Environmental Protection Technology Series
EXTRACTION OR DESTRUCTION OF CHEMICAL
POLLUTANTS FROM AQUEOUS WASTE STREAMS
Robert S. Kerr Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Ada, Oklahoma 74820
-------
EPA-600/2-77-148
July 1977
EXTRACTION OR DESTRUCTION OF CHEMICAL POLLUTANTS
FROM AQUEOUS WASTE STREAMS
by
R. R. Davison
Texas A&M University
College Station, Texas 77843
Grant No. R800947
Project Officer
Jack H. Hale
Source Management Branch
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
DISCLAIMER
This report has been reviewed by the Robert S. Kerr Environmental
Research Laboratory, U.S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
-------
FOREWORD
The Environmental Protection Agency was established to coordinate
administration of the major Federal programs designed to protect the quality
of our environment.
An important part of the agency's effort involves the search for in-
formation about environmental problems, management techniques and new techno-
logies through which optimum use of the nation's land and water resources can
be assured and the threat pollution poses to the welfare of the American people
can be minimized.
EPA's Office of Research and Development conducts this search through
a nationwide network of research facilities.
As one of these facilities, the Robert S. Kerr Environmental Research
Laboratory is responsible for the management of programs to: (a) investigate
the nature, transport, fate and management of pollutants in groundwater; (b)
develop and demonstrate methods for treating wastewaters with soil and other
natural systems; (c) develop and demonstrate pollution control technologies
for irrigation return flows; (d) develop and demonstrate pollution control
technologies for animal production wastes; (e) develop and demonstrate tech-
nologies to prevent, control or abate pollution from the petroleum refining
and petrochemical industries, and (f) develop and demonstrate technologies
to manage pollution resulting from combinations of industrial wastewaters or
industrial /municipal wastewaters.
This report contributes to the knowledge essential if the EPA is to meet
the requirements of environmental laws that it establish and enforce pollution
control standards which are reasonable, cost effective and provide adequate
protection for the American public.
W. C. Galegar
Director
Robert S. Kerr Environmental
Research Laboratory
ill
-------
ABSTRACT
The use of solvent extraction and ozonation to treat various industrial waste
waters was studied. Most were light chlorinated hydrocarbon, solvent wastes
and were principally extracted with a high molecular weight paraffin petrole-
um fraction. Distribution data on related pure chlorinated compounds were
also obtained. The economics of solvent extraction versus steam stripping
was examined. Though chlorinated solvents can be effectively removed by
extraction, stripping appears to be more economical.
A toluene diamine waste water was found treatable with benzene.
In all these wastes there are unextractable fractions.
Attempts were made to treat glycol, toluene diamine and light chlorinated
hydrocarbon waste waters with ozone, but results were not satisfactory.
iv
-------
CONTENTS
Foreword lii
Abstract -jv
Figures v1
Tables vi1
Acknowledgments ji
Section
l Introduction 1
2 Conclusions 3
3 Recommendations 4
4 Distribution Coefficients of Pure Components .... 5
5 Extraction Data on Waste Streams 9
6 Design of Solvent Extraction Process to Remove Chlorinated
Solvents from waste water 17
7 Design of Process for Removal of Chlorinated Solvents from
Waste Streams by Steam Stripping 23
8 Pilot Plant Design 32
9 Thermodynamic Analysis of Extraction VS Stripping
Volatile Dilute Components 38
10 Ozonation of Waste Water Streams 45
11 Chemical Analysis 57
12 References 59
-------
FIGURES
Number
1 Comparative Extraction Coefficients for Various Industrial
Waste streams 10
2 Extraction of Toluene Diamine 11
3 Extraction of Aromatic Waste 12
4 Extraction System 18
5 Steam Stripping System 24
6 Solubility of EDC in Water 25
7 Vapor Pressures of Various Substances 26
8 Vapor-Liquid Equilibrium Constants for EDC and Water
(Water Rich Phase) at 1 atm 28
9 Pilot Plant Design 34
10 Continuous Ozonation System 46
11 Ozonation of Propylene Glycol Solutions 47
12 Theoretical VS Actual TOC values for EDC 58
VI
-------
TABLES
Number
1 Pure Component Distribution Data 6
2 Chlorinated Solvent Data 8
3 Extraction of Industrial Samples 14
4 Purchase Cost of Major Items of Equipment for Extraction System 22
5 Purchase Cost of Major Items of Equipment for Stripping System 31
6 Extraction and Stripping Factors 44
7 Ozonation of Propylene Glycol (dilute) 48
8 Ozonation of Propylene Glycol (concentrated) 50
9 Continuous Ozonation of Propylene Glycol Streams 53
10 Ozonation of Toluene Diamine 55
11 Ozonation of Chlorinated Waste (sample 011A) 56
Vll
-------
LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
atm
Btu
EDC
COD
FRC
GPM
H.E.T.P.
I
mil eq
mgm
PG
ppm
TDA
TOC
TRC
SYMBOLS
A
D
E
H
AHV
K
L
N
p
p*
Q
S
t
At
U
V
VV2
X
Y
atmosphere
British thermal unit
ethylene dichloride
chemical oxygen demand
flow recording controller
gallons per minute
height equivalent to a theoretical plate
liter
mili equivalent
mi ligram
propylene glycol
parts per million
toluene diamine
total organic carbon
temperature recording controller
diameter, ft
extraction factor, K Lo/Lw
Henry's law constant, atm
latent heat of vaporization
equilibrium constant, Y /X or Xo/Xw
liquid flow rate, moles per unit time
number of theoretical stages or transfer units
partial pressure, atm
vapor pressure, atm
heat flow, Btu/hr
stripping factor, KV/L
temperature °F
temperature difference,°F
heat transfer coefficient, BTu/hr-ft -°F
vapor flow rate, moles per unit time
molar volumes
liquid mole fraction
vapor mole fraction
vm
-------
greek
Y -- activity coefficient
6 — Hildebrand's solubility parameter
TT — total pressure, atm
subscripts
E — in extractor
g -- gas phase
L -- liquid phase
o -- oil stripper or oil phase
Sat — at saturation
v -- volume basis
w — water stripper or water phase
-------
ACKNOWLEDGEMENTS
The support of the Industrial Advisory Council of the Lower Mississippi
Project (EPA Grant 8800773) is acknowledged with thanks to the following
companies and their representatives:
Allied Chemical Company - Baton Rouge, Louisiana
BASF-Wyandotte Corporation - Geismar, Louisiana
Dow Chemical Company - Plaquemine, Louisiana
Foster Grant Company - Baton Rouge, Louisiana
Hooker Chemical Corporation - Hahnville, Louisiana
Monochem, Inc., - Geismar, Louisiana
PPG Industries - Lake Charles, Louisiana
Vulcan Materials - Geismar, Louisiana
Support and help in the construction of the bench Ozonator, extraction
and ozone bench scale studies, analytical work, and report preparation by
C. G. Hewes and W. H. Smith are gratefully acknowledged.
-------
SECTION 1
INTRODUCTION
It is not the intention of this report to present a study of solvent extrac-
tion as a unit operation. This has been done in many books and articles
(1-4).
Similarly, for background information about the ozonation work, the reader is
referred to two recent papers by the Principal Investigator (5, 6).
The work described in this report is an attempt to apply the above two
methods, often lumped with others under the heading physical-chemical
treatment, to improve the quality of certain specific waste waters. This
work was done in cooperation with the state of Louisiana, with Gulf South
Research Institute (GSRI) of Baton Rouge and with several chemical companies
in Louisiana and with the Environmental Protection Agency's Robert S. Kerr
Environmental Research Laboratory of Ada, Oklahoma. Physical-chemical
methods are generally used when biological treatment is for some reason
inadequate, or to assist biological treatment by the prior removal of
refractory and toxic compounds, or the post removal of traces of undesirable
constituents.
Solvent extraction is best compared with other physical methods which actu-
ally remove certain substances from solution. The methods to which it is
most analogous and with which it must often compete are stripping and
adsorption. Each of these methods relies on a selective transfer between
phases - stripping by a favorable relative volability, extraction by a
favorable relative solubility, and adsorption, of course, by a favorable
adsorption equilibrium. Each of these methods offers the possibility of
recovery of the solute which, in some cases, can significantly effect the
economics of treatment.
Very little use has been made of solvent extraction in waste water treatment.
