United States
Environmental Protection
Agency
Hazardous Waste Engineering
Research Laboratory
Cincinnati, OH 45268
Research and Development EPA/600/M-87/002 June 1987
ENVIRONMENTAL
RESEARCH BRIEF
Evaluation of Innovative Technology for the
Treatment of Hazardous Aqueous Waste Streams
S. Garry Howell
Introduction
Dilute aqueous wastes have been a treatment problem for
many years. Even when the waste materials were readily
treatable, the large volumes of water which had to be
passed through a treatment system slowed the process,
and required inordinately high capital investments. As a
result, many such dilute wastes were dumped into streams,
or onto the ground where they eventually entered
groundwater supplies. These factors, and an anticipation of
the ban on landfilling of aqueous wastes prompted EPA to
initiate research;programs to concentrate, treat, or recover
organic or heavy metal pollutants from aqueous streams.
This report summarizes the results of four of the the most
recent projects initiated by EPA in the aqueous waste
treatment area. We recognize that there is no one solution
applicable to the whole gamut of aqueous waste streams,
and that several approaches must be tried; consequently,
four projects were initiated:
Concentration of wastes by absorbing the water and
excluding the solute using reversible gels;
Concentration by reverse osmosis, using newly
developed composite membranes;
Adsorption on "pristine lignin"; i.e., lignin derived from
steam exploded wood, as opposed to lignin produced by
sulfate or sulfite pulping;
Extraction of pollutants with supercritical carbon dioxide.
Purpose and Objectives
The purpose of the study was to find economical and
effective ways to concentrate aqueous waste streams or to
remove pollutants from them, yielding a relatively pure
water, and a concentrated waste which could be disposed
of in an accepted manner.
The objectives were to determine which types of wastes
(inorganic, nonpolar or polar organic) could be treated by
one or more of the four processes chosen, obtain
throughput rates, compare the processes to existing
technologies, and finally, estimate capital and operating
costs if enough data were available.
Technologies Evaluated
Four technologies were evaluated, three aimed at
concentrating or extracting the wastes, the fourth was a
novel adsorbent.
Reversible Gel Absorption
The feasibility of using crosslinked polymer gels in a
reversible process for extracting pure water from aqueous
systems was investigated by Drs. Maier and Cussler of the
University of Minnesota (1). Two types of gels were studied,
pH sensitive and temperature sensitive, (so called because
the reversibility of water absorption could be caused by
raising pH in the former, or temperature in the latter).
Objectives
The general objectives of this investigation were to
determine if either of those two types of gels, as described
below, could be adapted to preferentially absorb water from
a solution containing hazardous wastes, and if so, whether
this technology could be applied to detoxify "real world"
hazardous waste streams.
Test Results
pH Sensitive Gels
The pH sensitive gels were made by copolymerizing
acrylamide monomer and N.N'-methylenebisacrylamide
-------
crosslinker in water, after which they were treated with 0.5
M sodium carbonate at 60°C for 24 hours to hydrolyze the
amide groups. By varying the ratio of acrylamide and
crosslinker. the swelling capacity of the gel may be
controlled; the swelling or water absorptive capacity is a
function of ionizable group concentration and constriction of
elasticity by crosslinks.
When water of pH 6 or above containing a dissolved solute
is in contact with a gel of this type, the gel will swell to 10
to 120 times its starting volume, depending on the percent
crosslinker, as indicated in Figure 1. The absorption of
water concentrates the solution, which may be drained off if
a batch process is used; in a packed column, a continous
flow of concentrated solution would be produced. After the
gel has reached its maximum capacity, the water is simply
eluted by acidifying a rinse stream, as indicated ih the
lower pH range of Figure 1. Concentration of a number of
aqueous solutions or emulsions is illustrated in Table 1.
Note that all the solutes in Table 1 are neutral or negatively
charged at the pH of separation, and that while larger
molecules are excluded, smaller ones are not excluded,
and pass into the interior of the gel particles.
Table 2 lists several ionized, negatively charged solutes
along with their molecular weight and potential charges.
Solutes are more efficiently concentrated if they are highly
dilute, multiple charged, and have no salt present.
Predictions based on Donnan equilibrium theory confirm
these observations fairly closely, as noted in the table.
Temperature Sensitive Gels
Earlier work had indicated that modified polyacrylamide
gels exhibited abrupt changes in volume in response to
small changes in temperature. Utilizing this phenomenon
would allow concentration of wastes with much lower
energy expenditure than conventional evaporation
processes.
Two types of these gels were evaluated; one was a
terpolymer of diethyl acrylamide and sodium methacrylate,
with N.N'-methylenebisacrylamide (MBA) as crosslinker;
the second, a copolymer of N-isopropyl acrylamide with
MBA crosslinker. The variation of gel volume with
temperature shown in Figure 2 indicates that the
polyisopropyl acrylamide will lose its imbibed water at a
lower temperature, but will hold about 25% as much as the
other copolymer.
