
100000
g.32
80000 -
1
60000 4
40000 - \
20000
¦ 'i
I «
HEHBSAHE F2£D: SASH IV 8 = 9 ain: Etiyllsenzena
9 » 2 • JB1" f p-Xylena
'10,0 airs: Sttyrena
10.1 airs: o-Xylene
10.12
10.00
,V . Wia
Tiae -> 9.30
10.00
11.00 12.00 13.00
14.00
Abundance
120000 -}
100000 4
80000
£0000
2.is
TIC; GZ5-a~F2.0
4C0GC
20000
S.92
10.11
sljoly9 lo. do
Tina ->
9.00
10.00
11.00
12.00
13.00
Figure 7.2 Chromatographs from two injections of the same sample for a typical
GC/MS analysis.
40
-------
DATA REDUCTION, VALIDATION AND REPORTING
The reporting scheme for data collection, storage, validation and analysis is shown in Figure
7.3. Membrane performance results were periodically checked for standard NaCl rejections and
distilled water flux reproducibility. Finally, bulk concentration was correlated with an
adsorption resistance temi to predict solvent flux drop at different bulk concentration or to
predict solvent flux at different recovery.
DATA QUALITY INDICATORS
Table 7.1 shows the data quality indicators for the TOC analysis and Table 7.2 shows the
data quality indicators for other analytical results. These indicators were calculated using the
following relations.
Arithmetic Mean
n
Arithmetic Mean = x =——
n
where x, is any data and n is the number of data
Standard Deviation
(13)
Standard Deviation = s =
V ^ ^
(14)
i-1
Precision
If calculated from duplicate measurements:
Relative Percent Difference -¦
(x, - x2) x 100
(15)
x
If calculated from three or more replicates;
§
Relative Standard Deviation =— x 100
x
(16)
41
-------
Membrane Experiments
Data Collection
Analysis
I
Quality Control
Storage
f
Calibration
Figure 7.3 Reporting scheme for data collection, storage, validation and analysis.
42
-------
TABLE 7.1:
A TYPICAL OUTPUT FROM TOC ANALYZER
Sample # Header Channel
Avg
ppm
SD
ppm
var
%
FS
Vol
id
DEL
1 Test 50 ppm
TC
50.0
0.15
0.61
100
50
1:1
2 S4DW-F-6
TC
101
0.58
0.57
200
50
1:1 '
3 S4DW-R-6
TC
87.2
1.71
1.97
100
50
1:1
4 S4DW-P-6
TC
27.5
0.21
0.76
50
50
1:1
5 Test 100 ppm
TC
100.5
0.35
0.49
100
50
1:1
6 S4VW-F-14
TC
248
1.73
0.70
400
20
1:1
7 S4VW-R-14
TC
261
1.73
0.66
400
20
1:1
8 S4VW-P-14
TC
33
0.69
2.10
50
50 -
1:1
9 Test 400 ppm
TC
400.7
0.15
0.39
400
20
1:1
43
-------
TABLE 7.2: PRECISION IN CALIBRATION AND ANALYSIS
Description Measurements Relative SD
1 2 3 or. Relative % Diff.
Flux Measurement:
Fluxxl04cm/s I LI 10.9 1L1 1.05%
AA Analysis:
Cu (2 mg/1) 1.9 1.9 2.0 2.99 %
Typical GC/MS Analysis (Area Counts of feed):
m,p-xylene 7523189 6104658 - 21 %
Ethylbenzene 4471305 3812373 - 16%
Styrene 945656 768155 - 21 %
% Area rejection for m,p-xylene:
% 87% 84% - 3.51%
pH measurement:
pK 7 7.01 6.99 7.00 0.14%
Typical conductivity measurement:
mmoh/cm 1.75 1.72 1.74 0.88 %
Typical total suspended solids measurement:
mg/1 175 168 - 4.08 %
¦ Typical total dissolved solids measurement:
mg/1 1420 1410 - 0.71 %
Typical TOC analysis:
50 mg/1 50.1 50.0 50.2 0.30 %
44
-------
Accuracy
x — x
% Accuracy = x 100 (17)
x.
LINEAR REGRESSION STATISTICS
All linear regression was done using the 'REG' procedure in SAS software on IBM 3090.
Linear regression analysis was done for calibration curves and for the membrane flux drop
prediction model.
45
-------
SECTION 8
RESULTS AND DISCUSSION
Soil washing experiments were performed with SARM soils. The purpose of these experiments
was to produce a soil-wash leaehate that would serve as a feed for the membrane studies. The
leaehate typically contained volatiles, semivolatiles, heavy metals, suspended solids and
dissolved solids depending on the type of SARM soil and the type of washing solution.
Washing was done with distilled water, distilled water at different pH, distilled water +
surfactant and distilled water + chelant to determine the effect of different washing solutions on
flux drop and membrane separation characteristics. Studies were also done to study the effect
of turbidity constituents on membrane separation. Leaehate pH was also varied to study the
effect on flux drop and rejections. Ozonation of leaehate was also conducted to investigate its
effect in membrane flux.
SOIL WASHING STUDIES
As pointed out earlier, the purpose of these experiments were not to evaluate soil washing
itself but to produce a leaehate using different wash solutions, so that the leaehate could be used
to evaluate the performance of low pressure composite membranes. An overview of the soil
washing technology has already been presented in a previous section.
Soil Washing With Surfactant Solution
Table 8.1 shows the dissolved TOC, dissolved Cu and pH of the leaehate after washing
SARM I (high organics, low metal) with different concentration of non-ionic surfactant (Triton
X-100) solution. There was almost no apparent change in dissolved Cu teachability. There was
an increase in dissolved leaehate TOC for the initial increase in surfactant concentration. The
effectiveness of surfactant washing can be explained by the increased solubilization of the
hydrophobic organics in the surfactant micelles and the surfactant's ability to act as a detergent
and its associated low interfacial tension. The critical micelle concentration (CMC) is the
aqueous surfactant concentration at which the surfactant molecules become aggregates
(micelles) having a hydrophobic core and external hydrophilic headgroups (Armstrong, 1985).
The CMC for Triton X-100 is near 0.01 %. Surface tension for Triton X-100 at concentrations
near CMC is 30 dynes/cm as compared to 72 dynes/cm for distilled water, hence its ability to
disperse, or solubilize many water insoluble materials. There is some adsorption of non-ionic
surfactants like Triton X-100 on the soil (Abdul et al, 1991). This suggests that the increase in
leaehate TOC shown in Table 8.1 indicated some enhanced teachability of hydrophobic
organics. At high concentration (0.08 % by weight of surfactant) there was a drop in leaehate
TOC. At high concentration the expected increase in number of micelles could lead to pore
plugging and hence poor organics teachability from these pores. Moreover, there is an
increased adsorption of the surfactant on the soil at high concentration and this could change
the soil chemistry and increase the adsorption of organics (Abdul et al, 1991). For membrane
experiments, 0.04 % by weight surfactant solution was used to produce the soil-wash leaehate.
46
-------
TABLE 8.1: RESULTS OF SARMI WASHED WITH DIFFERENT
CONCENTRATION OF NON-IONIC SURFACTANT
SOLUTION
Surfactant Cocn. TOC in pH of Cu in
% by weight leachate (mg/I) leachate leachate (mg/I)
0 104 6.9 <0.1
0.01 117 6.9 <0.1
0.04 132 7.0 <0.1
0.08 99.9 6.9 <0.1
Surfactant = Triton X-I00 (TOC = 65.7 % )
SARM I: Wash Solution:; 1:10
47
-------
Soil Washing With EDTA Solution
EDTA has been shown to be effective in soil washing ( PEI Associates, 1989). Table 8.2
shows the dissolved TOC. dissolved Cu and pH of the leachate after washing SARM I (high
organics, low metal) soil. There was a increase in leachate TOC and a significant increase in
leachate Cu concentration with increase in Na4EDTA (Versene 100EP) concentration. EDTA is
a chelating agent with six metal-complexing sites. It forms a bond with all the reactive sites
(4-6) of a metal ion resulting in the formation of a ring structure incorporating the metal ion
(chelate complex). Chelation is an equilibrium reaction and the chelating agent reacts first with
metals which form more stable complex (generally have higher stability constant 'K'). For
example, consider the reaction between copper ion and EDTA
Cu^+EDTA4" = CuEDTA2" (18)
Stability Constant for Copper-EDTA complex is given by:
K = [CuEDTA2-] (19)
[Cu ] [EDTA ]
Figure 8.1 compares the effect of EDTA wash on teachability of dissolved Cu, Ni, Pb and Zn
with that of distilled water wash and surfactant wash of SARM IV soil. There was a maximum
increase in the teachability of Cu, This can be understood if the stability constants of different
metal-EDTA complex are taken into account. The stability constant of Cu-EDTA complex has
a higher value (log K = 18.8) as compared to EDTA chelates of Ni (log K = 18.6), Pb (log K =
18.0) and Zn (log K = 16.5) and this explains why there was the enhancement of Cu
teachability.
Ozonation Of S ARM-water Slurry
The purpose of this study was to combine soil washing with ozonation. Ozone has been
considered for in situ treatment of contaminated soils as discussed in the Chapter dealing with
literature review.
Tables 8.3 and 8.4 show the effect of ozonation on Blank SARM-water slurry and SARM
IV-water slurry. Figure 8.2 shows the TOC in the leachate with ozonation time. After 2 hrs. of
ozonation of Blank SARM-water slurry, there was almost a 3.6 times increase in dissolved
leachate TC, with about 1.5 times increase in dissolved leachate IC and about 40 times increase
in dissolved leachate TOC. However, in initial 5 minutes of ozonation there is only a small
increase in TOC teachability. This suggests that to prevent oxidation of the soil matter itself,
ozonation should not be carried out for long times,^ There is little change in pH even after 2 hrs.
of ozonation. This is because of the capacity of "the soil to act as an extremely good buffer.
Ozonation of SARM IV water was limited to 10 minutes to prevent much oxidation of the soil
organic matter. There was about a 15 % increase in dissolved leachate TOC and little change in
pH. The TOC increase could be due to method variability; our objective here was not to
investigate the effect of ozonation on soil but rather to study its effect on membrane
performance. Concentration of dissolved heavy metals (Cu, Ni, Pb and Zn) were also analyzed
and Table 8.4 shows that there was a small increase in Cu, Ni and Zn concentration in the
leachate after 10 min of ozonation.- This suggests oxidation of some heavy metals to higher
oxidation states leading to their increased mobility. One of the disadvantage of using ozone
with the soil is that much of the ozone would be wasted on oxidizing nontarget compounds
because of its high oxidizing potential.
48
-------
TABLE 8.2: RESULTS OF SARM I WASHED WITH DIFFERENT
CONCENTRATION OF CHELANT SOLUTION
Active Ingredient TOC in - pH of Cu in
(m mole^l) leachate (mg/1) leachate leachate (rng/1)
0 165 7.28 <0.1
0.21 174 7.35 2.7
0.41 210 7.58 2.9
0.82 254 7.75 4.1
SARM I: Wash Solution:: 1:10
Cfaelant = Versene 100EP (TOC = 13.1 % with 39 % Active Ingredient)
Active Ingredient =Na4EDTA.4H20
49
-------
c
k
u
C -Leachate Cocentrstion rng/f
Co =Leachate Coon. After D.W. Wash rng/l
pH = 6.0-7.0
Distilled Water Wash
1 moI/L EDTA Wash
0.04 % Surfactant Wash
m
Figure 8.1 Effect of wash solution on teachability of Cu, Ni, Pb And Zn.