The only common process is the recovery of phenolic substances. For years
phenol has been recovered from coke oven liquor with an aromatic oil. More
recently the Phenex (Exxon) process has been developed which uses a light
catalytic oil to extract phenol from refinery wastes. Other solvents have
been reported which give better distribution ratios (7, 8). In recent
reviews or symposiums on water reuse almost no mention is made of solvent
extraction.
-------
Solvent extraction processes can be categorized as follows:
1. Extraction of a solute having reasonable volatility into a
nonvolatile-water immiscible solvent.
2. Extraction of a nonvolatile solute into a volatile immiscible
solvent.
3. Cases 1 and 2 with a partially miscible solvent.
4. Extraction of a chemically reactive solute.
5. Extraction of water plus a valuable solute to produce a water more
concentrated in other components, followed by recovery of the solute
from water.
6. Extraction of water from a waste to produce a concentrated waste
which can be burned or otherwise disposed of. Pure water can be
recovered also.
In this work we will be concerned with the first three categories. Partial
miscibility of solvent and water means that there must also be a solvent
recovery step from the water, usually by steam stripping. The category of
nonvolatile solvent and volatile solute is most economical as a rule, because
of lower solvent recovery heat consumption.
Ozonation has been used for years in Europe, particularly to purify drinking
water. Research has shown that BOD and COD can be reduced by ozonation.
Recent work on ozone decomposition kinetics in water (5) and on the removal
of organics from waste water (6) has clarified the mechanism of ozone
reduction of organic content and has demonstrated that high utilization
efficiencies can be obtained.
-------
SECTION 2
CONCLUSIONS
1. Many light Chlorinated hydrocarbon solvents (1-3 carbon atoms) can be
effectively extracted from industrial wastes.
2. In most instances the same substances can be steam stripped from the
wastes with lower cost.
3. Solvent extraction is likely to compete with stripping when the water
solubility is higher and the volatility lower than usually encountered
in light chlorinated hydrocarbon wastes.
4. Toluene diamine can be extracted with benzene and the benzene plus other
volatile constituents can be steam stripped from the water.
5. After extraction with Ci2-l3 paraffin solvents, wastewaters in these
classes generally contained unidentified nonextractables ranging up to
500 ppm. These nonextractables may have included oxygenated compounds.
6. Ozone is not effective in treating glycol wastes.
7. Ozone is not effective at TOC (Total Organic Carbon) levels of several
hundred.
8. Ozone removes color and odor very much more rapidly than it removes TOC.
9. Ozone is generally effective in removing color.
10. At a given ozone level, the rate of reaction with organics is proportion-
al to the organic level. Even though the cost of treatment with ozone
decreases with decreasing organic load, the efficiency of ozone util-
ization also decreases.
11. As ozone has a short life, it is very important that proper contact be
made between the gas and the waste water.
-------
SECTION 3
RECOMMENDATIONS
The chlorinated wastes considered in this study should be treated by
stripping rather than extraction if only chlorinated hydrocarbons are to be
removed. However, the large amounts of material not extracted by paraffin
oil or benzene should be identified, and perhaps slightly miscible but
volatile solvents examined. Extraction should be considered in treating the
toluene diamine waste.
Dual Solvent systems should be studied in an effort to determine more
efficient and more flexible extraction systems.
Ozone contacting systems should be studied in more detail with partic-
ular attention to the gas-liquid, gas solid, and solid liquid interfaces.
The effects of these interfacial interactions on the chemical kinetics of
both ozone decomposition and the subsequent reaction with organics should
be studied.
Ozonation should not be used as a primary treatment for highly
contaminated wastes. Ozone has always been envisioned as a tertiary
treatment to polish wastes where treatment is less than adequate by other
methods.
-------
SECTION 4
DISTRIBUTION COEFFICIENTS FOR PURE COMPONENTS
A high percentage of the industrial waste streams studied in this research
program contain chlorinated hydrocarbons with EDC (1,2-dichloroethane)
being of principal concern. Though the waste streams are a complex mixture
containing unidentified components, some of which cannot be extracted, it
was felt that data on some pure components would be helpful. Since some of
the waste streams are quite acidic, data were also obtained in 15% HC1,
which as expected, had some effect on the results
The data are given in TABLE 1. These data are for screening purposes and
are not of high precision. To save time only the water phase was analyzed,
but the data are sufficient to show trends. Generally with distribution
coefficients as high as these, the precise value is of less significance
because the size of the extractor will increase only slightly with decreas-
ing distribution; and solvent recovery cost is independent of the distribu-
tion except as it affects solvent rate which is already small.
Generally, the extraction data were obtained by agitating the water samples
with the solvent sufficiently to obtain equilibrium. The phases were
separated and the water phase was analyzed by the total organic carbon
analyzer or the gas chromatograph.
From the data given in TABLE 2 several points may be noted. As would be
expected, the distribution coefficient increases with decreasing solubility
in water and decreasing Hildebrand solubility parameter. The coefficient
increases, generally, with chlorine content and with asymmetry in chlorine
distrubution; i.e., 1,1,1-trichloroethane is greater than 1,1,2-trich-
loroethane. Unsaturation increases the distribution coefficient.
-------
TABLE 1. PURE COMPONENT DISTRIBUTION DATA
Solvent Water = 1/10
Concentration in mgm/1
Compound
Methyl ene chloride
Chloroform
Carbon tetrachloride
Ethylene dichloride (EDC)
15% HC1
1 ,1 ,1-Trichloroethane
1 ,1 ,2-Trichloroethane
Water
feed
1,510
1,705*
170*
8,000
6,000
2,000
1,120
600
680
68
8,700
4,350
1 ,305
820
82
7,013*
721*
70*
6,992*
703*
70*
1,000
300
2,800
1,492
746
280
4,028
802*
Water
phase
888
687
66
1,096
1,176
272
118
109
0
0
3,340
1,397
469
230
20
1,450
176
16
2,750
260
45
78
0
582
292
152
73
446
143
Oil
phase
6,220
10,200
1,040
69,000
48,200
17,300
10,000
4,900
— _
—
53,600
29,500
8,400
6,900
620
55,600
5,450
540
42,350
4,400
250
9,200
—
22,200
12,000
5,400
2,100
35,800
6,590
Distribution
KVL
7
15
16
63
41
64
85
45
> 100
> 100
16
21
18
30
31
38
31
34
16
17
5.5
.118
--
38
41
39
28
80
46
-------
TABLE 1 (continued)
1,1,2-Trichloroethane
15% HC1
882*
270
6,120
23
1 ,1 ,2,2-Tetrachloroethane
1,2-Dichloroethane (CIS)
Trichloroethylene
Tetrachl oroethyl ene
Benzene
2,620
1,279
640
202
378
189
95
860
646
86
372
37
1,092*
475
346
156
50
62
27
16
129
52
0
0
0
66
15,500
9,300
4,800
1,500
3,200
1,600
785
7,300
5,900
10,300
33
27
31
30
51
60
49
57
114
> 100
> 100
150
*Analysis by TOC
-------
TABLE 2. CHLORINATED SOLVENT DATA
CH2C12
CHC13
cci4
1,2-C2H4C1
23
1,1,2-C2H3
1,1,2,2-C2
1,2-C2H2C1
C2HC13
r n
L2LI4
Distribution
Coefficient
KvL
7-16
60
>100
20-30
ci3 >ioo
C13 40
H2C14 30
2(CIS) 50
60-100
>100
Density Sol . H20 Sol .
( 3) / wt. s param- B.P.
cm ^fraction' eter °C
1.33 .02 9.8 40
1.484 .0073 9.2 61-62
1.589 .0008 8.6 132-134
1.246 .009 9.9 83-84
8.5 74.1
113-114
1.587 .0028 146.5
9.1 60
86.7
10"4 9.3 121
_.
-------
SECTION 5
EXTRACTION DATA ON WASTE STREAMS
The results of extracting various industrial wastes with a paraffinic
petroleum oil having approximately the properties of tridecane plus a few
data using benzene as solvent are shown in TABLE 3. Except for the last
data with a toluene diamine waste, the ratio of solvent to water was one to
ten. The results are confusing and sometimes contradictory. This probably
results from sample variability and the fact that large amounts of non-
extractables are in most wastes. These nonextractables are not chlorinated
hydrocarbons such as were investigated in the previous section. Part of the
difficulty also results from the lack of correlation between TOC (Total
Organic Carbon) and chromatograph analysis.