The temperature sensitive gels are not sensitive to pH. As
illustrated in Table 3, the pH sensitive gel decreases to 0.04
of its starting volume as acid is added, while the
temperature sensitive gel is relatively unaffected. Low salt
concentrations do not cause collapse of the temperature
sensitive gels; however, high concentrations will collapse
the particles as shown in Table 4.
Figure 1. Gel volume as a function of solution pH. (Source: Reference 1)
140
120 -
100
Relative
Volume
80
60
20
2.0%
O 3.5%
5%
D 70%
A. 75%
A 20%
Percent
Crosslinker
4ğ
O
D D
DEO
A
* *
;"#Ğ',
D
A
All
6
pH
10
11
-------
Table 1.
Concentration of Aqueous Solutions Using Hydrolyzed Polyacrlamide Gels (Source: Reference 1)
Solute
Polystyrene Latex
Polystyrene Latex
Silica
Bovine Serum Albumin
Hemoglobin
Polyethylene Glycol
Sucrose
Urea
Mol. Wt.
Daltons
-
--
-
66,000
64,500
3,000 - 3,700
342
60
Solute Size,
mm
990a
, 34.6a
5a
7.2b
6.2b
3.8°
0.84b
0.53b
Feed Cone.,
Wt.%
0.21
0.91
1.82
0.08
0.73
0.56
1.00
3.00
Raffinate Cone.,
wt.%
0.35
1.40
3.03
0.18
1.26
1.09
1.09
3.00
Percent
Efficiency
85
82
80
93
91
6
6
0
a Measured by electron microscopy.
b Estimated from the diffusion coefficient in water using the Stokes-Einstein equation.
cReported by the manufacturer from light scattering measurements.
Table 2. Separation of Charged Solutes with Hydrolyzed Polyacrlamide Gels3 (Source: Reference 1).
Monovalent Test
Solution
10-4 M NaCI
10-4 M Methyl Orange
10-4 M Brornocresol Green
2xlO-4 M Sodium salt of
pentachlorophenolb
10-4 M Methyl Orange
0.1 M NaCI
10-4 M Methyl Orange
Additional
NaCI
_
-
-
3 x 10-3 M
0.1 M
-
0.9 M
n
(expt)
97%
97%
94%
65%
55%
55%
14%
1
(Donnan)
>99%
>99%
>99%
69%
49%
50%
18%
Polyvalent Test
Solution
10-4 M Congo Red
1 0-4 M Trypan Blue
10-4 M Trypan Blue
10-4 M Trypan Blue
10-4 M Congo Red
10-4 M Trypan Blue
1 0-4 M Trypan Blue
10-4 M Trypan Blue
Additional
NaCI
--
--
0.05 M
0.1 M
0.1 M
0.1 M
0.2 M
0.3 M
0.9 M
Jl
(expt)
97%
96%
83&
73%
71%
69%
52%
34%
n
(Donnan)
>99%
>99%
>99%
97%
79%
92%
89%
36%
a Molecular weights and potential charges of the less familiar test solutes are as follows: methyl orange (327.-1); bromocresol green (720.-1);
pentachlorophenol anion (265.-1); congo red (697,-2); trypan blue (961,-4).
bThis experiment used a partially hydrolyzed polyacrylamide polymer synthesized with 5% crosslinks. The added electrolyte here is NaOH. All
other experiments used the hydrolyzed P-6 gel.
Conclusions and Recommendations
A serious but. not insurmountable deficiency of both
temperature arid pH sensitive gels is their inability to
exclude smaller molecules such as urea or methanol while
imbibing water from a solution. The selectivity of removal
might be enhanced if a semipermeable membrane similar
to those used for reverse osmosis were grafted onto the
exterior of the particles. This membrane could be
composed of the same monomers as the interior, but with a
higher ratio of crosslinker in the skin. The tighter molecular
structure would allow only smaller molecules or ions to
enter into the body of the gel. Another possible approach
would be to add formaldehyde to the particles suspended in
water, which would bridge amide groups on the particle's
surface.
This is an area recommended for development. Such
development would greatly broaden the applicability of this
technology to concentrate dilute solutions of heavy metal
wastes, and to organic separations perhaps; e.g., the
separation of ethanol and water, with large savings in
energy expenditure.
Low Pressure Composite Reverse Osmosis
Membranes
Concentration of dilute hazardous organic pollutants has
been effectively accomplished by Dr. D. Bhattacharyya of
the University of Kentucky (2), using FT-30 reverse
osmosis (RO) membranes developed by Film Tech Inc.
Bhattacharyya's work confirms results reported by Lynch,
et al in a I984 EPA report, but carries the study somewhat
further. Reverse osmosis processes use pressure to force a
fluid through a membrane in the opposite direction of
normal osmotic flow (solutions normally tend to become
dilute, a state that is the lowest energy state). RO has been
used for several years to desalinate water, but was limited
in application by the narrow pH and temperature
requirements of the cellulose acetate membranes available.