50
-------
TABLE 8.3:
EFFECT OF OZONATION ON CLEAN SARM SLURRY
Ozonation
TC
IC
TOC
pH
Time min
mg/1
mgfl
mg/1
0
27.5
26.0
1.5
/ .8
5
47.2
39.1
8.1
7.6
30
50.3
29.9
20.4 :
6.8
120
98.1
37.8
60.3
6.7
Blank SARM : Water:: 1:10
51
-------
TABLE 8.4:
EFFECT OF OZONATION ON SARMIV-WATER SLURRY
Ozonation Leachate pH Cu Ni Pfa Zn
Time mill TOCmg/1 mg/1 mg/1 mg/1 mg/1
0 130 6.2 1.58 10.5 1.72 224
10 150 5.8 4.8 11.6 1.23 242
SARM IV : Water:: 1:10
52
-------
160-
140-
120-
Ozcrne Rate=0.2 SLFM
2 % Ozone
P 100 j
^ 80-
o
O ¦
E- 60-
pH=8.7-7.8
¦ Blank SARM Slurry
* SARM IV Slurry
40-
20-
0 10 20 30 40 50 60 70 80 90 100 110 120
Ozonation Time (min)
Figure 8.2 TOC leachability from clean SARM and SARM IV with ozonation
time.
53
-------
MEMBRANE STUDIES
The purpose of these studies was to evaluate the performance of low pressure (< 2 MP a)
thin film, composite polyamide membranes (Film Tee's FT 30) for treating the leaehate
obtained from washing SARM soils. These type of membranes (pH tolerance 2-11) have high
water flux (10 - 13 x Iff"4 cm/s at 1.75 MPa) and have been shown to be effective in the
removal and concentration of dilute hazardous organics (Bhattacharyya et al , 1987, 1988,
1989). Membrane performance was evaluated in terms of membrane stability, permeate flux
behavior, TOC rejections, conductivity rejections, metal rejections, rejections of selected
organic compounds and the effect on these parameters by different wash solution, different feed
concentration, different feed pH, suspended solids, dissolved solids and feed pre-ozonation.
Membrane Stability
Membrane stability refers to the consistency of membrane performance and membrane
characteristics over the operating period. The general practice to check membrane stability is to
measure distilled water flux (DWF) before each experiment and to periodically measure
standard sodium chloride (1000 mg/i) rejections. Two set of FT30 membranes were used.
Figure 3.3 shows the distilled water flux for one of the membranes (membrane 1) over an
operating period of 133 days. The flux varied from 11.2 - 12.6 xHT4 cm/s (less than 12 %
variation) over the operating period. Sodium chloride rejections are shown in Figure 8.4 for
membrane 1. The rejections were in the range of 94-98 % (less than 5 % variation) over the
operating rime. Membrane cleaning was done by distilled water after each run. If the distilled
water flush did not help in recovering at least 90 % of the initial DWF, 10-30 % methanol
solution was used for cleaning. Methanol solution enhances the distilled water flux recovery
because it acts as a solvent to remove any of the organics that might be absorbed on the
membrane surface. Membrane 2 was used for a much longer period of about 230 days. Flux
values for distilled water are shown in Figure 8.5 with the flux varying from 9.0 - 11.5 xlO"4
cm/s. The major drop is only after 150 days indicating that methanol wash may not be
sufficient in recovering the initial DWF. However, the sodium chloride rejections as shown in
Figure 8.6 were still good (94-98 %) indicating good membrane stability.
Studies With Leaehate Of Different SARM Soils
Experiments were done with SARM I (high organic, low metal), SARM II (low organic,
low metal) and SARM IV (high organic, high metal) to study the effect of soil composition on
membrane performance in terms of flux drop and rejections. The soil washing procedure and
the filtration step was same as previously discussed. The filtrate was used as the feed for the
membrane runs. Tables 8.5 and 8.6 summarize the results of the membrane experiments for
this study. The runs were made at a pressure of 1.72 MPa for water recovery of 17-20 %.
Figure 8.7, 8.8 and 8.9 show the permeate flux and post-run distilled water flux for SARM I,
SARM II and SARM IV leaehate respectively.
Table 8.5 shows that washing different SARM types result in membrane feed of widely
different feed concentration but of almost same pH (6.1-7.1). The constant pH is due to the
buffering capacity of the soils. The permeate concentration was found to increase with an
increase in feed concentration as shown in Table 8.5. Since the solute concentration is
increasing on the high pressure side of the membrane, it is expected that the solute flux would
also increase (Equation 6). However, the magnitude of the feed concentrations are such that the
TOC rejections also increase with an increase in feed concentration as shown in Table 8.6. The
permeate flux behavior is almost same for all the three feeds as shown in Figures 8.7,8.8 and
54
-------
6 13-
o
System:Distilled "Water
AP:1.75 MPa
pH=5.8—6.8
% r = 10.0%
0
15
30
45
60
105 120 135
75
90
Operating time (days)
Figure 8.3 Distilled water flux with operating time for FT-30 ( Membrane 1).
55
-------
100
90
80
70
CI
60
.2
o
-------
7
System:Distilled Water
AP:1.75MPa
pH=5.8—6.8
% r = 10.0%
¦ 1 1 i ¦ • 1 1 i 1 ! 1 1 i 1 ' > ¦ i 1 • ' 1 i 1 ' '¦ • i • 1 1 1 i • 1 1 i i 1 ¦ 1 • i 1 • ¦ • i'
0 25 50 75 100 125 150 175 200 225 250
Operating time (days)
Figure 8.5 Distilled water flux with operating time for FT-30 ( Membrane 2 ).
5
3
1
57
-------
"SI
90
80
70
§ 60
¦ o
50
o
Q.£
M 40
30
20
10
0
System: 1000 mg/1 Nacl
AP:1.75MPa
pH=S.3
% r = 11.0 %
—I—:—[-1—I—II!'!—I I I I V I • | I I ! I f t f f I
I t ' I >' i I
I ' ' ' ' I
0 25 50 75 100 125 150 175 200 225 250
Operating time (days) . ¦
Figure 8.6 NaCl rejection with operating time for FT-30 (Membrane 2),
58
-------
TABLE 8.5;
SUMMARY OF BATCH EXPERIMENTAL RESULTS
WITH DIFFERENT SARM SOILS
SARM % r Feed Permeate Reteriate
Type' TOC ¦ pH TOC pH TOG pH
mg/L nag/1 tngfl
SARM I 18.1 64.7 6.23 18.2 5.9 55.8 6.9
SARM II 19.6 37.1 7.1 11.1 6.1 45.1 7.4
SARM IV 17.8 101 6.15 27.5 3.15 87,2 4.4
SARM: Water:: 1:10
AP= 1.75 MPa
59
-------
TABLE 8.6:
EFFECT OF SARM TYPE ON MEMBRANE PERFORMANCE
SARM TOC in % Dissolved % r % Flux % D.W. Flux % TOC % Unaccounted
Type Soil, mg/kg TOC Drop Recovery Rejection TOC
SARM I 1,828 3.54 18.1 32.3 89.2 72.0 -24.3
SARM II 100 37.17 19.6 26.5 97.2 70.1 +3.6
SARM IV 1,582 6.39 17.8 29.6 92.6 74.2 -24.2
SARM: Water:: 1:10
AP= 1.75 MPa
-------
8.9 with the lowest flux drop (26 %) being for lowest concentration feed (SARM II). The
figures also show that the best recovery of the post-run distilled water flux was for lowest
concentration feed - SARM II feed (97 %) as compared to higher concentration feeds - SARM I
and SARM IV (90 %). For SARM I and SARM IV runs, 30 % methanol wash of the
membrane was required to bring the distilled water flux back to its initial value. A mass
balance of the TOC was also done on the membrane system as indicated by % unaccounted
TOC in Table 8.6. The mass balance has a margin of error of within 5 %. There was almost no
unaccounted TOC for SARM II as compared to about 24 % unaccounted TOC for SARM I and
SARM IV runs. All this suggests that for high concentration feeds, there is some adsorption of
organics on the membrane surface resulting in membrane fouling.
Studies With Two Stage Wash Leachate
The purpose of this study was to compare the TOC teachability and membrane performance
of single stage wash leachate and two stage wash leachate. Two stage wash has been found to
be optimal for soil washing (PEI Associates, 1989). Experiments were done with SARM IV
soil. The soil washing procedures for both single and two stage wash has already been
explained in the section on experimental procedures. The leachates were filtered using vacuum
filtration as already discussed. The filtered leachates were the feeds for the membrane runs.
Table 8.7 and Table 8.8 compares the results of the membrane runs with leachate of single and
two stage wash of SARM IV. Figure 8.10 compares the permeate flux behavior for the two
cases.
Table 8.7 show that two stage wash resulted in almost a 40 % increase in TOC teachability
over single stage wash. An increase in TOC teachability is expected because theoretical
extraction efficiency of any washing solution increases with number of stages. Table 8.7 also
show about a 33 % increase in dissolved Cu teachability with two stage wash, but the actual
amount of dissolved Cu leached in both the cases is so small (1.58 mg/1 to 2.1 mg/1) that it is
difficult to draw conclusions. Table 8.8 compares the membrane performance in both cases.
Both the runs were made for a recovery of 16-17 %. The pH of the two feeds was almost same
(6.15-6.2). Cu rejections were high in both the cases (95-96 %). Table 8.8 and Figure 8.10
show that as compared to single stage wash ran, two stage wash run had a slightly higher flux
drop (about 3 % more), lower post-run distilled water flux recovery (about 8% lower) and
higher TOC rejections (about 10 % more). These results were expected because two stage wash
leachate has higher feed concentration and the effect of higher feed concentration on membrane
performance has already been explained in the discussion of membrane performance with
leachate of different SARM soils.
The feed and permeate of two stage wash run was analyzed for pentachlorophenol and
bis(2-ethylhexyl)phthalate using a GC/MS to establish their rejections by the FT 30 membrane.
Figure 8.11 show the chromatographs for feed and permeate samples and the relative
abundance of the peaks. Table 8.9 gives the results of GC/MS analysis for pentachlorophenol
and bisphthalate in feed and permeate samples of two stage wash leachate run.
Pentachlorophenol was found to be rejected more than 97 % and bis(2-ethylhxeyl)phthalate was
rejected more than 84 % (using area counts).
61
-------
4-j
System:SARM I Leachate
AP:1.75MPa
pH=6.23
2 -1 % r = 18-1%
¦ Permeate Flux
* Post—run D.W. Flux
0
. t »—i i , , i f » i j i i i t—r~
0 10 20 30 40 50 60 70 80 90
Time (min)
Figure 8.7 Permeate Flux vs. time for SARMI leachate as membrane feed.
62
-------
14i
o
System:SARM II Leaehate
AP: 1.75 MPa
pH=7.1
¦ Permeate Flux
Post—run D.W. Flux
0
10
20
30
40
50
60
70
Time (min)
Figure 8.8 ' Permeate Flux vs. time for SARMII leachate as membrane feed.
63
-------
14
_ 12
ca
.i 10
eo
* 8
X
P
4
2
04.