In order to obtain some meaning from these results we have assumed that the
TOC can be divided into two parts: the first part is extracted at constant
distribution coefficient while the second part is not extracted at all. For
that part that is extracted, the concentration remaining after each extract-
ion will plot as a straight line on semilog paper since each extraction
removes a constant percent of the remaining extractables. Therefore, by
trial and error, a number is found which when subtracted from all the TOC
values causes the remainder to plot in a line, Figure 1-3. The data in
Fig. 1 only involve three points and the results do not constitute a proof of
the assumption. For any three points, x, y,z, if x>y>z and (x-y) > (y-z).
Then a constant can be found to satisfy the relation log (x-c)/(y-c) =
log (y-c)/(z-c) which gives the desired straight line. The data in Figure 2
with five extractions is a better confirmation of the assumption. The data in
Figure 1 are plotted without regard to the magnitude of the ordinate to show
slopes only since the same slope indicates the same distribution coefficient.
We will begin with sample Oil A which is principally a high pH EDC waste.
The plot in Figure 1 gives a slope completely different from pure EDC and
most other EDC wastes, in spite of the fact that EDC analysis indicated that
nearly all the EDC was extracted. The final entry in TABLE 3 for Oil A was
an attempt to remove all the extractables. All the EDC was removed, but the
TOC data are contradictory and completely inconsistent with the EDC data.
For instance, over 3000 ppm of EDC was measured as against 350-550 ppm of
TOC, and the complete removal of the EDC only dropped the TOC by 200 ppm.
Sample 011 IB, also a high EDC waste, but acid instead of alkaline as was Oil A,
showed similar extraction characteristics with a large quantity of unident-
ified nonextractables.
-------
0 I 2
NO. OF EXTRACTIONS
Figure 1. Comparative extraction coefficients for various
industrial waste streams
10
-------
1000
<
Q
Q_
Q.
100
10
I
I
1234
NO. OF EXTRACTIONS
Figure 2. Extraction of toluene
11
-------
100
10
(f)
UJ
_l
00
o
<
cr
I-
x
UJ
Q.
0.
SOLVENT _J_
WATER " 10
K^IIO
I
I
0 I 2
NO. OF EXTRACTIONS
Figure 3. Extraction of aromatic waste
12
-------
Sample 011C, reported to be principally EDC plus some methyl chloride (which
would have been lost) is an alkaline sample but less so than 011A. Figure 1
indicates that the extractables behave much as EDC in sharp constrast to
011A. Sample 041A, another alkaline sample high in EDC, shows TOC removal
even more rapid that OllC. The data on EDC removal obtained by GC (Gas
Chromatography) rather than TOC typically show an even more rapid removal
than pure EDC as measured by TOC. Sample 041C(acid) shows a much lower
TOC than indicated by COD analysis, but is consistent with the chemical
analysis, which shows only small amounts of chlorinated solvents.
Sample 041D (ph = 7) contains aromatic constituents as well as EDC. A plot of
this waste is shown in Figure 3, and we see that all but 64 ppm of the TOC
is extracted at a very high rate.
Sample 041 F (pH>7) a chlorinated waste of fairly low TOC, shows a very rapid
extraction of about a third of the TOC.
Our analysis of sample 081A (pH >7) shows a low TOC about half of which
can be extracted. Samples 081B(acid) and OSlC(basic) shows very rapid
extraction of about two thirds of the TOC. These samples are interesting
in that 081B is very acid and 081C is very alkaline, but both show greater
removal of the extractables than obtained with pure EDC.
Sample 131A showed almost no extractables. Our analysis of sample 161A(pH=7)
showed a low TOC in contrast to a GSRI COD of about 100 ppm. Sample 161B
is an acid EDC containing stream with very rapid removal of nearly half of
the TOC. Sample 161C is a very acidic stream, supposedly containing
appreciable EDC and other chlorinated solvents, but our data show almost no
removal of the very high TOC.
Sample 221A is another highly acidic stream containing a high EDC content.
This sample shows very rapid removal of about 300 ppm TOC, but over 2000
ppm of nonextractables remain. Sample 221B, an alkaline high EDC waste,
shows TOC removal very similar to the highly acidic 221A. Again we found
a very high level of nonextractables. Sample 221C, with a high TOC but low
chlorinated solvent content, shows very little extractable content.
The final sample contains a high level of toluene diamine. Figure 2 shows
that about 2750 ppm of TOC can be extracted by benzene with a distribution
coefficient of a little over one.
13
-------
TABLE 3. EXTRACTION OF INDUSTRIAL SAMPLES
Solvent/Water=l/10
25°C
Concentrations in mgrn/1
Sample No. *£•
(analytical
parameter)
OllA(TOC) EDC
OllA(EDC) EDC
011A (Extracted with
OllB(TOC) EDC
OllC(TOC) EDC
041A(TOC) EDC
041A(EDC) EDC
041C(TOC) EDC+Arom
041D(TOC) EDC+Arom
Water
feed
1,233
3,964
Benzene 4 times
3,643(EDC)
3,292(EDC)
355(TOC)
522(TOC)
1,355
817
1,545
4,270
25
171
Water No. of
phase extractions
1,057
934
1,020
106
and oil 3 times)
9
0
79
312
1,224
1,144
403
301
714
522
739
277
21
19
73.2
65.3
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
14
-------
TABLE 3 (continued)
041F(TOC) methyl chloride 124
081A(TOC) tri + tetra
chlorohydrocarbons
081B(TOC) EDO
081B(TOC) EOC
081C(TOC) Tri +Tetra
chlorohydrocarbons
131A(TOC) Arom.
161A(TOC) EDC
161B(TOC) EDC
161C(TOC) EDC
221A(TOC) EDC
221B(TOC) EDC
124
35.0
IS
25.0
29
82
73
73
68
92
75
82
75
94
683
26
254
2,284
2,384
2,700
97
88
24
20
10
15
13
25
24
32
31
45
29
44
35
34
34
35
28
29
27
664
641
21
15
152
139
2,263
2,247
2,083
2,091
2,405
2,345
1
2
1
2
1
1
2
1
2
1
2
1
1
1
1
1
2
1
2
1
1
1
2
1
2
1
2
1
2
1
2
1
2
15
-------
TABLE 3. (continued)
2,766 2,394 1
2,387 2
2,691 2,346 1
221C EDO 1,880 1,813 1
1,797 2
1,534 1,402 1
1,394 2
1,612 1,548 1
Toluenediamine (TOC) Waste - Benzene/Water = 1/1
3,000 1,600 1
1,100 2
700 5
16
-------
Section VI
DESIGN OF SOLVENT EXTRACTION PROCESS TO REMOVE CHLORINATED SOLVENTS
FROM WASTE WATER
The purpose of this section is to give an approximate design for
extracting an idealized waste water to demonstrate general design procedures
and probable equipment and utility requirements. The initial concentration
of chlorinated solvent in the water was chosen a little below saturation.
The final concentration was chosen arbitrarily at a reasonably low value.
The design is based on using a paraffin, highly water insoluble oil
to extract the chlorinated hydrocarbons, assumed to be EDC, from the water.
A distribution coefficient at the lower end of the range found for the
industrial samples was chosen. The EDC is subsequently steam stripped from
the oil .
The extractor is an agitated column whose design and power consumption
are based on data obtained from du Pont for a process in which dimethyl
formamide is extracted from water. As oil and water are immiscible and as
the concentration of EDC is low, a constant extraction factor throughout
the column may be assumed leading to a very simple calculation of the
number of theoretical stages. For a completely stripped solvent, the
fraction extracted is /pn+1 rw(irn+l -,\ in which N is the number of
theoretical stages and E is the extraction factor, K-oil rate/water rate.
In the oil stripper there is considerable variation in both the
equilibrium constant and the vapor flow rate through the column, so the
number of transfer units is calculated using the stripping factor at each
end of the tower (Perry's Chemical Engineer's Handbook, 4th edition, p!4-
29). The height of a transfer unit is based on ranges recommended by U.S.
Stonewear.
Arbitrary and safe values were chosen for the oil rate in the extractor
and the stripper steam rate. The values are sufficiently small that costs
would not be greatly affected by reasonable changes.