The development of new membrane materials has greatly
broadened the applications of RO. Composites of two or
more polymers combine mechanical strength, the ability to
operate over a wider pH and temperature range, higher
water flux, and greater rejection for most solutes. Using RO
to concentrate dilute organic aqueous wastes greatly
-------
Figure 2. Gel Swelling vs. Temperature. Swelling is reported as a volume relative to that at high temperature. The
polyisopropylacrylamide shown on the left has a sharp change of swelling at 33°C. The copolymer of
diethylacrylamide and sodium methacrylate shown at the right shows a slower change, fastest around 55°C.
However, the volume change of the more highly swollen copolymer is larger. (Source: Reference 1)
12
10
Temperature, °C
30
50
70
10
Relative
Volume
Diethyl
Acrylamide/sodium
Methacyclate
Polyisopropyl-
Acrylamide
40
30
Relative
Volume
20
24
28
32 : 36
Temperature, °C
40
44
improves incineration or other treatments. Typical
applications of RO in this area would be in concentration of
leachate from landfills and contaminated groundwater, or
aqueous wastes from chemical processing.
Objectives
The objective of this investigation was to determine the
applicability of thin film composite membranes to
concentrate the following types of organic compounds:
Slightly soluble organics such as naphthalene,
anthracene, phenanthrene, and trichlorobenzene; these
organics were studied both as individual compounds and
as mixtures.
lonizable organics, including phenol, chlorophenols, and
nitrophenols. These were studied as mixtures in order to
determine if membrane performance was affected by
solute-solvent interactions.
Broader objectives were to find the effect of feed
concentration, pH, system pressure, and recovery on solute
rejection and water flux. The effects of these factors on
membrane stability were also studied.
-------
Table 3. Polyisopropylacrylamide Swelling vs. Added Acid
and Salta (Source: Referencel)
Added Solution
0 mM HCI/g gel
0.04
0.11
0.14 ;
0.19
0.23
0.32
excess 0.05 M NaCI
0.10
1.00
excess 0,05 M CaCI?
Relativeb
Volume
1.00
1.00
0.96
0.93
0.89
0.93
0.92
1.00
0.96
0.22
0.98
Volume of pHb.c
Sensitive Gel
1.00
1.00
0.15
0.11
0.07
0.05
0.04
1.00
0.93
0.80
a All experiments are at 25 °C. The isopropylacrylamide gel was made
with 1 % crosslinking.
b The relative volume is that in solution divided by that in water.
c The pH sensitive gel was a partially hydrolyzed polyacrylamide with
5% crosslinking. Similar gels have been reported elsewhere.
Results and Discussion
The apparatus used in the study is shown in Figure 3. The
FT-30 membranes are made up of three layers: a 0.05-
0.25 Jim layer of aromatic polyamide, 50 u,m of porous
polysulfone. and a 125 um polyester backing. Figure 4
illustrates the assembly into a batch test cell, using nitrogen
pressure to drive the liquid through the membrane.
Separation of selected classes of priority pollutants were
studied at 0.52-2.068 MPa (75-300 psig). The batch cell
experiments focused on the individual species and mixtures
of several slightly soluble PAH compounds (naphthalene,
anthracene, phenanthrene) chlorophenols, nitrophenols,
phthalates, and chlorobenzene. A wide range of pH values
were run on the chlorophenols and nitrophenols to establish
the rejection behavior of nonionized and ionized species.
Table 5 summarizes the results obtained with PAH
compounds. Note that permeate fluxes were nearly all in
the same range, and did not vary greatly from the permeate
flux of distilled water.
The rejection of chlorophenols is greatly affected by pH, as
shown in Figure 4. This effect is thought to be due to the
exertion of a repulsive electrostatic force between the
membrane and the chlorophenols. Figure 4 also shows the
effect of additional chlorine atoms on the molecules, as the
rejection increases with increasing chlorine content.
Mass balance calculations were made to determine the
adsorption of chlorophenol solutes on the membrane. The
calculated concentrations at low pH (5.5) differ significantly
from actual concentrations determined by TOG
determinations, indicating adsorption on the membrane; if
pH is raised to 10.8, adsorption is cut in half. This
adsorption was further confirmed by running unstirred vs.
stirred samples. Permeate flux dropped over 63% for
2,4,6-TCP as a result of a phenomenon called
concentration polarization.
Three chlorobenzene homologs were the last compounds
studied. These were monochlorobenzene, 1,4-
dichlorobenzene, and 1,2,4-trichlorobenzene, all run under
unstirred conditions. As might be expected, adsorption on
the membrane was even worse than with the nonionized
(low pH) chlorophenols when unstirred, with the dichloro
and trichloro compounds adsorbing more severely than the
monochlorobenzene. Rgure 5 compares unstirred and
stirred monochlorobenzene. When stirred, the permeate
flux remained high, decreasing only by about 2% during the
run, compared to the decrease in the unstirred flux.