System:SARM IV Leachate
AP:1.75 MPa
pH=5.15
% r = 17.8%
- Permeate Flux
« Post-run D.W. Flux
i ' • 1 1 i
i—i—j j > i i" j '"7—i—s—5—|—i—i—i—r-j—i—r~
I
o
10 20 30 40 50 60 70 80
Time (min)
• i
90
Figure 8.9 Permeate Flux vs. time for SARMIV leachate as membrane feed.
64
-------
TABLE 8.7:
SUMMARY OF BATCH EXPERIMENTAL RESULTS WITH SINGLE AND DOUBLE STAGE
WASHING
CT\
Ln
Washing % r Feed Permeate Retentate
Stage TOC Cu pH TOC Cu pH TOC Cu pH
mg/1 mg/1 mg/1 mg/1 mg/1 mg/1
Single 17.8 101 1.58 5.15 27.5 0.06 3.15 87.2 2.0 4.4
Double 16.4 144 2.1 6.2 25.8 <0.1 5.2 131 2.3 5.9
SARMIV: Water:: 1:10
AP= 1.75 MPa
-------
TABLE 8.8:
EFFECT OF TWO STAGE WASHING ON MEMBRANE PERFORMANCE
Washing % TOC % Cu % r % Flux % D.W. Flux % TOC % Cu % Unaccounted
Stage Dissolved Drop Recovery Rejection Rejection TOC
Single 6.4 0.2 17.8 29.6 92.6 72.8 96.2 -24.2
Double 9.4 0.2 16.4 32.3 84.7 82.1 >95.2 -21.0
CT>
ON
SARM IV: Water:: 1:10
AP = 1.75 MPa
-------
144
m
¦f
O
X
X
£
System:SARM IV Leaehate
AP: 1.75 MP a
pK=5.15-6.1
%r= 16.4-17.8
¦ 2—stage wash
* 1—stage wash
4:
0
10
20
30
40
60
70
60
50
90
Time (min)
Figure 8.10 Peimeate Flux vs. time for SARM IV leachate from single and double
stage washing.
67
-------
Abundance
i 2000020 J 12,75
I5QQ0G0 -i.
1000003
500000 4
i
Tine ->
HEHBRA.HE PEED: SARM TV ( 2 Stage )
12.75 ain: pentachloropftenoL
16.36 mint Bis-phtiiala-ca
I
16.96
13.00 14.00 15.00 16.00 17.00 18.00 13.00
Abundance
I 100000•
80000¦
60000
40000
20000
MEMBRANE PERMEATE: SARM IV (2 Stage)
12.75
12.75 ain; Fenteciilorophenoi
16.97 min: Bis-pbtoalate
16.97
liata ->
13.00 14.00 15.00 IS.00 17.00 IS.00 19.00
Figure 8.11 GC/MS ehromatograph of pentachlorphenol and bisphthaiate for feed
¦ " (double sage wash leachate of SARM IV) and permeate.
68
-------
TABLE 8.9:
GC/MS ANALYSIS OF PENTACHLOROPHENOL AND
BISPHTHALATE FOR FEED (TWO STAGE WASH LEACHATE
OF SARM IV) AND PERMEATE
Description
Area Counts
Area Counts
(12.8 min)
(16.9 min)
Feed
27274933
3419361
Permeate
548674
523958
Total Area Rejection >96 %
Pentachloropfaenol Area Rejection > 97 %, (12.8 min)
Bisphthalate Area Rejection > 84 %, (16.9 min)
Column : HP-1 (Crosslinked Methyl Silicone Gum) 12 ms 0.2 mm x 0.33 ji m film
thickness.
Operating Conditions: Temperature programming - hold at 40 C for 2 min then linear
increase to 100 C at 10 C/min, then linear increase to 250 C at 25 C/min and hold for 3
minutes; earner gas - Helium; sample injection volume - 1 fi L, solvent delay of 2 min
and mass scan range from 35 to 400.
69
-------
• Studies With Turbidity Constituents
This study was done to establish the effect of turbidity constituents (fine soil suspensions)
on the membrane performance. SARM IV soil was used for the experiment. The supernatant
obtained after settling the washed SARM IV-water slurry was divided into two halves. One
half was vacuum filtered and was labeled as filtered leachate. The other half was labeled as
non-filtered leachate and the suspended solids were not removed* The procedure followed to
measure suspended solids has already been reported in the chapter dealing with Analytical
Procedures. The suspended solids in the leachate were found to be 175.2 mg/L The filtered and
non-filtered leachates were used as feeds for the membrane runs. Table 8.10 and Table 8.11
show the effect of suspended solids on the membrane performance. Figure 8.12 compares the
permeate flux for filtered and non-filtered leachates for a recovery of 19-20 %.
The feed TOC for both the runs were same since samples were filtered before analysis and
both feeds were part of the same leachate. Table 8.10 show that conductivity and pH were
almost same for both the runs. Table 8.11 shows that the conductivity rejections were high
(about 96 %) for both the runs. However, it was found that non-filtered leachate as compared
to filtered leachate had a lower flux drop (about 10 % lower), lower % unaccounted TOC
(about 11 % lower) and slightly higher TOC rejections (about 4 % higher). This suggests that
presence of fine suspensions may actually be reducing the adsorption of organics on the
membrane surface and hence lower flux drop, lower % unaccounted TOC and higher TOC
rejection. However, this study was done with only 175.2 mg/1 of suspended solds and for a
recovery of only 19 %. Higher suspended solids concentration or a run at higher recovery may
plug feed-water passages, damage membrane, and damage pumps (Bates et al, 1988; Parise et
al, 1988).
Studies With Leachate Of Different pH Wash
This study was conducted to study if there is any effect of pH of the wash solution on the
teachability of organics and metals and thereby any effect on the membrane performance. The
experiments were done with SARM IV soil. Soil washing was done with distilled water
adjusted to different pH of 3.06, 6.8 and 10.3 and the leachate of each was used for membrane
run to evaluate membrane performance. Table 8.12 and 8.13 summarize the results of this
study. Figure 8.13 compares the permeate flux for leachate from different pH wash.
Table 8.12 shows that the pH of leachate from each wash was almost same (6.5-6.8). The
difference in % dissolved TOC in leachate for the three cases (about 3 % difference) could be
because of the difference in the soil samples itself. Table 8.13 shows that most of the indicators
of membrane performance for the three cases axe almost same: % flux drop (about 5 %
difference), % DWF recovery (about 4 % difference), % TOC rejection (about 4 % difference),
% conductivity rejection (about 3 % difference) and % unaccounted TOC (about 4 %
difference). This is because the pH of only the initial wash water was adjusted and the pH was
not continuously being adjusted during the washing itself. Soil has a high buffer capacity and
this results in the same pH of the leachate after soil washing (if the pH is not continuously
adjusted). This explains almost the same permeate flux behavior observed for the three runs as
shown in figure 8.13 for a recovery of 20-23 %.
70
-------
TABLE 8.10:
SUMMARY OF BATCH EXPERIMENTAL RESULTS WITH SUSPENDED SOLIDS
Feed % r Feed Permeate Retentate
Suspended Solids TOC k pH TOC lc pH TOC k pH
(mg/1) mg/1 mmho/cm mg/1 mm ho/cm mg/1 mmho/cm
0 18.9 141 1.56 5.52 22.2 0.065 4.81 132 1.77 5.9
175 19.71 141 1.51 5.53 16.9 0.062 4.9 155 1.88 6.43
SARM IV : Water:: 1:10
AP= 1.75 MPa
-------
TABLE 8.11:
EFFECT OF SUSPENDED SOLIDS ON MEMBRANE PERFORMANCE
Suspended Solids % TOC % r % Flux % D.W. Flux % TOC % Conductivity % Unaccounted
in Feed (mg/1) Dissolved Drop Recovery Rejection Rejection TOC
0 8.9 18.9 32.7 95.6 84.3 95.8 -21.7
175 8.9 19.7 22.8 95.4 88.0 95.9 -9.4
SARMIV .-Water:: 1:10
AP = 1.75 MPa
-------
14:
12-
m
s
o
T?
o
Systenr.SARM IV Leachate
AP: 1.75 MFa
pK=5.5 ¦
% r = 19 %
» NO Suspended Solids
*175.2 mg/1 S. Solids
60
70
80
90
0
10
20
30
40
50
Time (min)
Figure 8.12 Permeate Flux vs. time for filtered and non-filtered leachate of
S ARM IV.
73
-------
TABLE 8.12:
SUMMARY OF BATCH EXPERIMENTAL RESULTS WITH DIFFERENT pH OF WASH WATER
Wash Water % r
PH
Feed
TOC k pH
mg/1 mmho/cm
3.06 20.7
151 1.72 6.5
6.80 22.9
121 1.75 6.5
10.3 20.4
110 1.7
6.8
SARMIV: Water:: 1:10
Permeate
TOC k pH
mg/1 mmho/cm
Retentate
TOC lc pH
mg/1 mmho/cm
13.5 0.074 5.48
141 2.02
6.28
15.1 0.155 3.76
115 2.08
5.94
13.9 0.126 4.06
107 1.99
6.4
AP= 1.75 MPa
-------
TABLE 8.13: EFFECT OF WASH WATER pH ON MEMBRANE PERFORMANCE
Wash Water % TOC % r % Flux % D.W. Flux % TOC % Conductivity % Unaccounted
pH Dissolved Drop Recovery Rejection Rejection TOC
3.06 9.6 20.7 21.9 92.7 91.1 95.7 - 24.1
6.80 7.6 22.9 23.8 95.3 87.5 91.1 -23.9
10.3 6.9 20.4 27.2 91.0 87.4 92.6 - 20.0
Ln
SARM IV: Water:: 1:10
AP= 1.75 MPa
-------
3 10
System:SARM IV Leachate
AF:i.75 MPa
pH=8.3—6.8
% r = 20 %
* pH 3 wash
¦ pH 6 wash
* pH 10 wash
20 30 40 50 80
Time (min)
70
80
90
Figure 8.13 Permeate Flux vs. time for SARM P/ leachate washed at different pH.
76
-------
Studies With High Water Recovery
The motivation behind, this study was to study the effect of increasing concentration of the
solution on the high pressure side of the membrane. Membrane runs were done with filtered
leach ate of distilled water washed SARM TV soil. The soil washing procedure and the filtration
step was the same as previously discussed. Table 8.14 and Table 8.15 summarize the results of
the study. Figure 8.14 shows the behavior of permeate flux as a function of recovery. Figure
8.15 shows TOC rejection as a function of recovery and Figure 8.16 gives the results of
¦ pentaehlorophenol analysis of feed and permeate samples, (semi-quantitative analysis).