EXTRACTOR DESIGN
A sketch of this system is shown in Figure 4. The design bases and
calculations follow.
It is assumed that the feed contains 8000 ppm of EDC, and that the
product contains 15 ppm EDC.
The distribution coefficient K = Concentration in Oil/Concentration in
water, equals 15 on a volume basis. This is a lower value than that of
pure EDC but typical of some of the waste streams.
17
-------
CONDENSER.
00
FEED 100 GPM
^
r
cr
i^.
o
p
0
<
cr
H
X
UJ
1 *_
IOO"F_
II8°F
230°F \/ )
1
STEAM ^C^
2 ^/
HEAT
EXCHANGER
^B A\\^ 1 1 P^ 1 V \^^» 1 *
MIN v y
V -X
i
200°F
i
^
Q£
LJ
fc
/y
LL.
65
1
T
HgO 9 fl^/MIN.
EDC -6.67 ^/MIN.
OIL -1.8^/MIN.
STEAM
9 ^/MIN
1 OIL 20 GPM
100 GPM
Figure 4. Extraction system.
-------
The volumetric flow ratio of oil/water equal 1/5. Smaller ratios
might effect extractor efficiency. The oil is a paraffin hydrocarbon
approximated by tridecane. We will assume an agitated column with the
following characteristics:
H.E.T.P. = 5 /DTTnT
Combined flow = 300 gal/hr-ft
The extraction factor, EV = °wat*r = 3
With 5.3 stages we have the fraction extracted given by
pN+1 6.3,
fraction entracted = ,,,-. = " = 0.998
E -1 36'?!
which is the required recovery, where N is the number of theoretical stages.
Tower diameter is obtained by
12Q £§1 60 min _L _?£_ = PA ft2
'^U min x hr x 300 gal/hr ^ Tt
D = / 24 x 4 ^ 5 1/2'
IT.
Tower height becomes
H.E.T.P. = 5.3 / 5.5 x 12 = 43"
H = 4°'^X 5 = 18, adding 5' for separation gives 23'.
Power = 10 HP
OIL STRIPPER DESIGN
The stripper will be fed at 110°C or 230°F. We will assume that oil and
EDC form an ideal mixture. This will be conservative since there is some
positive deviation from ideality.
The moles of oil flowing is
20 jjjflx(. 79X8.33)^ ^/u>^ =0.716
. .
Use a steam rate of 3/2 minimum or 0.50 moles/min, or V = 0.5 x 18 x
60 = 540#/hr.
In stripping this mixture the moles per minute of EDC removed is
19
-------
100 x 8.33x 8000 x 1(T6
The heat required = 0.067 x 13,600 = 910 Btu/min
This cools the oil ,
t _ 910 Btu/min a, 12°F
132£ x.55Btu.
min #-°F
Therefore the oil leaves the stripper at 21 8° F.
At the top of the column
V2
V = .5 moles steam + .067 moles EDC = .567 mole/mi n
c _ VK _ .567 x 2 _ , ,Q
STOP " L~~ 7JW ~ ''58
At the bottom of the column
V1-55
Calculating the number of transfer units with 99.5% recovery gives
NTOP = 12
NBOT = 20
Assume seventeen transfer units and the height of a transfer unit to be
2 feet, then the column height is 34'. plus 1 foot at each end, or 36'.
Employing the US Stoneware charts for flooding in packed towers and
using 1" Intalox saddles we obtain a tower diameter of 1 ft.
HEAT EXCHANGER CALCULATIONS
Q = WC (Vt, ) = (20 x 8.33 x 60 x .79)#/hr x '51 1%
P C. I n~ r
= 400,000 Btu/hr
U = 75
At= 18
A = 400.000 =
M 75 x 18
20
-------
OIL HEATER
Q = 10,000#/hr x 30°F x -^°^ = 165,000 Btu
hr-ft
U = 100 Btu/hr-0F-ft2
At =30°
L2
A = = 55 ft
TOO x 30
CONDENSER
9# min x 60 = 540#/hr steam
400#/hr EDC
Q = 540 x 970 + 400 x 137 = 580,000 Btu/hr
U = 200 Btu/hr - °F - ft2
At = 75°F
A _ 580,000 Btu/hr _ 3g ft 2
200 x 75
STEAM CONSUMPTION
540#/hr to stripper
and 165,000 Btu/hr = }7Q#/. .
970 Btu/# i/u#/hr to heater
Total Steam = 710#/hr
OIL CARRY OVER:
at 230°F, K . = .017
' oil
Mol
es (EDC + H20) = 0.567 moles/min
Moles oil/mill = 0.567 x Q'983 - 0.01 moles/min
This is 1.84#/min of oil.
This would have only slight effect on previous calculations but would
require a small evaporator to remove the EDC from the oil and perhaps
another 60#/hr of steam.
21
-------
THE SIGNIFICANT EQUIPMENT LIST IS
5 1/2' x 23' agitated extractor
36' stripper
300 ft heat exchanger
55 ft2 heater
2
39 ft condenser
two 20 GPM oil pumps
one 100 GPM water pump
Oil surge tank, about 200 gal.
6,000 gal. feed tank
50 gal, decanter
The purchase costs of the major items of equipment are shown in TABLE
4. The total capital investment would be about 4 times as large. So we
have a total capital investment of perhaps $400,000 and a steam consumption
of about 700 Ib/hr.
TABLE 4
Purchase Cost of Major Items of Equipment for Extraction System (1973)
Extraction System
Stainless Carbon Steel
Extractor
Heat exchanger
Heater
Condenser
Pumps & Motors
Stripper
Tanks
68,000
9,000
3,700
3,000
3,800
14,000
13,300
68,000 (ss)
3,600
1,600
2,500
2,500
5,600
5,500
TOTAL 114,800 89,300
22
-------
SECTION 7
DESIGN OF PROCESS FOR REMOVAL OF CHLORINATED-SOLVENTS FROM WASTE STREAMS
BY STEAM STRIPPING
(A comparison of solvent extraction and stripping)
Though this report is concerned with solvent extraction, nearly all of the
chlorinated solvents for which extraction data are given in Table 1 can
also be removed by stream stripping. Since stripping is inherently a
simple operation it seems pertinent to design stripping and extraction
processes for the same components to get some measure of their relative
process and economic advantage.
STRIPPER SYSTEM DESIGN
A schematic diagram of this system is shown in Figure 5. The following
physical properties will be used:
C (H20) = 1 Btu/lb - °F
AHV(H20) = 970 Btu/lb = 17,500 Btu/mole
AH (EDC) = 137 Btu/lb = 13,600 Btu/mole
The mole fraction of EDC in the feed is 0.001465.
We do not have vapor liquid equilibrium data for the EDC-HpO system.
However, because of the low mutual solubility, it is possible to con-
struct useful equilibrium data from solubility and vapor pressure data.
The solubility of EDC in water at temperatures to 70°C is shown in
Figure 6. Vapor pressures for EDC and other substances are given in
Figure 7. We are only interested in the vapor-liquid equilibrium for the
water rich phase. Assuming that the saturated water phase is in equilibrium
with pure EDC we may write, assuming Henry's law.
p* = H y
EDC n'AEDC-Sat.
We want the Equilibrium Constant at 1 atm so dividing both sides by 1 atm
we obtain
23
-------
CONDENSER
FEED 100 GPM-
IOO°F
IIO°F
HEAT
EXCHANGER
I99°F
2I2°F
UJ
o_
o_
cr
»-
CO
EDC - 6.67
H20 - 4.25
STEAM
13.75
Figure 5. Steam stripping system.
-------
100
u
e
90
80
70
60
50
40
30
00
20
7000 8000 9000 10000 11000 12000 13000
EDC , ppm
Figure 6. Solubility of EDC in water.
25
-------
10
8
6
4
1.0
.8
.6
~ A
1
°- .2
O.I
.08
.06
£>4
.02
/T X O4' 24 25
I
26
I
27
I
28
29
I
TeC! 140 130 120 110 100 90 80
70
30
I
60
31
I
50
Figure 7. Vapor pressures of various substances.
26
-------
YEDC = P*71 = (H/1)x = KX
Klatm = P*/XEDC-Sat.
where
P* = Vapor pressure, atm
H = Henry's law constant, atm
V
EDC-Sat. = Mole fraction EDC at saturation
^EDC = ^°^e fraction EDC in vapor
K = Equilibrium constant, Y/X
Since the water phase is practically pure water, K for water at one at-
mosphere is equal to the vapor pressure in atmospheres.
