-------
are very effectively removed (>98%) by reactive foams containing
copper (II) salts or highly alkaline (pH>10.7) foams. The use of
foams to remove formaldehyde from air streams also looks very promising."
FOAM SCRUBBING OF PARTICULATES AND AEROSOLS
Table 2 presents an overview of studies that have been done on collection
of particulates and aerosols into foam.
Researchers have studied absorption of particles of various sizes into a
range of bubble sizes. In some of the studies, different surfactants were
compared, and wettable vs. nonwettable particulates were evaluated.
Damle (19) studied absorption of resolidified wax droplets, ranging from
0.1 to 5.0 microns in size. Some runs were made with hydrophillic methylene
blue for comparison.
Three surfactants were used, namely, hexadecyltrimethyl ammonium bromide
(cationic), sodium oleate (anionic), and Triton (nonionic). A range of bubble
sizes were studied, from 0.8 mm to 5.3 mm, by varying the size of the wire mesh
distributor.
Results show that collection efficiency increases for longer residence
times and for smaller bubble sizes. For a bubble size of 0.8 mm, the particle
removal was in the range of 80% to nearly 100% for a residence time of 90 sec.
Collection efficiency was found to be at a minimum in the 0.2 to 0.4-micron
particle-diameter range. This was found to occur due to a change in the
dominant collection mechanism from diffusion for small particles to
sedimentation for larger particles.
The two ionic surfactants were found to perform somewhat better than the
nonionic type. Finally, the hydrophillic methylene blue aerosol was found to
be collected more readily (up to 20% to 30% higher efficiency) than the wax
particulate.
VAPOR SUPPRESSION FROM LIQUID POOLS
A number of studies have been made of a related area of foam use, namely,
application for suppression of vapors from liquid pools of spilled chemicals.
In that application, the main purpose of the foam is to prevent or minimize
evaporation from the liquid. However, vapors that do escape are scrubbed to
some extent as they rise through the foam blanket. Vapors from cryogenic gases
(such as NH3) are also warmed by travel through foam, improving dispersion
(l).
The scrubbing that occurs in a foam blanket over liquid is different from
scrubbing, in which the contaminated air is used to blow the bubbles In a
foam blanket, the air within the bubbles is initially clean. Scrubbing occurs
as pockets of vapor make their way upward, either forming new bubbles,
channeling, or diffusing through the existing foam.
-107-
-------
o
u.
O
Z
1
i
2
S
1
<<
UJ
1—
1
a.
g
I
HI
g
£=
S
§
«
•
CM
Ul
w
i — *
i — '
I
o „
_rf
O) E 4-» ••-
O 3 4-* E
-* *C t- **
8 lag
= 5 0 3=
O
*~ C
X in
1 £
O *
ofl CO
^ S
£ 1
m >-
^
•S!
o
« 1 I
•»- "O C
CA t- tt-
o u c
C3 D O
i i
*
O ft?
o o
2 2
o
•? «•
CM Z
z*
= z
y-%
*«x
«—
1
i s
K)
— T c$
CO «—
4-» ••
fli c
£ 1
S ^
Q. -3
1
'o 5
3: 5'
t; " "g
V -CO
CO +-> .C
SCO
cu u
§OO •— CO 0.
i= •$ Si "^ *j- ^ o
4-* o co ro ^-* H~
CO i*~ — ' C (0 CI
ooo c c "• ** J2
M- "~ "~ CD H- 4-»
&«>?(- &« frS OC-4-*
O O 3 in IA D •*-
inr-W O 0 AW—*
E
OS ' w— O£ Q£
z t z z
Ctf
o z z z
C-
o; as oi co
z z z >
S e § S E
E 3 in •* 3 E
S- CM O O CO
ro o o o o «- co
V O O O O O O
X
01
m
~~*
I
S 0) 5"
'•— C U)
CO "D — •
C£ <- 0 0
z 13 •-• a.
CM
P ^
3 3> 2 2
*^ "* ° ^ ^
So *~ 1C ?S
«- C r- r-
2 D)
'C O 'E +J >*
d) > 5 co 2
•i ~ S b .1
I- v> ca co —i
i. "8
a tj
§ CO
11
>. (D
O O)
£ 2i
0 g
**-
O CO
"OT 7>
DC 4-*
i
=
K,
D)
O
£
i
i
1
T-
1
LL.
»**-
T~
|
"io
-108-
-------
^
"8
c
•—
c
o
o
\-f
rJ
Ul
ca
fE
CO
4-1
1
at
o>
— • 01
"8 N
5 "
I
4-> «
1
O
(D
£_
D
"*
a
N
CO
"o
O
t-
01
01
4-t
o
c u
"~ .5"
5*6 O)
) C
.. o>
0) — •
*•* >» •
JO Q. O
!c t- «
•i. - §
-r| o. c
4-f
.2 8 g
Q CO 1
i 1
CO O
0 (O
rvi
a:
f S
4V "S X
••- CO O
O) O U
L. &. 0)
01 0) 4J
1— < CO
=
^
^
CO
(r-
o
o
O D)
CO O
(0 D
JC CL
t3 >s
a £
1
k
ai
u
1
0
a> —•
u o CD t.
01 *»- o g
CM C U 0> C O •
a* £ S. 'i S § i§
S fe '1 .£ § ? "
§ S i « 2 ^ ~
in -D co o v o A
•*-* o
V) '• 4J
" I "Si
B ^> ^ .S „.
• o in
S« 9« >« ^
^ ** 5
CM in in o
d x o o ^
u
is
Sra S
a: a
o
1 1
D L.
•- jQ
C
8 IN 1
•2 4J 2 ffl
§_ 41 3 —
' S' B L* X
o oi •»- o .c
4J "O tO 4*
T» . ca
CO Of +•» O 0)
<; z iu tn m
«- CM 10 ~t in
8
0
-109-
-------
Liquid pool suppression studies have established important data on the
compatibilities and interactions of various gases with different foam types.
The information from these studies is readily available in the literature (1,
3,21-26) and is therefore not reiterated here.
COMPARISON TO ALTERNATIVE TECHNIQUES
There are currently a variety of techniques available to prevent, control,
or mitigate hazardous accidental releases of air toxics (EHSs). Preventive
techniques include, for example, alarms and interlocks that allow corrective
action to be taken before a release occurs. Control measures include
scrubbers, flares, and other devices intended to capture and destroy escaping
contaminants before a release can occur. Finally, if a release does occur,
mitigative measures such as water curtains can be used to reduce the off-site
impact.
Foam scrubbing falls into the categories of control and mitigation. A foam
generator can be used in the same manner as a scrubber to protect against
releases from process equipment or buildings; or it can be used to draw in a
gas cloud from, for example, a leaking pipeline or vessel. (Strategic
positioning of "at ready" units must be considered.)
In the first case, the advantages of the foam generator are:
o Equipment is relatively small and can be on standby until needed.
o Large volumes of gas can be controlled with relatively small amounts of
liquid.
o Residence time can be provided within the foam mass, rather than requiring
a large, permanent vessel.
o The foam generator can be readily sized for a large emergency flow (i.e.,
up to 30,000 cfm per unit).
Techniques currently used for mitigation of vapor clouds include water
sprays and curtains, steam curtains, and air curtains. Where liquid is also
present, techniques include dilution, neutralization, temporary covering,
freezing, and ignition (27, 28).
The water, steam, and air curtain techniques are primarily used to disperse
the released gases and have been shown to reduce concentrations by a factor of
from 1.5 to 4, depending on wind speed (27). Some scrubbing is also provided
by water curtains, but only when very large water volumes are used. In tests
with hydroflouric acid, up to 50% scrubbing was found (R. Hiltz, personal
communication, 1987). The main disadvantages of these techniques are: 1) their
relatively low effectiveness, and 2) the high levels of plant utilities
required.
Based on the foam scrubbing efficiencies measured by previous researchers,
it is felt that this technique can provide better mitigation, in many cases,
-110-
-------
than the currently available alternative techniques.
EXPERIMENTAL WORK
4. I?.dut£7 a set of Preliminary tests has been carried out to qualitatively
CACC^jsh the effectiveness of foam scrubbing. In these tests, smoke from
SAFE-VUE colored smoke candlesf-a dispersion of airborne dye particles-was
ducted into an MSA Fast Response1" high expansion foam generator using MSA
Fast Response1" high stability foam concentrate at the manufacturer's
recommended concentration. The Fast Response1" generator produces
approximately 225:1 expansion foam at a rate of approximately 35 cfm
Observation of the foam leaving the nozzle showed that essentially all of the
smoke was incorporated into the foam; no escaping smoke was seen.
The foam persisted for several minutes or more when piled on the ground
No smoke was seen to escape from the collapsing (draining) foam pile, and the
foam took on the color of the smoke. Closer observation showed the color to be
present in the liquid lamellae of the foam bubbles.
These preliminary results are very encouraging. More detailed testing is
planned, on both laboratory and demonstration scales, to quantify the
effectiveness of foam scrubbing on hazardous materials.
CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK
This study indicates that utilization of foam systems is a viable and
potentially cost-effective method for the capture and neutralization of larqe
volumes of released toxic vapors. There is certainly a need for the emergency
control capability that could be achieved using foam technology.
The results of previous work with foam are encouraging, showing generally
good absorption efficiency for both gases and particulates. There is, however
a need for clearly defined and described experimental work to bring the foam
concept to an industrial reality.
Therefore, we believe that foam technology could be a valuable addition to
the range of available techniques.
It is strongly recommended that a program of laboratory testing followed by
demonstration at pilot plant scale be carried out. It is expected that this
testing program will allow foam technology to develop rapidly from a concept
into a viable control option. •• . • H
-111-
-------
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Gross, S.S., and Hiltz, R.H. Evaluation of foams for mitigating air
pollution from hazardous spills. EPA-600/2-82-029. U.S. Environmental
Protection Agency, 1982, 73 pp.
Ctvrtnicek. Monsanto Research Corporation. Application of foam scrubbing
to fine particle control, Phase I and Phase II. EPA-600/2-76-125. U.S.
Environmental Protection Agency. 1976. 213 pp.
Hiltz, R.H. Mitigation of the vapor hazard from silicon tetrachloride
using water-based foams. J. Hazardous Materials. 5, 1982.
Jackson, J. Gas cleaning by the foam method. British Chemical
Engineering. 8:5, 1963.
Gillberg, G. Foaming technology - An overview. In: Proceedings of the
Eighth Technical Symposium: Nonwovens - Innovative Fabrics for the
Future. 1980.
Kouloheris, A.P. Foam: friend and foe. Chemical Engineering, 1987.
MSA Research Corporation. Information Package on Fire Protection Systems
and Components. Evans City, Pennsylvania, 1986.
Federal Register, April 22, 1987, pp 13403-13410
Ross, S. Foams. In: Kirk-Othmer Encyclopedia of Chemical Technology.
Vol.11, 1983. pp. 127-145.
National Fire Prevention Assoc. Standard for Medium and High Expansion
Foams. Philadelphia, Pennsylvania, 1983.
Shah, P.S. and Mahalingam, R. Mass transfer with chemical reaction in
liquid foam reactors. AIChE Journal. 30:6, 1984.
Nguyen Ly, L.A., et. al. Diffusion of gases through surfactant films:
Interfacial resistance to mass transfer. AIChE Journal. 25:6, 1979, pp.
1015-1024.
Ross, S. Bubbles and foam. Industrial and Engineering Chemistry. 61:10,
1969.
Kaldor, T.G. and Phillips, C.R. Aerosol scrubbing by foam. Ind. Eng.
Chem. Process Des. Dev. 15:1, 1976. pp. 199-206.
Metzner, A.B., et. al. Mass transfer in foams. Industrial and
Engineering Chemistry. 48:11, 1956. pp. 2040-2045.
Biswas, J. and Kumar, R. Mass transfer with chemical reaction in a foam
bed contactor. Chemical Engineering Science. 36:9, 1981.
-112-
-------
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Brander, S.M., et. al. Reactive foams for air purification.
Sci. Techno!. 18:4, 1984.
Environ.
Ohkawa, A., et. al. Bubble size, interfacial area, and volumetric liquid
- phase mass transfer coefficient in downflow bubble columns with gas
on^31?™^ * a li(Juid Jet- Journal of Chemical Engineering of Japan.
20: 1 ,
Darnle, A.S. Gas submicron particulate separation in a flowing liquid foam
1980 1X* PhD D1ssertati°n- Wash. State Univ. Pullman, Washington,
Ramsey, G.H.
particulate.
33 pp.
Evaluation of foam scrubbing as a method for collecting fine
EPA-600/2-77-197, U.S. Environmental Protection Agency, 1977
Greer, J.S., Gross, S.S., Hiltz, R.H., et. al. Modification of spill
factors affecting air pollution. Vol.2. The Control of the Vapor Hazard
from Spills of Liquid Rocket Fuels. EPA-600/2-81-215. U.S. Environmental
Protection Agency, Cincinnati, Ohio, 1981.
Takimoto, H.H., Lewis, S., and Hiltz, R. Vapor suppressant foam systems
for propel 1 ant spills. Jn: Proceedings of the 1984 Hazardous Material
Spills Conference. Nashville, Tennessee, 1984.~
Evans, M.A., and Carroll, H.A. Handbook for Using Foams to Control Vannr*
Aenc 1985°U2 Spi11s' Contract 68-03-3113, U.S. Environmental Protection
Hiltz, R.H. Control of the vapor hazard from reactive volatile hazardous
materials by foam. In: Proceedings of First Technical Seminar nn
Chemical Spills. 1983. pp. 63-74. '.
Norman, E.G. and Dowel!, H.A., National Foam Systems, Inc. Using aqueous
foams to lessen vaporization from hazardous chemical spills. Chemical
Engineering Progress, Technical Manual on "Loss Prevention," Vol. 13
1980, 7 pp. '
Norman, E.C. Vapor mitigation by the use of foam: Case history and
large-scale outdoor tests. In: Proceedings of the Hazardous Materials
Spills Conference. 1982. '
Brown, L.E., et. al. Hazard control methods for high volatility
chemicals. In: Proceedings of the International Symposium on Preventing
Ma.ior Chemical Accidents. Center for Chemical Process Safety of the
American Institute of Chemical Engineers. 1987.
Prugh, 1987. Post-release mitigation design for mitigation of releases
i"-: Proceedings of the International Symposium on Preventing Maior
Chemical Accidents. Center for Chemical Process Safety of the American
Institute of Chemical Engineers. 1987.
-113-
-------
TR ANSPORT OF INORGANIC COMPOUNDS
THROUGH COMPACTED CLAY SOIL
by: David E.Daniel
University of Texas ..
Austin, Texas 78712
Charles D. Shackelford
Colorado State University
Ft. Collins, Colorado 80523
Wing-Ping Liao
University of Texas
Austin, Texas 78712
ABSTRACT
The simplest approach to prediction of the rate of transport of solutes through compacted clay soil is to
compute the seepage velocity of the transporting liquid based on the hydraulic conductivity of the soil, the
SdSc gradiS, Ind the effective porosity of the soil. However, such analyses fail to account for mechanical
dispersion and molecular diffusion. With compacted clay soils having low hydraulic conductivity, molecular
diffusion is an important mechanism of chemical transport at the low hydraulic gradients that typify field values.
The purpose of this research was to study experimentally the migration of inorganic solutes through
compacted clay soil at realistic hydraulic gradients. A major portion of the research focused on molecular
diffusion It was found that the diffusion coefficient for compacted clay soil was insensitive to the
geomechanical factors that were studied (such as molding water content and method of compaction) but was very
sensitive to geochemical factors. Typical data are used to demonstrate that: (1) clay liners release contaminants
primarily by molecular diffusion if the hydraulic conductivity is on the order of 10* to 10-9 cm/s or less, and
(2) breakthrough of contaminants can occur much sooner than predicted by models that ignore dillusion.
-114-
-------
INTRODUCTION
Compacted soil liners are a vital component of nearly all solid waste disposal facilities and surface
impoundments for hazardous waste. Soil liners are intended to slow the release of contaminants and to retard the
migration of many of the dissolved contaminants that enter the liner (1). Predictions of the performance of soil
liners center on the volume, Q, of liquid that will leak through a liner with an area A and a hydraulic
conductivity k over a period of time t, when the hydraulic gradient driving flow is i. Based on Darcy's law:
Q = ki At
(1)
More sophisticated analyses of the release of contaminants through a soil liner are needed to account for
the movement of solutes. In general, two mechanisms are responsible for the migration of solutes in leachate
toough a soil hner: (A) adjective transport, which is the transport of solutes by flowing water in the soil, and
(B) molecular diffusion, which is the transport of solutes by thermal-kinetic energy gradients or gradients in the
concentration of a solute (2) For soil liners with low hydraulic
significant mode of transport of solutes through soil liners (3-7).
,, Prediction for kachate migrating through a clay liner is the temporal variation of the
™ T7(C) f °ne ^ TT S?1UutCS in ^ material ikying me soil liner (Fig. 1). Graphs such as the
one shown in Fig. 1 are called "breakthrough curves." Breakthrough curves can be measured in the laboratory by
permeating soil with liquid having a known and constant concentration (co) of a particular solute and measuring
the concentration (c) of the solute in the effluent reservoir over time. Breakthrough curves are often plotted in a
slightly different form; the abscissa is typically pore volumes of flow (P), which is the cumulative quantity of
inflow divided by the volume of the void space in the soil, and the ordinate is the relative concentration c/c0
Several such curves are shown in Fig. 2. In Fig. 2, "retardation" refers to the slowing of solute movement as a
result of ion exchange, precipitation, or other geochemical reactions.
,npff t- In.Soi1' P°res offl various sizes and degrees of connectivity exist. Some pores may be relatively
ineffective in conducting flow whereas other pores (especially large pores, cracks, or fissures) may conduct a
relatively large percentage of the flow. The "effective porosity" is a measure of the volume occupied by the
E5l?SS*£ rSff°f ?C fl°W dividCC\ by- the t0tal V0lume °f the soil <& If a11 the Pores conductflow
equally well, then the effective porosity (ne) will equal the total porosity (n). In compacted clay soil there is
evidence that n^n can be much less than one (8), indicating that a small percentage of the pores (the largest ones)
conduct most of the flow. The effective porosity may be determined from a breakthrough curve; nP is the pore
volumes of flow at c/c0 = 0.5 (9). Except when large advective velocities mask the effects of diffusion such
determinations of ne are not correct because they fail to account for the influence of molecular diffusion upon the
thr°Ugh aT1SOirr0lUnin- ?iffusion ™y cause breakthrough at less than one pore volume of
are equally efficient conductors.
MOLECULAR DIFFUSION
BACKGROUND
The Process of diffusion is usually assumed to occur in accord with Pick's first and second laws For
free solutions (i.e., no porous matrix), Pick's first law states that one-dimensional diffusion occurs as follows:
Jd = -I
.X ' <2>
where Jd is the diffusional mass flux (mass transported per unit area perpendicular to the direction of transport
per unit time), D0 is the free-solution diffusion coefficient, c is the concentration of a solute and x is the
direction in which the diffusion is occurring.
-115-
-------
• Concentration
o
g
'to
0)
o
o
o
Influent Liquid •
Effluent Liquid
Flow
Concentration c
Time
Figure 1. Breakthrough curve for column test.
-------
Solutes will not diffuse as quickly in soil as they will in free solutions. For soil, Eq. 2 becomes
Jd _ -
or
(3)
J = -
d
9x
(4)
where T is a dimensionless tortuosity factor, 6 is the volumetric moisture content of the soil, and D* is the
effective diffusion coefficient. It should be noted that some researchers include 8 in the tortuosity factor The
tortuosity factor accounts for the increased distance of flow and the more tortuous pathways experienced bv
S5S ^ r°Ugh SOilV ? refty' thCre my be °ther effects associated with S Ttortuos^ factor! e.g"
the variation in the viscosity of the solvent within the pore space and the negative adsorption (exclusion) of ions
law applk?'8
dCSCribeS steadv-state fl"x of solutes. For unsteady (transient) transport, Pick's second
(5)
Equation 5 is integrated for appropriate initial and boundary conditions to obtain a description of the solute
concentration changes with respect to time and space. Complementary error functions facilitate the integration
EXAMPLE
An example of the application of Pick's second law is as follows. A 1-m-thick clay liner retains
leachate containing a particular solute at a concentration of 10,000 mg/L. The underlying iounTwatS"
completely free of this solute. The effective diffusion coefficient is assumed to be 2 x 10-10 m2/s The
resulting distribution of solute in and at the bottom of the liner as a function of time is illustrated in Fig 3
=* 5 <">• ^ich assumes the leactafe
— = erfc
(6)
5 ? 40 fnH sn comPlementary error ^tion. The concentrations of solute reaching the base of the liner at 10
™L f h ^T ^ aPProximately 50, 460, 160, and 3200 mg/L, respectively. As illustrated in this
example of chemical transport through a liner by diffusion alone, the diffusion of chemicals through fine-grained
materials can be an important transport mechanism even over relatively short (10 - 30 years) periods.
EXPERIMENTAL PROGRAM
?hown. ?chematically in Fig. 4 was used to measure the diffusion coefficient of compacted
•• ^aolimte> whlch 1S a commercially-produced material, and Lufkin clay, which is a naturally-
occurring soil rich in smecute, were used in the experiments. Data on the soils are given in Table 1.
Test specimens were compacted, soaked in 0.01 N CaSO4 until advective flow ceased, and then exposed
on one side to a simulated waste liquid containing three anions (C1-, Br, and I-) and three cations (K+ Cd2+ and
-117-
-------
m
I
1
0.0
0.0
(UNER-LEACHATE INTERFACE)
01 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Relative Concentration, C/Co
Figure 3. Concentration profiles for 1-m-thick clay liner
releasing solute solely by molecular diffusion.
TOBURET
OPTIONAL T-SECTION
O-RING
4-
4 RODS
@90'
RESERVOIR
SAMPLE
LINE
\
OPTIONAL SOAK-
ING LINE
Figure 4. Schematic drawing of diffusion cell.
-118-
-------
TABLE 1. PROPERTIES OF SOIL
Property
Natural Water Content (g/g)
Dominant Clay Mineral
Optimum Moisture Content (g/g)
Max. Dry Unit Weight
Liquid Limit (g/g)
Plasticity Index (g/g)
Exchangeable Cations (meq/lOOg)
Na+
K+
Ca2+
Method of
Measurement
ASTMD2216
X-Ray Diffraction
ASTM D698
ASTM D698
ASTMD4318
ASTMD4318
Ref. 13
Value of Property
Kaolinite Lnflcin Clav
Dry
Kaolinite
32%
83 pcf
54%
23%
3.8
0.8
1.0
23%
Smectite
20%
102 pcf
56%
42%
6.4
2.7
19.1
RESULTS AND DISCUSSION
The results of the experiments on 0.04 N synthetic leachate are summarized in Table 2 The
experiments yielded effective diffusion coefficients generally in the range of 2 - 4 x lO'™ m2/s f
S~=^^^^^
JnmnP t ^ hf very.sensit.lve to subtle variations in geochemical factors (such as leachate sttengt?, mt S
common? W6re mSenSmVe t0 Se°Ph^al Paramet^ 0»«* as molding water content anTmefhod of
ADVECnVE-DIEFUSIVE TRANSPORT
THEORETICAL BACKGROUND
The advective transport of conservative (non-attenuated) solutes through soil is assumed m nrrnr in
accord with Darcy's law. The advective flux (Ja) may be computed with the fiZS/Sqi
(7)
-119-
-------
TABLE 2. RESULTS OF DIFFUSION EXPERIMENTS ON 0.04 N SYNTHETIC LEACHATE
Soil Compaction
Kaolinite Std. Proctor
11 "
.
„
„
" Kneading
" Static
Lufkin Clav Std. Proctor
Molding
Water
Content
24
27
30
33
36
33
33
22
Number
of
Tests
3
3
3
3
3
2
2
2
Average Effective Diffusion Coefficient, D*xl010 m2/s
ci-
4.1
3.9
3.8
2.8
3.6
3.8
3.3
1.5
Br
2.5
3.0
3.3
2.5
2.2
2.2
2.2
0.7
I-
3.0
1.1
3.7
3.6
2.2
3.7
2.1
0.5
K+
1.5
0.9
1.2
1.2
1.6
1.3
0.9
0.01
Cd2+
1.7
1.3
1.3
1.6
1.1
0.7
0.3
3.0
Zn2+
2.1
1.9
1.8
2.0
2.6
0.8
0.3
1.7
where h is the total hydraulic head and the other terms are as defined earlier. Many solutes in leachate will be
attenuated by soil eg., by ion exchange, sorption by organic carbon, precipitation, oxidation/reduction
SffaSiS^Ito retardation coefficient^, for a solute may be defined as the number of pore> volumes
of flow Sat is needed to achieve c/co = 0.5, divided by the number of pore volumes needed to achieve c/c0 -0.5
£^Conservative fracer, when in both instances the flow is purely advective. Thus, if n - ne, ^ conservaUve
tracer will break through a soil column (c/c0 = 0.5) at 1 pore volume of flow whereas a solute with R = 2 will
break through at 2 pore volumes of flow. The idea is illustrated in Fig. 5.
The retardation coefficient may be computed from the following relationship:
R= i + fii
(8)
where pd is the dry bulk density of the soil, 9 is the volumetric water content, and Kp is the partition coefficient
which relates mass of solute sorbed per unit mass of soil to the concentration of solute in solution at
equilibrium. Batch adsorption tests are often used to determine Kp (14).
EXPERIMENTAL PROGRAM
The same soils and ions mentioned earlier in the discussion of tests to determine effective diffusion
fficient are being utilized in column experiments with combined advective/diffusive transport. Hydraulic
Se S of 1 to 20 are being used for kaolinite, and gradients of 20 and 50 are being used for tests
nday Tome^Specimens have been soaked with standard solutions prior tc.permeation, and others
n pe7mea°ed without presoaking. Tests are still in progress. In general, significant diffusive flow has
been seen at the lowest gradients while flow seems to be primarily advective at the higher gradients.
-120-
-------
R =
= 2
rct
1.0
0.5
Conservative
Tracer (ct)
Solute (s)
0 1 2
Pore Volumes of Flow (P)
Figure 5. Definition of retardation coefficient (R) for a solute.
RELATIVE IMPORTANCE OF DIFFUSION
The data presented earlier, plus information in the literature (12), indicate that the effective diffusion
coefficient for anions diffusing through clay soils covers a fairly narrow range; a value of 2 x lO'10 m2/s may be
taken as typical. For landfills with leachate collection systems, the hydraulic gradient is likely to be slightly
greater than unity (ignoring suction pressures); a value of 1.3 may be taken as typical. A typical thickness for a
clay liner is 3 feet. The breakthrough curves shown in Fig. 6 have been computed for these assumed conditions
using the Ogata (15) solution. Without diffusion, all the curves would be the same, but with diffusion the pore
volume at which c/c0 = 0.5 is a function of the hydraulic conductivity of the soil. For these conditions,
diffusion is relatively unimportant for hydraulic conductivities greater than 1 x lO'7 cm/s but becomes important
for conductivities of 1 x 10"8 cm/s or less.
Another way to view the influence of diffusion is to plot the breakthrough time as a function of the
hydraulic conductivity of the soil. This has been done for the conditions outlined in the previous paragraph- the
results are shown in Fig. 7. At very low hydraulic conductivities, flow is purely diffusive, and at very high
values, flow is entirely advective.
The relative importance of advection and diffusion as mechanisms of contaminant transport through clay
liners may also be evaluated in terms of mass fluxes. The long-term, steady flux of a solute due to diffusion (Jd)
is given by Eq. 4. Typical values are assumed for a clay liner: D* = 2 x lO'10 m2/s, 0 = 0.5, concentration of
solute in leachate = c, concentration of solute in soil beneath the liner is zero, and thickness of liner = 1 meter
ror these conditions:
xlO~10m/s)
(9)
For the same liner, the advective flux for unit hydraulic gradient is:
Ja = kc
(10)
The total flux (J) is
-121-
-------
0.5 1.0 1.5
PORE VOLUMES OF FLOW
Figure 6. Breakthrough curves for a 3-ft-thick clay liner.
2.0
140
10
-10
10
10 ' 10 w 10 ' 10
Hydraulic Conductivity (cm/sec)
Figure 7. Breakthrough time versus hydraulic conductivity for a 3-ft-thick clay liner
-122-
-------
-(
1 x 10"10 m/s + k)
or:
- = 1 x 10"10 m/s + k
c
(11)
(12)
The mass flux per unit concentration (J/c) is plotted as a function of the hydraulic conductivity of the liner in
Fig. 8 For hydraulic conductivities less than lO'8 to 10'9 cm/s, the mass flux is essentially constant and is
controlled by Jd. The percent of total flux caused by diffusion is:
x 100% =
1 x 10'10 m/s
1 x 10"10 m/s + k
(13)
The percentage of flux caused by diffusion is plotted as a function of the hydraulic conductivity of the liner in
Fig. 9. For hydraulic conductivities less than 10'7 cm/s, advection is responsible for nearly all the mass flux
At a hydraulic conductivity of lO'8 cm/s, advection and diffusion each account for half of the mass flux At
hydraulic conductivities less than 10'9 cm/s, diffusion is the source for nearly all the mass flux.
CONCLUSIONS
Solutes are transported through compacted clay soil both by advection and diffusion. Experiments have
shown that the effective diffusion coefficient for anions diffusing through compacted clay soils is about 2 x 10-1°
m /s. Cations tend to diffuse more slowly due to ion exchange and other reactions. The method of compaction
and molding water content have little influence upon the effective diffusion coefficient of compacted clay soil
However, subde variations in geochemical factors, e.g., concentration of a solute in leachate or cation exchange
capacity of the soil, can cause significant changes in the rate of diffusive transport through clay soils.
A « liner thicknesses (3 ft or approximately 1 m) and hydraulic gradients (slightly greater
flian 1) advection dominates the flow of solutes through compacted clay liners having hydraulic conductivities of
1 x 10- cm/s or greater. Diffusion becomes the dominant mechanism of transport at hydraulic conductivities on
the order of 10-« to 10^ cm/s or less. Solutes can travel through a 3-ft-thick clay liner of zero hydraulic
conductivity purely by diffusion in periods of 10 - 30 years or less. To ignore molecular diffusion in clay liners
significantly less than 10"7 cm/s is to ignore m important' if not
REFERENCES
1.
2.
3.
4.
Daniel, D. E. Earthen Liners for Land Disposal Facilities. In: Geotechnical Practice for Waste Disposal
87. American Society of Civil Engineers, New York, 1987. pp. 21-39.
Freeze, R. A., and Cherry, J. A. Groundwater. Prentice-Hall, Englewood Cliffs, New Jersey, 1979
604 p.
Quigley, R. M., Yanful, E. K., and Fernandez, F. Ion Transfer by Diffusion through Clayey Barriers In-
Geotechnical Practice for Waste Disposal '87. American Society of Civil Engineers, New York, 198?!
pp. 137-158.
Goodall, D. E., and Quigley, R. M., Pollutant Migration from Two Sanitary Landfill Sites Near Sarnia
Ontario. Canadian Geotechnical Journal. Vol. 14, 1977. pp. 223-236.
-123-
-------
10
.-11
,-10
-8
-7
10 - 10 ~ 10 w 10 ' 10
Hydraulic Conductivity of Liner, cm/s
Figure 8. Mass flux versus hydraulic conductivity for a 1-m-thick liner
having an effective diffusion coefficient of 2 x 10'10 m2/s.
o
'en
i
Q
T3
CD
CO
1
O
x
ul
o
I
CD
CL
100
75
50
25
\
\
,-11
,-10
-9
,-8
-7
-6
10 •' 10 1W 10 " 10 w 10 ' 10
Hydraulic Conductivity of Liner (cm/s)
Figure 9. Percent of mass flux caused by diffusion versus hydraulic
conductivity for a 1-m-thick liner.
-124-
-------
5.
6.
7.
8.
9.
Gillham, R. W., Robin, M. J. L., Dytynyshyn, D. J., and Johnston, H. M. Diffusion of Nonreactive and
Reactive Solutes through Fine-Grained Barrier Materials. Canadian Geotechnical Journal. Vol. 21 1984
pp. 541-550.
Quigley, R. M., and Rowe, R. K. Leachate Migration through Clay below a Domestic Waste Landfill
Sarnia, Ontario, Canada. ASTM STP 933,1986. pp. 93-103.
Rowe, R. K. Pollutant Transport through Barriers. In: Geotechnical Practice for Waste Disposal '87.
American Society of Civil Engineers, New York, 1987. pp. 159-181.
Horton, R., Thompson, M. L., and McBride, J. F. Estimating Transit Times of Noninteracting
Pollutants through Compacted Soil Materials. In: Proceedings of the Eleventh Annual Research
Symposium on Land Disposal of Hazardous Waste, U.S. EPA, Cincinnati, Ohio, EPA/600/9-85/013
April, 1985. pp. 275-282.
Bowders, J. J., Daniel, D. E., Broderick, G. P., and Liljestrand, H. M. Methods for Testing the
Compatibility of Clay Liners with Landfill Leachate. In: Hazardous and Industrial Solid Waste Testing,
Fourth Symposium, STP 886, American Society for Testing and Materials, Philadelphia, Pennsylvania
1985. pp. 223-250.
10. Olsen, S. R., Kemper, W. D., and Van Schaik, J. C.. Self-Diffusion Coefficients of Phosphorous in Soil
Measured by Transient and Steady-State Methods. Proc. Soil Sci. Soc. America. 29:154. 1965.
11. Crank, J. The Mathematics of Diffusion. 2nd Edition. Clarendon Press, Oxford, England, 1974,414 p.
12. Shackelford, C D. Diffusion of Inorganic Chemical Wastes in Compacted Clay. Ph.D. Dissertation, The
University of Texas, Austin, Texas, 1987,449 p.
13. Rhoades, J. D. Soluble Salts. In: A. L. Page, R. H. Miller, and D. R. Keeney (eds.), Methods of Soil
Analysis, Part 2, Second Edition. American Society of Agronomy, Madison, Wisconsin, 1982. pp. 167-
14. U.S. Environmental Protection Agency. Batch-Type Adsorption Procedures for Estimating Soil
Attenuation of Chemicals. EPA/530-SW-006. Draft Technical Resource Document for Public Comment
Cincinnati, Ohio. (NTIS PB87-146-155).
15. Ogata, A. Theory of Dispersion in Granular Medium. U.S. Geological Survey Professional Paper 411-1
1970. *
-125-
-------
U. S. EPA RESEARCH IN SOLIDIFICATION/STABILIZATION OF WASTE MATERIAL
by:
Carlton C. Wiles and Hinton K. Howard
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Solidification and stabilization (S/S) technologies for the management
of hazardous wastes have a high level of importance, particularly with
regard to treating contaminated soils at Superfund sites and with regard to
treating residuals from other treatment technologies. This paper presents
a summary of recently completed, ongoing, and planned research by the U.S.
EPA to provide the best possible information regarding the utilization of
solidification/stabilization processes. Areas include studies of existing
protocols and processes, studies of basic chemical and physical mechanisms,
and applicability to a variety of wastes.
INTRODUCTION
The Resource Conservation and Recovery Act (RCRA) and the more recent
Hazardous and Solid Waste. Amendments have provisions aimed at reducing the
use of land disposal for the management of hazardous wastes. It is clear
that the most satisfactory solutions to hazardous waste problems are those
which either destroy or detoxify the wastes. However, for the near future,
there exist wastes which can not be destroyed or detoxified, but only re-
duced in volume or contained. These wastes and the residues from their
treatment will continue to be managed by land disposal until new technol-
ogies for their total destruction emerge. Solidification and stabilization
will continue to have a high level of importance for these wastes and in
the treatment of contaminated soils at Superfund sites.
The solidification/stabilization technologies are applicable to the
treatment of selected banned waste prior to landfilling. They are also
applicable to the treatment of residues from other treatment technologies.
The broad objectives are either to contain pollutants and prevent them from
entering the environment, or to transform the toxic materials into nontoxic
forms. The containment objective is addressed with processes which may
-126-
-------
produce a solid, decrease the surface area across which contaminants can be
transported, limit the mobility of contaminants exposed to leaching or
improve the handling characteristics of the waste. The detoxifying
objective is addressed with processes which involve a chemical change and
the formation of new compounds,
OBJECTIVE OF SOLIDIFICATION/STABILIZATION RESEARCH
The objective of research on solidification/ stabilization is to pro-
vide the technology which allows the complete transformation of toxic
wastes to non-toxic forms. This objective implies chemical transformation
and formation of new compounds which does not routinely occur with most of
today's available state-of-the-art S/S technologies. Some chemical con-
cepts have, however, been proposed which could improve the potential for
meeting the ideal objective. These concepts include passivation or armor-
ing reactions, elemental substitution or diadochy, chemisorption and pro-
duction of new insoluble compounds. Passivation is the chemical coating of
a substance with a rind that prevents .further chemical attack, niadochy is
a process that removes elements from the environment by substitution during
precipitation of commonly occurring compounds. Different ions have the
ability to occupy the same lattice position in a crystal structure.
Therefore, elements with similar sizes and charges can substitute for one
another in common crystal lattices. Toxic elements can be substituted in
stable crystal systems that can prevent release of the element to the
environment.
The most important technical needs, involve developing a scientific
understanding of the S/S processes. In order for S/S,technology to be
effective in managing hazardous waste, it is critical that process
candidates be properly selected, formulated, and used. There are several
factors important in the selection and subsequent performance of candidate
processes. These include waste chemical and physical characteristics;
binder physical and chemical characteristics; interactions of binder/waste,
and impurities in the matrix; product placement regime and others.
Knowledge,of these factors and the interactions which can be expected among
the -various candidate binders and waste types is critical to the successful
use of and evaluation of S/S for managing hazardous waste. This knowledge
is also critical to the successful formulation and implementation of a
regulatory strategy and technical guidance for the technology. .
Knowledge of the S/S binders and how they are expected to react under
different uses and with different wastes can provide valuable insight into
processing requirements, waste pretreatment requirements, waste-binder
interactions, and expected product performance. Governmental and other
research efforts have concentrated on cement-lime/fly ash-pozzolanic pro-
cesses. Additional efforts are required in this area. However, more basic
research needs to be conducted on other binders.
Although S/S technologies have been used for more than 20 years, there
exists little information on long-term physical durability and chemical
stability of,the S/S mass when placed in the ground. In addition, past S/S
-127-
-------
processors have not been driven by regulatory factors (i.e., waste banning,
liquids in landfills, etc.) which force current users to meet more strin-
gent performance requirements. Therefore, measuring the performance of
previous S/S waste applied to the land, although useful, may not be a com-
plete source of data, because future S/S processes will be required to meet
much more stringent performance criteria. Studies are required to deter-
mine the long-term physical-chemical stability of S/S residues when placed
on the land. Studies are also required to determine how and under what
conditions S/S residues should be placed in order to ensure environmental
protection over the long term.
Prediction and measurement of the performance of S/S processes and .
products is an area of concern, and needs to be more fully investigated.
Current efforts have begun to evaluate effectiveness of several protocols
to characterize S/S products. This work needs to be extended to provide
effective measurement techniques. More important, however, research needs
to correlate results of such tests with prediction of performance in the
field.
Generally, S/S technology is recognized effective for inorganic waste,
while organic wastes have the potential to cause problems. More informa-
tion is required on how much organic waste may be included in inorganic
waste without requiring pretreatment prior to S/S. The long term effects
of organics on S/S performance are important, however, little research has
been performed. This is particularly important for the remediation of high
volume, low level contaminated soils and debris.
SOLIDIFICATION/STABILIZATION RESEARCH PROGRAM
Currently, the U.S. EPA is evaluating S/S as a best demonstrated
available technology for treating hazardous waste. Also involved are eval-
uations for treating contaminated soil and debris, and research to develop
a more comprehensive scientific understanding of the technology.
The current U.S. EPA research (Table 1) has emphasized investigating
interferences to S/S, wastebinder interaction, waste disposition sites, and
methods for predicting performance of S/S products.
Research on interfering agents will provide data on the effects
interfering inorganics (Pb, Cd, Zn, sulfates, etc.) and organics (oil,
grease HCB, TCE, phenol) may have on commonly used pozzolanic binder
systems. The information will be useful in evaluating applications for
delisting hazardous waste and for permits to treat hazardous wastes with
S/S, particularly for those waste streams contaminated with organics. This
information and that from research on factors critical to S/S will also aid
decisions on potential waste pretreatment techniques for enhancing S/S
performance. Data from physical and chemical tests is being analyzed to
determine if a correlation exists between physical properties of the solid-
ified waste form and its ability to resist stresses when exposed to leach-
ing situations.
-128-
-------
An effort with Environment Canada emphasizing the evaluation of
several leaching tests will determine the extent of toxic constituent with-
in the binding. Protocols for examining physical properties are also being
evaluated. This research includes actual waste and synthetic sludges
solidified by vendors. Information on the performance of several different
solidified products will be compared. This research will provide important
data to compare current U.S. EPA regulatory leaching procedures with others.
being tested.
Electron scanning and x-ray defraction microscopy techniques along
with solvent extractions are being used to investigate waste/binder inter-
actions. The objective is to better understand S/S processes by identi-
fying binder reaction phases and where the waste components concentrate
their form and are bonded. Concurrent with this research, .specimens of
solidified waste from other research projects are being examined in efforts
to correlate results of physical and chemical tests with performance of S/S
products. Results from these specific studies have indicated that physical
entrapment of inorganic metals is a predominant containment mechanism.
However, some results are also indicating formation of altered or new
crystal structures in some phases which appear to be chemically bonding
some organics. Indications are that this type research could provide
information useful in preparing binder formulations better able to treat
a specific waste.
The physical and chemical characteristics of the waste affect the
performance, as do the climatic (temperature and humidity) conditions at
the time of placement and curing. The research on critical characteristics
is being expanded to develop information on what quality assurance/control
procedures are needed to better assure performance of S/S when applied in
the field. This information is especially important for the remediation of
contaminated soils with varying characteristics and field treatment pro-
jects requiring more than one season.
Because of the nature of many sol
air emissions are a potential problem
Investigations are being conducted to
emissions. Processes being evaluated
lime kiln dust - fly ash mixtures. As
air emissions. Some additives such as
which increases the release of volatil
of these emissions may be required to
ment, particularly in cases where the
volatile compounds.
idification processes, uncontrolled
to workers and the environment.
determine the magnitude of these air
are Portland cement - fly ash and
expected, mixing causes the greatest
lime result in exothermic reactions
e compounds. Capture and treatment
protect worker health and the environ-
waste or contaminated soils contain
A variety of leaching tests are used to predict how well the solidi-
fied waste form will prevent the release of contaminants over time. To
assure the proper use of these procedures, an effort (not shown on Table 1)
is being initiated to compile and evaluate available information on the
procedures to determine the correct use, how well each performs its
intended function, and which ones are best suited to better predict long
term performance. This is especially important for the proper design of
S/S processes for remediating contaminated soils and debris.
-129-
-------
TABLE 1. Current U.S. EPA Solidification/Stabilization
Research Projects ;
Project
Title
Objective
Status
Evaluation
of Factors
Affecting
Solidification/
Stabilization
of Toxic Waste
Investigation
of Test Methods
for Solidified -
Waste Character-
ization - A
Cooperative
Program
Study of
Morphology
and Micro-
chemistry of
Solidified/
Stabilized
Waste
Morphological
Examination
of S/S Samples
from Canadian -
US and other
studies
Critical
Characteristics
of Hazardous
Waste Solidi-
fication/
Stabilization
Air Emissions
from Waste
Stabilization
Evaluation
of S/S for
Treating
Ash Residues
Determine effects
of interfering
agents on perfor-
mance of S/S
Develop and
evaluate methods
for testing per-
formance of
solidification
processes
Investigate
bonding mechanisms
Morphological
examination
of solidified
waste specimens;
correlate results
with physical/
chemical test
of actual samples
Identify physical
and chemical
characteristics
of waste and
binders affecting
S/S
Determine air
emissions from
S/S processes
Evaluate S/S
of residues
from combusting
municipal solid
waste
Data
interpreta-
tion ongoing
Data inter-
pretation
ongoing
Research is
continuing
Ongoing
Being expanded
to provide
QA/QC informa-
tion
Ongoing
Ongoing
-130-
-------
Ashes from the combustion of municipal waste may contain unacceptable
levels of heavy metals, particularly lead and cadmium. Solidification/Sta-
bilization technology is being evaluated as an acceptable technique to
control leachate from the ashes if required.
EVALUATION OF S/S AS AN AVAILABLE TREATMENT TECHNOLOGY
The Hazardous and Solid Waste Amendments (HSWA), which amended the
Resource Conservation and Recovery Act (RCRA) provide detailed procedures
dictating how hazardous waste is defined, controlled, and managed. Wastes
classified as hazardous under RCRA are often referred to as RCRA hazardous
waste. Key provisions of HSWA are the ones which ban the land disposal of
hazardous waste unless it is proven to be more protective of the environ-
ment and human health than other alternatives. The legislation requires
that all hazardous waste be treated by the best demonstrated available
technology (BOAT) instead of and prior to land disposal. The U.S. EPA, is
required to determine and specify levels to which BOAT technologies can
treat RCRA waste. Solidification/Stabilization is one of several BOAT
technologies being evaluated. In this program selected hazardous wastes
are solidified/stabilized by Portland cement, cement kiln dust, and lime/
fly ash mixtures. Various ratios of waste to binder for each binder system
are then evaluated by the Unconfined Compressive Strength Test (UCS) after
a cure time of 7, 14, 21, and 28 days. Cured samples are then subjected to
the USEPA's Toxicity Characteristic Leaching Procedure (TCLP) extraction
test. Leachate from the TCLP is analyzed for the pollutants of concern
(Table 2), to determine how effective S/S can be for treating selected
hazardous waste. Results will be used to treat hazardous waste in the U.S.
The remediation of contaminated soils, hazardous waste uncontrolled
dump sites, etc., is controlled under legislation referred to as the Super-
fund Amendments and Reauthorization Act (SARA) which amended the Compre-
hensive Environmental Reclamation Compensation Liability Act (CERCLA).
Under SARA provisions, permanent treatment of the contaminated soil and
debris is being emphasized rather than the use of nontreatment containment
systems such as covers, grout walls, and similar methods. Because of this,
a program similar to the RCRA BOAT evaluations is being conducted for SARA
remediation technologies including S/S. Mixtures of soils contaminated
with selected chemicals (Table 3) are solidified/stabilized and tested to
evaluate performance of S/S technology for treating contaminated soils.
The test soil being used is a mixture of clay, sand, silt, topsoil and
aggregate.
In conjunction with other evaluation programs, laboratory facilities
and procedures have been established to test samples of solidified waste
extracted from bench and pilot treatability studies and field operations.
These facilities also assist in determining if a particular contaminated
soil or waste is capable of being effectively solidified/stabilized by
using a full range of leaching and physical tests. Data from this program
will help investigators to extrapolate laboratory test results to better
predict expected long term field performance.
-131-
-------
SUPERFUND INNOVATIVE TECHNOLOGY EVALUATIONS (SITE)
The U.S. EPA SITE program provides for demonstration and evaluation
of innovative technologies to remediate Superfund sites. Currently, six
S/S processes are being evaluated (Table 4). Results will provide informa-
tion on how well S/S can be expected to permanently treat contaminated
soils. The evaluations will also help make better extrapolation of labora-
tory tests results to field conditions.
PROCESS SELECTION CONSIDERATIONS
Factors important in the selection, design, implementation, and per-
formance of processes and products are: waste characteristics (chemical
and physical), processing requirements, S/S product management objective,
regulatory requirements, and economics. These and other site-specific
factors (i.e., location, condition, climate, hydrology, etc.) must be
carefully considered to ensure acceptable performance.
The chemical effects of some compounds can reduce the strength of the
binder/waste mix, while some compounds can accelerate or retard the S/S
curing rate. Temperature and humidity can also retard or accelerate cur-
ing. Size and shape of the waste particles can affect the viscosity of'the
mix. Overmixing or undermixing can affect the product's strength. i>'=><^
characteristics are among the most important factors affecting S/S
inorganics usually can be more easily solidified/ stabilized than
organics.
Waste
and
can
It is important to assess what process type and specific process
requirements are required before selecting an S/S technology. For
example, a wastebinder can be controlled and mixed more easily in a drum or
in a plant process than in the in-situ solidification of a pit, pond, or
lagoon.
How the solidified waste will be managed is important in selecting the
best process. Will it be disposed of in a landfill, or stored, or trans-
ported? Will it be placed in a plastic or metal drum below or above the
water table? Stored in a warehouse or underground mine?
Regulations in the United States will be critical to the success of
S/S. Processes can be altered to meet different performance criteria,
which will become increasingly stringent, as regulations become more
stringent. S/S will be competing with other treatment technologies to meet
these regulatory criteria.
Costs will depend on site-specific conditions. Important are the
waste's characteristics, transportation, type of process, and other special
factors. What is the physical form and chemical make-up of the waste? Is
pretreatment needed? Is transportation of raw materials and/or finished
S/S products required? Which S/S process is needed? What special health
and safety requirements are needed? What is the quality assurance/quality
-132-
-------
TABLE 2. Examples of U.S. RCRA Hazardous Wastes for
Which S/S is Being Evaluated as a Treatment
Technology
Waste
Code
Description
of Waste
Pollutant
of Concern
K048-52
K061
K046
Dissolved, air
flotation (DAF)
float from the
petroleum refining
industry
Emission control
dust/sludge from the
primary production of
steel in electric
furnaces
Wastewater treatment
sludges from manufac-
turing, formulation,
and loading of lead-
based initiating
compounds
Hexavalent
chromium, lead
Hexavalent
chromium, lead,
cadmi urn
Lead
TABLE 3. Contaminants Added to Test Soil Being Used
to Evaluate Effectiveness of Solidification
for Treating Soils
Metals
Volatiles
Semi-Volatiles
Lead
Zinc
Cadmi urn
Copper
Ni ckel
Chromium
Acetone
Ethyl benzene 1,2
Dichloroethane
Xylene
Tetrachloroethylene
Chlorobenzene
Benzo (a) pyrene
Pentachlorophenol
Naphthalene
control cost involved? What regulatory criteria must be met? Each of
these factors must be considered. As regulatory criteria become more
demanding, the costs of acceptable solidification processes may increase.
STUDIES NEEDED FOR S/S AS A TREATMENT TECHNOLOGY
For S/S technology to be effective in managing hazardous waste, pro-
posed processes must be properly selected, formulated, and used. Improved
-133-
-------
TABLE 4. Solidification/Stabilization Processes
Being Evaluated in the U.S. EPA SITE Program
Process Developer
Chemfix Technologies, Inc.
Soliditech, Inc.
Waste Chem Corporation
Battelle Pacific
Hazcon, Inc.
International
Waste Technologies
Description
Uses soluble silicate
reagents and silicate-
setting agents
Uses pozzolanic agents,
water and liquid reagents
including URRICHEM. Claims
to treat organics and
inorganics
Volume reduction and solid-
ification using asphalt
binder and extrusions.
Claims to treat organics
and inorganic metals
In Situ vitrification;
thermal treatment to
pyrolyze organics and
immobilize inorganics
in glass-like mass
Uses Portland cement,
fly ash, kiln dust and
proprietary chemicals
to treat organics and
inorganics
Uses silicate-colloid or
inorganic polymers as
medium; promotes adsorption
of organics; claims chemical
bonding of organics; treats
inorganics
knowledge of process selection considerations and the interactions among
the various candidate binders and waste types is critical to the successful
use as an acceptable treatment technology.
Studies are needed:
0 to determine the long-term physical-chemical
stability of S/S products when placed on the land;
0 to determine how and under what conditions S/S
-134-
-------
products should be placed on the land to ensure long
term environmental protection;
0 to more accurately predict and measure the performance
of S/S processes and products;
0 to provide a correlation between regulatory criteria
and real world situations;
0 to evaluate the effectiveness of protocols (e.g.,
TCLP, other leaching tests, UCS, etc.) to character-
ize S/S products, provide effective measurement
techniques, and correlate results of such tests with
performance in the field;
0 to determine the effectiveness of processes and
equipment to effectively solidify/stabilize contam-
inated soil and/or lagoons; the effectiveness of
mixing methods; and the resulting solidified/
stabilized soil performance at varying soil depths;
0 to determine the amounts of organic compounds that
can be included in inorganic waste streams without
requiring pretreatment before S/S;
0 to determine how effectively S/S processes treat
residuals compared to other alternative treatments.
SUMMARY
Solidification/Stabilization is being evaluated by the U.S. EPA as a
best demonstrated available technology for treating hazardous waste and
contaminated soils and debris. Future use of the technology in the United
States will depend on how well it performs compared to other available
treatment processes. Current research being conducted will provide some
answers regarding performance, however, additional studies are required for
a better scientific understanding of S/S. Whether or not S/S becomes an
important technology for treating hazardous waste and contaminated soils in
the U.S. ultimately depends upon regulatory requirements and the capability
of the technology to meet these requirements. As performance criteria
become more severe, S/S developers may need to improve their, processes.
The opportunity is very good. In the case of RCRA waste, S/S may be the
only acceptable method to treat selected inorganic waste and hazardous
residues from incinerators and other treatment processes. In the case of
contaminated soils and debris, S/S offers a relatively inexpensive method
for treating large areas in-situ. However, the capability of the technol-
ogy to perform satisfactorily over long periods of time has yet to be
determined. .
-135-
-------
SITE DEMONSTRATION OF HAZCON SQLIDJ11CATIQN/ST^ILIZATIQN..PROCE_SS
by: Paul R. de Percin
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Stephen Sawyer
Enviresponse Inc.
Edison, New Jersey
08837
ABSTRACT
Concern has been increasing in the United States over the use of land-
fills for the containment of hazardous wastes. In response to this concern,
the U.S. Environmental Protection Agency (USEPA) has established a formal
program to accelerate the development and demonstration of new and innovative
waste treatment technologies. This program is called the Superfund Innova-
Technology Evaluation Program or SITE.
One of the new innovative technologies demonstrated under the SITE pro-
gram was the HAZCON solidification/stabilization process. This process
blends contaminated soil or sludge with cement, pozzolans, and a proprietary
ingredient called Chloranan which neutralizes the inhibiting effects of
organics. A concrete-like mass is the result.
The demonstration of the HAZCON process took-piace in October 1987 at
the Oouglassville Superfund site along the Schuylkill River near Reading,
Pennsylvania. This rural site is a 50 acre plot that is contaminated with
high levels of oil and grease and low levels of PCBs, and volatile and
semivolatile organics from the operations of an oil reprocessing plant.
Soil samples from six different plant areas were processed to test the
capability of the HAZCON process on a diverse set of site wastes.
The primary criteria for evaluating the HAZCON process are: 1) reduced
contaminant mobility, and 2) integrity of the solidified soil mass. To
obtain the evaluation data, samples were taken from the untreated soil, the
blended slurry (after 7 days of curing), and from the solidified blocks
(after 28 days of curing). The samples were analyzed for soil characteris-
tics, Teachability, permeability, unconfined compressive strength, micro-
structural characteristics, and levels of contaminants.
-136-
-------
BACKGROUND
In response to the Superfund Amendments and Reauthorization Act of
1986 (SARA), the U.S. Environmental Protection Agency's Offices of Research
and Development (ORD) and Solid Waste and Emergency Response (OSWER) have
established a formal program to accelerate the development, demonstration,
and use of new or innovative technologies as alternatives to current contain-
ment systems for hazardous wastes. The new program is called the Superfund
Innovative Technology Evaluation program or SITE.
The major objective of a Demonstration Test Program is to develop
reliable cost and performance information. One candidate technology, which
was demonstrated in October 1987 at the Douglassville, Pennsylvania Superfund
Site, is the HAZCON proprietary solidification process. The process involves
the intimate mixing of hazardous waste material and cement with a patented
nontoxic chemical called Chloranan. Chloranan, according to HAZCON, neutral-
izes the inhibiting effects that organic contaminant materials have on
cement reactions. For this treatment, the wastes are immobilized and bound
into a hardened, leach-resistant concrete-like mass.
The Doug.lassville, Pennsylvania Superfund Site, No. 102 on the National
Priority List, was selected as the location for the Demonstration Test.
This is a 50 acre rural site that was an oil recovery facility. The site
includes: two large lagoons once filled with oily sludge, an oily filter
cake disposal area, an oil drum storage area, an area where generated
sludge was landfarmed into the soil and the plant processing area. An
estimated 250,000 cubic yards of soil may be contaminated.
PROGRAM OBJECTIVES
The major objectives of the SITE program are to determine the following:
1. The ability of the HAZCON process to immobilize the contaminants
at the Douglassville, Pennsylvania Superfund Site
2. The potential integrity of:the solidified soils, and
3. The field reliability of the equipment used to process the different
wastes at the site.
DEMONSTRATION PROGRAM
Contaminated soils at the Doug!assvilie, Pennsylvania Superfund Site
were processed by HAZCON using a truck mounted system. This system blended
the contaminated soils with port!and cement, water and Chloranan (a propri-
etary chemical) to create a solidified material.
-137-
-------
Six different contaminated soils at the Douglassville, Pennsylvania site
were processed by HAZCON from the following locations; 1) lagoon north
(LAN), 2) lagoon south (LAS), 3) filter cake storage area (FSA), 4) drum
storage area (DSA), 5) plant facility area (PFA), and 6) landfarm area
(LFA). The intent was to process enough soil to yield a treated volume of
5 cubic yards from each of 5 areas. An extended duration run for the sixth
area was to determine the reliability of the operating equipment. Approxi-
mately 25 cubic yards of LAS contaminated soil was to be treated. The
actual runs used less soil, but resulted in approximately 5 and 25 cubic
yards of treated soil.
The contaminated soil was excavated and screened to remove material
greater than 3 inches. It was then fed into the HAZCON Mobile Field Blending
Unit (MFU - the truck mounted system) along with #1 port!and cement, water
and Chloranan. Cement was used on an approximately 1:1 basis with soil and
the soil to Chloranan ratio was 10:1. The four feed components were blended
in a mixing screw and fed to 5 one-cubic yard wooden molds for the short
tests and 3 one-cubic-yard plus two 12 cubic yard pits for the LAS. During
the processing of the three soils lowest in organics (DSA, LFA and PFA),
toluene was injected into the soil to attain a concentration of 100 ppm.
It was felt that this would provide leachate concentrations of sufficient
magnitude to be above the analytical detection limits.
While the contaminated soil was processed and cured, the excavation
holes were enlarged, lined with an impervious plastic liner, and partially
filled with clean soil. After the one cubic yard blocks cured sufficiently
to be moved (48-96 hours), they were removed from the molds and placed into
the pits. The blocks were then covered with additional clean soil. After
28 days, the blocks were sampled. It is planned that 6 and 12 months after
burial the blocks be sampled again to develop long-term test results.
PROCESS DESCRIPTION
HAZCON Engineering, Inc.'s Mobile Field Blending Unit (MFU) operates
as a continuous processing unit. The Mobile Field Blending Unit Flow
Diagram followed by a description of each operation (number to number) is
shown in Figure 1. Operating capacities, though governed by predetermined
mix ratios set in the laboratory, are variable by up to 12 cubic yards of
processed raw waste per operating hour. The MFU has no external utility
requirements other than a standard water hook up and the attachment of a
"quick connect" line from a bulk cement carrier. The equipment obtains its
driving power from the diesel engine of the MFU vehicle.
Equipment calibration was performed each time a new waste feed material
entered the MFU. Calibration involves the determination of the waste and
cement feed by weight, for a set of given feed screw settings. Feed rates
were confirmed during the run by using a bag counter for cement, a flow
totalizer for water and tank level change for Chloranan.
-138-
-------
1.
H,o
6
\>V^
f •*
Jp^—- '
BLFNDIA/6-
F .
PROC£5S
T
'TREATED
OUTPUT
A
Waste material is introduced to the system through the use of pumps,
dredges, or earth moving equipment, whichever may be the most efficient
or practical .
2. The waste is moved through the process in a controlled flow, allowing
for precise measurement of the material.
3. Based upon these measurements, blending ratios, which are predetermined
in the laboratory, are set on a time weighted basis for both the Chloranan
and pozzonalic material, such as port! and cement. The Chloranan is a
proprietary additive that reduces the inhibiting effects of organics in
the crystallization of cement.
4. The ppzzolanic ingredient is stored in a hopper and then metered into
the mix. Typical ratios, on a weight basis, range from 1 part waste:
1 part pozzolan, to 3 parts waste: 1 part pozzolan. For the test at the
Douglassville site, a ratio of 1:1 was used.
5. Chloranan is stored in a holding tank, then pumped into the mixing
chamber. Through precise control of the flow rate, ratios of waste to
Chloranan can be accurately metered from a 10:1 to a 50:1 blend.
6. After initial combination of the primary ingredients, water may be
added as necessary to achieve the most desirable slump on the mix.
7. All additives are fed via pump or auger through a mixing chamber to
achieve a homogenous blend.
8. The resultant mass was extruded into either temporary or permanent
molds.
Note:
The injection point (A) for toluene addition, which was added to the
three least contaminated soils, was at the inlet of the blending auger,
Figure 1. Mobile Field Blending Unit Flow Diagram
-139-
-------
EQUIPMENT SPECIFICATIONS
The mobile field blending unit is shown in Figure 2. A separate bulk
cement carrier was provided for feeding to the cement bin on the equipment
trailer carrying the MFD. In addition, a water supply truck was provided
for process water and for feeding high pressure equipment decontamination
water. Also, Chloranan was supplied from a tank on the auxiliary trailer
provided by HAZCON.
SAMPLING AND ANALYSIS PROGRAM
Soil samples were taken before treatment, as a slurry exiting the MFU
for analysis after 7 days and as cores from the buried blocks after 28 days
of curing. For the 5 cubic yard blocks, two untreated soil composite samples
were taken along with 3 sets of slurry and solidified cores. For the extended
run at LAS, additional samples were taken for analysis.
The specific analyses performed on soil, slurry and core samples
included measuring physical properties, those being:
0 bulk density
0 moisture content
0 permeability (liquid)
0 unconfined compressive strength of the solidified cores
0 weathering tests for the 28-day core tests (wet/dry,' freeze/thaw)
Chemical analyses were performed to identify the volatile organics,
base neutral acid extractables (BNAs or semivolatiles), polychlorinated
biphenyls (PCBs) and metal contaminants in the processed soil. In addition,
three different leaching tests were run:
0 toxicity characteristic leaching procedure (TCLP) - standard leaching
procedure used for measuring Teachability of the organic and metal
contaminants.
0 ANS 16.1 - simulates leaching from the intact solidified core with
rapidly flowing groundwater (ANS - American Nuclear Society).
0 MCC-IP - simulates leaching from the intact solidified core in .
relatively stagnant groundwater regimes (MCC - Materials Character-
ization Center, Richland, Washington).
These latter two tests were drawn from the nuclear industry and modified to
hazardous waste analyses. Unlike the USEPA procedure that grinds the
solidified sample to a fine grain size, the ANS and MCC procedures leach
the intact sample.
Finally, microstructural studies were performed on the untreated soil
and solidified cores. These analyses included:
-140-
-------
E
cn
Q
3
o
C
OJ
3
CT
QJ
i-
CD
-141-
-------
0 X-ray diffractometry - characterizes porosity, hydration products,
fractures and presence of unreacted soil/waste material
0 Microscopy - scanning electron microscope and optical microscope -
identifies level of porosity, crystal appearance, agglomerates, and
fractures.
RESULTS - GENERAL
Not all of the analyses have been completed at this time, however, the
results obtained-to-date are as follows:
1. The six test areas offered a wide diversity of waste. The oil and
grease ranged from 1% by weight at the DSA to 25% at FSA.
Polychlorinated biphenyls (PCBs) were detected up to 80 ppm by weight
with the maximum concentration at LAS. Lead contamination concentrations
ranged up to 2.5% by weight. Volatiles and base neutral acid extractables
(BNAs - semivolatiles) organics reached levels of about 100 ppm in some
areas.
2. The volume of the solidified soil was approximately double that of the
undisturbed feedstock. Approximately an equal amounts of soil and
cement were used in the demonstration.
3. Permeabilities of the treated soil were very low, in the range 10-8 to
10 -9 cm/sec. A value of 10-7 or less is generally considered an
indication of an impermeable solid.
4. The unconfined compressive strengths (UCS) of the solidified soils
ranged from about 200 psi for FSA to 1500 psi for PFA and the values were
inversely proportional to the oil and grea'se concentration. These values
are quite satisfactory from a load bearing point of view, i.e., equipment
traffic.
5. Only the TCLP leaching tests, of the three leaching tests performed,
have been completed. The results shown in Table 1 are discussed below:
Metals - the leachates for the solidified soils showed metal levels
at or near the detection limits. The results were a factor of 500 to
1000 less than in the leachates for the untreated soils.
Volatiles - the primary compounds detected were trichloroethene,
tetrachloroethene, toluene and xylenes. Only the leachates for the
untreated soil and 7-day cores were analyzed. The levels of contaminants
were approximately the same in both the treated and untreated soils,
each at levels of hundreds of micrograms per liter.
BNAs - the compounds detected in the leachates were phthalates and
phenols. The phthalates were reduced to near their detection limits of
10 ug/1 in both the treated and untreated soil leachates. The total
-142-
-------
OO
UJ
h-
0
2j
g
^
u.
Q-
h-
U_
0
co
I—
1 1
o£
UJ
i
CQ
1—
^
CD
**
to
03
cu -o
03 . p^
O |
03
CU ,
1
•— 1 "r—
0
OO
4-3
03
CU
£_
| —
•—
CD
'^
CU ^
4-3 0
03 CO
-C
<-> T3
03 CU
CU 4-3
—1 03
CU
.|_3
c
E
Q.
CO
o
f n
4-3
c
O
^
O
o
r—
O
C/j
•^
03
SI
et
2*
cc
o
o
03
CU
i
^ji
— *^
CO
o
o
•a
03
cu
_!
O
CD
-
O
0
*
T3
CU
-^
03
CU
£_
<
to
' O CO
O i—
«^f- O
o co
0 0
o
Q •—
2 ^
ten co
I^ O
0 0
in
i— CO
CO
co o
i — CM
"a CM
0 0
CO \f>
CM CM
CO O
CO O
[^ O
• 0
0
CO «d-
.0 0
o o
CM Q
^ z
O CD
r- LO
CO 00
i — CM
*
*
CD Uf>
O CO
r^- o
LO ^
o o
CO 1 —
CO IO
•^J- CO
CM i —
^C ^C
OO LJ-
1— If
o c
oo en
ti> LT
i — CM
O r—
0 0
0 0
2= 2:
Q 0
CM CO
CM LO
OO O
•
O VO
0 0
CO CO
CT) CO
1^. •*
^™
•=r oo
U_ «C
^ i
I
-o
03
CU
to
to
r—
03
CU
CU
4-3
o
•r—
O
03
E~
03
CU
C.
CD
CU
c~
I—
*
.
CO
CU
c:
cu
^
A
CU
c
cu
N
CU
c
cu
^
^1
cu
cu
c:
cu
-d
cu
O ' .
£_
o
^:
o
03
£_
4-3
CU
4-3
CU
C 0)
cu c
JC CU
4-3 t—
CU • 03
t_ jc
r™ C"
-C 0.
O 03
"^* ^~"
4-3 .
to
CU "o TD
c: c: cu
CU CU 4-3
^ x: u
c — Q. CU
0 4-3
h- •> CU
tO T3
>> CU
•r- 03 O
03 03
E jc I
•r- 40
t- x: Q
CL. Q- s:
+ 0 C=
cu
£_
O
cu
u_
T3
03
<=c
u_
1
f1
OO
Q
£_
O
4-
to
CU
Q.
' 03
to
•r—
O
CO
"O
cu
4-3
03
CU
t_
4-3
C
O
4-3
C.
•P—
T3
CU
4-3
O
CU
•I—
to
03
5
CU
cu
3
r—
0
1—
ten
•
to
-o
•r—
o
CO
>,
03
-o
r-
c:
T3
CU
r
13
to
03
CU •
E
(
CU
CU
o
4-3
"(O
CU
T3
CU
O •
cu to
•1-3 i —
C O
•i- c:
cu
c .c:
O Q.
4-3 "O
03 C
£_ 03
4-3
C CU
cu c
o cu
cr i —
O 03
O J=.
4-3
JCZ
. Q-
O- 03
_l C
O
>r_
CD
C -C
•r- CD
O
4- >>
1_ £_
CU CU
Q. >
*
_J
CD
E
oo
CM
o
4-3
(_
CD
4-
O
CU
C
£_
CU
•^
c
to
cu
Q.
03
CO
c:
73
CU
o
0)
cu
T3
CU
CU
3
ro
^^
^_?
C.
O
0
0
-143-
-------
phenols in the leachates were in the range of hundreds of micrograms
per liter, with the same concentration levels seen in both the untreated
and treated soil leachates.
PCBs - PCB analyses are in progress and not yet available.
6. The microstructural studies provided the following information:
The mixing of the four process components was not always complete. At
times, microscopic analysis found globules of untreated organics and
un reacted cement.
RESULTS - SPECIFIC
The individual results from several test procedures were very interest-
ing. Examples are discussed below:
1. The filter storage area (FSA) had an oil and grease concentration of
25.3%. The unconfined compressive strength (UCS) of the 7 and 28-day
FSA samples was many times greater than the laboratory samples made
without the proprietary chemical Chloranan.
2. The oil and grease analyses of the untreated waste TCLP leachates were
near the detection limits of 0.2 ppm. Oil and grease concentrations
ranged from 1 to 25% for the six wastes.
CONCLUSIONS
The following conclusions were drawn from the preliminary analytical
data and operations observed at the Doug! assvi lie, Pennsylvania Superfund
Site:
1.
2.
3.
The volume of the solidified soils using the HAZCON process at
Douglassville resulted in an increase of approximately 100%, when
compared to the volume of the untreated soils,
The unconfined compressive strength ranged from 200 to 1500 psi and was
inversely proportional to the oil and grease concentration.
The TCLP leaching tests for the solidified soils produced very low
levels of metals, volatile organics, and semi volatile organics in the
leachate. However, for the organics, the leachate concentrations were
approximately equal for the treated and untreated soils,
-144-
-------
TRANSPORT OF'DISSOLVED ORGANICS FROM DILUTE AQUEOUS
SOLUTIONS THROUGH FLEXIBLE MEMBRANE LINERS
by: Henry E. Haxo, Jr.
Matrecon, Inc.
Alameda, CA 94501
ABSTRACT
_ ^ paper presents the results of experiments relating to the parti-
tioning of dissolved organics from dilute aqueous solutions to polymeric
flexible membrane liners (FMLs) and the permeation through FMLs of organics
from dilute aqueous solutions. The amount and rate of absorption and the
transport of these species through an FML are functions of such factors as
their relative solubility parameters, crystallinity content and thickness of
the FML, the molecular weight and concentration of the organic constituent
temperature, and their concentration gradients across the liner.
• In_order to simulate partitioning and the transport of waste constit-
uents within and out of a landfill, two sets of experiments were performed.
In the first set, samples of an FML that were either unexposed or saturated
with an organic were placed in test liquids, including deionized (DI) water
neat organics, and saturated aqueous solutions of organics; the concentra-
tions at equilibrium were measured. In the second set, the transport of
organics in aqueous solutions through an FML was measured in a three-
compartment closed apparatus. The distribution of the organics between the
test zones in these experiments was monitored by gas chromatography (GC).
The results of these experiments show that, even at low concentrations in an
aqueous leachate, some organics can be highly absorbed by a polymeric FML
and can permeate the liner. A multi-compartment apparatus, such as the one
described and used in this work, appears to be an appropriate and promising
means of assessing the effectiveness of an FML to contain the organic con-
stituents of a given leachate.
-145-.
-------
INTRODUCTION
Even though of polymeric flexible membrane liners (FMLs) are recognized
for their low. permeability to water and waste liquids, their permeability to
gases, vapors, and liquids can vary by orders of magnitude (1, 2, 3, 4, 5).
The amount and rate at which a constituent of a waste liquid can migrate
through an FML and the direction of migration depend on characteristics of
the consitituent, the FML, the temperature, and the concentration of the
constituent on both sides of the FML. Since polymeric FMLs are nonporous
materials, constituents permeate them by a three-step process that involves
the solution or absorption of the pemeant at the surface of the FML, trans-
port of the permeant through the membrane by diffusion, and evaporation or
desorption of the permeant on the downstream side. As a first approxima-
tion, the magnitude and the difference in concentration of an individual
species across a membrane determine the direction and the driving force for
the permeation of that species. Individual organics can permeate an FML at
substantially different rates. In a lined-waste disposal facility, an FML
generally encounters dilute solutions containing both organic and inorganic
species. The individual species can interact with one another and with the
FML in different ways, significantly affecting permeation rates of the
organic species through the membrane.
This paper begins with a discussion of the concept of partitioning,
presents the gas chromatographic procedures and materials used in the ex-
periments, discusses two types of experiments that illustrate the mechanism
of transport under conditions that simulate service conditions, and pre-
sents the results. These experiments measured:
o The partitioning or distribution of selected organics between water
and an FML.
o The transport permeation of selected organics through an FML in a
multi-compartment appartus that simulates the configuration of a
waste landfill.
PARTITIONING
The distribution of a dissolved substance between two phases that are
essentially immiscible has been an important subject for study in chemistry
-146-
-------
for more than a century (6). Some of the early investigations showed that
at equilibrium the ratio of the concentrations of a Substance dissolved in
two immiscible solvents in contact with one another is constant and does
not depend upon the relative volumes of the solutions involved. Later
studies showed that the distribution would be constant only if a single
molecular species was being considered as being distributed between the
two phases, that is, the molecular species would be the same in both phases
and would not either dissociate or associate when transferring from one
phase to the other. Thus, the partitioning could be treated by classical
thermodynamics as an equilibrium process and would be related to other
commonly measured functions such as partial pressure, osmotic pressure
and chemical potential. The partitioning of a dissolved substance, for
example, between water and an organic such as toluene, forms the basis
for the separation of organic compounds by extraction. When there are two
or more immiscible phases, an organic solute, for example, will partition
or distribute itself among phases depending upon its solubility in the
different phases. Many organic solutes will tend to be more soluble in the
organic phase and remain in that phase unless the solute is exceptionally
soluble in water. Nevertheless, a small amount of the organic will dissolve
in water when the two phases are in contact. The ratio at which a solute
partitions between the organic and aqueous phases is termed the partitioning
or the distribution coefficient and is determined by dividing the concentra-
tion of the dissolved substance in the organic phase by its concentration in
the aqueous phase.
The concept of partitioning is being used in tracking pollutants and
contaminants in the environment, e.g. pesticides and various organics which
resist biodegradation, including.the polychlorinated biphenyls (PCBs) (7
8). These chemical species can move through the food chain, resulting in
higher concentrations at each trophic level. This process which reflects
the greater solubility of an organic chemical in oily tissue than in water
has been termed "bioconcentration." The process of partitioning is also
involved in spill cleanup processes in which polymeric materials that are
used to absorb a spilled organic, e.g. an oil, can then be disposed of in
an environmentally safe manner.
In waste disposal containment, a similar situation exists with respect
to polymeric FMLs and waste liquids that contain dissolved organic sub-
stances. Organics dissolved in waste liquids can partition between the
waste liquid and the liner and thus enter and swell the liner. The organic
can then permeate and be transported through the liner at a rate which de-
pends on the material, its thickness, the temperature, and the concentration
gradient, or more precisely, its "chemical potential" gradient across the
liner. Inasmuch as FMLs are organic compounds and are at least partially
amorphous, they can behave much like liquids from a thermodynamic standpoint.
The major factors affecting the solubility of a solute in a polymeric
composition that is exposed to a liquid are:
o Solubility parameter value of the polymer with respect to that of
constituents of the liquid.
-147-
-------
o Crosslinking of the polymer.
o Crystallinity of the polymer.
o Filler content of the compound.
o Plasticizer content of the compound.
o Soluble constituents in the compound.
Due to differences in polymers and in compounding, some of these factors
do not apply or become important.
For rubber and noncrystalline or amorphous polymers the solubility
naraJSr is probably the most important factor in swelling and, as noted
above Is used by polymer scientists to measure the compatibility of an
amorphous polymeric composition with a liquid with which zt may be in
contact.
Crosslinking of a noncrystalline polymer or a rubber reduces its
ability to swell in a liquid which has similar solubility characteristics
to the polymer. The amount that a crosslinked polymer swells xn a good
solvent fof the raw polymer can be used to measure the degree of
Crosslinking: the less the swelling, the greater the crosslinkmg.
Crystallinity of a polymer acts much like crosslinking to restrict
the ability of a polymer to swell and dissolve. The crystalline domains
of a polymer do not Ssorb organics. Highly crystalline polymers, such
as SDra72ll swell slightly but will not dissolve in gasolxne even though
they are both hydrocarbons.
Two additional factors in FML compositions that also can affect the
1
ith thcrystalline *~i- 1" se.icrystalline
fillets do not absorb organics. In the latter case the plasticlzer is
generally e*«.ctable by organic solvents, but only slightly extractable
by water.
In previous work, Haxo et al (9) observed that slightly soluble
oreanics dissolved in an aqueous waste liquid can partition to an FML which
-148-
-------
Immersion testing of the FMLs was also performed in neat TBP (10). In
that experiment, the absorption of the TBP to equilibrium by various lining
materials took place in a relatively short time and, in most cases, ended at
considerably higher absorption values than the values in the tests with the
saturated solutions, indicating an effect of concentration.
EXPERIMENTAL DETAILS
Two sets of experiments were performed for this study. In the first
set, samples of an FML that were either unexposed or saturated with an
organic were placed in test liquids, including deionized (DI) water, neat
organics, and saturated aqueous solutions of organics. In the second set,
the transport of organics in aqueous solutions through an FML was measured
in a three-compartment closed apparatus. The distribution of the organics
among the test zones was monitored by gas chromatography (GC).
GAS CHROMATOGRAPHY PROCEDURES
A Perkin-Elmer Sigma Three Series gas chromatograph with a flame ion-
ization detector was used for the GC analyses. The instrument was fitted
with different columns as required to effect separation of the different
mixtures of volatile and semivolatile organics. Different temperature
programs for both the oven and the injection/detection ports were also
used to maximize separation and response factors for the mixtures. The
different conditions, injection volumes, carrier gas flow rates, etc., for
each of the experiments are all presented in Table 1.
The concentrations of the organics in the aqueous solutions and the
vapor zones of the test apparatus were determined by injecting solution
samples directly into the GC column. The experimental apparatuses for ex-
posing the FML samples featured septums through which the liquid and vapor
zones could be sampled during the test to determine whether equilibrium had
been reached. The concentrations of the organics in the FML samples were
determined either by the change in weight of the FML samples or by head-
space gas chromatography (HSGC). In the HSGC procedure, an exposed FML
specimen is placed in a small vapor-tight can provided with a septum through
which vapors from the specimen can be sampled. The can is placed in an oven
at 105°C and heated for approximately an hour. A sample of the vapors is
drawn from the can and injected into the GC. The FML specimen is removed
from the sampled can and placed in a new can which is then heated in a
105°C oven for approximately an hour. Once again, the vapors inside the
can are sampled and injected into the GC. The process of heating, sampl-
ing, and injecting is repeated until no organics are detected in the
sampled vapors by the GC.
The concentrations of the organics in the injected samples were cal-
culated by comparing peak height values with calibration curves... The cali-
bration curves for the leachates were determined by injecting 1 jjL of various
-149-
-------
•°^
CO
o
Si
>
Q
CO
O
M
O"
1-4
PK CO
O H
CO W
W Z
CO M
>i OS
Zj X
< w
0 SH
M H
ae M
CHROMATOGRAP
ING/PERMEABIL
CO !Z
<: o
H
2: M
M H
0 <
1*
M
CO
|
s
M
pyj
o
o
i
a
ZJ *r-
•~~. en
c-4 P.
O ^
CO OJt-i
01 01 co . =
•fi O3 tJ P
U V-> «r-i
£ SB
co en o
C3 co i— i iJ
60M-- T|
p
O Ol
4J 3 i 1
O) O *— '
C
M
0!
C £
•ft C O 4J
4J CO 4J CO /-v
0 U VJ O
Ol JJ CD 0) o
•r-} L 4-1 O-i^-^
c o 01 e
M ex. -a 01
"
B
1 -. t-
Oi 60
a. o
B •-<
ai a
4J
Ol
C P
01 r)
> 4J
^3 ^3
P
o
o
4J
p
01
^B
Ol
CX,
en
P
in o
O -ft
4-1
4-1 *H
01 TJ
CO P
O
o
CO
o
"*
o
co
fTo
0
in
C-i
.
C
*H
r-H Cd
J-. 0 C
1 o ° *n
M-i o ^S
O ^H O
0 0
O O O
\O AJ C"J
X w
w V^
p o o ex,
• rt 42 O O
>-. r- ( O
co eO *^ f '
^ O 0 0)
r-i oo ex.
S^ 3
!«! 0 P co
r-1 O
in ... 2 en
CO 0 01
\O CO CN P
en
01 o
P .n
01 P *O
r— i en T3 co Oi
>, p 0) 60 o fi
_f- Oi 4J r-l CO 0)
4J E CO O r-i 4J
0) .5 VJ Pu CO
>, 0 3 4= S
r- 1 OJ 4-1 4-1 13
o ex, co -H P P
pu en en 5 co -1-1
T-H
o
CO
O
co
• *. o
r-H O
, fH
•a ..
3 o
cr a
•r- CO
rJ >
in
, 0 CO
1 •= " « ^
>d 4J r— CO ^^
< Ol O O
S >^ 0 -r. 01
r-H >M-H 43
PQ O f-1 "f 3
O P. 60 en 4-1
I en
4J 3
)- 4J
CO CO 4=
ex, i- s- 4J
e co o; -r-i
o ex, 4J s
O CX, CO T3 Oi
1 CO S T3 P P
OJ Oi CO O)
O) 4J 4: & 3
)-. p 4J -H W fH
f Ol -ft P- O O
IH g IS en H 4-1
CN
O
CO
o
co
.-a
^H 0
in
CM
• n
O •
a P
° 'i 4-1
fH ~-, CO
O
O o •
4J O P O
CN -f o
0 E 0
O 4J VO
vO CO CN I-H
^
DB WAX-1, 15M
polyethylene
glycol, fused
silica open
tube (FSOT)
i en
Th r ee-c omp ar t'
ment apparatu
wi th water
spiked with
six organics
CO
.
o
4-1
O
01
4J
T3
p
0
• r-1
4-1
CO
N
•H
g
•r-i
Oi
i-H
Oi
4=
4J
ap.
hydrogen in
(H 01
4-1 "O
CO
P r-.
01 60
60 1
>, 4J
O OJ
60
P CO
CO V
60 01
3 4=
O 4-1
4= C
4J (-
3
"O 42
0)
CO O
en 4-1
CO
Ol
6 en
3 3
•r-t
fH P
Ol Ol
4: 60
*o ^
Ol O
MH 0!
• r. T3
V- CO
3 !-
P. 60
Ol pu
P co
CO 43
-150-
-------
solutions of known.concentrations of the different organics into the GC
column. Injections of each standard were performed five times to ensure re-
producibility of injection techniques. Standard deviation was 2%. The cali-
bration curves for the HSGC analyses were prepared by analyzing a specific
volume of vapor (either 100 PL or 400 PL) from headspace cans injected with
different volumes of a standard solution of organics.
ORGANICS USED IN EXPERIMENTS
Basic information for the organics used in these experiments is pre-
sented in Table 2.
TABLE 2. ORGANICS USED IN PARTITIONING/PERMEABILITY EXPERIMENTS
Organic
Vapor
Density Boiling pressure
Molecular at 20°C point at 25°C Solubility parameters3
weight (g cm-3) (°C) (mm Hg) Tfi^l (fid) (
-------
TABLE 3. SELECTED PROPERTIES OF POLYETHYLENE
FMLS USED IN STUDY
Polymer3
FML numberb
Nominal thickness (mil)
LLDPE
284
30
HOPE
483
100
Volatiles (%)
Carbon black content (%)
Specific gravity of FML
Density at 23°C
FML (g mL"1)
Polyethylene0 (g mL"1)
Crystallinity (%)
Melting point (°C)
2.5
0.927
0.924
0.913
36.3
119
0.1
2.5
0.945
0.936
59
121
aLLPE = linear low-density polyethylene;
HOPE = high-density polyethylene.
bMatrecon identification number.
^Calculated from the density of the FML by cor-
recting for the carbon black content.
DISTRIBUTION OF ORGANICS BETWEEN WATER AND AN FML
In the first set of experiments, samples of an FML were placed in the
test liquids to determine the distribution of a selected group of organxcs
between water and an FML. The FML samples that were placed xn the test
liquids were either unexposed or saturated with an organic. The unexposed
samples were placed in saturated aqueous solutions of the respective
organic, and the samples saturated with organxcs were placed xn DI water.
DISTRIBUTION OF ORGANICS BETWEEN ORGANIC-SATURATED
WATER AND A POLYETHYLENE FML
To assess the distribution of dissolved organxcs between an aqueous
acetone per liter (L) of solutxon.
-152-
-------
A specimen of polyethylene FML (No. 483) was immersed in each jar. At
intervals, each immersed specimen was removed and weighed until it reached
a maximum value; the concentration of the organic in the FML specimen was
calculated and reported in mg of organic per L of solution, i.e. per L of
FML. The values used for the concentrations of the organics in the saturated
aqueous solutions were obtained from the literature (12). The distribution
coefficients (DLW), i.e. the ratio of the concentration of the organic in
the FML (CFML) to the concentration in the aqueous solution (CH o) were
calculated. The results of the analyses and calculations are presented in
Table 4.
TABLE 4. DISTRIBUTION OF NINE ORGANICS BETWEEN SATURATED AQUEOUS SOLUTIONS AND A POLYETHYLENE FML*
Organic
Acetone
Methyl ethyl ketone
Trichloroethylene
1,1, 1-Trichloroethane
Benzene
Toluene
o-Xylene
m— Xylene
p-Xylene
Density
of ' organic
(g mlT1)
0.788
0.805
1.464
1.339
0.879
0.866
0.897
0.868
0.854
Solubility
in water
at room
temperature'3
(%o>
(mg L-l)
250,000C
240,000
1,100
1,320
1,180
515
175
196
190
Time to
maximum
swelling
(days)
44
46
43
52
30
50
40
38
49
Concen-
Weight t rat ion
increase of organic
of HDPE FML in swollen Distribution
specimen FML (CFML) coefficient (DLW)
(%) (mg L-l) (CpML/%0)
0.33
0.62
16.71
11.70
7.22
8.05
8.42
8.19
8.43
3,210
5,900
144,200
103,000
63,900
70,600
73,800
71,800
73,600
0.0128
0.0246
131.0
78.2
54.3
137
422
366
387
aHDPE FML (No. 483). [
bValues obtained from Riddick and Hunger (12).
clnitial concentration of the acetone-water solution; acetone is miscible with water in all proportions.
In all cases, more than 30 days were required for the specimen to
reach maximum swelling. The time required to reach maximum swelling ranged
from 30 days for the specimen immersed in benzene solution to 52 days for the
specimen immersed in the TCE solution. The distribution of all of the or-
ganics, except the acetone and MEK, was predominantly to the HDPE specimens.
The distribution coefficients ranged from 54 for benzene to 422 for o-xylene.
-153-
-------
DISTRIBUTION OF ORGANICS BETWEEN A SATURATED POLYETHYLENE
FML AND DEIONIZED WATER
In this experiment, individual specimens of the same polyethylene FML
(No. 483) were saturated with each of the organics. These specimens, ap-
proximately 1 x 3 in. in size and 2 g in weight, were immersed in the re-
spective organics and weighed daily until equilibrium was reached Once the
specimens were saturated, they were immersed in vapor-tight jars filled
with DI water. The vapor-tight jars consisted of 8-oz glass jars ^th
Jround and polished-top edges, and Teflon-lined phenolic resin tops, each
of wMch was fitted wi?h a^wagelock sampling port and Teflon-lined .silicone
rubber serums (Figure 1). The water in each jar was sampled periodically
^ndlnalyzed"y GCgfor the organic that transferred from the saturated FML
specimen! The water was sampled with a 1-yL syringe initially Jt 5-mn.
intervals and then hourly. Concentrations in these tests were determined in
tS GC under Set of Conditions 1 as described in Table 1. This procedure was
continued until the concentration of the organic in the water had plateaued,
^whichtime the specimen was withdrawn and weighed. The concentration of
the organic remaining in the FML specimen was determined by HSGC.
Teflon
Septum
Teflon-lined
Screw cap
Swagelock
Assembly
\
Nut
TOP ASSEMBLY
J jar with
ground and
polished
edge
8 OZ JAR
Figure 1. Schematic of the immersion test jar with septum for withdrawing
samples of water for GC analysis.
Data on the weight at maximum swelling of the specimens at saturation
as we?l as "he weighs of the swollen specimens after equilibrating ^th the
water in the jars, and the time required for each to reach equilibrium are
-154-
-------
presented in Table 5. The rates at which the absorbed organics diffused out
of the FML into the water varied considerably. It took 24 hours for the
water containing the specimen swollen with TCE to reach equilibrium con-
centration, whereas the water containing the specimen swollen with benzene
required 120 hours. It does not appear that equilibrium had been reached
with acetone and MEK when the specimens were withdrawn.
TABLE 5. DISTRIBUTION OF ORGANICS BETWEEN AN ORGANIC SATURATED FMLa AND DEIONIZED WATER
At equilibrium in deionized water
Saturated FML
Organic
Acetone
Methyl ethyl ketone
Trichloroethylene
1,1, 1-Trichloroe thane
Benzene
Toluene
o-Xylene
m— Xylene
p-Xylene
Maximum
swelling,
by weight
m
0.96
1.83
17.52
12.13
7.61
8.40
8.73
8.51
8.72
Concen-
tration
of organic
(mg L-l)
9,080
17,120
150,500
106,700
67,100
73,400
76,300
74,300
75,900
Time to
maximum
concen-
tration
in H20
(hours)
~120b
~120b
24
50
120
46
52
78
58
Swelling
of FML, by
weight
(%)
0.62
0.91
8.98
6.91
1.95
4.42
7.07
6.70
6.93
Concentration
of organic
in FML (CFML)
(mg L"1)
5,830
8,990
84,800
65,800
19,030
42,000
65,600
62,200
64,300,
Concen-
tration
of organic
in water
(CH20>
(mg L-l)
35
127
966
589
643
266
186
193
165
Distribution
coefficient
167
71
88
112
30
158
352
322
390
aHDPE FML (No. 483).