Table 8.14 compares the heavy metal rejections of the SARM TV-Distilled Water washed
leachate at recoveries of 18 % and 80 %. The rejections of Cu. Ni, Pb and Zn at 80 % recovery
were found to be almost as high as at 18 % recovery. The rejections were in the range of 94-95
¦ % indicating good membrane performance. The measured permeate concentrations (for both
Tables 8.14 and 8.15) were used to calculate rejections by Equation 3 (Section 4). Figure 8.14
shows the effect of increasing recovery on the permeate flux. Table 8.15 shows that the major
flux drop was during the initial 10 % recovery and then after 60 % recovery. The initial flux
drop for 10 % recovery was about 30 %. This drop was expected because of the higher osmotic
pressure of the feed solution as compared to distilled water. The additional flux drop for 10 to
40 % recovery was only about 5 %. This can be explained by the increasing concentration of
the solution on the high pressure side with recovery. Higher concentration on the high pressure
side means higher osmotic pressure and hence higher flux drop. For 40 - 60 % recovery, there
was an additional 6 % flux drop and for 60 - 80 % recovery, there was an additional 20 % flux
drop. To understand this we will have to consider the implications of high recovery. Figure
8.16 shows that the feed sample had about 10 mg/l pentachlorophenol and after 80 % recovery
the permeate had less than 1 mg/l pentachlorophenol. This high rejection of pentachlorophenol
(> 97.5 %) indicates that after 60 % recovery, the feed solution had about 23 mg/l
pentachlorophenol and after 80 % recovery, about 46 mg/l pentachlorophenol. The solubility
limit of pentachlorophenol in water is only about 14 mg/l. This suggests that because of
increasing concentration of solutes at high recovery, the solubility limits of sparingly soluble
solutes were exceeded resulting in two phase formation (increased hydrophobicity) on the
membrane surface. This could foul the membrane surface resulting in low solute flux. The
mass balance of TOC on the membrane system after 80 % recovery showed that about 37 %
TOC was unaccounted as compared to about 25 % unaccounted TOC observed for 20 %
recovery runs. This also suggests that there was an increased fouling of the membrane surface
at 80 % recovery.
Total dissolved solids (IDS) in the leachate of a distilled water washed SARM IV was
measured. TDS was found to be about 1420 mg/l. TDS would be highly rejected by these
membranes and this means that after 60 % recovery, the feed would have about 3,550 mg/l
TDS and after 80 % recovery, the feed would have 7,100 mg/l TDS. The osmotic pressure due
to TDS could be approximated by 0.07 MPa/1000 mg/l TDS. This means that there was about
0.25 MPa osmotic pressure from TDS alone after 60 % recovery and about 0.48 MPa osmotic
pressure from TDS alone after 80 % recovery. Thus % Flux Drop due to TDS at the operating
pressure of 1.75 MPa with initial 1,420 mg/l at recovery r can be approximated by
CT _ 0.07 x 1.42 inn
FIuxDrop = x 100 (20)
(1-r) (1.75)
Thus, TDS alone account for about 13 % flux drop at 60 % recoveiy and 27 % flux drop at 80
% recoveiy. Hence, total dissolved solids play an important role in terms of permeate flux at
high recovery.
77
-------
TABLE 8.14: SUMMARY OF BATCH EXPERIMENTAL METAL RESULTS
WITH DIFFERENT RECOVERY
% r Feed • Rejection
Cu Ni Pb Zn Cu Ni Pb Zn
. mg/1 mg/l mgfl mg/l % % % %
17.8 1.58 10.5 1.72 224 96.2 93.1 96.5 93.3
79.9 1.97 14.1 2.32 395 >94.9 >95.6 94.9 94.9
SARM IV : Water:: 1:10
AP = 1.75 MPa
78
-------
TABLE 8.15: BATCH EXPERIMENTAL STUDY OF FLUX DROP AND TOC
REJECTION WITH RECOVERY
r
Flux Drop
Permeate iOC
TOC Rejection
%
%
mg/1
%
1.29
25.91
30.9
82.93
9.82
29.64
33.3
81.60
20.5
32.76
34.4
80.99
29.43
33.36
35.6
80.33
39.75
34.23
36.4
79.89
60.25
40.12
38.4
. 78.78
79.93
59.97
45.2
75.03
SARMIV : Water:: 1:10
Feed TOC =181 mg/1
Retentate TOC = 456 mg/l
Unaccounted TOC = 37 %
AP = 1.75 MPa
79
-------
14-
13-
12-
oo
s
a
©
System:SARM IV Leachate
AP: 1.75 MPa
pH=8.3
0
30 40
80
90 100
10
20
50
60
70
% Recovery
• Figure 8.14 Permeate flux as a function of reeovexy for S ARM IV leachate as
membrane feed.
80
-------
c
.2
o
4)
100
90
80
70
60
£ 50
u
g "40 -j
30
20
10
0
o
System:SARM IV Leachate
AP:1.75 MPa
pH=8.3
Soil:Water::l:10
10
20
30
40 50 60
% Recovery
70
80
90 100
Figure 8.15 TOC rejection as a function of recovery for SARM IV leachate as
membrane feed.
81
-------
12
'
g Feed
10-
H Permeate
'
20% r 80% r
Figure 8.16 Relative concentration of pentachloroptienol in feed and permeate of 20
and 80 % recovery run (SARMIV - distilled water wash).
82
-------
TOC rejections were also monitored at different recoveries to evaluate the membrane
performance. Figure 8.15 shows the effect of increasing recovery on the TOC rejections.
There was a small drop in TOC rejections with increasing recovery. The drop is expected
because of increasing concentration on the high pressure side of the membrane with recovery
which results in higher solute flux (Equation ) and hence higher permeate concentration
resulting in lower rejections. The TOC rejections dropped about 3 % in going from 2 to 40 %
recovery, about 1 % in going from 40 to 60 % recovery and about 4 % in going from 60 to 80
% recovery. The trend in TOC rejections drop was thus found to be the same as the trend in
flux drop.
Studies With Leachate After Surfactant Wash
The purpose of this study was to see if the surfactant wash has any effect on the membrane
performance. Surfactant (0.04 % by weight -400 mg/1 Triton X-100) solution was used for
washing SARM IV soil. The soil washing procedure and the filtration step was the same as
previously discussed. Membrane runs were done with filtered leachate. Tables 8.16 and 8.17
summarize the results of this study. Table 8.18 shows the effect of surfactant wash on
membrane performance. Figure 8.17 compares the permeate flux behavior for leachate of
distilled water wash and leachate of surfactant wash for a recovery of 17-18 %.
Table 8.16 shows that there was a slight increase in teachability of heavy metals with the
surfactant wash but the metal rejections were still high (about 93 %). High metal rejections
indicate good membrane performance for heavy metals even after surfactant wash.
Figure 8.17 shows that there was a slightly higher flux drop (about 10 % higher) for
surfactant wash leachate as compared to distilled water wash leachate. However, Table 8.18
shows that there was less % unaccounted TOC (about 5 % lower) for surfactant wash leachate
as compared to distilled water wash leachate. This suggests that the surfactant concentration
has two effects on the membrane performance. First, the feed is higher in TOC which means
higher osmotic pressure on the feed side and hence higher flux drop. Secondly, because of the
surfactant's ability to act as a hard surface detergent and its ability to form micelles and bind
' hydrophobic organics there is less adsorption of organics on the membrane surface and hence
less % unaccounted TOC. Table 8.17 shows that TOC concentrations of feed, permeate and
retentate samples for both surfactant and distilled water wash are almost same (within 5 %).
The TOC rejections were found to be almost same (within 2 %) for both cases indicating good
membrane perfoimance. Figure 8.18 gives relative concentration of pentachlorophenol for
surfactant and distilled water wash. These analysis are analyzed by EPA support lab and were
semi-quantitative since the analysis of the samples were done after the required quality control
holding time. However, as expected, the results show a increase in pentachlorophenol leaching
. after surfactant wash and in both cases pentachlorophenol in permeate was below detection
limit (less than 1 mg/1).
Studies With Leachate After EDTA Wash
Experiments with Na4EDTA (Versene 100EP) had shown increased TOC and metal
concentration in the leachate as discussed in the section dealing in soil washing with Versene
1G0EP. Hence, the purpose of this study was to study the effect of EDTA wash on membrane
performance. 1.03 m moles/1 of Na^DTA^ap (1000 mg/1 Versene 100EP) was used for
washing SARM IV soil. The soil washing procedure and the filtration step was the same as
previously discussed. The filtered leachate was used as the feed for the membrane runs. Tables
8.19 and 8.20 summarize the results of the experiments done for this study. The rejection data
shown in Table 8.19 were calculated by Equation 3 and measured permeate concentration data.
83
-------
TABLE 8.16;
SUMMARY OF BATCH EXPERIMENTAL METAL RESULTS
WITH SURFACTANT WASHING
Surfactant Feed % r Rejection
Concentration Cu Ni Pb Zn Cu Ni : Pb Zn
% by wt. mg/i mg/1 mg/1 % % % %
0 1.58 10.5 1.72 224 17.8 96.2 93.1 96.5 93.3
0.04 1.93 14.9 2.31 386 17.0 >94.8 93.5 93.5 93.3
SARM IV ; Wash Solution:: 1:10
AP = 1.75 MPa
Surfactant = Triton X-100
84
-------
TABLE 8.17:
SUMMARY OF BATCH EXPERIMENTAL RESULTS WITH SURFACTANT WASHING
Surfactant % r Feed Permeate Retentate
Concentration TOC Cu pH TOC Cu pH TOC Cu pH
% by wt. mg/1 mg/1 mg/1 mg/1 mg/1 mg/1
0 17.8 101 1.58 5.15 27.5 0.06 3.15 87.2 2.0 4.4
0.04 17.0 94 1.93 5.9 26.0 <0.1 4.1 86.2 2.0 5.8
SARM IV :Wash Solution:: 1:10
AP= 1.75 MPa
Surfactant = Triton X-100
-------
TABLE 8.18:
EFFECT OF SURFACTANT WASHING ON MEMBRANE PERFORMANCE
Surfactant % Dissolved % r % Flux % D.W. Flux % TOC % Cu % Unaccounted
% by wt. TOC Cu Drop Recovery Rejection Rejection TOC
0 6.4 0.2 17.8 29.6 92.5 72.8 96.2 -24.2
0.04 5.5 0.2 17.0 41.6 80.0 72.3 >94.8 -19.2
SARM IV: Water:: 1:10
AP= 1.75 MPa
Surfactant = TritonX-100
-------
4
2
0
SystemrSARM IV Leachate
AP: 1.75 MPa
pH=5.15-6.0
% r = 17.0-17.8%
- No Triton X-100
* 400 ixig/1 Triton X—100
"1—'—I J" i
J
0
10
20
30 40 50 60
Time (miri)
70
60
90
Figure 8.17 Permeate flux vs. time for leachate after washing SARM IV with 0.04
% surfactant solution (400 mg/1 Triton X-100).
87
-------
20
CS
C
_o
I
C*
®
o
c
o
O
15
10
«1 ma'
Surfactant
Distilled Water
Figure 8.18 Relative concentration of peiuachlorophenol in feed and permeate for
distilled water and surfactant wash runs.
88
-------
Table 8.21 shows the effect of EDTA wash on the membrane performance. Figure 8.19
illustrates the effect of EDTA wash on the peimeate flux for a recovery of 17-18 %.
Table 8.20 shows that the TOC in leachate after EDTA wash was much higher as compared
to distilled water wash. 1.03 m moles/1 of EDTA (1000 mg/1 Versene 100EP) itself has about
133 mg/1 TOC but it is expected that some EDTA will get absorbed on the soil itself. Increase
in dissolved organics teachability has been reported after EDTA wash (PEI report, 1989).
Hewevers the objective was not to determine the effect of EDTA wash on organics teachability
but to study the effect on membrane performance. In light of the results reported in literature, it
is assumed that some increase in leachate TOC must be due to an increase in organics leaching.