Reading solubility from the extrapolated solubility seen in Figure 6 and
vapor pressures from Figure 7 the equilibrium data in Fiaure 8 were calcu-
lated. Assume the stripped water leaves the stripper at 212°F and is
cooled in the heat exchanger to 100°F. Assume the feed enters at 100°F.
A heat balance will determine the temperature into the column.
Assume 199°F = 365.7°R
From Figure 8 we obtain !<„„ = 540, and fy g = 0.74. A flash calculation
on 1 mole of feed yields 0.000482 moles of EDC remaining in the feed and
0.000983 moles of EDC, and 0.0028 moles of H20 flashed. A heat balance on
a mole of feed gives
18(102) = 18(t-100)+ 13,600 x 0.00098 + 17,500 x 0.0028
t = 198.5°F which is close
The maximum L/V ratio at the top of the column = 540. Take an actual
value of 2/3 this to obtain L/V = 360 = 0.00278 moles of vapor. The
total vapor leaving the top of the column is the sum of this plus the
flashed vapor or 0.00656 moles per mole of feed. Since essentially all
the EDC has been stripped, this vapor contains 0.00146 moles of EDC and
0.0051 moles of water.
An overall heat balance can now yield the amount of steam entering. The
heat leaving per mole by incomplete temperature approach in the heat ex-
changer is 18 x 10 = 180 Btu, so
27
-------
1000
800
600
O
O
UJ
200
1/TxlO '25
T°C '
26
27
28
29
100
90
80
70
30
.1
.8
.6
O
C\J
.4
.2
Figure 8. Vapor-H quid equilibrium constants for
EDC and Water (Water Rich Phase) at 1 atm
28
-------
180 = 17,500 (V - 0.0051) - 13,600 x 0.00146
V = 0 0165 mo]es steam
v u>ult)b moles of feed'
SIZING COLUMN
The equilibrium constants at top and bottom are
KTOP = 540
KBOTTOM = 640
V/LTOp = 1/360 = .00278
V/LBOTTOM = .0165
The stripping factors, KV/L, are
STOP = 540 x 0.00278 =1.5
SBOT = 640 x 0.0165 = 10.5
The mole fractions are
XTOP (after flashing)= 0.000482
XBOTTOM = 15 x 10"6 x 18/99 = 0.0000027
for a ratio XTOP/XBOT = 178
Calculating the number of transfer units from S np and SRnj we have
NTOP = 12
NBOT ' 5
Use N = 9 and Ht = 2'
Where N is the number of transfer units and H^is the height of a trans-
fer unit so the tower height = 18'. Add 6' for ends and distribution to
obtain 24'.
Calculation of the column diameter was made using flooding charts of U.S.
Stoneware and assuming 1" Intalox saddles. A flow rate of 70% of flood
was assumed yielding a 2' column.
29
-------
HEAT EXCHANGER
The heat transferred in the feed-bottom exchanger is given by
Q = 100 ^ x 60 $0. x 8.33 gi x 102°F = 5.1 x 106
U = 200 Btu/hr - °F - ft2
At = (10 + 13)/2 = 11.5°
c 1 x 10^ ?
A = J-' A IU = 2220ft
A 200 x 11.5 «*UTI
CONDENSER
#EDC/hr = 400
#H20/hr = 0.0051#/#Feed = 255#/hr.
Q = 400 x 127 + 255 x 970 = 302,000 Btu/hr
U = 200
At = 75°
A _ 302,000 _ 2n f 2
A 200x75 d() Tt
Assuming the feed is available at sufficient pressure to enter the column,
one 100 GPM pump is required for the column bottoms and a very small pump
for the recovered EDC: The condensed steam could be allowed to flow by
gravity back into the stripper feed pump suction.
THE SIGNIFICANT EQUIPMENT LIST BECOMES
2' x 18' packed column
2
2220 ft heat exchanger
2
20 ft condenser
2-100 GPM pump
6000 gal. feed tank
30 gal. decanter
The costs of the principal items of equipment are shown in TABLE 5. The
total installed cost would be about four times this much but a comparison
with the results in TABLE 4 indicates that stripping is clearly preferable
The steam consumption is about the same and the equipment is at least
50% more for the extraction system. Since in the extractor system, the
extractor cost is predominant and its diameter would only be slightly
30
-------
reduced by reduced oil flow, it would appear that extraction would most
likely compete with steam stripping only in systems of poor relative
volatility.
TABLE 5
PURCHASE COST OF MAJOR ITEMS OF EQUIPMENT FOR STRIPPING SYSTEM (1973)
Stripping System (Stainless)
Stripping Column 19,000
Heat exchanger 30,000
Condenser 2,000
Tanks 11,000
Pumps & Motors 2,000
64,000
31
-------
SECTION 7
PILOT PLANT DESIGN
We have designed a pilot plant to handle a mixture of volatile and less
volatile chlorinated compounds in water using a paraffin hydrocarbon oil
as the solvent. The design follows and is shown in Fig 9.
Feed rate = 1 GPM
Solvent rate = 0.2 GPM
Solvent Properties
Hydrocarbon paraffinic oil - approximately C-j2 - C13
Viscosity at 100°F = 2 Centipoise
Viscosity at 265°F = 0.6 Centipoise
Specific gravity = 0.76 - 0.78
Design Feed Composition:
Water Contaminated with:
Ethylenedichloride = 9,000 ppm
1,1,2,2 Tetrachloroethane = 500 ppm
EXTRACTOR DESIGN
On a volume basis
KEDC = 15
"1.1.2.2 =25
EEDC = 3
El,l,2,2 = 5
Based on constant oil to water ratio and distribution constant, tne relation
of these variables to the number of theoretical stages is given by
32
-------
Xl EN+1-E
XN+1-X* EN+1
where
k _ volume Concentration in Solvent
v ~ volume Concentration in Water
E = K L /L
v ov' wv
LQv = Solvent flow rate, volume
LWV = Feed water flow rate, volume
N = Number of stages
V
1 = Concentration of solute in water leaving extractor
= Concentration of solute in water entering extractor
X* = Concentration of solute in water in equilibrium with the
stripped solvent
For EDC, with essentially no recycle in the oil solvent, the left-hand side
is the fraction removed and is equal to 0.997 with 5 stages. For 1,1,2,2-
tetrachloroethane we will essentially have equilibrium with the recycle
oil and recovery depends on stripping efficiency.
Packed Column Extractor
Estimation of diameter based on U.S. Stoneware design charts indicates
7 1/2 - 8" with 1/2" Intalox saddles. The height is difficult to estimate.
Probably between 3 and 5 feet per stage will be required giving, say 25
feet of packing.
It is our recommendation, however, that manufacturers of mechanically
agitated equipment be contacted before an extractor is decided upon.
STRIPPER DESIGN
The Steam stripper is at one atmosphere with a Liquid/Vapor ratio of one
on a mole basis.
33
-------
CONDENSER
CRUDE PRODUCT
STILL
I GPM
FEED PUMP
RAFFINATE
RECYCLE TO
FEED
Figure 9. Pilot plant design
-------
1.27#/min _ nr,7/L7 moles oil
170#/mole " -UU/H/
OQ747 moles, steam = OJ34 £|tejm
at 265°F
KEDC = 3'5
Kl,l,2,2
_ mole fraction in vapor
mole fraction in liquid
The fraction solute removed is given by
SN+1 S
fraction removed = — rrpj —
S -1
in which S = T— = 3.5
Assume 6 stages
This gives essentially all EDC stripped. For 1, 1, 2, 2, tetrachloethane
the results from the above equation is indeterminate, but for this case the
fraction removed is N/N+1 or 6/7.
Packed Column Stripper
Diameter = 3"
Packing 1/4' Intalox saddles
Height at least 8' with redistribution of liquid at several points
Calculation of Final Raffinate Concentration
Concentration of 1,1,2,2 remaining in oil
•pa (~\r\
1/7 x 500 ppm in feed x 5/.76 wt. ratio ^— = 470 ppm
470
Concentration in equilibrium with oil = ~r-= 19 ppm
fraction tetrachloroethane removed = - - = 0.963
EDC concentration remaining = 9,000 x 0.003 = 27 ppm
35
-------
ft.
ft. (assuming at least 50 psig steam)
ft.