Conclusions
Low pressure composite membranes offer an efficient
method of concentrating certain dilute organic wastes. High
solute separations are achieved at relatively low pressures
of 1-2 MPa (145-290 psi) over the broad pH range of 2-
12. The aromatic polyamide membrane tested was Film
Tech's FT-30 which showed excellent stability over long
operating periods with respect to water flux and permeate
quality. Rejection of ionizable compounds such as
chlorophenols was 99 + % at pH 11, but only 77-89% at
pH 4.6; nitrophenols were slightly less affected by low pH.
While some compounds, notably chlorobenzenes, adsorbed
on the membrane, permeate rejection actually increased at
the expense of a drop in permeate flux.
Extraction with Critical or Near Critical Fluids
While many extractions of organic materials are
accomplished with common solvents such as hexane or
ethanol, a growing body of knowledge has developed on
the use of near critical or supercritical fluids (SCF) as
extractants. Perhaps the best known application of SCF is
the use of SC carbon dioxide to decaffeinate coffee,
replacing the trichloroethylene used previously.
When a liquid is heated in a closed pressure vessel with a
vapor space above it, it will boil, but there will be a clearly
defined interface between the two phases. If heating is
continued, the pressure will rise, and eventually a
temperature will be reached where the interface disappears;
this is the critical point, above which a single phase having
unique properties exists. Table 6 lists the critical properties
for a number of compounds, many of which have been
evaluated as solvents.
Using SCF as extraction solvents permits tailoring the
solvent properties by varying the pressure and/or
temperature above the critical point; this simple procedure
plus the wide choice of fluids available, gives considerable
control over an extraction process. Many variations of
solvent characteristics are available. As indicated in Table
6, a nonpolar solvent such as ethylene with low critical
pressures and temperatures may be selected for a specific
task: an extremely polar solvent such as water may be even
more useful. An extreme example of the ability to vary
solvency is evidenced in the case of water, which has a
dielectric constant of 80 at room temperature. As the
temperature is raised toward its critical point, the dielectric
constant decreases to about 2, which is very near the value
of nonpolar compounds such as hexane. Above the critical
point, salts which were soluble at lower temperatures
precipitate. Conversely, carbon dioxide (dielectric constant
1.0 at 100 °C and 0.1 MPa) rises to about 1.6 at 5.07 MPa.
Although this is below the critical pressure of 7.5 MPa,
carbon dioxide has been used as an extraction solvent in
this pressure region.
The projects summarized here are an extension of work
done on the regeneration of activated carbon by extracting
5
-------
Tablo 4. Selectivity of Extractions Using Temperature Sensitive Gels (Source: Reference 1)
i
Polyisoproply- Copolymer of Diethylacrylamide
Solute Mol. WL acrylamide and Sodium Methacrylate
Urea
Sodium Pentachtorophenolate
Vitamin B12
Ovalbumin
Potyetheytene Oxide
Galatin
Blue Doxtran
Polystyrene Latex
Polyethylene Gtycol
267
1,355
45,000
600,000
2,000,000
400
3,400
8,000
18,500
19
2
18
32
97
96
98
9?'
950
10
30
56
80
43
3
51
15
84
89
97
99
96b
5
19
25
61
Copolymer of Diethylacrylamide
and Sodium Methacrylate
2a
2
-
7
92
96
96
96°
11
16
48 ,
a Percent of crosslinkage used in preparation.
b This latex has a diameter of 0.06 um.
0 This latex has a diameter of 1.2 jim.
Figure 3. Batch membrane unit. (Source: Reference 2)
Feed Solution Inlet
Permeate
0
Cell Volume:1900 cm3
Adjustable Shaft
, Magnetic Stirrer
(stirring speed: 600 RPM)
N2
Tank
Membrane Area =9.6 cm2
-------
Figure 4. Effect of pH on rejections of chlorophenols at 2.068 mPa. (Source: Reference 2)
100
90
80
Solute
Rejection
(%)
70
60
50
AP = 2.068 MPa
Feed: 22 mgll of each solute
A Phenol
Q 2-Chlorophenol
O 2,4-Dichlophenol
2,4,6-Trichlor.ophenol
5.5
6.5
7.5
pH of Permeate Water
8.5
9.5
Table 5. Summary of Batch Experimental Results with PAH Compounds.3 (Source: Reference 2)
Feed Feed pH Permeate Flux, cm3/cm2 s Recovery, percent
Rejection, percent
Naphthalene (21 .4)
Naphthalene (2.22)
Antrhacene (0.12)
Phenanthrene (0.574)
Mixture of naphthalene (7.08)
anthracene (0.000999)
phenathrene (0.547)
Distilled water (averaged over 25 days)
6.5
5.5
6.3
6.2
5.5
7
9.45 x 10-4
8.63 x 10'4
9.08 x 10"4
8.92 x 10-"
10.25 x 10'4
11.00X10'4
79.3
82.9
84.2
83.5
82.8
98.01
97.99
> 99.10
> 99.96
88.37
94.14
99.24
a Ap = 1.72 MPa, stirred conditions.
with supercritical carbon dioxide reported by Di Filippi et al.
in 1980 (3). This summary covers basic work done at
Louisana State University (4) and the University of Illinois
(5), and applied studies on the extraction of steel mill
sludges, pesticide wastes, and toxic water soluble organics
done by Critical Fluid Systems, Inc. (6).