Figure 8.19 shows that the permeate flux behavior for EDTA wash run was very similar to
distilled water wash ran (less than 2 % difference in flux drop). Table 8.21 shows that the %
unaccounted TOC from TOO balance of the membrane system decreased about 14 % for the
EDTA wash ran as compared to distilled water wash run. The decrease in unaccounted TOC
and flux drop (as compared to the expected increase with an increase in feed concentration)
shows that the use of EDTA actually enhanced the membrane performance. This suggests that
because of EDTA's ability to form metal-EDTA complexes, there was reduced membrane
fouling. The metal-EDTA complexes would be highly rejected by the membrane as confirmed
by the high metal rejections (97-99.5 %) shown in Table 8.19. Figure 8.20 gives relative
concentration of pentachlorophenol for EDTA and distilled water wash. These analysis were
semi-quantitative since the analysis of the samples were done by EAP support lab after the
required quality control holding time. However, as expected, the results show a increase in
pentachlorophenol leaching after EDTA wash and in both cases pentachlorophenol in permeate
was below detection limit (less than 1 mg/1). ¦
Studies With Leachate Spiked With Model Compounds
An experiment was also done with distilled water washed SARMIV leachate after spiking
the leachate with selected hazardous pollutants. The soil washing and filtration step was the
same as previously discussed. The filtered leachate was spiked to make a feed solution of
leachate + 2,4-dinitrophenoi (75 ppm) + 4-chloroaniline (50 ppm) 4- 4-armnobiphenyl (50 ppm).
The purpose of this study was to evduate the membrane performance in the presence of these
spiked compounds.
Table 8.22 shows the results of this study. The pH of leachate dropped from 6.1 to 3.6 after
spiking the selected compounds. Figure 8.21 shows the permeate flux behavior during the run
and distilled water flux after the run. The flux drop was only during the initial 10 minutes and
after that the flux almost remained constant. After the run, the distilled water flux recovered
about 92 % of the initial distilled water flux. This indicates that most of the increase in flux
drop was due to the increase in organics concentration of the feed. Figure 8.22 shows the
GC/MS ehromatographs of feed and permeate samples and the relative abundance of
4-chloroanilne, 2,4-dinitrophenoi, 4-aminobiphenyl and pentachlorophenol peaks. As expected,
Table 8.22 shows that the TOC rejection was high (greater than 90 %) and the area % count
rejections of 4-chloroaniline (greater than 90 %), 2,4-dinitrophenoi (greater than 98 %),
4-aminobiphenyl (greater than 92 %) and pentachlorophenol (greater than 98 %) were also
high. The focus here was to find relative concentrations in the feed and permeate samples so
that the membrane performance could be evaluated.
89
-------
TABLE 8.19: SUMMARY OF BATCH EXPERIMENTAL METAL RESULTS
WTIH CHELANT WASHING
Active Feed % r Rejection
Ingredient Cu NI Pb Zn Cu Ni Pb Zn
mmoles/1 mg/I mg/1 mg/l mg/l % % % %
0
1.58
10.5
1.72
224
17.8
96.2
93.1
96.5
93.3
0.10
20.7
14.4
8.19
372
18.7
>99.5
>99.3
>98.8
97.5
SARM IV: Wash Solution:: 1:10
AP = 1.75 MPa
Cheiant = Versene 100EF (TOC = 13.1% with 39 % Active Ingredient)
Active Ingredient =Na4EDTA.4H20
90
-------
TABLE 8.20: SUMMARY OF BATCH EXPERIMENTAL RESULTS WITH CHELANT WASHING
Active
% r
Feed
Permeate
Retentate
Ingredient
TOC
Cu
pH
TOC
Cu
pH
TOC
Cu
PH
mmoles/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
0
17.8
101
1.58
5.15
27.5
0.06
3.15
87.2
2.0
4.4
0.1
18.72
248
20.7
6.5
33
<0.1
5.1
261
25.8
6.4
SARM IV : Wash Solution:: 1:10
AP = 1.75 MPa
Chelant = Versene 100EP (TOC = 13.1 % with 39 % Active Ingredient)
Active Ingredient =Na4EDTA.4H20
-------
TABLE 8.21:
EFFECT OF CHELANT WASHING ON MEMBRANE PERFORMANCE
Active Ingredient % Dissolved % r % Flux % D.W. Flux % TOC % Cu % Unaccounted
mmoles/1 TOC Cu Drop Recovery Rejection Rejection TOC
0 6.4 0.2 17.8 29.6 92.6 72.8 96.2 - 24.2
0.10 6.9 1.7 18.7 30.8 84.6 86.7 >99.5 -11.9
SARMIV : Wash Solution:: 1:10
AP = 1.75 MPa
Chelant = Versene 100EP (TOC = 13.1 % with 39 % Active Ingredient)
Active Ingredient =Na4EDTA.4H20
-------
14
^ 12
HI
B
3 10
"*2
* 8
*
S
fe
6
4
2
0
0 15 30 45 80 75 90 105 ,120 135 150
Time (min)
Figure 8.19 Permeate flux vs. time for leachate after washing SARM IV with 1.03
mmoies/l of Na4EDTA..4H20 (1000 mg/1 Versene 100 EP).
93
3 3
« -k —# "k -6
System:SAKM IV Leachate
AP: 1.75 MPa
pH=5.15-6.1
% r = 17.0-18.7%
¦ Distilled Water Wash.
* 1000 rag/1 Versene Wash
-------
D5
E
cT
.2
eg
©
o
c
o
Q
< 1 mo/i
< 1 rnc/l
EDTA
Distilled Water
Figure 8.20 Relative concentration of pentachlorophenol in feed and permeate for
distilled water and EDTA wash runs.
94
-------
TABLE 8.22:
RESULTS OF SARM IV LEACHATE SPIKED WITH 50 MG/L
4-CHLOROANELINE. 75 MG/L 2,4-DINITROPHENOL AND 50
MG/L 4-AMINOBIPHENYL
Retention time
Feed
Permeate
min
Area counts
Area counts
4.2
35285554
3232762
6.3
5180358
< 100000
6.8
4208374
<100000
7.9
25625460
1899562
8.1
8795143
< 100003
Total Area Rejection > 92,79 %
4-Chioroaniline area rejection (4,2 min) > 90 %
2,4-Dinitrophenol area rejection (6.3 min) > 98 %
2,3,5,6-Tetrachlorophenoi area rejection (6.8 min) >91 %
4-Aminobiphenyl area rejection (7.9 min) > 92 %
Pentachlorophenol are rejection (8.1 min) > 98 %
Column : HP-1 (Crosslinked Methyl Silicone Gum) 12 m x 0.2 mm x 0.33 j± m film
thickness.
Operating Conditions: Temperature programming - hold at 70 C for 1 min then linear
increase to 200 C at 20 C/tnin, hold for 2 min, then linear increase to 250 C at 25
C/min and hold for 5 minutes; carrier gas - Helium; samtjle injection volume - 1 |i L,
solvent delay of 2 min and mass scan range from 35 to 400. ¦
Leachate: TOC = 89 mg/1, pH = 6.1
Feed: TOC = 190 rag/1, pH = 3.58
Permeate: TOC = 18.6 mg/1, pH =4.89
TOC rejection = 90.2 %
95
-------
a_ Syst.em:Spiked SARM IV Leachate
AP:1.75MPa
3 -j pH=4.89
% r = 17.2 %
^ 1 ¦ Permeate Flux
^ j * Post—run D.W. Flux
0
0 20 ' 40 60 80 100 120 140 180
Time (min)
Figure 8.21 Permeate flux vs. time for leachate spiked with 50 mg/14-chIoroaniline,
50 mg/1 4-aminobiphenyl, and 75 mg/L 2,4-dinitrophenol (SARM IV -
distilled water wash).
96
-------
feijuridsnca
1400000
1200000
1000000
800000
600000
400000 J
200000
.ss ->
4,120
TIC: S4SF15—2.0
7,92
I
4.2 min: 4-Chlorosniline
6.3 min: 2,4-Dinitrcphencl
6.8 min: 2,3 ,5,6-Tetraclilorophsncl
7.9 min: 4-Aminobiphenyl
8.1 rain: Pentachlorophenol
6.26
6.79
07
Feed
J
4.00
6.00
3.00
11.71
10 .00
12.30
14.00
Abundance
I 40000 -
!
35000
30000
25QQ0
20000
150QQ
10000
S000
0
Sine ->
4.119
TIC: S4SF152.D
Permeate
7.SI
3.00 4.00 S.00 6.00 7.00 3.00 5.00 10.00 11.00
Figure 8.22 GC/MS chromatographs of 4-chIoroanilIne, 4-aminobiphenyi,
pentachlorophenol, and 2,4-dinitrophenoI (spiked in SARM IV -
distilled water wash) in feed and permeate.
97
-------
S todies With Leachate At Different pH Values
The leachate has metals (Ca and other alkaline earth metals found naturally in soil; Pb, Ni,
Cu, Zn and other heavy metals) and the solubility of these metals is dependent on pH. Hence,
the purpose of this study was to study the effect of leachate pH on membrane performance.
SARM IV was washed with distilled water. The soil washing and filtration step was the same
as previously discussed. Samples of the filtered leachate were taken and stored at 4° C. The
leachate was divided into three and used to make runs at three different pH values. The runs
were made for a recovery of 21-23 %. 5,1 ml of 1 N NaOH was added to 550 ml of the
leachate to adjust the pH to 8.18. pH adjustment to 8.18 resulted in the formation of precipitate.
This was expected because of low solubility of most heavy metals at high pH, The precipitate
was filtered out and the filtrate was used for the ran. For pH 3 run, 0.8 ml of 1 N Ii,S04 was
added to 500 ml of the leachate to adjust the pH to 3.03. pH adjustment to 3.03 resulted in the
formation of slight white cloudiness. This was expected because of formation of calcium
carbonates (at pH 3) from calcium bicarbonates that might have leached from the soil. This
was also filtered and the filtrate was used for the run.
Tables 8.23 and 8.24 summarize the results of this study. As compared to feed TOC at pH
6.5, the feed TOC at pH 3.03 was 21 % lower and the feed TOC at pH 8.18 was 29 % lower.
This suggests some loss of TOC with the precipitate. Table 8.24 and Figure 8.23 show that as
compared to % flux drop at pH 6.5, the % flux drop at pH 3.03 was 9 % higher and the % flux
drop at pH 8.18 was 4 % higher. Also Table 8.23 shows that pH of retentates for pH 3 and pH
8 runs remained around pH 3 and 8 respectively. This indicates that both sulfuric acid and
sodium hydroxide added to adjust pH were being rejected by the membrane. The drop in flux
for pH 8 run could be due to some precipitation of heavy metals on the membrane surface and
the drop in flux for pH 3 run could be due to some precipitation of calcium carbonates at low
pH. Table 8.24 shows that the TOC rejections for different pH were almost same (within 2 %)
indicating that pH did not affect significantly the interactions of the organics and the
membrane.
Studies After Leachate Ozonation
Feed pre-ozonation of chlorophenol-chloroethane mixtures has been shown to significantly
reduce flux drop and improve overall TOC rejections (Williams and Bhattacharyya, 1989).
Hence, the purpose of this study was to evaluate ozonatioo-membrane process to produce high
quality permeate water which could be recycled back for soil washing.