Pilot Plant Equipment
Figure 9 shows flow rate, temperatures where important, tank capacities,
and controls.
Equipment Required
Feed Storage
Feed Pump - 1 GPM
Two Solvent Pumps - 0.20 GPM
Condenser - 1.5 sq.
Steam Heater - 1 sq
Solvent Cooler - 5 sq.
Solvent Tank - 20 gal.
Extract Tank - 5 gal.
Decanter - 5 gal.
Extractor
Stripper
Crude Product Still
Controls
Feed Flow Control
Solvent Flow Control
Stripper Steam Flow Control
Stripper Feed Temperature Control
Extract Tank Level Control
Stripper Level Control
Discussion of Design
Extractor Though a packed column is specified, vender quotes for
agitated towers should be obtained.
Crude Product Still
Heat Exchangers
Solvent Tank
Not shown in this design is a crude product still.
The overhead from the stripper will contain a high
fraction of oil and some kind of recovery system
might be desirable. This could be a batch operation.
The condenser, cooler, and heater are so small that
fabricated double-pipe design could be used. *It
might be asked why a Solvent-Extract exchanger is not
specified. In the first place the system is too small
to make heat recovery important, and secondly, with
This exchanger the solvent pump would have to pump
265°F oil complicating the specifications of the pump.
This tank should contain about 20' of 1/2" tubing for
cooling water which will drop the temperature below
150°F at which almost any kind of solvent pump can be
used.
36
-------
Extract Tank This is strictly for surge capacity between the
extractor and stripper.
Stripper This unit is designed for feed containing some heavier
components as represented by 1,1,2,2 tetrachloroethane.
If not present, the stripper steam flow or temperature
could be reduced.
37
-------
SECTION 9
THERMODYNAMIC ANALYSIS OF EXTRACTION VS STRIPPING
OF VOLATILE DILUTE COMPONENTS
The following is an attempt to apply thermodynamic analysis to the case
in which a volatile, dilute solute - and in particular the chlorinated
hydrocarbons - is extracted by a nonvolatile solvent from water.
We will assume that the solute in water obeys Henry's law
p = HX.
where
p = partial pressure, atm
H = Henry's Constant, atm
X = mole fraction
As discussed earlier, at saturation we may assume the dilute solution in
water to be in equilibrium with pure solvent, so p, the partial pressure,
is equal to P* the vapor pressure.
P*= HX ,.
sat.
or dividing by the total pressure n
I • V = <£' "sat - Vsat
or the equilibrium constant for vapor-liquid equilibrium is
P*
1^ _ £
9 ' nxsat
Where K = Y/X = Vapor mole fraction/liquid mole fraction.
We will assume that the solute deviation from Raoult's law in the solvent
phase can be expressed by
p = P*YX
where y is the activity coefficient.
38
-------
In the solvent phase
For liquid-liquid equilibrium, the partial pressures are equal in each
phase, so Y in equilibrium with each phase is the same
or
X° - K -
Xw L "
where X and X are the mole fractions in the oil (assuming an oil solvent)
and water phase respectively. This is true for any slightly soluble
component and should apply to the extraction of all the chlorinated
hydrocarbons studied in section IV.
The ease of extraction is given by the extraction factor
Kl Ln
E = -L-°
Lw
The ease of stripping is measured by the stripping factor
w
The ease of solvent recovery is given by
K V
c go o
L
o
where
K. = Liquid-liquid equilibrium constant
K =Vapor-liquid equilibrium constant in the water stripper
39
-------
K = Vapor-liquid equilibrium constant in the oil or solvent stripper
L ,L,,= Oil and water flow rates in moles per unit time
o w
V = Steam rate in the water stripper, moles per unit time
w
V = Steam rate in the oil stripper, moles per unit time
Substituting the definitions of K
L
E =
YEXsat,ELw
c P* V
V w w
nxsat,w Lw
SQ PVoVQ
"Lo
where X . c is the mole fraction of EDC in water at saturation in the
sa u, t
extractor, X , is the mole fraction of EDC in water at saturation in
5 Q t 5 W
the water stripper, and P* and P* are the vapor pressures of EDC at the
temperature of the water and oil stripper respectively.
We see from the cost analysis of stripping vs. extraction that S and S
are at least the same order of magnitude. For simplification let us
equate them.
X ^ L |
sat,w w L0
reorganizing and substituting,
P* V E .
sa
Vr <. c
w w YEAsat.E _ v
y 00 0
sat,w
40
-------
P * X 4. V Y
E = ( ° \ ( sat.Wx , vo x , yo x
P * ' * X ' ^ M ' \ Y,- '
w Asat,E w yE
Since the cost estimate indicates the extractor is the major cost, E
should be large so
P _/ o \ i sat,w ) , o \ r ^o \ ,
E -(p*-; ( x ' ( \r ' ( T~) >> ]
w sat,E w Yw
but
W
perhaps as much as 2 or 3 by running up the oil stripper temperature
X
Ysat'w 'V 1.5 according to EDO data
Asat,E
y
— < 1 perhaps as low as 1/2
YE
y— -v 1 or less
w
One sees that we cannot have a large E unless we allow
S « s
o w
which would run up the cost of the oil stripper.
The problem may be summarized as follows:both extractor and stripper
efficiency are inversely proportional to solubility,
E - '/xsat,E
V ' 1/Xsat,w
41
-------
Similarly both steam stripping and solvent recovery stripping are pro-
portional to the vapor pressure
i/ „ p*
Kgw K w
Kgo ' p*o
While P* can be greater than P* and Xc r is a little larger than
0 W So t j t
xsat w' this 1S Partly offset by Y0 < YE- Tnis Plus tne fact tnat ex~
tractors tend to be larger and more expensive than strippers makes it
unlikely that these relatively small ratios will allow extraction to
best stripping in these systems.
To check the basic assumptions, let us calculate K for EDC and compare
it with the experimental value.
The activity coefficient of the oil-EDC system should agree fairly well
with the theory of Hildebrand (9) in which the activity coefficient at
infinite dilution is given by
v =
P
where
3
V = Molar volume, cm
<5 = Hildebrand's solubility parameter
MW = molecular weight
P = density, gm/cm3
V 99
EDC = — ^ — = 79
1.246 y<
V ., _ 184 _ ,„
oil - - = 233
.79
YEDC =3.4
42
-------
This agrees generally with Ypnr in oil calculated from vapor pressure
measurements.
Xsat,EDC = 900° P"1
on a mole basis.
Xsat, EDC=9000xMxl°~6 = 0-00164
K,= ] = 180
L 3.4 x .00164
This is on a mole basis. To convert to a volume basis we have
MW(H00) „
K = 180 x 18 x .79 ^ 14
L 184 x 1
This is a little below measured values.
In Table 6 the value of ^—,P*(100°C) and P*/XC3, , the factors that
sat. sat'
primarily determine the efficiency of extraction, solvent recovery, and
stripping are given. In the table, we have ignored the difference in P*
and P* and X . , since as shown, these factors are likely to cancel.
W 5 Q \* 3 W
Considering the cost of the EDC extractor, we should look for a system
with larger 1/X relative to P*/X. For perchloroethylene the solubility
is so low that extraction or stripping can hardly be justified. The
best system in Table 6 might be CpCl.Hp, but note that the stripping factor
for this compound is only slightly le"ss than for EDC, and the low pressure
would require a high oil stripper temperature, so even this system is
marginal. Incidentally, if both EDC and 1,1,2,2-tetrachlorethane were
present, the stripping system could handle it easily. With extraction,
we would have a very difficult situation in that EDC would limit the
extraction system and the heavy component would limit the solvent recovery
system.
43
-------
TABLE 6. EXTRACTION AND STRIPPING FACTORS
Solute
CH2C12
CHC13
C14
Cf~* "I TT / 1?T\ f~\ \
r» rt / v J-'A' v j
C2C14H2
C2C14
Xsat
0.00425
0.0011
9.3 x 10~5
0.00164
0.0003
1 x 10~5
P*(100°C)
7
3.1
2
1.6
0.24
0.50
(atm)l/Xsat
235
910
11,000
610
3,300
105
P*/X ^
sat
1,650
2,800
21,000
1,000
800
50,000
From this analysis it appears that there is little chance that solvent
extraction can compete with strioping for recovery of chlorinated solvents
from waste waters.