"Equilibrium did not appear to have been reached when specimen
was removed from test at 120 hours.
The distribution coefficients, also presented-in Table 5, show the
strong tendency of the chlorinated and aromatic organics to remain in the
polyethylene FML (No. 483).
DISTRIBUTION OF ORGANICS IN A TEST APPARATUS CONSISTING
OF THREE COMPARTMENTS SEPARATED BY POLYETHYLENE FMLS
The experiments initially performed in an apparatus consisting of two
compartments separated by an FML demonstrated the feasibility of this type
of apparatus in assessing the transmission of organics from dilute aqueous
solutions through polymeric FMLs. In these experiments, the transmission of
specific organics was monitored by GC (13). These experiments also in-
dicated the substantial differences in solubility of different organics in
a polymeric FML.
-155-
-------
THREE-COMPARTMENT TEST APPARATUS
In order to assess the permeation of organics from dilute aqueous
solution through polymeric FMLs under conditions that simulate service
conditions for an FML, a three-compartment closed apparatus was fabricated.
The test apparatus, shown schematically in Figure 2, can be divided into
seven zones, which are listed in Table 6. FML specimens separate the three
compartments (Zones 2 and 5). An aqueous solution containing organics
partially fills the middle compartment (Zone 4), and DI water can be placed
14/32-ln.
Septum
Figure 2. Schematic of the three-compartment test apparatus used in the
study of water/FML distribution and permeation of organics
from dilute solutions. Inside diameter of the compartment was
4 inches.
-156-
-------
TABLE 6. ZONES IN THREE-COMPARTMENT TEST APPARATUS
Zone
Compartment
Description
Volume
(mL)
1 Top
2 Barrier between
top and middle
3 Middle
4 Middle
5 Barrier between
middle and bottom
6 Bottom
7 Bottom
Airspace above "cover" 806
"Cover" FML (33-mil LLDPE): ~7
Area exposed to solution
Airspace above aqueous solution 306
containing organics
Dilute aqueous solution containing 500
organics
"Bottom liner" (33-mil LLDPE): ~7
Area exposed to solution
Airspace below "bottom liner" FML 506
Deionized water 3QO
Total 2,432
in the bottom compartment (Zone 7). The three compartments are clamped
tightly together to prevent leakage of any vapor from the apparatus. Thus,
the organics can either volatilize into the airspace above the solution and
then, permeating through the top FML specimen, enter the top compartment
or the organics can permeate through the lower FML specimen and into the
bottom compartment. The covers of one end of each of the top and the
bottom compartments are welded to the walls to prevent loss of volatiles.
The only potential leaks are those that might occur at the flanges between
which the FML specimens are mounted. Ports with Teflon silicone rubber
septums are incorporated in each of the three compartments for use in with-
drawing samples for GC analysis from the aqueous and airspace zones. The
two FML zones can be analyzed by HSGC after the apparatus is dismantled.
The three-compartment apparatus simulates the configuration of a
covered landfill as follows:
o The airspace in the top compartment simulates the airspace over a
"cover" liner. The FML specimen between the top and middle com-
partments simulates a "cover" liner.
-157-
-------
o The airspace in the middle compartment simulates the headspace
above a waste liquid, and the dilute solution containing organics
serves as the waste liquid. The FML specimen between the middle
and bottom compartments simulates the service conditions of a
single bottom liner.
o The airspace and the deionized water in the bottom compartment
simulate, respectively, pore space in the soil and the groundwater.
The configuration of the zones within each compartment can be modified
to assess double liners and various auxiliary materials, such as soil
covers, geotextiles, and drainage materials.
DISTRIBUTION OF TWO ORGANICS IN TEST APPARATUS
In the first experiment of this series, toluene and TCE were dissolved
in the water in the middle compartment. Both of these organics are com-
monly found in leachates and are easily identifiable and trackable by GC
analysis. An LLDPE FML (No. 284) separated the three compartments.
In this experiment, Zone 4 was filled with 500 mL DI water and spiked
with 191 mg each of toluene and TCE to yield concentrations of 382 mg each
per L of water. The five zones containing water or vapor were sampled and
analyzed by GC periodically to assess the changes in concentrations in
these zones. 'After 256 hours, when the concentrations appeared to remain
constant, the apparatus was dismantled and the FML samples were removed and
analyzed by HSGC to determine the concentrations of the organics in the
FMLs. The conditions of the GC analyses are presented in Table 1 (Set of
Conditions 2). The results of analyzing the samples taken at 24, 96, and
256 hours are reported in Table 7.
The results show that the water in the bottom compartment (Zone 7) had
absorbed organics. At the end of the test, the relative concentrations of
the two organics were the same in both aqueous .zones (Zones 4 and 7), demon-
strating the transport of these organics through the FML and airspace to the
water at the bottom. The data also show that, for each of the two organics,
the concentrations in the airspaces (Zones 1, 3, and 6) were similar in
value and the concentrations in the two FML specimens were essentially equal.
Overall, 92.6% of the TCE and 97.7% of the toluene were accounted for in the
final analyses of all zones. The HSGC analysis of the FML in the flange
area showed no organics, which indicated that there was no leakage through
the flanges. Overall, these results show that the apparatus had come to
equilibrium; they also show the high absorption of the two organics by the
FML At these equilibrium conditions, the distribution coefficients between
the LLDPE FML and water for TCE and toluene were 178 and 120, respectively.
-158-
-------
CO
ij
s
W
e
w
j
[M
EH
U
O
PM
m
O
W
EH
2
H
P4
CO
CO
a
EH
^
M
i
3 ofl
o B
§ ^^ '
C O^H
OJ «f-i 1
G -Ul (-3
C OJ
O V* OO
0 -U ^
ij
S Tx
3 W)
o g
a \^s
U
•H
n •
CO
00
0
QJ
g /— V
3 t~9
o ^~*
^>
c
o
QJ 4J
P 'G.
O -IH-
NI ^
O
CO
Q
O
s
-O O I-H m . in o O- •« I-H *d" r^
r^- oo r^. oo
i
ON oo *"~i CM f"^ co in t**« ^H CM (**• oo in "dr o^ ^^ vo r^*
• • • " ON 00 • '• - • • nQ i/"\ * • • • » •
CM «^ vO O • • • ' CM m vD O • • CO I-H \O vO CM f**-
t~— 1 i— H
mm •• om oo •• m3r o
' rH
.in oo ••'mo om •• CMCO tno ••
"-H ON • f • ON CO i-H i-H • • ^* C^ i-H • •
i— 1 - CO CO * r-H
CM ON • • f-n F-H - _ in CM •• \OVD mo \o •— H moo
- CT* ,r^ •• CMO* • • - oo • • m < — i
r-t •-! rH O CM r-H
i-H i-H
'
. • . -
,...-. .
OO OO OO 1-1 I-H OO OOOO I-H.-H OO
CT\ CT^ C^ CT» f-) CO
l-H l-H ,__, rH l-H l-H
2
CU ' Cv S co ^
> \ jo > o = -o
OpO >s ^ o , i -o i—t fa . P cu
CO § O fa CO "l-t •!-! E -Q 4_1 *r-l P
'faO. 3 3 fa CO rH 4J3
a; cucucrcr cujt cooo
OS =p CJ-rH.pl 'VJ O f-O 4J COO
COV-r-CUtOr-ir— CU CO PO fa
«>>>,t04J4J fa tO O
I-, Cu
rC T3
4-1 1
cu 3
o o
fa rH
o
I-H r*
rC to
O QJ
•H P
EH rJ
CQ pO
-159-
-------
DISTRIBUTION OF SIX ORGANICS IN TEST APPARATUS
In another experiment performed with the same type of test apparatus,
same configuration of zones, and same polyethylene FML (No. 284), 191 mg
each of six organics were injected into the 500 mL of water in the middle
compartment in Zone 4; the organics included n-octane, TCE, toluene,
o-xylene, n-butanol, and propionic acid. Table 2 presents data on these
organics. Due to the low solubility of n-octane in water, the amount of
this organic added to the water was excessive.
The five zones that contained either liquid or vapor were monitored
by sampling and analysis by GC. Exposure in the test apparatus was con-
tinued until essentially no change took place in the concentrations of
these chemicals in any given zone. At 28 days, the analyses indicated
that concentrations of the toluene and TCE had reached equilibrium, at
which time the test apparatus was dismantled and the FMLs were analyzed by
HSGC for the organics that they had absorbed. The conditions for the GC
analyses are presented in Table 1 (Set of Conditions 3). The results,
which include concentration data for 24 hours, 96 hours, and 28 days, are
presented in Table 8.
All of the organics, except the n-butanol, could be followed throughout
the exposure and appeared in different zones of the test apparatus in the
final analyses performed after 28 days of exposure. Seventy-five to 85%
of the aliphatic, aromatic, and chlorinated hydrocarbons, and 57% of the
propionic acid were accounted for in the final analyses. The n-butanol,
which was injected into Zone 4 at the beginning of the test, was not de-
tected after about a week and appeared to be replaced by n-butyric acid
which was detected by GC analysis. Based on a total conversion of n-butanol
to n-butyric acid, 11.5% of the n-butanol was accounted for as n-butyric
acid in the final analyses. A portion of the apparent loss of the organics
can probably be attributed to the large number of samples withdrawn from the
test apparatus during the 28 days of exposure for GC analysis to track the
organics.
The distribution and transport of each of the organics that were ini-
tially placed in the water in Zone 4 are discussed briefly in the following
subsections.
n-Octane
The n-octane was injected at a concentration substantially above its
solubility in water. It appeared within 24 hours in the airspace above the
spiked water and was substantially absorbed by the upper FML. In the final
analysis, essentially 75% of the spike was concentrated in the upper FML, a
small amount was detected in the lower FML, and the remainder was observed
in the airspaces, particularly in the upper airspace (Zone 1). Only a trace
amount was found in the liquid in the bottom compartment.
-160-
-------
TABLE 8. DISTRIBUTION OF ORGANICS FROM A DILUTE AQUEOUS SOLUTION IN TEST APPARATUS
CONSISTING OF THREE-COMPARTMENTS SEPARATED BY POLYETHYLENE FML BARRIERS
Zone
No. Description
1 Airspace above
"upper" FML
2 "Upper" poly-
ethylene FMLa
3 Airspace above
spiked water
4 Spiked water
5 "Lower" poly-
ethylene FMLa
6 Airspace below
"lower" FML
7 Bottom water
Volume
(mL) Organic
806 n-Octane
Trichloroethylene
Toluene
o-Xylene
Propionic acid
n-Butanol
7.02 n-Octane
Trichloroethylene
Toluene
o-Xylene
Propionic 'acid
n— Butanol
306 n-Octane
Trichloroethylene
Toluene
o-Xylene
Propionic acid
n— Butanol
500 n-Octane
Trichloroethylene
Toluene
o-Xylene
Propionic acid
'n-Butanol
n-Butyric acidb
7.02 n-Octane
Trichloroethylene
Toluene
o-Xylene
Propionic acid
n-Butanol
506
Trichloroethylene
Toluene
o-Xylene
Propionic acid
n-Butanol
300 n— Octane
Trichloroethylene
Toluene
o — Xyl ene
Propionic acid
n— Butanol
n-Butyric acid°
n-Octane
Trichloroethylene
Toluene
. o-Xylene
Propionic acid
'n-Butanol
n-Butyric acidb
Start of tes
it At 24 hours At 96 hours
Cpncen- Concen- Concen-
Amount tration Amount tration Amount tration
(mg) (mg L-l) (mg) (mg L-1) (mg) (mg L-1)
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 ' 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
191 382
191 382
191 382
191 382
191 382
191 382
0 0
.0 0
0 0
0 0
0 0
0 0
0 0
Of]
U
, 0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
191
191
191
191
191
'191
0
12.1 15 4.8 6
8.1 10 12.5 15.5
28.3 35.1 8.1 10.0
10.7 13.3 4.8 6
none none
none none
Analyzed by headspace GC
when the apparatus was
dismantled
20.8 68.0 18.7 61
12.2 40 4.90 16
11.3 37 3.21 10.5
2.1 6.8 1.04 3.4
none none
0.3 0.6 0.3 0.6
40.1 80 20.0 40
48.1 96 22.5 45
trace-: trace
106 212 102 204
140 , 280 110 220
none trace
Analyzed by headspace GC
when apparatus was
dismantled
0.4 0.8 2.6 5.2
0.3 0.6 3.8 6.5
0.03 0.015 0.126 0.25
none none
none none—:
0.15 0.5 4.23 14. T "
none — 0.468 1.56
none trace
none — ^-trace
At end of test ,
28 days
Concen-
Amount, tration,
(mg) (mg L"1)
2.58 3.2
19.3 24.0
5.64 7.0
4.19 5.2
none
126.7 18,051
37.8 5,885
49.3 7,024
76.0 10,829
none
0.98 3.2
6.23 20.1
' 2.14 7.0
1.59 5.2
none
0.3 0.6
17.1 34.2
15.0 30.0
0.5 1
99.7 199.5
none
20.75 41.25
2.60 370
47.60 6,780
61.15 8,712
76.04 10,833
none
none
1.29 '2.56
14.2 28.2
5.0 10.0
1.9 3.75
none
20.58 68.6
5.6 18.75
10.1 33.15
5.4 18.0
Total in all zonesc
155.03 (81.0)
162.8 (85.2)
143.8 (75.3)
160.2 (83.9)
109.8 (57.4)
0 -; (0)
26.15 (11.5)
aLinear low-density polyethylene FML (No. 284).
bn-Butanol appeared to have been oxidized to n-butyric acid which was detected in the GC as the amount of n-butanol detected
in the closed apparatus decreased.
cData in parentheses are the percent of initial organics accounted for in the final analysis.
-161-
-------
The n-octane did not migrate downwards into the lower FML and into the
airspace below the FML because of its low solubility in water. The absorp-
tion of the n-octane by the upper FML is attributed to its solubility in the
FML, both of which have essentially the same solubility parameter value.
Trichloroethylene
The spike of the trichloroethylene (TCE) was approximately 1/3 of the
solubility of TCE in water. Due to its volatility, the TCE was observed
quickly in the space above the spiked water in Zone 4 and also in the
airspace above the upper FML. In the final analysis, the TCE was present
in all zones, although approximately 50% of that measured was in the
FMLs; a substantial amount was transported into the liquid in the bottom
compartment.
These results show the tendency of the TCE to migrate and to be.trans-
ported through all phases and reflect the solubility of TCE in water as
well as in FMLs. At the end of the test, of the six organics in the study,
the TCE had the highest concentration in the water in the bottom compart-
ment. The 'results indicate that special efforts are needed to control this
organic.
Toluene
The amount of toluene injected into the water in Zone 4 was below its
solubility level in water. Toluene behaved similarly to the TCE, although it
did not migrate as rapidly. However, it was in the airspace above the upper
FML in essentially the same concentration as the TCE. As with the TCE, only
a small quantity of the toluene was present (0.3 mg) below the lower FML at
the end of 24 hours.
At the time of dismantling, the toluene was present in all seven zones,
with approximately 77% of the total in the two FML layers. As with the TCE,
it moved through all zones of the test apparatus. The results indicate
that special efforts are needed to control this organic.
o-Xylene
The amount of o-xylene injected into the water in Zone 4 was approxi-
mately twice its solubility level in water, i.e. 382 vs. 175 mg IT*, and it
was anticipated that the o-xylene would tend to migrate relatively rapidly.
The o-xylene appeared to have left Zone 4 quickly; it was detected at signi-
ficant concentrations (up to 13 mg L"1) in the airspaces after 24 hours;_
only a trace was detected in the water in Zone 4, and none in the water in
the bottom compartment.
In the final analyses, the o-xylene was concentrated principally in the
FMLs; almost all of it had migrated out of the spiked water. Also, the
-162-
-------
concentration in the airspaces was low in comparison with that in the FMLs,
which demonstrates the strong affinity of the o-xylene for the FML and
reflects its comparatively low solubility in water.
Propionic Acid
Propionic acid has a high solubility parameter value and is completely
miscible with water; therefore, it has a strong affinity for water. Only
57% of the propionic acid that was injected at the beginning of the experi-
ment was accounted for in the final analyses; approximately 75% of that was
still in the spiked water and the remainder was detected in the water in
the bottom compartment. No propionic acid was detected in the other zones
of the test apparatus.
n-Butanol
After the first few days of analyses, no n-butanol was observed in any
of the zones. However, n-butyric acid appeared in the analyses, indicating
that the n-butanol had been oxidized to n-butyric acid. No butyl propion-
ate appeared in the GC scan; thus, the possibility of esterification of the
n-butanol with the propionic acid could be discounted. Assuming that the
n-butanol was completely converted to n-butyric acid, 11.5% would have been
accounted for as n-butyric acid which was detected only in the aqueous
phases.
During the dismantling of the test apparatus, there was considerable
pressure in the bottom compartment, which may have been C02 generated from
the decomposition of the acids; C02 cannot be detected by flame-ionization
detection in the GC. Also, the bottom compartment fabricated of aluminum
exhibited some corrosion and darkening of the surface.
DISCUSSION
The experiments on the distribution of volatile organics from dilute
aqueous solutions show that, even at low concentrations, some organics can
be highly absorbed by a polymeric FML and can permeate the liner. Table 9
compares the swelling of an HOPE FML in a series of neat organics with the
swelling of the same FML in a series of saturated aqueous solutions of the
same organics. These results show that the amount of organic that an FML
can absorb during exposure to a saturated aqueous solution of an organic
can approach the amount an FML can absorb during exposure to a neat or-
ganic. Even though the maximum swelling values for the specimens exposed
to the neat organics were all somewhat higher, the maximum swelling values
obtained in the neat organics were approximately equal to those obtained
in the saturated dilute solutions.
-163-
-------
TABLE 9. COMPARISON OF THE SWELLING OF A POLYETHYLENE FMLa IMMERSED
IN NEAT ORGANICS AND IN SATURATED AQUEOUS SOLUTIONS OF ORGANICS
Saturated aqueous
solution
Organic
Acetone^
Methyl ethyl ketone
1,1, 1-Trichloroethane
Trichloroethylene
Benzene
Toluene
o-Xylene
m-Xylene
p-Xylene
Maximum
swelling ,
by weight
0.96
1.83
12.13
17.52
7.61
8.40
8.73
8.51
8.72
_ f-j—
Time to
maximum
swelling
(days)
18
21
17
15
6
6
9
6
7
Maximum
swelling,
by weight
0.33
0.62
11.70
16.71
7.22
8.05
8.42
8.19
8.43
Time to
maximum
swelling
(days)
44
46
52
43
30
50
40
38 .
49
&HDPE FML (NO. 483).
bAcetone is miscible with water in all proportions; the concentration
was set at 250,000 mg L"1 for the experiment with the aqueous
solution.
However, the results presented in Table 9 also show that the time to
reach maximum swelling values can be significantly affected by the concentra-
tion of the organic in the liquid in contact with the FML. The specimens
exposed to the neat organics reached maximum swelling values in considerably
shorter times than the specimens exposed to the saturated aqueous solutions.
The results of determining the distribution of organics between satu-
rated FML specimens and DI water indicate that the concentration of an or-
eanic in an in-service FML is affected by the concentration of the organic
in the liquid in direct contact with the FML and that the concentration_of
the organic in the FML will change as the concentration of the organic in
the liquid in contact with the FML changes.
-164-
-------
. The results of measuring the distribution of organics in the different
zones of a multi-compartment apparatus indicate that the amount and rate of
absorption and the transport of these species through a polymeric FML is a
function of such factors as relative solubility parameters of the FML and
the organic, crystallinity of the FML, and molecular weight and concentra-
tion of the organic constituent. A multi-compartment apparatus, such as
the one described and used in this work, appears to be an appropriate and
promising means of assessing the effectiveness of an FML to contain the
organic constituents of a given leachate.
REFERENCES
1. August, H., and R. Tatzky. Permeabilities of Commercially Available
Polymeric Liners for Hazardous Landfill Leachate Organic Constituents.
JEn: Proceedings of the International Conference on Geomembranes, June
20-24, 1984, Denver, CO. Vol. I. Industrial Fabrics Association In-
ternational, St. Paul, MN, 1984. pp 163-168.
2. Haxo, H. E. , and N. A. Nelson. Permeability Characteristics of Flexible
Membrane Liners Measured in Pouch Tests. In; Proceedings of the Tenth
Annual Research Symposium: Land Disposal of Hazardous Waste. EPA-600/
9-84-007, U.S. Environmental Protection Agency, Cincinnati, OH, 1984.
pp 230-251.
3. Haxo, H. E., J. A. Miedema, and N. A. Nelson. Permeability of Polymeric
Membrane Lining Materials. In; Proceedings of the International Con-
ference on Geomembranes, June 20-24, 1984, Denver, CO. Vol. I. In-
dustrial Fabrics Association International, St. Paul, MN. 1984. pp
151-156.
4. Haxo, H. E., J. A. Miedema, and N. A. Nelson. Permeability of
Polymeric Lining Materials for Waste Management Facilities. In: Migra-
tion of Gases, Liquids, and Solids in Elastomers. Education Symposium.
Fall Meeting - Denver, CO. Rubber Division, American Chemical Society.
The John H. Gifford Memorial Library and Information Center, The Uni-
versity of Akron, Akron, OH, 1984.
5. Matrecon, Inc. Lining of Waste Impoundment and Disposal Facilities.
SW-870 Revised, U.S. Environmental Protection Agency, Washington, D.C.,
1983. 448 pp. GPO #05500000231-2.
6. Leo, A., C. Hansch, and D. Elkins. Partition Coefficients and Their
Uses. Chemical Reviews 71(6):525-554, 1971.
7. Neely, W. B., D. R. Branson, and G. E. Blau. Partition Coefficient
to Measure Bioconcentration of Potential Organic Chemicals in Fish.
Environ. Sci. Technol. 8(13):1113-1115, 1974.
-165-
-------
Chiou, C. T. Partition Coefficient and Water Solubility in Environ-
mental Chemistry. In: J. Saxena and F. Fisher (eds.), Hazard Assessment
of Chemicals, Current Developments. Vol. 1. Academic Press, NY, 1981.
pp 117-153.
Haxo, H. E., R. S. Haxo, N. A. Nelson, P. D. Haxo, R. M. White, and S.
Dakessian. Liner Materials Exposed to Hazardous and Toxic Wastes.
EPA-600/2-84-169, U.S. Environmental Protection Agency, Cincinnati, OH,
1985. NTIS No. PB 85-121-333.
Haxo, H. E., N. A. Nelson, and J. A. Miedema. Solubility Parameters for
Predicting Membrane-Waste Liquid Compatibility. In: Proceedings of the
Eleventh Annual Research Symposium: Land Disposal of Hazardous Waste.
EPA/600/9-85/013, U.S. Environmental Protection Agency, Cincinnati, OH,
1985. pp 198-212.
11. Barton, A.F.M. Solubility Parameters and Other Cohesion Parameters
Handbook. CRC Press, Boca Raton, FL, 1983. 594 pp.
8.
9.
10.
12.
13.
Riddick, J., and W. Bunger. Techniques of Chemistry. Volume II -
Organic Solvents, Physical Properties and Methods of Purification.
Wiley-Interscience, NY, 1970.
Haxo, H. E., T. P. Lahey, and M. L. Rosenberg. Factors in Assessing
the Compatibility of FMLs and Waste Liquids. Contract No. 68-03-3213,
U.S. Environmental Protection Agency, Cincinnati, OH, 1987. 143 pp.
(In press).
-166-
-------
THE ELECTRICAL LEAK LOCATION METHOD FOR
GEOMEMBRANE LINERS
by: Glenn T. Darilek
Jorge 0. Parra
Southwest Research Institute
San Antonio, Texas 78284
ABSTRACT
An electrical method for locating leaks in geomembrane liners was devel-
oped and demonstrated for a wide variety of applications. Georaerabrane liners
are sheets of elastomer!c material used to prevent the leakage of waste and
to prevent rainwater from infiltrating landfills and surface impoundments.
When no leaks are present, a voltage applied between the material in the
liner and the earth under the liner produces a relatively uniform electrical
potential distribution in the material in the liner. Leaks are located by
mapping any anomaly in the potential distribution caused by current flowing
through a leak. A computer simulation model of layered earth sequences above
and below an insulating liner with a leak was developed to predict effi-
ciently the effect of a wide range of parameters on the leak signature.
Tests on a double-lined physical model demonstrated the applicability of
the method to a variety of drainage layers with various test parameters such
as leak size, electrode depth, and protective cover soil. Leaks with a diam-
eter smaller than 0.8 millimeters in the primary liner were reliably located
to within 10 millimeters. Leaks in the bottom liner could be detected, but
not located. The electrical leak location method was successful in finding
leaks in a full-scale impoundment that had been fully tested using the vacuum
box method.
The electrical leak location method was adapted for locating leaks in
the geomembrane liner of landfill cover systems. Scale model tests demon-
strated the applicability of the method under a wide range of cover soil
thicknesses and leak sizes. Special non-polarizing electrodes were used to
locate leaks with diameters as small as 3 millimeters under 600 millimeters
of cover soil.
-167-
-------
INTRODUCTION
The most common method of disposal of solid and hazardous wastes is in
landfills and surface impoundments. To prevent contamination, double geomem-
brane liner systems are installed beneath the landfill or impoundment to form
an essentially impermeable barrier that prevents the migration of contaminant
liquids. Geomembrane liner material is used for landfill final cover systems
to prevent rainwater from percolating through the waste and leaching chemi-
cals that could migrate into groundwater or surface water.
Installation practices and operational factors can result in leaks in
the form of punctures or separated seams. An electrical leak location method
was developed to effectively locate leaks in geomembrane liners to ensure
that liners have been installed and seamed properly and that the integrity of
the containment system is maintained.
The electrical leak location method can be used in liquid impoundments,
as a pre-service inspection of solid waste landfills, and to locate leaks in
the final cover for landfills or impoundments. The method will not damage
the liner.
TECHNICAL DISCUSSION
METHOD
Figure 1 shows the basic electrical leak location-method for detecting
and locating leaks in a geomembrane liner. The leak location method makes
use of the high electrical resistivity of the geomembrane liner material.
When no leaks are present, a voltage impressed across the liner produces a
relatively uniform voltage potential distribution in the material above the
liner. If the liner is punctured or separated physically, conductive fluid
flows through the leak establishing a conductive path for current flow, which
produces an anomaly in the measured potential in the vicinity of the leak.
REMOTE
CURRENT
RETURN
ELECTRODE
CURRENT SOURCE
ELECTRODE
MOVING
MEASUREMENT
ELECTRODES
LIQUID
LEAK
PATH
MEMBRANE
LINER
CURRENT
FLOW LINES
Figure 1. Diagram of the electrical leak location method.
-168-
-------
Therefore, leaks can be located by measuring the potential distribution pat-
terns in the material covering the liner.
COMPUTER SIMULATION MODEL
A computer model was developed to investigate the performance capabili-
ties of the electrical leak location method. The model can accommodate
various electrical and dimensional parameters in the three layers comprising
the lined impoundment or landfill. The electrical anomaly of a circular hole
in a thin, highly resistive layer was used to model the response of a geomem-
brane lined impoundment or landfill containing a damaged geomembrane liner.
The secondary potential for a leak in a geomembrane liner is in the form of
an integral equation, which includes a three-layer medium Green's function.*
The excellent agreement between experimental and synthetic model data veri-
fied the accuracy of the general solution for predicting leak signatures.
Model studies of the electrical leak detection survey technique were
made to characterize the performance of the method under various conditions
of the electrical parameters of the waste materials, the measurement elec-
trode array, the measurement dipole depth and proximity to the leak, the size
and number of leaks, and the impoundment depth. Figure 2 shows a typical
20
15
10
Key:
s
h
PW
PS
a.
electrode spacing
depth of the water
liquid resistivity
underlying soil
resistivity
leak radius
electrode depth
offset distance
ui
O 0
§
s -5
ui
O
Q.
Q-io
-15
s
h
Pw
Ps
a
2m
1m
15 Om
3011 m
0.0004m
zm = 0.9m
4 6 8 10 12 14 16
HORIZONTAL SCAN DISTANCE, y (m)
18
20
Figure 2. Plot of the leak anomaly versus electrode depth.
*A Green's function is the solution to the Laplace's equation with a unit
current source. The specific Green's function for this solution is found in
Darilek and Parra, March 1988, The Electrical Leak Location Method for Geo-
membrane Liners: Final Technical Report, EPA Contract No. CR-811771-01-3.
-169-
-------
family of leak anomaly responses illustrating the effects of various measure-
ment depths for a single leak located in a liquid waste impoundment. A sub-
stantial improvement in detection sensitivity is obtained when the potential
array is closer to the leak. The peak-to-peak anomaly amplitudes for dif-
ferent waste layer resistivity values were calculated. When a constant cur-
rent is injected, the leak detectability is increased linearly with the
resistivity of the waste material.
Figure 3 shows the peak-to-peak anomaly responses calculated for various
dipole offset distances from the leak center as a function of the survey
height above the liner. An improvement in leak detectability is observed for
survey lines located within a radius of 10 cm from the leak center when the
depth of the detector is increased.
h-zm
VT
LINER
x - 0.05 m
x = 0.08 m
x = 0.1 m
x • 0.2 m
x = 0.3 m
= 1 m
= 1 m
= 15 Urn
= 30 Jim
= 0.0004 m
0.01
0.05
0.25
Figure
0.1 0.15 0.2
HEIGHT ABOVE LINER, h - zm (m)
3. Peak-to-peak voltage anomaly versus electrode scan offset
distance.
0.3
Figure 4 shows that the horizontal dipole response is stronger than the
vertical dipole response. Figure 5 shows a typical vertical dipole anomaly
response of a leak. The leak is directly associated with the maximum anomaly
response. For practical reasons, leaks may be located more easily and accu-
rately with a vertical dipole because the leak is located at the peak of the
unipolar response rather than the bipolar anomaly with the horizontal dipole.
Multiple leaks can be resolved with less ambiguity when a vertical dipole is
used.
Leak signatures for various multiple leak geometries were computed.
Figure 6 shows horizontal dipole leak signatures computed for two survey
-170-
-------
200
= 1m
= 15m
= 30m
= sv = 0.30m
= 0.0004m
h-zn
LEAK
LINER
HORIZONTAL DIPOLE RESPONSE
VERTICAL DIPOLE RESPONSE
0.1 0.15 0.2
HEIGHT ABOVE LINER, h - zm (m)
0.25
Figure 4. Comparison of leak anomaly amplitudes for horizontal and vertical
measurement dipoles.
a.
O
LU
C9
§
5
Ul
O
0.
O
20
18
16
14
12
10
8
6
4
2
0
-2
! 1 ' i ' 1 ' 1 i ' i 1 i 1 1 1 1 1
h = 1m I ' -
sv = 0.3m
Pw = 15 ft-m I
Pa = 30O-m || ,
_ a = 0.0004m
;
-
-
'
.* — zm = 0.95 m
-
* zm = 0.9 m
•
l\ • •-
/ L
' J . 1 , ! , I . I . l . i . i . i .
18
20
0 2 4 6 8 10 12 14 16
HORIZONTAL SCAN DISTANCE, y (m)
Figure 5. Leak anomaly characteristic of a vertical dipole.
-171-
-------
0.
I
O
01
O
a
O
a.
5
HORIZONTAL SCAN DISTANCE, y (m)
Figure 6. Leak anomaly for multiple leaks.
depths when the leaks are spaced two meters apart. As expected, when the
horizontal separation between leaks becomes less than the horizontal dipole
spacing, separate resolution of the two leaks is lost. When leaks are
located at separations approximating the horizontal dipole detector spacing,
the resolution is poor. However, when measurements are acquired using a
small dipole detector spacing, the resolution is improved.
The derived geomembrane leak detection model can be used to aid in
planning surveys and processing leak survey data acquired in lined impound-
ments or landfills. The computed leak responses point out the practical
importance of performing the survey measurements near the bottom of the
impoundment. The results also indicate that the horizontal dipole detector
spacing must be less than the leak separation or a vertical dipole must be
used to improve leak resolution. The injected current must be increased to
offset the effect of lower measured leak anomaly attributed to lower resis-
tivity of the liquid.
DOUBLE LINER MODEL TESTS
By placing the current return electrode in electrical contact with the
liquid-saturated drainage layer located between the two liners, the electri-
cal leak location method is applicable for detecting and locating leaks in
the upper liner. Simple electrical continuity tests between the drainage
layer and the earth can determine the existence of leaks in the bottom liner,
but not the location of that leak.
-172-
-------
A scale model with dimensions of 3 m x 3 m was used to test the electri-
cal leak location method for locating leaks with various impoundment config-
urations, including different types of drainage layers, various types of
leaks, and a protective soil cover over the primary liner. Figure 7 is a
contour plot of the data for a leak with a diameter of 5.1 mm with a drainage
layer consisting of a sandy loam soil layer placed over geotextile mat, which
is then placed over the.geonet material. The location of the leak is clearly
indicated by the dipolar contour pattern. The potential gradient pattern
caused by the current injection electrode is also evident in the data. Other
tests indicated that a leak with a diameter of 25 mm and a 15-cm slit leak
produce anomaly characteristics very similar to the leak with a diameter of
5.1 mm. However, the larger leaks required less voltage to produce the same
anomaly amplitude.
N
CURRENT ELECTRODE
Figure 7. Contour plot of double liner model data.
Tests were conducted using various electrode materials and geome-
tries to determine the best and most practical electrode configurations for
electrical leak location surveys in liquid-filled impoundments. Figure 8
shows the relative leak anomaly amplitudes for various electrodes when the
centerlines of the electrodes were scanned directly over the leak, and 15 cm
offset from the leak. When the electrodes were scanned directly over the
leak, the anomaly amplitudes were inversely related to the length of the
electrodes. However, when the electrodes were scanned along a line 15 cm
from the leak, the 30-cm line electrode produced the largest anomaly. More
importantly, the leak anomaly was barely detected when the localized point
electrodes passed within 15 cm of the leak, where the longer electrodes pro-
duced easily detectable anomalies.
-173-
-------
til
Q
o.
<
111
1
UJ
EC
TYPE OF
ELECTRODE
100-
90-
80-
70-
60-
50-
40-
30-
20-
10-
XV
''/
'<<
'*,*"'"•
'',-<•
#;
p-'^_
',?•*
:"''
X** '
"•**
•. '*•»
«-;
s .-^
s
•/
.'
/
•f
J''r
••'
?> ,
S
s'
x;/
%
%
'//•
''//
S s'
x"'-
X/
:?<
^
/x
'xV
1. V 1
^_
xXl
'/,
'//
|:
'x*'-1
*" X
;^
>x
v^
$'j
•$-'
x \.'t
^:
vO
NX
'§
X vi
\v.
•-. V
^
s}\
X S
SN>;
x>
;<•;
x^ •
^%"
§
:^-
x\
••S
' f " ,
S s
' S
'/•.'
/^.-
.' X
//;
//
'Ss
-z^
^-^'
CARBON 15-CM 60-CM
POINT LINE LINE
STAINLESS
STEEL
POINT
pH
ELECTRODE
30-CM
LINE
PROTECTED
25-CM
LINE
SCAN OVER LEAK
SCAN OFFSET 15 CM
Figure 8. Relative anomaly amplitudes for various measurement electrodes.
LOCATING LEAKS IN COVER SYSTEMS
The electrical leak location method was adapted to make surface soil
potential measurements to locate leaks in final cover system geomembrane
liners. Half-cell electrodes were used to reduce the polarization noise
caused by electrochemical reactions at the interface between the soil and
metal electrodes. These electrodes typically consist of a plastic tube with
a porous ceramic tip. Electrical contact is made through a metal electrode
in a saturated salt solution in the half-cell.
Experiments were conducted using a physical model with dimensions of 5 m
x 5 m. Figure 9 is a plot of the measured leak anomaly for several soil
cover thicknesses. Although the peak-to-peak amplitude of the anomaly
decreases rapidly with increasing soil cover, the leak was easily detected
for all of the soil cover depths tested. Tests were performed with 60 cm of
soil cover to show that electrode contact noise is reduced significantly when
the electrodes are inserted in the ground to a depth of approximately 25 mm
or when the dry ground surface is scraped off prior to the measurements.
Other methods for detecting leaks in the geomembrane liner of cover sys-
tems, including ground-penetrating radar, tracer gas, the electromagnetic
induction method, encapsulated chemicals, and electronic transponders, were
analyzed. The most promising method studied for locating leaks in final
cover systems, other than the electrical leak location method, is ground-
penetrating radar. Limited testing using infrared imaging was unsuccessful
in detecting localized areas of low thermal conductivity caused by drainage
of soil moisture through a leak.
-174-
-------
LEAK DIAMETER = 3 MM
0123
DISTANCE (METERS)
+ 15.2 CM SOIL 0 25.4 CM SOIL A 30.5 CM SOIL x 61 CM SOIL
Figure 9. Measured leak anomaly for various soil cover thicknesses.
LINER RESISTIVITY TESTS
Tests were conducted to measure electrical resistance changes in liners
over a period of time to determine whether the electrical resistance of the
liner materials changes after exposure to waste liquids, thereby reducing the
usefulness of the survey technique. The tests were performed in triplicate
using polyvinyl chloride, high-density polyethylene, two thicknesses of
chlorosulfonated polyethylene, and chlorinated polyethylene. The liquids
used in the tests included sodium hydroxide solution, pH of 10; sulfuric acid
solution, pH of 1; sodium chloride solution, 10 percent by weight; and
deionized water.
The laboratory tests indicated that there was no significant decrease in
the resistivity of the liner materials during a 13-week exposure. Exposure
of these typical liner materials to these chemicals had no effect on the
applicability of the electrical leak location method.
FIELD DEMONSTRATION SURVEYS
Full-scale surveys at the Southwest Research Institute test impoundment
were performed to detect and locate four small circular leaks, each 0.79 mm
-175-
-------
in diameter. The impoundment was filled with water to a depth of approxi-
mately 46 cm and the measurement electrodes were immersed at a depth of 30
cm. The contour plot of the data shown in Figure 10 graphically indicates
the locations of the four leaks.
15 20
METERS
Figure 10. Contour plot of data at a full-scale impoundment.
The electrical leak location method was demonstrated at another full-
scale impoundment in the San Antonio, Texas area. Although the complete
liner had been tested previously using the vacuum box method, a 2.0-cm long
leak was found. The characteristic leak anomaly was clearly evident on scan
lines as far away from the leak as 1.5 m, and no false indications were
obtained.
CONCLUSIONS
An electrical method for detecting and locating leaks in geomembrane
liners for hazardous waste impoundments and landfills has been developed and
demonstrated successfully in a wide variety of applications. The project
demonstrates the validity and usefulness of the electrical leak location
method for testing the integrity of the geomembrane for single and double
liners and final cover systems. The technique is cost effective for con-
struction quality assurance and in-service performance monitoring. Addi-
tional development will bring the method into application for hazardous mate-
rial impoundments and for final cover systems.
-176-
-------
SUPERFUND STANDARD ANALYTICAL REFERENCE MATRIX PREPARATION
AND RESULTS OF PHYSICAL SOILS WASHING EXPERIMENTS'
by: M. Pat Esposito*, Barbara Bruce Locke, and Jack Greber
PEI Associates, Inc.
Cincinnati, Ohio 45246
and
Richard P. Traver
U.S. EPA, HWERL
ABSTRACT
In response to the RCRA Hazardous and Solid Waste Amendments of 1984 pro-
hibiting the continued land disposal of untreated hazardous wastes, .the EPA
has instituted a research program for establishing best demonstrated and
available technologies for RCRA and Superfund wastes. Under Phase I of EPA's
Superfund research program, several projects were initiated under which a
surrogate soil containing a wide range of chemical contaminants was prepared
for use in bench-scale and pilot-scale performance evaluations of five dif-
ferent treatment technologies. This paper covers one of the projects in which
the surrogate test soil was developed and bench-scale soil washing treatabi-
lity studies were completed. This work was conducted by PEI Associates under
EPA Contract No. 68-03-3413 during 1987. This paper,has.been reviewed in
accordance with the U.S. Environmental Protection Agency's peer and admini-
strative review policies and approved for presentation and publication.
Formerly with PEI, now with Bruck, Hartman & Esposito, Inc., Cincinnati
Ohio.
-177-
-------
INTRODUCTION
The RCRA Hazardous and Solid Waste Amendments of 1984 prohibit the con-
tinued land disposal of untreated hazardous wastes beyond specified dates.
The statute requires the U.S. Environmental Protection Agency (EPA) to set
"levels or methods of treatment, if any, which substantially diminish the tox-
icity of the waste or substantially reduce the likelihood of migration of
hazardous constituents from the waste so that short-term and long-term threats
to human health and the environment are minimized." The legislation sets
forth a series of deadlines beyond which further disposal of particular waste
types is prohibited if the Agency has not set treatment standards under Sec-
tion 3004 (m) or determined, based on a case-specific petition, that no further
migration of hazardous constituents will occur for as long as the wastes
remain hazardous.
In addition to addressing future land disposal of specific listed wastes,
the RCRA land disposal restrictions also address the disposal of soil and
debris from CERCLA site response actions. Sections 3004 (d) (3) and (e) V) or
RCRA state that the soil/debris waste material resulting from a Superfund-
financed response action or an enforcement authority response .action imple-
mented under Sections 104 and 106 of CERCLA, respectively, will not be subject
to the land ban until November 8, 1988.
Because Superfund soil/debris waste often differs, significantly from
other types of hazardous waste, the EPA is developing specific RCRA Section
3004 (m) standards or levels applying to the treatment of these wastes. These
standards will be developed through the evaluation of best demonstrated and
available technologies (BDAT) . In the future, Superfund wastes xn compliance
with these regulations may be deposited in land disposal units; wastes exceed-
ing these levels will be banned from land disposal unless a variance is is-
sued.
In early 1987, EPA's Hazardous Waste Engineering Research Laboratory, at
the request of the Office of Solid Waste, initiated a research program _ to
evaluate various treatment technologies for contaminated soil and debris from
Superfund sites. Under Phase I of this research program, which was conducted
from April to November 1987, a surrogate soil containing a' wide range of
chemical contaminants typically occurring at Superfund sites was prepared for
use across the board in the bench-scale or pilot-scale performance evaluations
of five available treatment technologies: 1) soil washing, 2) chemical treat-
ment (KPEG), 3) thermal desorption, 4) incineration, and 5) stabilization/fix-
ation. This report covers those segments of Phase I related to development of
the surrogate soil and experimental bench-scale tests on the potential effec-
tiveness of physical soil washing as a treatment technology.
PROCEDURES
SARM PREPARATION
The surrogate soil is referred to throughout the text as SARM, an acronym
for Synthetic Analytical Reference Matrix. More than 30,000 pounds of clean
(uncontaminated) SARM was prepared after considerable research into the types
of soils found at Superfund sites nationwide. The final composition selected
-178-
-------
consisted of 30 percent by volume clay (a mixture of montmorillinite and
kaolinite), 25 percent silt, 20 percent sand, 20 percent topsoil, and 5 per-
,cent gravel. The components were air-dried to minimize moisture and then
mixed together in two 15,000-lb batches in a standard truck-mounted 6-yd3
cement/concrete mixer. '
A prescribed list of chemicals found to be widely and frequently oc-
curring at Superfund sites was then added to the clean SARM in a series of
smaller-scale mixing operations utilizing a 15-ft3 mortar mixer. The organic
chemicals added included ethyl benzene, 1,2-dichloroethane, tetrachloroethyl-
ene, acetone, chlorobenzene, styrene, xylene, anthracene, pentachlorophenol,
and bis(2-ethylhexyl) phthalate. Salts or oxides of the following metals were
also added: lead, zinc, cadmium, arsenic, copper, chromium, and nickel.
Because concentrations of contaminants in soils vary widely, four different
SARM formulas containing either high or low levels of organics and metals were
prepared for use in subsequent treatability tests using the five technologies
named. Table 1 presents the target contaminant concentration of the four
SARMs prepared. Reserves of each SARM were also packaged and archived for
future use. The archived samples are being stored at EPA's R&D facility in
Edison, New Jersey.
PHYSICAL SOIL WASHING EXPERIMENTS
As part of the performance evaluation of soil washing as a potential
treatment candidate, samples of each SARM were physically washed in a series
of bench-scale experiments designed to simulate the EPA-developed Mobile Soils
Washing System (MSWS). This system can extract certain contaminants from
soils, which reduces the volume of the contaminated portion of the soils. The
MSWS is expected to be an economic alternative to the current practice of
hauling contaminated soils offsite to a landfill and replacing the excavated
volume with fresh soils.
Specifically, this project was designed to simulate the drum screen
washer segment of the MSWS as described by J.S. Shum in the Operation and
Maintenance Manual(1). This segment of the MSWS separates the +2 mm soil
fraction from the -2 mm soil fraction (fines) by use of a rotary drum screen.
A high-pressure water knife operates at the head of the system to break up
soil lumps and strip the contaminants off the soil particles. Both the design
of the MSWS and the design of the bench-scale experiments to simulate the MSWS
for cleanup of the SARMS samples are based on the following assumptions, which
underlie the volume reduction approach of physical soils washing:
1. A significant fraction of the contaminants are attached to the silt,
humus, and clay particles.
2. The silt and clay are attached to the sand and gravel by physical
processes (primarily compaction/adhesion).
3. Physical washing of the sand/gravel/rock fraction will effectively
remove the fine sand, silt, and clay-sized (less than 0.25 mm)
materials from the coarse material.
-179-
-------
TABLE 1. TARGET CONTAMINANT CONCENTRATIONS FOR SARMS
(mg/kg)
SARM I
(High
SARM II
SARM III
SARM IV
nraanic (Low organic, (Low organic, (High organic,
lowTelal) lowmftal) high metal) high metal)
Volatiles
Acetone
Chlorobenzene
1,2-Dichloroethane
Ethyl benzene
Styrene
Tetrachloroethylene
Xylene
Semivolatiles
Anthracene
Bis(2-ethylhexyl)
phthalate
Pentachlorophenol
Inorganics
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
6,800
400
600
3,200
1,000
600
8,200
680
40
60
320
100
60
820
680
40
60
320
100
60
820
6,800
400
600
3,200
1,000
600
8,200
6,500
2,500
1,000
650
250
100
650
250
100
6,500
2,500
1,000
10
20
30
190
280
20
450
10
20
30
190
280
20
450
500
1,000
1,500
9,500
14,000
1,000
22,500
500
1,000
1,500
9,500
14,000
1,000
22,500
-
-180-
-------
4. The contaminants will be removed to the same extent that the silt
and clay are separated (i.e., increasing the efficiency of the
washing process will directly increase the removal efficiency for
the majority of the contaminant mix).
versene 100 ), and 2) an anionic surfactant solution (phosphated formulafi™
from Procter and Gamble, Institutional Formula Tide5) Deferent pH and
v:nCr:ndroe Td±ti°nS Were evaluated *>r th.e wash solutions ^^^0!-
vents and oxidizing agents were considered, but were found not to be viable
soil-washing solutions because of material handling problems associated with
P°te"tlal-
^termination of the
of
..s.
leachable constituents by Toxicity Characteristic Leaching ProceiuJw (?CLP) .
RESULTS
SARM PREPARATION
Table r rLf Phy!±Cal ,tests ^ducted on the clean SARM are summarized in
tic of a sl?^7 ?/?S S ^dlf te that thC s^thetic soil is characteris-
tic of a slightly alkaline sandy loam with moderate clay and organic content
nated ^^ **** ^^ ^^^ CapaClty' Such a soil> wSn coSami-
nolog^. PreSe a reasonable challenge to any applied treatment tech-
Chemical analyses of samples of the four SARMs were conducted before
treatment to verify contaminant levels and moisture content. Sle 3 contains
SAlMr"!!^ C0n"ntrations obtai^d for each analyte in each of the four
SARMs. All numbers reported by each laboratory conducting the analyses
" " •
7^ i co^taminant levels (Table 1) are compared to the actual
levels found (Table 3) , the recovery efficiencies obtained are the highest and
most consistent for the metals, followed by the volatiles and the semfvola-
' Gene^lly> the SARMs containing the higher concentrations of volatiles
f-iwoi,"; i-h «- +-u "i~ " >-•&•» vuj.ai.j.ij.^aLj.on, adsorption), or alterna-
reliably d^, ^ 7** COncentrations of the organics were more difficult to
reliably detect and quantitate.
-181-
-------
TABLE 2. PHYSICAL CHARACTERISTICS OF CLEAN SARM
-a Range
Cation exchange capacity, meq/100 g
Total organic carbon, %
pH
Grain size distribution, weight %
Gravel (>4.75 mm)
Sand (4.75 mm - 0.075 mm)
Silt (0.074 mm to 0.005 mm)
Clay (<0.005 mm)
Average
132.7
(10)
3.2
(6)
8.5
(6)
3
(6)
56
(6)
28
(6)
12
(6)
77.5 to 155
2.7 to 3.9
8.0 to 9.0
2 to 4
54 to 58
27 to 30
11 to 14
a Values in parentheses indicate number of samples analyzed.
-182-
-------
TABLE 3. ANALYTICAL PROFILE OF SARMS: AVERAGE CONCENTRATION
FOUND UPON ANALYSIS3
(mg/kg)
Analyte
SARM I
SARM II
SARM III
SARM IV
(High
organic, (Low organic,
low metal) low metal)
(Low organic, (High organic,
high metal) high metal)
Volatiles
Acetone
Chlorobenzene
1,2-Dichloroethane
Ethyl benzene
Styrene
Tetrachl oroethyl ene
Xylene
Semivolatiles
Anthracene
Bis(2-ethylhexyl)
phthalate
Pentachlorophenol
Metals
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
Moisture, %
4,353 9)
316 9)
354 9)
3,329 (9)
707 (9)
408 (9)
5,555 (9)
5,361 (9)
1,958 (9)
254 (9)
18 (10)
22 (8)
24 (8)
231 (10)
236 (10)
32 (10)
484 (8)
20 (7)
356 8)
13 6)
7 8)
123 (8)
42 (8)
19 (8)
210 (8)
353 (7)
117 (7)
22 (7)
17 (7)
29 (6)
28 (6)
257 (8)
303 (8)
38 (8)
642 (6)
11 (7)
358
11
5
144
32
20
325
2)
1}
2)
2)
2)
(2)
181 (3)
114 (3)
30 (3)
652
2,260
1,207
9,082 (
14,318 (
1,489 (
31,871 (
4)
2)
4)
4)
4!
4)
4)
19 (3)
8,030 (2)
330 2)
490 2)
2,708 2)
630 2)
902 2)
.5,576 (2)
1,920 (3)
646 (3)
80 (3)
500 (4)
3,631 (2)
1,314 (4)
10,503 (4)
14,748 (4)
1,479 (4)
27,060 (4)
26 (2)
Values in parentheses indicate number of samples analyzed.
-183-
-------
PHYSICAL SOIL WASHING EXPERIMENTS
During the initial phase of these experiments, pH and temperature varia-
the PH of the chelant solution from its natural PH of 12 to 8.0 produced no
additional metal removal.
Reaction times of 5, 15 and 30 minutes were evaluated in a series of
scale-up operations
Next, surfactant concentrations of 1.5, 0.5, and 0 1 percen t^y weight)
removal tests.
-184-
-------
2000
•Q- Arsenic a
••- Cadmium
•*• Chromium
-*- Copper
•*- Lead
•o- Nickel
•*- Zinc
Time, mln.
Arsenic and nickel overlap in this figure.
Figure 1. Reaction time -1:1 molar chelant wash, SARM 111
TOX
10 20
Time, min.
a Total organic halogens
Figure 2. Reaction time - 0.1% surfactant wash, SARM
-185-
30
-------
to
O
V*
in
VI
in
*-< to o 10 en o
- cv;
co
o o ro in co
• o -
So
o
r- 10
CM •-<
t-lCvJCO - '
• • •
cn vo <£» •
eo •-« o co z o *r
o in r» o
ro *-< eo •-* CM o
csio O cs) o O •»•
CM v
§(
<
in
ID
en
en
*£. S
0.
ex
UL)
o:
CD
to
o
to
UI
CO
o o
So
IO
en
en
A
en
in
CO O O IO »-* CO ^
in »-« P> •«• o 10 •-« o
r-4 CO *T O 00
in cn o eo «-i •-< en in
'omr^ ^
t-l ««• OV
co co fo r>. co u-> ID
XS ca «*• oo o co CNJ
CO CO
CM —
**
CO
eo ro en
CM CM CO C3 CO
o o —«a o n
• • • o • •
oooozoo
«-< v
co \o *r en r» co in
o
en
cn
A
o CM
o en
CSJ
in O
vo «T
10
•NJ
10
vo
co
en
^•^ CO *^
vo In m
ro cn CM
co ro vo m co oo
*-* c\J o co
i—I CO CM
o ro in u> •-« CM co
to r^* *-« O
eo CM «tr »-* 10 CM ^r
SJ ca CM co co ro co
CM
en
c
o
c
o
u
cj 01 eg
O
cu
N
c •
OJ
CD XI
C O
O U
4-> O
f
u
eo
cn
I.
O
a
*~
o
>
1
oi
4) +•> o in •<-
x: «-^4-> •*-* c
c *^- o> 41 *o en
u
3
XI
£
CO
0)
cn
u
o
6
E 3
3 -f- U
-------
oo
_J
I—
UJ
2.
•
3»
O
oo
o
^^
§
o
o
•—4
I-H
OO— -^
Q.
.. CL
oo —
rs
00
UJ
C£.
CD
»— i
i
1
OO
*g
t-H
oo
•
in
UJ
•_j
CO
J.
1/1
10
*J
5
4J
U
"0
**-
!_
a
in
d
0.
i
c
«9
£
.c
u
i-
f«5
£
ro
.c
in
§
L.
CU
ro
re
c
E
O
«^k
CM
V
|
CM
C
0
in
CM
PJ
CM
A
E
;=.
o
irt
CM
V
M
CNJ
O
4J
E
g
CM
CM
A
E
O
in
CM
V
1
CM
O
+J
E
g
CM
CM
A
•°c
$.
!°
L.
C
U
c
o
u
Contaminant
(MOO O
CO ^ cs vo CM ro co
•-< z z d d d <-<
V
CM *»• in ro o
in o o -H — i ^j-
r~ o o o o o
O O O O O Z O
V
i CM en o co
10 0 O — « CM
«• o o o o
• .0. • .0 .
o o z o o z o
o ro en
CM O CO CM O *
r*» O O t— i ~H o **•
CM O O O O O O
vo ro oo vo <•
o o in vo o o
CM O O O O O CM
«-« o o oo oo
V
«r in vo <-<
COO 0 O CM
mo o o o
' • • O • • O •
o o z o o z o
o ro **• in CM CM
in CM <-H CM CM in
o o z o o o o
V V
ro «»• CM *»•
—« »-4 o OD ro o •— i
ro O O O *-" O ro
O O O O O O O
V V
CM ^ VO O
So 1-1 « «r
oo o o o
o o s: o o z o
vo ro r». ro CM en o
in I-H oj rr — i — <
ro 1-1 CM
c
C CU
(O ^-
53 £ £
•i- cu eu eu *J
C C O C CU
|Q CU I- CU O
en MOM s_
I- C i— C O
O CU JI CU •—
w ja o xs cu J=
CU CO"- C U CU
•— O I. O>— CU o i >,£.££
4J CU — CM JT >, *-> r—
•Z *<->•— UII/1H-X
o
*
s
en
en
X
en
en
en
X
en
en
en
fr<
vo
en
en
CO
en
en
A
en
en
en
A
CO
en
en
A
en
en
en
A
en
en
en
u
otal volatile organ
reduction
h-
82
P^ 9—t
O CM
CM ro
vo
«-< CO
CM
ss
VO CM
O «3-
ru ^
CO
CO O
OO
ro p~
oo ro
§*
CM
ro r~-
CM
ro r-
vn —
ro—<
u>
emivolatile organic
Anthracene
Bis(2-ethylhexyl)
phthalate
V)
o
z
CO
^
«
CM
O
t-"
in
a
z
vo
**"
oo
vo
§
CM
CM
Pentachlorophenol
ef
z
m
vo
S
ro
en
z
ro
r-T
o
en
Ue|
X
CM
in
X
en
ro
en
otal semivolatile
organic reduction
i-
oo r~- 10 oo
t** f*» VO CO t-* i-4 VO
CM co in r-. —4 «a- o
win en
vo «r in vo co co
ro en ro co <-i vo CM
CM ro •-"
9-4
O co r« o in CM co
ro v CM en co ro in
CM
ID in r-. o
CM t«. en oo o ro vo
^^ vo ro 9—t *cf *f
CM -H in
V O •»• vo vo o vo
•-« i— 1 VO
en o vo CM CM CM ro
ro CM 1-1 co vo «sr co
CM
CO VO «J- VO
^r in o CM o co o
CM in en m — < vo co
vo r«. ro
CM CM O «T O CM
«>• o «»• in en r- t~.
—I CM VO O
CO
m o co o o o o
CM vo o in «»• v —i
V CM
r«. en oo r- ro CO CM
»-i CM CM in o ro «s-
CM ro vo
in E
e c — E cu "a!
it) , CU E O Q.-O Jf '3
cu -o
CU *O CU
ro o o
(- c c
§ " "
u o o:
u. z z
9 JO U
-187-
-------
CO
CO
o
S
in
coojca
co
*S. E
CO Q.
Q.
CO
CO
Ul
O
to ro in co
^O CM O O m CO VO
CNJ O O CM O O CO
o o vo
o en en a;
oo r- v z
oen co cd
^^ »*^ 3C
in
S 5 in r-r- v
r-« en PJ •<»••*
rl CM <-> <
O CM
•-« vo en •
CM
• r-< co en
VD »—' O LO O O^
en O C3 00Q O O
s
ID O •» O
10 *+ o •*• vo **> co
t-i co in
VO CM O (O
r-. r-< o co m •"-" vo
t-< C3 O •-« O O «M
f»ca vo vo
y en
i o en z
i o m
vo •> rt» en o _ - _
,-< r- in co us in c\j
<-> CM o o «-« co
tsj co «•
CO
en
CO
en
a:
z
en
en
o "a,
•a- z
e\j us »
or**»i
r-< CM r-4
vo CM CM
x
o
"Eg
>— o
o
CO
vO
UI
co «-• in »»• CM o in
CM
CO
co i-« ro
«• CM n co evi co co
inr. VOCM-- in
co t-i in
CM o f-- c-J eo en ;-<
in vo o co !-" co r~
vo CM CM o ro «»• co
O
u
£
'o
OJ CJ ^-*
o c: CJ o>
*- « P ^
ON » ^^
^ c o *>
r* QJ f~ *&
o ^3 a> -c i—
.X c u a> o •
o *— oj •
I >, I- U O)
t-H UJ 00 I— X -g
o
a>
> 01 u
c j= *J o
a> *•> « •—
c
. o>
i/) •*•
c
^- IO
5?
o o
en
v.
o
o
«3
u
_
a.-a ^ u
o-nuc i—
10 .c o cu ••- ••-
ai
(I) -O
u
o
u
oc
-188-
-------
<^^>
OO
eg;
1 —
UJ
-TT
O
(— H
1C
4S
OO
0
•z.
o
o
cs
»— 1
t— 1
i~
oo E
Q.
Q.
GO
1 —
— 1
oo
UJ
C£
C3
z
I-H
3C
GO
<:
3
I-H
o
oo
.
r^
UJ
CO
t
r—
v>
10
^
1
u
-4
VO O O O OO O
1-4 V V
to
o coco o o o o
co vo v o in co o
CM CO
«}•
co in co 03 co
in <-4 o in co co in
v i-H y co
O CO O r-
CO CO CO i-H CO
CO O O 00 O ~H
mo o o z o o
v
sssggsss
O co *r f*» vo &* in
co co in
0)
cu c
c cu
IO r—
VI J= >s
<
c c o c cu
IQ CU ^ CU O
cn N o IM i-
i- C i— C O
o cu j= CD —
O aO VJ ^3 O ^Z
cu co— c o eu
f— O >- O .— O> 13 C
••- 4J o i >> u u eu
•— «< <_> 1-1 UJ CO t— X
o
^
X
uo
X
<~;
oi
A
X
cn
cn
X
f— 1
CO
X
cn
o>
A
cn
cn
cn
A
X
to
cn
vo
A
f-
cn
a\
A
S
cn
cn
A
U
«0
?
o
cu
•— o
O '*-
•— 3
IO T3
i—
U
•r-
C
ra
o
eu
*o
•f*
E
cu
I/)
§g
in
in 1-4
O CM
*1* VO
CM in
v
§8
CO CO
CO CO
ss
f-
1— 1
vo
00?
oo
0 O
CM — 1
in co
o «r
O CO
CO
eo
eo in
CO
o to
cn vo
1-1
__^
*:»,
thracene
s(2-ethylhe
hthalate
c — o.
et ca
o
o
•-4
CM
OO
CO
§
vH
V
CO
CM
„.
00
o
to
CO
cn
CO
ro
CO
0
00
T_
*o
c
cu
ntachloroph
cu
o.
0£
as
^4
it-
Co
k4
CO
CO*
cn
A
I
CO
CO
co
X
co
si
OL
Z
O
cc
z
X
CO
f-*
cn
cu c
•— o
•i— *^
1 semivolat
anic reduct
ganics
10 en u
4-> I. O
00 C
t— «•
COCJJOOOOOO CO
in i1-. in 1-1 o vo if i-!
1-1 1-4 in in
1-4 r-4 CO
X
in en co
o vo CM vo-eo co o I-H
•-H co co **• i-i vo *s- en
I-ICO «• 00 O
k4
cn 1-4 *j- «j-
o co in co co «»• co co
co o vo vo «-4 vo cn
PO *»•
X
o r» o o o w o •
co o co en vo co to co
1-4 »-4 •«»• cn co co i-i co
r* 1-4 i-H in
v^
P» CO C- CO
•-i o cn CM co o o o
cn 1-4 CM co r-^ f^ co cn
co co CM r-»
«r
X
«r w «r 1-1
co cn vo o co en eo co
vo t-~ co o 1-4 in cn
CM I-H in
vr co o o o o o
CO rf CO O 0 «3- O
cn vo 1-1 f cn CM CM sc
CM CO CO r-t VO
•-1 CM CO
0 «* S
o vo cn r» o vo* o o
1-1 co CM vo vo vn co cn
•-•co «r co co
•-1 CO
,y » 2
vo co f^ co co cn co r^
co «r w vo CM r~ cn
i-* CO ••* *-4 CO
Si-4 «»• co co cn o
co 1-1 o w r-» vo
in vo co in r» «• o
co 1-4 o «»• 1-1 f-
1-1 l-H CO
o
u
3
•o
o E 1 t 3
C — E CU CU E
CU E 0 0.-O.K 3
cu -o
eu T3 eu
(O O O
H- e c
§ " "
«- o ce
u. z z
10 xi
-------
general, the cleaning results of the water wash, the 3:1 molar chelant wash,
and 0.5 percent surfactant wash for the +2 mm soil fraction did not differ
significantly. As hypothesized, the silt and clay particles appeared to be
attached to the sand and gravel primarily by physical processes such as com-
paction and adhesion. These physical attractions are often related to the age
of the soil and the contact time between the contaminants and soil particles.
Because the SARM was a freshly prepared soil that had not been compacted,
weathered, and aged, the physical forces of attraction are believed to have
been relatively weak, a condition more typical of a spill site soil than an
older soil found at an abandoned CERCLA site. Consequently, the water wash
was as effective in cleaning the +2 mm soil fraction as the water-plus-addi-
tive solutions were.
Removal of contaminants from the medium-grained fraction (250 ym to 2 mm)
appears to entail both physical and chemical processes. By nature, this
middle soil fraction, which is composed of medium to fine sand, does not
absorb contaminants to the degree that clays and silts do. It has more sur-
face area, however, and should be somewhat harder to clean than the coarse +2
mm fraction. A comparison of the water wash with the 3:1 molar chelant wash
showed that the chelant wash reduced the residual concentration of metals in
the medium soil size class for each SARM subjected to the chelant wash (SARM
II, III, and IV). This trend is especially apparent in the data for SARM II
(Table 5) where the total residual metal reduction increased from 82.7 percent
for the water wash to 91.6 percent after the chelant wash. The organics show
less variation among experimental runs in this soil size class. For the most
part, water was as effective as the surfactant wash for reducing the level o±
organic contamination. The one anomaly was anthracene, which showed very high
concentrations in the medium soil class. The anthracene evidently was not
fully dissolved before it was added to the SARM; flakes of what was believed_
to be anthracene were observed on the 250 ym screen during the washing experi-
ments .
Reduction of contaminants appears to be affected more by the use of
different wash solutions in the fine soil fraction (less than 250 ym) than in
the other soil fractions. Contaminants are typically bound by both chemical
and physical processes in fine soil fraction. As shown in Tables 5 through 7,
the chelant wash significantly reduced metal contamination in the fine soil
fraction. This reduction is particularly evident in Tables 6 and 7, which
present the results for the SARMs initially high in metal content. Although
the spent wash water was not analyzed, it can be assumed that the chelant
effectively mobilized the metals into solution. Similarly, the surfactant
wash significantly reduced the volatile organic contamination in the fine soil
fraction, as evidenced by the results shown in Tables 4 arid 7 for the high-
organic-content SARMs. Again, the wash water was not analyzed; however, it
can be assumed that the surfactant successfully mobilized the organics into
solution.
The trends indicated by the results of the TCLP analysis were similar to
those shown in Tables 4 through 7. In general, reduction efficiencies ranging
from 93 to 99 percent were obtained in the TCLP analysis of volatile organics,
serai-volatile organics, and metals for the top two soil fractions (+2 mm and 2
mm to 250 urn). Most of the TCLP contaminants present in the +2 mm soil frac-
tions dropped below the proposed regulatory limit given in the Federal Register,
Volume 51, No. 114, June 13, 1986. In the SARMs .containing lower levels of
-190-
-------
metals (specifically SARM I and II), the middle soil fraction (2 mm to 250 ym)
also exhibited concentrations below the proposed TCLP levels.
CONCLUSIONS AND RECOMMENDATIONS
SARM PREPARATION
The preparation of a standard synthetic surrogate soil with physical
characteristics and contaminant levels representative of a wide range of
conditions typically found at Superfund sites was successfully completed. The
surrogate or SARM was subsequently utilized in evaluating the relative effec-
tiveness of five selected treatment technologies (physical soil washing,
chemical treatment, stabilization, low temperature thermal desorption, and
incineration), and a soil treatability data base has now been established.
Further studies comparing the treatability results that were obtained
with the SARM to results from similarly designed studies using actual site
soils are needed to further supplement the data base. Also, future studies in
which the SARM is used to evaluate the relative effectiveness of other pro-
posed treatment technologies at Superfund sites would be valuable.
PHYSICAL SOIL WASHING EXPERIMENTS
The soil washing results from this study appear to support the basic
assumptions underlying the volume-reduction approach to site remediation—that
a significant fraction of the,contaminants in contaminated soils are attached
to the smaller sized particles or fines (i.e., silt, humus, and clay) and that
the coarse material can be cleaned and returned to the site by physically
washing and separating it form the fines. The data indicate that water alone
can efficiently remove a significant portion of the contamination from the +2
mm soil fraction. Contaminant removal from the middle (2 mm to 250 pm) soil
fraction and the fine (<250 pm) soil fraction, however, can be generally
enhanced by chelant and surfactant solutions. Addition of a chelant to the
wash solution can improve metal reduction efficiencies for both the medium and
small particle size fractions. Addition of a surfactant to the wash solution
can lead to higher organic removals (compared with the water wash) from the
fine particles. In general, water appears to be more effective in mobilizing
the organics into solution than in mobilizing the metals.
In the preliminary bench-scale experiments, it was determined that the
SARM was approximately 13 percent (by weight) coarse material (i.e., >2 mm),
29 percent medium-grained material (250 pm to 2 mm), and 58 percent fines
(<250 pm). Therefore, the data presented in Tables 4 through 7 indicate
achievement of at least a 13 percent weight reduction of contaminated material
with a water wash alone. Addition of a chelant solution resulted in a 42
percent reduction by weight of the metal-contaminated SARM, and use of the
chelant and surfactant solutions resulted in lower metal and organic contam-
ination, respectively, in the fine particles.
The mix of contaminants in Superfund soils often lends itself to an
extraction or washing treatment technology such as that demonstrated in this
study. Although promising results have already been achieved at the pilot
scale at a number of lead-contaminated Superfund sites, additional research is
needed to demonstrate the cost-effectiveness of soil washing for full-scale
-191-
-------
treatment of a wide range of metal- and organic-contaminated soils. Specif-
ically, most of the research conducted to date has involved demonstration^of
the operation of various pieces of equipment for pretreatment and extraction
of the contaminants from the soil and for post-treatment of the extractant.
The effective separation of the wash solution from the soil, the recycling of
the regenerated wash solution, and the concentration/destruction of the con-
taminants, however, have not been demonstrated at a large-scale pilot facil-
ity (2). The following is a listing of areas in which future work is needed
with respect to the development of soil washing as a full-scale, viable treat-
ment option for Superfund soils:
1. Laboratory feasibility studies for evaluating removal of contaminants
from the wash water.
2. Laboratory-scale physical soil washing studies using actual Superfund
soils containing a mix of metal and organic contamination. (The first
study of this type is currently funded and should begin in the spring of
1988.)
3. Evaluation of sequential wash solutions for reducing combined organic and
metal contamination.
4. Additional pilot-scale studies on the use of the EPA Mobile Soil Washing
System.
5. Bench-scale feasibility studies evaluating stabilization/solidification
effectiveness as a treatment train option for the concentrated fines
remaining after soil washing.
6. Evaluation of feed stock preparation methods for the EPA Mobile Soil
Washing System.
1.
2.
REFERENCES
Shum, J. S. Drum Screen Washer Operation and Maintenance Manual. Prepared
for the U.S. Environmental Protection Agency, Hazardous Waste Engineering
Research Laboratory, Releases Control Branch, by Mason & Hanger-Silas
Mason Company, Inc., under Contract No. 68-03-3203. February 1987.
Dietz, D. H.3 et al. Cleaning Contaminated Excavated Soil Using Extraction
Agents (Draft). Prepared for the U.S. Environmental Protection Agency,
Hazardous Waste Engineering Research Laboratory, by Foster Wheeler Corpora-
tion, under Contract No. 68-03-3255. September 1986.
-192-
-------
EVALUATION OF PERSONAL COOLERS AND VITAL SIGNS MONITORS
FOR HAZARDOUS SUBSTANCE CLEAN-UP OPERATIONS
By: W. Thompson Lawrence, Sc.D.
Carla Mond
Arthur D. Little, Inc.
20 Acorn Park
Cambridge, Massachusetts 02140
Raymond Frederick
U. S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Woodbridge Avenue
Edison, New Jersey 08837-3679
ABSTRACT
Field tests were conducted in EPA's Region VII during the summer of 1986 and 1987.
The operation studied was the bagging of dioxin contaminated soil while wearing Level "C"
protective clothing. The objectives were to:
o evaluate various methods of obtaining vital signs of field workers;
o evaluate the feasibility of increasing productivity by using personal coolers to
reduce heat stress.
Environmental and physiological parameters were measured. Results of these tests
indicate that use of personal coolers can enable individuals to work for substantially longer
periods of time at relatively high ambient temperatures. In addition, recommendations were
developed for daily monitoring of worker vital signs at the beginning and end of each work
day.
-193-
-------
BACKGROUND
Under the direction of Dr. Ralph Goldman of Multi-Tech Corporation, two field tests
were performed at hazardous waste sites in Region VII. During these field studies, workers
were observed as they performed the "bagging" operation associated with dioxin clean-up.
During "bagging", two individuals stand beside a hopper with arms upraised, holding a bag in
position as a front-end loader operator moves contaminated dirt into the bag. Once the bag
is filled, the baggers tie it, and the crane operator moves the bagged dirt to another area.
During the operation, Level "C" protective clothing is worn: Tyvek coveralls with a hood, a
full-face respirator, rubber gloves and boots, and a hard hat.
The first field study, which took place in Fenton, Missouri during the summer of 1986,
provided data on the effectiveness of different automated vital signs monitoring techniques
in measuring heat stress danger in a field environment. The second study, at Castlewood,
Missouri during the summer of 1987, was undertaken to evaluate the use of selected personal
coolers in increasing productivity in field activities, by reducing heat stress.
OBJECTIVE AND SCOPE
The objectives of the field studies were:
o to evaluate various methods of obtaining vital signs of field workers; and
o
to evaluate the feasibility of safely increasing productivity by using personal
coolers to reduce heat stress.
These evaluations were uncontrolled field trials, and the following limitations applied:
o the pacing and exact nature of the work each day could not be controlled;
o climatic conditions were variable;
o participation was voluntary.
METHODOLOGY
The data collection procedure and physiological data collected were slightly different
for each study. The discussion that follows describes the type of instrumentation used and
the protocol followed for each test.
INSTRUMENTATION
The instrumentation consisted of a computer data acquisition system for on-line data
acquisition, meteorological instrumentation, physiological instrumentation, and personal
coolers. All instrumentation was calibrated and then assembled for a "dry-run" in the
climatic chamber at Multi-Tech. The equipment was then disassembled and shipped to the
field sites.
The meteorological instrumentation included: air temperature (Ta), relative humidity
(RH), solar load, black globe temperature (Tg), natural wet bulb temperature (Tnwb), wet
bulb globe temperature (WBGT), and wet globe temperature (WGT). The air temperature was
measured with Yellow Springs No. 405 air temperature sensor. Relative humidity was
monitored with a dew point hygrometer manufactured by Yellow Springs. The solar load
meter was Dodge Product No. 776. To obtain black globe temperature, a thermocouple was
inserted into a 6 in diameter black globe. The natural wet bulb temperature was collected
from a thermocouple covered by a wet cotton wick. A standard Botsball was used to obtain
-194-
-------
the wet globe temperature. Thermocouples were used for all temperature measurements
9807A data
^^^^^
A three-sensor array with a connecting cable was used to determine the skin
temperature at the chest, forearm, and calf of each subject. DuringTe ?986 fieW study a
reTut nS^vT^ rtv ST*""" ^ medlal thi§h te«- BaSd on'he
results ot this study, the medial thigh temperature was not measured during the 1987
evaluation, and the fourth thermistor was used to measure the temperamTof ^he ! personal
Monitoring body weight changes, day to day and between work shifts provided
PROTOCOL
1986 Field Study
Preparation of Test Subjects—
The, the inner aspect °f the
thp middle nf ti,» t™ i f .1. , arm. The third thermistor was affixed in
-195-
-------
MWST = .5 TcheSt + -34 Tieg + .16 Tarm
Since it has been suggestedM that the thigh temperature is the best single-point
representation for mean weighted skin temperature in the heat, this value was collected for
comparison purposes.
Vital signs monitors were checked and "investigator baseline VSM readings" were taken
on each subject using the "BATON", the telemeter unit and the electronic heart rate-BP
instrument. The portable Yellow Springs Instrumentation Telethermometer box was used to
read the thigh temperature.
Prior to initiating the study, each subject was instructed to drink only from his
assigned container. Values for fluid intake, body weight, heart rate and completed bags,
were recorded manually during the study and tabulated later.
Procedures—
The two "baggers" worked at either side of the "hopper" platform. They adjusted the
inner plastic bag, holding it up against the throat of the hopper, prior to dirt being deposited
from the front-end loader. The front-end loader operator had a slightly higher average
workload than the baggers due to the constant shifting of body position in response to
variation in the position of the front-end loader.
After approximately one hour (at the discretion of the Riedel Company Health and
Safety Officer), the first two baggers moved to the rest area. There, a wooden bench was
set-up on a rough wood pallet over which a tarpaulin had been erected on a crude frame to
shield them from the sun. The second pair of baggers replaced them at the hopper.
As soon as possible, after the first two baggers had moved to the rest area, removed
their respirator, and unzipped (and sometimes dropped their protective coverall to below
their knees), an intensive "round robin" of heart rate and blood pressure measurements was
initiated by Dr. Goldman and the Riedel Company nurse. These measurements, plus the
nurse's reading of oral temperature, were taken sequentially for about fifteen minutes and
provide the primary data base for comparison of the simple field VSM's with the
conventional heart rate and blood pressure determinations.
The first team of baggers then rested and performed light work at the site, such as
preparing bags, until shortly before the second pair were due for their rest break. They then
donned their full protective overgarment and respirator and moved to their original work
stations while the second pair of baggers moved to the rest area to begin the "round robin"
evaluation of their vital signs.
When the work was done for the shift, the workers returned to the dressing room and
were weighed fully clothed and nude. The time at which the weights were taken was
recorded. Final drinking fluid bottle weights were also obtained.
1987 Field Study
Preparation of Test Subjects—
Four workers from the clean-up contractor's work force were used as test subjects
during the evaluation. Program participation was voluntary and monetary incentives were
used to solicit program participation. All participants were briefed as to the purpose of the
study, the types of coolers being introduced, and the monitoring techniques.
-196-
-------
The four volunteer workers were split into two separate work teams (A and B) which
participated in the study on alternate days (e.g., Team A Week 1 W-F, Week 2 T-Th- Team
B Week 1 Th-Sat, Week 2 M-W).
The test subjects were asked to report to the test site roughly one hour before the start
of their afternoon work shift. Both workers were weighed nude and then outfitted with five
thermistors. Three of the thermistors were used to monitor a test subject's skin temperature.
These thermistors were placed on the test subject's forearm, chest and calf in the same
manner as in the 1986 field study. A fourth thermistor, of special design, was inserted into
the subject's rectum and used to monitor core temperature. The fifth thermistor was used to
monitor the temperature of the personal cooler. This thermistor was directly attached to the
ice vest, or positioned in the main delivery tube of the vortex cooler. These thermistors
were attached to a single plug connection which could be passed through an opening in the
protective clothing and connected to a computerized data acquisition system. One worker
was provided with a vortex unit and the other with an ice vest. Both workers then donned
protective clothing consistent with Level "C" protection requirements.
Procedures—
Each worker consented to modify his work cycle from a normal cycle of one hour
working and one hour resting to a single, three-hour work period, during which he
alternately performed his tasks with and without the aid of a personal cooler. The workers
were instructed that cooling would automatically be switched if their partner was
experiencing deep body temperature in excess of 101.5°F, or skin temperature in excess of
96.5°F.
The actual length of time each test subject worked without the aid of a cooler was
limited by his individual tolerance to heat stress. Whenever they decided that they could not
continue to work unaided, their cooling device was turned on. Initially, cooling was
provided alternately to the workers. However, due to the high environmental temperatures,
cooling was sometimes provided to both workers at the same time.
Test subject skin temperature and core temperature were monitored and recorded at
five minute intervals throughout the work cycle. Originally, heart rate was to be monitored
with a telemetering device. However, the unit failed to work in the field because of
interference from electric noise generated by the earth moving equipment. Thus, heart rate
was monitored manually when possible without interrupting work flow. Body weight was
measured at the beginning and end of the afternoon work shift. In addition, fluid intake
was monitored.
RESULTS
1986 FIELD STUDY ON VITAL SIGNS MONITORS
Summary
These tests provided data on the suitability of various devices for monitoring workers'
vital signs during hazardous waste clean-up activities at dioxin sites. The results indicated
that, for the work rates associated with the bagging task, skin temperature could be measured
with thermistors attached to the chest, arm, and calf, weight could be measured using a
digital "bathroom" type scale, and heart rate could be measured using a variety of
inexpensive sports type monitors including a telemetered unit and a "baton" shaped unit held
in both hands. No signs of heat stress were observed due to two key factors. First, the
environmental conditions encountered during the tests were relatively cool (15-20°C
-197-
-------
temperatures with light solar loads around 100 Btu/hr - ft.*) Second, the work rates were
relatively low (estimated at 150-225 Watts).
Heart Rate Monitors
The units evaluated were:
o
A telemetered unit in which a strap worn around the chest contained
sensors and a transmitter which relayed the pulse rate up to 35 feet to a
portable battery powered readout (Biosig #801).
o A baton unit which was grasped in two hands with a readout in the center
(Biosig #103).
o An electronic, portable home-type blood pressure monitor with a heart rate
readout (Marshall #85), and
o Manual readings taken by a nurse.
Data are shown in Figure 1. During the work period (indicated by the box above the
data), heart rate moves erratically as different bagging related tasks are performed. In the
rest cycle, the heart rate is more stable and approaches a resting value During the rest
period, the monitors agree within a few beats/min.
0)
IS
cc
r
CO
o>
X
170
160
150
140
130
120
110
100
90
80
D
Legend:
a Telemeter
+ Baton
O HR/HB
A Nurse
| | Work
r"""111 '1
1 1 Rest
n D
a
~
- a
-
n
a i
70 '
0
D
D
° D DO
0 DoA°
a D
a a •
... O
Q
SB
A
D
1
I 1 1 1 1
20 40 60 80
Minutes Into Test
Bagging Task Indicates that Subject Is Working
FIGURE 1. COMPARISON OF HEART RATE MONITORS FRIDAY AM/MAN # 2
-198-
-------
Mean Weighted Skin Temperature/Thigh Temperature
An example of data for these measurements is given in Figure 2. The MWST appears
to correlate somewhat with the work/rest cycle. However, variations due to changes in work
rate, dislocation of sensors, and orientation with respect to the sun also affect the data This
parameter is much less volatile than heart rate and therefore would be a more reliable
indicator of heat stress if these problems were overcome.
o
0)
0>
O)
0>
0)
•5
a
Q>
Legend:
a Tst
MWST
27
40 60
Minutes Into Test
100
Bagging Task Indicates that Subject Is Working
FIGURE 2. COMPARISON OF MWST AND THIGH TEMPERATURE WED. PM/MAN # 1
Thigh temperature was measured because it was believed by Dr. Goldman to be a
single point value which would track the MWST. The data indicate that this was not the case
tor these test conditions. In addition, problems were encountered with this thermistor
becoming dislodged particularly when the worker climbed up on the hopper to clear
blockages.
-199-
-------
Attempts at measuring sweat production by weighing clothes, fluid intake, and workers
at the start and end of a shift and before and after urination/defecation were unsuccessful.
Despite instructions to the contrary, workers shared water bottles, changed clothes, and
urinated/defecated without notifying the investigators. However, the use of a digital scale to
indicate day-to-day variations in a subject's condition provided useful information relative
to monitoring heat stress risk.
Conclusion
Based on these results, it was determined that the appropriate instrumentation for
monitoring participants during the 1987 field evaluation of personal coolers were MWST
measured with three thermistors, heart rate measured with the telemetered unit, and body
weight measured with a digital scale before and after a shift.
1987 FIELD STUDY ON PERSONAL COOLERS
Summary
These tests provided data on the use of ice vest and vortex personal coolers to reduce
the risk of heat stress while enhancing worker productivity during bagging. The units tested
were a Fyrepel Ice Vest #128 and a Fyrepel Vortex Cooler #167-5. The ice vest consisted of
a jacket with pouches containing packets of chemicals previously frozen in a conventional
freezer. The vortex cooler employs the Hilsch cooling method with compressed air to
provide cool air to a cooling vest and through tubes to the extremities. The environmental
conditions during the tests were varied and uncontrolled WBGT of 22.1°C - 32°C and
relative humidity of 40.4% - 49.3%.
Results of the tests using these coolers indicated that the use of either could permit
baggers to work safely and continuously for at least three hours with Wet Bulb Globe
Temperatures up to 31°C. Without coolers in use, skin and core temperatures rose indicating
that heat stress danger existed. When coolers were used, these temperatures declined
indicating that their use should make continuous work possible under these conditions.
The telemetered heart rate system did not perform satisfactorily during these tests,
possibly due to electromagnetic interference from the equipment. MWST measurement with
thermistors was acceptable although sensors occasionally became dislodged. Their use
required that the workers be tethered to the data acquisition system. Digital scales provided
acceptable weight data.
Ice Vest Performance
The performance of the ice vests on a typical day is illustrated in Figure 3. During the
tests, the subjects wore the vest until a new one was introduced, to avoid having to stop
work to remove the used-up vest. The tests indicate that useful cooling is supplied for about
75 minutes. In Figure 3, the MWST falls initially and then begins to rise as the vest melts.
The core temperature (rectal) drops slightly, lagging somewhat the MWST and then rises near
the end of the cycle.
Vortex Cooler Performance
Figure 4 presents data on the performance of the vortex cooler on a typical day.
MWST and core temperature rise initially without the use of a cooler. When the unit is
turned on the MWST begins to fall almost immediately while the core temperature levels off
and then falls 15 to 20 minutes later.
-200-
-------
o
in
Q>
D>
O
Q^
0)
=1
(3
0>
a.
E
a>
40
39
38
37 1
36
35
34
33
32
31
30
29
28
27
Legend:
n Rectal
•fMean Skin
On
Off
J I
I
I
I
I i
I
I
20
40
60
120
140
80 100
Time (Min.)
FIGURE 3. ICE VEST COOLER DAY 2
160
180
O
w
o>
2
O)
0)
Q
£2
0>
Q.
0)
40
39
38
37
36
35
34
33
32
31
30
29
28
27
Legend:
a Rectal
+ A/lean Skin
off
I
J_
I
I
I
I
I
I
20
40
60
120
80 100
Time (mln.)
FIGURE 4. VORTEX COOLER DAY 4
140
160
-201-
-------
CONCLUSIONS
Both coolers provided sufficient cooling to protect workers against heat stress while
working continuously at the test environmental conditions. In these tests, difficulties
encountered with the instrumentation indicated that improved equipment would be needed to
continuously monitor workers, particularly if they could not be tethered to a data collection
system. Monitoring worker weight daily, before and after a shift provides a basis for
assessing potential risk to heat illness.
-202-
-------
USE OF COLLOIDAL OAS APHRQMS FOR IN-SITU BIODEGRAnATTDN
OF CONTAMINATED CBOiran
Credits:
By: Donald L. Michelsen
Jeffrey ¥. Smith*
James A. Suggs**
Department of Chemical Engineering, Virginia Tech
Blacksburg, VA 24061
Project Funded Through Cooperative Agreement CR-813528-02-0
with Joint US E.P.A./Cine and U.S.A.F. (AFESC-Tyndall) Support
Dr. Ron Lewis and Captain Ed Heyse, Project Administrators
ABSTRACT
Laboratory and pilot scale work has been conducted on the scale-up of
ionnh HH?' *? ^ aPPlicafclon of a 50 to 65% dispersion of stabilized
microbubbles for hazardous waste treatment. A portable test facility has
been designed, fabricated and successfully tested using a packed bed of
millimeter sized glass beads to generate 40 1/min of the microbubble
dispersion with bubbles less than 90 microns in diameter. Microbubble
physical properties have been characterized and standardized through the
development of improved photographic and analysis methods to determine size
distribution, and through the development of a phase separation technique
to quickly determine a measure of stability and quality (* air in
dispsr*siony *
The use of this stabilized microbubble dispersion as a potential oxygen
source for in-situ biodegradation of flowing ground water has been studifd
S U7A10?- slice test cel17 ft- x 7 ft-x 5 in-from front to ba<*. A
50 to 65% dispersion of 50 ± 40 micron air bubbles has proven to be a
feasible means for delivery and retention of supplemental oxygen in the
saturated subsurface of sand as well as a simulated alluvial fan material
Downgradient dissolved oxygen measurements in the flowing ground water have
verified continued oxygen transfer from the retained bubbles to the flowing
ground water. The microbubble dispersion has proved superior in retention
and delivery compared to sparged air/water injections. At least 25% of the
oxygen in the microbubble is accounted for in the DO increase as anaerobic
groundwater flowed through the treatment zone. This utilization could
approach 50$, but about one-third of this oxygen retained is required to
biodegrade the surfactant used to form stable microbubbles.
* Now with PPG Corporation, Cleveland, Ohio
** Now with DuPont Corporation, Savannah River Plant
-203-
-------
INTRODUCTION
Aqueous foams are being used successfully throughout industry,
especially in the areas of flotation, dust suppression, petroleum
production, and fire fighting. However, the development of techniques for
producing consistent fine microbubbles could expand applications and
perhaps improve the uses of foams in industry. One new microbubble foam
development is known as Colloidal Gas Aphrons (CGA). The name "aphron"
comes from the Greek "acppov", a foam, and has been coined to mean a fluid
encapsulated in a thin aqueous shell, or true bubble. First investigated
by Sebba (1972), CGA are typically characterized as a highly stable, 50 to
65% dispersion (quality) of 50 micron sized bubbles. Quality is defined as
the % air as microbubbles in the total dispersion of air plus liquid. Thus
a 65? quality means 65% air bubbles dispersed in 35% water. CGAs are
typically a "wet" foam, meaning the air content is less than 70% of the
total volume, which allows the bubbles to be far enough apart that no
distortion is present. Because of their stabilized nature in part the
result of 200 to 500 ppm surface active agent present, the individual
bubbles do not coalesce. CGA viscosity is about 10 centipoise (cp) at room
temperature although as the air content rises to 70%, the viscosity
approaches 30 cp. These wet foams can be pumped with selected positive
displacement pumps.
In contrast, dry or conventional foams usually are characterized by a
high percentage of air, typically 9555, and the presence of distorted,
polyhedral cells of much greater size than wet foams. The wet foam behaves
like pseudoplastic fluids from 70% down to 50% quality CGAs. As the foam
qualities drop below 50%, they behave like Bingham plastics; that is , a
certain stress must be applied before any flow starts. The wide range of
potential applications for these wet microbubble foams has been identified
and include: flotation separation processes (removal of ash forming
material from coal, algae from waters), enhanced oxygen transfer (high
surface area) in gas-liquid systems (yeast generation in bioreactors),
removal of oils from sand (scouring), and, use as an in-situ oxygen source
(in-situ biodegradation of hazardous wastes).
Considerable CGA laboratory testing and some pilot-scale testing has
been completed to address these applications. However, large scale testing
has been restricted by the inability to produce larger quantities of these
microbubbles. CGA generation has been limited to a technique known as the
spinning disk method. This method involves the use of a small 1 to 2 in.
flat disk rotating in excess of 4000 RPM between two carefully positioned
baffles in a 2 liter beaker filled with dilute surfactant solution (Sebba
1985). With an overflow arrangement, this generator can operate in the
continous batch mode producing up to 4 L/min of CGAs. Efforts to scale-up
the spinning disk method have been hampered by the problem of generating
the necessary high shear in larger vessels. Thus, the first thrust of this
effort supported by USAF/EPA has been to develop and scale-up means for
generating production quantities of CGA's and to quantify CGA
characteristics.