Objectives
The objective of this investigation was to determine whether
critical or near critical carbon dioxide could be used to
extract and/or recover toxic organic compounds from
aqueous waste streams. A further objective, based on these
results, was to estimate the capital and operating costs of
an extraction system, and furnish a conceptual design.
Results and Discussion
Carbon dioxide is a convenient solvent; it is cheap,
nontoxic, and has a relatively low critical point. On the other
hand, when a more polar compound must be extracted,
carbon dioxide may not be very effective. Figure 6 shows
the results of extraction of a soil which had a starting
concentration of 1000 ug DDT/g soil. Extraction was done at
40°C and 10 MPa. Carbon dioxide alone extracts only
about half of the DDT, and was only slightly improved by
adding toluene, a nonpolar compound, as a cosolvent.
When 5 wt% methanol was added, the residual was
decreased to about 5% of its starting concentration. While
this experiment was run on undried soil, to which the DDT
might be quite strongly bound, cosolvents might be a better
choice for extracting certain aqueous wastes.
-------
Figure 5. Instantaneous permeate flux as a function of time for stirred and unstirred experimental conditions. (Source:
Reference 4)
Instantaneous
Permeate
Flux (x 10-4)
14
12
10
n .
Stirred Conditions
Unstirred Conditions
AP = 2.068 mPa
Feed = Chlorobenzene Mixture
100
200
\ Time (Minutes)
300
400
Table 6. Critical Properties for Selected Fluids (Source:
Reference 3) i
Critical Critical Critical
Temperature (To), Pressure Density (Pc),
Fluid °C (Pc), mPa g/cm3
Penlane
Butane
Sotvont-12
Propane
Ethane
Ethyleno
Carbon Dioxide
Sulfur Dioxide
Ammonia
Water
296.7
152.0
112.0
96.9
32.3
9.9
31.1
157.6
132.4
374.3
3.4
3.8
4.1
4.3
4.9
5.2
7.5
8.0
11
22
0.232
0.228
0.558
0.220
0.203
Q.22J
0.468
0.525
0.235
0.326
Dry mill scale had previously been deoiled with SC carbon
dioxide (4). Since a large portion of mill scale is disposed of
in lagoons, an attempt was made to extract sludge from the
lagoons, recovering iron and a high heating value oil. j The
sludge is 30 to 70% solids, and contains oil, water,[iron
fines and toxic and nontoxic residues. Oil and grease
content varies between 6 and 30% by weight of the solids.
One plant can produce up to 50,000 tons per year of this
sludge, and the pollution potential and economic losses are
substantial; an effective extraction method would allow
recycling the iron, and reuse of the oil as either fuel or
lubricant.
Sludge extraction was attempted with carbon dioxide in a
sieve tray extraction column, the only equipment available.
The dense sludge was very difficult to handle, plugging
lines, valves, and eventually the sieve trays themselves.
Solvent to feed ratios and dilution were increase^ in
attempts to keep continous flow and increase oil recovery,
but an oil reduction of 30% was the highest attained. To
handle the heavy abrasive sludge, a complete system
redesign was needed. One design considered was a multi-
stage mixer-settler operating at or above a 4-1 solvent to
sludge ratio. The economics of such a system have not
been studied, they may not justify recovery.
Sludge from pesticide manufacturing was tested in a stirred
laboratory reactor and later in the sieve tray pilot plant. The
sludges contained water, xylenes, carbon tetrachloride,
solids, fines, salts, and other insolubles. The purpose of this
test was to extract carbon tetrachloride and the mixed
xylenes from this waste.
Distribution coefficients determined in the stirred reactor
indicated an easy separation, but when tested in the
continous flow reactor with a much lower shear rate, four
passes were required to attain an overall reduction of 86%
of the carbon tetrachloride. Results with xylene extraction
were even more anomalous; the experimenters postulate
that soaps and dispersants present in the feed hindered
extraction of the xylene from the aqueous phase.
Extraction of Dissolved Polar Organic Compounds
Carbon dioxide extraction of aqueous solutions of
acrylonitrile and acetonitrile were much more successful
than any other tests. Much of this success was attributed to
the fact that these solutions were clear with a single phase,
thus the sieve tray column could function properly; i.e., it
could provide better contact between the extractant and the
aqueous feed. Using a solvent to feed ratio of 1.5-1
required multiple passes; i.e., the raffinate from the first
pass was used as feed for the next. A total of five passes
were used to get acetonitrile below the minimum detection
limits. Acrylonitrile, as predicted from distribution coefficient
data, was much easier to extract; only one pass was
required to remove it below the detection limit. The success
of these runs resulted in their use as a basis for a
conceptual plant design.
A hypothetical wastewater extraction system was designed
to process a stream with properties and compositions as
shown in Table 7. The economical number of trays for the
extraction was determined to be 30, using a tray efficiency
-------
Figure 6. Extraction results. (Source: Reference 4)
60 - 1000 itg DDTlg soil
0.6 r-
O Pure SC-C02
SC-CO2 with 5 wt% toluene
SC-CO2 with 5 wt% methanol
0.4 -
Relative
DDT
Cone.