SARM IV was washed with distilled water. The soil washing procedure and filtration step
was same as previously discussed. Samples of the filtered leachate were taken and stored at 4 *
C. The leachate was divided into three and used for ozonation for three different times (0, 10
and 30 min). One part (515 ml) of filtered leachate was ozonated for 10 min and one of the
other parts (485 ml) was ozonated for 30 min using the procedure described in the section on
Experimental Procedures. The pH dropped after ozonation to 3.3-3.4 and the ozonated
leaehates were stirred overnight to remove any residual ozone. After ozonation, the color of
¦ leachate changed to brownish and a brownish precipitate was formed. The precipitate was
removed by vacuum filtration and the filtered ozonated leaehates were pH adjusted to 6.5 to be
used as feed for the membrane runs. Tables 8.25 and 8.26 compare the results of membrane
runs with leaehates after different ozonation times. There have been numerous studies on the
oxidation of organic compounds by ozone. The studies with ozonation of phenol derivatives
(Wang et al, 1989) and benzene derivatives (Rice and Browning, 1980) have shown formation
of mostly organic acids as intermediates and TOC removal by carbon dioxide formation. It
98
-------
TABLE 8.23:
SUMMARY OF BATCH EXPERIMENTAL RESULTS WITH DIFFERENT pH OF FEED
-------
TABLE 8.24:
EFFECT OF FEED pi I ON MEMBRANE PERFORMANCE
Feed Feed % r % Flux % D.W. Flux % TOC % Conductivity % Unaccounted
pH TOC Drop Recovery Rejection Rejection TOC
3.03 94.7 20.7 33.3 95.3 86.6 90.0 - 14.2
6.50 121 22.9 23.7 95.3 87.5 91.1 -23.9
8.18 86.3 20.6 27.2 95.4 86.8 95.4 - 17.7
SARMIV: Water:: 1:10
AP = 1.75 MPa
-------
14:
12-
£ 10:
o
SystemrSARM IV Leachate
AP: 1.75 MPa
% r = 21-23 %
* pH 3 (T0C=94.7 mg/1)
¦ pH 6 (TGC=121 mg/1)
~pH 8 (TGC=86.3 mg/1)
0
50
10
20
30
40
60
70
80
Time (min)
Figure 8.23 Permeate flux vs. time for leachate at different pH values (SARM IV -
distilled water wash).
101
-------
TABLE 8.25: SUMMARY OF BATCH EXPERIMENTAL RESULTS WITH FEED OZONATION
Ozonation % r Feed Permeate Retentate
Time TOC k pH TOC k pH TOC k pH
mg/1 mmho/cm mg/1 mmho/cm mg/1 mm ho/cm
0 22.89 121 1.75 6.50 15.1 0.155 3.76 115 2.08 5.94
10 21.11 84.4 6.53 1.82 12.7 0.74 5.95 84.4 6.20 2.15
30 22.22 68.4 6.41 1.85 12.1 0.388 3.15 73.2 6.46 2.19
SARM IV: Water:: 1:10
AP = 1.75 MPa
-------
TABLE 8.26: EFFECT OF FEED OZONATION ON MEMBRANE PERFORMANCE
Ozonation Feed % r % Flux % D.W. Flux % TOC % Conductivity % Unaccounted
Time min TOC Drop Recovery Rejection Rejection TOC
0 121 22.9 23.7 95.3 87.5 91.1 -23.9
10 84.4 21.1 5.0 100 84.9 95.9 - 17.9
30 68.4 22.2 5.0 100 82.3 79.0 - 12.8
SARM IV: Water:: 1:10
AP = 1.72 MPa
-------
should be understood that the focus here was not to study ozonation but to evaluate its effect on
the membrane process. Hence, the 30 % reduction in TOC after 10 rain of ozonation and 43 %
reduction in TOC after 30 min of ozonation as shown in Table 8.25 could be accounted for by
carbon dioxide formation and some TOC removal with precipitate. The drop in pH from 6.5 to
3.3-3.4 found after ozonation indicates some formation of acids. Figure 8.24 compares the
permeate flux for leachate after different ozonation times. As indicated in Table 8.26, there was
only 5 % flux drop after ozonation as compared to 24 % flux drop with no ozonation.
Moreover, after the ran, distilled water flux recovered to its initial value for runs with ozonated
feeds as compared to 95 % recovery for non-ozonated feed. The flux behavior for 10 and 30
min feed ozonation was found to be same. These results suggest that ozonated products do not
interact strongly with the membrane as indicated by reduction in flux drop and % unaccounted
TOC from mass balance of the membrane system. The % TOC rejections for the FT30
membrane was found to be almost constant (82-87 %) but the overall TOC rejections for the
ozonation-membrane process were found to be high: 87.5 % for 0 min, 89.5 % for 10 min and
90 % for 30 min ozonation.
Figure 8.25 shows the GC/MS ehromatographs for feed and permeate samples and the
relative abundance of the peaks for non ozonated leachate, and Figure 8.26 shows the GC/MS
ehromatographs of leachate after ozonation for 5 and 30 min. Table 8.27 gives the % area
count rejections for the observed peaks for non ozonated leachate run and also shows the %
removal after ozonation. Xylene was found to be rejected more than 81 %, styrene and
ethylbenzene were not detected in the permeate samples. After 10 and 30 min ozonation
styrene, ethylbenzene and xylene were not detected in the samples indicating formation of
either some intermediates or carbon dioxide.
Studies At Different Flow Rates
The purpose of this study was to study the permeate flux behavior at different flow rates.
The setup for the continuous unit has already been explained in the section on Experimental
Procedures. Standard NaCl rejections were done for the membrane and were found to be 96.3
%. SARMIV was washed with distilled water and the supernatant was used as the feed for the
run. Suspended solids and dissolved solids content in the feed were determined using the
standard methods discussed in the section on Analytical Procedures.
Figure 8.27 shows the permeate flux behavior at different flow rates. The % flux drop for
batch ran for distilled water washed leachate with suspended solids (175 mg/1) was around 22
% for a recovery of 20 %. Amount of suspended solids for the continuous run was determined
to be 126 mg/1 and TDS was determined to be 1,420 mg/1. As explained in the discussion on
high recovery ran, TDS alone accounts for approximately 7 % flux drop for 20 % recovery (out
of a total 15 % flux drop). The total flux, drop for continuous ran was lower than for batch run.
Since, for batch system and for very low flow rates for a continuous run more solute is
accumulating at the membrane wall, there seems to be some relation of wall concentration to
pore blocking. As more solute reaches the membrane pore, there is a reduction of available
path for water resulting in increased flux drop (Jevtitch. and Bhattacharyya, 1986).
104
-------
14 :
m
£
a
SystemiSARM IV Leachate
AP: 1.75 MPa
ph=8.5
A no ozonation (T0C=121 mg/1)
¦ 10 min ozonation (T0C=84 mg/1)
^30 min ozonation (T0C=S8 mg/1)
30
50
60
0
10
20
40
70
Time (min)
Figure 8.24 Permeate flux vs. time for leachate ozonated for different times (SARM
IV - distilled water wash).
105
-------
MDunaarsce
120000
100C0G
80000
60GGG
40000
20000
Jims -> 9.00
is
8.92
KEHBRANE PEED: SASH IV 8.9 sin: Ethylbenzene
9 m sxn« bx— * p-Xylene
10.12
io. bo
10.0 min: Styrsne
10.1 mln: o-Xylene
10.00 11.00 12.00 13.00 14.00
iUsunsiaiice
1
45000 •"
MEMBRANE PERMEATE: SARM IV
9.2 sin: m-*,p-Xylene
10,1 min; a-Xyleme
40000
33000
9.19
30Q0Q
10.13
20QG0
15000
10000
10.00 11.00 12.00 13.00 14.00
lias -> 9 .00
Figure 8.25 GC/MS cbromatograpbs of xylene, ethylbenzene, and styrene in feed
and permeate before ozonation (SARM fV-disalled waterwash).
106
-------
TIC: GZ-lQ-f.D
5G000
Ozonation Tine = 10 min
SA5M IV Leachate
40000 -i
Peaks below detection limit
3GQQC -
20000
10000 -
10.00
12.00
14,00
16,00
iAbundance TIC: QZ-30-F.D
Ozonation Time = 30 lain
soqoq -
SASH IV Leachate
Peaks below detection limit
60000
40000
20000
12'. 00
Tiaa -»
8.00
10.00
16.00
6.00
14.00
Figure S.26 GC/MS chromatographs of xylene, ethylbenzene, and styrene in feed
after ozonation (SARM IV-distillcd waterwash).
107
-------
TABLE 8.27: GC/MS ANALYSIS OF XYLENE, ETHYLBENZENE AND
STYRENE IN FEED AND PERMEATE OF SARM IV
LEACHATE FOR OZONATION STUDIES
Description
Ozonation Feed
Time, min Area counts
Permeate
Area counts
% Rejection
Ethylbenzene
0
4471305
ND*
>97%
(8.9 min)
10
ND*
ND*
-
30
ND*
ND*
-
m,p- Xylene
0
7523189
980014
87%
(9.2 min)
10
ND*
ND*
-
30
ND*
ND*
-
Stynene
0
945656
ND*
> 89 %
(10.0 min)
10
ND*
ND"
-
30
ND*
ND*
ND = Not detected (< 100000 area counts)
Column : HP-5 (Crosslinked 5 % Phenyl Methyl Silicone ) 25 m x 0.2 mm x 0.5 )i m
film thickness.
Operating Conditions: Temperature programming - hold at 50 C for 1 min then
linear increase to 70 C at 20 C/min, hold for 2 min, then linear increase to 90 *C at
5 * C/min and hold for 10 minutes, then linear increase to 200 ° C at a rate of 20
C/min and Final hold for 5 min. earner gas - Helium; sample injection volume - 1 (i L,
solvent delay of 5 min and mass scan range from 5 to 400.
108
-------
14
"ji
o
X
X
p
fc
12
E
^ 10 -j
2
0
o
pH = 6,15
Dissolved Solids: 1,420 m.g/1
Suspended Solids: 126 mg/1
System:SAHM IV Leachate
AP:1.75 MPa
SoihWater:: 1:10
% r = 58.2 %
* Re=2,l26 (Q=0.95 1/min
¦ Re= 1,275 (Q = 0.57 1/min)
"&Ee=425 (Q=0.19 1/min)
10 20 30 40
% Recovery
50
80
70
Figure 3.27 Permeate flux vs. time for the continuous ran with suspended solids at
different flow rates (S ARM IV - distilled water wash).
109
-------
Summary
Figure 8.28 summarizes the membrane performance with leachates obtained by washing
SARM IV with different wash solutions and also of ozonated leachates. The membrane
performance was compared for % flux drop, % TOC rejections (% R TOC), % conductivity
rejections (% R Con<±), and % unaccounted TOC from the mass balance (% Unacc. TOC). The
distilled water flux for all the experiments were in the range of 9-12 x 1(T* em's and the pH of
the membranes feeds were in the range of 5.8-6.8. 1.01 mol/1 EDTA was used for EDTA wash,
0.04 % surfactant was used for surfactant wash, and ozonation time was 10 min for the
ozonation experiment. The results show that the TOC and conductivity rejections were high
and almost constant for all the experiments. Ozonation of leachates resulted hi minimal flux
drop suggesting formation of intermediates that do not interact strongly with the membrane.
The EDTA wash had the least % unaccounted TOC suggesting that the EDTA-metal complexes
reduce adsorption of organics on the membrane surface.
110
-------
c
CD
E
©
CL
0 EDTA
Surfact
D.W
0 Ozone
% Flux drop
% R TOO
% R Cond. % Unacc. TOG
Figure 8.28 Summary of membrane performance with different leachates (SARM
IV).