-204-
-------
For the treatment of hazardous waste and fuel releases to the ground
environment, conventional technologies involve pumping contaminated
groundwater to the surface for treatment. Since the solubility of organic
pollutants in water is very low, and organics tend to be sorbed or occluded
in, soil, the flushing process can take decades. In situ technologies take
the treatment.process to the source of contamination and have the potential
of accomplishing treatment faster. Enhanced aerobic biodegradation is
theoretically effective for treating nonchlorinated organics.
Biodegradation rates are limited by the availability of certain chemicals,
particularly oxygen limiting chemicals, are added to the subsurface.
Sparged air and hydrogen peroxide have been used to supply oxygen to the
groundwater. Air sparging can only supply low concentrations of oxygen to
water. Hydrogen peroxide is expensive. It also tends to degrade so
rapidly in an aquifer that oxygen is lost from solution of hydrogen
peroxide concentrations exceed 100 ppm.
This phase of the project investigated the potential to deliver a
microbubble dispersion (versus sparged air/liquid mixture) to the saturated
zone of an unconfined aquifer as the oxygen source for biodegradation of a
hydrocarbon contaminant. The experiments involved a series of oxygen
supplying injections to flowing ground water in a 7 ft. by 7 ft. by 5 in.
deep (front to back) vertical slice. The thrust of this second study was
to quantify the movement of these microbubbles through the soil matrix
transported by the flowing groundwater and to study the retention, and
cohesive properties of microbubbles (versus sparged air) -and their
propensity to form an active oxygen treatment zone.
BACKGROUND
Continuous Microbubble Generation
Currently the spinning disk technique (Sebba 1985) with its ability to
generate 4 L/min of CGA using a continuous overflow system has become the
standard CGA generation technique. This technique can provide sufficient
quantities of microbubbles for laboratory and pilot testing. Other
techniques for generating wet foams, gas sparging, and more recently high
shear generation techniques have received attention. Dissolved air
generation techniques also produce microbubbles of 40 to 150 micron size,
but their quality is normally limited to only 5 or 1055. Sparging is still
very popular for many application, but for the most part bubble size is
1000 microns (1mm) or greater unless shearing occurs. A number of shearing
devices have received attention. Deister Concentrator has used a venturi
approach to produce 40 to 50 micron bubbles (Parkinson, 1986). Shea and
Barnett (1979) have described an idea for obtaining microbubble foam by
using a hydrocyclone to separate the fine bubbles from conventional foam.
More recently Professor Jan Miller at University of Utah has been testing
an air-sparged hydrocyclone with a jacketed porous-wall tube to produce a
30 micron bubble size directly in the flotation reactor, and R. Yoon at
Virginia Tech has been producing 100 micron bubbles in a continuous
benchscale column (Parkinson, 1986). Finally Foot et al. (1985) and McKay
et al. (1987) have been using an external bubble generator which produces
-205-
-------
bubbles less than 200 microns and larger by a combination of pressure
dissolution and mechanical generation in a packed bed reactor. Much of the
recent effort in microbubble generation has come from the interest in fine
coal flotation/beneficiation, and most recent testing aims at producing a
80 to 95? quality dry foam.
After some consideration, we decided to study the system developed by
Don Foot and Jeff McKay with the Bureau of Mines. The goal was to scale up
a fine glass sphere packed bed system, and conduct testing to produce a 45
to 70% quality foam under conditions of high velocity (shear) and with
concentrations of surfactant greater than the typical 5 to 25 ppm used for
flotation.
The key parameters considered during this study were mean glass sphere
diameter (D ), Reynolds Number (velocity), quality of microbubble
dispersion ginerated, and sodium dodecyl benzene sulfonate, (surfactant)
concentration.
OXYGEN DELIVERY TO THE SATURATED SUBSURFACE
In the laboratory, Michelsen et al. (1984a) demonstrated that
colloidal gas aphrons, when sparged into various unconsolidated saturated
soil matrices, are "captured" and retained for prolonged periods of time.
During these tests, 70 to 82% of the air incorporated into CGA's, made
using a nonionic detergent (Tergitol 15-S-12), immediately adhered to a
saturated coarse sand when sparged into the sand with a fork-like probe.
Thirty days later, 70 to 80% of the initially captured microbubbles were
still retained in the saturated sand matrix as an air dispersion or as
coalesced aphrons. Similar tests were completed in a pilot test using a
sand-filled trough with results of somewhat lower microbubble retention in
the sand (Michelsen et al. 1984b, 1985). In similar tests, following the
same procedures but injection air instead of CGA's, the air rapidly
"bubbled up" and burst through to the surface with little gas retention in
the sand. These studies suggest that in situ biodegradation could be
enhanced using a combination of CGA's and nutrients injection into a
saturated aquifer matrix to activate the indigeneous microorganism culture.
A CGA injection study was conducted in a vertical slice test cell. The
microbubbles showed a lack of lateral movement in the medium sand tested.
The results suggest that in situ use of CGAs might be effective in a scheme
where a dispersed screen of microbubbles would be injected to serve as a
treatment zone through which the contaminants flow and are biodegraded
(Michelsen et al. 1986, 1987).
EXPERIMENTAL AND RESULTS
Continuous CGA Generation
To conduct the CGA generation tests, a portable unit and upflow packed
bed devise was designed and fabricated, see Figure 1. Air was added at
several points in the generator prior to passing into the packed bed. The
portable unit included means for sampling to check quality (% air in the
-206-
-------
Recycle Pump
(Screw-Type)
FIGURE 1. Portable Pilot Unit for Generating Mi.crobubble Foam
-207-
-------
total dispersion), stability (normalized rise % of clear liquid/bubble
interface to foam disappearance), and bubble size and distribution. The
microbubble dispersion sample line passed through a small viewing cell
mounted on a 10X microscope. The image was picked up with a black and
white videocamera and observed on a black and white high resolution TV
screen. To record, flow was stopped and the image was photographed. A
Kontron image analyzer was used to determine the bubble size and
distribution from the microphotographs (Suggs, 1987).
Table 1 shows the results of testing with water flow rate as the
variable. The stability increased with flow rate as evidenced by the lower
interface height after 1 minute. In addition the greater number of smaller
bubbles of less than 30 to 90 microns and fewer bubbles greater than 210
microns clearly show increased generation of microbubbles with increased
flow rate (turbulence, shear).
Table 2 shows that as the quality increases again the stability
improves and generally small microbubbles were formed. The generation of
smaller microbubbles at higher quality was not expected, and additional
testing would be necessary to clarify the results.
A number of tests with smaller sized packing were conducted, and a
general trend toward smaller bubble sizes and improved (lower) stability
was noted. Traded off against smaller microbubbles was pressure drop which
increased, as expected, as smaller glass spherical packings were tested.
An important element not investigated during testing was the void air (non
bubbles) passing through and not observed as bubbles.
The effect of surfactant concentration on bubble generation was also
evaluated. As surfactant concentration increased the stability increased
and the bubble distribution was altered. At low surfactant concentrations
the proportion of bubbles less than 30 microns was greater. As
concentration increased generally a more bimodal distribution of smaller
and larger bubbles appeared although the proportion of bubbles greater than
210 microns dropped. Some distortion in the analysis may have occurred as
a result of the long tubing distance between the test system and the
microscope mounted cell, and the selection of images on the screen to be
photographed(sample size was not large). Additional studies would be
desirable to explore the effect of surfactant level particularly on size
and distribution data.
The effect of 0.9 recycle ratio on microbubble generation
with the presence of the packed bed was also studied. The recycle ratio is
the fraction of microbubbles returned to the inlet of the recycle pump
compared to form (product) going to the holding tank. The stability
(interface height at one minute) was 25 and considerably lower than the
stabilities on Tables 1 and 2 and approaching the stability of 10 to 12 (1
min) observed for CGAs generated with the spinning disk generator. Also
with the recycle pump (Allweiler Pump Co., screw type positive displacement
pump) the faction of bubbles less than 90 microns was 96.455.
In summary, the packed bed can be used to scale-up and produce larger
-208-
-------
TABLE 1
EFFECT OF WATER FLOW RATE
Stability (500 iriL cylinders)
Run Water Quality
Flowrate
(L/min)
Interface height at: Slope
1 min 2 min
5
6
7
8
12
16
20
24
65.6
62.8
61.7
62.2
80
77
70
63
88
87
82
78
8
10
12
15
Size and Distribution
Run
5
6
7
8
Peaks (p )
45, 105
105,195,45
45, 165
45,105,195
Bubbles
< 30 p
0
0
6
3
Bubbles
> 210 p
8
6
6
3
Fraction
< 90 u (%)
31.8
26.2
50.9
46.2
Number of
Bubbles
44
42
53
52
Concentration
Packing Diam.
Recycle Rate
Peaks
500 ppn of NaDBS
1.5 mm
0
Midrange value of size range
eg. 45 p = 31-60 p.
Listed from largest to smallest*
-209-
-------
Run
Quality
TABLE 2
EFFECT OF QUALITY
Stability (500 itiL cylinders)
Interface height at:
2 min
Size and Distribution
Slope
11
7
12
53.6
61.7
72.3
83
70
46
92
82
67
9
12
21
Run
11
7
12
Peaks (p)
45,105
45,165
15,135,205
Bubbles
< 30 u
4
6
38
Bubbles
> 210 u
7
6
0
Fraction
< 90 w (%)
40.3
50.9
68.9
'lumber of
Bubbles
62
53
122
Concentration
Feed Flowrate
Recycle Rate
Packing Diam.
Peaks
500 ppm of.NaDBS
20 L/min
0
1.5 ram
Midrange value of size range,
e.g. 45p = 31-60u.
-210-
-------
quantities of microbubbles. A column packed with millimeter sized glass
beads is able to produce an effective microbubble foam that contains a
majority of bubbles less than 90 u , and few greater than 150 w. The ideal
operating conditions for producing these microbubbles are as follows- any
feed flowrate from 12 to 24 L/minj foam quality of 65% (34-68 L/min of
microbubble foam); surfactant solution of 200 ppm NaDBS; bead size of 3 0
mm: and, a recycle ratio of 1.0. If the inclusion of a few large, 200 to
300 moron sized bubbles does not affect the application for which the
proauct will be used, operation without recycle is recommended. For this
case, operating conditions require 20 L/min or more feed flowrate, 65% foam
quality (57+ L/min of microbubble foam), 200 ppm NaDBS, and a bead size of
1 .5 mm.
The operation of a packed bed microbubble foam generator was both
feasible and reliable, and economically viable. The estimated operating
cost is $0.58 to provide 1800 L/hr of bubbles using pressurized dissolved
air techniques with a continuous flow system. Cyclic batch flotation
systems could be cheaper. Dissolved air flotation (DAF) techniques are
widely used in the food, mineral, and petroleum industry for removing
particulars and dispersed oil from wastewaters.. However, the operating
cost is $.35 ($.33 for surfactant, 250 ppm solution) to provide an equal
Ioft«r 1 microbubbles U5inS the Pa^ed bed generating system (Suggs,
1988). The major concern, however, is that the packed bed generator will
plug if wastewater effluent containing solid matter is used for bubble
generation (typical operations). Current work is underway to minimize
surfactant utilization (cost) and the development of a new non-plugging; CGA
generator.
IN SITU
The on-going research in the vertical cell has been evaluating the
propensity for CG/'s (versus sparged air) to flow through a saturated
medium sand (10 to 10 u cm/sec hydraulic permeability) and an even
coarser alluvial fan consisting of a natural gravel/sand mixture layer
larger than 100 mesh but passing a 3/8 in. screen (10~T cm/sec hydraulic
permeability). These studies have evaluated changes of dissolved oxygen in
the flowing groundwater, movement of discrete microbubbles into the
groundwater, and effect of treatment injection on water permeability
(plugging) through the vertical test cell.
Figure 2 shows the front view of the vertical slice cell used for this
study. The groundwater table maintenance was controlled by establishing the
SfaUlf-C head ±n fche Srou^water feed tube on left with groundwater
effluent controlled by gravity overflow on the right. The solution feed
tube served as the injector for intermittently or continuously introducing
the microbubble dispersion or sparged air/water mixture. Details for
control of ground water introduction, the CGA generation/injector system,
MoflfitSt f11 Sample matrix in the back f&ce have been documented by Smith
U988J. A spinning disk generator operating in a constant overflow mode
supplied the CGA injection. Level indicator taps were tied into the
manometers (piezometers) mounted to the right of the cell. A matrix of
sample ports (back face) were used with quick release valves to withdraw
-211-
-------
Figure 2. Vertical Slice Test Coll Schematic
-212-
-------
liquid samples for microbubbles, dissolved oxygen (D.O.), pH, temperature
and for colorimetric analyses.
With either soil matrix tested the dissolved oxygen showed a sharp
increase with a step introduction of a sparged air or CGA's into the
gradient. With a medium sand matrix few microbubbles were observed down
gradient after introduction of sparged air or CGA's, and in both cases,
permeability changed little after injection. Based on laboratory studies
we expected to see microbubbles traversing through the coarse alluvial sand
matrix also tested in the vertical cell but few were observed. Instead,
CGA air bubbles and sparged air bubbles/saturated water caused the DO
downstream to increase. Figure 3 provides a profile of D.O. as a function
of cell position during the 24 hours after an initial pulse of CGA
injection. Groundwater flow was kept constant during the first 11.75 hours
with a velocity of about 0.75 ft. per hour. The vertical dashed line shows
the location of the groundwater wave front after pulse aeration assuming
uniform horizontal groundwater velocity (plug flows) through the vertical
slice. The flow was then stopped. Figure 4 provides a similar D.O.
profile to above but with an initial pulse of sparged air/water injection.
Groundwater flow for 8.5 hours after CGA injection was measured at a
velocity of 1.1 ft. per hour before being turned off. During both pulse
injections, some bubble and air loss was observed surging to the
unsaturated zone. However, the remaining air and especially CGA
microbubbles stabilized in place as documented in earlier testing The
fixed microbubbles supplemented the dissolved oxygen in the continually
flowing groundwater passing around the injection sparger for over a day.
The oxygen delivery of injected sparged air depleted much more rapidly, and
was less effective in transferring dissolved oxygen to the flowing
groundwater using a single pulse procedure Oxygen transfer for CGA
injection to the flowing groundwater was determined by multiplying the
groundwater flow (0.75 ft/hr) by the estimated A DO (4.0 ppm) pick-up
measured at injection point for the duration of the 11.75 hr test An
estimated 0.140 g. moles were transferred. The total CGA bubbles injected
were 7.2 L of air or 0.571 g. moles 0 or a 252 oxygen delivery to the
flowing groundwater. ^
The generation of microbubbles requires the use of a minimum 100 ppm
surfactant concentration. (500 ppm were used in Run 10, Figure 2) With
3.0 liters of liquid injected (to form the wet microbubbles) and assuming a
1.0 mg 0/1.0 mg surfactant is required for biodegradation of the
surfactant, this means that 0.094 moles of the 0.140 moles
injected/delivered oxygen (67?) could be required to biodegrade the
surfactant used for generation. On-going aquarium injection studies have
demonstrated that 50$ retentions are realistic, so in practice perhaps 66%
of the 02 retained (33% of the oxygen supplied) would be available for
pick-up and in-situ biodegradation of contaminants flowing through a
treatment zone.
Ann A* S/fila,T
-------
•A1-
—r~
8.9
1.2
—r~
2.3
4.5
—r~
6.7
1Q.1I
E'-
0'-
C'-
8'-
A1-
Movement of Ground-
water Front From
TIME=0
TIME 2Hr 56 Min
1.2
2.3
4.5
6.7
8.9
IQ.U
E'-
O1-
C-
B'-
A'-
—i—
1.2
—i—
2.3
4.5 . 6.7
8.9 10.11
E'-
0-
C'-
8'-
A'-
—r-
1.2
—i—
2.3
• 4.5
6.7
8,9
10.11
Figure 3
Dissolved Oxygen Concentration (mg/L) Within The
Vertical Slice Test Cell Using Injected CGA Pulse,
Run 10.
-214-
-------
Movement of Groundwater
Front from TIME=0
Slic
7.
Ftgure 4
Within
Air/Water
-215-
-------
Subsequent testing will continue to evaluate oxygen transfer to the
flowing ground water provided by intermittently injected CGA compared to
oontinuoual^ sparged air which could be justified economically. The CGAs
will be made from non-biodegradable 3M anionic surfactant, FC-100, to
eliminate the oxygen consumption due to surfactant biodegradation.
Blockage formed by the injected CGA or continuously sparged air will be
measured by injecting a solute pulse (NaCl addition) to increase the
groundwater electrical conductivity and following groundwater flow
patterns. If blockage occurred, no increase in conductivity would occur
downstream of the point of blockage or the increase would be significantly
reduced compared to when no blockage was present.
In summary, CGA, unlike sparged air/air saturated water, provided
supplemental oxygen which formed an "oxygen wall." CGA accumulate near the
injection tube in a stationary front which increased the dissolved oxygen
concentrations of groundwater flowing through the cell. Dissolved oxygen
data showed that CGA runs achieved consistently higher dissolved oxygen
concentrations than sparged air runs. Overall utilization of charged
oxygen was 25% assuming a 100 ppm surfactant solution used for CGA
generation. Laboratory and small aquarium studies are being conducted to
further reduce surfactant usage, to improve stability and injection
procedures, and to design the configuration for a treatment zone.
Bibliography
Bird, R. B., Stewart, W. E., and Lightfoot, E, N., 1960, Transport
Phenomena. John Wiley & Sons, New York, p. 200.
Foot, D. G., McKay, J. D., and Huiatt, J. L., 1985, "Column Flotation of
Chromite and Fluorite Ores," 24th Annual CIM Conf. of Metallurgists,
Vancouver, BC, Canada.
Fugate, E. J., D. L. Michelsen, D. A. Wallis and F. Sebba, 1985, "Use of
Colloidal Gas Aphrons to Facilitate Decontamination of Subsoil Hydrocarbon
Spills," Presented and preprinted, AIChE meeting, Chicago, IL, November.
Mackay, J. D., D. G. Foot, Jr., and M. B. Shirts, 1987, "Column Flotation,"
U. S. Bureau of Mines, Salt Lake Research Center.
Marsden, S. S. and Khan, S. A., 1966, "The flow of foam through shor porous
media and apparent viscosity measurements," Son. Pet, Ens. J, (March), pp.
17-25, AIME 237.
Michelsen, D. L., D. A. Wallis, and F. Sebba, 1984b, "The Use of a
Microdispersion of Air in Water for In Situ Treatment of Hazardous
Organics," pp. 398-403 in Proceedings of Fifth National Conference on
Management of Uncontrolled Haza^n.m Waste Sites. Hazardous Materials
Control REserach Institute, 1984.
Michelsen, D. L., D. A. Wallis and S. R. Lavinder, 1985, "In-Situ
-216-
-------
Biodegradation of Dispersed Organisms Using a Microdispersion of Air in
Water," pp. 291-298 in Proceedings of 6th National Conference on Management.
of Uncontrolled Hazardous Waste Site*, Hazardous Materials Control Research
Institute, Washington, D.C., November.
Michelsen, D. L., D. A. Wallis and F. Seb.ba, 1984a, "In-Situ Biological
Oxidation of Hazardous Organics Using Colloidal Gas Aphron Systems,"
Environmental Progress, Vol. 3, No. 2, pp. 103-107, May.
Pft7D\,TL'V?* ~ugate' &> A' ¥allis' J- Fisher> F- Sebba •
Paris, 1987, "In Situ Groundwater Mitigation of Hydrocarbon Contamination
Washin ton0^^ ^ AphrOnS'f' Final ReP°rt» American Petroleum Institute,
Michelsen, D. L., Wallis, D, , & Sebba, P., 1986, The Use of "Aph^n*
.Techniques for Treating Hazards wa*t-.*gj 3rd Intl. Congress of Chemical
Engineering, Tokyo, Japan: Sept.
Parkinson, G. (Reporter), 1986, "Improved Flotation Routes Get Separations
Tryouts."
,. D> W> Green' 6d" Perry's Chemical Engineers'
sixth edition, McGraw-Hill, New York, NY.
Sebba, F., 1971, "Microfoams— an unexploited colloid system « J Coll
Interfane Snl , t 35(4), pp. 643-6. - ^^
Sebba, F., 1985, "An improved generator for micron-sized bubbles," Chem
Ind,, pp. 91-2, Feb. —^^
Shea, P. T and Barnett, S. M., 1979, "Flotation using microgas
dispersions." Sep. Sol. Tenh. lltfol, rr 7^7-7^7. *
Suggs, J. A., 1987, Generation of Miorobubble Foam Uslnf? A Packed Column
Master's Thesis, Virginia Polytechnic Institute & State University
Blacksburg, VA, September.
-217-
-------
EXPERT SYSTEMS TO ASSIST IN REVIEW OF CLOSURE PLANS FOR LAND DISPOSAL SITES
by: Daniel 6. Greathouse
U. S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio 45268
BACKGROUND
The Resource Conservation and Recovery Act (RCRA) of 1974 mandates
that all hazardous waste land disposal facilities have an approved closure
plan by the fall of 1988. Included under this requirement are all new
hazardous waste land disposal facilities that have opened since 1974 in
addition to those that were in operation at the time of enactment of the
RCRA legislation. The principal objectives of these plans are protection
of the environment from hazardous waste contamination and minimization of
possible associated adverse health effects. To accomplish these objectives,
the Agency or designated states must thoroughly review proposed closure
plans for compliance with sound engineering principles and regulatory
policies. Review consistency is also an important factor in implementation
of this national regulatory requirement i.e., a closure plan should be
evaluated similarly by all regulatory officials within and among the EPA
regional offices and states. A package of four expert systems is being
developed to aid in accomplishing the objectives of the closure plan for
these reviews. Each expert system addresses a major component of a closure
plan, namely final cover, leachate collection, run-on/off control, and
vegetative cover. This paper will describe these four systems.
Closure plan review was identified as a high priority decision area in
discussions with EPA regional office permit review staffs and the EPA Office
of Solid Waste. Since relevant expertise and assistance was available for
decisions relating to closure plan review, the topic was selected for develop-
ment of expert systems. The Hazardous Waste Engineering Research Laboratory
(HWERL) and COM Federal Programs, under contract to HWERL, met further with
Agency decision makers and outside experts to identify the specific closure
issues amenable to expert systems development. Six issues, representing
the major components of a closure plan, were identified; the components were
final cover, vegetative cover, leachate collection, run-on/off, ground
water monitoring, and long term maintenance. Development of expert systems
-218-
-------
to address two of these areas was initiated in 1987. Each expert system
has been designed to operate as an independent module as well as in conjunc-
tion with the other systems. When used as a package the systems complement
each other by sharing information and conclusions, thus increasing user
productivity. These systems will be discussed in this paper.
MATERIALS AND METHODS
Expert systems are computer programs (software) designed to provide
advice in specialized areas. The design objective of these programs is to
emulate the advice of subject specialists by incorporating the experien-
tial^ derived decision rules or criteria that they use. A common way of
representing this knowledge is with IF—THEN statements. For example see
Mgure 1. As the different types of knowledge or the number of decision
rules increase, then more sophisticated knowledge representation techniques
such as objects, frames, demons, or neural nets, may be required. Expert
systems differ from traditional computer programs in several ways
1. Procedural versus goal oriented
2.
Know!edge/judgment under conditions of uncertainty versus information
manipulation
3. Separation of knowledge from inference mechanism and interface.
Expert systems offer the opportunity to effectively use state-of-the-
art technical knowledge, operational experiences, .and regulatory policies
in evaluation of the complex issues associated with closure plans. A long
xhi story of relevant hazardous waste research in addition to operational
experiences with different designs and site conditions have produced an
extensive literature base and expertise resource. This information and
these experiences can be synthesized along with the relevant regulations .
and policies into a decision aid via expert systems technology. Hence
these tools are a means of providing ready access to a knowledgeable,
consistent consultant for each permit reviewer in the ten EPA regional
offices and the states with regulatory authority. The need" for doing
literature searches, review of relevant regulations, and/or contacting
specialists with limited time is reduced. Hence these systems should
reduce the time for evaluation of closure plans in addition to improving
their quality. Other important benefits to the closure plan review process
include training new reviewers and documentation of review decisions and
the bases for them.
COM Federal Programs under contract to the EPA Hazardous Waste
Engineering Research Laboratory are developing the Closure Plan Review
Advisory Expert Systems (Advisory CES). COM has been responsible for each
stage in the development process including identifying the key issues-
determining the scope of the.issues; collecting information from experts
regulatory personnel, and targeted users (knowledge engineering); development
of the software; and field testing of the systems. An oversite group com-
prised of the ORD project officer and knowledgeable personnel from the
-219-
-------
IF
1) 1) the post-subsidence convex slope of the
vegetative layer is greater than or equal to 2,
and
2) the post-subsidence convex slope of the
vegetative layer is less than or equal to 5, or
2) 1) erosion in tons per acre per year is less than or
equal to 2, and
2) 1) the post-subsidence convex slope is less than
or equal to 33, or
2 the proposed vegetative cover is nonvegetative,
THEN
it is definite that the slope of the vegetative layer is
adequate.
Figure 1. Decision rule concerning adequacy of post subsidence slope
-220-
-------
Office of Solid Waste (OSW) was set up to monitor the overall direction of
these development efforts. The OSW staff has been reviewing each system to
ensure compliance with regulations and their knowledge of the scientific
literature. The systems are/have been being developed for use on an IBM
HL/AF, or compatible, microcomputer with a minimum of 640k of RAM and a CGA
graphics board with a color monitor. The PC Plus expert systems shell
w??hVepr°Ps?HFaMF ?istr1buted i?y. Texas Instruments) is being used in conjunction
with PC SCHEME (a version of the Lisp language by Texas Instruments) to
code the systems. '
The focus of the closure plan review systems is review of plans for
land disposal sites which are, or will be, closed as landfills. Hence
decision issues unique to clean closures or partially clean closures for
surface impoundments are not addressed. Also these systems do not provide
advice for plans that differ from prescribed regulatory criteria and EPA
9UDrD«Ce* F°r e?amP1e> tnere ^ no advice for review of an alternative to
a RCRA cap even if the applicant states that performance will be the same
or better. Advice, beyond that included in EPA regulations and guidance,
is limited to those situations and areas not specifically addressed bv
them. J
Initiation of this development effort involved identification of the
major closure related decision areas that would be amenable to expert
2f?hf"SFDre1oPmen*' JnPut was solicited from the targeted user community
within EPA regional offices, EPA headquarters, and states with regulatory
authority of hazardous wastes prior to area selection. It is very important
that the resulting systems meet the needs of the targeted user community.
The six areas selected for development were final cover, leachate collection,
run-on/off controls, final vegetative cover, ground water monitoring, and
long term maintenance. A guidance committee comprised of the work assignment
mana.ger, the ORD project officer and representatives of the EPA Office o
\
1. review the system prototypes during development
2. identify appropriate experts
3. review or identify other knowledge base reviewers
4. identify appropriate regulatory, guidance, and scientific documents
5. ensure the adequacy and compliance of the system conclusions with
current regulatory policies and documents.
The first system selected for development is the final cover advisory
system. The principal objective of this module is to determine whether a
§r f If? "ver system meets the regulatory criteria set forth in 40 CFR
Part 264 for the closure of landfills and surface impoundments. Based on
the recommendations of the Office of Solid Waste, the performance criteria
specified in their guidance documents are used to evaluate a proposed final
cover. This limits a cap to the standard "RCRA cap" endorsed by EPA. Hence
-221-
-------
the system verifies that the proposed cap is comprised of five layers;
vegetative, filter, drainage, flexible membrane liner, and soil liner layers.
For each layer, the system ascertains the acceptability of general design
criteria such as slope, permeability, construction practices or limitations,
and capacity to handle percolated fluids. The suitability of each layer is
determined as an independent component as well as part of an integral system.
Recommendations for approval (or disapproval) are provided which are based
on characteristics of the waste, layer designs, filling practices, etc.
The final cover advisory system determines if the cover design promotes
drainage and if migration of waste to surface and ground water are minimized.
This module is ready for extensive beta testing.
The second system examines some general issues concerning the vegetative
cover including its slope and the potential for run-off to cause gullies in
the cover. The system is broken down into three components; subsoil, top-
soil, and vegetation. In evaluating the soil layers, the system first
checks for appropriate textures and thicknesses. Properties of topsoil
that are analyzed include the organic content, pH, conductivity, and percent
of coarse fragments. Similar checks are done for the subsoil. The system
evaluates the proposed vegetation by first asking the user for the state
and season in which planting will occur. Once the plant species are identi-
fied, the system ensures that it is appropriate to plant each species in
the specified state during the season indicated. Two data bases were devel-
oped to do this evaluation. Criteria for approval of the plant mixture
include the elimination of any deep rooted species and the inclusion of at
least one drought tolerant species and one perennial. All legumes are
checked for inoculation. Finally, the system checks that the proposed
method of application is specified and that mulch and seed rates are ade-
quate. This system is currently in the demonstration prototype stage.
The third system is designed to evaluate a proposed or operating
leachate collection and detection system for completeness and compliance
with current RCRA requirements set forth in 40 CFR Part 264, 40 CFR Part
265, and in the Hazardous and Solid Waste Amendments of 1986. Performance
criteria which appear in guidance documents are incorporated in the system
to test for leachate collection system adequacy. Four elements of the
leachate collection system are evaluated; clogging, leachate head, storm
capacity, and pipe strength. The expert system determines whether the
overall design meets the regulatory objective and whether there is a poten-
tial for failure in any segment of the unit. No attempt was made to incor-
porate requirements set forth in the proposed May 29, 1987, regulations for
double liner systems. This system is ready for limited field testing by
representatives of the targeted user community.
The fourth system is being developed to assist in evaluation of the
proposed run-on/off control systems. It is assumed that the applicant has
accurately interpreted the topography of the site and has properly designated
drainage areas. The system determines if there is a potential for run-on,
and if so, how it is being handled with respect to run-off. Of the run-on/
off control structures, only the most down gradient conveyance and basin
are evaluated by the system. This assumes a worst case scenario with run-
on/off water reaching the most down gradient location. The user can ask
-222-
-------
the system to check the applicant's calculations for other run-on/off control
structures, but in this case it is the user's responsibility to identify
and track which conveyances and basins have been checked. For each convey-
ance, the system checks that the peak discharge flow is less than the convey-
ance flow capacity and that the fluid velocity in the conveyance is less
than its threshold velocity. The volume of each basin is calculated. A
checklist is also provided to ensure that the applicant has provided all
the supporting documentation detailed in the run-on/run-off calculations.
This system is currently under development.
Due to unexpected resource requirements for the first four systems,
the fifth system, ground water monitoring, and the sixth system, long term
maintenance, were not initiated.
Other expert systems are being developed by the Hazardous Waste
Engineering Research Laboratory to provide assistance in the review of
permits for land disposal facilities and to assist in Superfund clean-up
activities. Issues addressed by the systems include evaluation of method
9090 data for flexible membrane liners, evaluation of waste analysis plans,
screening of dike design features for surface impoundments, and screening
of technologies for the clean up of superfund sites. All of the systems
are operational. The first three are ready for extensive field testing and
the fourth system is ready for review by experts.
DISCUSSION-
Closure plan review was identified as a high priority decision area by
an EPA sponsored requirements study for land disposal related expert systems,
by permit reviewers in the EPA regional offices, and by staff involved with
land disposal regulations within the EPA Office of Solid Waste. Recognizing
this need for decision support and the availability of a large information
and expertise resource, the area was selected for expert system(s) develop-
ment. Component selection (final cover, leachate collection, etc.) was
based on the recommendations of knowledgeable experts and the EPA regional
office staff. Using instructions provided by the staff of the Office of
Solid Waste, the systems were designed to comply with appropriate OSW guid-
ance documents. For those situations where more than one guidance was
available, the OSW staff specified which to follow. By keeping the closure
plan rules within accepted OSW guidance, the systems were designed to meet
the advisory needs of the users. Experts were used to interpret and expand
the information in the guidance documents when necessary. In a limited
number of cases when EPA policy was unclear, information provided by experts
and not addressed by guidance were incorporated into the knowledge base.
The entire content of all knowledge bases were reviewed several times by
designated EPA staff.
In order for decision aids to be useful within a regulatory program
they must synthesize both the regulatory requirements/guidance of the
program and the relevant technical expertise in the area. The closure plan
advisory, review system illustrates this combination of policy and technical
information within a single system. To deviate from this practice and
-223-
-------
include technical expertise contrary to that contained in regulations or
official guidance documents would in effect provide approval for deviation
from the regulations or official guidance. The effect of limiting advice
to that authorized by the regulatory program may reduce the utility of the
systems to decision makers. This situation is illustrated by the final
cover advisory system. Final cover designs are commonly proposed which
differ from the prescribed RCRA cap but, according to the applicant, have
similar operating characteristics. Permit reviewers have requested that
the expert system assist in reviewing these designs. The system, however,
does not provide advice concerning alternatives to the RCRA cap. This
restriction (to a RCRA cap) has reduced the value of the expert system to
some permit reviewers. In spite of the potential loss of utility of the
systems they can not be supported by the Agency if they provide advice that
differs from regulations or official guidance.
CONCLUSION
Expert systems have the potential for providing advice concerning high
priority issues that must be evaluated by numerous decision makers within
the Agency and the states with regulatory authority.
In order to expedite completion of the systems, the development process
for the systems was serial rather than parallel. Hence the systems are not
all at same stage of completion. The current contract effort will see the
first system, final cover, to completion, i.e., field tested on a limited
scale and ready for extensive testing by the targeted user community. The
second system, vegetative cover, will be completed through the stage of
field testing prototype. The third system, leachate collection, will be
completed through demonstration prototype. The fourth and final system,
run-on/off will not be completed at this time. Completion of these systems
waits a revaluation of needs in relation to other priorities and availabil-
ity of resources. Hopefully these systems can all be fully field tested
within the next calendar year. Until this is completed and a user support
and systems maintenance function are established, the systems will not be
ready for general release or use.
-224-
-------
DEVELOPMENT OF A NOVEL SUPPORT AERATED BIOFILM REACTOR FOR THE
BIODEGRADATION OF TOXIC ORGANIC^OTffbUNDS ---
by: Sandra Woods1, Kenneth Williamson1, Stuart Strand2,
Kelly Ryan , Jonathan Polonsky1, Roger Ely1, Kurt Gardner1
and Philippe Defarges
Department of Civil Engineering
Oregon State University
Corvallis, Oregon 97331
2
College of Forest Resources
University of Washington
Seattle, Washington 98195
ABSTRACT
«3,,rr Pf H u pr°J"ect is to develop a biofilm process that
successfully degrades hazardous wastes. This is accomplished through the
selection of microorganisms by supplying appropriate gases directly to a
hSn T PP°rted °? a gas-permeable membrane. A gas-permeable membrane has
been used successfully as a support surface for the growth of bacterial
labor^toyy reactors in two applications: (1) an enrichment culture
and (2) an enrichment consortium of
**cteri* are enriched on the gas-permeable membrane
and °xy?en to the base of the biofilm. The biofilm
followin8 chlorinated one- and two- carbon compounds in order
for n K anaerobic/aer°bic consortium, the reactor is operated to enrich
for anaerobic microorganisms in the biofilm towards the bulk liquid by
maintaining an anaerobic bulk liquid, and aerobes toward the surface of the
portioTof ythe°M f"? ^ ^^ ^^ the membrane S»™°^' The -aerobic
portion of _ the biofilm should facilitate reductive dehalogenation of
dS?adatioiCofTh°U2dh'iand the/erobic P°rtio" of the biofilm should allow
degradation of the dehalogenated metabolic products. This reactor
configuration has been used to successfully treat chlorophenols and
chloronitrobenzenes and is being evaluated for the treatment of
•chlorobiphenyls, chlorobiphenylols, and chlorobeazenes.
-225-
-------
INTRODUCTION
A biofilm reactor has been developed that may increase the potential
for biodegradation of a wide variety of trace toxicants by facilitating the
growth of specific enrichment cultures. The gas-permeable-membrane-supported
(GPMS) reactor differs from traditional biofilm processes in that electron
donors and/or acceptors are provided to the microorganisms by diffusion ot
gases through a gas-permeable membrane. In addition to supplying substrates
directly to the biofilm, the membrane also physically supports the biofilm and
the overlying bulk liquid.
The treatment of low-molecular-weight chlorinated solvents by
an enrichment culture of methylotrophic bacteria and the treatment of
chlorinated aromatic compounds by an aerobic/anaerobic consortium are
described in this paper. The objective of this project was to determine the
feasibility of the GPMS reactor for the treatment of a variety of hazardous
compounds through the selection of appropriate groups of microorganisms, ine
reactor was fed synthetic wastewaters containing the following trace
toxicants: chlorophenols, chlorinated nitrobenzenes, polychlorinated
biphenyls, methylene chloride, chloroform, carbon tetrachloride, 1,2-di-
chloroethane and cis 1,2-dichloroethene.
In one reactor, methane oxidizing bacteria were selected for study due
to their ability to degrade a wide range of organic compounds.
Methylotrophs use single carbon compounds as electron donors and carbon
sources with oxygen serving as the electron acceptor. They oxidize methane
to carbon dioxide through the use of several enzymes in a series of
reactions involving sequential transfers of two electrons. The initial
oxidation of methane to methanol is catalyzed by the methane monooxygenase
enzyme (MMO). Methanol is further oxidized to form formaldehyde, formic
acid and C09. The methane monooxygenase enzyme is thought to be largely
responsible for the broad degradative competence of methylotrophs (1).
The\ methylotrophic biofilm was developed by providing methane and
oxygen through the membrane. Because methane and oxygen are sparingly
soluble gases, the growth of methylotrophs was facilitated in the GPMS
reactor by providing methane and oxygen directly to the microorganisms
through the membrane.
In another reactor, an anaerobic/aerobic consortium was developed to
encourage the biodegradation of haloaromatic compounds. These compounds are
biodegraded by different mechanisms in aerobic and anaerobic pathways.
Under aerobic conditions, many chlorophenols, chlorobenzoates and
chlorobiphenyls are degraded by hydroxylation followed by cleavage of the
aromatic ring. Chlorines are removed from the compound after the ring is
cleaved (2). In contrast, a very different degradative pathway exists under
anaerobic conditions. Chlorophenols'(3,4) chlorobenzoates (5), and
polychlorinated biphenyls (6) are reductively dehalogenated under anaerobic
conditions. In the GPMS reactor, the dechlorinating ability of anaerobic
consortia and the ability of aerobic consortia to hydroxylate and cleave the
ring are both present and potentially increase the biodegradation of these
compounds as the metabolic products are exchanged between the two bacterial
groups by diffusion.
-226-
-------
An aerobic/anaerobic biofilra was developed by providing oxygen at the
base of the biofilm to encourage the growth of aerobic organisms and by
providing high concentrations of electron donors to develop anaerobic
conditions in the bulk liquid. The reactor was fed chlorinated aromatic
compounds in a nutrient media consisting of inorganic minerals, vitamins and
carbon sources.
EXPERIMENTAL METHODS AND ANALYTICAL PROCEDURES
Several independent studies were conducted (Table 1). In each study,
the biofilm reactor was altered as experience was gained in operation. The
reactor design, nutrient solution and analytical methods for each individual
study is described in the listed references.
TABLE 1. GPMS BIOFILM STUDIES
Compounds
Reference
Methylotrophic Consortium
chlorinated methanes
dichloromethane, 1,2-dichloroethane
and cis 1,2-dichloroethene
halo -methanes, -ethanes, and -ethenes
Aerobic/Anaerobic Consortium
chlorophenols
chlorinated nitrobenzenes
polychlorinated biphenyls
chloroaromatic compounds
7,8
8,9
in progress
10
11,12
13
in progress
SYSTEM DESIGN
A schematic diagram of the GPMS reactor system used to treat
chlorinated, low-molecular-weight solvents is shown in Figure 1 (8)
Methylotrophs were enriched by providing a 50/50 volume/volume mixture of
methane and oxygen through the membrane. The aerobic/anaerobic consortium
was enriched by providing pure oxygen through the membrane
-227-
-------
Nutrient Solution
Pump
Pressure
/Equalizer
Solution
Teflon Paddle
Gas-permeable Membrane
Teflon Seal
Liquid
Sampling
Port
(Continuous Flow)
Gas
Dryer'
<0—"-txHD^
Lo
pressure
Regulator
Cylinder
Figure 1. Reactor Schematic for the Growth of Methylotrophs
The reactor was constructed of standard 4-inch ID glass beaded process
pipe. Glass ports for the addition and removal of nutrient solution and
gaseous substrates, and for sampling the headspace and gas compartments were
added as shown. The reactor was built in three sections which were
connected by flanged joints sealed with Teflon gaskets. The gas-permeable
membrane was installed across the flanged joint between the gas and liquid
compartments. The membrane material selected after evaluation of several
materials was a nylon-backed Teflon laminate manufactured by W.L. Gore and
Associates (Elkton, Maryland) and commonly called Goretex. The reactors were
held in a 30°C incubator.
Gases were provided to the biofilm through the ports in the gas
compartment. A low pressure regulator was used by Ryan (9) and Polonsky
(13) to maintain a constant pressure of approximately five inches of water
in the gas compartment. Gas consumption was monitored by measuring pressure.
loss in the gas storage tanks. In studies by Ely (7,8) and Gardner (11) a
small gas flow rate was maintained through the gas compartment.
-228-
-------
NUTRIENT MEDIA
The nutrient media for the growth of methylotrophs was a variation of
the nitrate mineral salts media described by Whittenbury et al. (14). It is
a defined media containing inorganic nutrients and EDTA. Nitrate serves as
the nitrogen source.
The nutrient feed solution for the aerobic/anaerobic consortia
consisted of mineral salts, trace elements, vitamins and carbon sources
necessary to sustain aerobic and anaerobic bacterial growth. The solution
used to treat polychlorinated biphenyls (13) and chlorinated nitrobenzenes
(11) were developed based on a modification of the nutrient media described
by Owen et al. (15). Glucose, methanol, acetic acid and acetone were
provided as alternate carbon sources to generate anaerobic conditions in the
bulk liquid.
ANALYTICAL METHODS
Methane, oxygen, nitrogen, and carbon dioxide were measured using a
Fisher Model 25V gas partitioner with two columns, one containing 30% bis
(2-ethoxyethyl) adipate (BEEA) on 60/80 Chromosorb PAW and the other
containing 45/60 Molecular Sieve 13X (Supelco, Inc., Bellefonte,
Pennsylvania). Gas samples (100 uL) were collected using a Pressure-Lok,
gas-tight syringe and immediately injected into the gas partitioner for
analysis. Standards curves were prepared daily.
Concentrations of the chlorinated compounds were determined using a
Hewlett-Packard Model 5890A gas chromatograph equipped with a Hewlett-
Packard Model 3392A integrator. Chloromethanes, chloroethanes and
chloroethenes were determined by direct injection of liquid or gas samples
and gas chromatography with flame ionization detection. A 1/4-inch glass
column, 8 feet in length and packed with 60/80 Carbopack B with 1% SP-1000,
was used for the separation.
Chlorinated aromatic compounds were measured with splitless injection
of hexane extracts and gas chromatography with electron capture detection.
A 30 m x 0.33 mm DB-5 with at 0.25 urn film thickness fused silica capillary
column (J&W Scientific, Deerfield, IL) was used to separate PCBs. A DB-1
fused silica capillary column (J&W Scientific, Deerfield, IL) was used for
the separation of chlorophenols and chlorinated nitrobenzenes.
RESULTS AND DISCUSSION
METHYLOTROPIC CONSORTIUM
The original bacterial inocula for the selection of methylotrophic
bacteria consisted of 10 ml of thickened trickling filter effluent from the
Corvallis, Oregon municipal wastewater treatment plant and 10 ml of
thickened sludge from a bench-scale anaerobic digester in operation at the
Oregon State University Environmental Engineering Laboratory. The culture
was enriched for methylotrophs by providing methane as the principal carbon
source and by maintaining a moderate flow of nutrient solution (without
carbon sources) to wash out undesired organisms.
-229-
-------
Once the biofilm was established, the reactor was fed a group of
chlorinated low-molecular-weight compounds in a series of experiments. Mass
balances were performed for each of the chlorinated compounds throughout the
experiments by measuring concentrations of the compounds in the reactor
influent, liquid phase, headspace and gas compartment.
The removal of dichloromethane, chloroform, and carbon tetrachloride
was evaluated in a batch test. A total mass of 20.6 umol dichloromethane,
19.6 umol chloroform, and 42.7 umol carbon tetrachloride were added together
to yield initial liquid phase concentrations of 10.9 umol/1 dichloromethane,
9 88 umol/1 chloroform, and 19.59 umol/1 carbon tetrachloride. The liquid
phase concentrations of each compound decreased rapidly. After,seven days
(168 hours), the concentration of dichloromethane in the liquid phase was
below detection; the liquid concentration of chloroform fell, to less than
3.2 umol/1; and the liquid concentration of carbon tetrachloride fell to
4.6 uraol/1.
The mass balance for dichloromethane is shown in Figure 2. The total
loss of dichloromethane was 16.2 umol of which 4.4 umol escaped through the
membrane in the small gas flow through the gas compartment. Although an
initial increase in the mass of dichloromethane in the headspace was
observed, this mass eventually fell to near zero. This was due to an initial
equilibration of the headspace with the bulk liquid, and reequilibration as
the dichloromethane present in the bulk liquid was biodegraded. Based on
these data, 16.2 umol of dichloromethane was removed.
Chloroform also was removed during this experiment; however, removal
rates were much slower than the rate for dichloromethane. After seven days,
6.0 umol chloroform of the 19.6 umol initial mass remained in the reactor of
which 10.3 umol left the reactor through the membrane and gas compartment
(Figure 3). Based on these data, a maximum of 3.3 umol may have been
biodegraded.
Carbon tetrachloride data were similar to data for chloroform. Of the
initial mass of 42.7 umol added to the reactor, 17.5 umol remained in the
reactor after seven days; 16.3 umol left the reactor through the membrane
and gas compartment. Based on these data, a maximum of 8.9 umol carbon
tetrachloride was removed during this test. The initial removal of 8.9 umol
was probably largely due to sorption since the mass decreased quickly within
the first three days and then remained constant.
In a second series of tests with the methylotrophic biofilm (8,9), the
removal of dichloromethane (DCM), dichloroethane (DCA) and cis 1,2-
dichloroethene (DCE) was evaluated. The masses of the chlorinated compounds
in each compartment (liquid,.headspace and gas compartment) are presented in
Figures 4, 5 and 6. Immediate substantial decreases in the masses of
chlorinated compounds in the liquid were observed in all cases with a
corresponding increase in the masses of compounds in the gas and headspace.
This initial transfer resulted from the volatilization of these compounds.
A small fraction of the masses was unaccounted for and is likely to be sorbed
to the biofilm and reactor vessel.
A control study was conducted by adding dichloromethane (DCM),
dichloroethane (DCA), and cis 1,2-dichloroethene (DCE) without the presence
-230-
-------
7
§
I
a
20
IS
18
17
18
ts
14
13
12 -
11 -
10 -
8 "
8 -
7 -
8 ~
8 -
4 -
3 -
2 -
1 -
0 -
3
D
a
D
O O Ua«« In Liquid
O Total Vetafll* Urn
Q + Uan In H«ad*p In liquid
O Total VolaflU Lo.«
~4- Ua«* In Kaad«pac«
• I I. I T I 1 I 1 1 1
20 40 60 80 100 120
Tbn«. hr
140
ISO 180
2OO
Figure 3. Chloroform Mass Balance
-231-
-------
28
24
2O
"o
6 16
| 12
S
8
4
n
A A Total
• Liquid
• • Headspace
• Gas
A
gl A,
• A
« A
» A
-•••.."*• A A
• *•* m • ** f __
w • • • «•
AA.A *A.A A A 1 » A A * J- *-
|O
20 30
TIME (hours)
4O
50
Figure 4. Mass Balance for the Batch Degradation
of Dichloromethane
35
30
25
'o
e 20
His
s
10
5
n
1 1 1 i
A A Total
• Liquid
• AA • Headspace
A *Gas
'• A
* • A A
• A
• A"
m-
".** *•* • ••• 9'
IO
20
TIME (hours)
50
Figure 5. Mass Balance for the Batch Degradation
of 1,2-Dichloroethane
-232-
-------
35
30 *
1 25
3-20
| 15
IO
5
t >**«
Total
Liquid
Headspace
,«•*••*
A A
•
«
A A
O 20 40 60 80 100 120
TIME (hours)
Figure 6. Mass Balance for the Batch Degradation of cis 1,2-Dichloroethene
v«wa
_l 0.07
\
E
0 0.06
J 0.05
£
§
£ O.O4
£
§ 0.03 -
C 0.02 -
a
o
0 0.01 -
o.oo -
« * BI-2,4
X
X B1~3.5
° I
x
•
o
o
X
0 20 ;0
Tim*. hr«
* 81 -2.4 x 81-3,5
Figure 7. Liquid Phase Dichloronitrobenzene Concentrations
During Batch Biodegradation
-233-
-------
of the biofilm. The concentrations of the compounds were monitored over a
40-hr period and no significant loss of the chlorinated compounds was
measured. These results showed the absence of significant abiotic removal
mechanism such as hydrolysis or adsorption to the reactor.
AEROBIC/ANAEROBIC CONSORTIUM
Similar reactors were operated with oxygen supplied at the base of the
Goretex support. These reactors were operated at moderately high organic
loading rates to generate an anaerobic bulk liquid and a thick biofilm; an
aerobic zone existed at the base of the biofilm.
Dichloronitrobenzene (DCNB) removal was evaluated in a series of batch
and continuous flow experiments (11). Two dichloronitrobenzenes (2,4-DCNB
and 3,5-DCNB) were tested to evaluate the effect of chlorine position on
removal. The concentration of both compounds in the reactor liquid rapidly
decreased (Figure 7). The ortho-para substituted dichloronitrobenzene was
removed slightly faster than the meta-meta substituted compound. No
metabolic products appeared in the batch reactor during this experiment.
However, subsequent continuous flow experiments indicated the presence of
chlorinated anilines, which result from reduction of the nitrate functional
group to an amine.
SUMMARY AND CONCLUSIONS
A variety of chlorinated compounds were successfully biodegraded in
with a gas-permeable membrane-supported biofilm under two applications: (1)
an enrichment for methane-oxidizing bacteria and (2) an enrichment
consortium of aerobic and anaerobic bacteria. By controlling the type of
gas transferred through the gas-permeable membrane, the electron donors and
electron acceptors and bacterial consorium was successfully selected for the
compounds to be removed.
A biofilm of methane oxidizing bacteria was grown on the gas-permeable
membrane by providing methane and oxygen to the base of the biofilm. The
biofilm degraded the following chlorinated one- and two- carbon compounds in
order of decreasing removal rates: dichloromethane, chloroform, 1,2-di-
chloroethane and cis 1,2-dichloroethene. Carbon tetrachloride was not
biodegraded.
A biofilra comprised of anaerobic and aerobic layers was developed by
providing high organic substrate levels in the liquid compartment and oxygen
to the gas-permeable membrane. The anaerobic/aerobic biofilm successfully
biodegraded chlorophenols and chloronitrobenzenes.
-234-
-------
REFERENCES
1. Higgins, I., Best, D., and Hammond, R. New findings in methane-
utilizing bacteria highlight their importance in the biosphere and their
commercial potential. Nature, 286: 561, 1980.
2. Hartmann, J., Reineke, Wl, Knackmuss, H.J. Metabolism of 3-chloro-, 4-
chloro-, and 3,5-dichlorobenzoate by a Pseudomonad. Applied and
Environmental Microbiology, 37: 421, 1979.
3. Woods, S. The fate of chlorinated, hydroxylated and methoxylated
benzenes in anaerobic wastewater treatment. Dissertation, University of
Washington, Environmental Engineering and Sciences, Seattle, Washington,
1985. 301 pp.
4. Boyd, S. and Shelton, D. Anaerobic biodegradation of chlorophenols in
fresh and acclimated sludge. Applied and Environmental Microbiology 47:
272, 1984.
5. Suflita, J., Horowitz, A., Shelton, D., and Tiedje, J. Dehalogenation:
a novel pathway for the anaerobic biodegradation of haloaromatic
compounds. Science 218: 1115, 1982.
6. Low, E. and Woods, S. Reductive dechlorination of chlorobiphenyls
during anaerobic treatment. Submitted to Environmental Science and
Technology, 1988.
7. Ely, R. Startup and performance of a gas-permeable-membrane-supported
(GPMS) biofilm system using a mixed culture of methylotrophs to degrade
methylene chloride, chloroform, and carbon tetrachloride. Thesis,
Department ,of Civil Engineering, Oregon State University, Corvallis,
Oregon, 1986.
8. Ely, R., Ryan, K., Woods, S., Williamson, K., and Strand, S. Treatment
of chlorinated one- and two- carbon compounds by methylotrophs using a
gas-permeable-membrane-supported-biofilm reactor. In preparation, 1988.
9. Ryan, K. Development of a method to elucidate biodegradation pathways
of chlorinated one— and two— carbon compounds using a gas—permeable—
membrane-supported raethylotrophic biofilra. Thesis, Oregon State
University, Department of Civil Engineering, Corvallis, Oregon, 1988.
10. Defarges, P. Development of a submerged aerated biofilm process for the
degradation of chlorophenols. Thesis, Oregon State University,
Corvallis, Oregon, 1987.
11. Gardner, K. Development of a method to accumulate metabolic products
from dichloronitrobenzenes using a gas—permeable—membrane—supported
biofilm. Thesis, Oregon State University, Department of Civil
Engineering, Corvallis, Oregon, 1987.
-235-
-------
12. Woods, S., Ryan, K., and Gardner, K. Development of a method to
accumulate metabolic products. In preparation, 1988.
13. Polonsky, J. Development of a gas-permeable-membrane-supported biofilm
reactor for the combined aerobic/anaerobic treatment of chlorinated
biphenyls. Thesis, Oregon State University, Department of Civil
Engineering, Corvallis, Oregon, 1988.
14. Whittenbury, R., Phillips, K., and Wilkinson, J. Enrichment, isolation,
and some properties of methane-utilizing bacteria. J. Gen. Microbiol..
61: 205, 1970.
15. Owen, W., Stuckey, D., Healy, J., Young, L., and McCarty, P. Bioassay
for monitoring biochemical methane potential and anaerobic toxicity.
Water Research, 13: 485, 1979.
-236-
-------
MICROBIAL MINERALIZATION OF ORGANIC CONTAMINANTS
ON SOILS AND SOIL FRACTIONS
D.S. Kosson, W.V. Black and R.C. Ablert
Rutgers, The State University of New Jersey
Department of Chemical and Biochemical Engineering
P.O. Box 909
Piscataway, NJ 08855-0909
ABSTRACT
The focus of this paper is preliminary evaluation of biodegradation
of organic contaminants on excavated soils and separated soil fines
using activated sludge populations in slurry reactor configurations.
Contaminated soils display a wide variety of sorbed and chemically bound
organic species; generally, only a limited number are present in high
concentrations. Priority pollutants usually represent only a small
percentage of total leachable or extractable contaminants capable of
impacts on public health and environmental quality. Given the diversity
of contaminants present, serial microbial processing, controlled to
maximize the rate of individual degradation steps is being investigated.
Several soil samples have been obtained from locations at a CERCIA-NPL
site^immediately adjacent to an impoundment used for disposal of
distillation bottoms from a benzene, toluene and xylene (BTX) production
process.
INTRODUCTION
An increasing number of dump and disposal areas for hazardous and
toxic substances have been identified. Within reasonable time, these
sites must be renovated and their potential for impacts on public health
and the environment mitigated completely. In typical circumstances, one
or more waste materials have been buried or discharged with little or no
provision for long-term containment. Many dump sites have received
wastes from several sources and, as a direct consequence, the
contaminant matrix is extremely complex and includes organic species,
heavy metals, inorganic salts and oxides, and/or inert materials under
strongly acidic, basic, oxidizing or reducing conditions.
The current physical and chemical state of a waste may not reflect
the initial state. The original states of abandoned wastes include
hydrophilic and hydrophobia liquids, gels, sludges, powders, dispersions
of particulates and/or liquids, emulsions, tarry or viscous materials,
'sorbed chemicals and agglomerates. The current state is a result of a
-237-
-------
site- specific scenario that could involve container failure, chemical
reactions between wastes, weathering (natural or enhanced acidic
rainfall, oxidation, etc.). aerobic/anaerobic microbial degradation, and
interactions with soil and water.
Once a waste disposal site has been identified, a series of actions
must be carried out to establish the chemical, physical, and biochemical
character of the contamination, the distribution and extent of dispersal
of contaminants, and the potential for rapid migration by diffusion or
gross movement via air or water. After the scope of the problem has
been outlined, a plan for containment, stabilization, clean-up and
ultimate destruction or permanent (reversible) isolation of recovered
waste, or waste by-products, can be developed.
Transfer off-site for complete destruction of wastes or
decontamination of soils is one available option. High-temperature
processing and oxidation, e.g., incineration, ozone, wet air oxidation,
etc., have been demonstrated. Disadvantages include high unit cost,
high capital cost, high containment and hauling costs, problems
associated with gaseous and solid residuals disposal and mechanical
problems relating to handling large quantities of inert material (soil).
On-site management of hazardous wastes avoids many transfer-related
problems. Personnel and environmental exposures during handling,
packaging and transportation are minimized and localized. Responses to
leaks, spills, equipment failures, gaseous emissions and other
fortuitous events can be rapid and comprehensive. In short,
contamination can often be confined to an area already judged to be
contaminated; personnel protection and training can be consistent with
the requirements of site activities. Physical, chemical, and biological
treatment methodologies can be made available at most sites. The goal
of treatment may be waste destruction, detoxification or stabilization
for long-term, local storage. In many instances extraction of water and
excavation of soil for treatment, as opposed to in-situ treatment, will
achieve better control, accountability and effluent quality assurance.
Many believe the ideal solution is permanent, on-site destruction.
APPROACH
Microbial degradation with activated sludge inocula has been used
with great success for total mineralization of organic residues. In
particular, an array of high-strength, hazardous, organic solutes in
aqueous systems has been destroyed by aerobic, anaerobic or a
combination of aerobic and anaerobic bioreaction steps. The
distribution of organic contaminants between soil and water is specific
to molecular type, soil characteristics, water content, pH and oxygen
tension. Factors of concern include:
i. Simple sorption equilibria with single solutes and soil
constituents acting independently.
-238-
-------
ii. Competitive sorption of organic solutes on individual or
interacting soil fractions.
iii. Chemical reactions with natural organic matter, i.e., hydrogen-
bonding, esterification, etc. . .
1V' S1°\anaerobic fermentation with fatty acid production, low pH
and highly reducing conditions.
v. Entrapment in clay mineral structures, such as irreversible
clathyration and catalytic conversion by specific clay mineral
ions or crystal configurations.
vi. Physical immobilization of contaminant molecules in small
pores and capillaries by dewatering, surface tension effects, or
restricted diffusion paths.
of t- ?trate/gy *S dictat*d in large measure by the combination
of those contaminant/soil/water interactions that influence behavior at
a site. Four general behavior patterns, or "Cases", can be cited to"
illustrate the options :
Case I;
Readily reversible sorption equilibria and partition
coefficients that favor groundwater and soil water.
Case II: Reversible sorption and/or reaction with kinetic, diffusion
rate limitations, or unfavorable partitioning.
Case III: Irreversible chemisorption, chemical binding, and physical
Case IV: Immiscible organic phase stabilization by surface tension
(capillary) effects.
A number of process steps can be combined into a complete scheme
for separation and elimination of polluting organic species. cKeTa
range of natural organic fractions and clay mineral species in soils
the organic contaminants can be anticipated to include normally volatile
and p!2 poljutant?- solvents, PCDDs and PCDFs, PCBs, pesticides, PNAs
and PAHs, and poorly defined complex solutions from manufacturing and
blending operations. Much effort has been expended to define the
efficiencies, costs, limitations, and final quantitative outputs from
individual operations and some combinations of operations. The major
exception is biological treatment of contaminated soil; this technology
is virtually unexplored. &J
Mno StUdy iS t0 characterize the cost effectiveness and
biochemical efficiency of microbial treatment for leachates and fine
fractions of excavated soil. Contaminants that occur at various actual
hazardous waste disposal sites and can be expected to reflect a
diversity of organic compounds, including many that are nominally
-239-
-------
xenobiotlc.
A treatment process for surficial soils will employ excavation,
component separation, chemical pretreatment, sequenced microbial
degradation processes, soil reconstitution, and soil replacement. Soil
excavation permits greater process control and variety of process
conditions than in-situ treatment. Soil component separation is an
attractive process step. Often a major portion of the contaminants are
present on a particular soil fraction, for example, contaminant binding
to a soil may be dictated by solute adsorption to the clay mineral
fraction or by partitioning with the natural organic (humic) fraction.
In many cases, acceptable remedial efficiency may be achieved
through special treatment of the small, quantities of highly
contaminated soil fines that are obtained through dry fractionation.
Alternately, the fine fraction may be generated as a slurry in a wash or
extraction solution formulated with surfactive agents and/or chemicals
to redistribute sorbed contaminants. The focus of the project is
development of a series of microbial reactors for treatment of soil
fines (containing clay and humic fractions) and slurries of fines in
leachate or extraction solutions.
Contaminated soil will display a wide variety of organic
contaminants with a limited number of species present in high
concentrations. In general, the sum of priority pollutants present
represents only a small percentage of total leachable or extractable
contaminants. Thus, key species and gross parameters (i.e., COD, TOC,
etc.) are usually employed as indicators of treatment effectiveness.
Given the diversity of contaminants potentially present, a series of
microbial processes using activated sludge inocula in separate reaction
vessels, controlled to maximize the rate of individual degradation steps
at particular conditions, is a promising treatment technology. Specific
organic species have been shown to be biodegraded through multiple
pathways. There is no intent to use genetically engineered or specially
cultured microorganisms or enzymes derived therefrom.
Specific sequences of aerobic and anaerobic treatment steps should
be based on (i) the nature and diversity of contaminants present, (ii)
the type of soil/contaminant binding, and (iii) regulatory expectations.
Use of enriched oxygen supplies or elevated temperatures and pressures
should also by included. An additional process variable is moisture
content. Moisture content may be adjusted to enhance microbial
processes or to facilitate materials handling. Most treatment steps
will require modification of nutrient levels and pH corrections to
accommodate the microorganisms used.
The technical organization of the project is a phased approach.
Phase One consists of the identification of a suitable matrix of soil
types, contaminant-soil binding types, and organic contaminant mixtures.
This matrix represents soil contamination occurrences at CERCLA-NPL
sites. Phase Two of the project is devoted to the investigation of the
-240-
-------
wnonerJ *icrob1ial process stePs- Batch and continuous reactors
will be operated to evaluate competitive effects resulting from mixed
organic species on microbial processes. This phase is orient" Sward
mav £S T and.evaluation of P^cess sequences. Pretreatment steps
afoxJdatio Sepa5af°nS ?f soil ^actions and chemical reactions, suchP
r«m * ° or hydrolysis, to selectively transform refractory organic
compounds into intermediates more amenable to biodegradation. Later
!omneSt- the act*vity wil1 addre** overall process selection criteria
applications?"8 dem°nstrations ' and scal«- P procedures for prototype
SITE DESCRIPTION
Soil samples were obtained from areas immediately adjacent to a
nrodu0^ J CERCLA-NPL site where distillation bottoms from a BTX
production facility were impounded. The lagoon is in excess of ten
The ?J * contains more than 100,000 yd* of impounded material.
The lagoon contents have separated into several layers including from
lagoon bottom to surface, a solid mixed organic/inorganic layer*' ™
MeratioTo/l °US """^ °rganlC layer« and fl°ati^ •W™
Migration of lagoon contents to neighboring soils has occurred.
n,™< °f the S°ils samPled are provided in Table 1. In
addition, the presence of polynuclear aromatic species (PNAs) was
' Mgh preSSU?e Hquif chroISography
with T mpleS Were seriaHy extracted fhree7
°
chromatogram and a PNA locator standard are provided in Figures la and
± •M«UB?ry ?f identified sPecies and order- of -magnitude^
Resent provided in Table 2; numerous unidentified species are
BIODEGRADATION IN SLURRY REACTORS
Initia?- studies have been focused on using an activated sludge
!PrS ^ bUlk S°n SampleS in an a^Ueous nutrientgmedia.
Tre f "dSe4samPles were obtained from the Somerset-Raritan Sewage
Treatment Authority, a secondary treatment facility that receives
significant inflow fraction from industrial wastewater. The intent of
these studies has been to establish a baseline for comparison of
emneovpd micr°bial Processes and reactor configurations . Sucrose was
employed as a co-substrate to avoid substrate-limiting initial
conditions. The reactor was operated in a batch mode. These
conditions were considered. to be the simplest meaningful configuration
-241-
-------
TABLE I
PROPERTIES OF SOIL SAMPLES
Property Sample 1-S
Sample 1-E
pH:
Texture (%):
Sand
Silt
Clay
Cation Exchange
Capacity (meq/gm):
Principle Exchange-
able Cations: .
(raeq/lOOgm)
3.6
66
26
8
6.6
3.5
70
20
10
5.7
Ca
Mg
K
Na
0.04
0.10
0.004
0.15
0.17
0.30
0.05
0.07
PAH Standard
0.40
0.3S
0.300
0.250
I"
' 4
T 1 r-
e 10 12
MINUTES
Flgur* 1«
14 tQ IB 20
-242-
-------
extract of ¥MI IK-T-I« to-a
4.OOO-
3.500-
3.000-
a.soo-
8.000-
1.500-
1.000-
0.500-
o.ooo-
* lo
«IHUTES
,
14
is
,
IB
ad
TABLE 2
Semi-Quantitative Soil Assays For PNAs
Peak
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Time
CMirO
7.97 ]
8.87 t
9.87 j
10.14 ]
10.76 1
11.34 t
12.04 I
12.44 I
14.11 I
14.45 C
15.81 B
16.49 B
16.93 B
18.04 D
18.31 B
18.86 I
Locator Mix
Napthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(A)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
10
20
10
2
1
1
2
1
1
1
,2
1
Benzo(a)pyrene 1
Dibenzo(a,h)anthracene 2
Benzo(g,h,i)perylene 2
Indeno(l,2,3-c,d)pyrene 1
5100
80
700
600
90
2
6
4
40
1
8
3
1
7
32
-243-
-------
a ±: SEE-.
5 min. 2mL/min. 100% acetonitrile
Typical process operating conditions and responses are £"«ibed as
«i« Initially a 5-L fermentor was charged with 500g of soil, i*S
sorpcive Equilibrium with dispersed soil J^ ""'"S™^ 7 0-7 1
temperature and PH were maintained at 4 L(STP)/nn.n, 30 C, and 7.0 7.1.
respectively using automated feedback controls.
Carbon dioxide evolution, TOG, and acid production was »°nitored
'
rate of 210 mg C/hr. The final TOG of 110 mg/L represented 94% TOG and
36% HOC reduction (on a. concentration basis). .............
CARBON DIOXIDE EVOLUTION
0.14
/
O.OB
I 1
20
HUE (HOURS)
t
30
•W
-244-
-------
TOTAL ORGANIC CARBON
•
£ o
ACID PRODUCTION
A mass balance on carbon was calculated at the conclusion of each
trial. For the typical case presented, 6.1 g C was removed from the
aqueous solution. Probable carbon end-states include cell mass
-245-
-------
production, carbon dioxide evolution and partitioning with
soils. A total of 1.5 g C was converted to carbon dioxide Currently,
analytical procedures are under development to quantitatively
differentiate between partitioned organic carbon and cell mass in the
dispersed heterogeneous solid phase. Post-fermentation extractions
followed by HPLC analysis yielded inconclusive results for indicator
PNAs.
CONCLUSIONS
Investigations are in progress to delineate the capability of
different activated sludge-based microbial process scenarios and reactor
configurations to degrade organic contaminants present' in or bound to
soils. A fractional factorial matrix with contaminant(s) type soil
type, microbial process, and reactor configuration *s primary fac^s
will serve as a basis for comparison. Subsequently, individual process
steps will be evaluated in series to provide a basis for overall process
evaluation.
Preliminary studies have been carried out to examine the
biodegradation of organic species in soils contaminated by BTX
distillation bottoms Results from a slurry reactor with dispersed_soil
and activated sludge culture will serve as a basis for comparison with
alternative microbial (aerobic and anaerobic) process steps.
Substantial methods development is necessary for clear evaluation ot
treatment effects on contaminants in the solid phase.
ACKNOWLEDGEMENT
This work was supported in part by USEPA Cooperative Agreement No.
CR-814055. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use by USEPA or by Rutgers
University.
-246-
-------
BIOLOGICAL TREATMENT OF LEACHATE/NEW LYME, OHIO
by: Edward d. Opatken, Hinton K. Howard, James J. Bond
U. S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio 45268
ABSTRACT
Studies have been completed on treating a leachate from New Lyme
Ohio. The Leachate was transported to Cincinnati, Ohio, where a pilot-
sized rotating biological contactor (RBC) was used for a treatment
evaluation.
^ pTJe b]om?ss was developed on the RBC discs with primary effluent from
the City of Cincinnati's Mill Creek Sewage Treatment Facility. Experiments
were then conducted to determine the effectiveness of treating a hazardous
waste leachate and to provide information on the following:
The kinetic rate of organics removal;
° The final effluent quality;'
° ™e fate of Priority pollutants and specific organic compounds:
The loss of volatiles via stripping in the RBC.
This paper reports on the results from these experiments and the
applicability of an RBC to treat a hazardous waste leachate from a
Superfund site.
BACKGROUND
Land disposal of organic wastes is no longer regarded as an ultimate
solution for solid wastes because of the potential for leachate formation
and/or groundwater contamination. Leachate releases have occurred at waste
disposal sites and technology is needed to remedy these conditions.
Biodegradation of the organics in leachates may provide an acceptable
and cost effective alternative for treating leachates. The rotating
biological contactor (RBC) was selected as a biological treatment process
to study its applicability for converting the organics in leachate into
innocuous products. A pilot sized unit was installed at the U S
Environmental Protection Agency's Test and Evaluation Facility (J&E) in
Cincinnati, Ohio, to conduct such an investigation.
-247-
-------
NEW LYME SUPERFUND SITE
The New Lyme Superfund site (1) was selected as the source for the
leachate. New Lyme is located in Ashtabula County, Ohio, approximately
40 kilometers (25 miles) north of the city of Youngstown Ohio. The site
consists of 14 to 16 hectares (35 to 40 acres) of landfill and began its
operations in 1969. It operated as a combined municipal and industrial
waste landfill where approximately 230,000 cubic meters (300,000 cubic
vards) of wastes were disposed of each year. The industrial wastes
included paints, solvents, oils, coal tar distallates and corrosive
liquids. In 1978 the landfill's license was revoked and the site was
subsequently placed on the National Priorities List under f^-P^nsive
Environmental Response, Compensation and Liability Act of 1980 (CERCLA). A
remedial investigation (2) was conducted in 1983 and concluded that:
0 the soils around the site;
0 the groundwater under the site;
0 the sediments in Lebanon Creek adjacent to the site;
had been contaminated by the landfill.
Samples were obtained from several seeps surrounding the landfill and
the results showed that the leachate contained up to 2000 mg/L of dissolved
organics. The leachate was pumped from a deep depression into a tank truck
and transported from New Lyme to Cincinnati for experimentation using the
RBC.
PROJECT DESCRIPTION (FIGURE 1)
A pilot sized RBC that contains 1000 m2 (11,000 square feet) of
surface area which is roughly 10% the scale of a full sized unit (100,000
square feet) is located at EPA's T&E Facility (3). The diameter is 3.6 m
(12 ft), identical to a full scale unit, but the length is only 1 m
(3.3 ft) whereas a full scale unit is 7.6 m (25 ft).
The RBC was operated in a batch mode. The leachate was transferred
from a storage tank to a mix tank where the volume was measured. The
nutrients were then added for specific experiments. The contents were
pumped into the RBC, and the RBC was operated at a constant speed of
1.5 rpm. When the reaction was complete, the contents were returned to the
mix tank and then fed to the clarifier for solids removal. The overflow
went to the receiving tank, where the treated leachate was stored until the
analytical results showed that the material was within the allowable
discharge limits.
Effluent limitations on the treated leachate were imposed by the
City of Cincinnati's Mill Creek Treatment Facility (MCTF as well as by
laboratory prescribed contaminant levels. Both sets of limitations had to
be met before the treated leachate was disposed of in the MCTF sewers. The
effluent limitations set by the MCTF were:
-248-
-------
15
y
tf>
U
CD
i_
D
u.
-249-
-------
total organic halides (TOX) <_ 5 mg/L;
0 vapor space organics (VSO) <_ 300 ppm (4);
0 6 < pH < 10.
The internally applied limits were based on achieving a soluble gross
organics level that would be equivalent to or less than a relatively high
strength raw wastewater:
o
o
o
soluble biochemical oxygen demand (SBOD) <_ 100 mg/L;
dissolved organic carbon (DOC) < 100 mg/L;
soluble chemical oxygen demand TSCOD) <^ 300 mg/L.
The underflow from the clarifier was analyzed for priority pollutants
and drummed for disposal at a permitted landfill. We did experience a
potentially unsafe incident with the drummed sludge. A drum bulged during
storage. A gas analysis showed a high concentration of methane, which was
probably caused by anaerobic degradation of the sludge. The pH of the
sludge was then raised to 12 by adding caustic to deactivate the biomass
and this action prevented a recurrence.
RESULTS
EXPERIMENTAL CONDITIONS
There were 8 experiments conducted with New Lyme leachate that varied
in:
o
o
o
o
volume of leachate fed to the RBC;
ratio of leachate to primary effluent (L/PE/);
nutrient addition to the leachate;
biomass condition (thickness, appearance, acclimation),
The initial runs used a combination of leachate and primary effluent to
allow for acclimation of the biomass which was formed with primary effluent
from the MCTP (5). The initial runs also used 2800 L (750 gal) instead of
3800 L (1000 gal) in order to conserve leachate. Nutrients were added to
the leachate for experimentation to study this effect on the reaction. A
fixed ratio of carbon (C)/nitrogen (N)/phosphorus (P) of 100/5/2 was used
in three of the runs. These experimental variations are summarized in
Table 1.
REACTION RATE
The reaction rate was determined by following the drop in the DOC with
time (6). A plot of experiments 1 and 4, are shown in figure 2, and
experiments 3 and 5 are displayed in figure 3. Experiments 2 and 7 are
shown in figure 4. Experiment 6 is not included in the figures because its
purpose was to determine the mass of volatile organics that was stripped by
the RBC without any biomass. Experiment 8 is omitted because the leachate
had undergone spontaneous biodegradation in the storage tank and was not
-250-
-------
l/Biu _ uoqjeo 01UB6JQ peA|oss|a
c
oc
o
11
I- DC
1/BlU -
0|UB6J0 PeA|088la
-251-
-------
(spuesnom)
l/Biu - uoqjeo oiueBjQ
-252-
-------
TABLE 1. EXPERIMENTAL CONDITIONS
Runs
01
02
03
04
05
06
07
08,
Total
Vol.
Gal.
750
750
750
750
1,000
1,000
1,000
1,100
Ratio
Leachate/PE
Gal /Gal
250/500
500/250
500/250
500/250
1,000/0
1,000/0
1,000/0
1,100/0
Nutrient
Addition
No
No
Yes
No
Yes
No
Yes-N/No-P
Yes
Rat i o
C/N/P
—
—
100/5/2
—
100/5/2
100/5/0
100/5/2
Biomass
No Acclimation
Poor
Satisfactory
Satisfactory
Satisfactory
None
No Acclimation
No Acclimation
Comments
Volatile organics
stripping run
representative of New Lyme leachate. The DOC at the start of experiment 8
had dropped to 280 mg/L, whereas the DOC for experiment 6 was 1780 rrig/L.
*++ - Sh°? an 1n1t1al dr°P in the DOC concentration and this is
attributed to- the volatile organics being stripped from solution (see
Figure 2, Experiment 4). There is evidence of this drop-off in each of the
experiments. To verify this assumption, experiment 6 was operated without
biomass and both the exhaust air from the RBC and the leachate in the RBC
were analyzed. Both analyses confirmed that volatile organics were being
stripped from the leachate. Provisions may be necessary to capture the
volatile organics from an RBC to prevent air contamination when treating
leachates with an RBC.
After stripping there was a period with a reduced reaction rate which
was attributed to acclimation of the biomass to the leachate (see Fiqure 2
Experiment 4). There was only a slight drop in DOC during this time inter-
val. The acclimation period varied but was especially evident in experi-
ments 2 .and 7 where the acclimation period lasted for 20 and 40 hours
respectively. The lengthy acclimation period for experiment 2 was attrib-
uted to an inadequate biomass. Prior to the start of experiment 2 the
biomass was observed to be reddish-brown and spotty with a noticeable
decrease in thickness, indicating that considerable sloughing-off of the
biomass had occurred. The extended acclimation period for experiment 7 was
attributed to the fact that it followed the volatile organics stepping run
where the biomass had been removed from the discs. Therefore there was no
prior acclimation of the biomass to the leachate.
-253-
-------
Following the acclimation period, the reaction progressed at a con-
stant rate, as shown in Figure 2, Experiment 4 by the disappearance of DOC.
The reaction rate varied for each experiment ranging from a high of JbU g
DOC concerted per hour to a low of 82 grams per hour. The two experimental
runs that had rates equal or greater than 300 grams per hour contained
adequate levels of both nitrogen and phosphorus and had an adequate accli-
mation period from previous runs. It appears that the nutrient addition
for experiments 3 and 5 may have contributed to the high organic removal
rate. However, additional cause and effect experiments are needed to
verify that nutrients may influence the reaction rate. The reaction rates
for the 6 experiments and the ratios of C/N/P are shown in Table 2.
TABLE 2. RATES OF REACTION AND NUTRIENT RATIOS
Expt.
1
2
3
4
5
7
DOC Removed
g/h
140
82
350
180
300
120
Ratios
100/5. 5/<. 2
100/l/<.2
100/5.5/1.4
100/1. l/<. 2
100/5/2
100/6. 7/<. 2
Note: The C/N/P ratios shown in Table 2
are different from the earlier values
given when describing the experimental
conditions. The values in this Table
reflect the contribution of the primary
effluent and the leachate.
POLLUTANT REMOVALS
The experiments with New Lyme leachate were analyzed for typical
biological parameters and priority pollutants. The removal of gross>
ics was determined by analyzing the influent and effluent for DOC, SBOD,
SCOD, TOC, total BOD (BODT) and COD . Other biological parameters that
were analyzed included sulpended solids (SS) and volatile suspended solids
(VSS), as well as the nutrient chemicals, both the nitrogen series and
phosphorus (7).
The priority pollutant levels for untreated New Lyme leachate were
compared with the effluents from the RBC after each experiment and on occa-
-254-
-------
si on compared with the sludge from the clarifier. Several volatile organ-
ics that were present in measurable quantities were identified during the
priority pollutant analyses and their removals were occasionally followed
during the RBC experiments. These results indicated that the removals of
the pollutant by the RBC was effective for all the contaminants. The gross
organics, as represented by DOC, SBOD and SCOD, were reduced to levels
approaching a treated municipal secondary effluent.
A summary chart is shown in Table 3 for the biological parameters
The data for experiments 2, 3 and 4 were averaged since they were all
run at the same leachate/PE ratio.
Only in the effluent from experiment 1 was a priority pollutant,
methylene chloride, detected. Otherwise the biological treatment was
effective for converting the priority pollutants into innocuous products.
Sludge samples were analyzed following experiments 2 and 5. Toluene
was the only priority pollutant identified both times as present in the
sludge. It was not determined whether the toluene was present in the
liquid or solid phase. However, since toluene was not detected in the
effluent, it is assumed that the toluene was associated with the solids
in the sludge.
CONCLUSIONS
, The study on treating the New Lyme leachate with an RBC showed very
good removals/conversions of the organics and several experiments had
reaction rates that were very satisfactory. The RBC is an effective
process for leachate treatment.
The RBC is an effective treatment process for reducing the
organics in New Lyme leachate.
Priority pollutants were converted and/or stripped from the
leachate during treatment.
The effluent quality is equivalent to a secondary treated
effluent from a municipal wastewater treatment plant.
Nutrient addition, at ratios of C/N/P of 100/5/2, shows
evidence of favorably influencing the reaction rate.
Volatile organics are stripped from the leachate by the RBC
action when operated in a batch mode.
The RBC experiences a period of inactivity near start-up
that is assumed to be caused by the need for the biomass
to acclimate to the leachate.
The RBC offers a high degree of treatment for processing leachates
from a hazardous waste site. Additional research is needed such as
-255-
-------
LU
z
ac
o
O
_l
*£.
UJ
oe
cc
o c o o o
o o o o o
vo *r *r co c
CM CM r-l i-l CO «3-
I i-l LO CM O O
CM «3- LTJ en CM
t-1 CM
co
• CO
O O O C O I
O O O C O I
C O i-< r~ t-<
CO CO ^H ^H CO •<*•
o o co CM o o
co co CD CM r-
C\J f—I CM CO
o o o o o o
.O O O O C O
cr|oo o CM cvj csj 10
CM CJ in m
r*. CT> tc en
•-< •-< vo vo
.0 O O O O C
cno O 10 co o o
— U3 t-» CO *-t CM
r-< CM CM
o
O
o o
o o
1-4 en
CM in
C CD UJ
a) j-v
3 «-
i— O 01
n- ^ <£
Ul 4-> t-
-------
determining the degree of organic stripping while the biodegradation
reaction is occurring, and the validation of rapid reaction rate kinetics
when the nutrients are controlled at their theoretical levels.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
Draft Report of "Treatabil ity Testing and Field Investigation Report
New-Lyme Landfill," U.S. Army Corps of Engineers, Omaha District,
January, 1987.
Final Remedial Investigation New Lyme Landfill Site, Ashtabula County,
Ohio, CH2M-Hill, February 6, 1985.
Brenner, R. C., Heidman, J. A., Opatken, E. J., Petrasek, A. C.,
"Design Information on Rotating Biological Contactors," EPA-600/2-
84-106, June 1984.
"Vapor Space Organics," In-House method developed by the Metropolitan
Sewer District of Greater Cincinnati, Industrial Waste Section,
Cincinnati, Ohio.
Opatken, E. J., Howard, H. K., Bond, J. J., "Biological Treatment of
Hazardous Aqueous Wastes," Second International Conference on New
Frontiers for Hazardous Waste Management, September 1987.
Opatken, E. J., "An Alternative RBC Design - Second Order Kinetics,"
Environmental Progress, Vol. 5, No. 1, February 1986.
Antonie, R. I.., Fixed Biological Surfaces - Hastewater Treatment,
CRC Press, 1976.
-257-
-------
TRIAL BURN REPORTING AND DEVELOPMENT OF PERMIT CONDITIONS
by: Sonya M. Stelmack
U.S. EPA, Office of Solid Waste
Washington, DC 20460
C. C. Lee
U.S. EPA, Office of Research & Development
Cincinnati, OH 45268
ABSTRACT
A technical guidance manual has been developed to assist
permit writers in one of the most complex and time-consuming
portions of the incinerator permitting process: evaluation of
trial burn results and translation of these results into
facility-specific operating conditions. The manual is also
intended to assist applicants in designing a trial burn
operational protocol and in reporting of trial burn data.
The manual sets forth the key operating parameters which
are important to maintaining compliance with the incinerator
performance standards. The parameters are divided into groups
depending on whether they are based on trial burn results or
manufacturer's specifications, and whether they should activate
automatic waste feed cutoff when exceeded. Methods for setting
levels for these permit conditions and for planning trial burns
are described.
The guidance also describes the information which should
be included in a trial burn report and provides suggested
standard reporting forms. These forms are expected to increase
uniformity of trial burn reports and reduce the need for the
permit writer to request additional information, thereby
speeding the review process.
DISCLAIMER
The guidance document described in this paper is a draft,
and has not been formally released by the U.S. Environmental
Protection Agency. Neither the guidance document nor this
paper should be construed to represent Agency policy. The
-258-
-------
guidance is being circulated for comments on its technical
merit and policy implications.
INTRODUCTION
Subtitle C of the Resource Conservation and Recovery Act
(RCRA) requires the U. S. EPA to develop, promulgate, and
implement regulations which control the generation,
transportation, and treatment, storage, and disposal (TSD) of
hazardous waste. Regulations promulgated under RCRA, at 40 CFR
Part 264, Subpart O, specify the following performance
standards which facilities treating hazardous waste by
incineration are required to meet:
1) 99.99% destruction and removal efficiency (ORE) for each
principal organic hazardous constituent (POHC) in its
permit for each waste feed (or 99.9999% for dioxin listed
wastes)
2) 99% removal efficiency of HC1 or 1.8 kg/hr of HC1
emissions, whichever is greater
3) particulate emissions less than 180 mg/dscm, corrected to
7% oxygen
The regulations also require that fugitive emissions be
controlled by keeping the combustion zone totally sealed,
maintaining negative draft, or an equivalent alternative'means
of control.
Facilities seeking a permit to incinerate hazardous waste
are required to demonstrate the unit's capability to meet the
performance standards during a trial burn. Since a permitted
incinerator's compliance with these performance standards
cannot be monitored over the long term, the conditions at which
the incinerator operated during the trial burn are included in
the incinerator's permit as conditions for continuing
operation. Compliance with these operating conditions is then
deemed to be compliance with the performance standards. An
incinerator must be operated with a system to automatically cut
off waste feed to the incinerator when operating conditions
deviate from limits established in the permit.
Although the regulations specify four operating parameters
which must be set as permit conditions based on the trial burn
(carbon monoxide level, waste feed rate, combustion
temperature, and an indicator of combustion gas velocity) it
is_left to the permit writer to determine how to translate the
5iai. rn data into Permit conditions, and to determine which
additional parameters should be set as permit conditions in
order to assure continued compliance. Because of the technical
complexity of setting permit conditions for hazardous waste
-259-
-------
incinerators, and the flexibility the regulations allow in
setting these conditions, there has been a lack of consistency
in the operational portions of incinerator permits issued
across the country. Further, an excessive number of permit
conditions may severely limit the incinerator's flexibility of
operation, while too few permit conditions may not provide
adequate assurance that the performance standards will continue
to be met.
APPROACH
The major goals in developing the guidance were to develop
a nationally consistent, sound technical approach to the
setting of operational conditions in incinerator permits which
would maintain proper performance while allowing a
reasonabledegree of operational flexibility. Technical
rationales were to be stated in the document so that it would
also serve as a training tool, and to enable the permit writer
to identify and address cases where specific portions of the
guidance may not apply. Various operating parameters thought
to have an effect on achievement of the incinerator performance
standards were considered for inclusion in the guidance. These
conditions were evaluated based on technical knowledge, and,
where necessary, consensus of engineering judgment, to develop
a set of operating parameters to be set in incinerator permits
which would meet with the above goals. "Back-up"'parameters
which would unnecessarily limit the permittee's flexibility to
operate the incinerator were avoided.
Management of the project was a cooperative effort by the
Office of Solid Waste and the Hazardous Waste Engineering
Research Laboratory (HWERL) of the Office of Research &
Development. The document was prepared by three contractors:
Acurex, general coordination and engineering aspects, MRI,
sampling and analysis aspects, and EER, engineering analysis.
Because of the desire to issue guidance which was both
technically sound, and practical from an implementation
standpoint, a significant amount of outside input was
incorporated.
DESCRIPTION
This paper reflects the content of the final draft
guidance. Minor changes may be made to the guidance prior to
publication, based on a final review. The guidance document
presents the key control parameters, shown in Table 1, which
should be monitored during the trial burn and for which limits
should be set in the incinerator permit. The parameters are
divided into three groups.
Group A parameters are continuously monitored parameters
interlocked to automatic waste feed cutoff. Most of these
-260-
-------
TABLE 1 CONTROL PARAMETERS
Group A
Continuously monitored parameters are interlocked with the
automatic waste feed cutoff. Interruption of waste feed is
automatic when specified limits are exceeded. The parameters
are applicable to all facilities.
1,
2
3,
4,
5,
Minumum temperature measured at each combustion chamber exit
Maximum CO emissions measured at the stack or other
appropriate location
Maximum flue gas flowrate or velocity measured at the stack
Design pressure maintained in primary combustion chamber
The following as applicable to the facility
o Minimum differential pressure across particulate venturi
scrubber
o Minimum water/liquor flowrate and pH to absorber and dry
scrubber
o Minimum KVA settings to ESP (wet/dry) and KV for ionized
wet scrubbers (IWS)
o Minimum and maximum pressure differential across baghouse
Group B
Parameters do not require continuous monitoring and are thus
not interlocked with waste feed cutoff systems. Operating
records are nevertheless required to ensure that trial burn
worst-case conditions are not exceeded.
7
8
9
Maximum total halides and ash feedrate to the incinerator
system
POHC incinerability limits
Maximum size of containerized to primary chamber
Minimum particulate scrubber blowdown rate
Group C
Limits on these parameters are set independently of trial burn
test conditions. Instead, limits are based on equipment
manufacturers' design and operating specifications and are thus
considered good operating practices. Selected parameters do
n°t require continuous monitoring and are not interlocked with
the waste feed cutoff.
10,
11.
12,
Maximum total heat input capacity for each chamber
Liquid injection chamber burner settings:
o Maximum viscosity of pump
o Maximum burner turndown
o Minimum atomization fluid pressure
Maximum APCE inlet gas temperature
-261-
-------
parameters are based on trial burn conditions. A minimal
amount of lag time may be incorporated into the limits for
these parameters by use of averaging times or delay times,
following guidelines in the document. This allows for normal
fluctuations in process parameters, and may allow time for
corrective action to be taken either by the automatic process
control system or manually. Group B parameters are set to
ensure that the "worst case" conditions demonstrated in the
trial burn are not,exceeded during continuing operation. These
parameters are not linked with automatic waste feed cutoff and
are not continuously monitored, but instead must be recorded in
the facility operating record. Group C parameters are set
independently of trial burn results, based on equipment
manufacturers' design and operating specifications. These
parameters are not continuously monitored or linked to
automatic waste feed cutoff.