44
-------
SECTION 10
OZONATION OF WASTE WATER STREAMS
It has been shown in earlier work (5, 6) that ozone is capable of
destroying many, refractory organic compounds in dilute water solutions.
These data indicate that at low concentration, ozonation can be an
economic tertiary treatment for producing high quality water. It was
found that such variables as pH, temperature, and 0~/0? ratios were im-
portant in obtaining efficient ozonation.
The experiments reported here are of two kinds. First there is a batch
system in which the water to be treated is placed in a 3£ flask and
ozone containing oxygen is bubbled through it continuously and samples
of water are withdrawn at time intervals for TOC analysis. Temperature
is controlled by immersion of the flask in a constant temperature bath.
The second kind of experiment was run in the continuous ozonation system
shown in Figure 10. This system consists of two reactors in series. Each
reactor is a 4" pipe, 5' long. Feed water is introduced into the first
reactor with a metering pump. The water in each reactor is circulated
at the rate of 2-3 gal/min through an aspirator which also recirculates
the gas in the reactor at a rate of about 10£/min. The combined gas-
liquid stream passes through a static mixer before re-entering the re-
actor. An ozone-oxygen stream, usually about U/min containing 70-80
mgm/£ of ozone, is introduced into the first reactor as shown in Fig. 10
by means of a separate aspirator. Gas from the first reactor flows to
the second reactor and out through a pressure regulator. Water from the
first reactor overflows to the second. By using a 2 stage cocurrent
flow, almost no ozone is present in the exiting gas.
RESULTS
Most of the ozone work was done on propylene glycol because two of the
waste streams were of this type and because glycol in general has been
found hard to treat. Figure 11 and Table 7 shows runs in the batch systems
with dilute propylene glycol at three temperatures. These results are
similar to those obtained with municipal wastes and various chemicals.
The rate is accelerated by increase in temperature, and the reaction
proceeds first order with respect to TOC concentration. The rate is less
than that experienced with municipal wastes.
The industrial wastes are all of higher concentration, so a number of
runs were made on water containing about 300 ppm of TOC in the form of
45
-------
ON
o
UJ
UJ
-1X3-
o:
o
o
<
UJ
cc
-OO—
o2-o3
UJ
f
Q
ASPIRATOR
-------
30
20
E
Q.
Q.
JO
1 8
or
z 6
UJ
o
o
O 4
O
O
A 30%
O 50%
Q 60%
15 30 45 60
90
120
180
TIME, MINUTES
Figure 11. Ozonation of propylene glycol solutions,
47
-------
propylene glycol, TABLE 8.
In runs with propylene glycol a rapid drop in pH was experienced, doubt-
less caused by the formation of organic acids. Various buffers were
tried. CaCU was used in the dilute data given in TABLE 7, but was
unsatisfactory in the concentrated runs. In runs 18, 21 and 22 in TABLE
TABLE 7. OZONATION OF PROPYLENE GLYCOL
(dilute)
#11
Temperature
Concentration
Buffer
Time
0
15
30
45
60
90
120
180
240
50°C
PG - 20.56 ppm
CaCl2
pH
7.30
5.50
5.60
6.20
6.52
6.90
7.19
7.30
7.30
TOC
28.04
19.07
7.63
5.05
4.30
0.34
3.13
3.05
3.85
#12
Temperature
Concentration
Buffer
60°C
PG - 20.56 ppm
CaCIo
Time
0
15
30
45
60
90
120
180
TOC
26.85
15.62
5.73
7.75
37
83
72
54
1.30
(continued)
48
-------
#13
Temperature 30°C
Concentration PG - 20.56 ppm
Buffer CaC"L
Time pH TOC
0 8.01 24.22
15 6.00 24.01
30 5.41 18.78
45 5.55 16.33
60 5.68 13.93
90 5.80 9.51
49
-------
TABLE 8. OZONATION OF PROPYLENE 6LYCOL
(concentrated)
RUN #18 Temperature:
Concentration:
Buffer:
Time
0
15
30
45
60
90
120
180
240
RUN #21 Temperature:
Concentration:
Buffer:
Time
0
30
60
90
120
150
180
210
240
RUN #22 Temperature:
Concentration:
Buffer:
Time
0
30
60
90
120
150
50°C
PG - 436.7 ppm
NaHC03 (2 gms)
60°C
PG - 436.7 ppm
NaHCO,
O
_2H_
8.6
8.5
8.5
8.6
8.6
8.7
8.9
9.2
9.4
50°C
PG - 391.55
None
TOC
322.8
315.0
304.2
304.8
287.3
274.2
TOC
334
280
289
293
290
308
288
190
31
TOC
273
247
261
242
204
136
54
34
29
ml of NaOh(n=0.0995)
0
9.1
17.5
21.7
14.45
4.9
50
-------
Table 8 (continued)
#15
Temperature: 50°C
Concentration: PG - 413.6 ppm
Buffer: CaCl0
Time ph TOC
0 8.00 347.0
15 4.35 329.1
30 3.70 326.0
45 3.55 323.2
60 3.60 311.7
90 3.85 280.0
120 4.22 249.6
180 4.85 174.9
240 6.90 42.5
51
-------
8, NaHC03 was used. In all data on concentrated systems an unexplained
phenomenon was encountered. There was a drop in TOC at the beginning.
If the system was unbuffered the pH dropped radically, and it is known
that ozonation is poor at low pH, then the TOC decreased very slowly
even in buffered solutions. After about 1-1/2 hrs the TOC dropped
rapidly. Run #22 is an attempt to measure the acid formation during
ozonation by stopping the ozonation periodically and titrating the acid
formed.
In this run,
mil eq of base - 67.65 ml x 0.0995N = 6.73
Assuming the acid comes from OH groups
r ^o -i nu ic i TOC i , 1 eqPG . 6.73 _ ,n TOC
6.73 mil eq OH = 36 (m11 eq PG) ( 2 eq OH > x "2T ' 6° ~T
That is, the TOC converted to acid is only slightly larger than the
48.6 mgm/£ of TOC that had disappeared, indicating that the conversion
to acid is the limiting step.
Late in the project, the continuous unit described above became operational
Runs were made on dilute glycol solutions and the results are given in
TABLE 9. Only pressure and flow were varied significantly in these runs
and it is difficult to draw any certain conclusion.
We may calculate the efficiency of ozone using a typical run, #24, as
follows:
mgm TOC destroyed/min = 0.150 £x (36.3 - 23.2) mgm0TOC = 1.91
III JO
mgm 0, fed/min = U x 77.1 = 77.1 mgm 0,/min
j O
mgm TOC destroyed
mgm 0, fed X 10 =
o
Eff = L9Z_x_LO =
77.1 ' ^
The factor of 10 comes from the fact that only 1/3 of the oxygen in
ozone is used and it takes 8 oxygens to destroy a molecule of propylene
glycol, or 128 mgm oxygen to destroy 36 mgm of TOC. Thus it takes
3 x 128/36 or about ten. These results are much poorer than previous
results for a variety of organic compounds (6) and it must be concluded
that glycol is relatively refractory to ozone.
52
-------
Batch Test on Toluene Diamine and Chlorinated Haste
The results of ozonating the toluene diamine waste is shown in TABLE 10.
The same drop in pH was obtained as in the glycol runs. In view of this
fact this waste was not tested further. It is noteworthy, however, that
the dark color of this waste was destroyed very rapidly.
TABLE 9. CONTINUOUS OZONATION OF PROPYLENE GLYCOL STREAMS
(Gas Flow U/min in all runs)
RUN #24
Reactors temperature: 51 C
Reactors pressure: 32 psig
Water feed rate: 150 ml per minute
Concentration of ozone in gas streams: mgm/a
Gas stream in: 77.1
Gas stream out: 1.3
pH and concentration: TOC values in mgm/2,
pH TOC
Feed 7.15 36.3
Reactor 1 6.20 24.3
Reactor 2 6.90 23.2
RUN #25
Reactors temperature : 53 C
Reactors pressure : 30 psig
Water feed rate: 150 ml per minute
Concentration of ozone in gas streams: mgm/£
Gas stream in: 77.2
Gas stream out: 0.8
pH and concentration: TOC values in mgm/£
TOC
Feed 7.10 33.5
Reactor 1 6.30 23.4
Reactor 2 6.40 23.0
(continued)
53
-------
RUN #27
Reactors temperature: 55°C
Reactors pressure : 12 psig
Water feed rate: 150 ml per minute
Concentration of ozone in gas streams: mgm/£
Gas stream in: 77.0
Gas stream out: 0.9
pH and concentration: TOC values in mgm/a
TOO
Feed 7.2 34.5
Reactor 1 6.2 23.5
Reactor 2 6.2 21.4
RUN # 28
. o .