(6/60)
0.2 h-
20
Time (Minutes)
40
60
of 0.44 and a solvent to feed ratio of 2-1. This system is
designed to reduce a waste stream having 1680 ppm
acetonitrile to less than 5 ppm, a reduction of 99.7%. A
process flow diagram is shown in Figure 7, where the waste
stream is contacted countercurrently with an upwardly
flowing stream of carbon dioxide at 7.6 MPa (1100 psi). The
extract stream has the recovered organics separated at
high pressure in a distillation column, thus minimizing the
work required to recompress the carbon dioxide, whose
heat of vaporization is furnished by the adiabatic heat of
compression of the recovered solvent.
Estimated annualized operating costs of a plant processing
20 gpm of an acrylonitrile/acetonitrile feed operating 8000
hours per year are given in Table 8. Processing cost would
be about $0.084/gal. with no credit for recovered product. If
the recovered products were purified, they would be worth
about $100,000 total, and would lower the overall
processing cost appreciably.
Conclusions
Supercritical or near critical extraction of organics from
aqueous wastes is applicable to certain waste streams,
particularly those having very toxic or valuable solutes.
Aqueous wastes with suspended solids and sludges must
be handled on a case-by-case basis, with very careful
consideration of the equipment requirements, and an
awareness of construction costs. As the art and science of
this process now stands, laboratory determinations of
partition coefficients and other parameters must be done
before any larger scale extractions are attempted.
Use of Pristine LJgnin to Treat Hazardous Waste
Conventional wood pulping processes are prodigous
chemical by-product producers. In addition to the lignin
extracted, which is not in its "pristine" form (in the Kraft
process lignin is converted to thiolignin, and in sulfite
processing, much is converted to lignin sulfonates)
hemicellulose sugars are produced; these are oxidized to
saccharinic acids in the Kraft process. Much of the lignin,
saccharinic acids, and hexoses are presently burned as fuel
for other parts of the papermaking operation. Some sources
have stated that the fermentable hexoses alone could
produce two to three times the amount of ethanol presently
produced in the U.S. (7).
Newer pulping processes beginning to be adopted use
solvents such as ethanol to extract lignin from the wood
chips, or explode chips by saturating them with steam in an
autoclave, then suddenly releasing the pressure. The lignin
thus produced is much nearer to its structure as it existed
in nature, and has been dubbed "pristine" lignin. A study of
the use of this material to adsorb both organic and
inorganic compounds from aqueous media, and to compare
its performance to activated carbon, has been performed by
Dr. D. J. O'Neil at Georgia Institute of Technology's Georgia
Tech Research Institute (8). Since lignin is much cheaper
than activated carbon ($0.03-0.13/kg vs. $1.00 + ) it could
compete if it were only 1/10 as adsorptive, and instead of
regenerating the adsorbent, it could, in many instances be
burned, as it has some fuel value. If the same assumption
of efficacy is applied to ion exchange resins, an even
greater cost advantage would be seen if pristine lignin were
used to remove heavy metals. A typical structure of lignin is
shown in Figure 8. The rich organic functionality is
apparent, with many carboxylic acid, phenolic, quinone
groups and others available.
Objectives
The principal objective of this study was to demonstrate
that "pristine" lignin could provide a low cost, yet effective
-------
Flguro 7. Flow diagram of proposed extraction process.
Distillation
Column
Flash Drum
Extract
Waste
and flexible material for the treatment of various aqueous
waste streams. To accomplish this, lignin was evaluated as
a physicochemical adsorbent of heavy metals and organic
compounds. Other objectives included an evaluation of
lignin's technical and cost effectiveness in comparison to
current treatment technologies such as activated carbon
adsorption, as well as a conceptual design of a full-scale
treatment plant.
Results and Discussion
Adsorption studies reported here used lignin derived from
tulip poplar, a relatively common, fast growing tree of the
Southeastern U.S. Two types of lignin were extracted, the
first, a light brown powder, was obtained from steam
exploded wood chips using 95% ethanol. Addition of
benzene in an attempt to change the character of the
extract produced a gummy precipitate upon drying. The
fraction of lignin recovered averaged 7.5%, had an average
surface area of 30 square meters per gram, and an average
molecular weight of 1380. The second type of extract was
made with dilute sodium hydroxide, and although an
average of 31.9% of the lignin was extracted, the surface
area was only about 0.1-0.3 square meters per gram, but
the molecular weight was 2300. with a much broader
distribution. The complex structure and presence of many
different polar groups presents great difficulties in
elucidating the structure of lignin for, as was the case here,
the extraction method defines the product.
The adsorptive behavior of ethanol extracted lignin was
studied using water spiked with several organic
"contaminants" and results are presented in Table 9.