Group A Parameters
Temperature is a key parameter of incinerator performance
due to its influence on reaction kinetics, and is a required
incinerator permit condition under RCRA regulations. The
minimum temperature limit is generally set from the lowest
temperature trial burn test at which compliance was
demonstrated. Combustion chamber temperatures are required by
the regulations to be tied to automatic waste feed cutoff. For
a two-chamber incinerator, minimum temperatures would be set
for each chamber. When minimum temperatures are not maintained
in the primary and secondary chambers, or in the secondary
chamber only, waste feed must be cutoff to both chambers.
However, if only the primary chamber falls below its minimum
temperature, waste may still be fed to the secondary chamber.
Carbon monoxide concentration in the stack gas is also a
parameter which the regulations specifically require. CO is
used as an indicator of the degree of mixing achieved in the
incinerator, and is related, by definition, to combustion
efficiency. Separate guidance on setting permit limits on CO
in order to minimize emissions of PIC's (products of incomplete
combustion) is being prepared by EPA.
The hazardous waste incinerator regulations require that
the permit specify limits for an indicator of combustion gas
velocity. Combustion gas velocity is directly related to the
gas residence time in the incinerator, which is known to be one
of the key parameters of combustion. Residence time becomes
more critical at lower combustion temperatures. For this
reason, the limit on maximum combustion gas velocity should be
based on the maximum trial burn value measured during the
lowest temperature test.
-262-
-------
A waste feed rate limitation is required by RCRA
regulations primarily to minimize the potential loss of
efficiency from overloading the combustion chambers. For low
heating value wastes, the limits are taken from the trial burn
test with the minimum temperature, since an increase in the
waste feed rate may cause a decrease in temperature. Maximum
waste feed rate for high or medium heating value wastes will be
set independently of temperature.
_ The requirement in the regulations to control fugitive
emissions is addressed by a permit requirement that the unit
design pressure is maintained. Incinerator chambers designedto
operate under negative draft (induced draft) would be required
by the permit to maintain negative draft. Forced draft or
positive pressure incinerators must be well sealed and a
maximum pressure would be set in the permit based on the unit
design pressure.
The guidance recommends that control parameters for air
pollution control equipment (APCE) be set to maintain the
particulate and acid scrubbing capability demonstrated during
the trial burn. For each type of APCE component, one key
parameter, was chosen to be tied to automatic waste feed
cutoff. For example, since the principal operating parameter
controlling ESP collection efficiency is the power utilization,
or KVA, the minimum KVA demonstrated during the trial burn at
the highest ash feed rate will be set as the permit limit.
Group B Parameters
One of the key principles behind conducting a trial burn
is that the incinerator should operate under the most severe
conditions it is expected to encounter for the duration of its
permitted operation. Group B parameters are included in the
guidance to assure that the incinerator will not operate at
more taxing conditions than those at which it demonstrated
compliance during the trial burn.
Parameters affecting APCE performance included in Group B
are total ash and chlorine loading to the incinerator. These
parameters affect the concentrations of particulate and HC1 at
the APCE inlet and the physical and chemical properties of the
gas. The ash and chlorine loadings are limited to the maximum
rate demonstrated in the trial burn. A minimum scrubber
blowdown rate will also be set based on the trial burn, since
suspended and dissolved solids in recycle water can be
re-entrained into the flue gas, contributing to particulate
emissions.
The Subpart O regulations require that POHCs (Principal
Organic Hazardous Constituents) be designated for each waste
-263-
-------
feed. The required DRE must then be demonstrated for^the POHCs
during the trial burn. Since the POHCs must be representative
of the waste feed, they are chosen based on factors such as
difficulty to incinerate and concentration in the waste feed.
The operator is then limited in the permit to burning only
waste containing hazardous constituents no more difficult to
incinerate than the POHCs for which compliance was demonstrated
during the trial burn. The heat of combustion of the hazardous
constituents has been used to rank the incinerability of
compounds on the premise that compounds'with a lower heat of
combustion are more difficult to burn. Field data indicate,
however, that other ranking systems may exhibit a better
correlation with incinerability. The guidance presents a draft
ranking of the incinerability of Appendix VIII compounds
prepared by the University of Dayton Research Institute based
on thermal stability at low oxygen (TSLoO2) conditions.
Additional laboratory and field testing of the TSLoO2 rankxng
are currently being conducted under a cooperative agreement
between EPA and UDRI.
s"
A limit on the maximum size of containerized waste fed to
the incinerator is also recommended in order to prevent oxygen
depletion from the sudden release of volatiles. The
containerized waste fed during the trial burn should be
representative, with respect to volatile content, of the waste
the facility will be burning under the permit.
Group C Parameters
Group C parameters were formulated based on the need to
ensure that incinerator operation adheres to good combustion
and APCE operating practices. To allow a reasonable degree of
flexibility and to avoid over-complication of the trial burn,
limits for these parameters are to be based on manufacturer's
design and operating specifications rather than the trial burn
settings.
To maintain proper automation of liquid waste and promote
efficient mixing, burner operational settings for liquid
injection and afterburner chambers will be limited to
manufacturer's specifications. These conditions include
maximum waste viscosity, minimum automation fluid pressure, and
maximum burner turndown. A minimum waste heating value will be
set in the permit for liquid injection chambers where 100% of
the heat input comes from the waste feed. Total heat input to
the incinerator will be limited to the incinerator design heat
input capacity.
The guidance recommends limiting APCE inlet gas
temperature due to its effect on the APCE performance as well
as to prevent equipment deterioration. For example, the
minimum inlet temperature to a baghouse would be based on the
-264-
-------
dew point of the incoming gases, while a maximum temperature
would be set based on manufacturer specifications for the type
of fabric used.
Other Permit Conditions
The guidance also includes additional conditions related
to waste feed cutoff. The permit should require that minimum
temperature be maintained in the secondary combustion chamber
after a waste feed cutoff until wastes remaining in the unit
are burned out. This would necessitate use of auxiliary fuel,
but must not conflict with the unit's flame safety management'
system. The guidance recommends a condition requiring
quarterly reporting of automatic waste feed cutoffs, reasons
for the cutoffs, and corrective actions taken.
Translating Trial Burn Results into Permit Conditions
The guidance presents a strategy for determining the limits
on operating parameters and converting them into permit
conditions. The goal in translating the trial burn results
into permit conditions is to assure the incinerator is
operating in a manner sufficiently similar to the successful
trial burn conditions to maintain compliance, but still allow
adequate operational flexibility. The approach commonly
employed is patterned around "mode-based" operation. The
permit contains a different set of operating conditions for
each waste combination the facility will burn. This approach
is best suited for a facility dedicated to treating a
well-defined set of uniform composition hazardous wastes.
The above approach, however, is not practical for
facilities such as commercial facilities, which burn a wide
variety of wastes. The guidance presents an approach to
developing a universal set of operating conditions which
defines the allowable range of operation for burning all of the
wastes in the facility permit. The difficulty in using the
trial burn data to develop a universal set of operating
conditions is that it is usually not possible to maintain
worst-case conditions for all permit parameters at a single
operating point. Thus, permit conditions must be developed
from data obtained at several operating points.
In the general approach set forth in the guidance, the
parameters are divided into three groups:
1.
2.
3.
Control parameters set from trial burn data
that are related to waste destruction
Control parameters set from trial burn data
that are related to APCE performance
Control parameters that are independent of
trial burn data
-265-
-------
Limits on parameters are set according to the above hierarchy.
The groupings of these parameters are shown in Table 2. Permit
limits must only be set from trial burn tests that show
compliance with the performance standards. Limits should be
set using these basic rules of thumb regarding "worst case"
conditions. The maximum combustion gas velocity should be set
from the trial burn test conducted at the minimum temperature.
The maximum feed rate of each low heating value waste stream to
each combustion chamber should be that demonstrated during the
minimum temperature test. The maximum feed rate of high
heating value wastes and the maximum combined feed rate should
be the maximums demonstrated at any point.
TABLE 2
Control Parameter Classifications
Waste Destruction-Related Parameters Set From Trial Burn Data
- minimum temperature
- maximum waste feed rate
- maximum combustion gas velocity
- maximum size of containerized waste
APCE Performance-Related Parameters Set From Trial Burn Data
- minimum differential pressure across venturi scrubber
- minimum water/liquor flow rate and pH to absorber
- minimum KVA for ESP and KV for ionized scrubber- minimum and
maximum baghouse pressure differential
- maximum chlorine and ash feed rate
- minimum particulate scrubber blowdown rate
Trial Burn-Independent Parameters
- maximum CO concentration
- maximum primary combustion chamber pressure
- maximum total heat input
- liquid injection chamber burner settings
- maximum APCE inlet gas temperature
Permit limits for APCE parameters relating to particulate
collection should be set from the trial burn test at the
maximum inorganic ash feed rate and the maximum flue gas flow
rate, since ash feed rate determines the load to the ACPE, and
increasing the flue gas flow rate may increase entrainment of
particulate matter. Minimum liquor flow rate to the absorber
and minimum pH to the absorber should be set from the trial
burn test at the maximum total halides feed rate and the
maximum flue gas flow rate.
-266-
-------
In some instances, it may not be possible to set the
conditions in the manner described above due to
interrelationships among parameters which prevent certain
conditions from being achieved at the same time. The guidance
presents an approach to estimate, through calculations, whether
the effect of setting the conditions based on less than
worst-case runs will be significant. For example, if the
permit limit for the maximum flue gas velocity is to be set
from a data point other than the minimum temperature test, then
the permit writer would do calculations to determine whether it
is likely that the flue gas flow rate at the minimum
temperature could be increased to the maximum flue gas flow
rate without causing ORE to decrease below 99.99%. This is
done by relating flue gas flow rate to residence time to DRE
assuming a first-order reaction.
The guidance emphasizes the importance of planning the
trial burn to obtain the desired permit conditions. The
applicant and permit writer should agree, prior to the trial
burn, on what permit conditions will result from the trial burn
as planned, assuming compliance is demonstrated. This will
allow the applicant to go back and make modifications to the
trial burn plan, if necessary, to obtain the desired operatina
conditions. ^
Trial Burn Reporting
The permit writer is often faced with reviewing a trial
burn report which is incomplete or which is not structured in a
way that is conducive to finding the information necessary to
evaluate compliance and set permit conditions. The permit
writer may need to go back to the applicant to request
clarification or additional data, which slows down the review
process. To assist both applicants and permit writers, the
guidance describes the information which should be included in
the trial burn report, and presents a trial burn report
format. Example reporting forms for the design, process, and
performance data required in a trial burn report have been
developed, and are presented in the document.
SUMMARY
The Guidance on Trial Burn Reporting and setting Permit
Conditions has been developed to assist permit writers in
translating trial burn results into site-specific operational
conditions in an incinerator permit. These parameters are
presented in the document, along with guidance on how to
develop permit operating conditions using the trial burn data.
The guidance will also assist applicants in planning their
trial burns to address the key operating parameters that must
-267-
-------
be measured and the necessity to test "worst-case" operations.
This will enable applicants to tailor their propose operating
conditions to the needs of their facility. One of the key
points made by the guidance is that the permit writer and
applicant should agree, prior to the trial burn, on what permit
conditions will result from the trial burn as planned. In this
way, it can be determined whether it is necessary to make
modifications to the plan in order to obtain the desired
operating conditions.
ACKNOWLEDGMENTS
This guidance document was prepared for the U.S.
Environmental Protection Agency's Office of Solid Waste (OSW)
and the Office of Research and Development (ORD) under the
overall direction of Ms. Sonya Stelmack of OSW Permits and
State Programs Division, with technical review by Dr. C.C. Lee
and Donald Oberacker of ORD Hazardous Waste Engineering
Research Laboratory. Ms. Robin Anderson also contributed to
the project.
The principal investigator was Mr. Carlo Castaldini,
assisted by Dr. Leo Weitzman and Mr. Andy Murphy, all of
Acurex. Technical contributors were Dr. Randall Seeker and Mr.
Wyman Clark of EERC, and Mr. Andrew Trenholm and Mr. Gary
Hinshaw of MRI.
Additional technical contributions and review were provided
by a panel of incineration experts from selected EPA Regional
Offices. This panel included Mr. Gary Gross of Region III, Ms.
Betty Willis of Region IV, Mr. Y.J. Kim of Region V, and Mr.
John Hart of Region IX.
EPA wishes to acknowledge and thank the American Society of
Mechanical Engineers for providing review comments on drafts of
the guidance document.
REFERENCES
A.D. Little, Inc. Sampling and Analysis Methods for Hazardous
Waste Combustion. First Edition. Prepared for EPA under EPA
Contract No. 68-02-3211 (124). PB84-155845, December 1983.
Mitre Corporation. Guidance Manual for Hazardous Waste
Incinerator Permits. EPA-SW-966, PB84-10000577, July 1983.
Monsanto Research Corporation. Engineering Handbook for
Hazardous Waste Incineration. EPA-SW-889, PB81-248163,
September 1981.
Clark, W.D., W. R. Seeker, and C.C. Lee. Engineering Analysis
of Hazardous Waste Incineration, Energy and Mass Balance. In
-268-
-------
Proceedings of the Twelfth Annual Research Symposium on Land
Disposal, Remedial Action, Incineration and Treatment of
Hazardous Waste. EPA/600/9-86/022, 1986. pp. 360-368.
Bellinger, B., et.al. Examination of Fundamental
Incinerability Indices for Hazardous Waste Destruction.
Proceedings on Incineration and Treatment of Hazardous Waste
/EPA-600/9-85/028, September 1985.
Hesketh, H.E., Air Pollution Control. Ann Arbor, Michigan, Ann
Harbor Science Publishers, 1980.
Brunner, Calvin R. Incineration Systems.
Reinhold Company, New York, 1984.
Van Nostrand
-269-
-------
ROTARY KILN INCINERATION:
THE EFFECT OF OXYGEN ENRICHMENT ON FORMATION OF TRANSIENT PUFFS
DURING BATCH INTRODUCTION OF HAZARDOUS WASTES
William P. Linak, Joseph A. McSorley
Combustion Research Branch, MD-65
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Jost O.L. Wendt
Department of Chemical Engineering
University of Arizona
Tucson, AZ 85721
Oames E. Dunn
Department of Mathematical Sciences
University of Arkansas
Fayetteville, AR 72701
ABSTRACT
Experiments on a 73 kW (250,000 Btu/hr) rotary kiln incinerator simulator
equipped with a prototype oxygen enrichment burner were performed to determine
the effect of oxygen enrichment on the magnitude and intensity of transient
puffs emitted during batch introduction of contained liquid surrogate waste
(toluene) on corncob sorbent. Results show that the effect of oxygen
enrichment at constant temperature is to decrease the transient puff size but
that increases in kiln temperature associated with oxygen enrichment can
offset this and cause net increases in the transient emissions. This effect
is likely due to increases in the rate of waste volatilization. These results
were especially true for particulate matter and are consistent with previous
results without oxygen enrichment.
-270-
-------
INTRODUCTION
Oxygen enrichment has long been applied to many industrial combustion
processes in order to increase product yields and thermal efficiency. The
use of oxygen in hazardous waste incineration to augment or enrich oxidant
levels in the combustion zones has the potential to improve the destruction
and removal efficiencies (DREs) of the principal organic hazardous
constituents (POHCs), minimize the formation of products of incomplete
combustion (PICs), and allow for better control of the .combustion process.
The use of oxygen enrichment in rotary kiln incinerators is of interest
as a potential method of minimizing transient failure conditions caused by
the batch introduction of waste material. Rotary kiln incinerators are
unique in that they are designed to allow a portion of their waste load to be
charged in batch rather than continuous mode. In this batch mode of operation,
solid wastes and "containerized" liquid wastes are introduced through entrance
chutes, typically cocurrent with the gas flow. Kiln angle and rotation speed
continuously expose fresh surface for oxidation, determine the residence
times of non-combustible materials, and provide for continuous ash removal.
Upon entry into the incinerator, the liquid waste container, typically
cardboard, plastic, or steel drums, ruptures or burns, exposing the contents
to the hot kiln environment. The liquid then rapidly vaporizes and reacts
with the excess oxygen present in the combustion gases from the continuous
primary flame. A possible kiln failure condition arises when the evolution
of waste gases from the ruptured container is more rapid than the rate at
which the stoichiometric amount of oxygen can be supplied from the primary
flame. This transient overcharging condition, unique to batch mode systems,
can lead to a "puff" of incompletely destroyed organic material which must
subsequently be destroyed by the system afterburner. Under circumstances
when the afterburner cannot destroy the quantity of organic material produced,
this puff leads to a system failure.
This paper discusses results of part of an in-house program at the U.S.
Environmental Protection Agency (EPA) directed toward gaining engineering
insight into potential problem areas related to rotary kiln incineration.
General objectives are to delineate fundamental combustion phenomena that
important in the generation of transient puffs from batch fed wastes. It
not within the scope of this work to do exhaustive POHC, PIC, or ORE analyses
The specific objectives of the work presented here are to determine the
effects of oxygen enrichment through the primary flame on the formation of
transient puffs, caused by the batch introduction of a contained surrogate
liquid waste (toluene) on sorbent into a rotary kiln incinerator simulator.
Further, the tests and analysis techniques attempt to separate the individual
effects of dilution, temperature, oxygen flow, and oxygen partial pressure.
Results from tests utilizing this simulator to examine transient puffs from
prototype solid plastic wastes and contained surrogate liquid wastes without
oxygen enrichment have been presented previously.1>2
This paper is divided into four sections. Following the Introduction,
the Experimental Approach is discussed. Next, Results and-Discussion of the
experimental study are presented. Finally, the findings from this study are
summarized in the Conclusions.
are
is
-271-
-------
EXPERIMENTAL APPROACH
The EPA rotary kiln incinerator simulator has been described in detail
previously,!>2 and therefore, is not presented again here. It has been
established that the simulator, although sized at only 73 kW (250,000 Btu/hr),
exhibits the salient features 'of full-scale units with thermal ratings 20 to
40 times larger. The simulator matches the volumetric heat release, gas-phase
residence time, and temperature profile of many full-scale units, and yet is
flexible enough to allow parametric testing. A schematic drawing of the
simulator is presented in Figure 1.
The water-cooled prototype burner is designed to produce stable flames
utilizing oxidant streams ranging from 22 to 100 percent oxygen. The burner
provides for separate introduction of natural gas, air, and oxygen with turn-
down ratios of at least 1 to 10. The burner is easily ignited and capable of
rapid response to changes in the fuel, air, or oxygen flows. Liquid oxygen
was provided in cryogenic Dewar tanks, with each tank containing approximately
127 m3 (4500 scf) of gaseous oxygen.
Contained batch charges to the rotary kiln are introduced through a
sliding gate/ram assembly located downstream of the continuous natural-gas-
fired oxygen-enriched main burner as shown in Figure 1. Quantification of
the transient puff produced requires the real-time measurement of system
response variables. These dependent variables include peak responses of 03,
CO, CO?, NOX, and total volatile hydrocarbon concentrations (THC) measured
by continuous flame ionization detection (FID) and reported in equivalent
parts per million methane. Other dependent variables include the time-
integrated responses of the THC analyzer and particulate filter weight. At
constant sampling conditions, these variables are proportional to the total
mass emitted in the transient puff. Weighed desiccated quartz filters and
the heated particulate filter system upstream of the THC analyzer were used to
collect particulate samples. Volatile hydrocarbons are defined as those in
the gas-phase at 420 K (300 °F), the temperature of the heated sample line.
Charges to the kiln were kept constant for all tests presented in this
paper. Only the natural gas, air, and oxygen flows to the prototype burner
were varied by test design to cause changes in the oxygen flow, partial
pressure, and stoichiometric ratio. Charges to the kiln consisted of 100 g
(0.22 Ib) of toluene on 135 g (0.30 Ib) of ground corncob sorbent sealed in a
45 g (0.10 Ib) cylindrical uncoated 1 qt (0.95 1) cardboard container. Similar
charges were used in previous work by this group.2 Toluene was chosen because
it yields large chlorine-free puffs to which the FID is very sensitive.
Toluene also produces large particulate loadings. A charge mass of 100 g
(0.22 Ib) was chosen because it provides large puffs which did not greatly
exceed the FID linear response region of zero to 10,000 ppm (reported as
methane) for the conditions tested.
As with the previous studies, a portion of the experimental design was
based on response surface experimentation.3 This methodology allows, on the
basis of one experiment involving a minimal number of trials, determination
of an empirical relationship between the response and the controlled variables
in the experimental region. For the first set of experiments, the flows of
-272-
-------
TO BLOWER AND STACK
I
BACKFIRE EXHAUST
=THERMOCOUPLE
SAMPLE PORT 4
SAMPLE PORT 3 v
'
SECTION 5
(CONTROL TEMP.
TOWER 2)
SECTION 4
(CONTROL TEMP.
TOWER 1)
SECTION 3
(BACKFIRE BURNER)
BACKFIRE BURNER
SAMPLE PORT 2
SECTION 1 (KILN)
ROTARY LEAF
MAIN BURNER SPRING SEAL
(MOVEABLE) \ _ \
I I . n t
gg-3. — AFTERBURNER
SLIDING GATE
•fftm — BACKFIRE EXHAUST
,:::.-::
ROLLERS
RAMROD
CHARGING BASKET
ROTARY LEAF
SPRING SEAL
SECTION 2
(TRANSITION/AFTERBURNER)
SCALE:
m 0
0.5
ft 0
Figure 1. EPA incinerator simulator, cutaway view.
-273-
-------
oxyqen and air were chosen as the controlled or independent variables. The
natural gas auxiliary fuel feed rate (kiln load) was held constant. These
variables are of intrinsic interest to an incinerator operator because they
are directly controlled. By defining these flow rates to the burner, several
other system parameters are defined. These calculated variables include;
stoichiometric ratio, primary flame oxygen enrichment, total flow
(approximately inversely proportional to residence time), primary flame
oxygen and nitrogen flows, adiabatic flame temperature, post primary flame
oxygen flow, and post primary flame oxygen partial pressure. Throughout this
first set of experiments, the natural gas flow was held constant at 5.bb
m3/hr (200 scfh), and the oxygen and air flows to the burner were varied by
experimental design to examine six combinations (plus one duplicate) oT
oxygen and air flows. Figure 2 presents the combinations of air (57.5 to
105.0 m3/hr, 2030 to 3707 scfh) and oxygen (1.13 to 9.74 m3/hr, 40 to 344
scfh) flows that form the second order rotatable pentagonal design (data
points denoted by + ). Figure 2 also shows that these air and oxygen flows
correspond to a range of calculated stoichiometric ratios from 1.3 to 2.4
(Figure 2a), post flame oxygen flows from 3.29 to 16.37 m3/hr (116 to 578
scfh) (Figure 2b), and post primary flame oxygen partial pressures from 5.0
to 16.0 percent (Figure 2c). A summary of the final trial matrix including
the controlled and derived system variables is presented in Table I.
The seven (six plus one duplicate) experiments, described in Figure 2
and Table I, were conducted in the following manner. The desired natural
aas air, and oxygen flows were set, and the kiln was allowed to thermally
equilibrate to the new condition for at least 5 hours. Charges were prepared,
sealed, and introduced to the kiln at approximately 10 minute intervals.
Between 10 and 12 replicate charges were recorded and averaged for subsequent
statistical analysis. Kiln rotation speed, kiln pressure, and burner Position
were kept constant at 0.5 rpm, -37.4 Pa (-0.15 in. H20), and -0.34 m (-13.5
in.) inside the kiln front wall, respectively. Note that the effect of the
Mln alone is being examined. There was neither afterburning nor external
heating of the combustion products. We are trying to determine the behavior
of the kiln and, subsequently, the relative role of the afterburner to ensure
adequate destruction.
Based on the result's of the first set of response surface experiments, a
second and third series of two dimensional tests were performed in an attempt
to separate the individual effects of stoichiometric ratio, post flame oxygen
flow, and post flame oxygen partial pressure. The design of^ response
surface experiments was limited. It is evident from Figure 2 that, for the
response surface experiments, the three oxygen parameters (stoichiometric
ratio, post flame oxygen flow, and post flame oxygen partial pressure) are
hiqhlv correlated with each other and vary in the same manner. In fact, due
to the constant auxiliary fuel load used for these experiments, specification
of either stoichiometric ratio or post flame oxygen flow defines the other.
To separate these effects, two sets of additional experiments were designed,
varying the auxiliary fuel load. These trial matrices are presented in
Tables II and III. The trial matrix in Table II, was designed to vary the
post flame oxygen flow and partial pressure independently while holding
stoichiometric ratio constant. Jhe trial matrix in Table III, varies
stoichiometric ratio and post flame oxygen partial pressure while holding
-274-
-------
(a) STOICHIOMETRIC RATIO
2.5
2000 2200 2400 2600 2800 3000 3200 3400 3600 3800
AIR, SCFH
(b) POST FLAME OXYGEN FLOW,SCFH
500
2000 2200 2400 2600 2800 3000 3200 3400 3600 ' 3800
AIR. SCFH
POST FLAME OXYGEN PARTIAL PRESSURES
2000 2200 2400 2600 2800 3000 3200 3400 3600 3800
AIR, SCFH
Figure 2. Contours^of constant (a) stoichiometric ratio, (b) post flame
post flame oxygen flow constant
The individual data sets were collected in
post fi
-275-
-------
TABLE I. TRIAL MATRIX FOR DATA SET 1
(RESPONSE SURFACE EXPERIMENTS, CONSTANT GAS FLOW).
Test
1
2
3
4
5
6
7
Gas
flow,
scfh
200
200
200
200
200
200
200
Air Oxygen Total
flow, flow, flow,
scfh scfh
3707
3066
2030
2030
3066
2780
2780
192
40
98
286
344
192
192
scfh
4099
3306
2328
2516
3610
3172
3172
Total
Oxygen oxygen Nitrogen
enrichment, flow, flow,
percent scfh
24
21
24
30
28
26
26
.8
.9
.5
.7
.9
.0
.0
Post flame
Post flame
Test
1
2
c*
3
4
"
5
*J
5
\J
7
SR*
2.4
1.7
1.3
1.7
J* • /
2.4
1.9
1.9
oxygen fl
ow,
scfh (F02)
560
274
116
304
578
367
367
oxygen partial
pressure
percent
13.7
8.3
5.0
12.1
16.0
11.6
11.6
3
(P02)
96/
681
522
710
985
773
773
Adiabatic
f 1 ame
temp,
K
1454
1687
2125
2017
1584
1732
1732
scrn
2932
2425
1606
1606
2425
2199
2199
Measured
kiln exit
temp,
K
1125
1172
1201
1200
1151
1181
1159
*stoichiometric ratio
RESULTS AND DISCUSSION
Results are presented in two parts. First, the measured puff is
considered without correction for any dilution effects. This represents what
must be handled by the afterburner and is of interest to operators and
designers of rotary kiln systems. Next, two sets of data are presented that
have been corrected for dilution effects. These represent the rates of
volatile release from the waste.
Set 1; Dilution Elf^ctsJN^t_^nsidered
Full quadratic models with respect to air and oxygen flows were fitted
to the data from the first set of experiments (Table I) using weighted least
squares, where the weights were equal to the number of replicates. These
models, obtained using the preprogrammed procedures in SAS (Statistical
Analysis System) should not be considered unique since other models might
attain fits essentially identical to those reported here. Once determined,
these models can be described pictorially through response surface plots.
The response surface plots are unique to the data they describe. Figure 3
presents these contour plots for the dependent variables: volatile hydrocarbon
-276-
-------
TABL£ II- TRIAL MATRIX FOR DATA SET 2
(TWO DIMENSIONAL EXPERIMENTS, CONSTANT STOICHIOMETRIC RATIO).
Gas
flow.
Test scfh
8 150
9 150
10 200
11 200
12(20) 250
13(21) 250
14 300
15 300
Air
flow,
scfh
2732
1249
3707
2224
4679
3196
5656
4173
Oxygen
flow,
scfh
40
350
40
350
40
350
40
350
Total
flow,
scfh
2922
1749
3947
2774
4969
3796
5996
4823
Oxygen
enrichment,
percent
22.0
38.2
21.7
31.7
21.6
28.7
21.5
27.0
Total
oxygen
flow.
scfh"
611
611
815
815
1018
1018
1222
1222
Nitrogen
f 1 ow ,
scfh
2161
988
293?
. C. J \J C—
1759
. 3701
2528
4474
.3301
Test SR*
8 2.0
9 2.0
10 2.0
11 2.0
12(20) 2.0
13(21) 2.0
14 2.0
15 2.0
Post flame
oxygen flow,
scfh (F02)
306
306
408
408
510
510
613
613
Post flame
oxygen partial
pressure,
percent (P02)
10.5
17.5
10.3
14.7
10.3
13.4
10.2
12.7
Adi a bat ic
flame
temp ,
K
1508
2107 .
1494
1891
1487
1782
1482
1716
Measured
kiln exit
temp,
K
1000
1047
1112
,1190
1213
1286
1247
1320
*stoichiometric ratio
hahtmn KaK V°lat11e ^rocarbon peak. concentration
(peak height) (Figure 3b), carbonaceous particulate filter weight /filter
weight) (Figure 3c), and CO peak concentration (peak CO) (Figure 3d versus
oxygen and air flows Full quadratic models were able to explain totaT
response variances (p2) Of 0.9485:, 0.8992, 0.9993, and 0.9610 for peak area
peak height, filter weight, and peak CO, respectively. The variance isT
sSr/H the.disPer|io" of the response, defined as the square of the
standard deviation. Evident from Figure 3 is that,, for all the response
variables, transient puffs are reduced at higher air and oxygen levels Peak
^ ^^ resP°nses at hi9h ^ and oxygen flows*(upper r ght
an aPP™"1**^ nine and five times less, respectively, than at
f fnnr ( IT left corners)' Filter weW response decreases by,a factor
to notP hoLv "if/^on, and peak CO by, a factor of six. It is important
can £ Pvni^ H'K 9t a P°?,lon (but not a11) of these response reductions
c^L ?hexp1ained ^ simple dilution. Total flow changes by a factor of 1.8
over this experimental region, but due to the fact that we are trying to
tn r^nnrnZe f. ™^ C0nd1tions to which an afterburner system must be able
to respond, we have not corrected these data for dilution.
-277-
-------
TABLE III. TRIAL MATRIX FOR DATA SET 3
(TWO DIMENSIONAL EXPERIMENTS, CONSTANT POST FLAME OXYGEN FLOW).
Gas
flow,
Test scfh
16 150
17 150
18 200
19 200
20(12) 250
21(13) 250
22 300
23 300
Air
flow,
scfh
3708
2225
4194
2711
4679
3196
5165
3682
Oxygen
f 1 ow ,
scfh
40
350
40
350
40
350
40
350
Total
flow.
scfh
3898
2725
4434
3261
4969
3796
5505
4332
Oxygen
enrichment,
percent
21.7
31.7
21.6
29.9
21.6
28.7
21.5
27.8
Total
oxygen
flow,
scfh
815
815
917
917
1018
1018
1119
1119
Nitrogen
flow,
scfh
2933
1760
3317
2155
3701
2528
4086
2912
Test SR*
16 2.7
17 2.7
18 2.2
19 2.2
20(12) 2.0
21(13) 2.0
22 1.8
23 1.8
Post flame
oxygen flow,
scfh (FOa)
510
510
510
510
510
510
510
510
Post flame
oxygen partial
pressure,
percent (P0£)
13.1
18.7
11.5
15.6
10.3
13.4
9.3
11.8
Adiabatic
flame
temp,
K
1239
1575
1380
1693
1487
1782
1571
1844
; Measured
kiln exit
temp,
K
931
1011
1071
1151
1213
1286
1255
1340
*stoichiometric ratio
The response curves in Figure 3 can be compared with the contours of
constant stoichiometric ratio, post flame oxygen flow, and post flame oxygen
partial pressure, presented in Figure 2. It is not surprising that puff
intensities should be reduced at higher oxidant levels and higher total
flows! In addition to the effect of dilution, providing higher oxygen flows
and higher oxygen partial pressures to the vaporizing waste cause fuel rich
regionsto be more difficult to achieve. However, without oxygen enrichment,
it is often difficult in practice to operate at high stoichiometnc ratios
while maintaining kiln temperatures and flame stability due ^ the quenching
effect of nitrogen. Also evident from Figure 3 is the fact that all three
Ssponses are more sensitive to changes in oxygen flow than to changes in air
flow. Small changes in oxygen flow cause large changes in stoictnometric
rat"o, post flame oxygen flow, and post flame oxygen partial pressure, without
significantly affecting the adiabatic flame temperature or gas-phase residence
time. Conversely, large changes in air flow are necessary to effect changes
in stoichiometric ratio or post flame oxygen flow. Without oxygen enrichment,
-278-
-------
(a) PEAK AREA, PPM-SEC
2900
AIR, SCFH
3350
AREA
15000
90000
30000 45000 60000
105000 120000 135000
(b) PEAK HEIGHT, PPM
3800
— 75000 HT 2000
6000
AIR. SCFH
3000 4000 5000
7000 8000 9000
(c) FILTER WEIGHT, MG
2900
AIR, SCFH
3350
3800
— 8
— 13
(d) PEAK CO, PERCENT
350
273
\
\
\
high
low
2450 2900 3350
AIR, SCFH
— 0.4 0.7 1.1 lit
— 1.8 2.1 2.4
.Figure 3.
Data set 1 response surface contour plots for (a) volatile
hydrocarbon peak area, (b) volatile hydrocarbon peak height, (c)
particulate filter weight, and (d) peak CO. Data are not corrected
for dilution.
oxygen partial pressure is limited by atmospheric conditions Additionally
p
oxygen flow, and post flame oxygen partial pressure n the response variables ,
-279-
-------
It has been shown previously^ that, with respect to transit puffs from
batch waste introduction, increasing kiln temperature increases puff intensity
and that this is likely due to increased vaporization rates. This result is
in contrast to typical kiln operating practice under steady-state operation,
where increased temperatures are desirable because they increase the kinetic
rates of oxidation reactions. Thus, for batch waste, while oxygen enrichment
may increase the flame stability, flame stoichiometry, post flame oxygen
flow, and post flame oxygen partial pressure increased temperatures may
drive the waste into the gas-phase more rapidly. We must try to quantify the
relative magnitudes of each of these mechanisms. Figure 4 presents the
calculated adiabatic flame temperatures and total flow rates through the
rotary kiln simulator over the range of air and oxygen flow rates investigated
by the first set of experiments (Table I). System residence time is roughly
proportional to the inverse of the total flow. Comparison of Figure 4 with
the response curves in Figure 3 shows that the regions of lowest pea; area
peak height, filter weight, and peak CO responses correspond to the regions
bf low adiabatic flame temperature and high total flow (low residence t me).
The effect of high total flow can be explained by simple dilution. Additionally,
the effect of increased temperature with respect to transient puffs is
counterintuitive to normal steady-state incineration practice but consistent
with increased vaporization rates.
(a) ADIABATIC FLAME TEMPERATURE, K
(b) TOTAL FLOW, SCFH
290
0240
X
Y
G
E 190
H
S
9 140
H
90
I 4000
\3400
\2800
3800
2000 2200 2*00 2600
2800 3000
MR, SCFH
3200 3400 3600 3800
Figure 4.
Contours of constant (a) adiabatic flame temperature and (b) total
flow over the region examined by data set 1 (response surface
experiments). Trial matrix points denoted by +.
To determine the relative importance of each of these controlled, measured,
and derived variables on the response variables, single term linear models
were constructed from the data of the first set of experiments (Table I).
Table IV presents the response variance explained for each independent variable
examined individually. The results show that the three derived variables of
most significance in explaining the variance in the response variables are
post flame oxygen flow, stoicMometrlc ratio, and post flame oxygen partial
pressure. As previously mentioned, the experimental design was such that all
-280-
-------
TABLE IV. FRACTION OF VARIANCE (p2)
BY SINGLE TERM LINEAR
Variable
Peak area Peak height
ppm-sec CH4 ppm CH4
OF DATA SET 1 EXPLAINED
MODELS.
FiIter weight Peak CO
mg ppm
Post flame
oxygen flow,
scfh (F02)
SR*
Post flame
oxygen partial
pressure,
percent (P02)
Total oxygen
flow, scfh
Residence
time, 1/scfh
Adiabatic
flame
temperature, K
0.8494
0.8510
0.9359
0.8389
0.8395
0.8632
Measured
kiln exit 0.3293 0.4614
temperature, K
Air flow,
scfh 0.2382 0.4513
*stoichiometric ratio
0.9460
0.9467
0.9374
0.5935
0.5114
0.8618
0.8636
0.7887
0.7633
0.5199
0.5148
0.8170
0.6204
0.6267
0.5519
0.6903
0.6938
0.5078
0.7284
0.7317
Total flow,
scfh
Oxygen
enrichment,
percent
0.4380
0.4759
,. 0.6089
0.5917
0.6596
0.1894
0.6920
.0.2094
0.5258
0.5423
three of these variables are highly correlated and indicate the oxygen available
to the vaporizing waste. Of lesser importance are the variables: total
oxygen flow, residence time, adiabatic flame temperature, total flow, primary
flame oxygen enrichment, measured kiln temperature, and total air flow.
These results indicate that, for the experimental region examined here,
transient puffs can be reduced by providing increased oxygen to the waste,
and that even though increasing the oxygen flow may change the temperature
and residence time, the effects of increased available oxygen dominate. It
-281-
-------
is important to note that these data indicate only a secondary effect due to
temperature, even though previous results1*2 indicated that temperature has a
major effect on the intensity of transient puffs from batch incineration.
Calculated adiabatic flame temperatures for this experimental set vary from
1450 to 2000 K (2150 to 3140 °F), yielding an experimental range of 550 K
(990 °F). In actuality, however, the measured kiln exit temperature varies
only from 1125 to 1200 K (1565 to 1700 °F); therefore, this actual experimental
range of 75 K (135 °F) may be insufficient to establish any temperature
effects.
Figures 5 and 6 present the single term linear models for post flame
oxygen flow and post flame oxygen partial pressure, respectively, versus peak
area (Figures 5a and 6a), peak height (Figures 5b and 6b), filter weight
(Figures 5c and 6c), and peak CO (Figures 5d and 6d). Again, due to the
experimental design of the first data set, similar plots for stoichiometric
ratio are identical to those for post flame oxygen flow. Figures 5 and 6
include the averaged data for each of the seven experimental conditions
(denoted by +), as well as the 95 percent confidence regions around each
linear model. All the data indicate decreasing response with increasing post
flame oxygen flow (,and stoichiometric ratio) and post flame oxygen partial
pressure. Attempts were made to add additional linear and quadratic terms
to the models; however, these additions did not significantly increase the
variance explained. Again, note that a portion of these responses can be
explained by simple dilution.
A possible disadvantage to oxygen enrichment for use in hazardous waste
incineration is the potential to produce high concentrations of nitrogen
oxides (NOX) due to the increased temperatures and increased available oxygen.
Effects of the latter may greatly exacerbate this problem if high nitrogen
containing wastes are incinerated, although even if the waste contains little
or no nitrogen, thermal fixation of ambient nitrogen may produce unacceptable
quantities of NOX. Figure 7 presents contours of baseline nitric oxide (NO)
concentration over the region examined by the first experimental set (Table
I). The data are presented as measured (Figure 7a) and corrected to 7 percent
oxygen (Figure 7b) in the exhaust gas. Concentrations as high as 1500 ppm
(at 7 percent oxygen) were determined for high oxygen operating conditions
using natural gas auxiliary fuel only.
Sets 2 and 3: Corrected for Dilution Effects
As stated previously, the second and third data sets (Tables II and III)
were intended to separate the individual effects of stoichiometric ratio,
post flame oxygen flow, and post flame oxygen partial pressure. By varying
the natural gas load to the kiln, experiments were designed in which these
three parameters vary independently. The purpose was to determine the relative
importance of the molar flow rate of oxygen to the waste charge (proportional
to post flame oxygen flow) and the volumetric oxygen concentration available
to the waste charge (equal to the post flame oxygen partial pressure).
Two-dimensional plots of post flame oxygen flow and post flame oxygen partial
pressure versus the response variables were prepared at constant stoichiometric
ratio (SR=2.0). Additionally, similar plots of stoichiometric ratio and post
flame oxygen partial pressure versus the response variables were prepared at
-282-
-------
(a) PEAK AREA PPM-SEC: R2, =0.849
(b) PEAK HEIGHT, PPM: R2 =0.839
200 300 400 500
0-
100
200 300 400
F02. SCFH
14
13
12
F"
i»
I '•
R ..
(c) FILTER WEIGHT, MG: R2 =0.946
300 400
F02, SCFH
(d) PEAK CO, PERCENT: R2 =0.862
p
£
A
K 2
100
JOO
400
F02, SCFH
500
Figure 5. Single term linear models describing response of fa) volatile
hydrocarbon peak area, (b) volatile hydrocarbon peak he ght (c)
particulate filter weight, and (d) peak CO versus bost flamp ( '
oxygen flow (F02). Averaged data Yrom data sell (rlspose surface
S^ri"?^ d!^e^b^+- So1;d-""«. indicate best f?nLr fit? 6
t confidence region.
constant post flame oxygen flow (F02=14.4 m3/hr, 510 scfh). These data were
mi/hr^S^fS1; d1^tion by normalization to a standard total flow of 113
m-Vhr (4000 scfh). These plots showed the data to be scattered, and no
consistent trend existed between all of the response variables. A likely
of^hanairin0^^856.1^0"51^6^ results 1s that' in addition to the effects
of changes in the post flame oxygen flow and partial pressure, confounding
influence.s due to large changes in kiln temperature are included. By changing
the kiln load, necessary to effect changes in post flame oxygen flow and
partial pressure, measured kiln exit temperatures vary over a wide range (320
-283-
-------
(a)
PEAK AREA PPM-SEC: R2 =0.936
(b) PEAK HEIGHT, PPM: R2 =0.863
MONO <
150000
140000
ijoooo
'. UOOM
70400
80000
50000
43000
MOM
20000
10000
7 8 9 10 It 12
P02, *
(c)
FILTER WEIGHT, MG: R2 =0.937
is
t«
13
. «
i 11
I"
I?
10
P02. «
11 12 13 14 15 16
(d)
PEAK CO, PERCENT: R2 =0.789
A
K 2
9 10 II
P02. »
12 13
15 16
Figure 6,
Single term linear models describing response of (a) volatile
hydrocarbon peak area, (b) volatile hydrocarbon peak height, (c)
participate filter weight, and (d) peak CO versus post flame
oxygen partial pressure (POe). Averaged data from data set 1
(response surface experiments) denoted by +. Solid lines indicate
best linear fit; dotted lines indicate 95 percent confidence
region. Data are not corrected for dilution.
K, 576 °F). It has been shown previously!»2 that the effect of increased
kiln temperature is increased transient emissions and that this is likely due
to increased volatility. Data have also shown that for toluene, a known soot
precursor, soot formation is disproportionately promoted over volatile
hydrocarbon emissions with increasing temperature.2 Due to the confounding
influences of temperature, no conclusions can be drawn that separate the
individual effects of post flame oxygen flow and partial pressure. However,
overall, the effect of oxygen enrichment via the primary combustion chamber
to minimize the emissions from transient puffs does not seem to be dramatic
-284-
-------
(a) BASELINE NO, PPM
(b) BASELINE NO, PPM AT 7% OXYGEN
273
o
X
y
G
E 195'
H
s
c
F
H
lit
40-
20
TAD
\ \ X> \
1 \ \ \
1 * » \
/high/ \ \ \
/ ' 1 *' \.
/ ' \ \
\ \
/ I \
•^ S '
^ ,'
,-'"''
.———*"*" -
low
•
0 2450 2900 3350 38
AIR, SCFH
100 200 300 400 500
350
27J
0
X
Y
E 195
N
F
H
118
40
0 20
' / / \ '' \ \
/ \ \ ' \
high / l \ \ \
// ( \
/ \
^ S ' *v'
"*" ^-^ /' \
--""*"" ___„--""' \
-
low
' " ' — — - •
00 2450 2900 3350 380
AIR, SCFH
IAO "2 330 530 730
Figure 7. Contours of constant baseline NO flue gas concentration (a) as
measured and (b) corrected to seven percent oxygen over the
examined by data set 1 (response surface experiments).
region
n 6J£er1!men*aJ re9i°n examined here. Indeed, it can be detrimental
when the effects of temperature are included.
.To illustrate that the effect of increasing kiln temperature can be
each ofthl ^rnnH051^"1!,5510^ Fl"9Ure 8 Presents ™ter weight data from
each of the second and third experiment sets (Tables II and III) versus
measured kiln exit temperature. Data from experiments taken at constant
/hY^n (S^.?MFi9^ 8a) and constant post flame oxygen flow
o-. m-Vhr, 510 scfh) (FTgure 8b) are presented. These data, relating
? (low/enr;chlTieflt (°Pe" symbols) and high oxygen flow/enrichment
mHn? ?Hi have.b??" corrected for dilution by normalization to a
standard total flow of 113 m3/hr (4000 scfh). Data taken at similar kiln
loads are connected. The data show fair correlation between filter weight
and measured kiln exit temperature, and indicate a trend of increasing
particulate emissions with increasing temperature over the experimental range
examined here. The data may indicate some slight reduction in particulate
emissions due to oxygen enrichment at similar kiln loads, as shown by the
IncrSLI1 ?!!%!!!! ^?Ur? Howev^, if the oxygen enrichment causes a dramatic
increase in the kiln temperature, particulate emissions are actually increased
This is also illustrated in Figure 8 by comparison of all open or solid
-285-
-------
15
SR-2.0
10
i
u. 5
I
i-O-H
(a)
2
Ul
oc
K
(
10
cc
111
FO2 - 510 scfh
*-&*"
0\—
800
_L
_L
(b)
900
1000 1100 1200
KILN EXIT TEMPERATURE, K
1300
1400
1500
Figure 8. Results from data sets 2 and 3 (two dimensional experiments)
" showing the relationship between filter weight and measured kiln
exit temperature for (a) constant stoichiometric ratio (SR) and
(b) constant post flame oxygen flow (FO^). Data are corrected to
4000 scfh total flow. Open symbols indicate tests at low oxygen
flow/enrichment (1.13 m3/hr, 40 scfh). Solid symbols indicate
tests at high oxygen flow/enrichment (9.91 m3/hr, 350 scfh).
Tests at similar kiln load (natural gas flow) are connected.
-286-
-------
CONCLUSIONS
The results of experiments involving the use of oxygen enrichment via
the primary flame (primary combustion chamber) of a rotary kiln incinerator
simulator feeding batch charges of contained liquid surrogate waste (toluene)
on corncob sorbent are consistent with results previously presented by the
authors.J-.^ A summary of the results from this study includes:
1. The phenomena associated with transient puff formation from the batch
introduction of contained, waste are fundamentally different from those
associated with steady-state incineration. The realization of these
differences is important in order to minimize organic emissions.
2. Although increased temperatures associated with oxygen enrichment may
provide potential benefits for steady-state incineration, these increased
temperatures are detrimental in controlling transient puffs from contained
wastes. This was especially true for particulate emissions.
3. It is difficult to draw conclusions regarding the individual effects of
stoichiometric ratio, or post flame oxygen flow, or post flame oxygen partial
pressure due to the confounding effects of temperature.
4. When the effects of temperature are removed, oxygen enrichment causes a
reduction, beyond simple dilution, in the magnitude and intensity of transient
puffs from contained wastes.
5. If temperature increases can be controlled (e.g. by using advanced'forms
of process control), oxygen enrichment may cause net decreases in transient
puff emissions. Results of planned tests on such a system at EPA's Combustion
Research Facility in Arkansas may provide these answers.
6. Oxygen enrichment may cause unacceptable emissions of nitrogen oxides
(NOX) even at baseline conditions burning clean auxiliary fuel due to thermal
fixation of ambient nitrogen. These emissions may be exacerbated if high
nitrogen containing wastes are incinerated.
7. The data indicate that the rotary kiln serves primarily to pyrolize the
waste into the gas-phase, and that complete oxidation is not likely in the
primary chamber even with the help of oxygen enrichment. It may be desirable
to operate the kiln at reduced temperatures in order to retard the rate of
volatile release of waste species .from the solid or liquid phase, and rely
on the afterburner for complete organic destruction.
8. Efficient design and operation of afterburner systems for these incinerators
are crucial in order to minimize organic emissions, especially with regard to
transient emissions.
9. Oxygen enriched burners may be more useful when utilized as part of an
afterburner system where increases in temperature are desired and do not
cause increases in the rate of solid or contained liquid waste volatilization.
-287-
-------
ACKNOWLEDGMENTS
The authors would like to thank EPA/AEERL's R.W. Lowans for his technical
advice and assistance in the data collection phases of this project. Special
thanks to R.E. Hall, Branch Chief, Combustion Research Branch, AEERL, for his
guidance and support. . The authors would also like to acknowledge the
contribution of American Combustion, Inc., of Norcross, GA, who designed,
fabricated, and installed a prototype version of their Pyretron burner on
EPA's rotary kiln incinerator simulator. Additionally, the authors would
like to acknowledge the contribution of Liquid Air Corporation of Countryside,
IL, who provided the liquid oxygen for the tests. Finally, the authors would
like to thank the EPA's Hazardous Waste Engineering Research Laboratory
(HWERL) in Cincinnati, OH, for partial support of this work.
DISCLAIMER
The research described in this article has been reviewed by the Air and
Energy Engineering Research Laboratory, U.S. Environmental Protection Agency,
and approved for publication. The contents of this article should not be
construed to represent Agency policy nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
REFERENCES
1. W.P. Linak, J.D. Kilgroe, J.A. McSorley, J.O.L. Wendt, J.E. Dunn, "On
the Occurrence of Transient Puffs in a Rotary Kiln Incinerator Simulator:
I. Prototype Solid Plastic Wastes," J. Air Pollut. Control Assoc. 37(1):54
(1987).
2. W.P. Linak, J.A. McSorley, J.O.L. Wendt, J.E. Dunn, "On the Occurrence
of Transient Puffs in a Rotary Kiln Incinerator Simulator: II. Contained
Liquid Wastes on Sorbent," J. Air Pollut. Control Assoc. 37(8):934
(1987).
3. C.R. Hicks, "Fundamental Concepts in the Design of Experiments," 2nd
edition, Holt, Rinehart, and Winston Inc., New York, NY (1973).
-288-
-------
PATHWAYS OF PIC FORMATION IN HAZARDOUS WASTE INCINERATORS
B. Dellinger, P. H. Taylor, D. A. Tirey, and J. C. Pan
Environmental Sciences Group
University of Dayton Research Institute
Dayton, Ohio 45469
and
C. C. Lee
Hazardous Waste Engineering Research Laboratory
US-EPA
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268
ABSTRACT
InS
™* -e
formation is
There has been considerable controversy concerning the potential
for emissions of toxic products of incomplete combustion (PICs) from
incinerators burning hazardous organic materials. It appears that the
leV?ls-,?f,beth ""destroyed components of the waste
contro1 ed by chemical reaction kinetics. Thus, a
vari°uns chemical reaction mechanisms that can lead to PIC
valuable in predicting and controlling PIC emissions from
Hpfnp At lea?J ^hre? general ^chanistic pathways may be
defined that can result in the formation of thermally stable PICs- 1)
concerted molecular elimination, 2) molecular attack by reactive '
*i!d,I??-S' 3E 3) radical-atom and radical-radical recombination
involving thermodynamically stable species. Each class of reaction
can result in the formation of PICs which have different "apparent"
stability behavior. Examples of each type of reaction are presented
and some general conclusions concerning PIC emissions are discussed.'
INTRODUCTION
h,,™nnohf emis_sions of toxic organic compounds from incinerators
burning hazardous materials is one of the major technical and socio-
logical issues surrounding the further implementation of incineration
as a waste disposal alternative. [1] Laboratory-, pilot-, and full
scale studies have produced data indicating that properly designed and
operated facilities can achieve the destruction of toxic organic waste
feed components to environmentally acceptable levels. [2-51 However
rh = Jl ?rgani$ emissions from the incinerator effluent are seldom fully
characterized This has led to both scientific and public concern
?PT?CI S P0fS]ble emissions of toxic products of incomplete combustion
PICs) resulting from incineration of toxic materials under less than
±<-M°np}r10n?-[(?'7] .Thuns> the ab11ity to Predict and test for
possible PIC emissions is clearly desirable.
i— ent®d d?^ \hat* str°n9^ suggest that the relative
cntln« fpnSr^C-eSJDEs) ?f Jhe principal organic hazardous
constituents (POHCs) in the waste feed, i.e., their relative inciner-
-289-
-------
ability, are controlled by high-temperature thermal decomposition
kinetics in the post-flame regions of the incinerator.[8,9] Along
with other researchers, we have suggested that "pyrolysis pockets in
the post-flame zone are responsible for emissions of undestroyed
POHCs.[8-101 This has led us to develop an index of POHC inciner-
ability based on thermal stability measurements under oxygen-starved
conditions which has successfully correlated with full-scale emissions
data in numerous cases.[10,11] Unfortunately, this index does not
address PIC emissions or the potential impact on apparent POHC DE when
a POHC is also formed as a PIC during the combustion process.
Available data strongly suggests compounds selected as POHCs can also
be formed as PICs from other components of the waste or fuel feed.
This can significantly impact POHC relative emission rates and explain
many of the anomalies in laboratory predicted versus full-scale
determined POHC incinerability.[9]
As a first step towards developing guidelines for predicting
full-scale PIC emissions, we are producing PIC data on simple
chlorinated hydrocarbons.[12] However, completion of a data base on
PIC formation from Appendix VIII compounds is far in the future.[13]
Consequently, we have begun to analyze theoretical routes for PIC
formation in an effort to determine which routes are dominant and lead
to common, thermally stable PICs that can be used for incinerator
trial burn testing. The following paragraphs present a theoretical
discussion of these routes along with exemplary experimental data and
reaction kinetic calculations.
EXPERIMENTAL TECHNIQUE
The data presented in this manuscript were generated primarily
with the Thermal Decomposition Analytical System (TDAS). Only a brief
description of its design and operation is presented here, as a more
detailed description is available in other reports and publi-
cations. [14,15] The thermal decomposition unit of the TDAS consists
of a high-temperature fused silica tubular reactor in which a flowing
gas stream, exhibiting a segregated flow pattern, can be exposed to
temperatures as high as 1200°C for mean residence times of 0.5 to 6.0
sec. Reactor design ensures that samples experience a square-wave
thermal pulse with a narrow residence time distribution. Heated fused
silica transfer lines connect the insertion chamber to the reactor and
the reactor to the effluent analysis system. The analytical function
of the TDAS is conducted by a Hewlett Packard 5890B programmed
temperature gas chromatograph (GC) and 5970A Mass Selective Detector
(MSD). Data reduction is-achieved with a HP 59970 ChemStation and the
accompanying system software which includes an on-line NIH-EPA mass
spectral library.
DISCUSSION OF PIC FORMATION MECHANISMS
From a purely chemical reaction kinetic viewpoint, one may
classify known mechanisms of PIC formation into three general groups:
1) concerted molecular eliminations, 2) radical-molecule reactions,
and 3) radical-atom or radical-radical recombination reactions. As
illustrated in Table 1, these mechanisms occur in different regions of
the incinerator and can result in PICs of different structure and
stability. Consequently, the nature of PICs formed and their
-290-
-------
Th be dePendent on the different exposure
The reaction rates for PIC formation processes increasp
nrnp t0 sec°ndT°^er in the radical concentrations on2
proceeds from mechanism 1 to mechanism 3. Thus, stoichiometry and
P r v?pfJ c°mP°slt on °f the waste/fuel feed may signif canSy ?mpact
PIC yields for radical-molecule and radical-radical reaction patKaJs,
TABLE 1
OF HAZARDOUS WASTE INCINERATOR PIC FORMATION ZONES
AND DOMINANT PIC FORMATION MECHANISMS
REACTION CQNnTTTnNS PIC MECHANISMS
T = 200-1000'C Mechanism lb
R.T. < 1 sec Mechanism 2^
[02] = 0-50% EAe Mechanism 3d
DESCRIPTION
ZONE
Preflame Zone
Flame Zone
Postflame Zone
(afterburner)
Cool Zone
(APCD and stack)
T = 1000-1800°C
R.T. < 1 sec
[02] = 50% EAe
T = 600-1100°C
R.T. = 1-3 sec
[02] = 50-100%
T = 80-600°C
R.T. = 2-20 sec
[02] = 3-9%
Mechanism 2
Mechanism 1
Mechanism 3
Mechanism 2
Mechanism 1
Mechanism 3
Mechanism 3
formation mechanisms listed in decreasing ordeT^f~T^^tance" ----
bConcerted molecular elimination reactions. importance.
^Radical -molecule reactions.
dRadical-atom or radical-radical recombination reactions
deViati°nS dUe to P°°^ixing'can result in
MECHANISM 1: CONCERTED MOLECULAR ELIMINATION REACTIONS
Concerted molecular elimination to form stable PICs is a
reactions, viz., reactions only involving the parent
a second molecule or radical is not involved in
yield is dependent only upon time, temperature
parameters of the molecular elimination reaction-
%PIC YIELD = [PIC]/[POHC]0 x 100 = 100[l-exp{-At exp(-Ea/RT)} (Eqn 1)
where: [POHC]o = concentration of parent POHC- FPICl =
concentration.of PIC, t - reaction time; A = Arrhen us
relctiSn'"? ^hp^ rea?t1o"5 Ea = the activation energy for the
constant.' = ^action temperature; and R = the universal gas
are refe^to0^^^^^
presents the conversion of pentachToroetfiane (cIScis) tS tetrichloro-.
-291-
-------
ethene (C2C14) and hydrochloric acid (HC1) as an example of a four
center process. Reaction 2 presents the conversion of diethyl
phthalate to phthalic acid and ethene (C2H4)as an example of a six
center process.[17]
CIH
Cl-t^t-CI
cm
HCI
C-OC2H5
r>0-"H,
< '*»*
^0--CH2
(R2)
+ C2H4
The former is of particular interest because it illustrates a pathway
where a fragile POHC may decompose to form large yields of a thermally
stable PIC.[18] The latter is also of interest because it demonstrates
how phthalates are easily degraded and should not be observed in incin-
erator effluents. In light of these degradation pathways, reported
observations of phthalate emissions are likely the result of samp ing or
analytical error or ubiquitous sources, e.g., outgassing of phthalates in
plasticizers.[4]
Table 2 presents a list of kinetic parameters for POHCs of
interest to incineration that have been experimentally shown to
decompose by four- or six-center processes. The reaction rates tor
these molecules are all energetically favorable underv incineration
conditions with an extent of reaction greater than -99% at temperatures
below 700°C. Because these PIC formation pathways occur so rapidly,
the resultant PIC can experience almost the full gas-phase residence
time experienced by the parent POHC. Thus, PICs produced by molecular
elimination may undergo decomposition resulting in secondary PIC
formation, or the reaction may proceed to the formation of thermo-
dynamically stable combustion products.
-292-
-------
TABLE 2
EXPERIMENTALLY DOCUMENTED MOLECULAR ELIMINATION PIC FORMATION PATHWAYS
POHC
l,2-Dibromo-3-
chloropropane
1,1-Dichloro-
ethane
1,2-Dichloro-
ethane
1,1-Dichloro-
propane
1,2-Dichloro-
propane
1,3-Dichloro-
propane
2,2-Dichloro-
propane
1,2-Dibromo-
ethane
DDT
Diethylphthalate
Di-n-butyl-
phthalate
Di-n-octyl-
phthalate ,
Pentachloro-
ethane
1,1,1,2-Tetra-
chloroethane
1,1,2,2-Tetra-
chloroethane
1,1,1-Trichloro-
ethane
1,1,2-Trichloro-
ethane
1,2,3-Trichloro-
propane
PIC
l-Bromo-2-
chloropropene
Chloroethene
Chloroethene
Chloropropene
Chloropropene
Chloropropene
Chloropropene
Bromoethene
DDE
Phthalic acid
Phthalic acid
Phthalic acid
Tetrachloro-
ethene
Trichloro-
ethene
Trichloro-
ethene
Dichloroethene
Dichloroethene
Dichloro-
propene
Ea
MECHANISM fkcal/mole)
HBr Elim.
HC1 Elim.
HC1 Elim.
HC1 Elim.
HC1 Elim.
HC1 Elim.
HC1 Elim.
HBr Elim.
HC1 Elim.
C2H4 Elim.
C4H8 Elim.
C4H8 E,lim.
HC1 Elim.
HC1 Elim.
HC1 Elim.
HC1 Elim.
HC1 Elim.
HC1 Elim.
51.3
53.5
53.5
53.5
53.7
56.6
51.5
51.3
53.2
48.0
46.0
46.0
56.6
56.6
56.6
54.0
54.0
55.0
A
(I/sec)
4.17E13
2.82E13
1.41E13
. 3.16E13
1.26E13
6.32E13
5.01 El3
4.17E13
7.94E12
7.78E12
3.08E12
3.08E12
5.27E13
6.32E13
2.10E13
1.00E14
5.00E13
8.45E13
T 99
MECHANISM .2: RADICAL-MOLECULE REACTIONS
572
620
641
616
648
647
570
572
654
564
556
556
652
647
681
592
611
613
Examples of bimolecular reactions that result in stable PIC
formation are radical-molecule reactions. Once the decomposition of
the parent POHC is initiated and a pool of reactive radicals is
formed, radical-molecule reactions will usually be the dominant
pathway for PIC formation. Hydroxyl (OH) radicals are the dominant
reactive species under stoichiometric and oxidative conditions, while
H atoms are the dominant reactive species under pyrolytic
conditions.[19] In addition, polyatomic organic radicals are also
believed to play an important role in the formation of PICs that are
of higher molecular weight than the parent compound.
The dominant route for the decomposition of benzene under
oxidative conditions has been suggested to involve attack by OH or 0
-293-
-------
may then further decompose to form carbon monoxide (CO) and
cyclopentadienyl radical (C5H5-)- C5H5- may then react with oxygen to
form cyclopentadienonyl radical (C5H50-). which will decompose to
butadienyl radical (C4H5-) and CO. Once ring opening has occurred,
other stable organic products such as acetylene (C2H2) and CzH4 may
form This pathway is representative of the oxidation of substituted
aromatic species after any side chains or other functional groups have
been removed.
Fiqure 1 depicts the formation of polychlorinated dibenzofurans
(PCDFs) from 2,3',4,4',5-pentachlorobiphenyl (PCB).[21] A relatively
simple mechanism involving OH attack is illustrated in reaction 3. We
currently believe that Cl displacement by OH is the first step. This
is followed with H abstraction by OH forming the resonance stabilized
tetrachlorobiphenoxyl radical. The third step is ring closure forming
2,3,7,8-PCDF. This mechanism occurs with percent yields of products.
n 2,3'4,4',5-PCB
+ TRI-CDFs
TETRA-CDFs
PENTA-CDFs
.01
900 1000 1100
200 300 400 500 600 700 800
TEMPERATURE (°C)
Figure 1. PCB Degradation and PCDF Formation/Destruction Profiles.
The Gas-Phase Residence Time was 2.0 Seconds and the Fuel/Air
Equivalence Ratio was 1.0.
-294-
-------
Cl + OH -> Cl
Cl + H «- Cl
Cl + Cl
(R3)
°f
interest is the formation of
««r'
JrSSd1 YhbS6rVh? intermediates such as dich1o?oacety1ene
2) and 1,3-hexachlorobutadiene (C4Cle) is illustrated n
ions 4 through 7. 22 uiusiratea in
reactions 4 through 7. [22]
-295-
-------
UJ
-J
O
s
10"y-:
10'
,-10
10
,-11
B-
n Chloroform
H Tetrachloroethene
« Carbontetrachloride
a, Trichloroethene
+ Hexachlorobutadiene
x Hexachlorobenzene
200
800
900
1 1 1 r-
300 400 500 600 700
TEMPERATURE (°C)
Figure 2. Chloroform Degradation and Selected PIC Formation/
Destruction Profiles. The Gas-Phase Residence Time was 2.0 Seconds
and the Fuel/Air Equivalence Ratio was 0.67.
1000
Cl
CI
=c
ci
CI
:c=c: + HCI
CI
CI
ci' ^ci
> CI-CsC-CI + CI
Cl CI
(R4)
(R5)
NCI
VCI
XCI
cr
v CIN
CK i x1
x=c-c=c; + ci (RS )
CI
CI
cr
ci,
cr
CI
CI
CI
ri r=r-ri
CI-C=C CI
ci
+ CS (R6)
+ CI CR7)
-296-
-------
These examples demonstrate the complexity of the formation of
PICs by radical-molecule pathways. It should be recognized that since
the rate of PIC formation is dependent on the concentration of both
the radical and molecule reaction partners, the observed PIC yield
will generally increase with increasing waste feed rate. It is also
important to note that many reaction steps may be required for the
formation of a given PIC, e.g., Cede. Thus, the high-temperature,
gas-phase residence time may be very short for such PICs, and their DE
be much less than expected based on the full residence time in the
incinerator. This is in contrast to PICs formed by concerted
molecular elimination and suggests that for identical PIC formation
yields for these pathways, the radical-molecule route will result in
greater PIC emission rates.
MECHANISM 3: RADICAL-RADICAL RECOMBINATION REACTIONS
Recombination reactions of radicals or atoms to form stable
molecules are expected to occur only when molecule concentrations are
very low or radical concentrations are high. These reaction rates are
second order with respect to radical concentration and are temperature
independent or exhibit small negative temperature dependencies. Since
radical concentrations in the combustion zone are at least a factor of
100 times less than stable molecule concentrations, it can be shown
that the overall forward reaction rate for radical-molecule reactions
are typically greater than 50 times the radical-radical reaction rates
at temperatures in excess of 1200°C. At lower temperatures, e g
800°C, due to the higher activation energy for the radical-molecule
reactions, the ratio of rates decreases to -10. As an example, these
results were obtained by evaluating the following equation which
compares relative reaction rates for recombination of trichloromethyl
(CCls-) radicals and Cl atoms versus displacement of a Cl atoms from
C2C14 by CC13-:
Rate of Displacement = A/exDf-EaVRTUC9n£l[rn_?l (Eqn 2)
Rate of Recombination A exp(-Ea/RT) [Cl][CCl3] ~
Since product distributions are exponentially dependent on the
reaction rate, radical-molecule reactions are clearly kinetically
favored routes as compared to recombination reactions at temperatures
greater than 600°C.
Since chemical reaction kinetics in the high-temperature and
post-flame zones of incinerators are extremely rapid, thermodynamic
equilibrium codes can be used to predict with reasonable accuracy the
free radical concentrations at these temperatures.[23] Thus, radicals
with small (or negative) free energies for overall reaction pathways
(AGrxn) may be expected to be present in significant concentrations,
i.e, thermodynamically stable, and exit the flame zone. Table 3 lists
a large number of radicals that may form in hazardous waste inciner-
ators along with their free energies (AGf) and heats of formation
(AHf). The heat of formation of these radicals may be used along with
the heat of formation of possible dissociation products (AHdiss) to
calculate the energy required forBbond homolysis, which represents an
alternative pathway for destruction of the radicals. The lower the
-297-
-------
values of AGrxn and the higher the values of AHdiss, the more likely
these species will exit the flame zone of an incinerator.
Once in the cool zone, temperatures may be sufficiently low that
radical -molecule reactions with stable combustion endproducts now
present in high concentrations, e.g., C02, H20, and HC1 , occur at a
much slower rate. As a result, radical -radical recombination routes
may now become kinetically significant. If the temperature is rapidly
quenched downstream of the post-flame zone, the concentrations of
these radicals may be predicted from equilibrium calculations at the
temperature .just upstream of the quenching boundary. With this
information, the recombination product concentrations downstream of
the quenching boundary can be generally related to the predicted
radical concentrations. Once formed, these molecules are not
subjected to high temperatures and may exit the incinerator
undestroyed.
TABLE 3
THERMODYNAMIC PROPERTIES OF SELECTED RADICALS PRODUCED
IN HAZARDOUS WASTE INCINERATION
SPECIES
H
OH
H02
0
F
FO
CFO
CF2
CF3
Cl
CIO
CC10
CC12
CC13
CHC12
CH2C1
CHS
C2F5
C2C15
C2HC14
C2H2C13
C2H5
C2C13
C2HC12
C2H3
C3H5
n-C3H7
1-C3H7
C6H5
C7H7
(300 K)
(kcal/mole)
52.103
9.432
4.996
59.562
18.863
26.000
-41.000
-43.501
-112.403
28.924
24.192
-15.000
57.000
19.002
24.058
28.260
34.814
-213.000
8.897
5.387
11.800
26.500
53.000
52.000
69.000
40.600
21.000
17.600 v
70.043
45.000
AGf (1000 K)
(kcal/mole)
39.562
5.706
16.021
44.875
4.522
22.884
-50.043
-50.947
-101.283
15.547
21.123
-26.205
48.001
28.838
27.807
30.807
38.179
-171.048
48.535
40.618
45.032
53.593
55.363
54.153
64.452
64.112
71.157
69.693
87.593
93.430
REFERENCE
24
24
24
24
24
24
24
24
24
24
24
24
24
24
16
16
24
16
16
16
16
16
16
16
16
16
16
16
25
16
-298-
-------
SUMMARV
We have discussed how three fundamentally different mechanisms
for gas-phase formation of PICs can occur in a hazardous waste
incinerator. These mechanisms are: 1) concerted molecular elimination
reactions, 2) radical-molecule reactions, and 3) radical-atom or
radical-radical recombination reactions.
Concerted molecular elimination occurs very early in the
incinerator and produces mostly olefinic species. Radical-molecule
routes also occur in this region and extend further into the high-
temperature flame and post-flame zone. Radical-molecule reactions can
lead to fragmentation of POHCs to form lower molecular weight PICs or
form higher molecular weight PICs by radical chain mechanisms.
Relatively slow recombination reactions involving stable radicals are
likely to occur in the cooler downstream regions of the incinerator
and tend to result in the formation of highly fluorinated and
chlorinated alkanes. As a result of their formation in successively
downstream regions, PICs produced by recombination reactions
experience a concomitant decrease in their high-temperature, gas-phase
residence times. Since the ultimate yield of PICs in the effluent
depends on both the extent of formation and the extent of destruction,
PICs formed in the cooler zones may exhibit greater emission rates
However, the reactivity of each PIC will be a major factor and much
uncertainty still exists over which reaction class results in the
greatest emissions.
We are continuing our experimental studies of each reaction class
and the factors affecting their reaction rates. We are combining
these studies with kinetic and thermodynamic computations to provide
subroutines that may be included in scale-up models of incinerators.
ACKNOWLEDGMENTS
This research was .partially supported by the US-EPA under
cooperative agreement CR-813938.
LITERATURE CITED
1. Oppelt, E. T., APCA Journal. 37, 1987, pp. 558-586.
2. Lee, K. C., APCA Journal. 37, 1987, pp. 1011-1017.
3. Dellinger, B., "Theory and Practice of the Development of a
Practical Index of Hazardous Waste Incinerability," Hazard
Assessment of Chemicals-Current Developments. J. Saxenor, ed., in
press. .
4. Midwest Research Institute, "Performance Evaluation of Full-
scale Incinerators," US-EPA HWERL Report. EPA-600/2-84-181, 1984.
5. Castaldini, C., et al., Haz. Waste Haz. Mat.. 1, 1984, pp.
159-165.
-299-
-------
6. EPA Science Advisory Board, "Report on the Incineration of
Liquid Hazardous Waste by the Environmental Effects, Transport,
and Fate Committee," April 1984.
7. Hershkowitz, A., Techno!. Review, July 26, 1987.
8. Dellinger, B., et al., Haz. Waste Haz. Mat.. 3, 1986, pp.
139-152.
9. Dellinger, B., et al., Haz. Waste Haz. Mat., 3, 1986, pp.
293-307.
10. Tsang, W., "Fundamental Aspects of Key Issues on Hazardous
Waste Incineration," to be published.
11. Taylor, P. H. and Bellinger, B., "A Thermal Stability Based
Ranking of Hazardous Organic Compound Incinerability,"
Proceedings of the International Symposium on Incineration of
Hazardous. Municipal, and other Wastes, Palm Springs, CA 1987.
12. Bellinger, B., et al., "Products of Incomplete Combustion
from the High-Temperature Pyrolysis of Chlorinated Methanes," to
be presented at the 195th National Meeting of the American
Chemical Society, Toronto, Ontario, Canada, June 1988.
13. 40 CFR, Part 261, Appendix VIII, pp. 373-378.
14. Rubey, W. A. and Carnes, R. A., Rev. Sci. Instrum., 56, 1985,
pp. 1795-1798.
15. Rubey, W. A., "Design Characteristics of the Thermal
Decomposition Analytical System," US-EPA HWERL Report, EPA-600/2-
80-098, 1980.
16. Benson, S. W., Thermochemical Kinetics, 2nd ed., John Wiley,
New York, 1976.
17. Benson S. W. and O'Neal, H. E., "Kinetic Data on Gas-Phase
Unimolecular Reactions," NSRDS-NBS 21, National Bureau of
Standards, Washington, D.C., 1970.
18. Dellinger, B., et al., Haz. Waste Haz. Mat., 1, No. 2, 1984,
pp. 137-157.
19. Tsang, W., "High-Temperature Chemical and Thermal Stability
of Chlorinated Benzenes," Proceedings of the International
Symposium on Incineration of Hazardous, Municipal, and other
Wastes, Palm Springs, CA 1987.
20. Brezinsky, K., Prog. Energy Combust. Sci., 12, 1986, pp. 1-
24.
21. Rubey, W. A., et al., Chemosohere, 14, No. 10, 1985, pp.
1483-1494.
-300-
-------
22. Taylor, P. H. and Del/linger,B., "Thermal Degradation
Characteristics of Chloromethane Mixtures," Environ. Sci.
Techno!.. Vol. 22, No. 4, p. 438, 1988.
23. Reynolds, W. C., "STANJAN Equilibrium Program, Version 3.0,"
Department of Mechanical Engineering, Stanford University,
Stanford, ,CA, 1986.
24. Stull, D. R. and Prophet, H., JANAF Thermochemical Tables.
2nd Edition, NSRDS-NBS 37, National Bureau of Standards,
Washington, D.C., 1971.
25. Bahn, G. S., "Approximate Thermochemical Tables for Some C-H
abd C-H-0 Species," NASA CR-2178, 1973.
-301-
-------
"Studies of POHC DE During Simulated Atomization Failure in a
1 Turbulent Flame Reactor"
by
Daniel P.Y. Chang
Professor and Visiting Research Engineer
Department of Civil Engineering
University of California, Davis
Marta K. Richards, Research Chemist
and
George L. Huffman, Chief
Thermal Processes Research Staff
Thermal Destruction Branch
U.S. EPA Hazardous Waste Engineering Research Laboratory
ABSTRACT
The failure of flames to destroy waste molecules deter-
mines the residual amounts of POHCs and PICs that have to be
destroyed by subsequent thermal reactions in the post-flame
zone of a combustor. Oversized "rogue droplets" have been
suggested as one means by which POHCs escape a flame. This
paper presents results of POHC destruction efficiency (DE) ex-
periments under conditions of a simulated atomization failure
leading to the generation of oversize- droplets.
Four pairs of POHCs were burned with heptane in a turbu-
lent flame reactor (TFR) at constant H:C1 ratio under a simu-
lated condition of atomization failure. Compound DE was mea-
sured principally by GC/FID. Differences in compound DE were
observed and relative DE of each pair of POHCs was determined.
The observed order of incinerability from most to least dif-
ficult was 1,2,4-trichloro-benzene (TCB) > tetrachloroethene
(PERC) > trichloroethene (TCE) > 1,1,2,2-tetrachloroethane
(TCA) » 1-chloropentane (CP) and heptane (HEPT). Such a rank-
ing is consistent with limited burning-rate and low oxygen
thermal stability data, and with theoretical flame mode kine-
tics reasoning. Some PICs (benzene, chlorobenzene, TCE,
toluene) were found to increase with increasing CO concen-
tration for the particular failure condition examined, while
others appeared to be present at a more or less constant back-
ground level.
-302-
-------
INTRODUCTION
concePtuallzed as a two-stage
-zone, near adiabatic flame temperatures
reactions' and high concentrations of reactivl'
wi u/r;StiC ?f flamSS' C°nvert reactants introduce!
T< 0 1 ^ fuel/wjste mixture on relatively short time scales
« 01 s) . In the -post-flame, high-temperature, thermal -decom-
position zone (characterized by slower Reaction rateT lowl?
concentrations of reactive species, but residence tiSes on the
siSs ASno°ndS ' aPPfeciable conversion of reactants pe?- •
from ^hA?« explanation for the escape of some reactants
from the flame-zone can be attributed to the "rogue droplet"
hypothesis (Mulholland, 1987), in which an oversize particle of
fuel/waste passes out of the flame-zone before completely eva-
porating. In such a circumstance, the oversize SpJet could
behave as an isolated burning droplet. It is known that indi-
vxdual droplets which have been ignited in the fllme- zone w?!l
eventually extinguish in a coolerpost-flameenv^onmeS? before
complete consumption of the reactants (Law, 1975) In such a
case only one out of a multitude of. droplets need escape the
flame to account for the low levels of residual organic sub-
stances emxssions (ROSE) observed in poorly oper'atS Scinera-
If the conceptual models described above are reasonabl v
accurate, then a flame-based ranking scale funded on droplet
mTS£n? Chara?teristics (burning-rate and/or flame extinction)
might be developed to complement the University of Dayton
Research Institute's (UDRI) gas-phase, thermal stability ?TS)
'' Vi985a) * ^ 10glC ln W1^^ such rank-
-a.t-3SS-S2 fri-in-K~'
hazardous organic^onstitieSs ^HC) of
In liquid- injection incinerators, sprays are typically '
generated, and burning occurs at the periphery of
clouds of droplets. This occurs because of rate"
The term residual organic substance emission (ROSE) is
nhl *** ^ ^^ tO th°SS organic compounds
wch ar inth
bv ?L oLi ^ ed/ e'g' POHCs' but not completely destroyed
by the combustion process, those compounds which are possiblv
nlF?AmHVr'?eWlY f°rmed ±n the ^cinerator (not Smi?e5 to ?he
2S?^ definitlon of Product of incomplete combustion, Pic,
which requires appearance in "Appendix VIII") , and those com-
"
-303-
-------
the mixing of oxidant into the cloud. A manor question to be
addressed as part of the current prolect was whether any of the
results of single droplet burning experiments could be related
to sprav combustion when oversize • droplets, a simulated
"failure-mode", were purposely introduced into a turbulent
diffusion flame. This paper deals with an experimental study
of oversize droplets burning in a turbulent flame reactor (TFR)
located at the U.S. EPA's Center Hill Laboratory (Staley et
al./ 1986) and concommitant development of sampling and ana-
lytical methods in support of those experiments.
DESCRIPTION OF RESEARCH CONDUCTED
The research program consisted of two phases: 1) pre-
liminary development of a sampling methodology for simultaneous
collection of volatile and semi-volatile organic compounds
using multi-sorbent traps (TAG - TENAX-TAR, AmbersorbR, char-
coal) ; 2) relative destruction efficiency (DE) tests of four
pairs of chlorinated hydrocarbons (CHC) blended with heptane
(KEPT) under simulated atomization failure conditions (oversize
droplets).
Semi-volatile organic compounds (SVO) were of interest in
this study because data on single droplet burning were avail-
able only for compounds whose boiling points were greater than
90 °C. Volatile organic compounds (VOC) were also of interest
because previous work carried out on the TFR had concentrated
on VOCs. For reasons of equipment costs, data comparability
and physical limitations imposed by a thermal desorption unit
interfaced to the analytical equipment (EnvirochemR UNACON
810A), it was desirable that 6 mm o.d. X 4 mm i.d. X 203 mm long
glass tubes be used to house the sorbent traps. Two additional
constraining problems had to be overcome: one was the transmis-
sion of higher boiling compounds through the sampling line, and
the second was the avoidance of water vapor condensation in the
sorbent traps. These problems were addressed by maintaining
the sampling line (see Figure 1) and a commercially available
multi-sorbent TAG trap at higher temperatures (150 to 160 °C
and 74 °C, respectively) than is specified for the volatile
organic sampling train (VOST). A preliminary validation of the
sampling method was undertaken for purposes of this research
project. This consisted of retention and recovery experiments
using known quantities of standards "spiked" onto the sorbent
tubes and subsequently analyzed by GC/FID.'
Five compounds 1-chloropentane (CP), 1,1,2,2-tetrachloro-
ethane (TECA) , trichloroethene (TCE), tetrachloroethene (PERC)
and 1,2,4-trichlorobenzene (TCB) were selected for this study.
These were burned as pairs in the TFR to determine relative DE.
Previous work with the TFR (Kramlich et al., 1985) had demon-
strated that oversize droplets could be produced by simply re-
-304-
-------
V)
co
CO
C
CO
C
i_
H—
LU
CO
o
DC
O
o«o-o"«i£i§!i'g
•r- 01 t- cu a> s_ cu r- .!± .2
•£ •£ +•> -4-> 3 .0
Sto_ nj to -M t_ __ —
RtOco-CtoJ-toXg. o
c JH « OT S — >
QJ dj cu Q- «
a = " 2. = ^ 5 o S^
f •Ss..ST,"^i*sa
S"^ti2i ^.^ <- •— *-
?5.ffTjS.gs|«5
o
o
cu "? Q-°
N I— *"
t * *^~ •
S
Q-..
tOT^jc
- o> c o
^-*4J O -r- -n-
10 V_ .„.«= _
i" .^^^3 = o-i-
oj-^c^; ^« g..
«- ^"^ '° 5=J *- eu &
CU «,_ g20l£° t
i 15 F^ <" w J " ^
cu ^ n3 Q d ^™ ••* ••* m
tOcoO ijr^OO-rj'oTJ
^lotoeu"?- o *- «t-s_
tO n3 "- -^ S _ !* T ••- -M -r-
I D)
LU
CU
c T' "ti "3 •—" *^, o
wg'cufeS^o X r; c S
^ ,2 — TJ CU t_.« IS
...cxg-ca-cu+jQ.. §u
ycfoE^icns-oi-i— >,
1— n3
-------
ducing the pressure supplied to the atomizing nozzle. Thus a
1.5 gph, DelevanR WDA 60° hollow-cone nozzle operated at 40
psig was used to produce the oversize droplets. In order to
maintain the thermal environment in the TFR as close to the
same as possible, identical fuel feed rate, stoichiometry and
turbulent mixing (swirl setting) were used in each run. To
reduce variability in stoichiometry, it was decided to concoct
the mixtures such that equimolar amounts of each POHC would be
blended with heptane to yield an overall H:C1 ratio of_100:1
(about 5.3 weight % Cl). It was undesirable that a primary
decomposition product of a POHC also be one of the POHCs in a
mixture, therefore, mixtures of TECA with TCE were avoided
since TCE was a known decomposition product of TECA. Tne
composition of the "cocktails" is shown in Table 1.
Table 1
POHC Mixture Compositions
Compound
b.p.
M.W.
I
|CP
JTECA
ITCB
IPERC
ITCE
IHEPT
1
chloropentane
1, 1, 2, 2-tetrachloroethane
1,2, 4-trichlorobenzene
tetrachloroethene
trichloroethene
heptane (technical grade)
107.8
146.0
231.5
121.0
87.0
98.4
106.6
167 . 9
181.5
165.8
131.4
100.2
Mixture
1
|CP:TECA:HEPT
| TECA :TCB: KEPT
|PERC:TCB:HEPT
| TCE :TCB: KEPT
1
average composition
C
1.00
1.00
1.00
1.00
H
2,27
2.24
2.24
2.23
Cl
0.023
0.022
0.022
0.022
Cl cone.
wt . %
5.3%
5.3%
5.3%
5.3%
Two sorbent tube samples were also analyzed by mass spec-
trometry (GC/MS) on a chromatograph equipped with an identical
analytical column (J & W Scientific, DB-1, 30 m X 0.32 mm id.,
stationary phase cross-linked and bonded, fused silica capil-
lary column) to the GC/FID used for DE tests. By careful
matching of identified peaks and retention times, certain ROSEs
were also tentatively identified and quantified retrospec-
tively. These are listed in Table 2. Further details
-306-
-------
experimental Procedures are given by Chang et al
Table 2
Boiling Point and Retention Times
Compound
Combined data
b.p. °C MS FID
[sj [s]
dichloromethane (DCM)
Freon 113 (F113)
1-hexene'
chloroform (CHC13)
1,2-dichloroethane (EDC)
1-chlorobutane
benzene (BZ)
CC14 (CARB TET)
2-methylhexane
pentanal
3-methylhexane
trichloroethene (TCE)
heptane (KEPT)
methylcyclohexane
1-chloropentane (CP)
toluene (TOL)
hexanal
3-methyleneheptane
4-octene (identity?)
tetrachloroethene (PERC)
4-octene (identity?)
chlorobenzene (C1BZ)
1-chlorohexane
tetrachloroethane (TECA)
nonane
bromofluorobenzene (BFB)
benzaldehyde
benzofuran
1, 4-dichlorobenzene
nonanol
1, 2-dichlorobenzene
1-decene
1-chlorononane
3, 7-dimethyloctanol
1,2, 4-trichlorobenzene
1-chlorodecane
1-bromodecane
tetradecane .
5, 7-dimethylundecane
2,3, 6-trimethyldecane
40
47
63
61
84
78
80
76
90
103
92
87
98
100
110
128
122
1-21
122.
132.
134.
146.
150.
152.
178.
174.
174.
212.
179.
170.
203.
212.
231.
223.
240.
253.
.5
.7
.4
.2
.0
.4
.1
.8
.0
.0
.0
.0
.4
.4
.6
.0
0
0
0
0
5
0
8
0
1
0
0
0
0
6
0
0
5
4
6
7
313
330
452
485
546
577
608
624
668
686
692
717
762
816
879
916
987
1017
1022
1048
1066
1132
1175
1275
1337
1338
1410
1534
1559
1565
1572
1588
1713
1808
1973
2185
2349
2470
2832
2833
265
447
557
569
639
662
712
832
863
937
971
986
1015
1069
1132
1215
1278
1895
2110
2422
-307-
-------
RESULTS OF TAG SORBENT TUBE RETENTION AND RECOVERY STUDY
A triplicate set of tubes, to which known amounts of stan-
dards had been delivered, was exposed to two liters of com-
bustion gases from the TFR while technical grade heptane was
being burned. The tubes were subsequently analyzed_for the
amount of each POHC remaining and compared to identically
"spiked" tubes which had not been exposed. It was assumed that
the internal standard recovery, bromofluorobenzene (BFB), was
100% in each case, and the other compound recoveries were
normalized to the amount of BFB recovered. The results are
shown in Table 3 (column labelled "retention efficiency"). The
results clearly indicate that considerable loss of dichloro-
methane, carbon tetrachloride, and trichloroethene occurred at
the 74 °C test condition.
Table 3
Retention and Recovery Efficiency Results
Standard
Retention
efficiency
*benzene interference
n.a. = not applicable
std.
dev.
Recovery std.
efficiency dev.
DCM
CC14
TCE
KEPT
CP
PERC
CLBZ
TECA
TCB
BFB
19%
54%
41%
107%
107%
65%
103%
79%
105%
100%
13%
9%
15%
17%
16%
19%
16%
16%
20%
n.a.
27%
125%*
50%
116%
112%
71%
105%
82%
91%
100%
14%
24%
13%
10%
9%
9%
8%
6%
16%
n.a.
The results of "spiking" known quantities of the standards
in methanol solution at the head of the sampling line are also
shown in Table 3 (column labelled "recovery efficiency").In
this case the "spike" was evaporated into two liters of TFR
combustion gas which was passed through the TAG sorbent tube.
Slightly higher recoveries of the "spiked" compounds were
noted. In the oversize droplet POHC DE studies described below,
DEs were corrected for recovery efficiencies shown in Table J.
The above assessment of the TAG tubes and sampling line
indicated that while operation at elevated temperatures ensured
transmission of the POHCs of interest, SVOs, and prevention of
water vapor/HCl condensation, recovery of VOCs normally trapped
-308-
-------
by the VOST methodology was less than desired. However, the
multi-sorbent tubes did extend the range of compounds that
could be successfully captured on a single trap to both VOCs
and some SVOs. SVOs could be quantified and qualitative data
on VOCs could be obtained simultaneously.
OVERSIZE DROPLET RESULTS AND DISCUSSION ' . .
Clear differences in the DE of the fuel/POHC cocktails
were observed. The differences were smaller for the most
refractory compounds TCB and PERC, but were several orders-of-
magnitude between the most and least refractory POHCs tested
TCB and CP, respectively. A relative ranking of compounds was
constructed from the observed absolute penetrations (Pt = 1 -
DE) by normalizing to the TCB penetration. No physical signi-
ficance should be attached to the numerical values of the
relative Pt reported. The ranking is shown in Table 4 along
with a ranking based on heat-of-combustion (He), gas phase
thermal stability under high O2 (TSHi02) and low Oo (TSLoOo),
: and lastly theoretical flame mode kinetics (TFMK) under low Oo
conditions. The actual DEs observed for one control test
(TCE:TCB:KEPT) with nominally optimum droplet atomization and
for oversize droplets is also shown for comparison in Table 5.
The results of the oversize droplet DE study suggest that
a relatively small number of chlorine atoms in comparison to
the_hydrogen atoms, 1:100 (5.3% Cl by weight), can produce
noticeable effects on the combustion process as evidenced by
increased CO production (several hundred ppm above heptane
alone - see Figure 2). This difference in CO production is
believed to have resulted primarily from, chemical effects since
the reduced specific energy of the POHC/heptane mixtures re-
sulted in overall exhaust temperature decreases of only 10 to
15 C, while the change in theoretical adiabatic flame tempera-
ture, assuming the same amount of air supplied to the POHC/fuel
mix as for fuel alone, was also calculated to be less than
about 35 °C.
ASSESSMENT OF RESIDUAL ORGANIC SUBSTANCE EMISSIONS
Complete identification of all emissions was not the
primary goal of the project. However using a retention time
correlation constructed from comparison of GC/FID and GC/MS
analyses it was possible to tentatively identify some peaks
observed in the chromatograms. The concentrations of selected
products of incomplete combustion (PIC), tabulated as GC/FID
area of the PIC normalized by area of the internal standard
(BFB), is shown in Table 6.