Reactors temperature : 54 C
Reactors pressure: 0 psig
Water feed rate: 150 ml per minute
Concentration of ozone in gas streams: mgm/2,
Gas stream in: 77:0
Gas stream out: 1.7
pH and concentration: TOC values in mgm/£
TOC
Feed 7.10 33.7
Reactor 1 6.45 28.7
Reactor 2 6.45 28.3
RUN #31
Reactors temperature: 52°C
Reactors pressure: 28 psig
Water feed rate: 75 ml per minute
Concentration of ozone in gas streams: mgm/Ji
Gas stream in: 76.8
Gas stream out: 0.2
pH and concentration: TOC values in mgm/Ji
pH TOC
Feed 7.10 17.42
Reactor 1 6.60 16.63
Reactor 2 6.40 13.99
54
-------
TABLE 10. OZONATION OF TOLUENE DIAMINE
Temperature: 50°c
Concentration: TDA - 170 ppm
Buffer: none
Time (min) pH TOC
0 7.85 415.8 (dark color)
30 3.63 318.0 (color gone)
60 3.28 284.6
90 3.20 264.2
120 3.10 245.7
150 3.11 234.9
180 3.20 219.0
210 3.20 211.2
240 3.20 204.7
55
-------
A similar test was run on sample 011A high in EDC. This is a high pH
sample and the pH remained high through the run. Probably the pH was too
high for efficient ozonation. The sample was first stripped with oxygen
for an hour to remove volatiles and then ozonated with very poor results
as seen in TABLE 11.
The rather uniformly poor results obtained with ozone was due to an un-
fortunate choice of wastes. Glycol turned out to be unusually refractory
and nearly all the wastes were too high in TOC for effective ozonation.
It has been demonstrated that many chemicals can be efficiently destroyed
in dilute solution. Ozone has promise as a tertiary treatment to make
either high quality water or to destroy small quantities of some refrac-
tory compounds.
TABLE 11. OZONATION OF CHLORINATED WASTE (SAMPLE Oil A)
02 stripping - 50°C
Time (min) TOC, ppm pH
0 652.6 11.6
15 580.9 11.7
30 553.3 11.8
45 551.9 11.8
60 547.8 11.9
0- reaction - 50° C
Time (min) TOC,ppm pH
0 531.2 11.4
15 532.3 (?) 11.3
30 530.7 11.3
45 529.6 11.2
60 528.9 1.1.2
75 528.0 11.2
90 527.1 11.2
120 526.0 11.2
56
-------
SECTION 11
CHEMICAL ANALYSIS
The concentration of ozone in all gas streams was measured iodo-metrical-
ly. A metered gas stream was bubbled through a porous disperser tube
into 500 milliliters of 4% potassium iodide solution. The iodide solu-
tion was acidified and titrated with 0.05N sodium thiosulfate with starch
as the indicator. The end point was detected by a color change, the
color going from blue-black to clear.
The total organic carbon content of the waste water was determined with
a Beckman Model 915 Total Organic Carbon Analyzer. The analyzer does
rapid analyses of microsamples of waste water solutions for determining
the total organic carbon content. The analysis actually consisted of
two analyses which were performed on an identical sample. The total
organic carbon content was the difference between the two obtained
values for the analyses. The first analysis was the total carbon which
consisted of organic carbon plus carbon in carbonates. The second
analysis was the inorganic carbon which consisted of carbon in carbon-
ates. Both analyses were based on the conversion of sample carbon into
carbon dioxide for measurement by a non-dispersive infrared analyzer.
The individual chlorinated hydrocarbons content were determined with a
Varian Aerograph Hy-Fi Model 600-D gas chromatography. The 5' x 1/8"
stainless steel analytical column which was used with a hydrogen flame
ionization detector was packed with 20% Silicone DC 550 on Chromosorb P.
The injection temperature was 105°C while the oven temperature was from
100° to 150°C. Helium was used as the carrier gas in the gas chroma-
tography. A Sol tec recorder with a disc integrator was used to measure
the amount of chlorinated hydrocarbon.
Many of the wastes contained hydrochloric acid which damaged the chromat-
ography. For this reason TOC Analysis was relied on where possible. We
discovered, however, that the ratio of TOC/GC was not theoretical. Fig.
12 shows the deviation of theoretical and actual TOC values for EDC in
distilled water. Many of the industrial samples showed great deviations
even from this graph. For instance in sample 011A, extraction of over
3000 ppm of EDC only reduced the TOC by about 250.
57
-------
1800
1600 -
200 -
1000 2000 3000 4000 5000 6000 7000 8000
EDC (PPM)
Figure 12. Theoretical vs actual TOG values for EDC.
58
-------
SECTION 12
REFERENCES
1. Treybal, R. E., Liquid Extraction., New York, McGraw-Hill Book Co.,
1963.
2. Hanson, C., Solvent Extraction, Chemical Engineering, 75:76-98,
Aug. 26, 1968.
3. Kemp, H. S. (Chairman) International Symposium on Solvent Extrac-
tion, Chemical Engineering Progr. 62:49-104, September 1966.
4. Mayhue, L. F., Solvent Extraction Status Report Environmental Pro-
tection Agency, Ada, Oklahoma; Publication Number EDA - R2-72-073.
5. Hewes, C. G. and R. R. Davison, Kinetics of Ozone Decomposition and
Reaction with Organics in Water. A.I.Ch.E. Jr. 17:141-147, January
1971
6. Hewes, C. G. and R. R. Davison, Renovation of Waste Water by
Ozonation. A.I.Ch.E. Symp. Ser. No. 129, Water-1972. 69:71-80, 1973.
7. Treyball, R. E., Liquid Extraction Review., Ind. & Eng. Chem. 51:
378-388, March, 1969.
8. Beychock, M. R. Aqueous Wastes, New York, John Wiley and Sons, 1967,
p. 97.
9. Hildebrand, J. H., J. M. Prausnitz, and R. L. Scott, Regular and
Related Solutions. New York, Van Nostrand Reinhold Co- 1970.
59
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-77-148
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Extraction or Destruction of Chemical Pollutants
from Aqueous Waste Streams
5. REPORT DATE
July 1977 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
R. R. Davison
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Texas A&M University
College Station, Texas 77843
10. PROGRAM ELEMENT NO.
1BB610
11. CONTRACT/GRANT NO.
R800947
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research Lab.- Ada, OK
Office of Research and Development
U.S. Environmental Protection Agency
Ada, Oklahoma 74820
13. TYPE OF REPORT AND PERIOD COVERED
Final Rpt. 3/72 - 1/75
14. SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The use of solvent extraction and ozonation to treat various industrial wastewaters
was studied. Most were light chlorinated hydrocarbon, solvent wastes and were
principally extracted with a high molecular weight paraffin petroleum fraction.
Distribution data on related pure chlorinated compounds were also obtained. The
economics of solvent extraction versus steam stripping was examined. Though
chlorinated solvents can be effectively removed by extraction, stripping appears
to be more economical.
A toluene diamine wastewaterwas found treatable with benzene.
In all these wastes there are unextractable fractions.
Attempts were made to treat glycol, toluene diamine and light chlorinated hydro-
carbon wastewaters with ozone, but results were not satisfactory.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Chlorinated Hydrocarbons, Pilot Plants,
Solvent Extraction, Ozonation
Chemical wastes, Refracto
organic compounds, Waste-
Water treatment, Steam
stripping, Propylene
glycol, Toluene Diamine
13B
8. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (ThisReport)
unclassified
21. NO. OF PAGES
70
20. SECURITY CLASS (Thispage)
unclassified
22. PRICE
EPA Form 2220-1 (9-73)
60
U.S. GOVERNMENT PRINTING OFFICE: 1977-757-056/6502 Region No. 5-11
------- |