Although smooth curves of adsorption were obtained, there
does not appear to be any pattern of structure dependency
of the adsorption isotherms. Rough estimates of the
capacity of the large particle size (30-40 mesh) lignin used
indicate that a smaller particle size adsorbent should
remove about 5 g/kg of the seven compound mixture in
Table 9, and about 0.3 g/kg of 2,4-D. Adsorption of lead
and chromium compounds (the investigator does not
specify the species) is shown in Figure 9. An adsorption
capacity of 1.3 g/kg of these metals is estimated under the
conditions of this test. Speed of adsorption is much higher
when higher ratios of adsorbent to both organic and
inorganic compounds was used, suggesting that an
10
-------
Table?. Properties and Composition of Acetonitrile/
Acrylonitrile Waste Streams (Source: Reference
6)
Feed Rate, gpm 20
Acetonitrile, ppm 1680
Acrylonitrile, ppm 1200
Dissolved Sulfates, weight % 2.5
Water, weight % 97.5
Temperature, °F 80
Specific Gravity 1.0
Characteristics of Recovered Organics
Organics: Acetonitrile, Acrylonitrile, weight % 98.5
Carbon Dioxide, weight % 0.5
Water, weight % 1.0
Temperature, °F 100
Characteristics of Wastewater Effluent
Water, weight % 97.4
Carbon Dioxide, weight % < 0.06
Acrylonitrile, ppm < 2
Acetonitrile, ppm <2
Sulfates, weight % 2.5
Temperature, °F 100
Tables. Annualized Operating Costs of 20-GPM
Acetonitrile - Acrylonitrile Extraction Plant.
(Source: Reference 6)
Utilities
Electricity (175 kWhr/hr @ $0.04/kW hr)
Well Water (18 gpm @ $1.50/1,000 ga!)
Steam (230 Ib/hr @ $10.00/1000 Ib)
Solvent Make-Up
Carbon Dioxide (15 Ib/hr @ $100.00/2000 Ib)
Labor
1 Operator @ $13.00/hr
Annual Processing Cost
Annual
Cost
$56,000
12,960
18,400
6,000
104,000
$802,360*
* Depreciation, overhead, and miscellaneous operating supplies of
$605,000 included.
increase in surface area would be of great benefit. A
comparison of the adsorptive capacity of ethanol vs. alkali
extracted lignin indicates that the latter might be as good or
better in capacity if the difference in particle size is taken
into account. The alkali extracted lignin had much larger
particles, partially accounting for its lower surface area per
gram. Another factor favoring alkali extracted lignin is the
higher yields; more adsorbent may be obtained per pound
of wood. Tables 10 and 11 indicate that after an adjustment
for surface area is made, alkali lignin would be about as
good on organics, and better as a metal adsorbent than
ethanol extracted lignin.
A comparison of the adsorption isotherms of alkali extracted
lignin and activated carbon, based on Freundlich
parameters indicates that approximately 30 times as much
lignin as activated carbon is required to achieve a
comparable reduction in phenol in a water solution. A
comparable figure was obtained for naphthalene. A much
more favorable comparison of lignin vs. activated carbon is
obtained when:Freundlich parameters for metal adsorption
are calculated (Figure 10). The ratio of lignin to carbon is
only about 4-1; taking cost into account, lignin might be
the preferred adsorbent in this case.
Although data gathered in this report are very preliminary, a
conceptual design of a one million gallon per day water
treatment plant is shown in Figure 9. Efficient operation of
an adsorption plant of this type will depend on developing
methods of controlling particle size and surface area of the
lignin adsorbents.
Conclusions and Recommendations
Pristine lignin, particularly the alkali extracted variety, shows
some promise as a low cost adsorbent for aqueous
hazardous wastes. This preliminary study indicates that
further work must be done to control particle size and
expand surface area (expressed in square meters per gram)
before a firm conclusion can be drawn. Evaluation of this
high surface-controlled particle size material might yield
enough data to allow the design of a pilot size test unit to
indicate the feasibility and cost of a full-size water
purification plant based on this technology.
References
1. Maier, W. J., and Cussler, E. L. Dewatering of Dilute
Aqueous Hazardous Wastes Using Reversible Gel
Absorption. EPA Contract 68-03-1957.
2. Bhattacharyya, D. et al. Concentration and Purification of
Dilute Hazardous Wastes Using Low Pressure
Composite Membranes. Cooperative Agreement CR
911976.
3. DeFilippi, R. P. et al. Supercritical Fluid Regeneration of
Activated Carbon for Adsorption of Pesticides, EPA
600/2-80-054.
4. Knopf, C. F. (Louisiana State University) Private
communication.
5. Eckert, C. Ai (University of Illinois) Private
communication.
6. Rice, P. N. et al. Supercritical Extraction of Aqueous
Hazardous Waste, EPA Contract 68-03-1956.
7. Mayerly, R. C. ef al. The Forest Refinery. Chemtech.
March 1981, page 186.
8. O'Neil, D. J. et al. Low Cost, High Efficiency Pristine
Lignin for Hazardous Waste Treatment. Cooperative
Agreement CR 812223-01-0.