-309-
-------
Table 4
Comparison of POHC DE with Various Ranking Methods
Relative
Compound Penetration Drop
Rank based upon:
He* TSHi02* TSLo02*
ITCB
1 * ^^*-»
j
IPERC
1
TCE
i — * ^^"-"
j
ITECA
j
CP
i ^^*»
1
IHEPT
1
l.OE+OO
6.0E-01
3.8E-01
7.6E-02
9.3E-04
5.2E-04
1
2
3
4
5
6
4
(64/272)
1
(11/272)
3
(22/272)
2
(16/272)
5
(NA)
6
3
(12/23)
1
(2/23)
2
(9/23)
4
(21/23) *
NA
(NA)
NA
1
(1/43)
2
3
4
(6/43)*
NA
' (NA)
NA
5 -i v, ^T-,Q
1
2
3
4
.5
6
cuo - ci .- ,..— . ..- -
reference sources used except for 1,1,2,2-TECA for which it was
assumed 1,1,1-trichloroethane would give similar ranking.
He - (CARB, 1986)
TsHi02 - (Dellinger, 1984) observed T99
TsLo02 - (Taylor, 1987)
TFMKL002 - (Tsang and Shaub, 1982; Tsang, 1987)
Table 5
Comparison of TCE:TCB:KEPT Penetration for Optimum and
Oversize Droplets
Penetration
Optimum Oversize
TCE
TCB
HEPT
6.2E-06
8.8E-06
5.9E-08
1
4
2
.4E-04 |
|
.IE-OS I
|
.6E-07 |
1
When burning TECA-containing mixtures, a chromatographic
peak assigned to TCE was clearly present. The ratio of TCE to
TECA in those cases ranged from about 1.5 to 2.5. TCE concen
trations were observed at levels about an order-of-magnitude
lower when TCE was a POHC in the mixture and TECA was absent.
-310-
-------
«U ••- O •
c -o ••-
•g 2 o ^
5 a. o -^ •
o> £=
j- _
-------
inn nf PICs that are specific to certain POHCs does
to occur in the oversize droplet model.
Table 6
Selected PIC Summary
GC/FID Relative Area Count
BZ +
POHC Mix DCM CHC13 EDC GARB TET** TCE TOL C1BZ
|CP:TECA:HEPT
1
|TECA:TCB:HEPT
1
|PERC:TCB:HEPT
1
| TCE :TCB: KEPT
0.61
0.14
0.12
NA
0.86
0.36
1.36
NA
0
0
0
. 55
.33
.35
NA
44.
22.
14.
21.
63
91
20
29
1
0
0
.41
.15
.33
0
0
.33
.28
NA
1.09
0.78
1.37
IHEPT* 0.04 0.67 0.44
|TCE:TCB:HEPT* 0.05 0.87 0.42
3.26 0.04 0.17 N.D.
2.68 0.02 0.26 N.D.
*0ptimum nozzle atomization pressure
**Believed to be primarily benzene
Benzene was observed in every POHC mix and in the heptane-
only sample? Senzene was apparently formed from one or more of
components of the fuel, possibly one of the cyclic alkanes
in abundance in the technical grade heptane. Whenever
amounts of benzene were formed, a smaller amount of
observed, independent of the POHCs contained
in the mixture. No chlorobenzene was observed when burning
heptane alone. This was interpreted to mean that commpji
r^mistrv ™« »i«n occurring, e.g., substitution
of a hydrogen atom on a ring by a chlorine atom.
With the limited data available, concentrations of PICs
were compared with CO concentrations. A consistent relation
ship was observed for CO and benzene and is shown in Figure 3.
ThS suacrests that the residual levels of CO and benzene ob-
served iTthl exhaust gases were related through a common phy-
mechanism, i.e., poor atomization, ^ -onnnnctron with a
che^Hr-al mechanism, i.e., interference of CO oxidation
chlorinated species . Note that the POHC mixture
spec .
TEc ?hat yielded the highest benzene level was not
the most difficult to destroy, but rather, relatively the
easiest.
-312-
-------
o>
u
c
o
o
O oi
o s-
a)
Q.
5
c: o
a>
tu
ni -i_» t ! .
to JO O X
XJ S_ O •!-
O 4-> n c
O)
en
-313-
-------
CONCLUSIONS
The following statements can be made on the basis of the
experiments presented in this paper (additional details are
presented in the parent reports, Chang et al.r 1988; Sorbo et
al., 1988):
1) The oversize droplet model, in a water-wall-
quenched TFR, provides a consistent picture of in-
complete combustion of large fuel/POHC droplets.
Flame-extinction phenomena observed in single droplet
combustion experiments and the "rogue droplet" hypo-
thesis are consistent with the observations.
However, the observations made are not conclusive.
2) The destruction efficiency of five POHCs was
established with the oversize droplet model. The
resulting ranking agreed with both gas-phase thermal
stability and theoretical flame-mode kinetics rank-
ings under low oxygen conditions. This was true even
though an overall fuel-lean stoichiometry prevailed
in the combustion unit under conditions of moderate
turbulent mixing. The order of difficulty of
destruction was TCB > PERC > TCE > TECA > CP > KEPT.
3) Given close attention to reproducing combustion
conditions as nearly as possible, increases in CO
level were apparent when simulating a "failure
condition" given relatively small additions of
chlorine (5% bv weight) into a hydrocarbon fuel.
4) Evidence was obtained for a relationship Between
CO and benzene emissions for the specific ^ failure
condition of oversize droplets combined with flame-
quenching under conditions of optimum stoichiometry.
5) Although the CP/TECA/HEPT mixture exhibited the
highest DE of the two POHCs in the mixture, it also
yielded the highest levels of the PIC benzene under
conditions of constant heat release rate and overall
stoichiometry.
6) Evidence for POHC-specific and POHC-independent
formation of PICs was obtained. TCE produced from
TCA was an example of the former, and chlorobenzene
was an example of the latter.
-314-
-------
REFERENCES
Chang, D.P.Y., Richards, M.K. and Huffman, G.L. "Spray
Combustion Studies of Surrogate Hazardous Waste Incineration."
Research Agreement
Dellinger B.> Torres, J.L., Rubey, W.A., Hall, D.L., -Graham,
?n « i ^t^Wtion °f the Thermal Decomposition Properties of
^U Selected Hazardous Organic Compounds." EPA-600/S2-84-138,
-i> y o ^r •
H i J-;',Hall, D. and Rubey, W. "Examination
of Fundamental Incinerability Indices for Hazardous Waste
Destruction." In: Incineration and Treatment of Ra7.ar-Ho11Q
Waste EPA/600/9-85/028, September, 1985a, p. 160. -
Dellinger, B., et al., "PIC Formation Under Pyrolytic and Air
Starved Conditions." Report on Grant No. CR 81-0783-01 to the
US EPA Hazardous Waste Engineering Research Laboratory,
Cincinnati, OH., 1985b. ,
Kramlich, J., Poncelet, E., Seeker, W.R. and Samuelsen, G.S.
»nrt S0^^ Study on the Effect of Atomization on Destruction
and Removal Efficiency for Liquid Hazardous Waste." In-
Incineration and Treatment of Hazardous Waste EPA/ 6 0~o79 -
85/028, September, 1985. p. 182. : ..
Law, C.K. "Asymptotic Theory for Ignition and Extinction in
Droplet Burning." Combustion and Flame. 24;«q-q«r 1975.
Mulholland, J.A. "Influence of Atomization Parameters on
Droplet Stream Trajectory and Incineration." In: Land
Disposal, Remedial Action. Incineration and Treatment of
Hazardous Waste, EPA-/600/9-87/015, 1987. p. 246.
Sorbo, N., Chang, D.P.Y., Steeper, R.R. and Law, C.K.
Engineering Evaluation and Control of Toxic Airborne Effluents
Part I: Single Droplet Studies of Surrogate Hazardous Waste
Incineration.," Draft Final Report, California Air Resources
Board Contract #A6-051-32, 1988.
Staley, L.J., Richards, M.K., Huffman, G.L. and Chang, D.P.Y
Incinerator Operating Parameters That Correlate With
Performance.", U.S. EPA Final Report on Project
D109/T/38/02/3435[A] Workplan f01269a, Hazardous Waste
Engineering Research Laboratory, Cincinnati, OH, June, 1986.
Taylor, P.H. and Dellinger, B. "A Thermal Stability Based
Ranking of Hazardous Organic Compound Incinerability." Paper
presented at the International Symposium on Incineration of
-315-
-------
Hazardous, Municipal and Other Wastes, American Flame Research,
Palm Springs, CA, Nov. 2-4, 1987.
Tsang, W. and Shaub, W. "Chemical Processes in the Incinera-
tion of Hazardous Materials." In; J. Exner (ed.), Detoxication
of Hazardous Waste, Science, Ann Arbor, MI, 1982. p.41.
Tsang, W. Personal Communication. Chemical Kinetics Division,
National Bureau of Standards, Gaithersburg, MD., September,
1987.
ACKNOWLEDGEMENT S
Many individuals contributed to this project. Without the
invaluable assistance of Mr. Albert F. Oberschlake, the experi-
mental program could not have been carried out. The efforts of
Ms. Laurel Staley and Mr. Lou Garcia in assisting with the
operation of the turbulent flame reactor (TFR) and collection
of the samples is also greatly appreciated. Drs. Barry
Dellinger and Wing Tsang supplied data for the low oxygen
environment ranking. Mr. Nelson Sorbo, Mr. Richard Steeper and
Professor C.K. Law at the University of California, Davis
carried out the study of single droplet burning of chlorinated
hydrocarbons which made it possible to interpret the results
obtained from the TFR. Ms. Shirley Tenhover assisted with the
day-to-day functioning of the Thermal Processes Research Group
and assisted with the final production of the report on which
this manuscript was based. The efforts of Mr. Seldon Heath in
the construction of and improvements made to the combustion
systems at the Center Hill Laboratory, before this series of
experiments began, and whose passing on occurred shortly after
his retirement from federal service are also remembered.
This study was made possible with the assistance of a
Cooperative Research Agreement #CR-813333-01 which supported
one author (Chang) during a sabbatic leave from the University
of California, Davis. For that opportunity he is grateful to
both the U.S. Environmental Protection Agency and the
University of California. The California Air Resources Board
supported the effort that was carried out at the University of
California, Davis under contract #A6-051-32.
-316-
-------
REVIEW OF RESEARCH TESTING ACTIVITIES
AT THE U.S. EPA COMBUSTION RESEARCH FACILITY
Robert E. Mournighan
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
and
Larry R. Waterland
Acurex Corporation
Environmental Systems Division
Combustion Research Facility
Jefferson, Arkansas 72079
ABSTRACT
During Fiscal Year 1987(FY87), 25 weeks of incineration testing took
place in which 53 individual incineration tests were completed under the
operations and research program at the EPA Combustion Research Facility in
tinn nfTh ^anS^' I"? Pro9rams completed included: an extended evalua-
tion of the fate of volatile trace metals fed to the liquid injection
incineration system (LIS) and testing to evaluate the valence state of
chromium discharges from the LIS; testing to support evaluation of the
composition .of residual discharges from the incineration of five listed
Jh?AmfH,WaSreSKfrr sP*cific sources; and preliminary evaluation testing of
the American Combustion Pyretron Thermal Destruction System as an innova-
tive Superfund site waste treatment technology.
INTRODUCTION
Far ™v U.S. Environmental Protection Agency's (EPA) Combustion Research
Facility (CRF) in Jefferson, Arkansas, is a 3,100 ft* experimental facility
which currently houses two pilot-scale incinerators, a rotary kiln inc ner-
ation system (RKS) and a liquid injection Incineration system(LIS)
and
and
. " * i "••" ~ ii^«iu mjev.u i mi i MI-i MSI du i on system(Llo)
their associated waste handling, emission control, process control and
safety equipment, with on-site laboratory facilities for waste characteriza-
tion and analysis of process performance samples.
The objectives of research projects conducted at the CRF have been and'
continue to be as follows:
* Iu d^ve1op technical information on the performance capability of
the hazardous waste incineration process to assist EPA Regional
Uffices and State environmental agencies in the review, assessment and
-317-
-------
issuance of reasonable and responsible permits for regulated hazardous
waste incineration facilities.
• To develop incinerator system performance data for regulated hazardous
wastes to support current incinerator regulations and performance
standards, and to provide a sound technical basis for any future
standards necessary.
* To develop an understanding of the hazardous waste incineration
process that enables the development of methods to predict the per-
formance of incinerators of varying scale and design for the major
classes of incinerable hazardous wastes as a function of key process
operating variables. These methods would also help to simplify and
perhaps reduce the cost of permit and compliance testing.
' To develop methods of improving the reliability and controllability
of the incineration process, including the use of destruction and
removal efficiency (ORE) surrogates.
of conducting specialized test burns (particularly
.. special waste materials) in support of specific
Regional Office permitting or enforcement actions or national priority
list site remediation efforts.
* To provide a means
for high hazard or
• To test the performance of new or advanced incinerator components or
sub-systems, or emission control devices.
Three major EPA Program Office regulatory development programs were
supported through test activities in FY87 (see Figure 1):
* The hazardous waste incinerator trace metal emission regulation
development program within the Office of Solid Waste via testing of
the fate of volatile trace metals fed to the liquid injection inciner-
ation system (LIS) and testing to evaluate the valence state of chro-
mium discharges from the LIS;
• Evaluating the composition of the residual discharges (scrubber blow-
down and kiln ash) from the incineration of listed hazardous wastes
from specific sources ("K-wastes"), to support the Agency's Land
Disposal Restriction Regulatory Program;
" The Superfund Innovative Technology Evaluation (SITE) program within
the Office of Remedial Action and Emergency Response via evaluating
the American Combustion (AC) Pyretron Thermal Destruction System as a
Superfund site waste treatment technology.
In addition, continuing support of the at-sea incineration program with-
in the Office of Water (OW) was supplied via a 1-week test program (the third
such program performed) to continue to validate the marine incineration bio-
logical assessment system (MIBAS).
-318-
-------
Week ending
Activity
Test days
1986
October
November
December
1987
January
February
March
April
May
June
July
August
September
. 3
10
17
24
31
7
14
21
26
5
12
19
23
31
9
16
23
30
6
13
20
27
6
13
20
27
3
I'D"
17
24
1
8
15
22
29
5
12
19
26
2
10
17
24
31
7
14
21,
28
4
11
18
25
30
BROS data reduction a
facility maintenanc
facility decontamin
<
Trace metal rangefind
Trace metal rangefind
Trace metal parametrii
nd report preparation,
e and repair, and
ation
ing
ng
: tests
Cr(+3)/Cr(+6) tests
K024 residuals characterization tests
MIBAS evaluation tests
Equipment malfunction (preparing for K001 tests)
Scrubber and RKS control system repair and
upgrade, AC burner system installation
AC burner system startup
Background burn and K001P characterization test
K001 residuals characterization tests
AC burner shakedown tests with K001P
. I
K086 residuals characterization tests
Ram feeder repair
K086 residuals characterization tests
RKS secondary air installation
K087 scoping tests
K087 residuals characterization tests
K037 scoping and BOAT tests
K037 residuals characterization tests
AC burner testing with K087
2 tests
2 tests
1 test
1 test
3 tests
2 tests
3 tests
3 tests
2 tests
3 tests
3 tests
7 tests
2 tests
2 tests
1 test
2 tests
1 test
3 tests
1 test
3 tests
2 tests
1 test
3 tests
Figure 1. CRF test schedule completed: FY87
-319-
-------
TRACE ELEMENT EMISSION AND DISCHARGE TESTING
Two series of incineration tests were conducted in FY87 specifically
to evaluate the fate of trace RCRA metals fed to the hazardous waste inciner-
ator The first series of tests was performed in the liquid injection incin-
eration system (LIS) at the CRF. These tests evaluated the fate of arsenic
and antimony fed in a synthetic waste mixture and changes in valance state
(trivalent/hexavalent) of chromium. Nineteen individual tests (including
several "range-finding" tests) were performed over an 11-week period.
In the first series of tests in the LIS the synthetic waste mixture
employed mixtures of methanol, chlorobenzene, and carbon tetrachlonde,
spiked with SbCls and As203. The specific objectives
of these tests were:
* To track arsenic and antimony emissions through the incinerator sys-
tem.
• To study the distribution of metals between flue gas and scrubber
blowdown water.
• To explore the relationship between metal emissions and incinerator
operating conditions and feed chlorine content.
• To evaluate whether the metals affect waste component destruction.
• To extend the data base on products of incomplete combustion (PIC)
emissions from the incineration of chlorinated hydrocarbons.
The test program studied the effects of three variables on the distri-
bution of the trace metals fed among discharge streams: feed chlorine con-
tent, incinerator temperature and combustion excess air.
Blends of methanol, carbon tetrachloride and chlorobenzene in the feed
produced the various inlet chlorine concentrations (0 to 33.8 percent).
Arsenic and antimony compounds were added to these blends in proportions
designed to give constant feed concentrations of 12 and 40 ppm, respec-
tively, for all tests.
The liquid feed entered the main combustion chamber of the LIS via a
steam-atomized nozzle at about 4b kg/hr (100 Ib/hr). Auxiliary propane
maintained the incinerator at temperatures that ranged from 1,137 to 1,4bU L
(2 079° to 2 642°F). Combustion air entered via swirl registers in tne bur-
ner to result in between 5.9 to 11.7 percent oxygen in the flue gas at the
incinerator exit.
EPA reference Method 5 was used to collect samples for particulate
loading and arsenic and antimony analyses. The Method 5 train was modified
to ensure collection of any antimony and arsenic that passed through the
particulate filter of the train. The sample train used consisted of a probe
and a qlass cyclone, followed by a filter and five impingers. The first
impinger contained 0.1 N NaOH. The second, third and fourth impingers con-
tained 0.2 M (NH4)2S208 + 0-02 M AgN03. The last impinger contained silica
-320-.
-------
gel. Following sample collection and digestion, furnace atomic absorption
(AA) methods were used to analyze for arsenic and antimony.
During the tests, Method 5 samples were obtained before and after the
packed tower (wet scrubber) to determine the effect of scrubbing on the
physical state of the compounds (particulate/vapor). Figure 2 shows the
results for antimony and Figure 3 shows the arsenic results.
The Figures show that, at the afterburner exit where temperatures were
above 760°C (1,400°F), the vapor phase (impinger samples) contained an aver-
age of 55 percent of the collected antimony and 35 of the collected arsenic.
At the scrubber exit where the temperature was generally at 74°C (165°F)
the vapor phase accounted for an average of 6 percent of the collected '
antimony and 3 percent of the arsenic. Figures 4 and 5 show the effect that
chlorine in the feed had on the metal distribution between the vapor
and particulate phases in the exhaust gases.
In addition, the data showed that the packed tower had an average
collection efficiency of 83 percent for antimony and 92 percent for arsenic.
Chromium Emissions(2)
The liquid injection system was used to determine chromium emissions in
the same manner as with arsenic and antimony. Chromium was spiked into the
feed for a set of three tests. In the last test, chlorinated compounds
and chlorobenzene) were spiked as well.
Figure 6 shows the collection efficiency of the scrubber system, with
and without chlorine. The data show that the chlorine content of the feed
may have considerable positive effect on scrubber efficiency.
Figure 7 shows the hexavalent chromium (Cr+6) content of the flue gas
particulate and the scrubber water. The values show that between 14 and 17
percent of the chrome measured in each stream consists of the hexavalent
species, with the remainder being trivalent chromium.
DESTRUCTION AND REMOVAL EFFICIENCIES (ORE)
The tests demonstrated acceptable DREs for both carbon tetrachloride and
chlorobenzene. Chlorobenzene DREs were higher than those for carbon tetra-
chloride. Carbon tetrachloride DREs were greater than 99.99 percent at the
afterburner exit. These increased to greater than 99.999 percent at the
scrubber exit. Similarly, chlorobenzene DREs ranged from 99.999 percent at
the afterburner exit to generally better than 99.9999 percent at the scrubber
exit. The available data do not suggest any discernible effect of tempera-
ture, feed composition, excess air, nor the presence of arsenic or antimony
on ORE. This is consitent with previous CRF data.
PRODUCTS OF INCOMPLETE COMBUSTION (PICs)
Incineration produces low levels of PICs. The CRF routinely analyzes
-321-
-------
LLJ
CC.
O
O
CC
LU
CD
o
CO
CE
LU
CD X
CC LU
LU
-322-
-------
CO
LU
CD
u_
\
QC
LLJ
GO
O
CO
cc
LU
QC LLJ
x
CE LU
LLJ
h-
U_
-323-
-------
lug
ocS
LJJ
CE
O
o
O
QC LU
O c/:)
bm
ICE
-324-
-------
pc
I
LO
LUCc
O
1
8
IT
O
O
LU
CL LU
CO
LLJ
m cc
-^ CO
Q_
-325-
-------
o
o
E
CCS
U_cp
*5
s
O
1
UJ
21
E
O
_j
O
X
' >
cc
LLJ
Lu
21
a:
g
I
o
o
I
o
-326-
-------
LU
DC
o
u-S
S!
o
CO
CO
03
LL
CL
LU
I
QC
•LJU.
00
CD
.D
-OE-
O
CO
'CO
<
CD'
LU
Z)
_l
LJL
I
o
-327-
-------
Volatile Organic Sampling Train (VOST) samples for 22 organic compounds with
GC/FID. For these tests at the afterburner exit, of these 22 compounds, •
carbon tetrachloride was present at the highest levels (>100 g/dscm).
Other chlorinated alkanes and alkenes were present in the 10 to 100 g/dscm
range. Chlorobenzene concentrations were similar to those of the common
chlorinated PICs.
At the scrubber exit, PIC concentrations are about one-tenth of those
at the scrubber inlet, i.e., afterburner exit. The levels of carbon tetra-
chloride levels ranged from 4.4 to 31 g/dscm. The highest concentrations
occurred at low excess air conditions. Except for one test, chlorobenzene
(the other POHC) levels ranged from 1.4 to 14 g/dscm and did not appear to
correlate with temperature or excess air. Other PICs include methylene
chloride which was present in all tests at relatively high levels (11 to
139 g/dscm). Other chlorinated alkanes and alkenes, hexane, benzene and
toluene were present at levels generally below 10 g/dscm.
LISTED WASTE INCINERATION
Incineration testing of five K-wastes was completed at the CRF from late
March through early September 1987. The wastes tested were:
' K024: Distillation bottoms from the production of phthalic anhydride
from napthalene
• K001: Bottom sediment sludge from the treatment of wastewater from
wood preserving processes that use creosote and/or penta-
chlorophenol (a pentachlorophenol process sludge was tested)
* K086: Solvent washes and sludges from cleaning equipment used in
the formation of inks containing chromium and lead
• K087: Decanter tank tar sludge from coking operations
* K037: Wastewater treatment sludges from the production of
disulfoton
The specific objectives of the tests were to:
* Establish that these wastes could be incinerated while complying
with the hazardous waste incinerator regulations
• Collect representative samples of the incineration residuals along
with waste feed samples leading to these residuals and establish the
hazardous characteristics (e.g., extraction procedure (EP) toxicity)
of the residuals, especially as compared to the parent waste
• Evaluate unit feed systems with actual hazardous wastes with diverse
handling properties
-328-
-------
Three tests were performed with the K024 waste in late March 1987.
Three tests were performed with the K001 waste in late June 1987. One test
with a K086 solvent wash waste was performed in late July 1987, and three
tests of three different K086 sludge wastes (one test per waste) were
performed in early August 1987. One shakedown test and three production
tests were performed with the K087 waste in late August. Finally,?one
shakedown and two production tests were performed with K037 waste in early
September 1987. The last tests were the most challenging in that all
incinerator operating personnel were required to don Level C protective
clothing and equipment. Level B protection for the incinerator room per-
sonnel sufficed for the other materials tested.
For all tests except those with the K086 solvent wash waste, waste
feed to the kiln was via the drum ram feed system. Thus, for these tests,
waste was packaged into 5.7 L (1.5 gal) "pilot-scale" fiberpack drums.
These drums were then batch fed, one at a time at a predetermined frequency,
into the kiln.
The K086 solvent wash waste was pumped into the kiln via a liquid feed
pump/water-cooled feed lance system.
The test period from late June to mid-September 1987 during which four
of the five K-waste test programs were conducted, along with two additional
test programs discussed in the next section, represented an intensive test
period during FY87. Over this period, the facility was in test operation
status for 12 weeks out of 14; 21 individual tests were completed.
SITE DEMONSTRATION TESTING
Under EPA1s Superfund Innovative Technologies Evaluation (SITE) program,
several innovative waste treatment technologies will be evaluated to determine
their applicability to Superfund site waste cleanup efforts. Under this
program, EPA will fund the evaluation while process vendors will finance
the treatment process construction and operation.
A demonstration of the American Combustion, Inc., (ACI) oxygen-enhanced
burner system (referred to as the Pyretron Thermal Destruction System)
retrofit to the CRF's rotary kiln incinerator system (RKS) is being performed
under the SITE program. ACI's proposal to EPA was one selected in the first
round of SITE program awards. This demonstration effort is being performed
with coordination from the EPA Office of Remedial Action and Emergency
Response.
Installation of the Pyretron system into the RKS at the facility occurred
during the RKS downtime to effect scrubber system and process control system
repairs in the period from mid-April through mid-June 1987. Startup and
initial system shakedown efforts were completed in early June. The new
burners installed were required to be capable of operating in a conventional
mode with air introduction as the only oxidant. Thus, despite the fact that
-329-
-------
most of the K-waste residuals characterization tests discussed above were
performed after installation of the new burner system, these tests were
conventional incineration technology with air-only combustion.
After the initial Pyretron system shakedown testing was completed with
no waste feed to the kiln (only propane fuel supplied heat input), two series
of shakedown, familiarization, and capabilities probing tests were completed.
The first of these employed K001 waste feed to the kiln and was completed in
early July 1987. Three tests were performed in this initial shakedown series.
The first test was performed at the same waste feedrate (50 kg/hr (110 Ib/hr)
and a drum fed every 3 min) as the K001 residuals characterization tests, but
with 35 percent oxygen participation in the combustion process. The second
test was performed using a programmed oxygen enhancement response at the
same overall waste feedrate, but with the feed mass per charge tripled and
the feed frequency at one-third that used in the first test (three 2.5 kg
(5.5 Ib) drums fed every 9 min). The objective of this test was to establish
that the increased transient puffs of uncombusted waste rising out of a batch
charge of the waste could be prevented, or their magnitude decreased, via use
of the Pyretron system.
The third test, also with programmed oxygen enhancement, was performed
at 50 percent increased waste feedrate but the same feed mass per charge as
the second test. The overall waste feedrate was increased by increasing the
feed frequency (three 2.5 kg (5.5 Ib) drums fed every 6 min). The objective
of this test was to show that increased waste feedrates could be achieved via
use of the Pyretron system.
During this first series of shakedown tests, the opportunity was taken
to support another EPA research program, the bioassay protocol evaluation
efforts being performed by the Health Effects Research Laboratory (HERL) in
Research Triangle Park (J. Lutix, HERL Coordinator, S. Nolan, AEERL Coordi-
nator). In this effort, the CRF provided access for the operation of a
dilution tunnel sampling train. The train is designed to collect samples
representative of the ambient samples resulting from flue gas discharges
from combustion sources.
The sampler was operated during the second two K001 shakedown tests.
Plans were to continue to use this sampler during the K086 residuals
characterization tests discussed in Section 4; however, scheduling problems
prevented this from being realized.
The second series of shakedown/finalization/capabilities probing tests
was performed in mid-September using K087 waste. These tests directly
followed the K037 waste residuals characterization tests. Two tests were
performed. The first test was performed at a waste feedrate of 36 kg/hr
(80 Ib/hr) with 35 percent oxygen participation in the combustion process.
One approximately 1.6 kg (3.5 Ib) drum was feed every 2.5 min. The second
test was performed with waste feedrate approaching 55 kg/hr (120 Ib/hr) with
programmed oxygen enhancement. The feed charge mass was doubled to about
1.6 kg (3.5 Ib) drums fed per charge. The charge frequency was decreased to
approach two drums fed every 3.5 min. The objective of this test was, again,
-330-
-------
to establish that transient puffs of uncombusted waste arising from each
batch charge could be prevented, or their magnitude decreased, and that an
increased waste feedrate could be achieved via the use of the Pyretron
system. The second test established that 55 kg/hr (120 Ib/hr) of this
relatively high heating value waste (up to 39 MJ/kg (17,000 Btu/lb)) was iust
over the maximum waste feedrate achievable in the RKS. Feedrates of this
waste at this magnitude result in kiln temperatures near the maximum tolerable.
The full demonstration program planned for this technology will be
completed in FY88. The demonstration tests will be conducted using waste
material excavated from the Stringfellow Superfund site near Riverside
California, mixed with K087 waste. The objective of the demonstration'
tests is to provide the data to evaluate three ACI claims regarding the
Pyretron system as follows:
'The Pyretron system with dynamic oxygen enhancement reduces the
magnitude of the transient high levels of organic emissions, CO
and soot ("puffs") that occur with repeated batch charging of
waste fed to a rotary kiln
The Pyretron system with oxygen enhancement is capable of achieving
the RCRA mandated 99.99 percent destruction and removal efficiency
(ORE) of principal organic hazardous constituents (POHCs) in wastes
incinerated at a higher,waste feedrate than conventional air-onlv
incineration y'
The Pyretron system is more economical than conventional incineration
REFERENCES
Lee, O.W., Ross, R.W., Vocque, R.H., Lewis, J. and Waterland, "Distribu-
tion of Trace Element Emissions' from the Liquid Injection Incinerator
System Combustion Research Facility," EPA/600/S2-87/054 Sept. 1987.
Draft Report: Lee, J.W. et al. "Distribution of Total and Hexavalent
Chromium in Emissions and Residuals from Pilot-Scale Liquid Injection
Incineration," March 1988.
-331-
-------
PILOT-SCALE BOILER COFIRING TESTS
TO INVESTIGATE NONSTEADY EFFECTS
H. B. Mason, J. A. Nicholson, R. J. DeRosier, and C. D. Wolbach
Acurex Corporation
Mountain View, California 94039
and
I. J. Licis
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
A simulated waste of semivolatile and volatile chlorinated organics was
cofired with gas, oil, and coal in a 2 million Btu/hr boiler simulator
operated under transient and off-design conditions. Nonsteady and off-design
operating modes tested include low excess air firing, atomization upsets,
transient firebox temperatures, and variable firing rate, residence time, and
waste of fuel flowrates. Waste destruction results showed uniformly high
destruction efficiencies (DE) even under major upset conditions where
combustion conditions were poor. The test average DE was 99.9993 percent for
volatile compounds and 99.9997 for semivolatiles compared to the average of
99.998 percent for various field tests. Waste destruction was generally less
efficient under upset conditions compared to normal operation. Products of
incomplete combustion (PICs) were emitted at average levels about seven times
the emissions of principal organic hazardous constituents (POHCs). The
analytical procedures accounted for about 90 percent of chlorinated volatile
PICs and about 80 percent of nonchlorinated hydrocarbons. Residual emissions
of POHCs and PICs were observed after waste firing was curtailed. The
hysteresis emission levels immediately after waste firing were of the same
order as emissions during normal operation; they were generally lower than
emissions during upsets. The severity of the hysteresis generally increased
as the boiler heat transfer surfaces became increasingly loaded with soot and
ash deposits. Emissions were higher than normal when deposits were being
removed by sootblowing. The severity of hysteresis was reduced after
sootblowing.
-332-
-------
INTRODUCTION
the t b,aliSff°r the Bending EPA boiler cofiring regulations is largely
the test data from a series of 14 industrial boiler cofiring field tests
conduc ed by EPA from 1982 to 1986. These tests were run to determ ne the
capabilities of industrial boilers for waste destruction under both
well-controlled and routine operation. The boiler test facility
specifications and operating conditions were selected to obtain a reasonable
coverage of the diverse range of industrial boiler designs and operating
conditions- Fi6ld ^ ^ W"e °btained for the following range of
• Firetube, package and field-erected watertube designs; 0.3 to greater
tnan 2 seconds residence time in main firebox
• 2,500 to 350,000 Ib/hr steam capacity
• Gas, oil, coal, and wood fuels
value from
0 to 100 percent heat input from wastes
The results of these 14 tests showed an overall mass-weighted waste
enerall10exhibiCienCy °f "'"8 percent- This high level of destruction was
.. oss broad range of designs, waste characteristics.
and operating conditions tested and prevailed even when the boilers were
intentionally operated in off-design upset conditions such as
irfacJ^fhr^^ air,' mal^nctionin9 atomizers, and waste flow transients.
olf n ^ruction efficiencies were so uniformly high that it was not
generally possible within the constraints of plant operation to identify and
run test conditions where the destruction was not acceptable.
Although the field test data base supporting the proposed standard
=ates uniformly high destruction efficiencies, additional information and
analysis ^needed to interpret several patterns observed in the data and to
generalize the effects of operation on destruction efficiency. One pattern
observed in the field tests was variability in DE results of up to an order of
magnitude, which was apparently not due to artifacts in the sampling and
analvsia scheme or to contamination. Additionally, the variations did not
with any boiler operational settings or with boiler performance
as CO or NOX emissions or smoke. It was also observed at
several sites that the stack concentrations of waste species continued for
=fter waste firing was curtailed> This hysteresis effect may be
deposition of trace concentrations of waste species on boiler
particularly on areas with carbonaceous deposits from oil or coal
* potential role of soot deposits in the observed hysteresis and
fh.t- «h ^H" ^ariabUifcv was also indicated by limited sootblowing results
that showed higher emissions during routine sootblowing from the boiler
convective section than during normal operation. To gain more confidence in
stack
-333-
-------
the field test data and assist in-the data interpretation, data are needed to
isolate the presence and extent of emission variability and to relate these
data to operational settings and the conditions of the boiler heat exchange
surfaces.
A second key issue from the field tests is a more comprehensive
characterization of the types and quantities of PICs emanating from the POHCs,
or from the conventional fuel. Several tests showed PIC emissions in boilers
as a higher fraction of the organic compounds in the flue gas than were
observed in incinerators. More data are also needed on the effects of
operating conditions on the relative concentrations of PICs and POHCs.
A third issue is the degree to which nonsteady or off-design boiler
operation affects destruction efficiency. The field tests were somewhat
limited in the types and severity of imposed upsets but showed a broad
operational window with high destruction efficiency. To generalize this
observation, more data are needed on the effects on destruction of various
types of upset imposed with more severity than was possible in the field.
An additional issue, not reported on in this paper, is the form and fate
of heavy metals contained in the waste or fuel.
The present pilot-scale tests addressed the above issues. The
pilot-scale approach,was selected because of wider operational flexibility,
tighter experimental control, and lower cost compared to field testing.
TEST PROGRAM
To address the issues cited above, seven test series were run, as
follows:
I Baseline — fuel only
II Baseline — cofiring
III Data variability, replication, and destruction
IV Evaluation of operational window for high destruction
V Form and fate of trace metals
VI Evaluation of PIC production
VII Confirmatory tests
The series I and II baseline tests were run to obtain reference levels of
waste and PIC emissions under weel-controlled operation and for clean heat
exchange surfaces in the facility. These results were used for reference-
level comparison to later runs in series III and IV as deposits continued to
build on heat exchange surfaces, thereby indicating the relative magnitude of
background levels or organics and the transient, hysteresis, buildup of the
-334-
-------
background. Series III was intended to gauge the confidence level in the
destruction data relative to variability and reproducibility effects. 'With
this level established, the remainder of the test plan was directed at
determining the effect of operating conditions on POHC, PIC, and metal
emissions.
Series TV provided for parametric testing of various off-design and upset
operating conditions to attempt to define the operating'boundary of the high
DE window. Operating parameters tested included low excess air; atomizer
upsets; heat exchange designs affecting residence time and firebox
temperature; firing capacity; and waste or fuel flow transients. Series V
focused on doping the flame with metal species of interest in the regulatory
proces: arsenic, cadmium, chromium, and lead, and sampling for metal
concentrations in several size fractions. Series VI featured a spot check of
substituting inorganic chloride for the chlorinated waste to determine if
waste identity influences PIC identity over and above the presence of
chlorine. The final runs in series VII were designed to test conditions
which, on the basis or prior results, showed potential for confirmation of- DE
levels and trends.
TEST FACILITY '
The pilot-scale facility used for these test series is a refractory-lined
furnace and convective section designed with geometric similarity and
time-temperature simulation of a tangentially or wall-fired watertube boiler,
as shown in Figure 1. For the present tests, the unit was fired in the
wall-fired mode, typically with gas, oil, or coal fired through three of the
burners and a simulated waste through the remaining burner. The fuel feed
system, cooling system, and post-combustion cleanup system are shown
. schematically in Figure 2. For most tests, the unit was fired at a rate of
1.5 million Btu/hr, which gives a residence time and temperature history in
the firebox representative of full-scale boilers in the radiative and early
convective sections, as shown in Figure 3.
The wall-fired burners were variable-swirl research burners equipped with
a single fuel injector tube. Combustion air was preheated to upstream of the
burner by an electric heater. Steam atomization was used for oil and waste
firing. The convective section consists of approximately 20 tube bundle heat
exchange drawers that are cooled to about 300°F with Dowtherm.
For oil firing, a 2.2 percent sulfur residual oil with a heating value of
18,300 Btu/lb was used for the three burners fired on conventional fuel. The
oil was heated to about 220°F by drum heaters and heat traced recirculation
loops. For coal firing, an Illinois No. 6 coal with a 3.6 percent sulfur
content was pneumatically conveyed to the burners.
The synthetic waste fired through the remaining burner was a blended
mixture of 80 percent by weight No. 2 distillate oil and 5 percent each carbon
tetrachloride, trichloroethylene, monochlorobenzene, and trichlorobenzene.
The mixture was recirculated continually with a pump to retard stratification
and fed to the steam-atomized burner at room temperature. For baseline
-335-
-------
testing in the absence of waste,
fuel.
all four burners were fired on conventional
Thermocouples, sampling ports, and viewports were positioned along the
furnace and convective section. Temperatures, flowrates, and emission
measurements were recorded and processed on a microcomputer or datalogger.
Continuous monitoring was done for 02, CO (two scales), C02, hydrocarbons,
S02» and opacity.
Volatile waste constituents and PICs were sampled in the flue gas with
the VOST. Two methods of using the VOST were employed. For periodic
comprehensive quantification, the full-VOST sampling protocol was used with
gas chromatography/mass spectroscopy (GC/MS) analysis following thermal
desorption. To reduce the cost and turnaround time for multiple samples, a
reduced protocol, termed "mini-VOST," was used. The sampling was done on a
single VOST trap for 10 minutes, and then the trap was taken to a nearby gas
chromatograph and desorbed to a parallel flame ionization detector and a Hall
detector to quantitate hydrocarbons and chlorinated species, respectively.
This procedure can provide near-real-time feedback on volatile organic waste
concentrations.
Semivolatile wastes and PICs were sampled by a Modified Method 5 train
fitted with an organics module containing 65 grams of XAD resin. The train
was typically run for approximately 4 hours to a total sample volume of
4.3 dscm (150 dscf). The expended XAD was Soxhlet extracted, along with other
catches from the Method 5, and analyzed with the GC/MS.
RESULTS
To address the test objectives, 30 test series were run over 28 test days
during early 1987. The resultant sampling and analysis for these runs
included 151 full-VOST or mini-VOST runs for volatile organic speciation and
18 Modified Method 5 runs for semivolatile organic speciation. Samples were
taken for normal routine operation, as well as for a range of induced upset
conditions, and during sootblowing. For several runs, samples were taken at
up to four stations ranging from the near flame region out to the flue gas.
Simultaneous mini-VOST and Modified Method 5 sampling was done at two stations
to detect differences due to residence time or temperature during upset
runs. Samples were also taken at specified intervals after waste firing was
stopped to detect the amount and duration of any hysteresis emission "tail."
Tables 1 and 2 summarize the overall test average DE for volatile and
semivolatile POHCs, respectivley. The tables also include a total POHC plus
PIC result. For these tables, PICs were specified as volatile or semivolatile
chlorinated organics. This designation was based on the fact that the POHCs
were all chlorinated. Some chlorine was also contained in the parent fuels
however, and that could yield some nonwaste-related PICs. Also, the
chlorinated POHCs could produce nonchlorinated PICs. These factors preclude a
precise identification or closure of PICs.
-336-
-------
The overall mass-weighted DE average was 99.9993 percent for the three
volatile POHCs and 99.9997 percent for the semivolatile POHC. These
destruction efficiencies are higher than the 99.998 percent average for the
^eilpefSf P°ssible because the pilot-scale system was intentionally clean
ILI ^ ?h8 T f thS °UtSet °f thS testin9- Th* differences cited in DE
among the three fuels and the volatile POHCs were not major; and, in all
cases, DE was well above 99.99 percent, even when total chlorinated POHC plus
PIC compounds were included as an aggregate.
The DE results on Figures 4 and 5 do not show a strong effect in going
from normal operation to upset operation. This observation requires some
interpretation in relation to what constitutes an upset. In this regard,
Figures 4 and 5 show DE as a function of average flue gas CO emissions
(corrected to 7 percent oxygen) during a run. Figure 4 shows total
chlorinated volatile compounds, POHCs plus PlCs, and Figure 5 shows only the
carbon tetrachloride results. These results show some correspondence between-
higher destruction at low CO and fewer high destruction results at high CO
These plots are useful to discuss the concept of upset operation. Some of'the
very low CO results shown in Figures 4 and 5 were operationally classified as
upsets since they involved atomization maladjustments, waste flow surges, and
r^e^°- PS60 °r 'ranisent settings. in some cases, these settings did not
result in CO or other emission excursions and would not be classified as
f-f", ^ S emissi°n-based criteria were used to segregate "normal" operation
rrom "upset" operation.
_ Figures 6 and 7 show the normal and upset results when the upset
designation is arbitrarily assigned to any run with an average CO level above
200 ppm. This procedure thereby reclassifies as normal the runs initially
designated as upsets but with low CO. This approach to segregating runs by CO
emissions shows a generally significant and consistent trend of lowered
destruction efficiency resulting from upsets producing combustible
emissions. This result for the pilot-scale tests rs more pronounced and
consistent than was observed in the field tests, possibly because of higher
!rC"r"OU"r"emiSf °^ an* hvsteresis effects with the field boilers. In Figure
-MM ; G , and O- refer to coal, gas, and oil, respectively, and "U» and '
N refer to upset and normal. The only situation that deviated from the
trend cited above was for monochlorobenzene cofired with coal. All other
combinations showed lower destruction efficiency under upset conditions.
One of the purposes of the present tests was to examine the nature and
extent of background emission variability or noise and hysteresis. The
facility was initially free of POHC exposure and waste firing was introduced
only after fuel baselines. The results showed that during the fuel baseline
and the early days of cofiring, the POHC and PIC emission levels were very
low, frequently nondetectable. As the upset operating settings were
introduced and the deposits built up on the convective section, background
emissions began routinely appearing. These background emissions were
quantified by sampling after cofiring was stopped, with the unit operating on
gas only. The level of the background appears to depend on the elapsed time
since cofiring had been curtailed, on the types of upsets imposed prior to
stopping cofiring, and on the total number of days of facility exposure to
-337-
-------
POHCs. The background, or hysteresis, thus appears to have both short-term
and long-term components.
Figure 8 shows the residual emissions at various intervals after cofiring
was curtailed for total volatile chlorinated POHCs and PICs. There is a
general trend for the background to decay with elapsed time since cofiring.
At the 15- and 65-hour samples, there are numerous data points on the abcissa
that outweigh the outliers. It should be'emphasized that this facility was
operated on waste for 6 to 8 hours per day, 5 days per week, and on gas only
during the remaining time. The deposits, therefore, had an opportunity to
bake-out or desorb during nights and weekends. Alternate duty cycles could ,
cause the hysteresis results to differ. , .
Figure 9 shows the historical variation of'the hysteresis effect by
comparing emissions sampled at 15 hours after'cofiring cessation, but at
different total number of days' operation of the facility on waste. These
results show a general pendency for the background emission levels to increase
with exposure of the facility to POHCs. Soot was blown from the tubes on day
50, which may account for the higher backgrounds around day 40.
Emissions were measured during sootblowing, at 5 separate stations of the
coiwective section. Results indicate that significant organics were emitted
during sootblowing from the first three high-temperature stations. Emissions
during sootblowing were up to an order of magnitude higher than normal.
Subsequent tests showed lower emissions after sootblowing than during normal
operation. Hysteresis results also were lower after sootblowing than during
operation with deposits on tubes.
The identification and quantification of PICs generally showed a
reasonable level of closure. Mini-VOST chromatograms were run for the flame
ionization detector, which shows primarily nonchlorinated hydrocarbons and the
Hall detector, which speciates chlorinated organics. The PICs identified were
similar to the compounds identified in the field tests from the full-VOST
analysis. The total emission loading quantified for chlorinated organics was
generally over 90 percent of the total chromatogram response. The closure on
PIC identification for hydrocarbons was about 80 percent of the emission
loading basis. The closure on the number of peaks identified was lower
because of the large number of hydrocarbon species emitted in small
quantities. For the semivolatile chromatogram, most of the large peaks were
identified, but due to a strong background emission level, the. closure on
total emissions was only about 50 percent.
The mass-weighted ratios of PICs to POHCs in the. flue gas are listed in
Table 3 for volatile and semivolatile chlorinated compounds. The
mass-weighted average for volatile compounds was about 7, somewhat lower than
the average of about 15 observed in the field. Table 3 and Figure 10 also
show a pseudo PIC-based DE which compares volatile chlorinated PIC emissions
to POHC feedrate to the boiler. The PIC with increasing CO trends are similar
to the decreasing destruction efficiency observed for POHCs.
-338-
-------
Con vect I've jtct 1 on
Mot'sampling sorts
To btghouse
Ash pit
•*«HMU
Figure 1. Pilot-scale experimental furnace
Domhffli
cooling rddutor
In Out
(cooling tMter)
Figure 2. Pilot-scale research furnace system
-339-
-------
• Coal-fired utility
• Industrial package
watertube
Pilot-scale no cooling
1,000
1 2
Residence time (sec)
Figure 3. Time-temperature simulation
100-
89.99999-
"i^
53
ll
U)&
g
89.9999-
89.999
89.99
99.9
;
f
I a
Mo ° a ° 0
KB" BB» n«
b+o «o__n a •
r + BOB n B
' * ° D B
— 1 ' ' '. ' .V, ' I'K •> 2.4 2J
0.4 03
o Upsets
CO PPM
(thousands)
+ Steady-state
Figure 4. Effect of upsets on DRE: total volatile chlorinated species
-340-
-------
»^
ii-s
jl
*-» *>
8*
100-
99.99999-
99.9999 -
99.999 -
99.99 .
99.9-
e
M n — f
jL .- .:
r- * *
a + +
I * *
o + * . *+ * *
3
I
i
" 01 1Jt U t i.4 U
c^T*'
O »tt«W-JTOt + UHfa
Figure 5. Effect of upset on ORE: CC1
Coal Normal
Gas Normal
Gas Upset
Oil Normal
Oil Upset
Figure 6. Effect of upsets: total volatile chlorinated species
-341-
-------
EZ CU
Figure 7. Effect of upsets: POHC and fuel effect
Gas, Oil Hours After CF
te
1
J
1
"c
2
ft
*?
g
H
22-t
21-
20-
19-
18-
17-
16-
15-
14-
13-
12-
11-
10-
9-
B-
7-
6-
6-
4-
3-
2-
1-
W
0
D
•
0
i
B °
0 20 40 60
Hour After Cessation of Cofiring
Tbtal (PIC + POHC)
Figure 8. Hysteresis: total volatile organic species measured after cofiring
-342-
-------
5
o
1
1
(3
I
'
6-
6-
4-
3-
2'-
1-
0-
- — Q _
°
n
o
' A
$Q - +
•' . •. * A
0 20 40 60
Day of firing
0 Total (PIC +POHC) + CCI4 • TCE « MCB
Figure 9. Hysteresis: samples taken 15 hours after cofiring
AV\J -
^
!. . ~~" " B 7 1
D
ff . ;
n
fro a
99.9999 4| %
r s° „. .v
V ° ° a ° a ft 0°
99.999 -i D D Da
I d °
99.99 "
99.9 -
c
0
e
i i : i i i i i , j , — ; — | — _,
°^ a8 '"rtv^--*.''* * t* «
CO ppm (ave)
Figure 10. PIC-CO correlation
-343-
-------
TABLE 1. VOLATILE ORE SUMMARY
Gas
Oil
Coal
Normal
Upset
Overall
POHC
+ PIC CC114
99.9952 99.9991
99.9960 99.9997
99.9963 99.9998
99.9995 99.9997
99.9982 99.9995
average 99.9961 99.9997
TABLE 2. SEMIVOLATILE ORE
POHC
+ PIC
Gas 99.99945
Oil 99.9988
Coal 99.99921
Normal 99.99938
Upset 99.99928
TCE
99.9977
99.9995
99.9993
99.9998
99.9981
99.9993
SUMMARY
TCB
99.99994
99.99993
99.99917
99.9998
99.99976
MCB
99.9994
99.9995
99.9973
99.9996
99.9983
99.9992
Overall average 99.9993 99.99978
-344-
-------
TABLE 3. PRODUCTS OF INCOMPLETE COMBUSTION
Volatile
ORE PIC/POHC
Gas 99.99906 4.1
Oil 99.99967 11.5
Coal 99.99912 3.2
Normal 99.99977 7.8
Upset 99.99895 11.2
Overall average 99.99953 7.2
Semivolatile
ORE
99.99953
99.99941
99.99921
99.99973
99.99933
99.99945
PIC/POHC
9.8
2.8
1.0
4.0
2.1
2.6
-345-
-------
OXIDATION OF POLYCHLORINATED BIPHENYLS BY PSEUDOMONAS PUTIDA LB400
\ Sadhana Singh., Frank K. Higson, Louise M. Nadim and David T. Gibson,
Center for Applied Microbiology
The University of Texas at Austin
Austin, Texas 78712
ABSTRACT
Pseudomonas putida LB400 was originally isolated from an environment
contaminated with polychlorinated biphenyls. Gas chromatography/mass
spectrometry analyses revealed that this organism has the ability to
degrade a wide range of individual chlorinated biphenyls. However, little
information is available on the initial reactions involved in the degrada-
tion of these compounds. Biphenyl-grown cells were shown to oxidize 2,2'-,
2,5,2'S1-, 2,4,5,2',5'-, and 2,4,4'-chlorobiphenyIs and some reaction pro-
ducts were tentatively identified by gas chromatography/mass spectrometry.
Under the same experimental conditions 4,4'- and 2,4,3',4'-chlorobiphenyIs
were not metabolized. Similar results were obtained with an enzyme prepara-
tion from biphenyl-grown cells of P_. putida LB400. Current studies are
directed towards identifying the reaction products formed from the above
chlorinated biphenyls by intact cells and purified enzymes.
-346-
-------
INTRODUCTION
Polychlorinated biphenyls (PCBs) are extremely stable compounds and
have been used extensively by industry in capacitors, transformers,
dielectric fluids, fire retardants and plasticizers (8,10). Although these
compounds •• (Aro.clors) are no longer synthesized for use in open systems
considerable amounts of PCBs have been introduced into the environment
where they tend to persist for long periods' of time.
The resistance of highly chlorinated PCBs to biodegradation by aerobic
organisms has been well documented (5,6) and their lipophilic properties
have led to their concentration in organisms at the end of food chains.
Nevertheless, evidence has been pre; inted to show that PCBs, particularly
the more highly chlorinated isomers and congeners, can undergo microbial
reductive dehalogenation in anoxic sediments (4). Since many of the lower
(less heavily chlorinated) PCBs are known to be susceptible to degradation
by aerobic bacteria it is not inconceivable that protocols could be devel-
oped for the removal of PCBs from certain polluted environments.
In the present study we have focused on the ability of biphenyl-grown
cells of P_. putida LB400 to degrade 2,2'-, 4,4'-^ 2,4,4'- 2 4 3' 4'-
2,5,2',5'-, and 2 ,4 ,5 ,2 ' ,5 '-chlorobiphenyls (CBs). This' organism is
thought to ^metabolize biphenyl by the reactions shown in Figure 1 (9) and
it was anticipated that chlorinated analogs of these metabolites would be
produced during the degradation of the individual PCBs.
, OH
COOH
Cg Acid
2,3-Dihydroxybipheny! Benzole acid
cte-Biphenyl dihydrodlo! Ring fission product
© Biphenyl dioxygenase
(D ejls-Biphenyl dihydrodiol dehydrogenasa
(§) 2,3-Dihydroxybiphenyl dioxygenase
@ Hydrolyase
Figure 1. Initial reactions in the oxidation of biphenyl by P putida
LB400. - *
-347-
-------
MATERIALS AND METHODS
MICROORGANISMS AND GROWTH CONDITIONS
Pseudomonas putida strain LB400 was originally isolated from a PCB-con-
taminated site by enrichment culture with biphenyl as the sole source of
carbon and energy (3). The organism was grown in 50 ml of mineral salts
medium containing sodium succinate (0.2% w/v) as the source of carbon.
After growth at 30° C for 9 hours biphenyl (0.1% w/v) was added to the cul-
ture medium and growth allowed to continue for a further 9 hours. At this
time the cell suspension was filtered through glass wool to remove residue
biphenyl and centrifuged at 10,000 x g for 15 min. The cells were washed
twice with 0.05M KH2P04 buffer pH 7.5 and resuspended in 50 ml of the
same buffer.
TRANSFORMATION EXPERIMENTS
Transformations of individual PCBs were carried out in 250 ml conical
flasks containing 50 ml of biphenyl-induced cells at a turbidity of 1.0 at
600 nm. Substrates (0.5 mg in 0.05 ml of methanol) were added to each
flask. At designated times the contents of each flask were extracted three
times with ethylacetate. The organic phase was dried over anhydrous sodium
sulfate and the solvent removed in vacuo at 30° C. The residue remaining
in the flask was dissolved in 1.5 ml of methanol and stored in 3.0 ml vials
until required for analysis by high pressure liquid chromatography (HPLC)
or gas chromatography/mass spectrometry (GC/MS).
PREPARATION OF CELL EXTRACTS
Biphenyl-grown cells of P_. putida LB400 were disrupted by sonication
and centrifuged at 100,000 x g for one hour. The clear supernatant solu-
tion was used as a source of crude cell extract. The ability of cell ex-
tracts to oxidize biphenyl and PCBs was determined polarographically using
a Clarke-type oxygen electrode.
RESULTS
OXIDATION OF SELECTED PCBs BY BIPHENYL-INDUCED CELLS OF P_. PUTIDA LB400.
Table 1 shows that there is considerable variation in the rates of
-348-
-------
degradation of certain PCBs. The rates observed appear to be a function of
the^position, rather than the number, of chlorine substituents. Thus 2,2'-
CB is rapidly metabolized whereas in the same time period 4,4'-CB was'not
attacked. In contrast to 4,4'-CB and 2,4,3',4'-CB, the trichlorinated
2,4,4 -congener was significantly metabolized during the one hour incuba-
tion period. The presence of a 2,5-dichlorinated phenyl ring has been
reported to facilitate degradation (2) and this is supported by the results
with 2,5,2',5'-CB and 2,4,5,2',5'-CB which were degraded by 60 and 44
percent respectively.
TABLE 1.
DEGRADATION OF SELECTED PCBs BY P. PUTIDA LB400*
Time
(minutes)
*
PCB
% Remaining
2,2'- 4,4'- 2,4,4'- 2,4,3',4'- 2,5,2',5'- 2,4,5,2',5'-
0 100
30 0
60 0
100
100
100
100
85
74
100
98
93
100
69
40
100
75
66
Individual PCBs isomers were incubated with biphenyl-induced cells of P.
putida LB400 for the times shown. PCB concentrations were determined by
gas chromatography after work up as described in Materials and Methods.
DEGRADATION OF 2,2'-CB
Biphenyl-induced cells of P_. putida LB400 rapidly oxidized 2,2'-CB. Most
of the substrate (86%) was metabolized during the first 10 minutes of the
incubation period and after 30 minutes no 2,2'-CB could be detected in the
reaction mixture. Analysis of the reaction products by GC/MS revealed the
presence of four metabolites with molecular ions at m/e 156, 170, 238 and
256 respectively. The identity of these compounds designated A, B, C, and D
respectively are given in Table 2.
The formation of the methyl ester of 2-chlorobenzoic acid was unexpected
and was attributed to the use of methanol as a solvent for 2,2'-CB. This
was confirmed by repeating the experiment with 2,2'-CB dissolved in deuter-
ated methanol (CD3OD). The methyl ester formed gave a parent ion at M+
173. Nevertheless, small amounts of the methyl ester of 2-chlorobenzoic
acid were detected when 2,2'-CB was added in dimethylformamide.
-349-
-------
TABLE 2.
IDENTITY OF COMPOUNDS A, B, C, AND g FORMED FROM 2,2 '-CB BY
P. PUTIDA LB400
Compound
M
ci35/ci37
Structure
A
B
156/158
170/172
238/242
256/260
2-Chlorobenzoic acid
Methyl ester of 2-chlorobenzoic acid
2,2'-Dichlorobiphenylol
Dihydr.odiol derivative of 2,2 '-dichlorobi-
phenyl
* Products were isolated and identified by GC/MS.
f It was not possible to assign the position(s) of the hydroxyl groups in
these metabolites.
In addition, methyl ester formation was not detected when 2-chlorobenzoic
acid was incubated with biphenyl-induced cells of P.. putida LB400. These
results could indicate that 2-chlorobenzoic acid is not taken up by P.
putida LB400 and that ester formation from endogenous methyl donors may be
catalyzed by an intracellular enzyme.
DEGRADATION OF 2,5,2',5'-CB
Significant metabolism of 2,5,2',5'-CB was observed over a 1.0 hour
time period (Table 1). At 30 min a major metabolite with a molecular ion
weight of m/z 324 was detected. Over the next 30 min this compound began
to disappear and was replaced with a second major metabolite (molecular
weight 358). These compounds have tentatively been identified as 3,4-dihy-
droxy-3,4-dihydro-2,5,2',5'-tetrachlorobiphenyl and 3,4,3',4'-tetrahydroxy-
3431 4'-tetrahydro-2,5,2',5'-tetrachlorobiphenyl respectively. In addi-
tion to these major products small amounts of 2,5-dichloroacetophenone and
a compound tentatively identified as 2,2[21,5'-dichlorophenyl] ethanol were
also detected. Similar results were presented at last year s conference
(9).
-350-
-------
DEGRADATION OF 2,4,5,2',5'-CB
This pentachlorobiphenyl was rapidly oxidized by biphenyl-induced cells
of P_. putida LB400. After one hour 34% of the parent compound had disap-
peared and a major neutral metabolite with a retention time of 20.5 min was
detected by HPLC. The exact structure of this reaction product which ap-
pears to be a phenolic metabolite has yet to be determined.
DEGRADATION OF 4,4'-CB, 2,4,3',4'-CB AND 2,4,4'-CB
^ Biphenyl-induced cells of P_. putida LB400 failed to oxidize either
4,4'-CB or 2,4,3',4'-CB under the experimental conditions described in the
Materials and Methods section. Even after 48 hours incubation no degrada-
tion products were detected by GC/MS. In contrast, 2,4,4'-CB was oxidized
under the same conditions. The only metabolite detected by GC/MS gave a
molecular ion at M 272, and was tentatively identified as a phenolic
derivative of 2,4,4'-CB.
ENZYMATIC OXIDATION OF BIPHENYL AND SELECTED PCBs
Cell extracts prepared from biphenyl-grown cells of P_. putida 39/D
oxidized biphenyl and the substrates shown in Table 3.
TABLE 3.
OXIDATION OF BIPHENYL AND PCBs BY CELL EXTRACTS
Substrate
Activity
Biphenyl
2,2'-CB
4,4'-CB
2,5,2',5-CB
2,4,5,2',5'-CB
2,4,4'-CB
2,4,3',4'-CB
2,3,4,5',6'-CB
100
68
0
45
53
21
0
0
Enzyme activity was measured polarographi-
cally as described in Materials and Methods,
-351-
-------
DISCUSSION
The results shown in Table 1 confirm previous investigations which have
shown that P_. putida LB400 has the capability of oxidizing a wide range of
PCBs (1,3,9). However, most of these studies did not determine the rate of
PCB degradation at time periods as short as one hour. When one considers
that 2,2'-CB completely disappears after 30 minutes incubation with P_.
putida LB400 whereas there is no detectable metabolism of 4,4'-CB over the
same time period it seems clear that the position of the chlorine substi-
tuents plays a major role in the susceptibility of different PCB isomers
and congeners to microbial degradation. In this context it is of interest
to note that chlorination at the 2 position enhances the degradation of
4,4'-CB. Bedard et al. (2) have clearly shown that a 2,5-chlorination
pattern enhances the degradation of certain PCB congeners and this is sup-
ported in the present study by the results obtained with 2,5,2',5'-CB and
2,4,5,2I,5'-CB.
The metabolites detected as degradation products from 2,2'-CB suggest
that this molecule is oxidized at the 5,6-position to form a dihydrodiol
which can then be further metabolized to 2-chlorobenzoic acid. These
reactions are analogous to those shown in Figure 1 for the degradation ^ of
biphenyl. However, the position of the hydroxyl groups in the dihydrodiol
metabolite have not been established. It is of interest to note that a
mutant strain (FM408) of P_. putida LB400 oxidizes biphenyl to
cis-2,3-dihydroxy-2,3-dihydrobiphenyl. A genetic lesion in the structural
"gene for cis-2,3-dihydroxy-2,3-dihydrobiphenyl dehydrogenase causes the
accumulation~~of the dihydrodiol metabolite (7). However when 2,2'-CB was
incubated with strain FM408 the major metabolite formed was 2-chlorobenzoic
acid (data not shown). Thus there may be more than one enzyme in P_. put ida
LB400 that can degrade PCBs. For example, the oxidation of 2,5,2',5'-CB at
the 3,4-positions may be catalyzed by an enzyme different to that respon-
sible for initiating the degradation of biphenyl.
A comparison of the metabolites identified in the present study with
those reported in reference 5 is shown in Table 4. The failure to detect a
specific metabolite does not mean that the metabolite is not formed. Thus
the table reveals the limitations of the information that can be derived
from whole cell studies and GC/MS analyses. We believe that more specific
details of PCB degradation can be provided by studies with purified enzymes
and the results shown in Table 3 represent our preliminary studies on the
enzyme(s) responsible for the initial attack on the PCB ring. It _ is
anticipated that further investigations with purified enzymes will provide
more specific information on the mechanisms involved in PCB degradation.
-352-
-------
TABLE 4. COMPARISON OF THE METABOLITES FORMED FROM SELECTED CB CONGENERS
BY JP. PUTIDA LB400, ALCALIGENES EUTROPHUS SP. Y423 AND
ACINETOBACTER SP. P6a
2,2
CB
isomer
i _
Dihydro-
diol
mb—
mm
Dihydroxy
compound
c d
- -O
Chloroben-
zoate
e
x o monoh
Other
ydroxv compound (•)
4,4'-
2,4,4'-
2,5,2',5'-
2,4,3',4'-
'-
- o
— o
— - o
X O
X O
- - o
benzole acid methyl ester (•)
monohydroxy compound (•,o)
2,5-dichloroacetophenone (•)
2,2[2' ,5'-dichlorophenyl]
ethanol 0")
monohydroxy compound (o)
monohydroxy compound (o)
a Data taken from reference 5
b P_. putida LB400 (•)
c none detected
d Acinetobacter sp. P6 (o)
e Alcaligenes sp. Y42 (x)
f metabolites formed but not identified
ACKNOWLEDGMENTS
This work was supported by grant CR812727 from the Office of Research
and Development, the Environmental Protection Agency. We thank Dr. P.R.
Sferra, EPA Project Officer for his interest and suggestions. We thank
Catherine Potter for her patience and skill in organizing and typing the
manuscript.
REFERENCES
1.
Bedard, D.L., Unterman, R., Bopp, L.H., Brennan, M.J., Haberl, M.L.
and Johnson, C. Rapid assay for screening and characterizing
microorganisms for the ability to degrade polychlorinated biphenyls.
Appl. Environ. Microbiol. 51: 761, 1986.
-353-
-------
Bedard, D.L., Haberl, M.L., May, R.J. and Brennan, M.J. Evidence for
novel mechanisms of polychlorinated biphenyl metabolism in Alcaligenes
eutrophus H850. Appl. Environ. Microbiol. 53: 1103, 1987.
Bopp, L.H. Degradation of highly chlorinated PCBs
strain LB400. J. Ind. Microbiol. 1: 23, 1986.
by Pseudomonas
4. Brown, J.F., Jr., Bedard, D.L., Brennan, M.J., Carnahan, J.C., Feng,
H. and Wagner, R.E. Polychlorinated biphenyl degradation in aquatic
sediments. Science 236; 709, 1987.
5. Furukawa, K., Tomizuka, N., Kamibayashi, A. Effect of chlorine sub-
stitution on the bacterial metabolism of various polychlorinated
biphenyls. Appl. Environ. Microbiol. 38: 301, 1979.
6. Furukawa, K. Microbial degradation of polychlorinated biphenyls
(PCBs) In; A.M. Chakrabarty (ed.), Biodegradation and Detoxification
of Environmental Pollutants. CRC Press, Boca Raton, Florida, 1982.
p. 33.
7. Gibson, D.T., Roberts, R.L., Wells, M.C. and Kobal, V.M. Oxidation of
biphenyl by a Beijerinckia sp. Biochem. Biophys. Res. Commun. 50:
211, 1973.
8. Hutzinger, O.H.O., Safe, S.H., Zitko, V. Commercial PCS preparations,
properties and composition. In; The Chemistry of PCBs. CRC Press,
Cleveland, Ohio, 1974. p. 8.
9. Nadim, L.M., Schocken, M.J., Higson, F.K., Gibson, D.T., Bedard, D.L.,
Bopp, L.H. and MondellOi F.J. Bacterial oxidation of polychlorinated
biphenyls. In: Proceedings of the Second International Conference on
New Frontiers for Hazardous Waste Management. U.S.E.P.A., Cincinnati
1987. p. .
10. Safe, S.H. Microbial degradation of polychlorinated biphenyls. In;
D.T. Gibson (ed.), Microbial Degradation of Organic Compounds. Marcel
Dekker, Inc., New York, 1984. p. 361.
-354-
-------
Biological Oxidations of Organic Compounds
Enzymes from a White Rot Fungus
by: J.A. Bumpus, G. Mileski, B. Brock, W. Ashbaugh,
S.D. Aust
Biotechnology Center
Utah State University
Logan, UT 84322-4430
ABSTRACT
The ability of the white rot fungus Phanerochaete chrvsosporium to
degrade a wide variety of structurally diverse organopollutants is dependent
upon the ligmn degrading system of this microorganism. In part, the lignin
degrading system consists of a family of peroxidases, which are able to
catalyze the initial oxidative depolymerization of lignin. In addition
these enzymes catalyze the initial oxidation of many "hard-to-degrade"
xenobiotics. Current research is focused on ways to increase ligninase
activity in culture and on understanding the enzymology of this system as it
affects xenobiotic oxidation. Ligninase activity is expressed by the funqus
in response to nutrient nitrogen limiting conditions. In agitated cultures
(1 L total volume in 2.8 L Fernbach flasks), ligninase activities up to 446
units (/ancles of veratryl alcohol oxidized) per liter, representing
approximately a 20-fold increase over standard stationary cultures, were
observed. _ Furthermore, ligninase activity was shown to be cyclic, reaching
peak activities at about 30 hour intervals. Substantial ligninase activity
(200 units/liter) was also observed in 100 ml stationary cultures containing
I ^ veratzyl alcohol representing approximately a 10 fold increase over
standard stationary cultures, m these cultures, veratryl alcohol may serve
to induce ligninase biosynthesis and/or it may protect the enzyme from
subsequent onactivation. Ligninase H-2 was the predominant or second most
predominant ligninase isozyme produced in culture and was purified to
electrophoretic homogeneity by Fast Protein Liquid Chromatography (FPLC)
In addition to veratryl alcohol oxidation (oxidation of a benzylic carbon)
^^•"f36 H~2 catalyzed the oxidative dechlorination of pentachlorophenol '
(PGP) to form 2,3,5,6-tetrachloro-2,5-cyclohexadiene-l,4-dione, N-demethyl-
ation of crystal violet to form N,N,N',N',N»-pentamethylpararosaniline and
the oxidation of pyrene, presumably forming the quinones of this polycyclic
aromatic hydrocarbon.
-355-
-------
INIRDEUCITON
Hie ability of the white rot fungus Kianerodtiaete dhrysosporium to
degrade a wide variety of structurally diverse organopollutants to carbon
dioxide has been shown to be dependent upon the lignin degrading system of
this fungus, that is expressed under idiqphasic conditions (1-8) as a
consequence of nutrient (nitrogen, carbon or sulfur) limitation (1-10). In
part, the lignin degrading system consists of a family of enzymes, commonly
referred to as ligninases, that are able to catalyze the oxidative
depolymerization of lignin (11,12). Following depolymerization, the _ smaller
and more water soluble lignin-derived metabolites may be further modified by
these enzymes or by a family of manganese-dependent peroxidases that are
also secreted by the fungus during idiophasic metabolism (13-15). The
lignin degrading system also contains an aromatic ring cleavage system (16).
It is unclear at present whether or not cleavage of the aromatic ring
moieties of lignin, IB vivo, takes place before, after or during lignin
depolymerization. However, regardless of this, it is apparent that aromatic
ring cleavage of many different low molecular weight compounds readily
occurs in ligninolytic cultures of this fungus (1-8,16). Intermediates of
lignin and xenobiotic degradation are metabolized to Krebs cycle
intermediates and, eventually, to carbon dioxide to complete the
mineralization process.
The enzymology and protein chemistry of ligninases has been the subject
of considerable attention since their discovery in 1983 (11,17). It is
known that ligninases are more properly known as lignin peroxidases in that
they contain protoporphyrin IX as a prosthetic group and form intermediates
similar to the compound I and compound II intermediates of other peroxidases
(18,19,20). All ligninases characterized to date are glycoproteins having
molecular weights between 38,000 and 46,000 (19). Ligninase isozymes are
designated as H-l through H-10. The letter H refers to the fact that the
isozyme in question is a heme-protein while the number designates the order
in which the isozyme elutes from a mono Q (Hiarmacia, Uppsala, Sweden) anion
exchange column during Fast Protein Liquid Chromatography (21). To date
most research has been performed using ligninase H-8, the most abundant
ligninase expressed in standard stationary cultures of P.. dhrysosporium.
This enzyme is able to catalyze benzylic alcohol oxidation, carbon-^carbon
bond cleavage, hydroxylation, phenol dimerization and O-demethylation of
lignin model compounds (12). A complementary ENA (cENA) of ligninase H-8
has also been cloned and expressed in Escherichia coli and the cloned cENA
has been seguenced revealing the primary amino acid sequence of ligninase H-
8 (22).
Less attention has been focused on the biodegradation of environmental
pollutants by ligninases. However, it has been shown that ligninase H-8 is
able to oxidize benzo[a]pyrene as well as a number of other polyaromatic
hydrocarbons. In the case of benzo[a]pyrene, the benzo[a]pyrene 1,6-, 3,6-
and 6,12-quinones are formed (23) whereas pyrene is oxidized to the pyrene-
1,6- and 1,8-diones (24). Dibenzo[p]dioxin and 2-chlorodibenzo[p]dioxin are
-356-
-------
also oxidized by ligninase H-8 (24). Of importance is the fact that
ligninase catalyzed oxidation of dibenzo[p]dioxin occurs via a cation
radical intermediate (24) as is observed during the degradation of lignin
model compounds, (25). ihis suggests that the non-specific free radical
mechanisms active in lignin biodegradation are operative in ligninase-
catalyzed oxidation of environmental pollutants (24).
The use of P.. chrvsosporium and its biodegradative enzymes for the
treatment of organochemical wastes holds considerable potential. However a
more thorough understanding of the enzymology of biodegradation as it occurs
m this fungus is required if this potential is to be fully realized. The
present study focuses on the production of ligninases in cultures of P
chrysosporium. It also describes the purification and characterization of
ligninase H-2 with special attention to its ability to oxidize several model
organopollutants.
METHODS AND MATERIALS
MICROORGANISM
Phanerochaete chrysosporium (EKM-F-1767) was obtained from the United
States _ Department of Agriculture, Forest Products Laboratory (Madison, WI)
and maintained on malt agar slants at room temperature. Subcultures were
routinely made every 30-60 days.
CULTURE CONDITIONS
K chrysosporium was incubated at 37°C in the liquid culture medium (10
ml) previously described (26) in 250 ml Wheaton bottles equipped with caps
having a Teflon seal. This medium consists of 56 mM glucose, 1.2 mM
ammonium tartrate, mineral salts and thiamine (1 mg/L) in 20 mM
dajnethylsuccinate (sodium) buffer, pH 4.5. Cultures were established by
inoculating the medium with spores as described (27). During the first
three days of incubation, cultures were allowed to grow in an atmosphere of
air. After three days, the cultures were grown in an atmosphere of 100%
oxygen. Under these conditions, which are referred to as "basal
conditions", ligninase activities of 15-20 /zmoles/L were obtained after 5-6
days of incubation. P.. chrvsosporium was also cultured in 100 ml of the
basal media supplemented with 1.5 mM veratyl alcohol in 2.8 L Fernbach
flasks. As with the 10 ml stationary cultures, these cultures were grown
under an atmosphere of air for the first three days after which they were
grown under an atmosphere of 100% oxygen. Large scale (1 L) agitated
cultures of P.. chrysosporium were also used. Agitated cultures were
initiated from two day old mat cultures as described (28) except that the
culture medium of the mat cultures consisted of 10 mM dimethylsuccinate
(sodium), pH 4.5, containing 56 mM glucose, 1.2 mM ammonium tartrate,
thiarnine and a mineral supplement (27) and the cultures were grown under
ambient atmosphere rather than 100% oxygen. The agitated culture medium
consisted of 10 mM dimethylsuccinate (sodium), pH 4.5 containing 56 mM
-357-
-------
glucose, 1.2 mM ammonium tartrate, 0.4 itM veratryl alcohol, 0.6 mM benzyl
alcohol, 0.1% Tween 80, thiamine (1 mg/L) and a mineral supplement that was
7 fold the concentration of that used in the mat culture (27) . Cultures
were incubated on an incubator shaker (200 rpm) at 37°C.
PURIFICATION
Ligninase H-2 was purified from the extracellular fluid of agitated (1
L) or stationary (100 ml) cultures of P.. chrysosoorium by Fast Protein
liquid Qiromatography using procedures similar to those previously described
(28) . Briefly, purification procedures entailed the harvesting of the
extracellular fluid of ligninolytic cultures of P.. chrvsosporium followed by
concentration using a Millipore Minitan concentrator equipped with a filter
with a molecular exclusion limit of 10,000. The preparation was then
frozen, thawed and centrifuged (15,000 xg, 20 min) twice in order to remove
mucilaginous material that interfered with subsequent chromatography. The
preparation was then dialyzed overnight against 10 mM sodium acetate, pH
6.0, and the ligninases were purified by FPIC using a Mono Q HR 10/10 column
(Pharmacia, Uppsala, Sweden) . The sample was applied in 10 mM sodium
acetate, pH 6.0, and washed with 25 ml of the same buffer. The column was
eluted with a linear gradient from 10 mM sodium acetate, pH 6.0, to 0.7 M
sodium acetate, pH 6.0 (2 ml/min, 70 min) . Ligninase H-2 eluted at 0.18 M
sodium acetate. This peak was collected and stored at -20°C until used.
ENZYME ASSAYS
Ligninase activity (veratryl alcohol oxidase activity) was measured by
the procedure described by Tien and Kirk (12) . One unit of enzyme activity
oxidized one /zmol of veratryl alcohol in one minute at room temperature.
Oxidation of pyrene by ligninase H-2 was assayed as described _ for ligninase
H-8 (24) . Oxidation of crystal violet was assayed by monitoring the
decrease in absorption at 591 nm, its visible wavelength maximum. The
reaction mixture contained crystal violet (2 /M - 25 /M) , 50 mM sodium
succinate buffer, pH 4.5, and 20 [M or 400 juM H2O2. Metabolites of crystal
violet oxidation were separated by High Performance Liquid Chromatography
(HPLC) . Following incubation, 0.5 ml aliquots of the reaction mixture were
extracted with 0.5 ml of methylene chloride and 50 \il aliquots of the
methylene chloride extracts were used for HPLC analysis.
Reverse phase HPLC was performed using a system equipped with an Altex
model 110A pump (Anspec Co. , Inc. , Ann Arbor, MI) , a Rheodyne injector
(Rheodyne, Inc., Cotati, CA) , an R-Sil C-18 reverse phase column (4.6 x 250
mm) (Alltech Associates Inc. , Deerfield, IL) and a Schoef fel Model 770
detector (ABI Analytical, Kratos Div., Ramsey, NJ) . Isocratic elution was
performed using acetonitrile/water (80/20) containing 0.01 M 2-naphthalene-
sulfonic acid and 0.01 M trifluoroacetic acid as described (29) . Elution
was monitored at 560 nm. Crystal violet and its metabolites (N,N;N! ,N' ,N"-
pentamethylpararosaniline, N,N,N' ^''-tetxamethylpararosaniline and N,N' ,N"-
-358-
-------
trimethylpararosaniline) were identified by cranigration with authentic
standards.
PGP oxidation was assayed by monitoring the increase in absorption at
266.5 nm. The reaction mixture contained 20 mM sodium tartrate pH 3.0, 20
/^ H2O2, 20 iM PCP and 25 nM ligninase H-2. The reaction was initiated by
addition of H2O2. Metabolites of PCP oxidation were separated by HPLC. PCP
and PCP metabolites were extracted from aliquots (10 ml) of the reaction
mixture with hexane (1 ml). HPLC was performed using the Beckman System
Gold HPDC equipped with an Econosphere (Alltech Associates Inc., Deerfield,
IL) reversed phase (C-18) column (4.6 mm x 250 mm). Isocratic elution was
performed using acetonitrile:H20:glacial acetate acid (75:25:0.125).
Elution of PCP and its metabolite were monitored at 238 nm and 290 nm,
respectively.
RESULTS AND DISCUSSION
ENHANCED PRDDUCTION OF LIGNINASES
Ligninases have been shown to catalyze the oxidative depolymerization of
lignin (11) as well as the initial oxidation of certain xenobiotics (22,23).
The initial oxidation of many compounds is oftentimes the most difficult one
and, in some cases, may be rate limiting. Therefore, a logical approach to
increasing the biodegradative abilities of this fungus would be to increase
net production of ligninases. Although it is, conceptually, a straight
forward approach, in practice it is not necessarily easy to scale up
ligninase production from the 10 ml cultures used in many laboratories (19).
Ligninase activity appears to depend upon oxygen partial pressure (19). It
also depends upon culture and flask volume (19). Thus increased culture and
flask volume does not necessarily result in a proportionate increase in
ligninase activity (19). Increased production of ligninases by altering the
incubation conditions has been reported, however, by several research
groups. For example, Leisola et al. (30) reported that 1-2 mM veratryl
alcohol added to nitrogen limited stationary and carbon limited agitated
cultures of P.. chrvsosporium BKM-F-1767 resulted in veratryl alcohol oxidase
(ligninase) activities up to 180 and 670 units/liter, respectively.
Similarly, Jager et al. (28) reported that ligninase activities as high as
131-144 units/liter could be achieved in agitated cultures supplemented with
0.05-0.1% Tween 80, 0.4 mM veratryl alcohol and increased (7X) levels of
trace elements. The highest levels of ligninase activity reported to date
are those of Asther et al. (31) who used a different strain of P..
chrysosporium (INA-12). These investigators reported ligninase activities
of 1,344 units/L and 1,158 units/L in stationary and agitated cultures,
respectively, supplemented with oleic acid emulsified with Tween 80 (0.04%,
w/v). Most recently, Kirkpatrick and Palmer (32) developed conditions
whereby P.. chrvsosporium BKM-F-1767 ijnmobilized on polyurethane foam enabled
ligninases to be produced in high amounts (550-950 units/liter) on a semi-
continuous basis.
-359-
-------
We have also investigated ways to increase ligninase production. Using
culture conditions similar to those reported by Jager et aL. (28), we have
been successful in achieving ligninase concentrations as high as 446
units/liter in agitated cultures of P.. chrvsosporium BKM-F-1767. We also
showed that under these conditions, ligninase production was cyclic (Figure
1) with peak ligninase activity occurring at about 30 hour intervals
following the initial ligninase peak. These results are significant because
they show that in order to maximize ligninase production methods, ligninase
activity must be monitored at f requent intervals so that cultures may be
harvested at peak ligninase activity.
400
300
200
min/1
100 ••
100
150
Time (hours)
200
250
300
Figure 1. Idgninase activity in agitated cultures of P.. chrysosporium.
Cultures (1 L) were incubated in 2.8 L Fernbach flasks on a
rotary shaker (200 rpm) at 37°C as described in Methods and
Materials. After its onset, ligninase activity was monitored at
2 hr intervals.
We have also investigated ways to enhance ligninase production in
stationary 100 ml cultures. Our results were similar to those of leisola et
3l. (30) in that the addition of 1.5 iM veratryl alcohol resulted in
cultures containing a 9-10 fold increase of ligninase activity relative to
basal conditions.
OXIDATION OF XENOBIOZECS BY UGNINASES FROM P., chrysosporium
It is now known that P., chrvsosporium has remarkable biodegradative
abilities. In addition to its ability to degrade naturally occurring
bicpolymers, it is known that this fungus has the ability to degrade a wide
variety of chlorinated and nonchlorinated environmental pollutants as well
-360-
-------
as many other compounds. Table I is a summary of the chemicals reported to
be degraded by P.. chrysosporium.
Most biodegradation studies using P.. chrysosporium have been performed
using the organism in culture (1-8). These studies suggested that enzymes
involved in ligninase degradation were also involved in xenobiotic
degradation and recent evidence (23,24) has shown this hypothesis to be
correct. As noted above, ligninase H-8 has been shown to oxidize
dibenzo[p]dioxin, 2-chlorodibenzo[p]dioxin, benzo[a]pyrene and pyrene as
well as a number of other polycyclic aromatic hydrocarbons (PAH) (23,24).
An interesting facet of the results of Hammel et aL. (24) is the
observation that ligninase H-8 is able to oxidize PAH with ionization
potentials up to -7.55 eV whereas the well-studied horseradish peroxidase is
unable to oxidize PAH with ionization potentials greater than 7.35 eV. This
is the basis for the suggestion that the H2O2 oxidized intermediates of
lignin peroxidases (compound I and II) are more electropositive (i.e.,
better oxidants) than comparable intermediates of other peroxidases. It was
also shown that PAH compounds with ionization potentials higher than -7.55
eV were not oxidized by ligninase H-8. This is interesting because we have
shown (5) that phenanthrene, which has an ionization potential greater than
8.0 eV (33,34), is mineralized by P. chrysosporium, thus suggesting that
other enzymes, possibly other ligninases, may exist in P.. chrysosporium that
have even greater oxidative capacities than the well-studied ligninase H-8.
In the present study, we purified ligninase H-2 from nitrogen-limited
cultures of P. chrysosporium and partially characterized the isozyme with
special regard to its ability to oxidize several xenobiotics. The molecular
weight of this hemeprotein was estimated to be approximately 41,000. H2O2
was required as an oxidizing co-substrate. The specific activity of the
enzyme was variable from preparation to preparation. An average specific
activity of 39.9 + 16.1 (N = 6) vnits/imin/wg was observed.
In addition to veratryl alcohol, ligninase H-2 oxidized
pentachlorophenol, crystal violet, and pyrene (Figures 2, 3, 5). These
chemicals were selected as "model" compounds for several reasons. For
example, pentachlorophenol (PCP) has been widely used as a wood
preservative, is an acknowledged toxic environmental pollutant, and is
relatively resistant to microbial degradation (35). Although crystal violet
is not considered to be as great an environmental problem as PCP or PAH, it
is used as a biological stain, as a topical treatment in human and
veterinary medicine, and is used as a dye (36). It is not considered to be
carcinogenic. However, crystal violet has been shown to be a potent
clastogen (37,38) and it is toxic to many bacteria, a fact that makes its
biodegradation difficult in the environment and in waste treatment systems
(39). Pyrene was chosen as a representative PAH. Although pyrene itself is
non-mutagenic, many PAH are mutagenic, carcinogenic and environmentally
persistent (40). Pyrene is found in creosote (41), a PAH mixture derived
from coal tar that has seen extensive use as a wood preservative. Pyrene
has also been shown to be oxidized by ligninase H-8 (24).
-361-
-------
' TABLE I. ORGANIC COMPOUNDS DEGRADED BY P.. CHRYSOSFORIUM
Polycyclic Aromatic osrpounds
Benzo [a]pyrene
Biphenyl
2-Mathylnaphthalene
Phenanthrene
Bsnzo[a] anthracene
Pyrene
Anthracene
Perylene
Dibenzo[p]dioxin
Chlorinated Alkyhalides
Lindane
Chlordane
Biopolymers
Lignin
Cellulose
Kraft Lignin
3-chloroaniline-lignin conjugate
3,4-dichloroaniline-lignin conjugate
TriphenylJtethane Dyes
Crystal violet
Pararosaniline
Cresol red
Bromphenol blue
Ethyl violet
Malachite green
Brilliant green
Chlorinated Aromatic Compounds
4-Chlorobenzoic Acid
Dichlorobenzoic Acid
2,4,6-Jrrichlorobenzoic Acid
4,5-Dichloroguaiacol
6-Chlorovanillin
4,5,6-^Erichloroguaiacol
Tetrachloroguaiacol
Pentachlorcphenol
3-Chloroaniline
3,4-Dichloroaniline
2,4/5-JIrichlorophenoj^acetic acid
Polycyclic Chlorinated Aromatic Compounds
DDT (l,l,l-tric3iloro-2,2-bis(4--chloro-
phenyl)ethane
2,3,7,8-Ttetrachlorodibenzo-p-dioxin
3,4,3',4'-Tetrachlorobiphenyl
2,4,5,2',4',5'-Hexachlorobiphenyl
Aroclor 1254
Aroclor 1242
2-Chlorodibenzo [p] dioxin
Dicof ol (2,2,2-trichloro-l,1-bis(4-chlorophenyl)ethanol
•Die above is a partial list of compounds degraded by P.. chrysosporium
conpiled from ref. 1, 3, 4, 5, 7, 8, 23, 24, 44 and from unpublished
observations in our laboratory.
-362-
-------
0
4 8
TIME (MIN)
12
Figure 2. Reversed phase HPLC of the products of PCP oxidation by ligninase
H-2. Conditions were as described under Methods and Materials.
Product (2,3,5,6-tetrachloro-2,5-cyclohexadiene-l,4-dione(TCD))
was identified by comigration with an authentic standard.
Figure 2 shows that PCP was oxidized by ligninase H-2. HPDC analysis
of enzyme reaction mixtures showed that PCP underwent oxidation and
dechlorination at C-4 to form 2,3,5,6-tetrachloro-2/5-cyclohexadiene-l,4--
dione. This is important because it demonstrates that at least one enzyme
from P.. chrvsosporium has the ability to catalyze dechlorination of
chlorinated organic compounds. It has been known for some time that certain
bacteria and fungi degrade PCP and that the pathway involves dechlorination
and hydroxylation either ortho or para to the hydroxyl group to form a
catechol or quinone (or hydroquinone) (42). However, until now, no
bacterial or fungal enzyme has been shown to catalyze either of these
reactions (42).
Figure 3 shows the separation of crystal violet and its metabolites by
?PIC f°llowir|g incubation with ligninase H-2 and H2O2. Three products were
identified as N,N,N' ,N' ^"-pentamethylpararosaniline, N,N,N' ,N"-
tetramethylpararosaniline and N/Nl/N"-trimethylpararosaniline. These
results indicate that the initial step in the biodegradation of crystal
violet by P.. chrvsosporium is the N-dealkylation of crystal violet to form
N,N,N« ,N' ,N"-pentainethylpararosaniline which may then undergo subsequent N-
demethylation to form N,N,N' ^"-tetramethylpararosaniline which may also
serve as a substrate to form N,N« ,N"-triinethylpararosaniline. H2O2
irreversibly inactivated the enzyme during crystal violet oxidation.
Results (Figure 4) indicated that in addition to being a required oxidizing
substrate, H2O2 is also a competitive inhibitor. This interpretation is
-363-
-------
0
8 12 16
TIME (MINUTES)
Figure 3. Keversed phase HPIC of the products of crystal violet oxidation
by ligninase H-2. Peaks I, II, III and IV were identified as
crystal violet (N,N,N',N',N»,N»-hexaTnethylpararosaniline),
N,N,N' ,N« ,N"-pentamethylpararosaniline, N,N,N' ,N"-tetramethyl-
pararosaniline, and N,N« ,N"-tr3jnethylpararosaniline, t
respectively. Product identifications were made on the basis ot
(^emigration with authentic standards.
supported by the data in Figure 4 which show that increasing amountsof the
rSScing co^substrate (crystal violet) abolished the inhibitory effects of
high concentrations of H2O2.
Ohe PAH pyrene was also shown to be oxidized by ligninase H-2 (Figure
5). These results are similar to those of Hammel et al. (24) who, as noted
above, showed that pyrene is oxidized by ligninase H-8.
Ihese results, together with others in the literature, show that
liqninases purified from P. dhrvsosporium are able to catalyze an unpressive
array of biological oxidations of xenobiotics. Included among the reactions
catalyzed are: 1, the oxidation of PAH; 2, oxidative dechlorination of
chlorcphenols; 3, oxidation of dibenzo[p]dioxin and 2-chlorodibenzo[p]di-
oxin? and 4, N-deaUcylation of arylamines.
COHER OC*fSIDEE»TICNS
In addition to results obtained with pure enzyme, considerable indirect
evidence exists concerning the enzymatic capabilities of this microorganism.
For example, it is known that in cultures of P^ chrysosponum dicofol and
DDT both undergo reductive dechlorination of their chlorine bearing alkyl
carbon to form FW-152 and ODD, respectively (1,3). It is also known that
-364-
-------
DDT undergoes benzylic carbon hydroxylation to form dicofol (1,3) and that
FW-152 and dicofol undergo cleavage between C-l and C-2 to form DBF (3). it
is also apparent that P.. chrvsosporium must possess an enzyme system able to
catalyze aromatic ring cleavage of xenobiotic metabolites as evidenced by
the fact that many xenobiotics that are 14C-labeled in the aromatic ring (s)
are degraded to ^CO2 by this fungus (1-8). Hie mechanisms involved in the
above reactions and the extent to which ligninases catalyze these reactions
is the subject of ongoing research in this laboratory.