Ill
-------
SECTION 9
MEMBRANE WATER FLUX PREDICTION AND CORRELATION
OVERVIEW
The objectives of this section are to explain solute adsorption phenomena by mathematical
equations, and use them in predicting solvent flux or flux decline based on concentration on the
high pressure side of the membrane. The assumptions made include negligible concentration
polarization conditions and no back-diffusion of the solute. The concentration at the solution
membrane interface on the high pressure side of the membrane was considered to be same as
the bulk concentration. These assumptions can be made because high mixing conditions were
maintained in the cell to avoid concentration polarization.
ADSORPTION RESISTANCE TERM
Some organic solutes can adsorb on the membrane surface and in the pores. The adsorbed
solutes nmy cause large drops in water flux due to reduction in the path available for water
transport through the membrane. The drop in flux drop caused by organic adsorption has been
by modeled by Bhattaehaiyya et al. (1987) by including an adsorption resistance term in the
solution-diffusion model:
m Ads
where Rm is the membrane resistance (Rffl=l/A) and RAds is the resistance due to solute
adsorption on the membrane. For the case of no solute adsorption, equation (21) reduces to
equation (6). The adsorption resistance term can be found by rearranging equation (21) to
R = ( AP - Arc ) _ R (22)
Acs J m ^ '
w
The organic solutes and total dissolved solids (IBS) in the solution contribute to the osmotic
pressures. For the soil wash leachates, the osmotic pressure due to organic solutes were
negligible (< 0.007 MPa). The osmotic pressure due to total dissolved solids can be
approximated by 0.07 MFa/1000 mg/l of total dissolved solids (Weber, • 1985) and dissolved,
solids are rejected almost 99.99 % by the FT30 membrane. The osmotic pressure of this system
for water recovery r is thus given by
. TDS x 0.001 x 0.07
Ait = (23)
112
-------
. The water flux Jw, is determined from the system studied, and the membrane resistance, 1/A, is
found from the DWF (DWF = A ( AP). Thus, the adsorption resistance term at recovery r
reduces to
D 1 , ao TDS x 0.001 x 0.07 . AP
R., =— ( AP ) - (24)
Ms j V 1 — r DWF
W
Table 9.1 list results found for the adsorption resistance term for a high recovery ran with
distilled water washed SARM IV leachate. The concentration in the high pressure side of the
membrane was calculated from initial feed concentration and permeate concentration at
different recoveries. Higher adsorption resistance terms were found for higher concentration.
This was expected because there was a larger flux drop at higher concentrations.
CORRELATION OF ADSORPTION TERM WITH BULK CONCENTRATION
Correlation of adsorption resistance term with bulk concentration would be useful because it
could be used for predicting solvent flux drop based on desired bulk concentration. Even
though such a correlation would be system specific, it would give insight into how the flux drop
could be modeled for any such system. Since, the adsorption resistance term is a measure of
solute adsorption on the membrane surface, and the bulk concentration is same as the
adsorbable solute equilibrium concentration under negligible concentration polarization
conditions, the adsorption resistance term could be related to bulk concentration as a form of
Freundlicfa adsorption isotherm. The Freundlich adsorption equation is perhaps the most
widely used mathematical description of adsorption in aqueous systems (Faust and Aly, 1987).
It should be noted, however, that Rsds may be a function of pH also, but this system was around
pH 6 and all solutes were non-ionized. The effect of pH could be correlated in a similar
fashion. Thus, the adsorption resistance could be expressed as
Ra* =a cb"° <25>
where 'a* and 'l/n' are constant characteristics of the system. For linearization of the data, the
above equation could be written in logarithmic form
In (RAds) = In (a) + j- In (CB) (26)
This linear equation can then be used to find the slope l/n and intercept In (a). Since, the bulk
concentration influences but is not influenced by the adsorption resistance term, both variables
are wel related over the range by a linear model and the errors in determining both the
variables are statistically independent, linear regression analysis is appropriate here. Figure 9.1
shows the linear fit of ln(RAl33) with ln(CB) using least-squares estimators method. This method
minimizes the residual sum of squares (Mason et al., 1989). For linear regression, the root
mean square, also called standard deviation, was small (0.12969) as compared to the average
response (-2.81652) indicating that the observed responses are tightly clustered around the
fitted line. C.V., the coefficient of variation, was around 4 % indicating a good fit. The
R-square value was high, around 88 %. indicating strong linear association of the variables.
The standard error for the estimated parameters were small. From the estimated slope and
intercept, the constants 'l/n' and 'a' of the system were found to be —¦ = 0.663926
n
113
-------
TABLE 9.1: ADSORPTION RESISTANCE TERM CALCULATED FROM
EXPERIMENTAL VALUES
r
Flux xlO4
Ra„
Cb
%
cia/s
MPas/cm
TOC mg/l
1.29
8.55
0.0413
¦ 183
9.82
8.12
0.0503
197.1
20.5
7.76
0.0578
241.6
39.75
7.85
0.0503
276.4
79.93
4.62
0.1200
721.8
114
-------
1.50-
1.65 -
1.80 -
1.95-
2.10-
¦ •
2.25-
2.40-
2.55-
2.70-
y/ m
2.85-
a i
3.00-
3.15-
3.30-
3.45-
* ¦
AP:1.75 MPa
pH=8.3
TDS=i420 mg/1
¦ Experimental
—Regressed
3.60-
t 1 i 1 i 1 i 1 r
4.75 5.00 5.25 5.50 5.75 6.00 8.25 6.50 8.75 7,00
log (Bulk Concentration)
Figure 9.1 Log( RAd3 ) vs. log (Bulk concentration).
115
-------
and 'a' = 0.001406. Using equation (26) and the calculated values of 'a' and *l/n', RAds can be
calculated at different values of bulk concentration. Figure 9.2 shows the RA
-------
£
Q
W
50
ft,
3
rt
©
m
<8
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
¦
a 1
m
¦ a
SystemrSARM IV Leachate
¦
A.F:1.75 MFa
pH=8.3
TDS—1420 mg/1
¦ Experimental
-Calculated
100 200 300 400 500 600
Bulk Concentration. (TOC mg/1)
700
800
Figure 9,2 Experimental and calculated vs. bulk concentration.
117
-------
70
60
o, 50
o
Sm
Q
X
E 40
30
20
100 200 300 400 500 600 700 800
Bulk Concentration (TOC mg/1)
m
/ ¦
SystemrSARM IV Leachate
¦
/ g
AP:1.75 MPa
n /
pK=6.3
TDS=1420 mg/1
¦
¦ Experimental -
— Predicted
Figure 9.3 Experimental and predicted % flux drop vs. bulk concentration.
118
-------
14
13
12
m
o
x
*
S3
8
7
6
5
4
3
2
1
0
¦
¦
m «
a
pH=8.3
SystemrSARM IV Leachate
AP:1.75 MPa
TDS=1420 mg/1
¦ Experimental
-Predicted
0
10
20 30
40 50 60
% Recovery
70
80 90 100
Figure 9.4 Experimental and predicted water flux with % recovery.
119
-------
REFERENCES
Abdul, A.S. and Gibson, T.L., "Laboratory Studies of Surfactant-Enhanced Washing of
Poly chlorinated Biphenyl From Sandy Material", Environ. Sci. Technol,, 25,665, (1991).
Amdurer, M.R., Fellman, R.T., Roetzer, J., and Russ, C, "Systems to Accelerate In Situ
Stabilization of Waste Deposits", EPA Report. EPA/540/2-86/002, (1986).
Andrews, A.T., Ahlert, R.C, and Kosson, D.S., "Supercritical Fluid Extraction of Aromatic
Contaminants from a Sandy Loam Soil", Environmental Progress, 9, pp. 204-210, (1990).
Arienti, M., Wilk, L., JasinsM, M., and Prominski, N., "Technical Resource Document:
Treatment Technologies for Dioxin-Containing Waste, EPA Report EPA/600/2-86/096,
(1986).
Ball, R.,Wolf, S., "Design Considerations for Soil Cleanup by Soil Vapor Extraction",
Environmental Progress, 9, 187, (1990).
Basu, S.. Sapkal, V., "Membrane Separation Technique in Simplification of Soda Recovery
Process in Pulp and Paper Industry", Desalination, 67:371, (1987).
Belfort, G., Synthetic Membrane Processes: Fundamentals and Water Application, Academic
Press, Inc., New York, NY (1984).
Bhattachaiyya, D., Garrison, K., and Grieves, R., "Membrane Ultrafiltration for Treatment and
Water Reuse of TNT-Manufacturing Wastes", Journal Water Pollution Control
Federation, Vol. , 800 (1977).
Bhattachaiyya, D., Moffitt, M., and Grieves, R., "Charged Membrane Ultrafiltration of Toxic
Metal Oxyanions and Cations from Single- and Multisalt Aqueous Solutions", Separation
Science and Technology, 13,449 (1978).
Bhattacharyya» D., Shelton, S., and Grieves, R., "Charged Membrane Ultrafiltration of Multisalt
Systems: Application to Acid Mine Waters", Separation Science and Technology, 14,
193 (1979).
Bhattachaiyya, D., Jevtitch, M., Ghosal. J., and Kozminski, J., "Reverse Osmosis Membrane
for Treating Coal-Liquefaction Wastewater", Environmental Progress, 3,95 (1984).
Bhattachaiyya, D., and Cheng, C., "Separation of Metal Chelates by Charged Composite
Membranes", Recent Developments in Separation. Sciences, vol. IX, pp. 107-133, CRC
Press, Florida (1986).
Bhattachaiyya, D., Barranger, T., Jevtitch, M., and Greenleaf, S.s "Separation of Dilute
Hazardous Organics by Low Pressure Composite Membranes", EPA Report
EPA/600/87/053 (1987).
Bhattachaiyya, D., Adams R., and Williams M., "Separation of Selected Inorganic Solutes by
Low Pressure Reverse Osmosis Membranes", Biological and Synthetic Membranes, pp.
153-167, Alan R. Liss, Inc., New York, NY (1989)
120
-------
Bindoff, A., Davies, C., Kerr, C., and Bukley, C., "The Nanofiltration and Reuse of Effluent
from the Caustic Extraction Stage of Wood Pulping", Desalination, 67,453 (1987).
Bird, E.B., Stewart, W.E., Lightfoot, E.N., Transport Phenomena, John Wiley & Sons, Inc.,
(1960).
Bruschke, H.E.A., "State of The Art Of Pervaporation", in Proceedings of 3rd International
Conference on Pervaporation in the Chemical Industry, Englewoood, N.J. (1988).
Buckley, C., Tref&y-Goatley, K., Simpson, M., Bindoff, A., Groves, G., "Pretreatment,
Fouling, and Cleaning in the Membrane Processing of Industrial Effluents", ACS
Symposium Series N. 261, New York, (1985).
Cadotte, J., Peterson, R., Larsen, R., and Erickson, E., "A New Thin-Film Composite Seawater
Reverse Osmosis Membrane", Desalination, 32,25 (1980).
Cadotte, J., Forester, R., Kim, M., Petersen, R., and Stocker, T., "Nanofiltration Membranes
Broaden the Use of Membrane Separation Technology", Desalination, 70,77 (1988).
Cartwright, P., "The Application of Membrane Separation Technologies to Industrial
Processes", 43rd Purdue Industrial Waste Conference Proceedings, Chelsea, Michigan:
Lewis Publishers, Inc., (1989).