11
-------
Figure 8. Typical tlgnin structure. (Copyright 1967 by John Wiley & Sons, Inc. Reprinted by permission of John Wiley & Sons,
Inc. From Klrk-Othmer Encylcopedia of Chemical Technology, Anthony Standen, Executive Editor.)
.0
H(CaH,A).0.
CH.O
OCHa
12
-------
Table 9.
Concentrations of
Reference 8)
Model Compounds in the Supernatant After Adsorption on the Pristine Lignin (ppm). (Source:
Time
Ohr
1 hr
2hr
4 hr
8 hr
1 day
2 days
4 days
4 daysc
Aniline
5.000
4.09
3.20
3.29
3.34
3.75
4.09
4.14
3.97
3.52
3.25
1.56
2.62
2.88
1.47
0.71
: 1.28
Phenol
5.000
2.62
2.56
2.73
2.50
2.56
2.66
2.48
2.22
1.94
2.38
2.26
2.16
1.77
2.16
2.05
2.06
Isophorone
1.560
1.66
1.37
1.29
1.41
1.39
1.49
1.38
1.61
1.31
1.23
1.22
1.06
1.16
1.02
0.68
0.81
Naphthalene
1.300
0.97
0.74
0.61
0.65
0.63
0.65
0.57
0.64
0.56
0.51
0.53
0.41
0.48
0.45
0.21
0.24
Pentachlo-
rophenol
1.540
0.70
0.70
0.54
0.52
0.49
0.49
BDt_a
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
Bis-(EtHex)
phthalate
1.560
0.33
0.36
0.29
0.23
0.22
0.23
0.14
0.14
0.18
BDL
BDL
BDL
BDL
BDL
BDL
2,4-D
1.250
.08
.27
.21
.28
.05
.28
0.74
1.03
0.67
-b
0.63
0.57
0.65
0.60
0.75
Trichloro
ethylene
2.500
1.97
1.55
1.54
1.52
1.93
1.52
0.36
1.78
1.63
1.73
1.97
1.94
1.45
0.54
1.26
- Below detection limit.
bSamples were lost.
cLignin dose was 0.75 g for organic mixture, 0.5 g for 2,4-D, and metal kinetics study.
Table 10. Test Results - Second Size Study (Organics). (Source: Reference 8)
Concentrations of Model Compounds in the Supernatant After Adsorption on the Pristine Lignin (ppm)
Equilibrium Time: 4 days
Liqnin
C0 Nominal
C0 Actual
ETOH Extract
30x40
40x60
60x100
NaOH Extract
60x100
Aniline
4.758
4-447
4.130
3.130
3.222
2.917
2.164
1.577
1.620
1.230
1.420
Phenol
5.337
4.346
4.298
4.346
4.036
4.209
4.357
3.340
3.832
3.987
4.099
Isophorone
2.043
1.985
2.002
1.467
1.500
0.969
1.003
1.434
1.340
1.695
1.587
Naphthalene
2.259
2.113
1.856
0.660
0.453
0.665
0.654
6.428
0.597
1.053
0.857
Pentachlo-
rophenol
2.610
1.833
2.123
0.360
__
0.488
0.665
Bis-(EtHex)
phthalate
4.031
2.790
3.500
1.263
1.349
1.514
1.143
0.774
0.854
2.392
2.184
Trichloro-
ethylene
3.336
2.583
2.604
2.058
1.847
1.333
1.398
1.577
1.520
1.348
1.477
Note: Sample volumes of 100 ml were used (except for trichloro-ethylene: 120 ml).
13
-------
Figure 9. Conceptual design of a 1 MGD treatment plant. (Source: Reference 8)
Float Oil
to Incinerator
Sediment and Waste Adsorbent
to Incinerator
Course
Basket
Strainer
Back
Wash Water
10 gpm/ft2
ireate
Effluent
Storage
Surge
Sewer System
(POTW) or Receiving
Water
14
-------
Table 11. Test Results - Initial and Second Size Studies.
(Source: Reference 8)
Adsorption of Lead and Chromium on Lignin, pom (4 days)
Lead
Lignin
C0 Nominal
C0 Actual
Ethanol Extracted
30x40
40x60
60x100
NaOH Extracted
60x100
Lead
10
8
8
7.83
7.3
7.37
7.2
6.4
7.5
5.5
7
(duplicate)
10
11.3
11.02
7.5
8.00
6.5
8.50
6.50
6.45
2.50
2.00
Chromii
10
9.89
9.9
4.19
5.34
4.55
4.56
4.93
4.65
0.86
0.92
Figure 10. Kinetic study of the adsorption of heavy metals on pristine lignin (Cr and Pb). (Source: Reference 8)
Cone, (ppm)
10 Jl Pristine Lignin (30/40 Mesh)
pH = 4.0
Temp. = 19"C
- Lead/0.25 g Lignin
D - Lead/0,50 g Lignin
- Chromium/0.25 g Ugnin
O - ChmmiumlO.50 g Lignin
Time (Days)
15
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
EPA/600/M-87/002
------- |