30 r
V
-1
-0.2
0.2 0.4 0.6 0.8 1.0
(CRYSTAL VIOLET)"1 CuM]'1
Figure 4.
Effect of H2O2 on the initial velocity of crystal violet
oxidation by ligninase H-2 as a function of crystal violet
concentration. Ihe reaction mixtures (1.0 ml) contained either
20 iM (Closed circles) or 400 juM (Closed squares) H2O2, 2.0 ua
ligninase H-2, and variable amounts of crystal violet in 50 mM
sodium succinate, pH 4.5. Reactions were performed at room
temperature and were initiated with H2O2.
-365-
-------
0.600
0.400
0.200
0.000
250
300
Wavelength
350
Figure 5. Oxidation of pyrene by ligninase H-2. The reaction mixture
contained 20 fM pyrene, 8.54 jug ligninase H-2, and 200 juM H2O2 in
H20/N,N-dimethylforinamide 8:2 containing 20 mM sodium tartrate,
pH 2.5. Olie reaction, which was initiated with H2O2, was scanned
between 240 nm and 380 nm prior to addition of H2O2, immediately
after addition and after 14 minutes.
ACKNOWIEDGEMENTS
This work was supported by Cooperative Agreement CR-814448 from the
Hazardous Waste Engineering Laboratory, Office of Research and Development,
U.S. Environmental Protection Agency, Cincinnati, Ohio, P.R. Sferra, project
officer. The authors thank Terri Maughan for her expert secretarial
assistance and Toni Patik for technical assistance.
REFERENCES
1. Bumpus, J.A., Tien, M.. Wright, D., and Aust, S.D. Oxidation of
persistent environmental pollutants by a white rot fungus. Science
228: 1434, 1985.
2. Bumpus, J.A. and Aust, S.D. Biodegradation of environmental pollutants
by the white rot fungus Phanerochaete chrysosporium; Involvement of
the lignin degrading system. BioEssays 6: 166, 1987.
-366-
-------
3. Bumpus, J.A. and Aust, S.D. Biodegradation of DDT (l,l>l-Tridtiloro-2,2-
Bis(4-Chlorophenyl) Ethane) by the white rot fungus Rhanerochaete
chrvsosporium. Appl. Environ. Microbiol. 59: 2001, 1987.
4. Bumpus, J.A. and Aust, S.D. Biodegradation of chlorinated organic
compounds by Phanerochaete chrvsosporiumi. a wood-rotting fungus. In:
Solving Hazardous Waste Problems: Learning From Dioxins (Jurgen H.
Exner, ed.). ACS Symposium Series 338, American Chemical Society,
Washington, D.C., 1987.
5. Bumpus, J.A. and Aust, S.D. Mineralization of recalcitrant environ-
mental pollutants by a white rot fungus. Proceedings of the National
Conference on Hazardous Wastes and Hazardous Materials. Hazardous
Materials Control Research Institute, Silver Spring, MD. Library of
Congress Catalog No. 87-80469, 1987. p. 146.
6. Sanglard, D., Leisola, M.S.A., and Fiechter, A. Role of extracellular
ligninases in biodegradation of benzo[a]pyrene by Rhanerochaete
chrvsosporium. Enzyme Microb. Technol. 8: 209, 1986.
7. Arjmand, M. and Sandermann> Jr., H. Mineralization of chloroani-
line/lignin conjugates and of free chloroanilines by the white rot
fungus Fhanerochaete chrvsosporium. J. Agric. Food Chem. 33: 1055,
1985.
8. Eaton, D.C. Mineralization of polychlorinated biphenyls by Rhanero-
chaete chrvsosporium; a ligninolytic fungus. Enzyme Microb. Technol.
7: 194, 1985.
9.
10.
11.
12.
13.
3effri.es, T.W., Choi, S., and Kirk, T.K. Nutritional regulation of
lignin degradation by Phanerochaete chrvsosporium. Appl. Environ.
Microbiol. 42: 290, 1981.
Leisola, M.S.A., Kozulic, B., Meussduerffer, F., and Fiechter, A.
Homology among multiple extracellular peroxidases from Phanerochaete
chrvsosporium. J. Biol. Chem. 262: 419, 1987.
Tien, M. and Kirk, T.K. lignin-degrading enzyme from the Hymenomycete
Phanerochaete chrvsosporium Burds. Science 221: 661, 1983.
Tien, M. and Kirk, T.K. Lignin-degrading enzyme from Rhanerochaete
chrvsosporium: purification, characterization and catalytic properties
of a unique H20-reguiring oxygenase. Proc. Natl. Acad. Sci. 81: 2280,
1984.
Kuwahara, M., Glenn, J.K., Morgan, M.A., and Gold, M.H. Separation and
characterization of two extracellular ^C^-dependent oxidases from
ligninolytic cultures of Rhanerochaete chrvsosporium. FEE Letters 169-
247,.1984.
-367-
-------
14.
15.
16,
17.
18.
19.
20.
21.
22.
Glenn, J.K., Akileswaran, L. , and Gold, M.H. Mi (II) oxidation is the
principal function of the extracellular Mn-peroxidase from
Phanerochaete (Arysosporium. Archives Biochem. Biophys. 251: 688,
1986.
Paszczynski, A., Huynh, V.-B., and Crawford, R.L. Conparison of lig-
ninase-I and peroxidase M-2 from the white rot fungus Phanerochaete
chrvsosporium. Arch. Biochem. Biophys. 244: 750, 1985.
leatham, G.F., Crawford, R.L., and Kirk, T.K. Degradation of phenolic
compounds and ring cleavage of catechol by Phanerochaete chrysosporium.
Appl. Environ. Microbiol. 46: 191, 1983.
Glenn, J.K., Morgan, M.A., Mayfield, M.B., Kuwahara, M., and Gold, M.H.
An extracellular H2O2-requiring enzyme preparation in lignm
biodegradation by the white rot basidiomycete Phanerochaete
chrysosporium. Biochem. Biophys. Res. Comm. 114: 1077, 1983.
Renganathan, V. and Gold, M.H. Spectral characterization of the
oxidized states of lignin peroxidase, an extracellular heme enzyme from
the white rot basidiomycete Hianerochaete chrvsosoorium. Biochem. 25:
1626, 1986.
Tien, M. Properties of ligninase from Pfranerochaete chrYsosporium and
their possible applications. CRC Critical Reviews in Microbiology 15:
141, 1987.
Andrawis, A., Johnson, K.A., and Tien, M. Studies on compound 1^
formation of the lignin peroxidase from Phanerochaete chrysosporium.
J. Biol. Chem. 263: 1195, 1988.
Kirk, T.K., Groan, S., Tien, M., Murtagh, K.E., and Farrell, R.L.
Production of multiple ligninases by Phanerochaete chrysosporium;
Effect of selected growth conditions and use of a mutant strain.
Enzyme Microb. Technol. 8: 27, 1986.
Tien, M. and Tu, C.-P.D. Cloning and sequencing of a cENA from a
from Phanerochaete chrysosporium. Nature 326: 520, 1987.
23.
24.
Haemmerli, S.D., leisola, M.S.A., Sanglard, D., and Fiechter, A.
Oxidation of benzo[a]pyrene by extracellular ligninases _ of
Phanerochaete chrvsosporium; Veratryl alcohol and stability of
ligninase. J. Biol. Chem. 261: 6900, 1986.
Hammel, K.E. , Kalyanaraman, B. , and Kirk, T.K. Oxidation of polycyclic
aromatic hydrocarbons and dibenzo[p]dioxins by Phanerochaete
cihrysosporium ligninase. J. Biol. Chem. 261: 16948, 1986.
-368-
-------
27.
28.
25. Hammel, K.E., Tien, M., Kalyanaraman, B., and Kirk, T.K. Mechanism of
oxidative Ca-CQ cleavage of a lignin model dimer by Fhanerochaete
chrysosporium ligninase: Stoichiometry and involvement of free
radicals. J. Biol. Chem. 260: 8348, 1985.
26. Fenn, P. and Kirk, T.K. Ligninolytic system of Phanerochaete
chrysosporium; Inhibition by o-phthalate. Arch. Microbiol. 123: 307,
1979.
Kirk, T.K., Schultz, E., Connors, W.J., lorenz, L.F., and Zeikus, J.G.
Influence of culture parameters on lignin metabolism by Phanerochaete
chrvsosporium. Arch. Microbiol. 117: 277, 1978.
Jager, A., Groan, S., and Kirk. T.K. Production of ligninases and
degradation of lignin in agitated submerged cultures of Phanerochaete
chrysosporium. Appl. Environ. Microbiol. 50: 1274, 1985"! "
29. Abidi, S.L. High-performance liquid ctaomatography of quinoidal
unnunium compounds derived from triphenylmethanes. J. Chromatog. 255:
101, 1983.
30. leisola, M.S.A., Thanei-Wyss, U., and Fiechter, A. Strategies for
production of high ligninase activities by P.. chrvsosporium. j
Biotechnol. 3: 97, 1985.
31. Asther, K., Corrieu, G., Drapron, R., and Odier, E. Effect of Tween 80
and oleic acid on ligninase production by Phanerochaete chrvsosporium
INA-12. Enzyme Microb. Technol. 9: 245.
32. Kirkoatrick, N. and Palmer, J.M. Semi-continuous ligninase production
vising foam-immobilized PhanerochaetQ chrvsosporium.. Appl. Microbiol
Biotechnol. 27: 129, 1987.
33.
34
Pysh, E.S. and Yang, N.C. Polarographic oxidation potentials of
aromatic compounds. J. Am. Chem. Soc. 85: 2124, 1963.
Cavalieri, E.L., Rogan, E.G., Roth, R.W., Saugier, R.K., and Hakam, A.
The relationship between ionization potential and horseradish
peroxidase/hydrogen peroxide catalyzed binding of aromatic hydrocarbons
to DNA. Chem. -Biol. Interactions 47: 87, 1983.
35. Rao, K.R. (Ed.) Pentachlorophenol : Chemistry, Pharmacology and
Environmental Toxicology, Plenum Press, New York.
36. The Merck Index (1983) Tenth Edition, Merck and Co., Inc., Rahway, NJ.
37. AU, W., Butler, M.A., Bloom, S.E., and Matmey , T.S. Further study of
the genetic toxicology of gentian violet. Mutat. Res. 66: 103, 1979.
38. Au, W., Pathak, S., Collie, C.J., and Hsu, T.C. Cytogenic toxicity of
gentian violet and crystal violet on mammalian cells in vitro.
Mutation Res. 58: 269, 1978.
-369-
-------
39. Michaels, G.B. and Lewis, D.L. Mcrobial
and triphenylmethane dyes. Environ. Toxicol. Qiem. 5.
aZ°
40. Cemiglia, C.E. Aromatic hydrocarbons:
and algae. Reviews in Biochemical Toxicology 3: 321,
41 lAjinsky, W., Dormsky, I., Mason, G., Ramahi, H.Y., and Safari, T. The
cSSStographic determination of trace amounts of V°ly™^
SdrSarbons in petrolatum, mineral oil and coal tar. Anal. Qiem. 35.
952, 1963.
42 Rochkind, M.L., Blackburn, J.W., and Sayler, G.S. Microbial decom-
SdSon of chlorinated aromatic compounds. EPA/600/2-86/090, U.S.
Environmental Protection Agency, Cincinnati, CH, 1986.
K. and Kirk, T.K. De ngvo synthesis and decomposition of
c^oTby a lignin-degrading basidiomycete. Photochemistry
17: 1676, 1978.
44. Huynh, V.-B., Cbang, H.-m, Joyce T.W., and Kirk TK Dechlorination
of chioro-organics by a white rot fungus. T.A.P.P.I. J. 66. 9, 1985.
370-
-------
DETOXIFICATION OF CONTAMINATED SLUDGES USING COTVmTNEn
MICROBIOLOGICAL AND PHQTOLYTIC DEGRADATTVE APPROACHES
' by: Ralph J. Portier, Kuniko Fujisaki, and Charles B..Henry
Institute For Environmental Studies, Louisiana State University
Baton Rouge, La. 70803
ABSTRACT
Dioxins have long been recognized as major carcinogenic substances associated with
creosote and pentachlorophenol contaminated sludges and soils. These hazardous waste materials
have posed a significant threat not only to groundwater supplies in the southern United States but
also have been recognized as a major potential global source of dioxin contamination worldwide
via transport mechanisms such as volatization and paniculate transport. Research will be presented
on the feasibility of biologically degrading pentachlorophenol and creosote sludge materials using a
contact reactor method incorporating the use of acclimated biomass in a highly aerated stirred-tank
reactor system. To date, data indicate that significant biotransforniation of these listed materials is
possible at a loading rate of 20% solids. Significant concern remains with regard to the residual
dioxin components associated with this reactor technology. Preliminary laboratory data will also be
presented which indicate the possibility of photolytic decomposition occurring as a primary
degradative pathway for OCDD and related dioxin contaminants associated with pentachlorophenol
biotransformation in these reactors.A secondary pathway involves the use of selected bacterial and
lungal strains which appear to further facilitate mineralization. The fate of OCDF and related furan
contaminants also found in these wastes will also be discussed.
-371-
-------
INTRODUCTION
Dioxins have long been recognized as major carcinogenic substances associated with
creosote and pentachlorophenol contaminated sludges and soils. These hazardous waste materials
have posed a sfenificaot toeat not only to groundwater supplies in the southern United States but
a\so have been recognized as a major potential global source of dioxin contamination worldwide
via transport mechanisms such as volatization and particulate transport. The Institute for
Envkonmental Studies has for the past four years been involved in extensive research on the
feasibility of biologically degrading pentachlorophenol and creosote sludge materials using a
SSSSSSSJwd mcor^rathfgAe use of acclimated biomass in *.^™*
reactor system. These sludge materials, referred to as K001 materials, is listed i
Appendix VTEL
Preliminary data from a biological closure of an RCRA site in Tennessee indicated the
possibility of photolytic decomposition occurring as a primary degradative pathway associated with
Sntachlorophenol biotransformation. The objective of our investigation is to evaluate the
contribution of photolytic decomposition of dioxins in liquid/ solid contact reactors as a
consequence of irradiation, % solids, transport and acclimated biomass introduction Data^wm be
presented on roughing cell experiments wherein the soils/sludges are suspended and introduced to
acclimated biomass.
MATERIALS AND METHODS
SAMPLING PROCEDURES.
Samples for analysis, both water and sludge residuals, were properly identified, labelled
and recorded in the lab notebook immediately after collection. All samples were collected,
preserved, and transported following the provisions of EPA/SW -846.
LIQUID/SOLIDS CONTACT REACTORS.
Chlorinated dioxins and furans are known to exist in many sludges, contaminated soils,
and contaminated slurries of materials having significant hydrocarbon content. Of particular
wncVrn toState and Federal agencies are the dioxins and furans found in chloroahphaticwastes
such alcreosote waste materials, particularly those materials containing high concentrations of
Pentachlorophenol (PCP) (14). Liquid /solids contact (LSC) reactors, as shown in Figure 1,
w?rf coScted aid placed in an environmentally closed laminar flow hood, and rnoculated, with
acclimated microbial populations developed in earlier pentachlorophenol studies (12). Specific
-372-
-------
processes, and post-treatment residual eermnaons o on-
PHOTOLYSIS OF DIOXIN ISOMERS.
28or T 3°°-315nm> was "PPfcd to laboratory reactors for 8 h/day at
: ™«ACHLOROraENOL- CONTAMINATED
LSC CREOSOTE STUDIES STEP #1 ROUGHING CELL REACTOR TESTS
-373-
-------
.predominantly K001 constituents, and also provided indications of fate of OCDD congeners
associated with these wastes. After seven days of high energy contact, the supernatant was
transferred to a polishing biological reactor cell where additional biological treatment was again
performed for a seven day period(see Figure 1). Over these time frames, GC /MS determinations
were made of the primary K001 constituents as well as the chlorinated dioxin and chlorinated furan
contaminants.
CORRELATIVE MICROBIOLOGICAL APPROACHES.
To identify microbial contributions to dioxin degradation, two approaches were considered.
Sterile L/S contact tests were conducted using the aforementioned laboratory approach.
Antibiotics were used to hinder microbial growth and kinetic response. Comparisons were made
between abiotic and biotic tests for OCDD removal.
MICROBIAL DENSITY
The Standard Plate Count Method (SPC) is a direct quantitative measurement of the
viable aerobic and facultative anaerobic microflora. Four general groups of microorganisms,
i.e., bacteria, actinomycetes, yeasts and filamentous fungi were enumerated using colony
forming units (CPU) and SPC. Replicate 1.0-ml aliquots were inoculated on Jensen's agar
medium (2) supplemented with 40 |J,g/ml with cycloheximide (Sigma) , to inhibit growth of
filamentous fungi. Filamentous fungi and yeasts were enumerated on Martin's agar medium (3),
with 30 Jig/ml streptomycin (Sigma) and 30(J.l/ml chlortetracycline (Sigma) to retard bacterial
growth. All plates were then incubated for four days at 30°C. Bacteria and actinomycetes were
counted on a Biotran El Automatic Colony Counter (New Brunswick Scientific). Yeasts and
fungi were enumerated manually.
ADENOSINE 5' TRIPHOSPHATE (ATP)
A modification of the adenosine 5' triphosphate (ATP) assay as advanced by Holm-
Hansen and Booth (4), and further presented by Van de Werf and Verstraete (5) and Karl (6),
was used for determination of microbial biomass. 1.0 g aliquots from all mesocosms (wet
weight) were transferred into dilution bottles containing 99 ml sterile distilled and deionized
water and homogenized on a homogenizer (Janke &Kunkel Ultra Turrax SD 45) for about 45 s.
A 100 jil aliquot of that suspension was then transferred to a 3 ml plastic vial. The vial was
inserted into aLumac 3M Biocounter (Biocounter M2010, Lumac Systems, USA) and 100 |Jl of
buffer and 100 |il NRB reagent (Nucleotide releasing agent for bacterial cells) was added. ATP
was released from microbial cells by adding NRB to a 100 |jl sample. After application of this
reagent, ATP was measured using the following reactions:
luciferin + luciferase +ATP -Mg++--> (luciferin* luciferase * AMP) + ppj
(luciferin*luciferase *AMP)-O2™> decarboxyluciferin + luciferase +CO2+ AMP + light
Following the injection of 100 (il of a luciferin-luciferase solution (Lumit, Lumac
systems) into the vial, light outputs expressed as relative light units (RLU) were determined over
a 10 s integration period. Relative light units were expressed as p,g ATP/g dry net weight using
-374-
-------
standardized 10 pi aliquots of a known ATP standard. Quench corrections for each
3SSS?SS!?EIef7ere^tablished with 100 m aHquots of ATP standard /buffer added in
place of the 100 |il buffer addition (internal standard method).
RECEIPT AND STORAGE OF SAMPLES.
"in hmfSSfeT® *^sfTed f™m *? source to to6 analytical laboratory accompanied by an
in house chain of custody form. Since the samples never leave the department, the form
Schedule work kthe laboratory. Samples are stored at 5°C
SAMPLE PREPARATION.
The samples, each approximately 20 ml subsampled into scintillation vials from the
experimental reactor vessel or other apparatus, were allowed to clarify into water and
w^^tiw6 fr^0118'1116 two Actions were analyzed separately. The following volume and
weight determinations were made: total volume, water fraction volume, wet weight
sediment/sludge, dry weight sediment/sludge.
SEDIMENT/SLUDGE EXTRACTION METHOD.
0«{t A ASfiAe^ater.wa? decan,ted md wet weight determinations made, the samples were
spiked with the following internal (surrogate) standards: 10 ^g d-10 acenaphthene^ 10 |ig d-12
perylene and 25 ug hexamethylbenzene. The samples were oven dried at 40°C for 3-6 hours At
this point the sample vial was again weighed for dry weight determination. Approximately 1 g of
sodium s^fate was added to ensure complete drying and to aid in extraction efficiency 15 mis of
fi^iV^^- ^e (DCM) was added and extracted with the aid of bath sonication for 1*2 min. The
final extraction volume of 15 ml was subsampled and fractionated on silica-gel columns.
The subsampled extract, typically 250 ^1, was reduced to 100 jil for fractionation. The
swerefractionated on activated silica-gel columns into two fractions: a slightly rx>lar
fraction eluted off the ctfumn with 3.5 mis DCM and a polar fraction eluted off SeSmn
with 3.5 mis methanol. The fractions are labeled F-2 and F-3 respectively. The fractions werTthen
reduced to a final volume of 100 ul and were ready for GC/MS analysis.
Water Fraction Extraction Method,
Thej water fraction was extracted by a slightly modified EPA method 625- only
acid/neutrals were extracted. In addition to the internal surrogate standards listed in the above
h"<^*^l^^*» ^^1 WTi-in n.J.JA~J A_ *J^ s+ .-* ..._ . -
add fraction' in
GAS CHROMATOGRAPHY / MASS SPECTROSCOPY ANALYSIS.
The following is the standard methodology developed for the GC/MS analysis of
polynuclear aromatic hydrocarbons (PAHs), heptachlorodibenzofurans (HpCDF),
-375-
-------
octachlorodibenzofurans (OCDF), heptachlorodibenzodioxins (HpCDD), and
octachlorodibenzodioxin (OCDD) in complex sample matrices: i.e., sludge/water samples which
have undergone partial microbial decomposition.(Note: High levels of microbial activity and
biomass can cause interferences not normally encountered with typical creosote sludge samples.)
Qualitative and quantitative gas chromatography/mass spectrometry (GC/MS) analysis
was performed on a Hewlett Packard 5890 GC equipped with a 12 meter Ultra-2 capillary column
(0.2 MM ID; 0.33|i film thickness) directly interfaced to a Hewlett Packard 5970B mass
spectrometer. The F-2 fraction was analyzed twice. First in the scanning mode and then by selected
ion monitoring (SIM) for detection and quantitation of the the PCDDs and PCDFs.
SCAN Method.
When operating in the scanning mode the oven was held at 50°C for 3 min then increased
to 260°C at a rate of 8°C/min and held for 8 min. The mass spectrometer was scanned between 45-
500 amu at a rate of approximately 1 scan/sec. The electron multiplier voltage (EMV) was set 200
volts above the tuning value. The inlet system was operated in the splifless mode with the purge oft
until 30 seconds into the analysis. The column flow is 0.68 ml/min at 100 C.
vSTM Method.
When operating in the SIM mode the GC was initially at 100°C for 3 min and increased to
260°C at 20°C/min. until 180°C where the rate is changed to 6°C/min until the final temperature of
260°C was reached then held for 10 min. Detailed information about the characteristic ions used
and their time windows is given in the Table 1.
-376-
-------
Table 1 : SIM Method Characteristic Ions.
WINDOW RT ION
1
2
3
S
Q
C
5.8 147
6.1 164
15.0 322
15.0 320
16.3 430
16.3 432
21.6 264
408
410
424
426
26.2 458
26.2 460
26.2 442
26.2 444
surrogate standard
quantitation ion
confirmation ion
COMPOUND
HMB
d-10 acenaphthalene
1,2,3,4-TCDD
1,2,3,4-TCDD
OCBP
OCBP
d-12 perylene
HpCDF
HpCDF
HpCDD
HeptCDD
OCDD
OCDD
OCDF
OCDF
ADDITION AT . TNFOT?MATTnM
M-CH3
M
M+2
M
M+4
M+6
M
M+2
M+4
M+2
M+4
M+2
M+4
M+2
M+4
S
S
0
^c
c
o
^t
c
S
c
Q
C
Q
c
o
^c
c
Q
Analytes of interest were quantitated by an internal standard method calibrated by a 5-point
external standard curve. Daily, or during each 8 hr analysis period, a single concentration of the
mixed standard was analyzed to verify the validity of the calibration curve. If the daily standard
differed by greater than 20 %, a new standard curve was made ^adra
RESULTS AND DISCUSSION
Biotransformation Of Pentachlorophenol-Contaminated Surface Water.
As shown in Figure 2, pentachlorophenol contaminated waters were biologically reacted
over a time frame of 21 days. Control and three experimental contact reactors were monitored for
JSfSS Pentachlorophenol concentration expressed in parts per billion (ppb). As determined by
GC/MS analyses, reactors had initial concentrations of 3,000-4,000 ppb PCP. Experimental
reactor #1 had an initial concentration of .5,222 ppb. After three days of biological contact, all
experimental reactors, #1, #2 and #3, indicated significant reductions in total PCP. After seven
-377-
-------
days of continuous contact, concentrations of PCP approached < lOppb. At the conclusion of the
study at 21 days, residual concentrations were again < lOppb. However, the control reactor
showed some fluctuations in PCP over time but remained rather consistent in terms of minimal
reduction and total concentration.
Figure 2 shows the same experiment in which liquid chromatography (HPLC) analyses
were performed. Again, concentrations were approximately the same at time. Experimental
reactors were significantly reduced in concentrations over the first seven days with minimal
residual levels noted for days 14 and 21. Control reactors showed minimal, if any, losses in PCP
concentration. PCP residuals noted at day 21 reflected concentration variables due to evaporative
loss of water from the reactors, i.e., a concentrating factor.
GC/MS SIM determinations of OCDD residuals in control and experimental contact
reactors for the pentachlorophenol contaminated water are shown in Figure 3. The initial
concentrations of OCDD ranged from 3.4 ppb to 13.9ppb for all LSC reactors tested. After three
days of biological contact, both control and experimental bioreactors showed significant reductions
in OCDD concentrations. The most significant removal rate of OCDD occurred in experimental
reactors where significant biological activity was taking place. The photolytic control reactors,
however, also showed significant reductions in OCDD concentrations. This reflects perhaps the
contribution of ultra violet radiation exposing suspended water droplets in the reactor during high
energy mixing. Abiotic control reactors without UV showed no loss in total OCDD concentrations.
Subsequent analysis of volatiles traps of all experimental and control reactors show no
volatilization and no volatile loss of OCDD or related dioxin congeners from the reactor within
limits of detection. Again, GC/MS detection limits for OCDD are LOppb. Subsequent follow-up
studies with higher concentrations of pentachlorophenol have also shown the similar relationships
for OCDD removal and abiotic photolytic decomposition experiments where only ultra violet light
was illuminating the pentachlorophenol contaminated wash waters. A more complete presentation
of these data will be presented in a subsequent manuscript. An evaluation of filtered biomass from
these reactors did not show detectable OCDD residuals.
giotransformation Of Creosote Waste KOQ1 Constituents.
Figures 4 and 5 provide information on the residual levels of key K001 constituents for all
reactors for roughing cell phase one biological treatment. The roughing step involves the actual
resuspension and solubilization of creosote and pentachlorophenol materials over a seven day
period.
This key initial step forces the solubilization of the K001 constituents as a result of the
addition of surfactant (Triton X100, Sigma) and pH adjustment (to 7.3) resulting in the
increased availability of these materials for biological attack.
The data presented show high concentrations of fluorene, phenanthrene and fluoranthrene
for initial waste loading. Reactor cell #1 had the highest levels of these K001 constituents in
concentrations exceeding 7,000 ppm. Reactors cells #2 - #4 had concentrations approaching 8,000
ppm or less with the exception of phenanthrene, averaging 15,000 ppm. It is important to note that
the initial concentrations varied in terms of chemical content, however, they all represented a 20%
loading rate based on solids for all reactors. With combined microbial addition and surfactant
-378-
-------
/* ft6 r£SidUal leVel/°r fl"°rene' Phenanthrene, fluoranthrene and pyrene were greatly
reduced after seven days of continuous aeration and agitation. .&™*y
to
to
rTCef trati°nS fro™. this ^ughing cell step, which would normally be transferred
reactor step , averaged in concentration between 500 and 4,000 ppm with
- ^P6/™2 to be ** most resistant to the continuous agitation over a seven day
treatment Period. In reactor test #3, minimal microbial levels were noted as determined by direct
8§01 ™* ^d^0^ AJP^tmiates. As a consequence of this, minimal removal levels for all
K001 constituents were noted. In particular, pyrene and crysene resulted in negligible
biodegradation Phenanthrene was MarginaUy reduced from 13,000 ppm to approximately 10,000
ppm Fluoranthrene and fluorene appeared to be the most significantly reduced of the K001
constituents .
j«* * i?elCt°r tl5having the LhiShest K001 constituents-loading rates, had the greatest reduction
m total hydrocarbon content. In particular, pyrene and crysene were more dnmaicaUynSST
reactor cell #1 as compared to reactor ceU #3. High biomass levels were noted in reactor cell #1
Microbial ATP levels exceeded 109 cells per ml for continuous treatment.
Discussion.
Optimal mixing of creosote waste materials can result in significant reductions in
nv ? eaS°r CeU #3 h^d minimal P^fonnance in K001 reductions. mSdate
poor microbial performance and mixing. Thus, a key component in the ability to biologically
transform these materials rests with the ability to sufficiently mix and suspend by wet weight the
rnSSoftrS-mf1 qi?S-T * su£se
-------
However, the OCDD concentrations were removed at much higher levels. These reductions
ranged from 25% to 45% for a 7 day biotreatment period.
Similar phenomena were noted for the chlorinated dibenzofurans (see Figures 10 and 11)
HpCDF concentrations varied in overall treatment efficiency. In reactor cell #3, the HpCDF
concentrations increased from 2.2 ppm to over 3.4 ppm. OCDF concentrations, however, varied
depending upon reactor initial concentration and the contact period. The greatest improvement in
OCDF removal occurred in reactors #2 and #4 with trace quantities detected in both instances.
CONCLUSIONS
The focus of these preliminary investigations has been to evaluate the consequences of
mixing and microbiological addition to creosote/ pentachlorophenol waste materials during the
initial steps of liquid/solid contact technology. Data indicate that changes are, indeed, occurring in
terms of the total quantity of all primary K001 constituents involved. However, the five ring
polynuclear aromatics are marginally reduced. Mixing performance plays an important role in
terms of the total residual materials over a seven day mixing period.
Figure 12 provides a composite summary of this initial treatment step for liquid/solids
contact reactor processing. For the chlorinated dioxins and chlorinated dibenzo furans. HpCDD
and HpCDF concentrations were not as significantly reduced in the initial seven day treatment
period. However, OCDD and OCDF concentrations were dramatically reduced. At this time, it is
difficult to determine the primary contributor to the OCDD and OCDF removal rates. High
biomass concentrations were realized in reactor cells #1, #2 and #4. However, reactor cell #3 had
only marginal biomass growth. Nevertheless, the OCDD levels of removal in reactor cell #3 were
comparable to tiie other three reactors in terms of total reduction rates. Abiotic control reactors
showed essentially no reduction in K001 constituents.
The investigations are incomplete at this writing with regard to the actual contributions of
ultra violet light in the reactor process. Pure culture studies have been completed where ultra violet
light was the primary mechanism for degradation of the prime chlorinated dioxin/furan
constituents. Experiments on the contributions of the incident ultra violet light in polishing reactor
cells (Step 2) wUl be completed shortly.
Nevertheless, it can be stated that significant biological activity is achievable for
creosote/pentachlorophenol waste in the reactor configurations as outlined. Biomass
concentrations reflected incomplete mixing and also reflected high oil and grease loading rates.
Analysis of volatiles exported from the reactor cells by mixing and air sparging did not contain
dioxin or furan congeners. Comparisons of biotic reactor cells to abiotic reactor cells showed
reductions in volume and specific toxicants, notably PNAs, occurred only in biotic reactors.
As reported by Crosby, 1983, the reduction of TCDD, a dioxin congener associated with
Agent Orange herbicide formulations, is rapid and complete within a matter of hours for UV light
All dioxins absorb light above 295 nm, so sunlight can serve as a light source. However, three
criteria must be met for dioxin photolysis to be practical: the wave lengths of incident light must
-380-
-------
correspond to appreciable absorption by TCDD/HpCDD/OCDD, the light must penetrate the
SfthS0 T ' ? ^oxin component, and a source of extractable hydrocarbons must be present
(although water is not very effective, formulating solvents and related hydrocarbon material?, i.e.,
K001 constituents would suffice) (Crosby and Wong, 1977). The roughing cell reactors provided
situations in which these criteria have been met. c«iou«& provurea
Finally, a recent paper by Pereira, 1986, suggests that in the presence of elevated
concentrations of dioxins and furans, particularly OCDD and HpCDD, microbial attack may be
indicated .under aerobic conditions. Specific variations in congener profiles over time indicated
that specific microorganisms may be participating in primary or secondary deeradative
mechanisms for dioxin removal in chlorinated hydrocarbon sludges. Data generated to date in our
investigations tend to support these observations.
ACKNOWLEDGEMENTS
The research presented in this manuscript was supported by funding from the Hazardous Waste
1 13888 ^ O^S^PA) U1Smna University' Baton RouSe> La- (Cooperative Agreement CR
LITERATURE CITED
1984 B°SSert' L' Kachel'W-M- and Bartha, R. Applied and Environmental Microbiology April
763-767.
[2] Jensen, H.Z.. Soil Science. Vol. 30, 1930, pp. 59-77
[3] Martin, J.P..Soil Science. Vol. 69, 1950, pp. 215-233.
510-5 193olm'"Hansen' °' and Booth' C-R- Limnology and Oceanography Vol. 11, 1966, pp.
Verstraete' W* in Proceedings, International Symposium
[5] Van de
Applications
Of Bioluminescencp. and Chemilurmnesr.p.nr.f. Schramet al. Eds., 1979., pp. 333-338.
[6] Karl, D.M., Microbiology Reviews. Vol. 44, 1980, pp. 739-796.
m^JPA/530-SW-84-015 Dec. 1984. Permit Guidance Manual on Hazardous Waste. 1
Demonstrations.
Register/ Vol. 49 No. 209. Oct. 19847 Rules and Regulations, Method 610,
-381-
-------
[9] Alexander, M. in Microbial Degradation of Pollutants in Marine Environments. A.W.
Bourquin and
P.H. Pritchard, Eds., EPA-600/9-79-012.
[10] Portier, RJ. and Meyers, S.P., Developments in Industrial Microbiology. Vol. 22,1982,
pp. 459-
478.
[11] Portier, R.J., Chen, H.M. and Meyers, S.P. Developments in Industrial Microbiology.
Vol. 24,
1983, pp. 409-424.
[12] Portier, RJ. and Meyers, S.P. in Toxicitv Screening Procedures Using Bagterial Systems,
Liu, D.
and Dutka, B., Eds., Marcel Dekker, Inc., New York, 1984, pp. 345-379.
[13] Porriety RJ. in Validation and Predictability of Laboratory Methods for Assessing the Fate
and Effects
of Contaminants in Aquatic Ecosystems. ASTM STP 865, T.P. Boyle,.Ed., American
Society for
Testing and Materials, Philadelphia, 1985, pp. 14-30.
[14] Crosby, D.G. and Wong, A.S. 1977. Environmental degradation of 2,3,7,8-
tetrachlorodibenzo-p-
dioxin (TCDD). Science 195:1337-1338.
[15] Pereira, W.E., Rostad, C.E., and Sisak, M.E. 1985. Geochemical investigations of
polychlorinated
dibenzo-p-dioxins in the subsurface environment at an abandoned wood-treatment facility.
Environ.
Toxicol. Chem. 4:629-639.
[16] Crosby, D.G. 1978. Conquering the monster: The photochemical destruction of
chlorinated
dioxins. ACS Symposium Ser. 73:1-12.
-382-
-------
PROBE ACCESS
SAMPLE PORT
AIR
RESIN TRAP
7
SOIL/WATER
SUSPENDED
SLURRY
1 .EXCAVATION
- Jj»— WATER
2.ROUGHING CELL x!
4.POLISHING CELL
•' i.
3. BIOLOGICAL CONTACT
RESIDUAL
SOLIDS
LSC BIOREACTOR
Igure 1 Appended drawing
Figure 1. General Design and Process Flow Diagram of LSC Reactor Approaches
-383-
-------
LSC Control
LSCEXP#1
LSCEXP#2
LSC EXP #3
GC/MS Dala
Time (Days)
Figure 2. Gas Chromatography Biotransformation Profiles Of Pentachlorophenol-Contaminated
Wastewaters
"to
d
Q
Q
8
15-C
10-
LSC Control
LSC EXP #1
LSC EXP #2
LSC EXP #3
1 4
2 1
TIME (Days)
Figure 3. Chlorinated Dioxin Residuals Of Biologically- Treated Pentachlorophenol Wastewaters
-384-
-------
£3
CO
1
•s
(Wdd) SlN3fUIJSNOO IOOM
O
O
I
(Wdd) SJ.N3nj.llSNOO 100>I
UJ
s
(Wdd)SlN3njUiSNOOlOO>l
SB
§
I
•1
(Wdd)SJ.N3niliSNOOlOOX
I
.§»
PL,
-385-
-------
(fldd) SOtlVlTOVUVaiOnNAIOd SNIU S
(Wdd) SOIJAfflOHV HVaiOflNATOd 9NIH S
• SO
I
ui
I
T-J-T'| « | I I'f 1'1-p-p1"!' f.-pr-T I I '"I"1'
ioQoogoog<
IOO OOOOO O
)U>*TC4Oa>U>XTC4
(Wdd) SOU.VHOUV bV JICMINAlOd ONIH 9
UJ
•s.
(Wdd) SOIiVWOHVUVaiOflNAlOd SNIU 9
I
B?
-386-
-------
UJ
S
a
I
(Wdd) SNIXOM Q31VNIU01HO
(Wdd) SNIXOIO Q3J.VNIU01HO
u
(Hdd) SNIXOIO Q3JLVNIH01HO
(Weld) SNIXOICIQ31VNIU01HO
-387-
-------
s
g
(Hdd) SNVUndOZN38ia
I
B?
a
to
P
S
o
i
I
1
^
P4
s
u
<+-<
o
ca
§
PH
tl
a
^
-388-
-------
STUDY #1: COMPOSITE
60 -
40 -
*>n-
n-
1
I
1
•%
i
H
__.
1
i.: A
?
1
*
*
;
s
;
R1
R2
R3
R4
• HEPTACDD
E3 OCDD
H HEPTACDD
0 OCDD
LJ HEPTACDD
Bi OCDD
m HEPTACDD
HI OCDD
I
5
0 7
TIME (DAYS)
STUDY #1: COMPOSITE
o 7
TIME (DAYS)
R1
R2
R3
R4
• HEPTA CDF
0 OCDF
11 HEPTA CDF
0 OCDF
n
m
m
HEPTA CDF
OCDF
HEPTA CDF
OCDF
Figure 12. Composite Of Chlorinated Furan/Dioxin Residuals In Roughing Cell Experiments
-389-
-------
TESTING OF TREATMENT SYSTEMS FOR CYANIDE BEARING HAZARDOUS WASTES
by: Douglas W. Grosse
HWERL, U.S. EPA
Cincinnati, OH 45268
Sardar Q. Hassan
Michael P. Vitello
Margaret K. Koczwara
Dept. of Civil & Environmental Engineering
University of Cincinnati
Cincinnati, OH 45221-49TE
ABSTRACT
As a part of the program to perform hazardous waste treatment technology
assessments, the U.S. EPA has been directed to evaluate technologies
applicable to the treatment of metal and cyanide bearing wastes. Bench- and
pilot-scale studies have been conducted on spent process electroplating wastes
(RCRA Code F007) at U.S. EPA's Test and Evaluation Facility. A series of
unit processes have been evaluated: alkaline chlorination, precipitation,
filtration, resin adsorption and activated carbon. Several different
treatment trains have been evaluated for the purpose of determining the
optimum treatment.
-390-
-------
INTRODUCTION
Alkaline chlorination is a treatment technology which has shown to
be successful in destroying free cyanides from aqueous metal/cyanide waste
resulting from cyanide electroplating operations. However, complexed
(non-amenable) cyanides are often left untreated and may pass into precipi-
tation processes which generate metal hydroxide/cyanide sludges or are
discharged as final effluents. It is desirable to destroy all the cyanide
rather than to have it mixed with metal hydroxide sludges which, in most
cases, are land disposed. Further treatment may be required to remove the
residual cyanides. The objective of this study is to evaluate the most
effective way to remove (destroy) hazardous components from aqueous metal/
cyanide waste streams using conventional alkaline chlorination in .conjunction
with additional polishing or removal steps. These additional technologies
are anionic exchange and precipitation of ferrocyanide complexes in the form
of ferri-ferrocyanide (Prussian- Blue). Metals treatment consisted of lime
precipitation and cationic exchange.
APPROACH
This study was conducted in conjunction with the University of
Cincinnati (UC) at the U.S. EPA's Test and Evaluation (T&E) Facility, in
Cincinnati, Ohio. Results from bench- and pilot-scale research were used to
evaluate the treatment of an electroplating cyanide spent process stream
(RCRA Code F007). Configurations (treatment trains) of unit processes were
evaluated based upon the ability to optimize treatment. Performance was
evaluated at selected design parameters, e.g., ORP, pH, dosage and retention
times. It is hoped that data generated by this study will benefit commercial
manufacturing operations utilizing on-site waste treatment systems, off-site
treatment storage and disposal facilities (TSDFs) and regulatory agencies
(federal, state and local) in understanding the parameters which influence
treatment of cyanide bearing electroplating wastes.
Initially, bench-scale studies were conducted for the purpose of
optimizing unit process parameters. Then, a conventional pilot system was
installed to test the operation of the various unit processes employed.
Lastly, treatment technology modifications were evaluated at bench-scale to
meet the following treatment objectives:
Highest quality final
constituents, and
effluent devoid of metal and/or cyanide
o Least amount of cyanide present in the metal hydroxide sludges.
Results are presented from each phase of study.
WASTE DESCRIPTION
In order to perform this work it was necessary to obtain the appropriate
permits (TSDF and RD&D permits) to legally import an electroplating process
waste (synthesized F007). This waste is defined as a "spent cyanide plating
bath solution from electroplating operations (except for precious metals
-391-
-------
electroplating spent cyanide plating bath solutions)." (1) The source of the
wastewater for the experiments described in this paper was a local Cincinnati
metal finisher who has been involved in joint EPA-industry demonstration
projects in the past. The metal finishing plant is a multi-process facility
which incorporates zinc, brass, chromium, tin, bright nickel and cadmium
plating. Twelve 55-gallon drums of synthesized F007 electroplating waste_
were acquired. Drag-out from plating operations was added to give specific
concentrations of cyanide and heavy metals. Results from the characterization
of this waste are provided in Table 1.
TABLE 1. CHARACTERIZATION DATA FOR THE RAW WASTE
Symbol
Concentrations
(mg/L)
Cyanide (total)
Cyanide (not amenable to chlorination)
Sulfide
Fluoride
Arsenic
Barium
Cadmi urn
Chromium (total)
Copper
Nickel
Sil ver
Zinc
Methyl ene chloride
Chloroform
Bi s- ( 2-ethyl hexyl ) phthal ate
TOC
TSS
CN(T)
CN(NA)
S
F
As
Ba
Cd
Cr
Cu
Ml
Ag
Zn
418-462
3-10
0.70-0.90
1.80-2.10
0.02-0.03
0.04
24.8-26.2
0.06-0.12
52.5-58.5
11.9-12.8
0.01
11.2-11.9
0.01
Of\ O
.02
0.02
31'5
44.7
The raw waste was analyzed for cyanide, inorganics
determine applicable treatment options.
and organics in order to
The rationale behind the waste selection was based upon obtaining a raw
waste that would challenge the efficiency of the conventional alkaline
chlorination/precipitation treatment process. A profile for an initial
wastestream was developed taking into consideration waste characteristics,
permit limitations and safety. In order to achieve our objectives, a
prescribed composition of both cyanide (amenable and non-amenable)
and heavy metals was required. It was determined that a synthesized
wastestream was preferable. Although the selected wastestream would be
atypical of those wastestreams generated by electroplating job-shops, it
may be more representative of mixed aqueous metal/cyanide wastestreams that
would be treated at off-site commercial treatment, storage and disposal
facilities (TSDF). Future studies will examine different waste matrices and
constituent concentrations.
-392-
-------
TREATMENT TECHNOLOGY EVALUATIONS
The assessment of bench- and pilot-scale treatment of the subject waste
was based upon (1) bench-scale process optimization studies; (2) monitoring
of significant process parameters during pilot-scale operation; and (3)
evaluation of treatment technology modifications. In total, five treatment
trains were evaluated as having potential for the treatment of the
characterized waste. All sampling and analyses were performed in accordance
to standard procedures as described in the EPA's Office of Solid Waste
Manual (2). Metal analysis was performed on an Inductively Coupled Plasma
(ICP) spectrometer while organics were analyzed on a gas chromatoqraph/mass
spectrometer (GC/MS). For bench-scale testing alkaline chlorination and lime
precipitation were performed in a batch mode while ion-exchange operations
were performed in a continuous mode by use of laboratory burettes All
sludges and solids were settled and filtered prior to ion exchange to avoid
fouling and clogging of the resins. Upon decanting of the supernatant, the
residual sludge slurry was filtered. The resulting filter cake was analyzed
for cyanide and metals concentrations.
OPTIMIZATION OF UNIT PROCESS PARAMETERS
As a result of the raw waste analysis (Table. 1) an initial conventional
treatment strategy was identified as having potential for treatment of the
characterized waste. This treatment train included two-stage alkaline
chlonnation for cyanide destruction, lime precipitation for metals removal,
sulfide precipitation, multi-media filtration, cation exchange and carbon
adsorption for organics removal. Figure 1 shows a schematic of this initial
treatment configuration.
CYflNIDE DESTRUCTION
METflLS REMOURL
1st Stage
Alkaline
Chlorination
2nd Stage
fllkallne
Chlorination
Lime
Precipitation
jjfa$mm%®>
METflLS POLISHING
Sulfide
Precipitation
Multi-Media
Filtration
Multi-Media
Filtration
Cation
Exchange
Rctiuated
Carbon
Figure 1. Treatment train A - conventional treatment train
cyanide and metal bearing waste.
for
-393-
-------
The bench unit process optimization studies determined optimum
conditions for two-stage alkaline chlorination (dose, retention time and pH).
Hrne and sulfide precipitation (dose jar tests at selected pH values) and
polyelectrolytic addition (dose, pH, turbidity and conductivity). Break-
through conditions were determined for filtration (TSS), ion-exchange
(metals) and carbon adsorption (TOC). Results of these studies are shown in
Table 2.
TABLE 2. OPTIMUM UNIT PROCESS PARAMETERS
First Stage Alkaline
Chlorination:
Second Stage Alkaline
Chlorination:
**I_ime Precipitation:
**Sulfide Precipitation
Flocculation:
Clarification:
*Cation Exchange for
Metals Polishing:
PH
ORP
Residence time
PH
ORP
Residence time
pH
Residence time
pH
Residence time
Dose
Residence time
Polyelectrolyte type
Polyelectrolyte dose
Residence time
Loading rate
Loading rate
Resin type
Resin capacity
11.5
300 mV
60 min.
8.5
750 mV
60 min.
10.5
15 min.
10.5
15 min.
333 mg/L
15 min.
cationic
20 mg/L
30 min. ?
0.104 gpm/f1T
1 gpm/ft3
Amber!ite IRC-718
1 meq/ml. wet resin
*Supplemented from literature and vendor information.
**The flocculation and settling for these precipitation processes is the same.
Upon analysis of the data gathered from the preliminary bench tests, it
was determined that sulfide precipitation and carbon adsorption offered little
improvement in the treatment of this wastestream. As a result, these two
unit processes were omitted from the preliminary pilot test.
-394-
-------
PILOT SCALE TESTING
Based upon.the results generated from bench-scale optimization of unit
processes, a pilot-scale treatment train (Figure 1) was prepared for
operation. The system incorporated four unit processes: alkaline
chlorlnation, lime precipitation (incorporating coagulation, flocculation and
settling), multi-media filtration and cationic exchange. The pilot system
was designed for a flow rate of 1 gpm, however, due to waste quantity limita-
tions, the flow rate in the alkaline chlorination process was reduced to 0.5
gpm. A complete pilot run was conducted in two days; with 2-stage alkaline
chlorination performed on the first day and metals treatment on the following
day.
Sampling was performed allowing for unit process hydraulic retention
times. For example, the effluent sample collected following two-stage
alkaline chlorination was taken two hours (the residence time for 2-stage
alkaline chlorination) after the corresponding influent sample. Detailed
QA/QC was adhered to which included the collection of field duplicates and
spikes. Data resulting from this run is presented in Table 3.
TABLE 3. CYANIDE & METALS ANALYSES FOR PILOT TESTING
Concentration (mg/L)
Sample
Source
Raw waste
Alk. chl.
Lime prec.
Cation exch.
CN
(T)
365
17
21
18.7
CN
(NA)
25.
11.
14.
17.
0
8
5
2
Cd
29.3
29.6
0.7
0.01
Cr(T)
0.04
0.04
0.06
0.07
Cu
93.7
66.0
2.1
0.02
Ni
14,6
14.5
0.8
0.2
V
0.01
0.02
ND
ND
In
13.1
12.7
0.3
ND
TOC
315
241
165
TSS
44.7
459*
31.7
* This increase in TSS can be
alkaline chlorination step.
CN(T) is total cyanide.
CN(NA) is cyanide not amenable
Cr(T) is total chromium.
attributed to solids generated during the
to chlorination.
Although metals concentrations were significantly reduced following
cation exchange (<.2 mg/L) total cyanide was still passing through the final
effluent in concentrations exceeding the local allowable discharge limit of
10 mg/L (at 18.7 mg/L). Upon further analyses of the treated waste, iron
(Fe) concentrations in the raw waste (13 mg/L) and final effluent (7 mg/L)
suggested that the 18.7 mg/L cyanide in the.final effluent was complexed with
the Fe in the form of ferrocyanide (Fe(CN)g ). An additional polishing step
would be required to remove the remaining cyanide complex. In order to better
understand the treatment of this waste, a series of alternative treatment
trains were investigated in bench studies.
-395-
-------
In observing the values of the cyanide concentrations (both total and
non-amenable), there is a noticeable degree of variability. This may be
attributed to matrix effects (interferences), variation of waste composition
from drum to drum (although the synthesized waste was pre-mixed in a single
process tank prior to separation into drums) and problems encountered with
the development of analytical methods. This variability is more prevalent
with the non-amenable analysis than with the total cyanide analysis.
This phenomenon was documented by the quality control (QC) and quality assurance
(QA) data reported for the pilot-scale testing, which is as follows:
Non-amenable
Precision (RPD):
Spike Recovery:
Standard Recovery:
Total Cyanide
Precision (RPD):
Spike Recovery:
Standard Recovery:
52.5% (j+45.3%)
68.0% (+20.0%)
71.0% (+17.7%)
29.6% (+23.3%)
80.0% (±22.6%)
75.6% (+11.5%)
EVALUATION OF TREATMENT TECHNOLOGY MODIFICATIONS
Based upon the results of the pilot run, four additional treatment
trains (B, C, D, & E) were developed to effect treatment technology
modifications in order to achieve lower cyanide concentrations in the final
effluent and the lowest amount of cyanide settling out in the hydroxide
sludge. These treatment trains included all or some of the following unit
processes: alkaline chlorination, lime precipitation, anion exchange for
complexed cyanide removal, cationic exchange for metals removal and
precipitation of ferrocyanide complexes in the formation of Prussian Blue.
These units were operated in a manner which simulates the pilot-scale
operating conditions listed in Table 2. For cvanide polishing Amberlite IRA-
van ide
loading rate of 1 gpm/ft . Capacity of the resin is
1.2 - 2.3 Ibs CN/ft^. Results of applying these modified treatment trains
to the synthesized F007 waste are presented below.
operating
958 was selected at
A schematic of Treatment Train B is shown in Figure 2. This treatment
train evaluates the effectiveness of cation exchange resin in removing
metals from the raw waste and segregating metals from sludges produced during
alkaline chlorination. Alkaline chlorination and anion exchange follow the
initial cation exchange metal removal step. Sludges produced during this
step are settled and the supernatant filtered before anion exchange.
Performance of this treatment train is shown in Table 4. Cation exchange in
Treatment Train B failed to remove all the metals (Cu - 63.2 mg/L; Ni - 11.1
mg/L) and the majority of these unadsorbed metals precipitated in the alkaline
chlorination sludge (as is illustrated in Table 4). Initially, it was
presumed that the cationic exchange resin was not contacting properly with the
waste. As a result, a separate stirred beaker test was performed maximizing
the contact of the waste with the cation resin. However, similar results were
obtained refuting our initial presumption. It is quite possible that Cu and
-396-
-------
MCTfUS BEMDURL
Figure 2. Treatment train B - metals removal by cation exchange
prior to alkaline chlorination.
Ni were complexed with the cyanide prior to alkaline chlorination, thus,
inhibiting ion exchange removal. The succeeding alkaline chlorination
step then broke the Cu and Ni cyanide complexes precipitating the two
metals in the following clarification step. The effluent from this clarifier
showed a substantial reduction in Cu (1.36 mg/L) and Ni (1.45 mg/L)
concentrations. Effluent from the anion exchange column contained 26 mg/L of
total cyanide.
" TABLE 4. CYANIDE AND METALS ANALYSES FOR TREATMENT TRAIN B
Sample
Source
Raw waste
Cation exchange
Alk. chl.
Clarifier
Anion exchange
CN
(T)
179
288
22.0
29.0
26.0
CN
(NA)
31.
30.
**
26.
Concentrations (mg/L)
Cd Cr(T) Cu Ni
5
5
0
25.2
0.58
0.89
0.28
0.04
0.01
0.02
0.02
58.5
63.2
61.1
1.36
11.4
11.1
11.3
1.45
V
0.00
0.00
0.00
0.00
Zn
10.3
0.17
0.22
0.03
Ag
0.04
0.05
0.06
0.04
Fe
5.49
2.99
4.92
6.48
**CN(NA) was determined to be greater than CN(T); actual number is
not available.
-397-
-------
Treatment Train C converts free cyanide into a ferrocyanide complex
(Fe(CN)g ) for subsequent removal by an anion exchange resin (IRA-958).
Then, ITme precipitation and cation exchange are employed for metal removal
and polishing, respectively. A schematic of this treatment train in shown
in Figure 3.
The objective of this treatment strategy was to completely remove the
cyanide prior to lime precipitation so that cyanide would be prevented from
accumulating in the hydroxide sludge. Conversion of free cyanide into the
ferrocyanide complex was achieved by dosing with ferrous sulfate (FeSO^) at
a pH of 8.5. Excess ferrous sulfate forms ferrous hydroxide precipitate
(Fe(OH)?) which was settled and filtered prior to anion exchange. Results
of this treatment train are presented in Table 5.
REMOVAL OF EXCESS Fe(OH)2
Figure 3. Treatment train C - initial removal of all cyanide in
the form of a ferrocyanide complex using anion exchange,
In Treatment Train C, again the anion exchange resin failed to remove
all the cyanide (25 mg/L CN in the effluent), part of which then precipitated
with the hydroxide sludge. The Fe(OH)2 sludge
sludge 48.1 mg Cd/kg wet sludge, 38.7 mg Cr/kg
sludge and 1800 mg Zn/kg wet sludge. With the
cyanide in the hydroxide sludge this treatment
in achieving desirable residuals separation of
contained 9310 mg CN/kg wet
wet sludge, 68.7 mg Ni/kg wet
accumulation of metals and
option proved to be deficient
cyanide and metals.
-398-
-------
TABLE 5. CYANIDE AND METALS ANALYSES FOR TREATMENT TRAIN C •
Concentrations (mg/L)
Sample
. Source
Raw waste
Anion exchange ;
Lime prec.
CN
(T)
350
25.0
12.0
CN
(NA)
33.0
**
**
Cd
29.5
1.41
1.49
Cr(T)
0.04
0.00
0.01
Cu
68.7
0.30
0.32
Ni
14.8
0.00
0.07
V
0.01
0.00
0.01
Zn
12.2
0.05
0.05
0.
0.
0.
Aq
07
04
06
Fe
13.3
0.50
0.65
Fe(OH)9 sludge 9310 **
mg/kg wet sludge
48.1 38.7 8.70 68.7 6.90
1800 4.40 38500
**CN(NA) was determined to be greater than CN(T); actual numbers are not
available.
CN to Fe(CN),
4-
PRECIPITflTION OF PRUSSIHN BLUE
—
FeS04
addition
METflLS REMOUflL
Lime
Precipitation
FeCI3
Precipitation
METflLS POLISHINI
Multi-Media
Filtration
Cation
EKchange
Figure 4.
Treatment train D - initial removal of cyanide in the form of
Prussian Blue.
Figure 4 shows a schematic of Treatment Train D. The purpose of this
option was to initially convert all free cyanide into the ferrocyanide com-
plex, with subsequent precipitation in the form of Prussian Blue,
(Fe4[Fe(CN)6]~). Ferrous sulfate and ferric chloride (FeCl~) are used in
the two reactions at pH values of 8.5 and 6.5, respectively; Lime precipita-
tion and cation exchange follow as metal removal processes. Table 6 exhibits
the results generated from running this option. In Treatment Train D, the
Prussian Blue precipitation step failed to remove all the cyanide (83 mg/L CN
in effluent). Furthermore, the sludge concentrations were (mg/kg wet sludge)
1690 for Cd, 42.5 for Cr, 53 for Cu, 106 for Ni and 567 for Zn in addition to
the 15,600 for CN and 62,600 for Fe.
-399-
-------
TABLE 6. CYANIDE AND METALS ANALYSES FOR TREATMENT TRAIN D
Sample
Source
Raw waste
Pr. Blue prec.
Lime prec.
CN
(T)
350
83.0
83.0
CN
(NA)
33.0
73.0
72.5
Concentrations
Cd
29.5
0.44
0.70
Cr(T)
0.
0.
0.
04
00
04
Cu
68.7
55.8
58.7
(mg/L)
Ni
14.8
13.6
11.6
V
0.01
0.01
0.05
Zn
12.2
0.15
0.17
Ag
0.07
0.04
0,09
Fe
13.3
1.93
1.49
(mg/kg wet sludge)
Pr. Blue sludge 15600 15500 1690 42.5 53.0 106 1.50 567 4.10 62600
In Treatment Train E, as shown in Figure 5, a conventional treatment
train for cyanide bearing metal plating waste (similar to Figure 1) has been
augmented with polishing for complex cyanide and metals using ion exchange
columns. Anion exchange is employed for the polishing of cyanide, (particu-
larly the stable ferrocyanide complexes), while cation exchange was used for
metal polishing. As Table 7 shows, this treatment train achieved significant
reduction of constituent metals and cyanide concentrations in the effluent,
with total cyanide and metal concentrations at or below 1 mg/L.
CVBNIDE DESTRUCTION
Figure 5. Treatment train E - cyanide polishing with anion exchange
resins after lime precipitation.
-400-
-------
TABLE 7. CYANIDE AND METALS ANALYSES FOR TREATMENT TRAIN
Sample
Source
Raw waste
Alk. chl.
Lime prec.
An ion exchange
Cation exchange
CN
(T)
513
36.0
30.0
0.10
0.15
CN
(NA)
30.0
21.5
23.5
0.00
0.00
Cd
25.3
24.6
0.48
0.07
0.05
Concentrations (mg/L)
Cr(T) Cu Ni V
0.06
0.06
0,05
0.04
0.05
55.8
57.8
1.08
0.14
*1.04
11.9
11.9
0.32
0.07
0.07
0.01
0.00
0.03
0.02
0.01
In
10. .4
10.6
0.24
0.-05
0.05
Aq
0.
0.
0.
0.
0.
02
06
05
04
02
Fe
3.72
4.62
8.69
0.09
0.03
(mg/kg wet sludge)
Lime prec. slud. 2530 878 8410 3.70 22400 4270 4.00 3610 12.7 275
* This increase is due to copper screening being used as an underdrain for
the bench-scale cation exchange columns.
DISCUSSION '
After analyzing the treatment data generated' from the four treatment
trains (B, C, D, & E) it was concluded that Treatment Train E provided the
best treatment for this particular waste. Although unsuccessful in completely
separating CN from the metals in the hydroxide sludge (2530 mg/kg wet sludge
CN ), Treatment Train E exhibited the lowest CN" concentration in the effluent
(0.15 mg/L). Metals concentrations in the cationic exchange effluent were
below sewage discharge limits (POTW).
Future work will confirm the results of these bench studies on treatment
technology modifications. Again, the pilot treatment system will be modified
to run on -the optimum treatment train (E). Results of these studies will be
presented as soon as they become available. Other studies are also in
progress to determine how the cyanide-bearing sludge formed by Treatment
Train E can be further treated to reduce its CN~ content prior to land disposal
1.
2.
3.
CONCLUSIONS
Although alkaline chlorination was unable to lower the total cyanide
below 17 mg/L in the synthetic F007 waste studied, anionic exchange
polishing reduced the total cyanide concentration to less than 1 mg/L.
Lime precipitation were able to reduce metal concentrations in the waste
to local discharge limits, but with polishing (multi-media filtration,
anion and cation exchange) metals concentrations were lowered to <0.1 mq/L
for each metal except copper.
A complete separation of cyanide and metals in residual sludges was not
achieved by using the unit processes studied in this research.
-401-
-------
REFERENCES
1. Code of Federal Regulations 40, Part 261, revised as of July 1987.
2 U.S. EPA, Office of Solid Waste, "Test Methods for Evaluating Solid
Wastes" SW-846, 3rd Edition, November 1986.
-402-
-------
FIELD EXPERIENCE WITH THE KPEG REAGENT
by: Alfred Kornel
Charles J. Rogers
Harold Sparks
Hazardous Waste Engineering Research Laboratory
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Chemical reagents prepared from polyethylene glycols and potassium
hydroxide (KPEGs) have been demonstrated under mild conditions (25° - 14u <;,
to dehalogenate PCBs, PCDDs, and PCDFs with laboratory destruction efficiencies
exceeding 99.9999%. The reaction mechanism is nucleophilic substitution at an
aromatic carbon.
140°C)
Bench scale studies have already established conditions for PCB destruc-
tion to less than 1 ppm and for PCDDs and PCDFs to less than 1 ppb. Toxico-
logical tests have established that arylpolyglycol by-products from KPEG
reactions are non-toxic. The non-toxic property of the by-products may allow
fqL 9?nng *?? °n~site disPosal of treated materials. In July and August
1986, a 2700 gallon KPEG reactor was used in Butte, Montana, on a wood pre-
serving site and in Kent, Washington, on a waste disposal site to successfully
detoxify PCDDs and PCDFs (120 ppb - 200 ppm) in 17,000 gallons of liquid waste
to non-detectable levels. A reactor designed to treat both liquids and solids
has been tested on selected Superfund and Department of Defense sites. These
field studies have validated conditions for destruction of PCBs PCDDs and
PCDFs to acceptable levels required by the regulations. This presentation will
review treatment data, regulations for treated materials, costs, and the
potentials of KPEG for the destruction of a variety of halogenated pollutants.
-403-
-------
INTRODUCTION
Chlorinated dibenzo-p-dioxins (PCDDs), polychlorinated biphenyls (PCBs),
and chlorinated dibenzofurans (PCDFs) are three series of related compounds
that gained notoriety for their high toxicity and persistence in the environ-
ment. In the last few years PCDDs and PCDFs have increasingly been identified
in chemical product waste streams as well as in effluents from incineration
processes.
Most of our knowledge of PCDDs and PCBs as environmental contaminants is
associated with their occurrence in soils, sediments, combustion particulates,
and in fish. In most cases, this contamination stems from improper waste
disposal of highly toxic PCDDs in such products as hexachlorophene or 2,4,5-
trichlorophenoxy acid herbicides (2,4,5-T, herbicide orange). The occurrence
of PCDDs in fish is considered to be the major source of these compounds in
humans. While PCDDs arise principally from dimerization of chlorophenols,
PCDFs are primarily produced from pyrolysis of PCBs.
Although the toxicological profiles with aquatic organisms are limited, it
has been determined that short-term exposure of fish to low ppb and ppt of PCB
and PCDD concentrations respectively, cause decreased growth rate, poor sur-
vival and increased mortality. PCDDs, however, are considered to be a probable
human carcinogen.
The accumulation of PCDDs, PCBs, and other toxic halogenated compounds in
the environment and living systems is a serious problem that has been well
documented. Although a great amount of work has been done by many groups on
the area of direct chemical decomposition of halogenated organics, relatively
little effort has been directed toward on-site chemical detoxification.
The "cleanup" of a contaminated site, which often appears in the news
media, is not really a permanent detoxification but rather a transfer of a
toxic spill from one region to another. As an example, PCB-contaminated soil
along some of the highways of North Carolina has been dug up, and has only been
removed, at great expense, to another area of that state and landfilled. These
PCBs are still in the environment and will persist there until they are removed
and destroyed.
The chemical stability of PCDDs, PCBs, and other haloorganics precludes
their destruction by conventional refuse incineration methods. Most municipal
incinerators cannot achieve the high temperatures necessary to destroy these
chemicals in refuse. The surprisingly high volatilization rates of PCBs and
other chlorinated aromatic compounds raises questions over the use of land
disposal for these materials.
Currently, some commercial chemical methods are available to chemically
alter or destroy PCBs and other haloorganics in contaminated oils. The chemi-
cal methods developed by Acurex, Goodyear, and Sun Ohio involved dispersion of
metallic sodium in oil or the use of sodium-biphenyl or naphthalene mixtures.
Because of the reactivity of sodium with water, some of these reagents cannot
be used efficiently to directly decompose PCDDs or PCBs in soils, sludges,
sediments and dredgings. Other chemical reactions have been evaluated for
-404-
-------
dan .°f/nvi™™ental pollutants but have not been found to be
adaptable to field conditions (1,2,3,).
Biological treatment of PCDDs , PCBs , and other hazardous pollutants is
also receiving attention. The efficacy of microbes to destroy toxic halo-
laboratories S n0t beSn fUlly evaluated b? U'S- EPA or independent
KPEG PROCESS
to ^ summer of 1978 a new chemical reagent was synthesized and used
to effectively dechlonnate PCB-contaminated oils (4). Since that time a
PrePared from Potassium hydroxide and polyethylene
of
reaSent Preparation, potassium hydroxide reacts with poly-
En r1^ We±ght approximately = 400) to form an alkoxide (see
Equation 1). The alkoxide in turn reacts initially with one or more of the
chlorine atoms on the aryl ring to produce an ether and potassium chloride salt
isSeaddqri0n }' /n S°me KPEG reaS6nt f°mU lati°ns> dlLthylsulf oxide (DMSO)
is added as a cosolvent to enhance reaction rate kinetics by improving rates of
extraction of aryl halide wastes into the alkoxide phase (6).
HO PEG + KOH > KO PEG + HO
Aryl-Cl + KO PEG > Aryl -0- PEG +
KC1
(1)
(2)
In 1982, detailed investigations were initiated to determine the effects
ofasoiira9?10I1TMaramet:erS T thS rate ^ eXtent °f ^^^ decontamm
of soils (9). This research focused almost exclusively on the direct
T ^PCDD-contaminated soil. The first field investigation,
January 1986, was aimed at identifying treatment conditions for
"^ "*
PURPOSE
Research and field investigation studies were initiated in January 1986
ethv^nrT f a Ch^^al reagent, prepared from potassium hydroxide and poly-
ethylene glycol, could be used to treat PCDD and PCDF contaminated oil at an
industrial wood preserving site near Butte, Montana. The wood preserving site
contained approximately 9000 gallons of light petroleum oil collected prf-
viously from groundwater over a period of two years. The oil contained 3 5%
ITlTf^^ ' PCDD ^ PCDF h0m0l°SS ranglng fr°m 422 PPb of tet?a-isomers
cM^i I PPb of octa-isomers. Because of the presence of these highly toxic
chlorinated dioxins and furans, the oil could not be transported off-site for
incineration Bringing in and operating a mobile incinerator for on-site
destruction of contaminated oil was rejected because of high costs
-405-
-------
In April 1986, U.S. EPA Region 8 agreed, after a review of laboratory
data, that the chemical process, based upon a potassium polyethylene glycol
(KMC) reagent, could be used to decontaminate the PCDD/PCDF contaminated
oil on-site.
The mobile field equipment employed to implement the previous chemical
process comprises of a 2700-gallon batch reactor mounted on a 45-foot
trailer equipped with a boiler/cooling system and a laboratory/control room
area. Heating of the raw oily waste/APEG reagent mixture was achieved by
the recirculation of the oil and reagent through a pump, a high shear
mixer, and a tubeheat exchanger which was heated using a boiler or cooled
through a series of fin-type air coolers. A schematic is shown below:
UAINRECIRCULATON
PUMP
HTFRECIRCULAT10HPUMP
The process was employed in July 1986, to successfully destroy PCDDs
and PCDFs (1000 ppm) in 9000 gallons of oil waste to non-detectable levels
(Table 1).
Table 1. TREATMENT OF CONTAMINATED OIL, BUTTE, MONTANA
Contaminants
Concentration in
Concentration in
Treated Residue (ppb)
70°C, 15 min. 100°C. 30 min.
\jUUf \jU\ »•"
TCDD (2,3,7,8-)
TCDD (total)
PeCDD
HxCDD
TCDF (2,3,7,8-)
TCDF (total)
PeCDF
HxCDF
HpCDF
OCDF
i \> i v. w» v *. w ** • • \ri '
28.2
422
822
2982
23.1
147
504
3918
5404
6230
~
•"
~
12.1
33.3
™"
4.91
5.84
~
*MDC
0.65
0.37
0.71
2.13
0.28
0.35
0.36
0.76
1.06
2.62
*Minimum detectable concentration in parts per billion.
-406-
-------
In May 1987, the KPEG was used at the request of U.S. EPA Region 7 to
destroy TCDD in 20 gallons of 2,4-D/2,4,5-T formulation stored in a 55-gallon
drum at an abandoned pesticide formulation facility in Omaha, Nebraska.
On January 22, 1987, documentation was provided to U. S. EPA's Director
Office of Emergency and Remedial Response, recommending the "Establishment of
KPEG as the Best Developed Alternative Technology (BOAT) for Four RCRA
Hazardous Waste Streams: PCBs and Ethylene Dibromide; Pentachlorophenol-oil
and Spent Solvents Contaminated with Dioxins and Furans" (11). This recom-
mendation was approved in 1987 by the Office of Emergency and Remedial Response
in its Superfund cleanup efforts. Future efforts are directed at decontamina-
tion of soils, sediments, and sludges.
KPEG TREATMENT OF PCB-CONTAMINATED SOIL
In July 1987, a small KPEG reactor was .transported to Moreau, New York at
J™ ^quest of u's- EPA- Region 2, and used successfully to treat approximately
400 Ibs. of PCB-contaminated soil. •
The purpose of conducting the demonstration at Moreau was two fold- 1) to
confirm that the KPEG reaction is effective in a 40 gallon reactor in the
dechlorination of PCBs in soil to acceptable levels « 2 ppm) and to gather
data that will be used to design the larger 2 cubic yard reactor which will be
jointly tested by U.S. EPA and the U.S. Navy in Guam. The results from the
Moreau field tests (Table 2) clearly demonstrated that PCBs in soil ranging
from 138 to 7012 ppm can be dechlorinated by KPEG to less than 10 ppm.
In November 1987, a new improved 400 gallon Littleford mixer was pur-
chased, modified as a KPEG reactor, and will be field tested on 30 tons of soil
on the U.S. Navy's Public Work Center in Guam. The PCB concentration in this
contaminated soil range from 2500-4500 ppm. Field tests with the new reactor
will commence on March 28 and will be completed by May 15, 1988.
Soil samples have been received from the Guam site, treated with KPEG and
analyzed to establish treatment conditions to lower PCBs to less than 2 ppm
The laboratory testing and analysis for the Guam site follows.
GUAM PCB-CONTAMINATED SOIL TREATMENT AND REAGENT RECOVERY
The APEG treatment of PCB-contaminated soil is a rather straightforward
process. It consists essentially of placing the contaminated soil into a
reactor followed by a 50% by weight portion of the KPEG reagent. After load-
ing, stirring and heating is commenced. Low speed stirring is required for
intimate contact of reagent and contaminated soil. Heating causes two major
effects, the first being distillation of water from the reaction mixture and
secondly, to increase reaction rates. Typically, the vat temperature of the
reactor remains near 105-110°C until the majority of water is distilled off
(30-45 minutes) after which the reactor temperature slowly rises to ca 135-
150 C. The total time for a typical reaction is from 5-6 hours.
After the required elapsed time, any condensate is removed and extracted
for residual PCBs. The flask containing the treated soil/reagent mixture is
-407-
-------
CO
CO
Ul
o
o
O-
o
Ul
D_
u.
o
z:
o
1 — 1
1—
1—
co
o
Ul
o
Ul
1 — 1
U.
=3
Ul
0
az
U.
CO
1 —
CO
Ul
C£.
h-
co
Ul
CM
Ul
_J
CO
1—
«=J-
3
c.
0
•r-
4-*
ea £- £=
0^0.
Q- c: Q-
o
c
o
o
o •—•
JJ g t_
O-c-3
n3 +J O
4-> E «-
0 -r- 3
(O 4-> O
O) ^=
o; —
CM
=*te
c
3
<*•
I— 1
=tt
c
™
Cti
0
'.J3
(O • — -
co t_ e
O 4-> Q-
0- c: CL
o
£=
0
o
c:
o — -
•r-
O -i- 3
03 4-> O
01 J^
f>f *>— -*
c
o
•p—
Q.
o_ c: Q.
cu —
o
f~*
o
o
c:
o ^-~ -
•I— Ol t/5
•P E t-
O -I- 3
<8 4-> O
cu J=
C£
c^ r^» LO eft
00 • • •
f — 1
LO
o «-i CM r--
co
CM CO CO LO
t-1 CM CX! •
O CM O1
LO LO
O i-l CM CM
CO <£>
LO VO VO
LO • «=t-
1 — O •
LO <— 1
LO
CM
o « •=*-
o
CC CM
CO CM
r-H •
o
LO
O CM
•
1— 1
-408-
-------
cooled to 40-50°C and filtered on a Buchner funnel. An additional 25-30 ml of
water is slurried with the reactor contents to aid in filtration.
After filtration and partial drying to 10-20% moisture, samples of the
treated soil are removed and extracted for PCS (Aroclor) determination.
Recovered reagent/water is saved for reuse on subsequent decontamination
runs. In this case recovered reagent from 2-3 runs is pooled, an additional
quantity of 60% KOH solution or KOH pellets are added, and the used reagent is
added to said reactor containing the contaminated soil. The process is then
continued.
PCS ANALYSIS OF TREATED SOIL
The soil coming from the reactor is extracted for residual PCB or Aroclor
determination. This process is as follows: 10 gram aliquots of the soil are
placed into 125 ml screw-top Erlenmeyer flasks to which is added 30-35 ml of
hexane/acetone (1:10), this is placed on a gyrorotary mixer at 120-130 rpm for
one hour. The extract is carefully decanted into a 250 ml separatory funnel
through a small funnel loosly packed with glass wool. This extraction is
repeated twice again using hexane/acetone 1:1 and 10:1. All extracts are
combined in said separatory funnel and the extracts are washed three times with
50 ml of water. The washed remaining hexane extract is placed into a KD
apparatus fitted with a 10 ml receiver and is concentrated to ca 2-5 ml. The
equipment is internally washed with hexane ca 2-4 ml and then the receiver is
removed, filled to the 10 cc level and agitated. Approximately 2-3 ml of this
sample is removed, placed into a 3.5 ml septa sealed glass sample container,
and subjected to gas chromatographic analysis, either to an electron capture
detector (EDC) or mass spectrometer (MS).
ANALYTICAL PROCEDURES
Generally, the in-house analytical procedure analyzed for residual
Aroclor(s) via capillary GC-ECD. In this system a 30 meter 0.32 mm I.D. DB-5
column, using an SGE cold on-column injector and a electron capture detector
with Nitrogen make-up gas is employed for routine PCB analysis. However, for
the Guam PCB/KPEG process, a HP GC-MSD with related data system for the analyt-
ical work has been utilized. In this case, the GC is equipped with a split/
splitless injector set to the splitless mode. The pumping requirements of the
MSD require either use of a narrow bore capillary column (0.22 mm I.D.), or a
jet seperator for use with packed columns. The narrow bore DB-5 column was
used in this work.
The MSD is set to acquire data from M/Z 250-500 over the 16 to 32 minute
range. The^emperature program used for the GC is 40°C for 5 minutes ramp to
180 C at 25 C per minute, hold 2.5 minutes then ramp to 280° C at 5°C per
minute, and hold 10 minutes. The total time per run is approximately 42
minutes. This method permits us to resolve the Aroclor mixture and is reliable
for as low as 10 ppm of the Aroclor mixture.
For the analytical requirements to be used for the Guam PCB detoxification
utilizing APEG, we have been requested to use the Dry Color Manufacturers
-409-
-------
Association (DCMA), PCB mixture. This mixture contains mono-thru-deca chloro-
biphenyls. Any residual PCB peaks will be quantitated by comparison to this
mixture. For example, were a hexachlorobiphenyl to remain after soil treat-
ment it would be quantified by comparison to the DCMA hexachlorobiphenyl. The
maximum residual PCB levels which are permissible in this work are to be 2 ppm
per resolvable PCB component. Therefore, all PCB peaks detected after treat-
ment must be below 2 ppm (as shown in Figures 1,2,3).
CONCLUSION
As shown, the KPEG reagent has demonstrated its capability to reduce PCDDs
and PCDFs to non-detectable levels in a variety of matrices. Further, the
reagent can be used to reduce PCB levels from the thousands of ppm occurring
from Aroclors 1248, 1254, 1260 and 1262 to levels below 2 ppm within a reason-
able time frame. This demonstrates the use of KPEG systems applicability to
these pollutants in a variety of matricies.
REFERENCES
1. Miller, J. Nucleophilic Aromatic Substitution. Elsevier Press,
Amsterdam, 1968.
2. Yoshikazu, K. and Regen, S. L. Poly (ethylene glycols) are Extraordinary
Catalysts in Liquid/Liquid Two Phase Dehydrohalogenations. J. Org. Chem.
47, 1982, (12) 2493-2494.
3. Andrews, A., Cremonesi, P., del Buttero, P., Licondra, E. and Malorano, S.
Nucleophilic Aromatic Substitution of Cr(Co)^Complex Dihaloarenes with
Thiolates. J. Org. Chem. 1983, 48 3114-3116.
4. Pytlewski, L. L., et al. Method for Decomposition of Halogenated Organic
Compounds. U.S. Patent 4,400,552 and 4,337,368.
5. Kernel, A., Rogers, C. J. PCB Destruction: A Novel Dehalogenation
Reagent. J. Hazardous Materials. 12 (1986) 161-176.
6. Peterson, R. L. Method For Decontaminating Soil. Patent Number
4,574,013, March 4, 1986.
7. Brunelle, D. J. and Singleton, D. A. Destruction/Removal of Polychlori-
nated Biphenyls from Non Polar Media Reaction of PCB with Poly (ethylene
glycol)/KOH. Chemosphere, 12, (2), 1983, 183-196.
8 Li and Alper H. Poly (ethylene glycol) Promote Reactions of Vinylic
Dibromides. Dehydrohalogenation and Palladium - Catalyzed Formal Oxi-
dative Homologation. J. Org. Chem. 1986, 51, 4353-4356.
9. Rogers, C. J. Chemical Treatment of PCBs in the Environment. EPA-600/
9-83-003, 197-201.
-410-
-------
10. Peterson, R. Potassium Polyethylene Gycol Treatment of PCDD/PCDF - Con-
taminated Oil in Butte, Montana. IT Corp./Galson Research Corp., Project
#86-706, July 1986. U.S. EPA Contract 68-03-3219.
11. Rogers, C.J., Kornel, A. Chemical Destruction of Halogenated Aliphatic
Hydrocarbons. U.S. Patent 4,675,464, June 23, 1987.
TIC of
£. 5E5-:
2. 0E5-:
1 . 5E5-;
1. BE5:
5. BE4:
Bj
TIC of
2. 5E5-:
2. BE5-:
1. 5E5:
1. 0E5:
5. 0ET4-J
BJ
DRTR:
t j
DRTR:
IE
RCBB1 . D
jtll
RHIBi . D
2B
J
2
J
;
2
j
III 1
Jill, til , , . .
24 2E 2E 3B 3S
•2'. 5E5
f2. BE5
'-1. 5E5
:1. BE5
f5. BE4
«3
-2. 5E5
f2. BE5
•1 . 5E5
-1. BE5
•5. BE4
*
Figure 1
UPPER: Aroclor 1260 extracted from Guam soil.
LOWER-: Guam soil after treatment.
-411-
-------
TIC nf DRTR: RBB10T1.D
BBB-l
TIC of DRTR:RH1B1.D
3BBB-]
Figure 2
UPPER: Aroclor 1260 at 10 ppm.
LOWER: Guam soil after treatment.
-412-
-------
TIC D
1 . 5E4-
1 0000-
5000-
0-
TIC D
1 . 5E4-]
f DRTR: RH1 01 . D
t
.^w^- -~>— *^A—
f QRTR: DCMF
1
10000-J
J
5000J
• <
101 . D
1 1 i
IE 20 22 24 2B 2B 30 31
ri.m
^1000,
7500Q
H3
f 1 . 5E4
:.«..
75000
-a
3
Figure 3
UPPER: Guam soil after treatment.
LOWER: CDMA PCB mixture
(left to right = tetra, penta, hexa at 10 ppm
and hepta, octa, nona and deca chlorobiphenyl at 5 ppm.)
-413-
-------
DEMETALLATION OF USED OIL TO FACILITATE USE
AS A NON-HAZARDOUS FUEL .
By: J. H. Kang and A. R. Tarrer
Chemical Engineering Department
Auburn University, AL 36849
Edward R. Bates
U. S. Environmental Protection Agency
Cincinnati, OH 45268
Joe Kaminski
The Office of the Assistant Secretary of Defense
Alexandria, VA 22314
Mike Dombkowski
Defense Reutilization and Marketing Service (DRMS)
Battle Creek, MI 49017
ABSTRACT
Used oil invariably contains soluble and insoluble metal-bearing
compounds. The presence of these compounds complicates the recycling of
used oil as a fuel, by causing boiler maintenance problems and by making it
difficult to comply with particulate emission standards and RCRA
regulations. Demetallating a used oil prior to its burning minimizes
maintenance problems and facilitates meeting the regulatory requirements
controlling its burning. Different physical and chemical methods can be
used to demetallate used oils. In this work, a chemical demetallation agent
was used to convert entrained non-filterable metals into a form which could
be effectively removed by filtration. Two different types of chemical
demetallation agents were used: metallic borohydrides (NaBify and KBH4> and
ammonium salts (DAP, (NH^SO^ etc.) The activity and selectivity of these
demetallation agents under different reaction conditions in various types
and sizes of reactors (thereby varying mass transfer rates) were compared in
bench-scale studies as well as in pilot-plant runs. Several types of used
oils having different metal contents were examined. Both""sedimentation and
filtration were found to be practical for removing solid metal products,
with filtration rates being significantly accelerated by the addition of low
percentages of No. 2 fuel oil.
KEY WORDS: Waste oil, demetallation, ash content, lead content, metallic
borohydrides, diammonium phosphate, lead reduction.
-414-
-------
iTitx odvxc.-ti.on.
In the United States, automotive and other oil-consuming industries
generate about 1.1 billion gallons of used oil each year [1]. This
voluminuous amount constitutes a serious waste oil disposal problem. Recent
Environmental Protection Agency (EPA) regulations and growing public concern
over hazardous wastes make development of a viable alternative to land-
filling very desirable. The objective of this work was to develop a
practical and feasible waste oil demetallation process to remove the ash-
forming contaminants, especially organolead compounds, from the oil so that
the demetallated used oil would qualify as a specification-grade fuel oil
Waste oil, although contaminated, has a high energy value. A major
operation in reclaiming waste oil is the removal of water and various ash-
forming impurities that remain dispersed in the used oil. The ash-forming
impurities consist of very fine carbon particles, organometallic
particulates (e.g., organolead compounds), and inorganic materials (e.g.,
dust).
Reuse of an untreated used oil having a high concentration of lead and
other metallic contaminants may lead to serious air pollution and boiler
maintenance problems. To qualify for minimal regulatory requirements, i.e.,
specification-g-ade fuel oil requirements under EPA regulations, the lead
content of a waste oil to be recycled as a fuel in a commercial boiler can
not exceed 100 ppm. Since most of the used oil from the automotive industry
contains several times the permissible EPA limit, it is imperative that an
economic process be developed to reduce the lead content of the oil to an
acceptable level so that the reclaimed waste oil can safely be reused as a
fuel oil without endangering public health or the environment.
Waste oils were demetallated using two types of demetallating reagents:
metallic borohydrides and ammonium salts. First, a detailed parametric study
was undertaken on a bench scale to map out the process variables so as to
identify the most efficient demetallation conditions. An evaluation of
solid-liquid separation techniques was also done to design an efficient
solids removal operation for the subsequent pilot plant studies.
Experimental Procedure (Bench-Scale Studies)
Materials -- All oil was collected from the Auburn University Waste Oil
Reprocessing Pilot Plant. The demetallating reagents used were sodium
borohydride, potassium borohydride, sodium borohydride aqueous solution,
diammonium phosphate (DAP), and ammonium sulfate. The phase transfer
catalyst used was tri-n-butyl-methyl ammonium chloride.
Equipment -- Three types of reactors were used. They were: an autoclave
(3785 ml and 300 ml), a microreactor (45 ml), and a distillation flask (300 .
ml). The autoclave was equipped with a turbine agitator, a cooling coil,
baffles, and a thermowell, all made of 316 stainless steel. -The autoclave
was heated by an electric furnace and agitation was provided by a magnedrive
system. The tubing bomb microreactor (TBMR) was constructed of 316
stainless steel tubing (1.9 cm O.D. with 0.165 cm wall thickness and a
length of 20.3 cm). Agitation was provided by a shaker assembly as shown in
Figure 1. For a typical run, the TBMR was immersed in a preheated fluidized
sand bath. As for the distillation flask, the reaction mixture was agitated
using a magnetic stirrer, and the reactor was heated in a constant
temperature paraffin.oil bath.
Analysis -- The experimental results were verified in accordance with an EPA
approved quality assurance project plan. The properties of the demetallated
-415-
-------
used oil were evaluated to assess its potential as an EPA specif ication-
erade fuel oil using the following test methods: water content, ASTM D4006-
81 (Water in crude oil by distillation); ash content, a modification of ASTM
D482-80 (Ash from petroleum products);- lead content, standard atomic
absorption (AA) analysis. .
In each ash analysis, an oil specimen (about 5 g) in a preweighed
crucible was dried first on a hot plate and then heated in a Model F-6000
Thermolyne muffler furnace at a -rate of 373K (100°C) per hour. Once the
temperature reached 873K (600°G) , the crucible was allowed to stay in the
furnace for an additional 12 hours. Then, the crucible was removed and
cooled in a desiccator. The weight of ash remaining in the crucible was
determined, and its percent concentration in the original oil sample was
computed. _ in
In each lead analysis, the ash was dissolved in an acid solution ot 10
ml of HN03 solution (1:1) and 2 ml of concentrated HC1. The lead content
was determined using AA spectroscopy. A lead working standard was prepared
by serial dilution of a lead standard solution (1000 ± 1% ppm) supplied by
Fisher Scientific. A blind-spiked oil sample was occasionally prepared to
verify the results. An organic lead standard (lead cyclohexanebutyric acid,
was used for this PurP<>se.
Results and Discussion
Use of Metal Borohydride As A Demetalling Reagent
The properties of metal borohydrides , such as NaBH4, KBH4, and SWS
solution (consisting of 12 wt% NaBH4, 40 wt% NaOH, 48 wt% water), are
described in detail in technical literature published by Morton Thiokol,
Inc., Ventron Division [2,3].
Metallic borohydrides (e.g. sodium borohydride) are strong reducing
agents. The following reaction is typical of the metal reduction that occurs
with sodium borohydride [2] .
8MX + NaBH4 + 2 H20 ---- > 8M° + NaB02 + 8 HX
where M - the metal (valence 1+)
X - the anion
In this chemical reaction, stoichiometrically, sodium borohydride
(NaBH4) has 8 reducing equivalents per molecule. Based on bench- scale
experimental results, Morton Thiokol, Inc. reported that a stabilized water
solution (i.e., SWS solution) was an effective reagent for removing lead
from used crankcase oil [4] . It should be noted that sodium borohydride
(NaBH4) degrades in the presence of water or acid via hydrolysis to liberate
hydrogen and a salt (NaB02) when the pH of the solution is low. A
representative chemical reaction is as follows :
Water Hydrolysis:
NaBH4 + 2H20 ---> NaB02 + 4 H2
A parametric study was done to investigate the lead reduction
efficiency of two Morton Thiokol products: Venpure powders (NaBH4 and KBH4)
and SWS solution. A tubing bomb microreactor (TBMR) and the reaction
conditions listed in Table 1 were used in this study. A concentrated NaOH
aqueous solution (50 wt%) was usually added to maintain a high pH and thus a
low hydrolysis activity of the metal borohydrides.
In a typical run, the TBMR was charged with 20 g of waste oil, the
specified amounts of 50 wt% NaOH aqueous solution and VenPure product. The
reactor was then sealed. Subsequently, the TBMR was attached to the
vertical shaft of the agitation assembly, as shown in Fig. 1, and agitated
at ambient temperature for 3 minutes at 860 rpm before lowering it into the
-416-
-------
preheated fluidized sand bath. The heat-up time was about 60 seconds. At
the end of the chosen reaction time, the TBMR was removed from the sand
bath, immediately quenched in cold tap water, and checked for any leaks.
The gases were released slowly from the bomb under a hood and the liquid
product was then collected and vacuum filtered. The filtrate (or product
oil) was collected for ash and lead analyses.
With 0.65 wt% SWS and 1.2 wt% NaOH solutions, the lead content of the
Type B oil was markedly reduced from 480 to 70 ppm (Run 1) as shown in Table
2. In other words, an 85.4% lead reduction was achieved. This is well
within the allowable maximum lead content of 100 ppm for specification grade
used oil set by EPA. However, a. high sludge production (16 wt%) resulted
from this run.
Since metal borohydrides are expensive, the charge of SWS solution
(containing 12 wt% sodium borohydride) was reduced to 0.43 wt% in Run 2.
Also, the charge of NaOH concentrated solution was reduced from 1.2 wt% to
0.8 wt%. The result was that the lead content of the oil decreased by only
about 50%, to 240 ppm. The ash content of the product oil was 0.58 wt%.