Cecille, L., and Toussaint, J., Future Industrial Prospects of Membrane Processes, New York,
Elsevier, (1989).
Chakravorty, B., "Effluent Treatment by Membrane Technology - Opportunities and
Challenges", Biological and Synthetic Membranes, pp. 169-180, Alan R. Liss, Inc., New
York, NY (1989)
Chakravorty, B„ and Srivastava, A., "Application of Membrane Technologies for Recovery of
Water from Pulp and Paper Mill effluents", Desalination, 67:363, (1987).
Chian, E., Bruce, W., and Fang, H., "Removal of Pesticides by Reverse Osmosis",
* Environmental Science and Technology, 9,364, (1975).
Cullinane, M.J., Jr., Jones,L.W., Malonc, P.G., "Handbook of Stabilization / Solidification of
Hazardous Waste", EPA Report, EPA/540/2-86/001, (1986).
Dev, H, Sresty, G.C., Bridges, J.E., and Downey, D., "Field Test of the Radio Frequency in In
Situ Soil Decontamination Process", In: Superfiind'88, Proceedings of the 9th National
Conference, The Hazardous Materials Control Research Institute, Silver Spring, MD, pp.
498-502, (1988).
Dickson, J., Matsuuia, T., Blais, P., and Sourirajan, S., "Reverse Osmosis Separations of Some
Organic and Inorganic Solutes in Aqueous Solutions Using Aromatic Polyamide
Membranes", Journal of Applied Polymer Science, 19,801 (1975).
DOE, "Membrane Separation Systems", DOE Report, DOE/ER/30133-H1, P5-1,1990
Dorcia, J., "Ultrafiltration of Bleach Plant Effluents - A Pilot Plant Study, Journal of Pulp and
Paper Science, 12 172 (1986).
121
-------
Downey, D.C., Elliott, M.G., "Perfomiane of Selected In Situ Soil Decontamination
Technologies: An Air Force Perspective", Environmental Progress, 9, 169, (1990).
Drioli, E., and Nakagaki. M., Membrane and Membrane Processes, New York, Plenum Press,
(1986).
Dyke, C.A., Bartels, C.R., "Removal of Organics from Offshore Produced Waters Using
Nanofiltration Membrane Techmology", Environmental Progress. 9 183, (1990).
Edwards, V. and Schubert, P., "Removal of 2.4-D and Other Persistent Organic Molecules from
¦ Water Supplies by Reverse Osmosis", Journal AWWA. (Oct, 1974).
Ellis, W.D., Payne, J.R., and McNabb G.D., "Treatment of Contaminated Soils with Aqueous
Surfactants". EPA Report. 600/52-85/129, (1985).
EPA, "Handbook on In Situ Treatment of Hazardous Waste-Contaminated Soils", EPA Report,
EPA/540/2-90/002, (1990).
Eriksson, P., "Nanofiltration Extends the Range of Membrane Filtration", Environmental
Progress, 7,58 (1988).
Esposito, P., Kessling, J., Locke, B.B., Taylor, M., Szabo, M., Thumau, R., Rogers, C., Traver,
R., and Earth, E., "Results of Treatment Evaluations of a Contaminated Synthetis Soil",
JAPCA, 39:294-304, (1989).
Fitzpatrick, V.F., Timmerman, C.L., Buelr, J.L., "In Situ Vitrification - An Innovative Thermal
Treatment Technology", In: Proceedings. 2nd International Conference on New Frontiers
for Hazardous Waste Enginnering Research Laboratory, Cincinnati, OH,
EPA/600/9-87/018F, pp. 305-322, (1987).
Fox, R.D., Alperin, E.S., Huls, H.H., "Thermal Treatment for the Removal of PCbs and other
Organics from Soil", Environmental Progress. 10 40, (1991).
Franson. M. (ed.)5 Standard Methods for the Examination of Water and Wastewater. 16, APHA,
AWWA, WPCF, Washington, DC (1985).
Fumkawa, T., and Brindley, G.W., "Adsorption and Oxidation of Benzidine and Aniline by
Montmorillonite and Hectorite", Clay Miner, 21:279-288, (1973).
Hess, M., Jones, G., Micheletti, W., and Tomlinson, J., "Wastewater Concentration by Reverse
Osmosis: a Field Demonstration in the Electric Power Industry", Environmental Progress,
7:7,(1988).
Hsiue, G., Pung, L., Chu, M., and Shieh, M., "Treatment of Uranium Effluent by Reverse
Osmosis Membrane", Desalination, 71 -.35, (1989).
Hutzler, N., Murphy, B.E., and Gierke, J.S., "State of Technology Review - Soil Vapor
Extraction Systems", EPA Report, EPA/600/2-89/024, (1989).
Imasu, K., "Wastewater Recycle in the Plating Industry using Brackish Water Reverse Osmosis
Elements, Desalination, 56; 137, (1985).
122
-------
Jonsson, A,, and Wimmerstedt, R., "The Application of Membrane Technology in the Pulp and
Paper Industry", Desalination. 53:181, (1985).
Kedem, O., Katchalsky, A., "Thermodynamic Analysis of the Permeability of Biological
Membranes to Non-Electrolytes", Biochem. Biophys. Acta., 27,229-246, (1958).
Kurihara, M., Harumiya, N., Kanamaru, N., Tonomura, T., and Nakasatomi, M.. "Development
of the PEC-1QG0 Composite Membrane for Single-StEge Seawater Desalination and the
Concentration of Dilute Aqueous Solutions Containing Valuable Materials", -
Desalination, 38 449 (1981).
Lipski, C, Cote, P., "The Use of Pervaporation for the Removal of Organic Contaminants from
Water", Environmental Progress, 9, pp. 254-261, (1990).
Lonsdale, K., Merten, U., and Riley. "Transport Properties of Cellulose Acetate Osmotic
Membranes". Journal of Applied Polymer Science. 9.1341 (1965).
Matsuura, T., and Sourirajan, S,, "Reverse Osmosis Transport Through Capillary Pores Under
the Influence of Surface Forces", Industrial Engineering Chemical Process Design and
Development 20,273 (1981).
Nash, J.H., "Project Summary: Field Studies on In Sith Soil Washing", EPA Report
EPA/600/S2-87/110, (1989).
Nielson, R.K., and Cosmos, M.G., "Low Temperature Thermal Treatment of Volatile Organic
Compounds from soil: A Technology Demonstrated", Environmental Progress. 8, 139,
(1989).
Nusbaum, I., and Argo, D., "Design, Operation, and Maintenance of a 5-mgd Wastewater
Reclamation Reverse Osmosis Plant", Synthetic Membrane Processes: Fundamentals and
Water Applications, ed. Belfort, G., New York: Academic Press, Inc., (1984).
Parekh, B„ Reverse Osmosis Technology. New York: Marcel Dekkar, Inc., (1988).
Pfaeiffer, T.H., Nunno, T.J., Walters, J.S., "EPA's Assessment of European Contaminated Soil
Treatment Techniques", Environmental Progress, 9, pp. 79-86, (1990).
Puseh, W.,"Measurement Techniques of Transport Through Membranes",
Desalination, 59:105. (1986).
Rautenbach, R., and Aibrecht, R., Membrane Processes, John Wiley & Sons Ltd., (1989).
Rice, R.G., "Purification and Recycling of Groundwater Contaminated with Petroleum Products
and Cyanide - The Karlsruhe (Federal Republic of Germany) Drinking Water Treatment
Plant", In: The 5th National Conference on Management of Uncontrolled Hazardous
Waste Sites. November 7-9, Washington, D.C., (1984).
' Robison, G., "Recovery Pays off for Chicago Job Shop Plater", Products Finishing, (June,
1989).
Sanning, D.E., and Black, M.I., "Protecting the Environment at Supcrfund Sites Through
Chemistry". EPA Report. EPA/60Q/D-87/222, (1987).
123
-------
Sherwood, T., Brian, P., and Fisher, R., "Desalination by Reverse Osmosis", Industrial and
Engineering Chemistry Fundamentals, 6,2 (1967).
Shuckrow, A., Pajak, A., Osheka, J,, "Concentration Technologies for Hazardous Aqueous
Waste Treatment", EPA Report, EPA-600/S2-81-019 (1981).
Siier, I. and Bfaattscfaaryya, D., "Low Pressure Reverse Osmosis Membranes: Concentration
and Treatment of Hazardous Wastes", Hazardous Waste and Hazardous Materials, 2, 45
(1985).
Slater, C., Ahlert, R., Uchrin, C., "Applications of Reverse Osmosis to Complex Industrial
Wastewater Treatment", Desalination, 48 171 (1983).
Slater, C., Ahlert, R., and Uehrin, C., "Reverse Osmosis Process for the Renovation and Reuse
of Hazardous Industrial Wastewaters", Current Practices in Environm. Sci, 3:1, (1987)
Slater, C, Ferrari, A., WisniewsM, P., "Removal of Cadmium from Metal Processing
Wastewaters By Reverse Osmosis", Journal of Environmental Science and Health, A
22:707, (1987)
Soltanieh, M. and Gill, W., "Review of Reverse Osmosis Membranes and Transport Models",
' Chemical Engineering Communications, 12,279 (1981).
Sourirajan, S., Reverse Osmosis, Academic Press, New York, NY (1970).
Souri? ¦ajan, S., and Matsuura, Reverse Osmosis/Ultrafiltraaon Principles, National Research
Council of Canada Publications, Ottawa, Canada, K1AOR6 (1985).
Staiff, D.C., Butler, L.C., and Davis, IE., "A Field Study of the Chemical Degradation of
Paraquat Dichloride Following Simulated Spillage on Soil", Bull Environ. Contam.
Toxicol., 26:16-21,1981
Traver, R.P., PEI Associates, Inc., "Development and Use of EPA's Synthetic Soil Matrix",
EPA Report, Risk Reduction Engineering Laboratory, Edison, N.J., (1989).
Truett, J.B., Holberger, R.L., and Sanning, D.E., "In Sim Treatment of Uncontrolled Hazardous
Waste Sites",' Proceedings of the 3rd National Conference, the Hazardous Materials
Control Research Institute, Silver Springs, MD, pp. 451-457, (1982).
USEPA, "Technology Evaluation Report: SITE Program Demonstration Test, International
Waste Technologies, In Situ Stabilization/ Solidification", EPA Report
EPA/54G/5-89/004a, (1989).
Wang, Y., Pai, P., and Latchaw, J., "Effects of Preozonation on the Methanogenic Toxicity of
2,5-Dichlorophenol", Journal Water Pollution Control Federation, 61, 320 (1989).
Weber, W.J., Jr., Phvsicochemical Processes for Water Quality Control, John Wiley & Sons,
(1972).
Wijmans, J.G., Kaschemekat, J,, Davidson, J.E., and Baker, R.W., "Treatment of
Organic-Contaminated Wastewater Streams By Pervaporation", Environmental Progress,
9, pp. 262-268, (1990).
124
-------
Williams, M., "Separation and Purification of Dilute Hazardous Qrganies By Ozonation-Low
Pressure Composite Membrane Process", M.S. Thesis, D. Bhattacharyya, Director,
Department of Chemical Engineering, University of Kentucky, Lexington, Kentucky,
(1989)
Williams, E.B., "Contamination Containment by In Situ Polymerization", 2nd National
Symposium on Acmifier Restoration and Groundwater Monitoring. National Water Well
Association, Worthington, OH, (1982).
125
------- |