The amount of sludge production was about the same as that from Run 1.
In order to reduce the sludge production, in Run 3, only 0.05 wt% NaBH4
powder was used and no NaOH was added. The amount of lead reduction was
only 10.4 percent; however, little sludge was produced (2 wt%). To enhance
the demetallation activity, 0.8 wt% of 50 wt% NaOH was added in Run 4 and
the other reactions conditions were kept the same. The lead removal
efficiency increased from 10.4% for Run 3 to 37.5% for Run 4.
To improve the solubility of NaBH4 in the oil phase, 0.10 wt% of phase ^.
transfer catalyst (tri-n-butyl-methyl ammonium chloride) was also added to
the oil (Run 5). Comparing the results of Run 5 with those of Run 3 (in
both runs no NaOH was added), the addition of the phase transfer catalyst
did result in an increase in lead removal efficiency from 10.4 to 27.1%.
With 0.8 wt% NaOH solution added in Run 6, the lead reduction efficiency
improved slightly more to 31.3 wt%. On the whole, there seemed to be a
trade-off between using either SWS or NaBH4 powder as a demetailing reagent;
the former performed better in terms of lead reduction efficiency; whereas
the latter produced less sludge.
To determine the reproducibility of the demetallation experiment, Run 1
was duplicated by Run 7, and the results were essentially the same for both
runs, confirming the accuracy of the experiment.
Since waste oil is derived from various sources, it is plausible that
the lead removal efficiency using a demetalling reagent might respond
differently to different types of waste oil. Thus, a different oil, viz.,
Type A oil (lead content - 660 ppm, ash content - 0.59%) was used in Runs 8-
12.
In Run 8, the reaction conditions were almost the same as those in Run
1 only a different oil was used. The lead content of Type A oil was
reduced by about 24% to 500 ppm; whereas for Type B oil the lead content
decreased by 85% to 70 ppm (Table 3).
When the charge of SWS was doubled (Run 9), the lead content of Type A
oil was reduced by about 58% (from 660 ppm to 280 ppm) as shown in Table 4.
SWS contained 12 wt% sodium borohydride, which implied that the sodium
borohydride concentration in the oil in Run 9 was 1,560 ppm. A sodium
borohydride concentration of this level would not be economically feasible
for reclaiming waste oil.
As mentioned earlier, the runs with caustic and SWS solution were
characterized by high sludge production. A gel-like product was
-417-
-------
occasionally observed, which was difficult to filter, making the
determination of sludge content very difficult. The gel formation was
probably due to the presence of sodium hydroxide which acts to saponify
fatty acids (e.g. detergents) present in the used oil. In order to inhibit
the gel formation, sodium borohydride was added in a powder form (Runs 10
and 11) instead of in a solution; NaOH was still introduced as a 50% aqueous
solution to maintain the activity of sodium borohydride.
In Run 10, 0.15 wt% (i.e., 1500 ppm) NaBIfy was added, which was
approximately the same amount of NaBify added in Run 9; the lead reduction
efficiency was slightly lower than that of Run 9. This was probably due to
the greater solubility of NaBlfy in Run 9. In Run 11, a phase transfer
catalyst (tri-n-butyl methyl ammonium chloride) was added, and the lead
content of Type A oil was reduced by about 73% to 180 ppm.
When potassium borohydride (KBH4) was used in Run 12, the lead
reduction was 59% (i.e., from 660 to 270 ppm). Thus the final lead content
was still above the 100 ppm allowable maximum limit.
Examination of the ash content of the product oil (shown in Tables 2
and 4) reveals that, for both types of waste oil, the ash content of the
product oil was not markedly decreased by metal borohydride treatment,
although the lead content was selectively reduced in some cases. This might
be attributed to the formation of other fine salts which could not be
separated by regular filtration. A major disadvantage of using metal
borohydrides as a demetalling reagent is that the pH must be kept high for
the borohydrides to be stable, and this results in a very viscous (gel)
product.
In the previous runs, the reaction temperature used was 110°C. When
the reaction temperature was increased from 110°C to 150°C, keeping the
other conditions the same as those used in Run 1, the product was a gel, and
as a result, poor lead reduction efficiency was obtained (Runs 13 and 14,
Table 5).
Investigation of Gel Formation -- Several experiments were conducted in
which a cotton seed oil or a non-detergent virgin motor oil was reacted with
either"SWS solution or NaBH4 powder to confirm that saponification does
occur when fatty acids are present.
A cotton seed oil (20g) was blended with 1.2 wt% NaOH in 50% aqueous
solution, and then reacted with 0.65 wt% SWS in a TBMR (Run 15). The
reaction was conducted under reaction conditions given in Table 1 (similar
to those used for the TBMR demetallation runs). The reaction product was
collected in a 2.5 cm O.D. x 15 cm length test tube. The sludge produced
was about 35 volume % (Table 6). This sludge was probably the salt
resulting from the saponification of fatty acids, in the cotton seed oil,
with NaOH. The cotton seed oil was reacted with 0.05 wt% NaBH4 powder in
the absence of sodium hydroxide (Run 16). In this case no sludge was
formed, confirming that the sludge (gel) was due to the reaction of fatty
acids with NaOH. The amount of NaBH4 was doubled, and still, there was no
sludge formed (Run 17). The effect of NaOH on non-detergent motor oil was
observed by reacting an SAE-30 non-detergent motor oil with SWS (Run 18).
After the reaction, two layers of liquid were formed. The top layer was
reddish brown and the bottom layer was bluish green in color. A trace
amount of suspended solids was also observed. In summary, the gel product
formed while treating used oil with borohydrides with NaOH present was
probably due to the saponification of fatty acids.
Effect of Caustic -- In an attempt to reduce the amount of sludge
production, two reactions (Runs 19 and 20, Table 7) were performed in which
-418-
-------
1.5 wt% NH4OH was added instead of NaOH, using the reaction conditions shown
in Table 1. Also, to help stabilize the NH4OH, 25 psig of NH3 was used
(Run 19). As compared to a 0.8 wt% NaOH addition (Run 4), the amounts of
lead reduction and sludge production were about the same.
Effect of 'Water --As noted earlier, the reducing ability of metallic
borohydrides can be hampered substantially by water hydrolysis. Since the
waste oils obtained were used as-received (which had a water content of
about 5 volume %), the poor performance for those runs with metallic
borohydrides, in terms of lead reduction efficiency and ash content
remaining in the oil, may possibly be attributed to the presence of water
during the reaction. To investigate the effect of water on lead removal a
sufficient amount of Type A oil (containing 5 vol % water) was blended with
0.25 wt% and 0.50 wt% NaOH solutions (50 wt% concentrated solution)
respectively. Water along with some light ends were then distilled off via
atmospheric distillation. The residual oils were termed as Type C and Type
D, respectively. The properties of Type C & D oils are given in Table 8.
In Run 21, 20 g of dry Type C oil was reacted with 0.9 wt% of SWS
solution in a TBMR under reaction conditions given in Table 1. As shown in
Table 9, the lead content of the Type C oil was successfully reduced from
700 to 100 ppm; i.e. about 86% lead reduction was achieved. This suggests
that a much better demetalling performance can be achieved with a dry oil
(e.g., type C) than with a wet oil (e.g. Type A).
To further minimize the use of SWS solution, the charge of SWS solution
was reduced to 0.4 wt% in Run 22. As a result, about 74% lead reduction was
obtained.
In Runs 23 and 24, 0.1 wt% of a phase transfer catalyst (tri-n-butyl-
methyl ammonium chloride) was added to see if the demetalling performance
could be improved. Compared to run 21 (without catalyst), catalyst
addition (e.g., Run 23) did not improve the results.
To study the effect of NaOH charge on lead reduction, Type D oil was
used in Runs 25 and 26. The difference between Type C and Type D oil was
the amount of NaOH in the oil; the latter had twice the amount of NaOH than
the former (-1250 ppm). In comparison to Run 22, with the same charge of
SWS solution, Run 26 gave a lower lead reduction efficiency.
On the whole, Table 9 shows that the presence of water in the oil has
an adverse effect on demetallation when metallic borohydrides are used
This can be avoided by distilling off the water before the demetallation
reaction. It should be noted that the results presented so far were obtained
for a 45 ml tubing bomb microreactor (TBMR). Knowing that these bench-scale
demetallation results were to be incorporated in the scale-up and design of
the Auburn University Waste Oil Reprocessing Pilot Plant, experiments were
extended to a larger batch reactor, i.e., a 3785 ml autoclave, to simulate
actual operating conditions as much as possible.
As mentioned above, water can decrease the demetalling activity of
metallic borohydrides. Therefore, the as-received Type A oil, containing
about 5 vol.% of water, was blended with 0.25 wt% NaOH aqueous solution (50
wt%) in the following runs and then distilled to a specified temperature
under atmospheric pressure. The residual oil was assumed to be moisture-
free The moisture-free oil (or dry oil) was then treated with metallic
borohydride products in the one-gallon autoclave. The results are given in
Table 10 and discussed below:
In Run AS1, 890 g of dry oil (lead content - 700 ppm) was reacted with
0.9 wt% of SWS solution in the 3785 ml autoclave using the reaction
conditions shown in Table 11. For Run AS1, only a 37% lead reduction
-419-
-------
efficiency was obtained. The poor demetailing performance was probably due
to the low distillation cut-off temperature (129°C), which resulted in a
certain amount of water being left behind in the residual oil. In other
words, the residual oil obtained in Run ASl was not dry enough.
Thus, in Run AS2, the distillation cut-off temperature was further
raised to'l75°C. The oil and SWS solution charges to the autoclave were 900
g and 0.91 wt%, respectively, which were about the same as in Run AS1. With
the same reaction conditions, the lead content of the oil was markedly
reduced from 700 ppm to 25 ppm; i.e., a 96% lead reduction was obtained
compared to 37% in Run ASl.
Although a promising result was obtained in Run AS2, the amount of SWS
solution (containing 12 wt% NaBH4) used was considerable, and hence,
uneconomical. Thus Run AS2 was duplicated by Run AS3 except that the SWS
solution charge was halved. As a result, only a 40% lead reduction was
obtained; this suggested that more SWS solution was required, if a better
demetailing performance was to be achieved.
To compare the demetalling effectiveness of SWS solution and NaBH4
powder, Run AS4 was performed in which the same amount of NaBH4 was used as
in Run ASl (i.e. 0.10 wt% NaBH4 power) and 0.20 wt% phase transfer catalyst
(tri-n-butyl-methyl ammonium chloride) were reacted with 545 g of dry oil in
the autoclave. Coincidentally, the same level of lead reduction was
obtained. However, with NaBH4 powder.(Run AS4) the product oil had a lower
ash content as compared to Run ASl with SWS solution; also,- less sludge was
obtained in Run AS4.
To study the demetalling ability of KBH4 powder, Run AS4 was duplicated
(Run ASS) except that KBH4 powder was used. As shown in Table 10, the lead
reduction efficiency increased from 37% to 51%, but the ash content of the
oil was not altered significantly.
Since KBH4 powder is more expensive than NaBH4 powder, Run AS4 was
duplicated again in Run AS6, in which the NaBH4 charge was increased from
0.10 to 0.15 wt% while maintaining the rest of conditions the same as in Run
AS4 to determine whether a better lead reduction efficiency could be
obtained. It was found that the lead reduction efficiency increased^from 37
to 71%, but the ash content of the product oil did not change appreciably,
as shown in Table 10. Although the amount of NaBH4 used can be further
increased to improve lead reduction performance, it is uneconomical to do so
when the operating costs for reprocessing waste oil are considered.
Therefore, another demetalling method was explored in an effort to seek the
most economical process.
Use of Diamnonivm Phosphate As a Demetallation Reagent
As previously described, the major problems in using metallic
borohydrides as demetalling reagents were: (1) a soap-like sludge was
usually formed after the reaction, which was very difficult to filter; (2) a
caustic oil was produced (this could result in corrosion problems); and (3)
the treated oil after filtration usually had a higher ash content than the
original waste oil. Because of these problems, another demetallation
process, which had been proposed by the Phillips Petroleum Company as a
pretreatment step in converting waste oil into a lubricating oil was
investigated [1,5]. In this process, waste oil was reacted with a
demetalling reagent, diammonium phosphate (DAP) (both with and without the
addition of excess water). Different types of reactors were used in this
investigation to permit dominant mechanisms over different operational
ranges to be identified.
-420-
-------
In the bench scale experiments, three types of reactors were used: (1)
A tubing bomb microreactor (TBMR) as described previously; (2) A 500 ml
distillation flask equipped with a magnetic stirrer; and (3) A 3785 ml
autoclave reactor. The TBMR was operated under closed conditions, whereas
the distillation was done under open conditions; i.e. the flask was open to
the atmosphere to distill off water and trace amounts of light components.
TBMR Reaction Studies
Effect of Reaction Time and Temperature -- The variation in lead
content of the DAP-treated oil with reaction time was observed as shown in
Figure 2. The demetallation experiments were done at 150°C. From an
initial value of about 210 ppm, the lead concentration of the product oil
decreased with increasing reaction time to a final value of 16.5 ppm for the
90 minute run at an 800 cpm agitation rate. From this figure, it can also
be deduced that agitation has a significant effect on the lead content of
the oil. For example, for a 45 minute reaction time, when the agitation
rate was decreased from 800 to 400 cpm, the lead content of the treated oil
increased from 70 to 125 ppm. An increase in agitation is thought to
increase the interfacial contact area for reaction between the oil and
aqueous reactants.
To investigate the effect of reaction time on other types of oil at
various reaction temperatures, several experiments were performed in which
20g of Type B oil and 8 wt% DAP (in 40 wt% aqueous solution) were reacted in
a tubing bomb microreactor at temperatures of 200°C and 250°C, for 15 to 60
minutes. The results are shown in Figure 3. With a reaction temperature of
200°C, the lead content of type B oil was reduced from 480 to 230 ppm for
the 60 minutes run, i.e., a 52% lead reduction. As temperature was
increased to 250°C, the lead content was further reduced to about 30 ppm
(i.e. a 94% lead reduction). Consequently, it can be inferred from Figures
2 and 3 that either a higher reaction temperature or a longer reaction time
or both facilitate oil demetallation, as seen in Figure 4.
Distillation Flask Reaction Studies -- Run DF1 was performed in which 60 g
of Type A oil (lead content - 660 ppm, water content - 5%) and 8 wt% DAP
powder were mixed and heated in the 500 ml distillation flask to 150°C and
the reaction was allowed to continue for an additional hour. The heat-up
time was about 70 min. During the reaction, water and trace amounts of
light components were distilled off. A notable result was obtained in this
run (DF1) i.e. the lead content of the waste oil was successfully decreased
from 660 to 24 ppm (Table 12). In other words, a 96% lead reduction
efficiency was obtained. More promisingly, the ash content of the oil was
also reduced to 0.02 wt% which is much lower compared to that from metallic
borohydride runs ("0.5 wt%).
One may speculate that the lead reduction obtained in Run DF1 was due
to the high reaction temperature and good separation by filtration. Miller
found that some part of lead can be removed just by filtration [5]. Run DF2
was thus performed in which oil A alone, without adding any DAP, was heated
in a distillation flask for the same period of time as in Run DF1. The
result was that only a 3% lead reduction was obtained, indicating that DAP
plays a very important role in demetallation. In addition, without DAP, the
ash content of the waste oil was not altered appreciably. In an attempt to
reduce the DAP^charge, Run JDF1 was duplicated by Run DF3 except that the
amount of DAP used was cut from 8 to 2 wt%. Again, a similar lead reduction
(97%) was obtained.
Effect of Reactor Configuration on Lead Reduction -- Table 12 shows that
satisfactory demetallation results under mild conditions can be obtained
-421-
-------
using a distillation flask as a reactor. It should be noted that the
reactor was open to the atmosphere. To study the effect of the reactor
configuration on the demetallation performance, a tubing bomb microreactor
(TBMR) was used as a closed system reactor. The mass transfer effect for
this type of reactor was also studied. A detailed discussion of the
experimental results is given below.
As shown in Table 12, with 2 wt% DAP, 97% lead reduction could be
obtained (Run DF3). The reaction time for this run was 1 hr. To determine
the effect of the reactor configuration on lead removal, Run Dl was
performed in which 20 g of the same type of oil used in Run DF3 and 2 wt%
DAP were charged to a TBMR and reacted for 1 hr under a closed system. Only
a 70% lead reduction was obtained in Run Dl compared to 97% in Run DF3.
In Run D2, three steel balls (1/4" diameter) were added to the TBMR in
order to increase the mixing effect and interfacial contact area while
maintaining the rest of conditions the same as in Run Dl. The lead
reduction efficiency increased from 70 to 77%. When six steel balls were
added in Run D3, the lead reduction efficiency increased further to 81%.
To study the effect of water presence on lead reduction, Run D3 was
duplicated by D4 in which dry oil (obtained by distilling off water from
Type A oil) was used instead of wet oil (i.e. Type A oil). As shown in
Table 13, the lead reduction efficiency decreased from 81 to 78%. This
difference was attributed to a lower interfacial contact area available for
the dry oil. It should be noted that this difference was small because of
the excellent dispersion obtained in this type of reactor; a much larger
difference was observed later, using a reactor (autoclave) that did not have
as good particle dispersion properties.
The inconsistent results (Table 12 & 13) obtained from the two types of
reactors (distillation flask and TBMR) may predominantly be due to the
difference in heat-up time; i.e., for a 150°C demetallation run, the
distillation flask had a much longer heat-up time (75 min) than the TBMR (1
rain). In other words, some demetallation may have occurred during the heat-
up.
To study heat-up effects, Run D5 was performed to duplicate Run 1 (see
Table 14) except that the TBMR was preheated in the fluidized sand bath (see
Figure 1) so that the total heat-up time from room temperature to 150°C was
about 75 min. (It should be noted that in a typical TBMR run, the sand bath
alone was preheated to 150°C before the TBMR was immersed; the heat-up time
was then about 1 minute.) Table 14 shows that, with the same heat-up time
as that for distillation (Table 12), a satisfactory result was obtained from
Run D5; i.e. 91% lead reduction could be achieved from a TBMR run. In
comparison with the lead reduction obtained from the distillation run, (91%
versus 97% lead reduction), this result was acceptable for a reactor with a
different configuration and mixing pattern.
As discussed above, with 75 minutes of preheating time and an
additional hour for reaction (i.e. total reaction time, 2 hr), 91% lead
reaction could be obtained in Run D5. It would be interesting to see what
the lead reduction efficiency would be with a 2 hr reaction time, 160°C
reaction temperature and a regular TBMR 1 minute heat-up time. Run D8 was
thus performed and the results given in Table 14 show that about the same
level of lead reduction as Run D5 was obtained.
It was thus concluded that with a total reaction time of 2 hrs at
160°C, a desirable demetallation result could be achieved for various types
of reactor configurations (distillation flask or tubing bomb microreactor).
This agreement is mostly due to the fact that thermodynamic equilibrium (=
-422-
-------
complete lead removable) was achieved under these reaction conditions.
Since the distillation flask was used for the evaporation of water and
light^components, so that the hydroxyl phosphate droplets could shrink to
facilitate removal of lead particulates, an open-type reactor such as a
distillation flask (Run-DF3) may be more suitable for oil demetallation as
compared to a closed reactor, such as TBMR (Run D5).
To further evaluate the effectiveness of DAP demetallation with respect
to the configuration of the reactor used (i.e., the importance of mass
transfer), a certain amount of Type A oil (lead content - 660 ppm, water
content - 5 vol%) and 1.5-2.0 wt% DAP were reacted in 45 ml, 300 ml, and
3,785 ml (i.e., 1 gallon) autoclaves, at 160°C respectively. The 45 ml
microautoclave, as a matter of fact, was the TBMR (tubing bomb
microreactor). The results are given in Table 15.
One should note the main difference between a TBMR and a regular
autoclave in terms of agitation and heating; the former was vertically
agitated; whereas the latter was horizontally stirred. The mass transfer
rates in the TBMR are generally better than those in the 300 ml and 3,785 ml
autoclaves [7,8]. Also, the heat-up time varied a great deal; for
example,the heat-up time for a TBMR heated by a preheated fluidized sand
bath was about 1 minute, whereas, for a 300 ml or a 3785 ml autoclave
(heated by a electrical furnace from ambieat temperature) the heat-up time
was approximately 2 hours. Table 15 shows that regardless of the reactor
configuration, above 90 percent lead reduction was obtained using DAP as a
demetalling reagent under mild reaction conditions for a reaction time of
about 2 hours. The amount of solids produced was less than 1 wt%, excluding
the DAP added. These observations implied that DAP is more suitable for
used oil demetallation than sodium borohydride.
Effect of Water -- As described earlier, water present at the early stage of
a demetallation reaction can facilitate lead removal by making more surface
contact area available for the DAP.
A comparative study was carried out using a 1-gallon autoclave. The
effect of water on ash and lead removal was investigated using the reaction
conditions given in Table 16. Unlike the earlier studies with a TBMR (Table
13), the dispersion of the DAP was probably much less in this reactor (a
stirred autoclave).
In Run ADI, with dry oil obtained by distilling off water from Type A
oil, and 4 wt% DAP, a 74% lead reduction was obtained as shown in Table 17-
the ash content of the product oil from this run was 0.310 wt%.
To discern the effect of water on the demetalling performance, Run ADI
was duplicated by AD2, except that a wet oil (i.e., Type A oil) was used
As a result, 99.7% lead reduction was achieved and the ash content of the
product oil was very low (0.020 wt%). This observed increase in lead
removal efficiency that resulted with the addition of water was
significantly larger than that observed before with a TBMR and was felt to
be due to the difference in the mass transfer characteristics of the
reactors; namely, the degree to which the DAP.was dispersed.
In Run AD3, the DAP charge was reduced from 4.0 to 1 5 wt% to reduce
the operating cost and sludge generation and one part of the top of the
autoclave was opened to vent water and the lights to facilitate lead
removal. In so doing, the lead reduction was 98.8% and ash content was
U.UJU wt%. It should be noted that for all the DAP runs, the sludge
generation, excluding the DAP added, was no more than 1 wt%. A duplicate
run was done in AD4 and similar results were obtained as evidenced in Table
-423-
-------
Examination of the demetallation results (Runs AD1-AD4) for Type A oil
as shown in Table 17 reveal that: 1) much better lead and ash removal could
be achieved when used oil contained a certain amount of water (apparently,
the water aids in dispersing the DAP during initial period of reaction); 2)
total metal contaminant content (i.e. ash content) was successfully reduced
to an acceptable level « 0.1 wt%); 3) a good quality of product oil could
be produced with only 1.5 wt% diammonium phosphate (DAP) under mild reaction
conditions; and 4) reproducibility of the DAP demetallation experiment
seemed to be acceptable.
Other types of used oils (Type B & E) were also investigated using the
same operating conditions as Run AD3. The lead contents for Type B and Type
E oils were 480 ppm and 180 ppm, respectively and the ash contents were 0.52
wt% and 0.65 wt%, respectively. _
For Type B oil, a 96% lead reduction was obtained in Run ADS while a
99% lead reduction was obtained for Type E oil in Run AD6. The ash content
of the product oil from both the runs was very low. Table 17 also shows
that with 1.5 wt% DAP treatment, used oil (e.g. Type A, B, or E) seemed to
be fairly well demetallated.
Use of Ammonium Sulfate as a. Demetallation Reagent
Another promising ammonium reagent investigated was ammonium sulfate,
(NHA)2SO However, due to time constraints these experimental studies were
limited to a 300 ml distillation flask equipped with a magnetic stirrer. The
flask was open to the atmosphere to permit the escape of water and trace
amounts of light components. ., T ,, „ „ _
Effect of Reagent (i.e., (m^2so^ Amount on Product Oil Lead Content
-To study the efficacy of ammonium sulfate, different amounts of 30 wt%
ammonium sulfate solution in water were added to the waste oil as a
demetallation reagent. The reaction was carried out in a 300 ml distillation
flask for 1 hour at 160°C. The results, as seen in Table 18, are extremelly
encouraging. With 2 wt% of the reagent, the lead content of the product oil
was substantially reduced (approx. 96% reduction) from 660 ppm to 25 ppm.
This final lead content is well within the maximum permissible limit of 100
ppm for specification grade used oil. Another favorable outcome of these
experiments was the low sludge production (<1 wt%).
Effect of Temperature on Product Oil Lead Content -- As was the case
with other demetallation reagents, as the temperature was increased - the
reaction rate increased and consequently the product oil lead content
decreased substantially, for a fixed reaction time of 1 hour. Table 19 lists
these results, showing the effect of temperature on the final lead content
of the treated oil.
In conclusion it can be said, that in comparison with DAP, ammonium
sulfate seems to be just as potent in demetallation, producing a low,
quantity of non-caustic, sludge. Ammonium sulfate is however considerably
less expensive than DAP, making its usage more attractive economically.
Hence, additional studies investigating its demetallation efficiency were
performed in the pilot-plant.
Liquid-Solid Separation Studies -- The objective of these studies was to
evaluate the responses in terms of separation ease to the addition of filter
aid and the use of a diluent (No. 2 fuel oil). As has been done throughout
this work, both gravity sedimentation and filtration were studied for
removing solids.
Two DAP demetallation runs (Runs ADS and AD9) were performed in a 3785
ml autoclave to study the effects of filter aid and diluent on the settling
time and filtration rates. The response of ash and lead content of the DAP-
-424-
-------
treated oil to the settling time were obtained. The effects of dilution
with No. 2 fuel oil on settling time and filtration rate were also studied.
In Run ADS, 1 wt% filter aid (diatomaceous earth) was added to the reactor
with 700 g of Type A oil and. 2 wt% DAP before the DAP reaction, whereas, in
Run AD9, the same amount of the filter aid was added after the DAP reaction.
Table 20 shows that an oil of better quality was obtained by adding 1
wt% filter aid before the DAP demetallation (e.g. Run 8) if gravity
sedimentation was adopted; for example, in Run ADS, after 24 hour of
settling at 60°C without any diluents, the lead content of the product oil
(i.e., top-layer oil) was greatly reduced from 660 to 20 ppm (i.e., 97% lead
reduction)', whereas, in Run 9, with the addition of 1 wt% filter aid after
the DAP reaction, the lead content of the product oil was reduced from 660
to 46 ppm (i.e., 93% lead reduction). A higher ash content for the product
oil was also obtained when the filter aid was added after the reaction
(e.g., Run AD9) as seen in Table 20. When filtration was used, both runs
produced a similar oil quality. In comparison with gravity sedimentation,
the solids removal efficiency with filtration was higher. However, with
gravity sedimentation, the product oil had acceptable ash and lead contents,
and this method does have its advantages in lower operation and maintenance
costs.
With filtration j. good solids-removing efficiency was obtained; however
the filtration rates were very low for DAP-treated used oil. In order to
evaluate the use of a diluent (No. 2 fuel oil) to improve filtration rates,
another DAP demetallation run was carried out in a 1000 nil distillation
flask under the reaction conditions shown in Table 12 (Run DF3). After the
reaction, the reaction product including spent DAP was evenly divided into
four fractions; a specified amount of filter aid or No. 2 fuel oil or both
were thoroughly blended in three of the four fractions. Viscosity of each
fraction was first measured using a Cannon-Manning viscometer before
filtration (using the house vacuum). The filtration rate for each fraction
was determined, and so was the viscosity of the filtrate from each fraction.
The results showed that the viscosity, of the product oil drastically
decreased from 90 to 33 cP at 25°C when diluted with 20% No. 2 fuel oil by
weight of total liquid as shown in Table 21 (No. 2 fuel oil has a viscosity
of 3.5 cP at 25°C). The filtration rate, after dilution, approximately
doubled (from 11.5 to 25.0 ml/min). With the addition of 1% by weight of
the total liquid, the filtration rate was further increased from 25.0 to
34.0 cnrymin; however, the viscosity of the resulting liquid increased from
32.8 to 38.4 cP due to the presence of extra solids (i.e., filter aid).
Conclusion (Bench-Scale Studies)
Used oil treated with NaBH4 aqueous solution (i.e., SWS solution) was
characterized by high sludge production and ash content. High sludge
production could be circumvented by substituting SWS solution with NaBH4 or
KBH4 powder. However, the lead reduction efficiency decreased.
Demetallation performance with metal borohydride treatment could be greatly
improved by distilling off water contained in the used oil before the
reaction.
As for DAP demetallation, water present (about 5 wt%) during the early
stages of the demetallation reaction was found to be beneficial. This
beneficial effect was felt to be due to the additional interfacial contact
area available for DAP particles to react with oil. Overall, mass transfer
(DAP dispersion) was observed to be important; however, by extending the
reaction time (= 2 hr), even with mild reaction conditions, thermodynamic
-425-
-------
equilibrium (<* 100% lead removal) could be approached, thereby minimizing
the significance of any mass transfer related parameters. Ammonium sulfate
seemed to be just as efficacious as DAP, without producing a large amount of
sludge. It is also more attractive economically than DAP, however, to date
the range of experiments using it have been more limited.
In liquid/solid separation, a better quality of oil was obtained by
filtration as compared to gravity sedimentation. (Nevertheless, an EPA
specification-grade oil was produced using gravity sedimentation.) The oil
filtration rate was doubled by dilution with 20 wt% of No. 2 fuel oil.
Pilot Plant Studies
Pilot plant studies have been subdivided into two stages. The ongoing
intermediate pilot plant studies which have recently begun constitute the
first stage. The magnitude (in terms of through-puts, reactor sizes etc.) of
this stage is more limited. The principal objective of the intermediate
pilot plant study is to ascertain the reproducibility of data gathered from
the bench scale studies. Of particular importance is the role of mass
transfer or interfacial surface area between the aqueous and organic phases.
To this end, studies, using the previously mentioned demetallation reagents,
are directed toward determining the importance of the dynamics of water
removal rates versus reaction rates. The second stage of pilot plant testing
will be the operation of the final pilot plant. Its principal objective is
to fine tune the demetallation process to define the optimum operating
parameters and demetallation reagent.
Process Description of the Intermediate Pilot Plant
Used oil from the feed tank (ST-1) is pumped into the two mixing tanks
(MK-1, MK-2). MK-1 has a volume of 14 gallons whereas MK-2 holds 85 gallons
(Refer to fig. 5). In the mixing tanks, oil is thoroughly mixed with an
aqueous solution (30 wt.%) of the demetallation agent (DAP or (NH4)2 804).
The DAP solution constitutes about 2 wt% of used oil. Mixing is achieved by
an electrical agitator. Oil to be demetallated is feed to the reactor from
MK-2 by means of a metering pump. The contents of MK-1 serve as a back-up
for MK-2, in the event that MK-2 runs dry. The demetallation reactor is a
continuously stirred tank reactor (CSTR). The reactor vessel has a total
volume of 30 gallons. It is heated by means of a gas burner at the bottom
and is provided with a variable speed agitator. During operation, the
reactor temperature is maintained at 320°F, while the pressure is
atmospheric. A feed rate of 7.5 gallons/hr is used to permit a 2 hour
residence time in the reactor. The reactor is adequately insulated to
minimize heat loss. During reaction a small fraction of volatile components
is recovered in a condensation unit and the reaction product (i.e.,
demetallated oil) is pumped via a metering device to a 120 gallon settling
tank (SL-1). The settling tank facilitates separation of the product oil
from the solid waste sludge material. The product oil is subsequently
stored in a storage tank, whereas the sludge material, withdrawn from the
bottom of the settling tank, is stored in drums prior to its disposal.
Results from Intermediate Pilot Plant Studies
Initially, the biggest impedement in this scale-up operation was the
foaming of oil (>50 vol.%) in the reactor. This caused substantial loss of
oil as it would exit the reactor in the form of slugs and foam via the
outlet for the lights and flood the condenser. It is felt that the foaming
was due to rapid vaporization of lights resulting from a high heat-up rate
caused by the direct heating.of the reactor; the vapor velocity in the linch
line between the reactor and the vapor fractionation column was estimated to
-426-
-------
be 70 ft/sec.
To circumvent this problem various design changes were incorporated. As
a first step, the reactor agitator was fitted with lateral wire like
filaments at various heights along the agitator shaft. These filaments
partially helped break the foam in the reactor itself. A second
modification, was the introduction of a larger tappered cylindrical column
through which the lights, and escaping foam, would have to pass prior to
reaching the condenser. This column was 4 ft. long, with a top diameter of
1.5 ft. and a bottom diameter of 1 ft. The column was packed with ceramic
packings at the bottom and fitted with mechanical foam arresters at regular
intervals along its height. These mechanical defoamers were wire meshes
(screens) attached to the column wall. A recycle line fitted to the bottom
of the column directed the defoamed oil back to the reactor.
With the above modifications in place demetallation reactions were
carried out. No foaming occurred with excellent demetallation and deashing
resulting. These results are tabulated in Table 22. Complete material
balance data are also shown in Table 22.
Final Pilot Plant Description >
The principal units of the final pilot plant, which is designed for a
through-put of 3 gallons per minute of waste oil, consist of the storage
facility, the demetallation units, the separation units and the waste water
treatment facility. These basic units, shown in Figure 6, are described
below.
The storage facility includes separate storage tanks for the untreated
waste oil, light fuel oil recovered as a distillate from the demetallation
process, and heavy fuel oil (i.e. the demetallated used oil). Drums to store
the solid sludge-material produced from the demetallation process constitute
the remainder of the storage facility.
The demetallation unit consists of a screen filter to arrest dirt to
avoid clogging of the heat exchanger tubes in the pre-heater, a gas fired
pre-heater to heat the used oil to the reaction temperature, a continuous
stirred tank reactor (CSTR) in which the demetallation reaction takes place
and a light duty condenser to capture the volatile organics leaving the top'
of the reactor. F
The separation unit consists of a distillate tank to separate the
lights from waste water, a sedimentation unit to facilitate separation of
the demetallated oil from the oily sludge, a plate and frame filter press
which serves as a final step in removing any solid particles from the
product oil, and two skim/wash tanks in series to enable recovery of as much
entrapped oil as possible from the oily sludge material.
The waste water treatment facility incorporates two treatment tanks
The first tank facilitates precipitation of soluble contaminants (inorganic
and organic) by chemical reaction with reagents, whereas the second involves
biological treatment of the waste water to comply with municipal BOD and COD
requirements prior to its discharge to a sewer.
Results from Final Plant Studies
It is expected that the pilot plant tests will be completed during the Fall
of 1988 and the results published during the spring or summer of 1989
Demetallation Process Waste Disposal Options
The solid waste extracted from the demetallation process is
concentrated with heavy metals and consequently, is classified as an EP
Toxic hazardous waste. Added to this, growing public environmental
awareness, coupled.with constantly emerging stricter regulations, make the
disposal of this waste from the demetallation process a subject of immense
-427-
-------
importance. Nevertheless, a couple of viable options that afford a safe way
of disposing of this waste do exist.
1) Despite the high concentration of some heavy metals in the sludge, it
does have sufficient fuel value (>5,000 Btu/lb) and hence would qualify as a
hazardous waste fuel. The waste could therefore be burned in approved
cement kilns by blending, in a low ratio, with virgin fuel. A number of
other industrial furnaces, such as, lime kilns, coke ovens, blast furnaces,
aggregate kilns, phosphate kilns etc., are also authorized to burn such
hazardous waste fuel.
2) The sludge could be treated to extract all residual, entrapped oil,
allowed to dry to form a hardened cake, and then be disposed of in drums at
permitted landfills.
Acknowledgements
The authors would like to thank Morton Thiokol, Inc. for providing
metal borohydride chemicals. The information in this document has been
funded in part by the United Stated Environmental Protection Agency (EPA)
under cooperative agreements CR 812090 and CR 814635 and by the Department
of Defense under interagency grant RW 97931972 to Auburn University.
-428-
-------
References
[1] M.M. Johnson, "Reclaiming Used Oil by Chemical Treatment with
Ammonium Phosphate," U.S. Patent 3,879,282, 22 April 1975.
[2] Code of Federal Regulations, Title 40, Part 266.40.
[3] Process Stream Purification With Sodium Borohydride Technical Manual
and Users Guide, Morton Thiokol, Inc., Ventron Division.
[4] Metal Removal & Recovery With Sodium Borohydride, Morton Thiokol Inc
Ventrol Division. ' ' '
[5] Personal communication with Ventrol Division of Morton Thiokol, Inc.
[6] Miller, T.M., "An Investigation of the Demetallization Chemistry
Associated with the Re-refining of Used Motor Oil," M S Thesis North
Carolina State University, Raleigh, N.C. (1983).
[7] "The Use of Disposable Catalysts in Coal Liquefaction Activities for
Oil Production", presented at the National AIChE Meeting, Houston
Texas, March, 1981. '
[8] Gollakota, S.R., "An Investigation of Mass Transfer Phenomena in Coal
Liquefaction: Assessment of Resistances and Reactor Types," Ph.D
Dissertation, Auburn University, Auburn, AL (1984).
-429-
-------
Table I—Reaction conditions for typical TBMR runs
Demetallation reagent
Reaction temperature,°C
Heat-up time, sec
Reaction time, hr
Agitation rate, cpm
Metallic Borohydrides
110, 150
60
2.0
860
DAP
150-160
60
0.5-2.0
860
Table 2--Demetallation results using metal borohydrides (for type B oil)
Run No,
50% NaOH Added to
the Oila, wt%
SWS, wt%
NaBH4 Powder, wt%
Phase Transfer
Catalystb, wt%
After Reaction;
Sludge Production,
wt%
Product Oil Analysis:
Ash Content, wt%
Lead Content, ppm
Lead Reduction, %
1
"
1.2
0.65
0
0
16
0.36
70
85.4
7
1.2
0.65
0
0
16
0.36
69
85.6
2
0.8
0.43
0
0
15
0.58
240
50.0
3
0
0
0.05
0
2
-
430
10.4
4
0.8
0
0.05
0
-
0.77
300
37.5
5
0
0
0.05
0.1
6
0.40
350
27.1
6
0.8
0
0.05
0.1
7
0.53
330
31.3
aOil charge - 20g Oil Type: B (lead content - 480 ppm, ash content - 0.52
wt% water content — 1.8 vol%).
^Tri-n-butyl-methyl ammonium chloride.
-430-
-------
Table 3--Effect of oil type on demetallation
Run No. :
Oil Type
Lead Content, ppm
Oil Charge, g
50% NaOH Added to
the Oil, wt%
SWS, wt%
Product Oil Analvsis:
Ash Content, wt%
Lead Content, ppm
Lead Reduction, %
1
B
480
20
1.2
0.65
0.36
70
85.4
8
U
A
660
20
1.2
0.65
0.57
500
24.2
Table 4--Demetallation results using metal borohydrides (for type A oil)
Run No.
50% NaOH. Added to the
oila, wt%
SWS, wt%
KBH4 Powder, wt%
NaBH4 Powder, wt%
Phase Transfer Catalystb,
wt%
Product Oil Analysis-
Ash Content, wt%
Lead Content, ppm
Lead Reduction, %
8
1.2
0.65
0
0
0
500
24.2
9
1.2
1.3
0
0
0
0.62
280
57.6
10
1.8
0
0
0.15
0
0.56
300
54.5
11
1.8
o
0
0.15
0.10
0.50
180
72.7
12
1.8
0.15
0
0.35
0.68
270
59.1
aOil charge = 20 g Oil Type: B (lead content = 660 ppm, ash content = 0 59%
water content - 5 volume %). v.jy*,
Tri-n-butyl-methyl ammonium chloride.
-431-
-------
Table 5--Effect of reaction temperature on sludge formation
Run No. :
Oil Type
Reaction Temp . , °G
50% NaOH, wt%
SWS in the
oil, wt%
After Reaction:
Sludge Production, wt%
product Oil Analysis;
Lead Content, ppm
Lead Reduction, %
1
B
110
1.2
0.65
15.5
70
85.4
13
B
150
1.2
0.65
100% (Gel)
400
16.7
14
B
150 .
In
. 2
0.65
100% (Gel)
-
Table '--Effect of caustic solution on gel formation
Run No.;
Oil Type
50% NaOH in
the oil, wt%
SWS, wt%
NaBH4 Powder, wt%
After Reaction;
Sludge Production,
Vol%
15
Cotton Seed
Oila
1.2
0.65
0
35
16
Cotton Seed
Oil
0
0
0.05
17 18
Cotton Seed Motor Oil"
Oil
0
0
0.10
1.2
0.65
0
trace1
aMade from winterized cotton seed oil with TBHQ and citric acid in propylene
glycol to retard oxidation with dimethyIpolysiloxane added as an
antifearning agent. Made by Lou Ana Foods, Inc., Opelousas, LA 70570.
bGRC, G-100 Non-Detergent Motor Oil, SAE-30, made from Gurley Refining
Company, Memphis, Tennessee 38101.
°Two phases were formed with the top layer in reddish-brown and the bottom
layer in dark bluish-green with trace amount of suspended particle.
-432-
-------
Table 7--Influence of NH^OH on sludge production
Run No. :
Oil Type
Oil Charge, g
50% NaOH, wt%
NH4OH, wt%
NaBH4 Powder, wt%
NH3, psig
Before Reaction:
Lead Content, ppm
Ash Content, wt%
After Reaction:
Sludge Production, wt%
Product Oil Analysis!
Ash Content, wt%
Lead Content, ppm
Lead Reduction, %
4
B
20
0.8
0
0.05
0
480
0.52
4.3
0.77
300
37.5
19
JL «*
B
20
0
1.5
0.05
25
480
0.52
3.4
0.31
260
45.8
on
zu
B
20
0
1.5
0.05
o
480
0.52
3.2
0.37
,* *r *
400
16.7
Table 8--Properties of used oils
Oil Type
Water content
Ash content,
Lead Content,
, vol.%
wt.%
ppm
A
5.0
0.59
660
B
1.8
0.52
480
C
0
0.75
700
D
0
1.02
640
E
• 0.8
0.65
180
-433-
-------
Table 9--Influence of water presence on lead removal using SWS solution
Jlun no . :
Oil type
Oil charge
50% NaOH, wt%
SWS, wt%
Phase transfer
catalyst8-, wt%
9
A
20
1.2
1.3
0
21
C
20
-
0.90
0
22
C
20
-
0.40
0
23
C
20
-
0.90
0.1
24
C
20
-
0.25
0.1
product 011° Analysis;
Ash content, wt%
Lead content, ppm
Lead reduction, %
0.62
280
57.6
0.53
100
85.7
0.68
180
74.3
0.63
120
82.9
0.77
300
57.1
25
D
20
~
0.35
0
0.93
280
56.3
26
D
20
~
0.40
0
0.69
220
65.6
aTri-n-butyl-methyl ammonium chloride.
^Sludge production in all the runs was high (about 15-20 wt%).
-434-
-------
Table 10-Results for 3785 ml autoclave runs using metallic borohydrides
Run No . :
Atmoscheric Distillation Conditir
used for oil preparation!
Type A* oil charge, g
50% NaOH, wt%
Distillation cut point, °C
Reaction Mixture;
Residue5 from distillation g
SWS wt%
NaBH4 Powder, wt%
KBH4 Powder, wt%
Phase transfer catalyst0, wt%
Product Oild Analysis;
Ash content, wt%
Lead content, wt%
Lead reduction, %
AO1
AoJ.
>ns
890
0.25
129
890
0.90
0
0
0
0.74
440
37.1
AS2
1000
0.25
175
900
0.91
0
0
0
0.50
25
96.4
ASS
1000
0.25
175
900
0.46
0
0
0
0.51
420
40.0
AS 4
600
0.25
150
545
0
0.10
0
0.20
0.57
440
37.1
ASS
600
0.25
150
545
0
0
0.10
0.20
0.56
345
50.7
AS 6
600
0.25
150
545
0
0.15
0
0.20
0.53
200
71.4
aWater content =5.0%
Lead content =660 ppm
Ash content = 0.59%
Lead content - 700 ppm
Ash content - 0.86%
ctri-n-butyl-methyl ammonium chloride
dSludge production in all the runs was high (about 15-20 wt%)
Table 11--Reaction conditions for 3785 ml autoclave runs
Demetallation Reagent
Reaction Temperature, °C
Heat-Up Time} hr
Reaction Time, hr
Agitation Rate, rpm
Metallic Borohydrides
110
1.5
2
2000
DAP
160
2
1
2000
-435-
-------
Table 12--Wet oil DAP demetallation study using distillation flask
Run No.
Type A Oila charge, g
Water content in
the oil, vol %
Percent DAP added
Product Oilb Analyses
Lead Content, ppm
Lead reduction, %
Ash content, wt%
DF2
60
5
0
640
3
0.47
DF3
60
5
2
18
97
0.06
DF1
60
5
8
24
96
0.02
aType A oil:
Lead Content — 660 ppm
Ash Content - 0.59 % (no filtration)
^Reaction Conditions:
Reaction temp, °C 150
Heat-up time, min 75
Reaction time, hr 1
Reaction Pressure - Atmospheric
°Less than 1 wt% solids (excluding DAP added) was produced in all the runs
Table 13--E£fect of DAP dispersion (steel ball addition) for TBMR runs
Run No,
No. of Steel Balls Added
Product Oilp Analyses :
Lead Content, ppm
Lead Reduction, %
TBMR Reaction Conditions:
Reaction Temperature , UC
Reaction Time, hr.
Agitation, cpm
Type A oil charge, g
DAP added, wt.%
Dl D2
0 3
200 153
70 77
150
1
780
20
2
D3 D4
6 6a
127 156
81 78
aDry Oil (lead content - 700 ppm)
bLess than 1 wt% solids (excluding DAP added) was produced in all the runs
-436-
-------
Table 14--Demetallation results for TBMR runs
Run No . :
Type A oil Charge, g
Percent DAP Added, %
TBMR Reaction Conditions :
Reaction Temp, °C
Heat -Up Time, min
Reaction Time, hr
Agitation, cpm
Product Oilb Analyses :
Lead Content, ppm
Lead Reduction, %
Dl
20
2
150
1
1
800
200
70
D5
20
2
150
75
1
800
62
91
D8
20
2
160
1
2
800
64
90
D9
20a
2
160
1
-L.
2
£m
800
81
88
D10
20
0
t~
200
1
4-
I
J-
800
76
/ W
88
aDry Oil (lead content - 700 ppm)
No. of steel balls used - 0
bLess than 1 wt% solids (excluding DAP added) was produced in all the runs
Table 15--Oil demetallation using various sizes of autoclave
Run No . :
Oil Type
Autoclave Size, ml
Oil Charge' g
DAP Added, wt%
Reaction Conditions-
Reaction Temp,°C
Heat -Up Time, hr
Reaction Time, hr
Reaction Pressure, psi
Agitation, rpm
Product Olla Analysis:
Lead Content, ppm
Lead Reduction, %
no
UQ
A
45
20
2
160
0
2
N/A
800
64
90.0
A r\~7
AU/
A
300
60
2
160
2
1
150
2000
21
96.8
A T\O
AD2
A
3,785 (1 gal)
554
1 5
X • «/
160
y
£,
i
,L
atmospheric
2000
o
£,
99.7
aLess than 1 wt% solids (excluding DAP added) was produced in all the runs
-437-
-------
Table 16--Reaction conditions for 3785 ml autoclave run using DAP
Temperature
Heat-Up Time
Time
Agitation
160°C
2 hr
1 hr
2000 rpm
Table 17--DAP demetallation study using 3785 ml autoclave
Run No.at
Oil Type
Oil Charge, g
Percent DAP Added, %
Product Oilc Analvses:
Lead Content, ppra
Lead Reduction, %
Ash Content, wt%
ADI
Ab
541
4.0
180
74.3
0.310
AD2
A
554
4.0
0
100
0.020
ADS
A
554
1.5
0
100
0.030
AD4
A
554
1.5
2
99.7
0.026
ADS
B
554
1.5
18
96.3
0.016
AD6
554
1.5
2
98.9
0.036
AD7
fib
A
560
1.5
580
17.1
0.464
aRuns ADI And AD2: The autoclave was completely sealed.
Runs ADS-7: One port of the autoclave was intentionally opened
to allow the water vapor to escape in an effort to
simulate the distillation process.
bDry Oil: Lead Content - 700 ppm
°Less than 1 wt% solids (exlcuding DAP added) was produced in all runs
-438-
-------
Table IB-Demetallation with ammonium sulfate using distillation flask (300 ml)
Run No.:
Type A Oil Charge, g
Ammonium Sulfate
Reagent3 Added, %
Reaction Conditinns•
Reaction Temp, °C
Reaction Time, hr
Product 011b Analyses:
Lead Content, ppm
Lead Reduction, %
AS1
100
1.0
160
1
300
55
AS 2
100
1.5
160
1
60
91
aAmmonium sulfate was added as a 30 wt% solution in water
Less than 1 wt% sludge was produced in all runs
Run No T
Type A Oil Charge, g
Ammonium Sulfate
Reagenta Added, %
Reaction Conditions-
Reaction Temp, UC
Reaction Time, hr
Product 0-nk Analvses-
Lead Content, ppm
Lead Reduction, %
Hi'ffl M 1 ifTT»l» M Ilium,, , , , |i--LJ
AS4
100
2.0
100
1
475
28
rimnii rir~ 11 I i • i . m | m ., . „_.
ASS
i no
O n
130
1
fiO
91
aAmmonium sulfate was added as a 30 wt% solution in water
bLess than 1 wt% sludge was produced in all runs
AS 3
100
2.0
160
1
25
96
AS 6
100
2.0
160
1
25
96
-439-
-------
Table 20--Effect of filter aid and/or diluent on oil quality
Jlun No , :
I.
11.
Sedimentation in 60°-C:
Water bath for 24 hr
Oil only
Oil+1 wt% filter aid
Oil+20 wt% No. 2 fuel oil
Oil+1 wt% filter aid +
20 wt% No. 2 oil
Vacuum Filtration:
Oil only
Oil+1 wt% filter aid
Oil+20 wt% No. 2 fuel oil
Oil+1 wt% filter aid+20
wt% No. 2 fuel oil
AD *
Ash Content,
wt%
0.046
0.012
0.012
0.008
Pb Content,
ppm
20
10
0
0
AD 9
Ash Content, Pb
wt%
0.066
0.124
0.050
0.066
0.012
0.014
Oooi,
t W-r
0.002
Content ,
ppm
52
46
. 36
32
4
6
2
4
Table 21--Effect of filter aid and diluent on filtration rate and oil viscosity
Viscosity at 25°C
(cP)
Oil Only 89.9
Oil +
1 wt% Filter
Aid 95.5
Oil +
20 wt% No. 2
Fuel Oil 32.8
Oil +
20 wt% No. 2
Fuel Oil +
1 wt% Filter
Oil 38.4
Filtration Rate
(ml/min)
11.5
12.0
25.0
34.0
Filtrate Viscosity
at 25°C (cP)
87.7
87.3
37.3
37.6
-440-
-------
Table 22-Batch run of intermediate pilot unit: Percentage removal of ash and lead
Oil Charge, g ig
DAP Soln.a Added, wt.% 2
Before Reaction:
Lead Content, ppm 155
Ash Content, wt.% 0.4
Reaction Conditions:
Temperature, °F 300
Pressure, atm. i
Reaction Time at 300°F, hr. 1
Heat-up Time, hr. 1.5
Productb Oil Analysts;
Lead Content, ppm 1Q
Lead Reduction, % 94.5
Ash Content, wt.% Q.05
Ash Reduction, % 90
a DAP soln. (30 wt.% DAP, 70 wt.% H20)
b Material Balance:
Oil Balance (gal):
Input
Oil Feed —. 18.8
Output
Heavy Oil Product - 16.8
Light Oil Product - 2.0
Net Losses =0.0
Water Balance (gal):
Input
Water in Feed Oil - 1.2
Water in DAP Soln. = 0.8
Output
Water in Lights - 2.0
Net Losses = 0.0
-441-
-------
VARIABLE
SPEED
ELECTRIC
MOTOR
.RING
TACHOMETER
•AGITATOR SHAFT
•MICROREACTOR (TUBING BOMB)
^-—FLUIDIZEO SAND BATH
TEMPERATURE CONTROLLER
Figure 1 — Schematic diagram of tubing
bomb microreactor agitation assembly
240 -
O 8OO CPM
A 40O CPM
w 20 40 60 80 100
Reaction Time, Min
Figure 2 — Effect of reaction time and
agitation rate on lead removal
Oil Type- Type B .
Reaction Time- 45 mm.
Reaclor Volume- 45 ml
DAP % in Oil' 6%
Water added' 14%
Oil' Type B
Reaclor Volume' 45 ml
OAP % in Oil' 8%
Water Added- 12%
•~30 45 60
Reaclion Time, mln.
i5O
Reaclion Temperature, 'C
Figure 3 — Effect of reaction time on
lead removal at various temperatures
Figure-4'-- Effect of reaction temperature|
on lead removal for type B oil
-442-
-------
OH. FfiQU FEEO._TAHK. ST-I
SOLUTION
OF OEuCTALLlNG
AGENT
SOLUTION OF OQETALIJNO AGENT
Figure 5 — Intermediate pilot-plant flowsheet
SEPARATION
UNIT
I SKttHW*
DEMETALLAT10N
UNIT
SKWWASH TANKS
Figure 6 — Final pilot-plant flowsheet
STORAGE TANKS
1 4TTTm
ffcS?
UASTE WATER
TREATMENT UNIT
*• SCLC-R1CH UATERIAL
(DISPOSAL BY LANDFILL
Of RESLUP.RY1NG AMD
DISPOSAL AS A UETAL-
Rtt^ FUEL)
PRCOPtTATKJH
-443-
-------
THE INCOMPLETE COMBUSTION OF CARBON TETRACHLORIDE DURING
NORMAL/ABNORMAL HAZARDOUS WASTE INCINERATlM
by: Robert C. Thurnau
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Land disposal of hazardous waste is the subject of the requirements
of Subtitle C of the Resource Conservation and Recovery Act (RCRA) of
1976. This act requires that the treatment, storage or disposal of
hazardous waste be carried out in accordance with the RCRA regulations.
Incineration of hazardous waste falls under these regulations and is
covered by the permitting requirements. When used for hazardous waste
treatment, incineration must demonstrate that at least 99.99% of the
principal organic hazardous components (POHC) are destroyed. However,
as efficient as incineration is at destroying POHCs, by-products of
combustion are formed and in some cases may be just as much of a health
hazard as the parent compound.
The incineration of carbon tetrachloride has been successfully
demonstrated at numerous hazardous waste incinerators. One of the
products of incomplete combustion of carbon tetrachloride is chloroform
and this paper presents a comparison of the observed chloroform values
with those predicted from the unimolecular thermal decomposition reaction
for carbon tetrachloride.
-444-
-------
INTRODUCTION
The disposal of hazardous waste on land is the subject of Subtitle C of
the Resource Conservation and Recovery Act (RCRA) of 1976 and its subsequent
Amendments in 1984. These statues require that the storage or disposal of
hazardous waste be carried out in accordance with Federal regulations. The
incineration of hazardous waste falls under these regulations which specify
that in addition to particulate and HC1 emission limits, incinerators must
rreduce specified principal organic hazardous constituents (POHCs) by at least
99.99%, based on the ratio of stack gas emissions to feed concentrations.
However, it has been reported that the combustion of hazardous waste1 even at
efficiencies of 99.99% can generate products of incomplete combustion (PICs),
and some of these compounds are just as undesirable as the parent compounds.
This paper deals with the combustion of carbon tetrachloride and one of its
thermal degradation products, chloroform.
BACKGROUND
To help comply with the statute requirements of 99.99% destruction
removal efficiency (ORE), indicator compounds such as carbon tetrachloride
(CC14) are spiked into the feedstock and incinerated along with the 'specified
wastes. The data generated from CC14 spiked incinerator tests, supplies
additional information on the incinerator's ability to destroy a thermally
stable compound, and adds credence to the overall combustion efficiency of
the POHCs.
The incineration of hazardous wastes containing CC14 has been docu-
mented in at least four different hazardous waste incinerators^. It has also
been reported-^ that a family of chlorinated methane compounds (chloroform,
methylene chloride dichloromethane, monochloromethane) was generated when a
hazardous waste spiked with CC14 was incinerated. From a total mass
emissions (TME) standpoint, the rate at which (PICs) are generated is
important to the overall performance of the incinerator. If the mechanism
and rates of formation and destruction are known, the magnitude of the PIC
emissions can be determined, and controlled (if necessary) by regulating the
feed materials and/or changing incinerator conditions.
UNIMOLECULAR DECOMPOSITION MECHANISM
The thermal stability of numerous organic compounds has been studied in
great detail by Dellinger4»5 and associates at the University of Dayton
Research Institute. Carbon tetrachloride and chloroform are two of the
compounds studied and data on pseudo-first-order kinetic activation energy
and Arrhenius coefficent were determined and reported^. The thermodynamic
work was done under laboratory plug flow reactor conditions and was not
-445-
-------
applied to full-scale incinerators. However, after an extensive review of a
XEently comp eted EPA incinerator test, it was concluded that the oxygen
profiles, temperature and residence time distribution were_not so radically
different as to preclude the laboratory work from application to the field
test data.
It was postulated that CC14 decomposed thermally via the rupture of the
carbon/chloride bond and this reaction was characteristic of a unimolecular
decomposition mechanism. During the operation of a hazardous waste inciner-
ator, the probability of generating gas phase free radicals of this type is
sufficiently large to suggest the following mechanism for the destruction o
carbon tetrachloride.
of
CC13 + Cl
CHC13 + OH
HCL + OH
Fast
The overall reaction for carbon tetrachloride can be written as:
CCL4 + 2H20 __,_ CHCL3 + HCL + 20H Eq. 1
Once the chloroform has been generated, its oxidation mechanism can take
over and may be represented by the following global reaction:
CHC13 + 1/202 + 2H20
C02 + 3HC1 + H20
Eq. 2
The operating conditions recorded during the field test were applied to
the unimolecular decomposition mechanism, and the concentration of the POHC
(carbon tetrachloride) and one of the PICs (chloroform) was calculated. This
information was compared with the stack data collected during the test.
EXPERIMENTAL
The test was conducted on a rotary kiln incinerator system that rou-
tinely burned both conventional and hazardous waste generated by a large
petrochemical manufacturing plant. The kiln can receive solid waste, slurry
feed and liquid waste. During the test the slurry feed was not used. The
primary combustion chamber gases exit from the end of the kiln, and are
directed upward to the secondary combustion chamber. The gases leaving the
secondary combustion chamber are passed through a vertical quench section
and a three stage ionizing wet scrubber before being exhausted to the
atmosphere. A schematic diagram of the incineration system with its various
inputs is presented in Figure 1. The following data was taken on the
feedstocks presented to the kiln for incineration:
Solid Waste Feedstock
1) Chlorinated pyridine
2) Polyethylene wax
3) Substituted cellulose
3,000 Ibs/hr
24,000 BTU
TF~
-446-
-------
-447-
-------
Liquid Waste - Organic Fraction
Liquid Waste - Aqueous Fraction
1341 Ibs/hr
-------
Using the pressure, temperature and volume relationships in the Ideal Gas
Law the expanded volume of 0.081 moles was calculated to be 7.23 liters.
The volume of the combustion chamber was calculated as lOObft^ or
28462 liters, and when the carbon tetrachloride was dispersed in the
combustion chamber its concentration was calculated to be 254 ppm or 2.85 x
10-6 moles/L.
The other variable in the pyrolysis component of the equation was the
concentration of water vapor. The aqueous waste stream supplied water to
the combustion chamber at the rate of 10 gallons per minute. A calculation
similar to that carried out for carbon tetrachloride was done for the water
component and found to be 109,000 ppm or 1.23 x 10~3 moles/L. The concen-
tration of water vapor was about 400 times larger than the carbon tetra-
chloride, thus the limiting factor for the'pyrolysis reaction was the amount
of CC14 available for reaction. The concentration of water was assumed to be
constant and was Incorporated into the rate constant. Substituting these
values into the rate expression and assuming first order kinetics, the amount
of unreacted carbon tetrachloride remaining after undergoing unimolecular
decomposition mechanism was calculated:
rate = -dCCU = K [0014]
dt
-dCCl4 = 6.12 x 10~12 moles/L/sec
dt
If the decomposition rate for carbon tetrachloride is compared with the
charging rate, it is found that a very small amount of the initial CC14
(28 ppb) was projected to remain unreacted when the combustion gases exited
the incinerator. There were no samples of the incinerator exit gas taken
and therefore no direct way to compare the actual carbon tetrachloride
concentrations with the predicted values. It was also assumed that the
chloroform formed (difference between initial and final carbon tetrachloride
concentrations) was also incinerated with the rest of the waste.
The combustion gases derived from the incineration of the spiked waste
material pass from the primary combustion chamber to the secondary combustion
chamber where the reactions initiated in the primary are thermally encouraged
to go to completion. The products of incomplete combustion generated in the
kiln would be destroyed in the secondary combustion chamber with little if
anything remaining in the stack gas. However, a significant amount of the
spiked organic waste was also injected into the secondary combustion chamber
(1244 grams/min) and the kinetics of its thermal decomposition and PIC
formation would be reflected in the measurements taken in the stock.
The total volume of the secondary combustion chamber is about 6,000 ft-*,
but the location of the waste injection port and the velocity of the gases
reduces the effective reaction volume to about 3,000 ft-*. During the
collection of the first set of samples on the volatile organic sampling train
(VOST), the secondary combustion chamber was maintained at 1853° _+ 31°F.
Using the same approach that was applied to the CC14 in the primary com-
-449-
-------
bustion chamber, and accounting for the additional CC14 injected as well as
the extra reaction volume, the following kinetic data can be calculated:
rate constant = k = Ae
-Ea/RT
k = 288000 sec-1 e
cal
-26,000 mole
1.99 cal
mole
x 1285°K
k = 11.06 (for 1853°F) sec'1
[CCL4] = 1.58 x 10'6 moles/L or 142 ppm
[H20] = 2.46 x 10" 3 moles/L
Substituting the rate and concentration data into Equation #1, the amount of
carbon tetrachloride left unreacted by the unimolecular decomposition
portion of the overall reaction can be calculated:
dt
= k1 [CC14]
k1 = pseudo first order
rate constant
d = 1.06 x 10"10 moles/L/sec
dt
Again, if the decomposition rate for carbon tetrachloride was compared with
the charging rate, it was found that only a small portion of the initial CC14
remained unreacted (2 ppb) and was quenched as the gas exited the secondary
combustion chamber. The stack test data collected during this test indicates
a carbon tetrachloride concentration of 5 ppb and is in agreement with that
predicted by the unimolecular decomposition mechanism. Table 1 compares the
predicted carbon tetrachloride concentrations with the measured carbon
tetrachloride concentrations for the normal operating conditions, and Table 2
summarizes the comparisons for the upset conditions.
-450-
-------
The stack test data also showed significant concentrations of chloro-
form, and it is suggested that the chloroform was generated from the thermal
decomposition of carbon tetrachloride. If, as the unimolecular mechanism
suggests, all of the carbon tetrachloride is converted to chloroform, the
same type of kinetic analysis could be carried out on the chloroform as on
carbon tetrachloride and its exit concentration could also be predicted:
rate constant = k = Ae-Ea/RT
cal
x 1285°K
-49,000 mole
k = 3.16 x 1011 sec'1 e 1.99 cal
mole °K
k = 15.06 x 102 sec-1
[CHC13] = 142,000 ppb - 2 = 141,998 = 142 ppm
[H20] = 1.23 x ID'3 moles/L
[02] = 1.84 x 10~2 moles/L
rate = ki [CHC13] ki
rate = 41.72 x 10-6 [1>58 x 10-6]
rate = 6.592 x 10-H moles/L/sec
If the decomposition rate for chloroform is compared with the charging
rate, it is found that only a small amount of chloroform (6 ppb) was pre-
dicted to remain unreacted. Table 3 compares the predicted chloroform
emissions with the measured chloroform emissions taken under normal
operating conditions.
= integrated rate constant
-451-
-------
Table 1. Comparison of Theoretical and Actual Carbon Tetrachloride
Emissions From a Hazardous Waste Incinerator - Normal Operation
Mean
Calculated
Concentration
ppb
Measured
Concentration
ppb
Sample 1
Sample 1
Sample 1
Sample 2
Sample 2
Sample 2
Sample 3
Sample 3
Sample 3
Pair 1
Pair 2
Pair 3
Pair 1
Pair 2
Pair 3
Pair 1
Pair 2
Pair 3
2
2
2
2
2
2
2
2
2
5
7
NA
2
1
1
3.4 + 2.2
Table 2. Comparison of Theoretical and Actual Carbon Tetrachloride Emissions
From a Hazardous Waste Incinerator - Operating with Periodic Upsets
Calculated
Concentration
ppb
Measured
Concentration
ppb
Sample 4
Sample 4
Sample 4
Sample 5
Sample 5
Sample 5
Sample 6
Sample 6
Sample 6
Pair 1
Pair 2
Pair 3
Pair 1
Pair 2
Pair 3
Pair 1
Pair 2
Pair 3
2
3
2
1
1
2
1
1
7
7
/
12
21
16
7'
6
O
*J
Mean
9.7 + 5.7
-452-
-------
Table 3. Comparison of Theoretical and Actual
From a Hazardous Waste Incinerator -
Chloroform Emissions
Normal Operation
SAMPLE
Test 1 Pair 1
Test 1 Pair 2
Test 1 Pair 3
Test 2 Pair 1
Test 2 Pair 2
Test 2 Pair 3
Test 3 Pair 1
Test 3 Pair 2
Test 3 Pair 3
Mean
Calculated Concentration
ppb
6
7
6
2
2
3
2
2
1
Actual Concentration
ppb
63
51
63
18
11
NA
72
57
66
50.1 + 22.9
45.7% RSD
After the baseline experiments were completed, the incinerator was
periodically operated in a way that would simulate upset or abnormal com-
bustion conditions i.e. burner failure, barrel of concentrated waste etc.
These conditions are undesirable from many standpoints, and with proper
operator control usually last for short periods. The transient upset
condition was simulated by spiking a large dose of waste organic solvent
into the secondary combustion chamber. The upset resulted in elevated
carbon monoxide and total hydrocarbon concentrations, but lasted for only
2 to 3 minutes.
The same battery of samples were collected during the upset conditions
as were collected in the baseline experiments and the collected data was
treated in the same manner. Table 4 summarizes the chloroform concen-
trations predicted by the unimolecular decomposition mechanism and the
chloroform concentrations found in the stack.
-453-
-------
Table 4. Comparison of Theoretical and Actual Chloroform Emissions From
a Hazardous Waste Incinerator - Operating with Periodic Upsets
SAMPLE
Test 4 Pair 1
Test 4 Pai r 2
Test 4 Pair 3
Test 5 Pair 1
Test 5 Pair 2
Test 5 Pair 3
Test 6 Pair 1
Test 6 Pair 2
Test 6 Pair 3
Mean
Calculated Concentration
ppb
3
3
3
3
3
3
7
6
6
Actual Concentration
. PPb
75
64
82
23
31
19
14
74
21
44.8 + 28.
63% RSD
2
DISCUSSION AND CONCLUSION
The application of'thermal degradation data generated under closely
controlled laboratory conditions to a full-scale commercial hazardous waste
incinerator must be done in a careful manner. The unimolecular decomposition
mechanism predicted that under both normal and upset conditions very low
concentrations of carbon tetrachloride would result when waste spiked with
carbon tetrachloride was incinerated. The results measured under normal,
conditions fit well with the predicted values, but are slightly higher for
the upset conditions. The higher carbon tetrachloride values resulted when
the combustion conditions in the secondary combustion chamber were altered to
simulate transient upset situations.
The application of a global oxidation mechanism to the chloroform gen-
erated from carbon tetrachloride thermal decomposition also predicted that
small amounts of chloroform would be present in the stack gas. The chloro-
form values found in the stack gas were significantly higher than predicted
and may have been due to improper assumptions in the reaction mechanisms,
erroneous pseudo-first-order kinetic parameters, or the assimilation of
additional CHCls from the combustion train (i.e. scrubber water).
Under normal combustion conditions the ORE for carbon tetrachloride was
calculated at >99.99%. The same calculation for carbon tetrachloride under
upset conditions also yielded a DRE of >99.99% and supported earlier data
that suggests that upsets do not result in significant reductions in POHC
DRE.
-454-
-------
If the thermal degradation of carbon tetrachl.orlde takes place as out-
lined earlier in this paper, a considerable amount of chloroform will be
generated and oxidized. If the residual products of incomplete combustion
(CHCls) are accounted for in the carbon tetrachloride calculation (the ORE)
the level of destruction drops to 99.96% for both upset and normal combustion
conditions.
The carbon tetrachloride/chloroform results summarized in Tables 1
through 4 were taken under relatively constant temperatures (1771° +_ 71°F
and 1805° jL42°F). The temperature in the secondary combustion chamber was
consistent between the two sets of data and should not be a significant
factor in contributing to the differences observed between the emissions
measured under normal, and upset operating conditions.
REFERENCES
1. Trenholm, A.; Hathaway, R.; Oberacker, D.; "Products of Incomplete
Combustion From Hazardous Waste Incinerators; Incineration and Treatment
of Hazardous Waste", Proceedings of the Tenth Annual Research Symposium,
NTIS PB 85116291, EPA-600/9-84-022, Sept. 84.
2. Castaldini, C.; Mason, H.; DeRosier, R.; Unnasch, S.; "Field Tests of
Industrial Boilers Cofiring Hazardous Wastes", Hazardous Waste, Vol. 1
No. 2, 1984.
3. Trenholm, A; Thurnau, R.; "Total Mass Emissions From a Hazardous Waste
Incinerator", Proceedings of the Thirteenth Annual Research Symposium,
EPA-600/9-87/015.
4. Dellinger, B.; Torres, J.; Rubey, W.; Hall, D.; Graham, J. and
Carnes, R.; "Determination of the Thermal Stability of Selected
Hazardous Organic Compounds", Hazardous Waste Vol. 1, No. 2 1984
p. 137-157.
5. Graham, J.; Hall, D.; Dellinger, B.; "Laboratory Investigation of
. Thermal Degradation of a Mixture of Hazardous Organic Compounds",
Enviro. Sci. Techno!. Vol. 20, No. 7, 1986, p. 703-10.
-455-
-------
HAZARDOUS WASTE INCINERATION
PRIOR TO LAND DISPOSAL
by
Ronald Turner
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268
and
Robert Hoye and Fred Hall
PEI Associates, Inc.
11499 Chester Road
Cincinnati, OH 45246
INTRODUCTION
The EPA's Hazardous Waste Engineering Research Laboratory (HWERL),
Cincinnati, Ohio, is providing the Office of Solid Waste and Emergency Re-
soonse with data on various hazardous waste treatment technologies to assist
in the development of land disposal restriction standards under the Resource
Conservation and Recovery Act (RCRA). Manv of the RCRA listed wastes which
may have been land disposed in the past are or may be incinerated. The
incinerable wastes range from highly concentrated organic liquids to sludges
and low-concentration solid wastes. The HWERL had collected stack gas, ash,
and other analytical data from incineration of mixed RCRA and non-RCRA
wastes. The purpose of the tests described in the paper is to characterize
the residuals from incineration of specific RCRA wastes.
Earlier HWERL studies of incineration cases have examined the perform-
ance of combustion systems relative to destruction and removal of organic*
and metals in the feed. Standards have been promulgated for stack gas from
-456-
-------
hazardous waste incinerators. They require a ORE of 99.99 percent for each
principal organic hazardous constituent (POHC), removal of 99 percent of
hydrogen chloride from the exhaust gas, and limiting of particulate matter
to 180 milligrams per dry standard cubic meter. The other two residual
streams which may result from incineration are bottom ash and air pollution
control device (APCD) effluent which are not specifically regulated under
RCRA. Many of these residues are hazardous wastes and may also contain heavy
metals and undestroyed organic material. Additional treatment may be neces-
sary to remove or stabilize the hazardous constituents in the ash or efflu-
ents before ultimate disposal.
To determine the quality of incineration residuals, the HWERL recently
sampled the ash and APCD effluents at 10 commercial facilities incinerating
mixtures of hazardous wastes. Stack tests and feed sampling were not con-
ducted. Metals, volatile, organic compounds, and semivolatile organic com-
pounds were detected in the scrubber waters and bottom ash. Overall, the
data indicated that the facilities were apparently capable of achieving high
levels of RCRA organic hazardous material destruction as very small amounts
of residual organic compounds remained in the incineration ash and air pollu-
tion control effluents.*
Metals such as arsenic, barium, beryllium, chromium, cadmium, lead,
mercury, nickel, and zinc are of concern in incineration. The principal
environmental concern is the form of the metals; i.e., bottom ash, stack
emissions, or APCD residues. In general, data on metal emissions and parti-
tioning are limited and often incomplete. Therefore, it was decided that
additional tests would be necessary to characterize the organics and metals
in the residues for specific RCRA waste-incinerator combinations.
This paper presents the results of a complete analysis of the feed,
residues, and effluents from the incineration of two unadulterated RCRA
hazardous wastes:
K001 - Bottom sediment sludge from the treatment of wastewater from wood
preserving processes that use pentachlorophenol(PCP)/rotary kiln
0 K015 - Still bottoms from the distillation of benzyl chloride/liquid
injection
Although K001 can consist of sludges generated from wood-preserving processes
that use PCP or creosote, only the treatment of K001-PCP is discussed in this
paper. KOOl-creosote was also incinerated in a rotary kiln test. The K001-C
waste was similar to the K001-PCP in its soil and water content, but the
organic concentrations and heating values were about twice'as high in the
creosote waste. The residual concentrations of organics and inorganics were
similar to those obtained for the K001-PCP test.
Van r«uren, D., G. Poe, C. Castaldini. Characterization of Hazardous Waste
Incineration Residuals. U.S. EPA. January 1987.
-457-
-------
WASTE CHARACTERISTICS
K001-PCP
Wood-preserving processes that use pentachlorophenol (PCP) and creosote
qenerate v/astewater containing phenolic compounds, including penta- and
tetrachlorophenol, volatile organic solvents such as benzene and toluene, and
polynuclear aromatic (PNA) components of creosote. Treatment of this waste-
water results in a bottom-sediment sludge (listed as K001) that must be re-
moved for ultimate disposal. Over 170 wood treatment plants using creosote
and/or PCP generate approximately 3 million gallons of wastewater treatment
residuals per year.* Wood treaters periodically remove the K001 from their
water treatment pond and generally send it to a hazardous waste landfill or
incinerator for disposal.
The approximate concentrations of the major constituents of the untreat-
ed K001-PCP waste are given below.
Constituent
Soil
Water
Wood chips
Organic compounds
Concentration, percent
40
30
10
20 (naphthalene, phenanthrene
fluoranthene, PCP, others)
100
Significant composition parameters of the sampled K001-PCP waste are
presented in the following listing. These parameters represent waste charac-
teristics that would affect treatment performance. The wide range of values
indicates the variations in composition that can be expected in K001-PLP,
even from the same source.
Parameter
Ash content
Heating value
Water
Pentachlorophenol
Range of determined values
12 to 51% (30% average)
3800 to 8300 Btu/lb (6000 Btu/lb average)
8 to 41% (25% average)
920 to 3000 ppm (2000 ppm average)
A rotary kiln incinerator was used for destruction of the K001-PCP due
to the semi-solid form of the waste and its high ash content.
K015
RCRA waste K015 is still bottoms generated during the production of
benzyl chloride. In this process, toluene is chlorinated in the presence of
ultraviolet light to produce benzyl chloride, which is then separated from
Radian Corporation. Draft Final Engineering Analysis of Wood Preservation
"and Surface Protection-Volume 1. Prepared for U.S. Environmental Protec-
tion Agency under Contract No. 68-01-7287, Washington, D.C. October 1986.
-458-
-------
unreacted toluene and refined. Two U.S. facilities generate a total of about
3 million pounds of K015 per year. The K015 is either incinerated at a RCRA
permitted treatment, storage, and disposal facility (TSDF) or sold as a raw
material. The approximate concentrations of the major constituents-of the
untreated K015 waste are indicated below.
Constituent
Benzal chloride (C6H5CHC12)
Benzyl chloride (C6H5CH2C1)
Benzotrichloride (C6H5CC13)
Water
Other constituents
Concentration,
percent
88-100
<1
3.7-4.5
Other significant parameters of the sampled KOI5 waste are presented
below. These parameters represent waste characteristics that would affect
incinerator treatment performance.
Parameter
Ash content
Heating value
Carbon
Dry loss
Sulfur
Range of determined values
0.01 to 0.29% (0.09% average)
10,000 Btu/lb
51.0 to 51.3% (51.1% average)
96.0 to 99.02% (97.2% average)
0.03 to 0.32% (0.22% average)
The K015 waste.contains approximately 43 percent chlorine. It was
transported to the treatment facility in 55-gallon drums. A liquid injection
incinerator system was used for destruction of the K015 because of the mate-
rial's high organic/low ash content.
-459-
-------
TEST FACILITY DESCRIPTIONS
EPA COMBUSTION RESEARCH FACILITY (CRF)
The K001-PCP test burn was conducted at the U.S. EPA Combustion Research
Facility located in Jefferson, Arkansas, in a pilot-scale rotary kiln incin-
erator rated at 2.5 million Btu/h and 1 hour retention time at 0.25 rpm.
Figure 1 presents a schematic diagram of the incineration system. A ram
feeder was used to inject 1.5-gallon, cylindrical fiber packs, each contain-
ing about 5 Ib of the K001-PCP. The kiln is designed to operate at tempera-
tures up to 1000°C (1832°F); however, the kiln has occasionally reached
temperatures of 1150°C (2100°F).
The combustion gases from the kiln pass through an afterburner for fur-
ther incineration. The afterburner's design temperature is 1200°C (2200°F),
with a 2-second residence time. Both the kiln and the afterburner use pro-
pane as startup fuel and as supplementary fuel during the waste burn.
The hot combustion gases- leaving the afterburner enter a venturi scrub-
ber, which is followed by a packed tower, a carbon bed, and a high efficiency
particulate air (HEPA) filter in series. An induced-draft (I.D.) fan is in
line after the HEPA filter. The carbon bed and HEPA filter were added to the
CRF system because of operating permit requirements.
Sodium hydroxide is added to the scrubbing system (venturi and packed
tower) to maintain a pH greater than 7. Makeup water is added at a rate of 5
to 10 gallons per minute, and the water system is blown down continuously at
a rate of 2.0 to 2.5 gallons per minute.
Strip charts continuously monitor and record oxygen (02), carbon monox-
ide (CO), and carbon dioxide (C02) at the afterburner outlet and after the
scrubber system. An operator also manually recorded 02, CO, C02, tempera-
ture, scrubber pressure drop, feed rate, and makeup and blowdown water flows
every 15 minutes. U.S. EPA Modified Method 5 was used to collect pentachlo-
rophenol samples in the stack gases after the scrubbing system but before the
charcoal bed. EPA Method 8270 was used to measure concentrations.
The temperature in the kiln is controlled by increasing or decreasing
waste feed rate, combustion air, and/or supplementary fuel. Waste feed rate
was manually controlled by changing the rate of waste filled fiber packs, fed
to the incinerator. The feed rate was about 86 pounds per hour for the
tests. Combustion air and supplementary fuel are also manually controlled by
valve adjustments.
During the three tests, the rotary kiln operating temperatures ranged
from about 1650° to 2000°F and afterburner temperatures ranged from about
1800° to 2100°F. Stack gas oxygen concentrations averaged about 6 percent
with measured CO concentrations of less than 1 ppm to spikes exceeding 100
ppm.
-460-
-------
in
o
Q.
cn
c
to
(O
0)
o
c
•I—
IB
O
O)
cn
-461-
-------
JOHN ZINK LIQUID INJECTION
The John Zink inciner ation test facilltyln Tulsa Oklahoma,
Btu/hr and a maximum of 3
s SMS*
by adding more
Hlgh-mtensny Vortex
burner.
a...
ly prior to the quench pot sprays
I. addition, the OHdK. State ^part^t o, „ eaHh
that a fume Incinerator be •«eddunn| all t est s in vo 9 combustion sys-
Tt^sTlSePras and ^^1^^ SSSTrtWr'.!* a residence tl. of
2 seconds.
The hot combustion gases leaving tne primary "mbustion
carbonate also was added to
neutralization The scrubber and the
pot water systems were blown
pot wate y
feed rates of .4.27 to 4.64 pounds per . ^^^ ^
rates were maintained at 24 to 27 P^cent^ Tests 2 and 3
tained at a pH of 8 duncn?Je^n£'rf ^bbS was about 40 inches of water.
IRe rHe^ehedrqrcChr°PoSt %£%£"* Spring Test 1 and 7 to 8 dunng
Tests 2 and 3.
-462-
-------
LU p;
. 2> DC
1 1 1
z
O
LU Q.
CO
B
LU
-
"C
^
&
» -^
35 "B
Z co
.£
^0
1
CO
I 1
O
1
CL
0
0
D_
o
O)
o
ench Pot Water
a
a.
0
1
S
CO
s
-g
ench Pot Water
a
D
0)
-l->
I/)
>>
to
o
(B
S- *
O) ID
C i-l
•r- O
u m
c
o
C M-
O
•r- I/)
-l-> OJ
O •*->
QJ ••-
•r-j CO
c
•i- en
c:
=3 Q-
cr E
•r- tO
i — I/)
C (B
CO =5
CLTD
O to
CJ O)
^
C T3
•i- C
rs 10
C T3
J= 0)
O CU
O t-
O) -O
Cv)
OJ
3
-463-
-------
SAMPLING AMD ANALYSIS PROTOCOLS
K001-PCP TEST
The following four streams were sampled during the K001-PCP test burn:
Sampling Point
A
B
C
D
Description
K001-PCP waste feed
Rotary kiln ash
Scrubber makeup water
Scrubber blowdown
The process schematic diagram presented in Figure 1 shows the sampling points
for the K001-PCP burn. A total of 1200 Ib of waste was consumed in the test.
Table 1 presents the compounds targeted for each sample. Samples were
prepared and analyzed in accordance with SW-846, 3rd edition. Ash samples
were prepared using TCLP. To optimize quantitation for PCP, a slightly
modified EPA Method 8040 (SW-846) was used for PCP extraction (an isothermal
column oven was maintained at 170°C) and analysis (electron capture detec-
tor).
Pentachlorophenol was selected as the principal organic hazardous con-
stituent (POHC) for stack sampling. The stack sampling point was between the
scrubber-packed tower and the charcoal bed.
K015 TEST
K015:
The following five streams were sampled during incineration tests of
Sample Point
A
B
C
DP
D
Description
K015 waste feed
Scrubber water recycle - Pretest
Scrubber water blowdown
Quench pot water - pretest
Quench pot water
The sample points are also noted in Figure 2. A total of 1600 Ib of K015 was
used in the three tests. No ash was generated during this test. Samples ot
K015 feed, scrubber water, and quench pot water were analyzed for volatile
and semivolatile organics and purgeable and nonpurgeable organic carbon. The
waters were also analyzed for metals. Specific compounds quantified are
identified in Table 1.
The POHC's for this test were benzyl chloride and benzal chloride. The
stack sampling point was between the water separator and the fume incinerator
(Figure 2).
-464-
-------
E
o
> N ID :
8
I *
ffffJil
S-t If o o
,. E Ei Siillll*
vhlm, §1111^-
3zlJ5lll4 gllll^lS
O
UJ
UJ
5?
o
CO
«=c
a.
c . « a • •ilD3
<o &mffloim4a)
-465-
-------
TEST DATA
K001-PCP TEST DATA
Four samples of untreated K001-PCP, three samples of incinerator ash,
and three samples of scrubber blowdown were taken and analyzed. Xylenes and
toluene were found in the feed at 73 to 110 yg/g and 14 to 32 vg/g. respec-
tively The K001-PCP feed samples also contained polynuclear aromatic hydro-
carbons (e.g., anthracene, fluoranthene, and naphthalene) ranging up to
50,000 yg/g (ppm,. Pentachlorophenol concentrations in the feed ranged from
920 to 3000 ppm.
Methylene chloride was detected in one scrubber effluent sample at 61
vg/liter (ppb). No PCP or other semivolatile compounds were detected in any
of the scrubber effluent/influent or ash samples.
The metals analysis showed zinc as the major metal present in these
samples. Zinc ranged from 30 to 64 ppm in the feed and up to 11 ppm in the
ash Barium in the feed ranged from 17 to 30 ppm and concentrations were up
to 74 ppm in the ash. Only barium (up to 0.32 ppm), lead (0.02 ppm in one
sample), and zinc (0.03 ppm in one sample) were detected in the ash Toxicity
Characteristic Leaching Procedure (TCLP) extracts.
No target dioxin/furan analytes were detected in any of the scrubber
effluent/influent, waste feed, or ash samples.
Stack gas data are not currently available.
K015 TEST DATA
Six samples of untreated K015, three samples of quench pot water and
three samples of scrubber water were taken and analyzed. Emission tests were
conducted to determine the DRE.
The K015 feed used in the test burns consisted almost entirely of benzal
chloride (88 to 100%). Benzyl chloride (3500 to 6400 yg/g) and benzotn-
chloride (3700 to 4500 yg/g) were also present. No other semivclatile or-
ganics were detected. This finding is consistent with information supplied
by the generator. No toluene or other volatile organics were detected except
for the following minor constituents in one sample of untreated K015:
Chloromethane (6.0 vg/g)
1,1,1,2-Tetrachloroethane (2.1 vg/g)
Tetrachloroethene (2.6 yg/g)
Toluene concentrations [15 to 59 parts per billion (ppb)] were detected
in each of the three samples of quench water taken during the incineration
tests. Each sample of quench water was found to contain anthracene (14 to
210 ppb), 2,4-dinitrophenol (15 to 44 ppb), phenanthrene (17 to 110 ppb), and
ohenol (12 to 21 ppb). Two samples also contained benzal chloride (66 and 94
ppb), and one contained benzotrichloride (16 ppb). The quench water also
-466-
-------
contained several trace metals. Chromium was found in the range of 4 to 34
parts per million (ppm); nickel, in the range of 2 to 25 ppm; and copper, in
the range of 0.6 to 3.5 ppm. Other metals were present in lower concentra-
tions. No volatile or semi volatile compounds attributable to incineration of
K015 were detected in the scrubber water.,
An emission test indicated a particulate rate of 1.689 Ib/h at a concen-
tration of 0.348 gr/dscf (equivalent to 0.311 gr/dscf at 7 percent oxygen).
Much of this particulate (approximately 62 percent) was found to be sodium
chloride, apparently formed in the scrubber from the interaction of sodium
ions from the sodium carbonate and chloride ions from the HC1 in the combus-
tion gases. Trace metals were present in the particulate. Chromium, esti-
mated at 0.03 percent of the particulate, was the major trace element found.
Hydrogen chloride emissions, determined during one test, were less than
0.9 ppm and the removal was greater than 99 percent.
The designated principal hazardous organic compounds (POHC's), i.e.,
benzal chloride and benzyl chloride, were found in the stack gas stream. The
destruction and removal efficiencies for benzal chloride and benzyl chloride
averaged 99.990 and 99.917 percent, respectively, based on two tests. Chlo-
robenzene was also found in the stack gas, but was below the detectable limit
in the feed (less than 10 ppm); therefore, no ORE calculation was made.
Carbon monoxide emissions were generally quite low (less than 10 ppm)
except for a few spikes (less than 1 minute in duration) when concentrations
exceeded 500 ppm.
CONCLUSIONS
Incineration appears to be an effective method for treatment of both
K015 and K001-PCP as indicated by the two tests. The residual ash and
scrubber water samples contained only trace concentrations of organic or
inorganic compounds. The excessive particulate emissions encountered during
the K015 test could be reduced with a more effective air pollution control
device. The low levels of volatile and semivolatile organics present in the
K001-PCP and K015 scrubber waters may be related to the amount of wastes
processed because somewhat higher organics were found in the tests of commer-
cial incinerator air pollution control device effluents. The relatively low
concentration of benzyl chloride in the feed, compared with the concentration
of benzal chloride, affected the accuracy of the feed input rate for this
compound and thus decreased the accuracy of the ORE determination.
-467-
-------
ASSESSMENT OF VOLATILE ORGANIC AIR EMISSIONS FROM AN
INDUSTRIAL AERATED WASTEWATER TREATMENT TANK
by: Bart Eklund
Radian Corporation
Austin, Texas 78720-1088
David Green
Research Triangle Institute
Research Triangle Park, North Carolina 27709
Dr. Benjamin Blaney and Lisa Brown
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Volatile organics (VOs) may be effectively removed from wastewater prior
to discharge by treatment in activated sludge systems containing aeration _
tanks The relative contributions of air stripping/volatilization, biological
oxidation, and solids sorption to total disappearance of VOs in such systems
have not been extensively characterized. Studies were conducted at a full-
scale treatment facility to determine the relative extent to which specific
compounds are lost to the atmosphere. Direct measurements of air emissions
were made via collection and chemical analysis of off-gases from an aerated
tank of an activated sludge unit. Multiple air emission measurements were
made to determine the total off-gas flow rate, the emission rate of specific
compounds, and the associated spatial variabilities. The emissions were _
compared to the overall volatile organics losses as determined from analysis
of influent and effluent liquid streams. This paper describes these .
measurement techniques and the results of the study.
-468-
-------
INTRODUCTION
The U.S. Environmental Protection Agency (EPA), through its Hazardous
Waste Engineering Research Laboratory (HWERL) and other offices, conducts
projects in support of the U.S. EPA implementation of PL 94-580, "The Resource
Conservation and Recovery Act" and the 1984 Amendments. To adequately protect
human health and the environment, regulations are being considered for air
emissions from aerated and quiescent surface impoundments, tanks, and basins.
Emissions from quiescent impoundments have been characterized to a limited
extent, while those from aerated units have been studied less. In both cases
there is limited information available to determine the magnitude of emissions
from these processes or to model those emissions.
Industrial wastewater treatment (WWT) systems typically include surface
or subsurface aeration units with an active bioculture to remove dissolved
organic compounds prior to discharge or reuse of the water. Aerobic organisms
can use certain organic compounds in dilute concentrations as a food source.
For maximum cost effectiveness, oxygen must be supplied to these organisms by
bubbling air through the wastewater and/or agitating the surface of the
wastewater. A large mass of the aerobic organisms is necessary to achieve
high removal rates, so the biomass, or activated sludge, is normally allowed
to recycle and build up to a concentrated level. Besides providing oxygen,
the aeration system acts to keep this biomass in suspension and provides
mixing.
Wastewater typically requires several hours of residence time within the
aeration unit for the biological oxidation of the dissolved organic compounds
to near completion. If volatile organics (VOs) or semivolatile organic
compounds are present in the wastewater, then these compounds may also be
removed via mass transfer to the atmosphere. A third potential pathway for
organic compound removal is via sorption onto the biomass and ultimate
disposal in the wasted solids. The relative contributions of air
stripping/volatilization, biological oxidation, and solids sorption to total
disappearance of VOs in aerated, activated sludge systems have not been
extensively characterized. This paper describes a study conducted at a full-
scale treatment facility to determine the relative extent to which specific
compounds are lost to the atmosphere. Radian Corporation and Research
Triangle Institute (RTI) jointly performed the field work during August 1987.
DESCRIPTION OF THE TEST SITE
The test site was located at a large petroleum refinery which produces
fuels and chemicals. The site was chosen because its subsurface aeration
pattern appeared to be uniform, and because several volatile compounds were
present at levels over 100 ppb at the influent to the aeration tank. The
wastewater treatment plant treats an average of 23 to 26 million liters of
wastewater per day, primarily from continuous processes. Two separate process
streams are present. The wastewater first goes to skim and surge tanks,
followed by API separators. One process stream then empties directly to the
oxidation tanks, while the second process stream is further treated in another
separator, roughing filters, and primary clarifiers before entering the
oxidation tanks. The wastewater is split between two oxidation tanks, is
-469-
-------
recombined, then split again between two final clarifiers. before finally
being filtered and discharged to a nearby river. Sludge is recycled from the
clarifiers back to the oxidation tanks.
One of the two identical oxidation (aeration) tanks was studied. Each
tank is 36.6 m in diameter, 5.5 m deep, and has a capacity of 5.7 million
liters. Typical flow through the system is 19,000 to 23,000 liters per
minute. The wastewater retention time in the oxidation tanks is typically 8
hours, and the total retention time upstream of the oxidation tanks is about
one day.
Oxidation air from a 260 cubic meters per minute (9,250 cfm) capacity
compressor is split evenly to the two oxidation tanks. The air enters the
tanks through dispersion rings located near the tank bottom that have 2,000
diffusers per tank. The dissolved oxygen in the tanks is monitored and
maintained above 1 mg/L. The biological oxygen demand (BOD) removal is
routinely monitored. The influent values are usually 25-100 mg/L. but may
reach 200 mg/L. The effluent BOD is typically 1-2 mg/L.
EXPERIMENTAL APPROACH
The sampling strategy, sampling techniques, and analytical methods are
summarized below. Further detail is available in the project reports (1,2).
Sampling Strategy
The oxidation tank was divided into five concentric circles, each having
an equal area. Emissions from the tank were measured along a transect line at
the midpoint of each ring. In addition to these ten points, the tank s
centroid and three other points were also sampled as shown in Figure 1.
three additional sampling points were selected to coincide with the *
rings. Seven sampling points (2, 5. 7, 10, 12, 13, and 14 were directly over
the aeration rings. The seven remaining sampling points (1. J. 4-, o. o, y,
and 11) were relatively quiescent.
.vcrna
\
The
Figure 1.
Depiction of sampling locations for oxidation tank.
-470-
-------
An enclosure device, the isolation emission flux chamber, was used to
measure the of£-gas -flow rate and to collect samples of the emit'd Jal Off
gas rate measurements and VO emission rate measurements were made concur-
vTnd^s Si curvl^ ^ '" ^ «"«*-«* - estimated froHhe
itv Of +£*-?*,,*** .°f''.aW*J:i** W3S USed t0 assess the radial ^Patial variabil
ity of the tank emissions in- terms of VO emissions and air flow rate To
(SS^onc ^' t -°W rate— ements and total non-methane hydrocarbons
CTNMHC concentration measurements were made at seven points along the tran-
upon this ^S.tank- ^ on~site analytical data were developed. Based
distinct r£? 1rina5T •»**••. the tank was determined to contain no separate
distinct radial zones of emissions, i.e., emissions were homogeneous! ,
^cond;and .third Campling days were used to characterize the emis-
..
the influent and effluent streams to the oxidation tank. ejected from
Sampling Procedures .
The chosen air sampling approach used an enclosure device, referred to as
was needed. A pump was used to withdraw sample gas from the flux chamber t
the same rate it entered the aerators as indicated by the chamber pressure!
The air flow rate was adjusted until the average chamber pressure, as measured
lL^ m*nometer, ™s Zero. The volumetric flow rate of air through the
chamber was manually recorded from visual readings of a rotameter
• at the exit of the chamber- The
where:
E.R.
C
£ - emission rate of species, i (ug/m2min)j
£ = measured concentration of species i (ppmv converted to
ug/mj);
Q = air flow rate (m3/min) ; and
A = exposed surface area (m2) .
area ri T** ^ & ^at~^^& acrylic cylinder that enclosed an
area of 0.29 m and had an internal volume of 0.130 m3 assuming a 2 5 cm decth
^113-011- ^^ (1A lnCh USed to withdraw the of f-
chamber and
-471-
-------
The sample gas stream passed through a calibrated rotameter with a preci-
sion flow controller. A console was constructed with three rotameters capable
of monitoring a range of off-gas rates from 0 to 80 L/min. All components in
contact with the gas were glass, teflon, or stainless steel. A sampling mani-
fold was used for gas sample collection. Samples for on-site analyses were
collected in precision lock, 100 mL teflon and glass syringes. The samples
were shielded from direct sunlight and analyzed within one hour of collection.
Gas samples for off-site analyses were collected in evacuated, 2.8 liter Summa
polished, stainless steel canisters.
A variety of procedures was used to collect liquid samples. Grab samples
of influent streams were collected from existing valves, and effluent samples
were bailed from the tank at an overflow weir using a teflon bailer. All grab
samples were collected in 40 mL zero headspace septum vials, preserved with
HC1 to a pH below 2. and refrigerated until the time of analysis. Composite
samples of the influent, recycle influent, and effluent were collected over a
six-hour or longer period using a syringe pump that gradually filled two 50 mL
syringes from a slipstream of the process water. The entire sampling appara-
tus was cooled throughout the sampling period. When full, the syringes were
decanted into acidified 40 mL septum vials.
Analytical Procedures
The on-site analyses were used to provide feedback to assess the sampling
strategy. They were limited to gas-phase analyses of the air samples col-
lected in gas-tight syringes from the outlet of the flux chambers. A Shimadzu
Mini-2 gas chromatograph (GC) with a flame ionization detector was used to
determine the concentrations of total non-methane hydrocarbons (TNMHC) and
methane (CH4) in the effluent air samples from the flux chamber.
The off-site analyses included the chemical speciation of the flux cham-
ber air samples collected in stainless steel canisters and chemical speciation
of the various liquid samples. Additional sampling and analyses were per-
formed that are not addressed in this paper.
Chemical speciation of the air and some selected liquid samples for C2 to
C10 hydrocarbons was performed with a Varian 3700 gas chromatograph fitted^
with a flame ionization detector and a photoionization detector. The chemical
species were cryogenically concentrated to increase the sensitivity of the
analyses. Liquid samples were analyzed using a purge-and-trap technique modi-
fied to utilize the cryogenic trap. For both air and liquid samples, total
volatile organics concentrations were obtained as the sum of the species con-
centrations.
Grab samples from the influent and effluent streams were analyzed for
benzene, toluene, and xylenes (BTX) by EPA Method 602. The EPA Method 602 _
results were confirmed by mass spectroscopy using EPA Method 624. A composite
of the oxidation tank influent was analyzed for extractable organic priority
pollutants by EPA Method 625 to establish the background of organics present
in the wastewater. Finally, the volatile suspended solids content of recycle
and effluent sample splits were determined by Standard Method Number 209E.
-472-
-------
SAMPLING AND ANALYSIS RESULTS
Fifteen off-gas rate measurements and thirteen emission rate measurements
were made at the oxidation tank. The results of these measurements are
summarized in Table 1. The experimental coefficients of variability were
34.4% and 62.9% for the off-gas and TNMHC measurements, respectively.
TABLE 1. SUMMARY OF FINDINGS FROM OFF-GAS AND EMISSION RATE MEASUREMENTS
Measurement.
Weighted Average
Total
Off-Gas Rate
(m3/min'm2)
TNMHC* Emission Rate
(ug/m *min)
TNMHC* Emissions
(ug/min)*
Methane Emission Rate
(ug/m min)
Methane Emissions
(ug/min)^
0.119
2410
1,050
2,530,000
1,110,000
$f ""•———— — —-—• •'••.•.—.— —.•.•.,.. —.-.—.-•. . . .. _, r.[ , r_i
..Total non-methane hydrocarbons (excluding oxygenated species).
7r Surface area = 1050 m .
The total air emissions for the major individual compounds that were
detected are given in Table 2. Methane and straight-chain aliphatic com-
pounds account for the majority of the emissions. The total non-methane
hydrocarbon emission rate corresponds to a daily emission of less than 4 kg.
TABLE 2. TOTAL EMISSIONS FOR SELECTED INDIVIDUAL COMPOUNDS
Compound
Average Emission Rate
(ug/m -min)
Total Emissions
(mg/min)
Methane
C-3 VOC
n-Heptane
n-Octane
n— Nonane
n-Decane
n-Undecane
3-Methylheptane
Methyl— cyclohexane
Toluene
Cyclopentane
Isoheptane
Benzene
p,m-Xylene
o-Xylene
Ethylbenzene
TNMHC
1054
35.1
58.2
110
144
124
73.1
47.3
40.9
52.5
5.74
26.4
5.20
11.8
8.96
3.03
2410
1106
36 8
•J *J • \J
61.1
116
151
130
J- ^> W
76 8
f \J a \J
49 7
~ ./ • /
43 0
~-~f • W
55 1
•J ~f • X
6 0^
\j • \j*j
27 7
^ / » /
•5 46
•J • "W
12 4
J. £j • *T
O A1
y • *TX
3.18
2530
-473-
-------
The average measured air flow converts to 125 m3/min of off-gassing from
the entire tank. The spatial variability in the off-gas rate was +12%, but
the accuracy of the sampling technique has not been assessed. A total air
flow of 141 m3/min was estimated from the fan curve for the compressor in
service. The measured value is 11% less than this estimate.
The average concentrations of selected species in the influent and
effluent streams of the tank are given in Table 3. The total mass of
individual compounds entering the oxidation tank was determined by multiplying
the average concentration in the wastewater by the average flow rate for each
of the three influent streams for the time period when air sampling was
conducted. These data are shown in Table 4. The liquid concentration data
were from the 2-3 samples per influent stream analyzed using the Varian GC.
The estimated percentage of each compound that was volatilized is also shown.
The values agree well with other published data for aromatic compounds (6).
TABLE 3. AVERAGE LIQUID-PHASE CONCENTRATIONS FOR WASTEWATER STREAMS
Compound
Benzene
Toluene
m»p-Xylene
o-Xylene
n-Nonane
n-Decane
n-Undecane
Methylcyclohexane
TNMHC
4F
Influent
Stream
(ug/L)
440
274
466
209
122
105
107
114
5,010
2B
Influent
Stream
(ug/L)
9.67
19.0
49.1
25.1
11.5
23.4
24.2
BDL
663
Recycle
Influent
Stream
(ug/L)
0.21
0.26
0.42
0.35
0.94
BDL
0.58
0.26
41.0
Effluent
from Tank
(ug/L)
BDL*
— - V
BDL*
BDL
BDLff
™7t
BDL'
BDLT
^Effluent was analyzed using a separate analytical protocol and analytical
.system.
*BDL = below detection limit (0.3-3.8 ug/L).
tDetection limit = 10 ug/L.
TABLE 4. INFLUENT WASTEWATER MASS LOADINGS AND ESTIMATED VOLATILE LOSSES
Compound
Total Mass Loading
of Influent Wastewater
(g/min)
Estimated Volatile Losses
(% of Influent Cone.)
Benzene
Toluene
p,m-Xylene
o-Xylene
n-Nonane
n-Decane
n-Undecane
Methylcyclohexane
TNMHC
2.50
1.59
2.76
1.24
0.72
0.66
0.68
0.64
30.1
0.2
3.5
0.4
0.8
21
20
11 •
6.7
8.4
-474-
-------
The analysis of the influent streams to the oxidation tank indicated that
the recycle sludge had negligible VOs (i.e., no loss of 70s via sorption to
the biomass). Likewise, the effluent liquid from the tank had no detectable
BTX; indicating that essentially complete removal of these compounds occurred
in the oxidation tank.
CONCLUSIONS
The tank was found to contain a single zone of emissions (i.e., no clear
trends in spatial variability were determined). Overall, the air emissions
from the tank were low, both on an absolute basis and as a percentage of the
influent VOs. The ratio of methane to TNMHC emissions was about 1:2. The
source of the methane has not been determined, though an anaerobic layer at
the bottom of the tank is one possibility. Relatively large methane emissions
have also been measured at another aerated tank (3). The flux chamber sam-
pling method exhibited good precision and appears to be accurate based on com-
parisons of total air flow to fan curve data.
NOTICE
The information in this document has been funded wholly or in part by the
U. S. EPA under contract to Radian Corporation and Research Triangle Insti-
tute. It has been subject to the Agency's peer and administrative review, and
it has been approved for publication. Mention of tradenames or commercial
products does not constitute an endorsement or recommendation for use.
REFERENCES
1. Eklund, B., et al. Results of Radian's air emission measurement at plant
A3's wastewater treatment system. Draft Report to U. S. EPA/HWERL
October 9, 1987.
2. Green, D. Field assessment of air emissions from an industrial waste-
water treatment system at plant A-3. Draft Report to U.S. EPA/HWERL
February 18, 1988.
3.
4.
5.
6.
Green, D. and B. Eklund. Field assessment of the fate of organics in
aerated waste treatment systems. Paper presented at the 13th Annual EP..
Symposium on Land Disposal, Remedial Action, Incineration and Treatment
of Hazardous Waste, May 1987.
Eklund, B. M., W. D. Balfour, and C. E. Schmidt. Measurement of fugitive
volatile organic emission rate. Environmental Progress. 4s 3, 1985.
Wetherold, R. G., B. M. Eklund, B. L. Blaney, and S. A. Thornloe.
Assessment of volatile organic emissions from a petroleum refinery
landtreatment site.. Paper presented at the Hazardous Materials Control
Research Institute's 3rd National Conference on Hazardous Wastes and
Hazardous Materials, March 1986.
\
Namkung, E. and B. E. Rittmann. Estimating VOC emissions from publicly
owned treatment work. Journal WPCF, 59: 7, July 1987.
-475-
-------
CALIFORNIA DEPARTMENT OF HEALTH SERVICES/EPA
STATUS UPDATE OF THE CALIFORNIA LAND DISPOSAL RESTRICTIONS
AND IMPACT/IMPLEMENTATION PROGRAM
Robert Ludwig and Benjamin Fries
California Department of Health Services
Sacramento, CA 95814
ABSTRACT
The California Department of Health Services (DHS), in co-
operation with the U.S. EPA, is conducting demonstrations of
alternative treatment technologies and studies on hazardous waste
management. The overall objective of the California Waste Manage-
ment Program is to reduce the amount and eventually eliminate land
disposal of untreated hazardous wastes. Strategies being studied
include source reduction, recycling, and treatment.
This paper provides an update of EPA-funded projects illustra-
ting California's Waste Reduction Program. Specific studies
include: 1) Research, Development, and Demonstration Projects
(RD&D); 2) Waste Management Information Transfer; 3) Metal
Finishing Waste Audit; 4) Waste Stream Information Accumulation and
Analysis; and 5) Source Reduction Resource Partnership.
BACKGROUND
In December, 1982, California's Land Disposal Restrictions
Program was initiated by DHS which implemented a regulatory program
to phase out land disposal.of specific hazardous wastes. This
program provided a schedule of land disposal restrictions for
certain hazardous wastes, the California List, which includes
liquids with high levels of cyanides, metals, PCB's, halogenated
organic compounds, or a pH of less than 2.
In order to assist and move industry towards alternatives to
land disposal, DHS initiated the Waste Reduction Program_in July,
1984. The various program elements are designed to provide tech-
nical assistance, information/technology transfer, economic and
regulatory incentives.
Part of California's Waste Reduction Program has involved a
Cooperative Agreement with the EPA entitled "California Land
Disposal Restrictions Impact and Implementation Evaluation
Program." This 3-year project started in December, 1985, has
provided DHS with staff and contractual support to effectively
evaluate new waste generation, reduction, recycling, treatment, and
destruction technologies capable of processing hazardous wastes.
-476-
-------
INTRODUCTION
The purpose of this paper is to provide a status update on the
EPA-funded projects. Program areas include: 1) RD&D; 2) Waste
Management Information Transfer; 3) Metal Finishing Waste Audit;
4) Waste Stream Information Accumulation Analysis; and 5) Source
Reduction Research Partnership. A summary of these projects is
presented in Table 1.
RESEARCH, DEMONSTRATION, AND DEVELOPMENT PROJECTS
The status of 4 RD&D projects for the treatment or on-site
recycling of hazardous wastes is presented below.
Aerobic Composting System for Pesticide Rinsewaters
In December, 1987, California Agricultural Research set up an
Aerobic Composting System at the Chemical Waste Management Inc.
facility in Kettleman City, CA, to determine the feasibility of
treating pesticide rinsewaters. Two 25-gallon additions of a
Simizine/carbofuran-water mixture have been applied to the 14' x 5"
x 30' compost system consisting of cotton gin wastes and chicken
manure. A similar compost system receiving only water is serving
as a control. Both compost systems are being maintained aerobi-
cally via a forced air perforated pipe network.
Preliminary results from the first application indicate a
degradation of carbofuran from 50-1 mg/L over a 7 day period.
Simizine degradation from 70-3 mg/L occurred over 40 days.
Temperatures in the compost system ranged from 11-67° with a mean
and standard deviation of 48 + 10°, (n=40).
A second application of carbofuran/simizine was then added to
the same treated compost pile. The pesticides did not degrade as
quickly as the first application and the temperature in the system
was considerably lower. Carbofuran degradation occurred from 87-50
mg/L after 21 days while simizine degraded after 21 days from 50-8
mg/L. Temperatures in the system ranged from 6-18° C with a mean
and standard deviation of 8 + 6° C, (n=21).
Chlorinated Solvent Adsorption Using Ambersorb XE-340
Woodward-Clyde Consultants are currently evaluating Ambersorb
XE-340, a synthetic carbonaceous polymer as an alternative to
activated carbon, for the removal of chlorinated solvents from
contaminated groundwaters. In addition, an evaluation of adsorbent
regeneration within the column utilizing steam will be performed.
Two resin columns (10 x 150 cm) containing 11 liters of
Ambersorb XE-340 adsorbent, are being operated in parallel. The 2
contaminants in the feed water are trichloroethylene (TCE) at 10.7
ppm and 1,1,1-trichloroethane (TCA) at 0.58 ppm. Complete TCE
-477-
-------
Cn
§
c2
CD
Q
P4
§
H
q"i
Q
pa
CO
H
3
rQ
3
g
o
w.
CO
H
3
tx
*~j
EH
CO
*33
EH
CO
g
1
Q
i
•H 01
jj
01 4J
0) 01
ft 0
|._i rt
O O
o
O T!
•H a)
•p -p
O M
-H Q)
rH id
^
JJ .
treatmeni
rinseates
r-H
Biologica:
Pesticide
01 0)
a) id
O 01
i4< rE< Q)
H • Q)
Id H H T3
id o id J3
3 -* > id
H n 0
(d 1 g W Q) U
W X fc EH
01
• O
C -H
O 1 (3
•H 42 id
-P P 01 U>
id 01 p M
M O 0
id fi d)
ft o p • O
O ft -P
01 o id P
>i01 M H
43 "d -P O
0< ri! 01 01
• 01
(1) JH
I>s (tf
vS
'C *rH
(0 fij
T^
o •
O M
•
a)
cH
id
H
•H
id
id
tJ
o
ft
0)
01
a)
CM -P
o -P id
CM (3 Q)
ffi 0) 01
t>-P-H
p id M
0)
m M 01
O -P Q)
•d
o o'o
-H 1-1 -r\
4J 4J
Cd g 01
P -l -rl
EH C3 rH
. . Q)
> 01 -P
f3 -rl J3
pr"| ^ «|-J
id !3£
• Q
•P >i
ft - |d
<0 O M
Q P !S
C 4->
•H M
4-> O •
01 ftCO
Q) Q) CO
| 1 |V] |
CO
r?-d*Q)
O 0) i-H
id Q)
1 r \ *w
M o ro
i
i -P
i -rH id
•d -d Q)
(D TS JH
-P id -P
p
O Ti Q)
ft C 01 •
01 id id >i
O O
fi t tt t i *"J
o o o a)
•P J3-H
C 01 -H O
OP -H
•rH 42 O C
> (U-rH 0)
W 42 -P g
01
rH (D
id -P
g 01 01
P id a)
4-> C tP S3
•rH $3 Id •
^ g -H >i 01
S3 P S3 O Q)
O H -H T! '
•H r«3£] i — 1 tJ*"H
-p -P S3 M
(d • O -H O
SH M ft $3 P
Q) O -H rH
C 4-> -P id M rt
O S
•d
o
g
•H
•P
C
o
o
-478-
-------
g
*J
OH
O
1
o
REDUCTI
H
EH
CO
g.
CO
i fd
rH CQ .
(d c
> >t O
0) 5H-H
a) .p
c« C 0
•H 3
C &
(d
C •
td CQ
^
0) -H
4_) -i-i
CQ (d
(d c
15 SM •
0) CQ
rH Hp rH
•H rH -H
O (d O
•
CQ C
1 0) O
> PS CQ
G rH
H rH 0)
(d s
13 4J
C 4J
>i a) td
J-l fn
0) O •
C r) fM
W-H g
C -P
O M
-H O
-P ft
O Q)
QJ fr4
rH
rH • CO
O CJ^CO
0 G 1
•H CO
(d o
-P tr> a)
(d C 3
DOTS
4J
0) • rJ
a >t o
•p ^-i ft
•P 0)
r< CQ rH
O 3
IH ""cj t?j
>1-H rH
*t3 (d
•P C fi
CQ -H fd
X! g
-P CQ
•H -H -p
T3 C-H
3 -H -3
(d rH
(d 0) 0)
12 g CO
.
K*lTj •
•a c CQ
3
CO &i CQ
f"i jrt
4J-H >
-H 4->
'O (d CT1
r^ I r-*
fJ r~1 >_|
rt! ft-H
Q) H CQ
4-> (d-H
CQ 4-> C
(d
ci li
o o
•H C ft
-P-H 0)
O PH
O CQ
rH -H • CO
rH CQ CQ CO
O >i CQ 1
O rH 0) 0
jrt 1 1 (_|
(d C CTI
-p i Q)
O rH fi >,
5-1 1
O O (N
M rH rH
fd f"*i fd
-P CO)
fd C -H 3
Q -H fe ^3
1 CQ •
O Q) i
(dO 4J
O CQ C C
•H O3
O
-P -P O
C fd 3 vo
fd C Tf
3 0) Q) C
CQ
C5
0)
C >
0 rH
•H O
-P CQ
3 T3
73 0)
Q) 4->
M • fd
CQ C
0) C -H
O O r4
rH-H O
3 -P rH
O P! <~{
CO O U
(
C rH .
O 0)
•H C
-p -p m
O 5-1 -rH
3 fd *-^ o
ro CM CM S
O PH
PH Xi PH ^*l
0 CO -P
a) 5n ^- fd
o fd «'
5n QJ ft
3 CQ -H •
O 0) X! rH
CO PH CQ Q
-479-
-------
saturation (inf. equaling eff. concentration) occurred after 7,000
bed volumes of water at a feed rate of 1.64 LPM (9.2 bed
volumes/hour). The final TCE loading on the adsorbent was about
40% of the predicted equilibrium adsorption isotherm. Complete TCA
saturation occurred after treatment of 4,500 bed volumes of water
with the final TCA loading about 13% of the loading predicted by
the equilibrium adsorption isotherm.
The first indication of TCA breakthrough (eff. greater than
0.01 ppm) occurred after 800-1,000 bed volumes at a feed rate of
0.44 L/minute (2.4 bed volumes/hour). It appears that TCE break-
through was just about to begin after 1,200 bed volumes of water.
It was estimated that complete TCA and TCE saturation would occur
after an additional 4,000 bed volumes over a 3 month period. A
flow rate of 1.65 LPM (5 GPM) was determined to be acceptable for
future studies to provide an acceptable loading rate for TCE
removal. Furthermore, there appeared little process advantage to a
lower flow rate which would be too low for practical scale-up.
The next step of the study will involve the evaluation of a
steam generator for the regeneration process of the Ambersorb
XE-340 within the columns.
UV/Hvdrogen Peroxide Treatment For Destruction of Pesticide Waste
The Dept. of Environmental Toxicology of UC Davis, has com-
pleted an evaluation of a small ultraviolet (UV) and oxidation
system, Perox Pure Model SQ, manufactured by Peroxidation Systems,
Inc., to treat a broad range of pesticides and study variables
related to the treatment process^. Specific compounds and pesti-
cides treated were m-xylene, Captan, Pentachloronitrobenzene
(PCNB), and Propazine.
The unit has a reactor and a reservoir capacity of about 20
liters each. The reactor/oxidation chamber is a stainless steel
cylinder (24" x 8.6") and receives UV light by an axially mounted
high pressure mercury vapor lamp delivering radiation in the
190-400 nanometer range. For the majority of experiments, H2O2 and
a proprietary catalyst were fed continuously into the reactor.
Wastewater was circulated at 12 LPM through the system and returned
to the holding reservoir. The residence time in the reactor was
1.67 minutes.
Some of the test results are presented in Table 2-Pesticide
Degradation Utilizing a UV/H2O2 System. The unit rapidly degraded
all of the compounds in optically clear water. It is important to
note that the rates of breakdown of these compounds were not
significantly dependent on the presence of H2O2 or the catalysts.
The m-xylene was used as a test chemical since xylenes are
common components in emulsifiable concentrate pesticide formula-
-480-
-------
tions. Two different concentrations of m-xylene were degraded to
about the same level after 37-45 minutes w/ or w/o H2O2 or cata-
lyst. Due to the low water solubility of Captan, a carrier solvent
(5% methanol v/v) was added to the system. Captan breakdown from
10-0.03 ppm occurred after 16 minutes w/ or w/o H2O2 or catalysts.
Higher Captan concentrations of 60 and 125 ppm (using a commer-
cially available wetable powder formulation which made a cloudy
solution for the 125 ppm solution) resulted in final concentrations
of 0.61 and 0.01 ppm after 30 and 89 minutes. The slower degrada-
tion rate of the latter solution may be caused by the cloudiness of
the solution.
TABLE 2,
Chemical
PESTICIDE DEGRADATION UTILIZING A UV/H2O2 SYSTEM
Initial H2O2 Catalyst Final Time of
Cone. Cone. Cone. Operation
(PPm)
m-xylene
m-xylene
Captan
Captan
Captan
Captan WP*
PCNB
PCNB
Propazine
Propazine
23
50
10
10
60
125
2
2
7
12
0
115
0
115
115
115
0
115
0 ,
115
None
None
None
Yes
Yes
None
None
Yes
, None
None
0.37
0.50
0.03
0.03
0.61
0.01
0.005
0.005
0.04
0.03
37
45
16
16
30
89
15
28
30
30
Water solubility of PCNB is extremely low, 0.44 ppm, so most
of the experiments were conducted at 2 ppm with a 5% solution of
isopropanol as a carrier solvent. PCNB was rapidly degraded in
clear water by the UV light without H2O2 or any catalysts. UV
light_only proved to be the best condition for the degradation of
PCNB in aqueous and isopropanol solutions. PCNB broke down after
15.5 minutes from 2-0.01 ppm under these conditions.
Propazine also had low water solubility (3 ppm) but solutions
of 7.4 & 12.6 ppm were generated using 5% isopropanol. H2O2 had no
effect on the degradation rate of propazine and it took 30 minutes
to degrade the 2 solutions to 0.04 ppm.
In conclusion, the unit was found to be applicable for treat-
ment of low levels of pesticides and other pollutants in clear
water was not shown to be effective use for degrading high
concentrations in cloudy solutions.
Circulating Bed Combustion Of Spent Potliners
Ogden Environmental Services Inc., is conducting a series of
20 parametric combustion tests in a small-scale, 2-inch diameter
-481-.
-------
spouted-bed combustor system to determine the effectiveness of
additives in preventing the agglomeration of spent potliner types.
Aluminum spent potliners (SPL) are a solid waste by-product of
aluminum smelters that contain soluble species of cyanides and
fluorides. Over 2 million tons of SPL are stockpiled in the U.S.
awaiting development of an effective treatment technology.
This spouting bed combustor (SBC) constitutes an intermediate
testing phase between muffle furnace and the pilot-plant circula-
ting bed combustor (CBC) unit. The spouting bed unit will be used
to study the fluidization and agglomeration behavior of potential
waste feeds and to determine the performance of agglomeration-
inhibiting additives during fluidized bed combustion of wastes with
low melting temperatures. For the proposed series of tests, a test
matrix was established with 1) SPL type; 2) additive type; 3) addi-
tive ratio; and 4) initial combustor temperature. Also, 1 test was
varied by using sand instead of alumina as an inert bed material.
Tests showed that potliners produced by different aluminum
manufacturers behaved differently in the SBC. The 3 additives
produced agglomeration temperatures ranging from 825-900° with the
4 different SPL types. An additive ratio of 20% appeared to be the
minimum, while a maximum of 40% increased the agglomeration
temperature to 930° C.
Leachability tests were conducted on the ash samples for
cyanide and fluoride. Cyanide levels were reduced from 2,000-0.76
and 0.06 mg/L, using the EPTOX and the California Waste Extraction
Test. Fluoride leachability was effectively reduced several orders
of magnitude by the addition of 15 wt percent CaCl2 powder to the
SPL ash. The resultant blend of 35 mg/L was below the levels
required by the EPA and DHS.
Overall, the SPG test program demonstrated that several
different types of additives can be used to control agglomeration,
with distinct differences in resultant agglomeration temperatures.
Potliners from different aluminum producers behaved differently,
but agglomeration was controlled with sufficient quantities of
additives.
WASTE MANAGEMENT INFOPvMATION TRANSFER
Oil Waste Management Alternatives Symposium
A report and symposium on oil waste management was completed
by Energy and Environmental Research Corporation. _The study and
symposium seek to interest, encourage, and assist industries with
improving their management of oil wastes. It identified effective
alternatives that could lead to a reduction in oil generated and/or
increase the recyclability of oil waste.
-482-
-------
The report focused on alternatives that can be implemented by
industry and recommendations that can be made by the DHS to genera-
tors. Types of oil wastes resulting from oil usage included
lubricants, coolants, cutting and machining oils, and hydraulic
fluids. PCB oil or oils from other major process residues result-
ing from petroleum refining, production of oil products, petro-
chemical manufacture, and other major process residues where oil is
a raw material or a product were not included. The report and
symposium also presented technological, economic, and regulatory
aspects.
An in-depth discussion of a used oil re-refinery system was
also presented. The system demonstrates hydro-refining technology
at a full-scale production capacity. Hydro-refining is an estab-
lished technology in crude oil refining, but has not been utilized
in used oil re-refining. This information can establish whether
the system has economic viability over an extended time period. It
could possibly lead to wide-spread oil collection and
hydro-refining as an ongoing waste management technology.
The results of the report were presented to industries,
regulatory agencies, and environmental consultants, and the public
at 2 two-day technical symposia in April, 1988. Total attendance
was around 700 for the two seminars.
METAL FINISHING WASTE AUDIT PROJECT
A waste audit study for the metal finishing industry is now
underway by PRC Environmental Management, Inc. The study will
identify waste reduction technologies available to the industry and
develop a waste reduction audit protocol that can be used by metal
finishers to assess their own waste reduction opportunities.
To meet the objectives, PRC is performing waste audits at 3
small to medium_sized metal finishing plants in the San Francisco
Bay area. Initial visits have been conducted, background informa-
tion is now being reviewed, and areas where waste reduction
technologies may be implemented are being identified. A second
visit will be conducted to fill in data gaps identified during the
review of the background and to gain more detailed information
about process requirements, work procedures, and hazardous waste
generation.
PRC will then identify and evaluate the applicability of
implementing various waste reduction technologies into the plant
manufacturing processes. The costs of implementing these technol-
ogies will be compared with the potential savings in waste reduc-
tion. The findings of each audit will be summarized in a report
and submitted to each plant. After review by plant personnel, a
third visit will be made to meet and discuss management's reactions
to findings and recommendations of the audit. At that time, PRC
will find out if the company is planning on implementing the
-483-
-------
recommendations made in the report and if not, why such a decision
was made.
After completion of individual waste audits, a general report
will be prepared containing a recommended standard format, or audit
protocol, for performing waste reduction audits at metal finishing
plants. This protocol will serve as a guide for self audits by
plant managers to undertake at their own plants.
WASTE STREAM INFORMATION ACCUMULATION AND ANALYSIS PROJECT
In order to better assess the potential for waste reduction,
treatment, storage, disposal, and recycling facility needs, and
compliance with the land disposal restrictions, DHS is setting up a
program to collect and analyze data accumulated from the 1987
Biennial Generator and Annual Facility Reports. The information
will provide DHS and industry with a better idea of compositions
and volumes of hazardous wastes, remaining landfill capacities,
existing treatment and storage capacities, and the potential for
waste minimization activities in the State.
One project of this program would involve an update and
revision to the Handbook of Industrial Waste Compositions in
California. The Handbook, last prepared in 1978, will organize
waste compositions and volumes according to Standard Industrial
Classification (SIC) codes, RCRA waste codes, and California Waste
Codes (CWC's) and present this in a final report.
A second project will determine the treatment, storage, and
disposal capacity of on- and off-site facilities by county. Total
tonnages of wastes handled and the total capacity in tons for each
of the handling methods employed will be collected and analyzed. A
final report summarizing these results will be produced.
A third project will provide information necessary for the
development of treatment standards. Information related to quanti-
ties of hazardous wastes generated by RCRA and/or CWCs, and the
individual or sequential handling methods employed in the manage-
ment of those wastes will be collected and analyzed. This data
will also be used in conjunction with the component information to
establish a baseline for assessing the impact of future disposal
restrictions on resource recovery, product substitution, and other
waste minimization activities.
SOURCE REDUCTION PROJECT
A 2-year project entitled "Quantification of Halogenated
Solvent Waste Reduction Potential" was started by the Source
Reduction Resource Partnership (SRRP), a joint venture of the
Metropolitan Water District of Southern California (MET) and the
Environmental Defense Fund (EOF). The primary aim of the study is
to estimate the potential for source reduction of chlorinated
-484-
-------
solvents of the principal solvent using industries in the 6 county
area served by the MET.
Profiles on 14 high solvent using industries will be developed
by gathering information from published literature sources and
interviews with industry representatives. The profile will include
detailed descriptions of chlorinated solvent use, waste generation,
and waste management. Some of the industries to be investigated
include vapor degreasing, electronics, adhesives, textile process-
ing, paint stripping, and pesticides.
An initial list of source reduction options based on the
industry profile will be generated for each of the industries.
Based on the industry profiles and the source reduction options, a
survey for field visits and telephone interviews will be developed
and around 100 firms in the high solvent using industries will be
visited and interviewed. The survey will then be updated and
additional source reduction options will be documented and consi-
dered. An additional 100-200 solvent using firms will be surveyed
with the results of all interviews and site visits being placed on
a computerized data base. The source reduction options will then
be analyzed and an estimate of costs associated with the implemen-
tation of these options will be made. These cost estimates will
take into account the variation in the size of the industry.
Finally, institutional and other issues related to waste
management and the regulatory environment which may affect the
future implementation of source reduction practices will be ana-
lyzed. A partial list of issues to be examined include military
specifications, small quantity generators, on-site recycling
impediments, and assistance from large to small and medium firms in
waste reduction.
SUMMARY
A better assessment of the waste reduction potential can be
made after the collection and analysis of data related to the types
and quantities of hazardous wastes produced and those methods
available to handle the wastes. The information obtained from
these EPA-funded projects will contribute to the DHS' more effec-
tive implementation of California's Waste Reduction Program.
REFERENCES
1. Winterlin, Wray, and Peterson, D. UV/hydrogen Peroxide
Treatment for Destruction of Pesticide Laden Waste. Sept. 30,
1987. Dept. of Env. Tox., U. California, Davis. Report
submitted to CA Dept. of Health Services, Toxic Subs. Cont.
Div., Alt. Tech., Sacramento, CA.
-485-
-------
THE EPA MANUAL FOR WASTE MINIMIZATION OPPORTUNITY
ASSESSMENTS
by: Gregory A. Lorton, P.E., Jacobs Engineering Group, Pasadena, California
Carl H. Fromm, P.E., Jacobs Engineering Group, Pasadena, California
Harry M. Freeman, US EPA, Hazardous Waste Engineering Research Laboratory,
Cincinnati, Ohio
Waste minimization (WM) is fast gaining recognition as a means of contending with the
nation's hazardous waste problem and other forms of environmental pollution. Opportunities exist
for waste minimization throughout industry and government. The waste minimization assessment
procedure described in this paper offers a means of determining a facility's waste situation, and
identifying and evaluating potential viable options for reducing waste. This procedure has been
developed by the authors for the US EPA and will be published in the EPA Manual for Waste
Minimization Opportunity Assessments.
WHAT IS WASTE MINIMIZATION?
Waste minimization is comprised of source reduction and recycling. Source reduction is
defined as any activity that reduces or eliminates the generation of waste at the source, usually
within a process. Recycling is defined as the recovery and/or reuse of what would otherwise be a
waste material. Figure 1 illustrates the various categories of waste minimization techniques.
The emphasis in this paper is on "hazardous waste". However, all waste streams must be
considered when conducting an assessment. This includes air emissions, wastewater, and non-
hazardous solid waste. The transfer of pollutants from one medium to another is not waste
minimization.
INCENTIVES
There are a variety of incentives for minimizing wastes. These include the following
• Attractive economics (including reducing waste treatment and disposal costs, and savings in
raw material costs)
• Increasing regulations (including landfill disposal regulations, reporting requirements, and
permitting requirements for waste treatment)
* Reduced liability (including liability for environmental problems and workplace safety)
• Improved public image and environmental concern
The economic performance of WM projects has been enhanced in recent years by the
dramatically increasing costs of waste disposal. Environmental regulations, especially RCRA
(Resource Conservation and Recovery Act), have had a major effect on treatment and disposal
costs.
-486-
-------
THE WASTE MINIMIZATION ASSESSMENT PROCEDURE
The waste minimization assessment procedure presented here is a systematic framework that
can be used by a facility's own employees to identify WM opportunities. As a structured program,
it provides intermediate milestones and a step-by-step procedure to (1) understand the facility's
wastes and processes, (2) identify options for reducing waste, and (3) determine which of the
options are technically and economically feasible to justify implementation. On the other hand, the
procedure should be modified to meet the specific needs of the individual company. As such, this
manual should be viewed as a source of ideas and concepts, rather than a rigorous prescription of
how to do assessments.
Figure 2 illustrates the WM assessment procedure. The WM assessment procedure is one part
of a larger waste minimization program, which is required of hazardous waste generators. Careful
planning and organization precedes the assessment itself. The assessment procedure can be split
into two major phases:
• Assessment phase (collect information, and identify and screen potential WM options)
• Feasibility analysis step (technical and economic evaluation of the options)
Implementation of the recommended options follows the assessment The WM program should be
viewed as a continuing program, rather than a one-time effort.
PLANNING AND ORGANIZATION
Careful planning and organization is necessary to bring about a successful WM program. To
start the program and maintain momentum and control, it is necessary to obtain management
commitment. The program should set general goals by which to measure its effectiveness.
Selecting a good program staff is critical to the ultimate success of the program. Since the program
is a project organization within the company, a task force provides an effective way of carrying out
the program. Table 1 describes important considerations involved in planning and organizing the
program.
ASSESSMENT PHASE
The assessment serves to identify the best options for minimizing waste through a thorough
understanding of the waste-generating processes, waste streams, and operating procedures.
Therefore, the assessment task force's first major tasks are to collect information about the
facility's waste streams, processes, and operations.
Collecting and compiling facility information
Information about the facility's waste streams can come from a variety of sources, such as
hazardous waste manifests, biennial reports, environmental audits, emission inventories, waste
assays, and permits. Mass balances should be developed for each of the important waste-
generating operations to identify sources and gain a better understanding of the wastes' origins.
Collecting waste stream data and constructing mass balances will create a basis by which the
assessment task can track the flow and characteristics of the waste streams over time. This will be
useful in identifying trends in waste generation and will also be critical in the task of measuring the
performance of implemented WM options later. The result of this activity is a catalog of waste
streams that provides a description of each waste, including quantities, frequency of discharge,
composition, cost of management, and other important information.
-407-
-------
WASTE MINIMIZATION TECHNIQUES
RECYCLING
(ONSITE AND OFFSITE)
SOURCE REDUCTION
USE AND REUSE
PRODUCT CHANCES
SOURCE CONTROL
- Processed for
resource recovery
- Proceuedaia
by-product
- Return to original process
- Raw material substitute
for another process
. Product mbaiBiuon
• PlUUUCt CCCSCrVllMi
. Caanfa kt product
composition
TECHNOLOGY
CHANGES
GOOD OPERATING
PRACTICES
INPUT MATERIAL
CHANGES
Proaxfetnlmeuuie*
Loaprevcrdm
practice*
Waste arena segregitlao
piping, oc
Ityoat chi&gec
Chmgca in opetitlooid
PrtxtadlcG BcbeduUag
Figure 1. Waste Minimization Techniques
PLANNING AND ORGANIZATION
Set ovenll
program go«l«
AHSMneiTt OTgffTlly^Mf"?
ASSESSMENT PHASE
• Collect proceu tat ftclBy d«l«
• PrioritaJ mi jctect tactsaeniOlfO*
• Select peopto for tliMtmrra «ami
• Review dm and imped 3ilr
• GocnUooptlota
* Screen mdaelectopilom for further »mdy
AMeumeot report of
•elected option ir
Select now
ujeumeat tiigets
mdnevtlurie
previous options
FEASIBILITY ANALYSIS PHASE
• Technical evafeution
• Economic evaluation
• Select option* for implementation
Final report. Including
recommended optical
IMPLEMENTATION
i justify project! and obtain fending
i Installation (equ^xoenl)
> Imptamitmcn (procedure)
• Evaluate pafamance
Repeat the process
Successfully Implemented
waste minimization projects
Figure 2. The Waste Minimization Assessment Procedure
-488-
-------
TABLE 1. Planning and Organization Activities Summary
SETTING UP THE PROGRAM
Get management commitment to:
• Establish waste minimization as a company goal.
• Establish a waste minimization program to meet this goal.
• Give authority to the program task force to implement this program.
Set overall goals for the program. These goals should be:
ACCEPTABLE to those who will work to achieve them.
FLEXIBLE to adapt to changing requirements.
MEASURABLE over time.
MOTIVATIONAL.
SUITABLE to the overall corporate goals.
UNDERSTANDABLE. . .
ACHIEVABLE with a practical level of effort.
STAFFING THE PROGRAM TASK FORCE
Find a "cause champion", with the following attributes:
Familiar with the facility, its production processes, and its waste management operations.
Familiar with the people.
Familiar with quality control requirements.
Good rapport with management
Familiar with new production and waste management technology.
Familiar with WM principles and techniques, and environmental regulations.
Aggressive managerial style.
Get people who know the facility, processes, and procedures.
Get people from the affected departments or groups.
• Production.
• Facilities/Maintenance.
• Process Engineering.
• Quality Control.
• Environmental.
• Research and Development
• Safety/Health.
• Marketing/Client Relations.
• Purchasing.
• Material Control/Inventory.
• Legal.
• Finance/Accounting.
• Information Systems.
GETTING COMPANY-WIDE COMMITMENT
Incorporate the company's WM goals into departmental goals.
Solicit employee cooperation and participation.
Develop incentives and/or awards for managers and employees.
-489-
-------
In addition to data about waste streams, other information is needed to fully understand the
facility's operations. This includes the following items:
• Process, equipment, and facility design information
• Environmental reports, assays, manifests, documents, and permits
• Raw material and production information
• Operating cost information
* Policy and organizational information
Prioritizing and selecting waste streams to assess
Ideally assessments should be conducted on all of the waste-generating operations in a plant.
However in larger plants this often is not practical, considering the limited resources (money,
time and expertise) available. In this case, the assessment program task force should prioritize the
streams. Important criteria to consider in prioritizing waste streams and/or facility areas to assess
include the following:
Compliance with current and future environmental regulations
Disposal cost and/or quantity of the waste
Hazardous nature of the waste, and other safety considerations
Potential for (and ease of) minimization
Potential for removing production or waste treatment bottlenecks
Available budget and expertise for the waste minimization assessment program
A practical consideration in selecting waste streams for the first assessment is to find those that
can be reduced with a good likelihood of success. A successfully implemented WM project will
insure the acceptance of further waste minimization efforts within the organization.
Select assessment team members
The assessment team must include people who are familiar with the area of the facility to be
assessed. Including first line operators and production supervisors is recommended. /I hese
people may or may not already be on the assessment program task force. (In a large facility, the
task force should have a broad understanding of the facility's operations, while the assessment
team should have a specific understanding of the area to be assessed.) It may be advisable to
include people from other parts of the facility that regularly interact with the area to be assessed.
$ite Inspection
Although collecting and reviewing data is important in the assessment, the assessment team
must be familiar with the actual operation at the site. To do this requires that the assessment team
visit the site during the various stages or cycles of an operation. If all of the assessment team
members work at the facility (or are located relatively close by) it is easy for the team members to
visit the site. However, if one or more members are from outside of the facility, it is recommended
that a formal site inspection be carried out
The formal inspection serves to resolve all questions raised during the review and to
complement that information already obtained and reviewed earlier. The inspection also confirms
whether or not the facility actually operates in the way it was originally intended to. inis
inspection concentrates on understanding how the wastes are generated.
-490-
-------
The assessment team should "walk theline" from the beginning of the process to the point
where products and wastes leave the facility. Since waste can be generated in receiving and
storage areas as well as the production areas, all areas within the site should be visited. The
following guidelines will help in organizing an effective site inspection:
Prepare an agenda in advance
Schedule the inspection to coincide with the particular operation of interest.
Monitor operations at different times during the shift.
Interview operators, foremen, and supervisors. Assess the operating personnel's awareness of
the waste generation aspects of the operation.
Observe the housekeeping aspects of the operation. Assess the overall cleanliness of the site.
Review the organizational structure and level of coordination of waste-related activities between
the assessed facility area and other related areas.
• Assess the administrative controls.
Generating WM options
Following the collection of data during the assessment preparation step and the site inspection,
the members of the assessment team will have begun to identify possible ways of reducing waste
in the assessed area. The generation of options is both a creative and analytical process. While the
individual assessment team members may be able to suggest many potential WM options on their
own, the process can be enhanced by using some of the common group decision techniques, such
as brainstorming. These techniques allow the team to identify options that the individual members
might not have come up with on their own.
Identifying potential options requires the expertise of the assessment team members. Much of
this knowledge comes from their education and on-the-job experience. Other sources of
background information on potential options include the following:
Trade associations
Published literature
Environmental conferences and exhibits
Equipment vendors
Plant personnel (especially the operators)
Federal, state, and local government environmental agencies
Consultants and/or employees from other facilities
Screening and selecting the most promising options for more detailed evaluation
A successful assessment will result in many WM options being proposed. At this point it is
necessary to identify those options which offer a real potential to minimize waste and reduce costs.
The screening procedure serves to eliminate those suggested options that are perceived as marginal,
impractical, or inferior, without the detailed and more costly feasibility study. The procedures for
screening these options can range from an informal decision made by the assessment program
manager or a vote of the assessment team, to a weighted sum method that combines relative
weights of such factors as operating cost reduction, capital cost requirement, reduction in waste
hazard, etc.
Some options (such as procedural changes) may involve no capital costs and can be
implemented quickly. The screening procedure should account for the ease of implementation for
an option. If such an option is clearly desirable and indicates a potential cost savings, it should be
considered for further study or outright implementation.
-491-
-------
In screening the options, the assessment team determines what the important criteria are in
terms of the WM assessment program goals and constraints, and the overall corporate goals and
constraints. Examples of criteria that can be used include the following:
Does the necessary technology exist to develop the option?
How much will the option reduce waste quantity, hazard, and treatment/disposal costs.'
How much will the option reduce safety hazards?
How much will the option reduce the use of input materials?
What will the impact be on liability and insurance costs?
How much does it cost? Is it cost effective? .
Can the option be implemented within a reasonable amount of time/ ... „ „ .
Does the option have a good "track record"? If not, is there evidence that the option can work
in this case?
• What other benefits will occur?
FEASIBILITY ANALYSIS PHASE
The WM options that are successfully screened in the assessment step then undergo a more
detailed feasibility analysis. The feasibility analysis is not unlike that earned put for any new
projecVwithin most organizations. However, there are some important characteristics; to^consider
when evaluating waste minimization projects that are not necessarily considered with other types ot
projects.
Technical Evaluation
The purpose of the technical evaluation is to be sure that the option will really work as
intended, and whether it can be implemented within specific facility constraints and product
requirements. Typical criteria for the technical evaluation include the following:
Will the option work in this application?
How has it worked in similar applications? .
Is space available? Are utilities available? Or must new utility systems be installed?
Is the new equipment or procedure compatible with the facility's operating procedures, work
flow, and production rates?
How long will production be stopped in order to install the system/
Will product quality be maintained or improved?
Is special expertise required to operate or maintain the new system? Does the vendor provide
acceptable service?
Does the system or procedure create or remove safety hazards /
Does the system or procedure create other environmental problems?
All affected groups in the facility should contribute to and review the results of the technical
evaluation. Prior consultation and review with the affected groups is needed to ensure_ the viability
and acceptance of the option. If the option calls for a change in production methods, the effects on
the quality of the final product must be determined. If the project appears mfeasible or impratical
after the technical evaluation, it is dropped.
Economic Evaluation
The economic evaluation is carried out using the standard measures of profitability, such as
payback period or discounted cash flow techniques (internal rate of return and net present value).
-492-
-------
Each company uses its own economic evaluation procedures and criteria for selecting projects for
implementation. In performing the economic evaluation, various costs and savings must be
considered. As in any project, the cost elements can be broken down into capital costs and
operating costs.
Capital costs for WM projects are similar to most other projects. These costs include not only
the fixed capital costs for designing, purchasing, and installing equipment, but also costs for
working capital, permitting, training, start-up, and financing charges. As mentioned earlier, it is
important to realize that some WM options, such as procedural or materials changes, will not have
any capital costs. Also, many source reduction options have the advantage of not requiring
environmental permitting in order to be implemented.
WM projects need to show a savings in operating costs to be economically effective.
Operating costs and savings typically associated with WM projects include the following:
Reduced waste treatment, disposal, and reporting costs
Raw material cost savings
Insurance and liability savings
Increased costs (or savings) associated with product quality
Decreased (or increased) utilities, operating and maintenance costs, and overhead costs
Increased (or decreased) revenues from changes in production marketable by-products.
Once the capital and operating cost savings have been determined, the project's profitability
can be determined using the profitability measures. These methods are discussed in virtually all
financial management, cost accounting, and engineering economics textbooks. Those options that
require no capital costs should be implemented as soon as savings in operating costs can be
shown.
An important consideration for WM projects is their potential to reduce the risk of
environmental and safety liabilities for a company. Although these risks can be identified, it is
difficult to predict if and when liability problems will occur and the financial impact It is important
that the managers within the company who decide to fund the company's projects be aware of the
significance of these risks and factor the risk reduction benefits of waste minimization into these
projects. Also, while the profitability of a WM assessment program is important in deciding
whether to implement a project, compliance with environmental regulations may be more
important, since violation may ultimately result in shutting down the facility, and possible criminal
penalties for the company's responsible people.
Final Report
The product of a WM assessment is a report that presents the results of the assessment and the
technical and economic feasibility analyses. It also contains recommendations to implement the
feasible options. A good final report can be an important tool for getting an attractive project
implemented. The report should include not only how much the project will cost and its expected
performance, but also how it will be done. Important topics to discuss include the following:
whether the technology or procedure is established, with a mention of successful applications.
the required resources (money, expertise, and manpower) available in-house, and those
resources that must be brought in from outside.
the estimated construction period and production downtime.
the means by which performance can be evaluated after the project has been implemented.
the reductions in environmental and safety liability.
-493-
-------
Before the report is finalized, be sure to review the results with the affected groups. It is
important to solicit the support of the affected groups. By having people from these groups assist
in preparing and reviewing the report, the chances are increased that the attractive projects are
successfully implemented.
IMPLEMENTATION
The implementation of the WM project is not unlike any other project that involves new
equipment or procedures. It may be necessary to overcome inertia or resistance to change within
the organization. The commitment of management to waste minimization is important at this time.
Once the project has been implemented and operating, it is important to evaluate its
performance. Is it performing as expected? If not, should it be abandoned, or is its use still
beneficial? What other potential options have been identified through the operation of this option?
ONGOING PROGRAM
The WM program should be viewed as a continuing one. As WM options are implemented,
the task force should continue to look for new opportunities, assess other waste streams, and
consider attractive options that were not pursued earlier. The ultimate goal is to reduce wastes to
the maximum extent practical.
CONCLUSION
The waste minimization assessment offers opportunities to reduce operating costs, reduce
potential liability, and improve the environment, while improving regulatory compliance. The WM
assessment procedure results in a careful review of a plant's operations toward reducing wastes.
The WM program task force should strive to build a waste minimization philosophy within the
company. In doing so, the entire company can help to minimize waste.
-494-
-------
CASE STUDIES OF WASTE MINIMIZATION ASSESSMENTS FOR
CYANIDE WASTES FROM ELECTROPLATING OPERATIONS
' by Deborah Hanlon and Michael Callahan
Jacobs Engineering Group
251 S. Lake Avenue
Pasadena, CA 91101
and
Thuy Le
ABSTRACT
*™ Hazai;dous wfste minimization opportunity assessments were conducted at
two electroplating facilities generating cyanide -bear ing wastes. The
2) ?oSTd,nMfl r asseffflents wer>e D'to test assessment procedures and
2) to identify and prioritize waste minimization options for the cyanide-
bearing wastes generated at the facilities. The procedures used for ?hese
assessments constitute those identified in the EPA Manual for
Minimization Opportunity Assessments. 1988. — ~
This paper summarizes the waste minimization options
ssss. ^xrir,
"""
and
»«
to improve upon eir
°f bath
through nitration, redu<=tio
The involvement of plant
assessment was found to be
waste minimization options
minimization programs within
human interaction as .well as
found to-be essential for the
personnel in all phases of a waste minimization
essential to the acceptance of the recommended
and to the implementation of ongoing waste
the facilities. Effective communication and
technical insight and engineering skills were
success of the assessments.
-495-
-------
TABLE 1
Summary of Options for Minimizing Cyanide Wastes at Facilities 1 and 2
| Control Category | Control Methods fop Facility 1
Control Methods for Facility 2
f. ....-.--.--- — -
[Drag-out
Minimization
(Bath Life
Extension
1
1
+.
(Rinse Water
Minimization
+ —
(Substitution of
Non- cyanide
Solutions
1
1
|1. Proper positioning of parts on rack
(2. Lower bath constituents concentration
(3. Reduce speed of withdrawal of parts
(4. Increase bath temperature
[5. Use of surfactants
(6. Improve dragout recovery
|1. Dragin reduction by better rinsing
|2. Use deionized or distilled water
(3. Bath impurity removal
(4. Mechanical vs. air agitation
|5. Proper rack design & maintenance
(6. Purer anodes
|7. Proper anode removal
(8. Return spent solution to manufacturer
(1. Automatic flow control
(2. Rinse bath agitation '
|3. Multiple rinse tanks
|4. Fog nozzles and sprays
(5. Close- loop rinsing t wastewater reuse
(6. Proper parrel rinsing arrangement
|1. Pyrophosphate copper plating bath
(2. Non-cyanide cadmium plating bath
(3. Non-cyanide silver stripping solution
I , I
[1. Proper positioning of parts on rack |
[2. Lower bath constituents concentration
|3. Reduce speed of withdrawal of parts |
|4. Increase bath temperature
| 5. Use of surfactants I
[6. Improve dragout recovery
|1. Dragin reduction by better rinsing
[2. Use deionized or distilled water
(3. Bath impurity removal
I*.
(5. Proper rack design & maintenance
(6. Purer anodes
|8. Return spent solution to manufacturer
|1. I
|3. Multiple rinse tanks
I*-
|8. Spray rinsing
(1. Pyrophosphate copper plating solution •
|2-
(3. Non-cyanide silver stripping solution |
(Substitution of
Plating Method
|1. Cadmium ion vapor deposition
(2. Aluminum ion vapor deposition
(Good Operating
Practices
(6,
Waste stream segregation
Operator training/ closer supervision
Spills and leaks prevention
Preventive/ corrective maintenance
Management initiatives
Material/ waste tracking
(1. Waste stream segregation
|2. Operator training/ closer supervision
[3. Spills and leaks prevention •
|4. Preventive/ corrective maintenance
[6.
T. Removal of salts and dust
I. Wastewater routing
(Alter Plant Layout|1.
-496-
-------
BACKGROUND
. set or
ASSESSMENT SUMMARY FOR FACILITY 1
DESCRIPTION OF FACILITY 1 PLATING PROCESS AND WASTE STREAMS
Zt ±S esti^ed that at
-497-
-------
Figure 1 represents a typical process flow diagrao i for
the
was?e, spe'nt striping solution, cyanide-containing rinse water, and
inadvertent spills.
Cvanide losses occur primarily because of drag-out of plating/striping
s
recover precipitated metals from the sludge produced.
WASTE MINIMIZATION OPTIONS
After the site inspection, the asaessment team identified numerous WM
selected for additional evaluation and analysis.
technical and economic feasibility evaluation was
economic analyses associated with the selected
Figure 1
3i
RM3C
-------
Table 2
Summary of Economic Feasibility Study for Facility 1
Description
of
Options
1. Drag-out
Minimization
2. Bath Life
Extension
3. Rln«« Watw
Minimization
Waste
Reduction
Method
Us* of Drain Board*
Us* of Drain Boards/
impurity rempvai
Us* of Drain Boards/
Mechanical Agitation
Still Rinsing
• Coppar
- Cadmium
- Silver
Impurity Removal
Mechanical Agitation
Spray Rinsing
Waste Capital Monthly
Reduction Cost Cost
(%) ($) ($/month)
so
90
90
40
40
40
•
-
50
890
1103 1820
7030
560
1680
2800
213 1820
6140
1168S
Monthly
Savings
($/month)
784
784
784
202
58
89
•
•
440
Pay-Back
Period
(month)
1.1
-
9.0
2.8
2.4
2.6
•
m
2.2
yrs
yrs
yrs
yrs
Drag-Out Minimi2iatiQn
Is and treatment/disposal
included the following:
Surface tension. The use of surfactants will lower the surface
tension of the plating solutions and will reduce liquid clingagef
o Viscosity. Increasing the bath temperature
9alnd redUCS -
the
will reduce
method wmS
h h H . water from a sti11 <*i*se to the
i, t0 the higher ~te of evaporation. Care must be
taken, however, to ensure that the increase in temperature does not
lead to break-down of the bath constituents.
iirfh« and,,s"rfa°e area °f ^e workpiece. Rack mounted
items should be positioned so that all recesses can drain properly.
h n Sfffd° Wifchd^«ing the workpiece slowly from
the plating solution will provide ample time for the part to drain.
-499-
-------
Drain boards. Drain boards can be installed to recover plating
solution that would otherwise drag into the first rinse tank.
Further investigation of these options revealed that some of the drag-
sss. -srs? of
.
The
saving of $784 from raw materials savings and avoided disposal costs.
resulting payback period would be 1.1 month.
Bath Life Extension
Extending the bath life of the plating solutions was identified as a
option by the assessment team. The lifetime of a plating
soluion i controlled, among other factors, by the input rate of
™n?aminants and/or by the rate of depletion of constituents due to drag-out.
DeSeSon of constituents due to drag-out can be reduced by minimizing drag-
out (discussed previously). However, many operators rely on excessive drag-
o"ut tf remove b^h impurities so some other means of impurity removal must be
included when a facility employs drag-out reduction techniques. The
following discusJes the major sources of impurities and the ways to control
or reduce them.
o Racks. Proper rack design and maintenance can help reduce the
amount of corrosion .that enters the bath.
o Anodes. Use of purer anodes, use of anode bags, and removing the
anodes from the bath when not in use can reduce dissolution of the
anode material into the bath.
o Drag-in. Proper rinsing of the parts before entering the bath can
reduce drag-in of impurities and contaminants.
o Water make-up. Use of deionized water helps to prevent the build-
up of certain minerals such as calcium or magnesium.
o Air. Use mechanical agitation instead of air agitation to avoid
the build-up of carbon dioxide /carbonates.
After reducing the amount of impurities that enter the bath, additional
-
.
Jor reduciag plating bath waste is to return the solution to the manufacturer
for reprocessing.
The WM options recommended for Facility 1 to extend bath life included
conversion from air to mechanical agitation of the bath and to chill the bath
solution to freeze out carbonates and other *»™"£*'*£
for installation of mechanical agitation was estimated to be
-500-
-------
product quality.
Rinse Water Minimization
. h installation.
since both must be implemented to assure
.adequately rinsing the workpTece? ?hese are
-till
o
o
o
o
o
o
Automatic flow controls.
Rinse bath agitation.
Multiple rinse tanks.
Spray and fog nozzles.
Closed-loop rinsing and rinse water reuse,
Proper barrel rinsing arrangement.
Non-Cyanide Substitutes
previously untried, on a small scale.
solutions,
o
o
o
Pyrophosphate copper plating solution.
Non-cyanide silver stripping solutions.
Non-cyanide cadmium plating baths.
Plating or stripping operation
have a much lowe'r tole?anoe
solution3-
step whioh precedes the
non-°^1<'«
requirement for ths USe of
specifications is regarded to
Good Operating Practices
^anide-based to non-
MHWhl°h Spe11 out the
££%£" °*
-501-
-------
GOP recommended for this facility by the assessment team included the
following:
o keep cyanide waste streams separate from other wastes;
o provide more operator training and closer supervision;
equip all tanks with overflow alarms;
routinely inspect for leaks, improve operational control for
loading/unloading and transfer operations;
o
o
o
o
isolate equipment or process lines that are not in service; and
document the spillage and related dollar values for future
improvements in clean-up procedures and prevention measures.
ASSESSMENT SUMMARY FOR FACILITY 2
DESCRIPTION OF FACILITY 2 PLATING PROCESS AND WASTE STREAMS
Facility 2 is a small privately owned electroplating shop located in
Southern California. Their main business is refinishing deooratwe item.
The principal metals plated at this facility are nickel, brass, silver, and
gold.
Thirty tanks are used in cleaning and electroplating operations. The
configuration of a typical plating unit includes a plating bath, followed by
one or two still rinse tanks and a continuous rinse tank. Except for nickel
plating, all plating/stripping solutions presently used at the facility are
cyanidelbased. All plating operations performed at the facility are
performed manually.
Cyanide waste is generated from silver stripping; from silver, gold,
brass, and copper electroplating; and from associated rinsing operations.
The principal "ate streams are wastewater (which includes overflows from the
continuous rinse tanks and water used for floor washings), and plating tank
filter waste. Aqueous streams generated from paint stripping, from metal
stripping/electroplating, and from floor washings are routed to a common
sump, then to the sanitary sewer.
In addition to sump sludge, solid waste is also generated due to the
accumulation of metallic sludge inside the plating tanks, *** " ^"ered
out from the plating solution once a month using a Potable dual cartridge
filter. Two filter cartridges are used for each plating tank. Each filter
cartridge is replaced approximately every 2 to 3 months. The contents of the
sump Approximately 300-400 gallons of sludge containing dirt, stripped
paint, cyanides, and. heavy metals) are vacuumed out and disposed of as
hazardous" waste 'approximately once every 6 months. The total cost of waste
collection, transportation, and disposal (per visit) increased from $2,000 to
$4,000 after September 1986.
-502-
-------
PROPOSED WASTE MINIMIZATION OPTIONS
minimization options for the
during the meeting:
infection was
"881_°a to . "entity potential waste
The following options were proposed
o Reduce solution drag-out from the plating tanks by:
L— -
Extend plating solution bath life by:
Reducing drag -in by better rinsing.
Using deionized make-up water.
Using purer anodes.
Returning spent solutions to the suppliers.
Reduce the use of rinse water by:
n countercurrent rinse tanks.
Using still rinsing.
Using spray or fog rinsing.
for the feasibility analysis:
o Reduce drag-out by using drain boards.
o Extend bath life using deionized water for make-up.
o Use spray rinsing to reduce rinsewater usage.
o Segregate hazardous waste from unhazardous waste.
FEASIBILITY ANALYSIS
the following options
2 is shown in Table 3.
economic feasibility for
-503-
-------
Table 3
Summary of Economic Feasibility Study for Facility 2
Description
of
Options
Drag-out
Minimization
Bath Ufa
Extension
Rlnsa Water
Waste
Reduction
Method
Usa of Drain Boards
Use of Dolonlzed H,O
Usa of Spray Rinsing
Hut of Plastic Media
Waste
Reduction
(%)
50
50
50
90
Capital
Cost
($)
315
582
2825
17900
Monthly
Cost
($/month)
-•
38
••
2519/yr
Monthly
Savings
($/month)
241
241
29
6607/yr
Pay-Back
Period
(month)
1.3
2.9
8.1 yrs
4.4 yrs
Practices
Blasting
Segregate Hazardous Wastes
The assessment team recognized that segregating hazardous wastes from
nonhazardous wastes could, be implemented at virtually no cost and would save
money immediately. There were no identified technical problems.
Use Drain Boards to Reduce Drag-Out
Drain boards could be used to collect plating solution that drips off the
rack and the workpiece after they are pulled out of the plating tank. The
plating solution dryers from the part onto the board and drains back into the
plating tank. This option reduces the concentration of cyanide in the rinse
water and reduces make-up chemical consumption.
The purchase price of drain boards is estimated at $115, with installation
costs of $200, for a total capital cost of $315. This option is expected to
reduce rinse water disposal costs by $500 per year, and reduce make-up
chemicals costs by $400 per year. The resulting payback period is 0.35
years, or about 4 months.
Use Deionized Water for Make-up Solutions and Rinse Water
Since the use of drian boards tends to dreduce the rate of impurity removal,
use of DI water will help reduce the build-up or introduction of impurities
into the plating bath. In particular, the build-up of minerals from tap
water will be avoided. This, in turn, will avoid the precipitation of
carbonates in the plating tanks.
The assessment team decided to combine the evaluation of this option with the
previous option of using drain boards. The initial purchase and installation
of a rented deionizer was $267. When adding the cost of the drain boards,
the total capital cost of this option is $582. The deionizer is rented and
-504-
-------
serviced by an outside water treating service company for $450 per vear The
in *1Sp°Sal costs and "ato-up chemical costs is%900 per *yelr
opratins cosfc savings ls
Install Spray Rinses
Installing spray rinses would reduce the amount of rinse water reauired to
clean the ltems. With spray rinse nozzles and controls, rinsing can be done
on demand. Rinse water usage was estimated to be reduced fly 50% The
5°*
The assessment team determined that four spray rinse units would cost 42 120
resulted in a payback of over 8 years. wastewafer. This
Implementation
• systel" "as online two months later. The results of
CONCLUDING OBSERVATIONS
"to
""S
SJg.~E?
minimizati°» P^S^am. Some non-technical sSlls of
t .
facilitator for the assessment team and host facility personnel alike In
e
-505-
-------
in some cases, only partially. The availability and quality of information
often varied significantly. It is important to make allowance for this
possibility and to have a fall-back position. For example, if a piping and
instrument diagram (P&ID) is not available, often times a piping layout plan
will be. Similarly, if the information is not available from the facility,
it does not mean that it cannot be obtained.
Much information is available from outside sources such as trade
journals and vendors, e.g., the costs of bath make-up chemicals or the
physical design of the process equipment. If information is truly needed but
is not readily available, it can be obtained with proper initiative and
ingenuity, but not without a possible detriment to project cost and schedule.
In this light, it is important that only the information truly required to
understand the process, to allow for delineation of waste sources and current
waste management techniques, and to characterize waste generation quanti-
tatively should be sought. Requesting unnecessary information burdens both
the provider and user (auditor) and slows down the work.
The importance of the pre-assessment activities, particularly the pre-
assessment site visits, were revealed as being extremely important in
facilitating the assessment process. At a facility where the assessment team
spent a little more time getting to know the host facility staff and how
their organization functioned, the assessment process moved more smoothly.
The assessment team found it easier and faster to acquire needed data because
they knew better where (e.g., from whom) to obtain what data. Likewise,
because they knew the operation and the people a little better, the level of
cooperation by plant staff was improved.
It was observed during the assessments that while waste stream selection
helped to sharpen the focus of the effort, such selection should always be
open to review and reassessment. An excellent example is provided by the
assessment of Facility 2. During the waste stream selection stage, the
generation of the cyanide-bear ing sludge in the sump was attributed to the
electroplating operation. As waste minimization options were formulated, it
was determined that the actual source of the sludge was from buffing and
paint stripping operations (operations where cyanides are not used). Since
total segregation of the cyanide-bearing electroplating waste and the non-
cyanide-bearing sludges was impractical, the scope of the assessment was
expanded to include options regarding sludge reduction and elimination.
The pilot studies experience led to a modification of the initially
proposed assessment method. The initial approach of having the host facility
personnel merely review and discuss the assessment team's ratings following
the presentation of the option ratings resulted in a relatively casual
attitude and less involved behavior by the host facility staff. To overcome
this barrier, the modified approach required host facility personnel to
independently develop ratings of each of the waste reduction options under
consideration. The assessment -team's ratings of the options and the host
facility's independent ratings can then be reviewed and reconciled in a group
session.
-506-
-------
ACKNOWLEDGMENT
-Sfc
-507-
-------
BIBLIOGRAPHY
Durney, L.J., ed. 1984. Electroplating engineering handbook. 4th ed. New
York: Nostrand Reinhold Co.
Fromm, C.H., and Callahan, M.S. 1986. "Waste Reduction Audit Procedure." A
methodology for identification, assessment and screening of waste
minimization options. Conference Proceedings. Atlanta, GA.: Hazardous
Materials Control Research Institute.
Kahane, S.W. 19&6. Waste minimization assessments. In Symposium on solvent
waste reduction. Los Angeles, Calif.: ICF Consulting Assoc.
Lowenheim, F.A. 1979. Electroplating. In Kirk-Othmer encyclopedia of
chemical technology. 3rd ed. Vol. 8. pp. 826-829.
McRae, G.F. 1985. In-process waste reduction: part 1. Plat. Surf. Finish.
72(6):14.
Olsen, A.E. 1973. Upgrading metal finishing facilities to reduce pollution.
EPA-625-3-73-002. Washington, B.C.: U.S. Environmental Protection Agency.
Parkinson, G. 1979.
pp. 25-27.
Presenting - the energy audit. Chem. Eng. October 11.
Pojasek, R.B. 1986. Waste minimization - planning, auditing and
implementation. In Monograph on Hazardous and Solid Waste Minimization.
Washington, D.C.: Government Institutes Inc.
U.S. Congress 1986. Serious reduction of hazardous waste for pollution
prevention and industrial efficiency. OTA-ITE-313. Washington, D.C.: U.S.
Government Printing Office.
USEPA 1981. U.S. Environmental Protection Agency. Inplant changes for metal
finishers. Cincinnati, Ohio: Industrial Environmental Research Lab.
1981. U.S. Environmental Protection Agency, Office of Water Regulation
and Standards. Development document for efflient limitation, guidelines, and
standards for the metal finishing point source category. EPA-440-1-83-091.
Washington, D.C.: U.S. Environmental Protection Agency.
1986a. U.S. Environmental Protection Agency, Office of Solid Waste and
Emergency Response. Report to Congress - minimization of hazardous waste.
EPA-530-SW-86-042. Washington, D.C.: U.S. Government Printing Office.
1986b. U.Si Environmental Protection Agency. Waste minimization,
issues and options. Vol. 1. EPA-530-SW-86-041. Washington, D.C.:
Government Printing Office.
Williams, M.A. 1976. Organizing an energy conservation program. Chem. Eng.
October 11. pp. 149-152.
-508-
-------
TCLP AS A MEASURE OF TREATMENT EFFECTIVENESS: RESULTS OF TCLP WORK
COMPLETED ON DIFFERENT TREATMENT TECHNOLOGIES FOR CERCLA SOILS
by: Robert C. Thurnau
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
M. Pat Esposito
PEI Associates, Inc.
Cincinnati, Ohio 45246
ABSTRACT
The 1984 Hazardous and Solid Waste Amendments (HSWA) of the Resource
Conservation and Recovery Act (RCRA) require that EPA either ban the
disposal of hazardous wastes to the land or ascertain that such wastes
are acceptable for land disposal. The soil and debris associated with
the clean up of Superfund sites also fall under these statutes and must
be addressed. A significant part of the regulatory strategy adopted by
EPA involved the determination of best demonstrated available technology
for contaminated soils and debris. A series of soil treatment tech-
nologies that were considered as candidates for Superfund sites
(physical, chemical, thermal sol idifcation) were tested on a laboratory
prepared feed sample and the waste product streams generated were pro-
cessed by the Toxicity Characteristic Leaching Procedure (TCLP).
The TCLP chemicals and mechanism have been compared to some of the
most severe leaching conditions experienced in the land disposal of
hazardous wastes, and therefore the results of these tests should
simulate the worst case situations. In this context, TCLP is being
studied as an indicator of treatment effectiveness, and may be one
of the criteria employed to determined if a waste is banned or land
disposed. This paper presents the TCLP data generated from the five
(5) different BOAT treatment technologies tested and helps to put
the use of this technique into practical perspective
-509-
-------
INTRODUCTION
Under section 3001 of the Resource Conservation and Recovery Act (RCRA)
EPA was charged with identifying those wastes which, if improperly managed,
would pose a hazard to human health and the environment. The statute also
specified that EPA identify such wastes through the development of lists of
hazardous waste characteristics. Characteristics are those properties
which, if exhibited by a waste, identify it as a hazardous waste, and are
established for levels of which there is a high degree of certainty that the
waste needs to be managed in a controlled manner. The Extraction Procedure
Toxicity Characteristic (EPTox) was developed to determine if specified
metals, insecticides and herbicides could be mobilized from a simulated
municipal sanitary landfill environment. As with most first generation
rules, improvement and expansion followed. The Toxicity Characteristic
Leaching Procedure (TCLP) was subsequently developed to address the
mobility of a broad range of both organic and inorganic compounds and to
solve the operational problems of the EPTC protocol.
The work reported in this paper centers around a set of soils that
were synthetically prepared to simulate the soils found at a typical
Superfund site. The soils referred to herein as Synthetic Analytical
Reference Matrixes (SARM) were processed through five different treatment
technologies: incineration, low temperature thermal desorption, chemical
treatment, physical treatment and stabilization. The performance of each
technology was evaluated by comparing total waste analysis (TWA) and TCLP
analyses of the starting materials to the treated, residues.
TOXICITY CHARACTERISTIC LEACHING PROCEDURE
A brief description of the TCLP is as follows: The sample is classi-
fied as liquid or solid by the percentage of solid material, reduced in
size if necessary «9.5 mm), weighed and mixed with an acidic solution at
least 20 times the weight of the solid phase. The mixture is filtered and
the extract is retained for chemical analysis. The procedure is modified
somewhat when volatile organics are involved in that a zero headspace
specific extraction vessel is required. A TCLP flow chart is presented as
Figure 1.
EXPERIMENTAL
The basic composition of the soil used in the technology evaluations
was determined as a result of an extensive literature search of Superfund
records. The final soil composition selected consisted of (by volume):
30$ by volume of clay (montmorillinite and kaolinite), 25% silt, Z(U sand,
20% top soil and 5% gravel. The components were air dried, and then mixed
-510-
-------
FIGURE 1. TCLP Flowchart
Wet Waste Sample
Contains 0.5%
Non-Filterable
Solids
! Representative Waste
Sample
Wet Waste Sample
Contains 0.5%
Non-Filterable
Solids
t
Liquid/Solid
Separation
0.6-0.8 urn
Glass Fiber
Filters
Dry Waste
Sample
Discard
Solid
-Solid-
Solid
Liquid/Solid
Separation
0.6-0.8 um
Glass Fiber
Filters
f
I Reduce Particle Size If 9.5 mm I
| Or Surface Area 3.1 cm2 I
t
TGLP Extraction*
of Solid '
O-Headspace Extractor
Required For Volatiles
Liquid
Store At
Liquid/Solid
Separation
0.6-0.8 um Glass
Fiber Filters
Discard
Solid
I
Liquid
- TCLP Extract -
TCLP Extract
i
I
Analytical
Methods
--TCLP ExtractJ
* The extraction fluid employed is a function of the alkalinity of the solid phase of the waste.
-511-
-------
together in two 15,000-1b batches in a standard truck mounted cement mixer.
A prescribed list of chemicals found to be widely and frequently occurring
at Superfund sites was added to the clean soil to produce the SARMs in a
series of small scale mixing operations utilizing a 15 ft-3 mortar mixer.
The organic components added were: acetone, chlorobenzene, 1,2-dichloro-
ethane, ethyl benzene, styrene, tetrachloroethylene, xylene, anthracene,
bis(2-ethylhexyl) phthalate and pentachlorophenol. The metals added to
the clean soil were either salts or oxides of: arsenic, cadmium, chromium,
copper, lead, nickel and zinc. Due to the fact that the contaminant pro-
files of Superfund soils differ widely from site to site in composition and
concentration, four different SARM formulas were prepared. Table 1 presents
the target contaminant concentration level of the four SARM samples as well
as the actual level achieved. The SARMs showed representative consistency
and homogeneity between formulations and approached the target levels
outlined on Table I.
The SARMs produced in the first phase of this project were then pro-
cessed through five treatment technologies that were thought to be most
readily available and to have the greatest applicability to CERCLA site
restoration activities. The technologies studied were: incineration, low
temperature thermal desorption, chemical treatment (KPEG), physical treat-
ment (soil washing) and solidification/stabilization. The effectiveness
of the five different treatment technologies was measured by observing the
change in concentration of the identified compounds and elements. TWA was
used to measure changes in the total contamination levels in the treated
SARM residuals as a result of treatment. TCLP on the other hand was used
to track potential changes in leachate composition as a result of treatment.
The logic behind the TCLP aspect of the project was that if the treatment
reduces the total level of contamination or immobilizes it to a point
where it will not migrate with the acidic leachate, the residual will be a
good candidate for land disposal. Thus, residuals from each of the treat-
ment technologies were collected and processed through the TCLP, and
analyzed for the metals and organic compounds listed above.
RESULTS
Table II presents the results of the TCLP on the untreated soil SARM I.
The data indicate that the semivolatile organic compounds were not effi-
ciently extracted from the untreated soil. The volatile organic compounds
were extracted at higher concentrations than the semivolatile organic
compounds with the most volatile compounds (acetone and 1,2-dichloroethane)
being extracted completely. Figure 2 shows a plot of the percent of vola-
tile compound extracted by TCLP vs. the compound's boiling point. There
appears to be a trend toward lower extraction of volatile organic compounds
with increasing boiling point. The relatively lower water solubilities and
high boiling points of the semivolatiles are a logical extension of this
trend. This type of information would be useful in determining the minimum
soil concentration levels that would be necessary for the use of TCLP as a
predicting tool for organic compounds in residuals earmarked for land
disposal.
-512-
-------
CC
CO
V)
CC
c?
10
§1
If
<- ^
LJJ
CC
CO
m
O
c <
O^
_ eo
_—
CC
CO
CO
J3
r2
_
CO
cu
en
(2
"5
3
"*
en
H~
~s
1
(2
"5
*••
(2
Compound
ro O ^- CO T— O CO
in CN ro o oo *t r^
o co >t r^ co in in
oo CM in
Ooooooo
ooooooo
CO <^- CD CM O CD CM
CO CO «- 00
O O ro O CO ,- O
co s; * CO CN , CN
\*j
OOOOOOO
00 •* CD CN O CD CN
CO CO «- CD
•* **• CD °- °- o 9
00 O' • ^ CO ._: CM
<- »- co <- co ^ r^
tl -H tl 4-1 tl -H tl.
O CM CD O O O O
r^ co •* co co en in
in i- CM -t <- co '
.CO r- f—
ooooooo
00 ^- CO CN O CD CM
CD CO ^ CO
CO ,_
co co •£ r^ CM
-' § 2 CD en 2 ®
tl +1 fl -H t-l fl fl
co co ^ en r^ 05 m
in »- in CN o in in
oo co co co r- yj- m
"t co in
So o o o o o
o o o o o o
CO t CD CN O CO CM
CD CO <~ CO
« C
C 0
Hi !
Ilifiii
< o ^-- ULI w (£ x
t— CN co *t in co r^
CM
o *d-
en co co
r- CM CM
H fl tl
O CD O
CM -t CO
O) CO
O O 0
000
in in o
CD CM t—
O O o
J^ •* O
co p co
o o o
in in o
CD CM *~
9 § q -
r^ CN in
w * — LO
tl -fl tl
O O O
co CD cri
in co co .
O o o
in in o
CD CN .»-
s s «--
r^ T— co
CO T— CM
•HI '+1 +1
•^ CO CD
co in CM
CO CO CO
in T-
O o O
O o O
in in o
CO CN *~
S
JO
03
£• o
. 1 1 •
§ 1- 1
ro o
c "to g
< bo a:
S ro 2
ro •- PJ ^
I^ TO r^f rn CM ^^ C?
«- CN ^- v- CN co in
tl +1 fl tl fl fl f|
S § 5 S S g S
8 8 88 8 8-8
in o in in o O in
•— "- ro <* «- CM
T- CM
CM r- 1^ °0
3 in co in o oo r~
g co CD m m r» co
+1 tl fl fl tl fl f|
^^ ^5 ^D ^5 c^ ("5 r^
S g g S £ 2 £
r~ oj CN »— CD •* »-
-^ ^- ro S! t- co
^ CO
§000000
o o o o o o
in o in in o o in
"— *- ro ^- v— CM
T- CM
S's S 5 § 3.1
4-1 fl fl fl tl fl fl
^" ^" O O O ^t O
^ ^t co ro in uj co
•~ CM oo in o co *-
CM CM CO "*
o p o o o o o
•- CN co ro co co in
'—CM -t
o co CN r-, ?2
r^- '— r^ in CD ^- CM
-fl +1 fl fl 4-1 4-1 4-1
ro co co ro in co CM
i- CN CM CO •* CO O
CM CM in
ooooooo
<- CM co ro co co in
*- CM •*
* 1 1 .
1 1 1 1 1 1 8
< o o o 3 ± K5
^- CM co •* in co r~
-513-
-------
pOJOBJJX3
Q.
(0
*o
B
CO
Q>
DC
'
3 SJUQUOdlJUOQ
dlOl %
UOIJOEJ1X3 JOJ
BAV lunoiixeiAl
s3|diuEg
11V J° Uea|/\|
joafo-id
UOIJBJ3UIOU|
qO8_l_ UE/^\ 88"|
\.# ajduJBS OD
UOai UB^^ 88"|
2# ajdtuBS OD
sajeioossv (3d
\,# 8|duiBS OD
BumsBM IPS
•o
CM C
E
I 0
888 8888
LO CD CD OO 1I~ ^J ®*
0) f CM CM CO CM CM
800 >— o o P
,5 10 CM CM 00 C5 CO
41 41 41 41 41 41 f]
000 0 o 0 g
o co r-* in ,3- o .
co to K -. r^ d
r- o in co co «- ^
41 41 +' 41 41 41 41
800 0 0 0 o
co co r— o
co *~~ (f ^
|og P P g |
000 0 0 0 0
nf K "fr -rf CM P- CO
co --.•*'- i>
P S P | | g |
o- S" 8 § 8" g" p
g <- 55 § S2 •- P:
o o o o
d 0 0 0 „ P
in 0 K CM p g co
M S N t - 5 ^
g o- 8 8 o- o 8
" to .T* _J O CO tn
01 r*- CM ^- T^ ^ ^
§ g 8 8 8 ° P
CM CO CD O o •* P
CM CM "3- «- r-
O O O •_ O
0- o p p g P
S 5 CM 5 52 g S
g g 8 8 g g §
CD r«* oo to ^5 ^ -.
s s m s
" 2 S o
5 — c ^
§a "o § S = "5 c
_ 5 =5 •>. 2 2 §
g S CM -S & S 5-
^? U i-^ UJ CO I— X
o
o>
)
t- CM "0
Poo
o d co
V
I'll
41 41 41
g g g
m p: co
CM p <-
52 CM CO
41 41 41
9 g §
2 CM •*
SCO
O 0)
9 g
III
• c?
8 8 1
° CM
° 8 S
0 P CO
9 ° :
.0
§O m
P ^
O ^
o o o
o o o
CO CO ^f
o" o" o"
q p o
CO CO ^
o
111
8 8 1
IO r-| CO
o"
O O O
o q o
C5 CO Ifi
^- rf o"
p O CD
If) l£) OO
O CD
g 8 '-
Anthracene •
Bis(2 ethylexyljphthalate
Pentachlorophenol
. — — o
OD OJ ^_
* *
0
8888888
O> CD 10 O> CO "tf O>
IT) CN *— CO ^J-
5 | S 8 S 1 1
41 S 4| 41 41 41 4|
S <» g 8 g S 9
d *-= ^ CM CM T-' m
'-' "- *- CM.
r^- ^- -^- o o c^ °
O CM o oo oo ^ m
41 ft 41 fl H 41 44
05 r~ CD o o "0 °
o cq q in to in 10
in CM ^- ^- CD •* CD
;:; CD o co "^ °? °?
d o o t-1 d d r^
0
CM r~ f-\ c~i ^~ CO
m oo ^- 9 S i*- in
g -d ° in CM o «-
^ oo 2 CM" CM" °" g"
d d d iri CM d i£5
- S „ S ° ° |
3 o 3 in ^_ °_ -
_r Q - — «r O O o"
S co j: 9 r- r- ^
d d d 10 CM d co
CM co o o ' m in
o co o 05 S P co
d rt d co T-: o T-
CM CD" d d CM" o
O CO O 05 CO CD CO
d ^ d co --' o J2
CM CM O 00 *fr o
O CO O CO O O o
d d d d d d «-=
o
CM CM 10 _ Tf °0
o r~ o o g CD co
do d • ^: ^' o
» - «• • o
2! £^ S o" o" S o>
P f~: P co 05 P co
o op oi "-1 ° ^
CM" CM" 10 o- o ^ o"
d d o 't v-' o co
,r, CO »— r~ O5 r^ O
;2 in o co •* CM CM
o" d d d d d en
,-o i | ...
c c e g. • -53
03 .£ O ^ "O Ji O
J2 ^ £ o g .9 .E
< O CJ U J 2: M
^ CN co ^- to to r^.
O
o
CN
-514-
-------
CO
+
X
00
"^ 00
o d
0)
'o
i
o
o
a
_o
+3
a
CO
4-<
X
LU
i
s
O)
CQ
CM
I
g>
uu
o
o
o
00
o
CO
o
CM
O
O
O
00
O
CD
O
o
D)
CO
o
CM
O
O
O
00
o
CO
o
CM
%) iuaioi.yaoo
-515-
-------
The third area reported was the release of metals by TCLP. Although,
the procedure was designed to simulate the worst case leachate, in these
studies the TCLP only dissolved about 30 percent of the original metal into
the leachate sample. Again, this is important from a sample selection and
analytical detection limit standpoint. A sample like SARM I which has a
relatively a high clay composition and a relatively low concentration of
metals would not be a good choice for evaluating treatment effectiveness or
metals as measured by metal mobility on the basis of TCLP. As the
contamination level rises', the easier it becomes to evaluate both treatment
effectiveness and TCLP mobility.
The main thrust of the project was to evaluate the treatment effective-
ness of the five different technologies, by either destroying, removing or
containing the pollutants of interest. One of the ways to judge the treat-
ment effectiveness was to compare the Teachability of the target compounds
after treatment with similar data before treatment. TCLP was the vehicle
selected to make this comparison. In making this selection it should be
remembered that because the TCLP is not totally effective in extracting the
different classes of compounds, the initial TCLP values for untreated SARMs
were in several cases quite small and often near the limit of analytical
detection. Thus in many cases the treatment effectiveness was judged on
the difference between two small numbers. With these thoughts in mind, the
TCLP data generated for the five CERCLA tratment technologies is presented.
Table 3 presents some typical TCLP data collected for the solidifica-
tion experiments. SARM I was treated with portland cement, kiln dust and
lime/fly ash and cured for 28 days. The stabilized materials were sampled,
examined by TCLP and summarized. Each of the individual chemicals in the
TCLP extract from the treated residuals were compared to the initial TCLP
concentrations for the untreated SARMs and the individual removal effi-
ciencies were calculated. Appropriate adjustments were made to account for
dilution when the binders were added to the SARM. These efficiencies were
summed and averaged for each class of compounds (volatiles etc.) and the
average number taken as the treatment effectiveness attributable to TCLP.
The bar graphs show the relationships for the different binders toward the
same class of compounds.
The type of data illustrated in Table 3 can be expanded across all the
technologies. Table 4 illustrates a typical TCLP data set for SARM I and
how it was affected by each of the five treatment technologies tested. The
TCLP data indicate that incineration did an excellent job of reducing the
Teachable organics in the residues. Surprisingly, the TCLP data for the
metals in incineration ash are 83 to 99+ percent Tower than the untreated
SARM TCLP vaTues, indicating that either metals were removed from the ash
during the incineration process or that the ash was altered so as not to
release as much metal in the TCLP test. Low Temperature Thermal Desorption
at 150°F was only moderately effective on the volatiles, but at 350°F and
was very effective. Semivolatile removal results exceeded 95% for
anthracene and bis(2-ethyl hexyl) phthalate at all 3 temperatures; results
for pentachlorophenol are somewhat erratic but seem to indicate the best
removal rate (about 90%) at the 550° temperature. The metals data point
toward a change in the soil matrix during heating which results in higher
-516-
-------
Eco
0}
CO
W —
iss
% orie
OcTS6
% orss
pue|jjcy
% OS'96 MSV A|-j/3Lun
08'86
% OL'Zd
08'S£ MSV Ay/aiun
O9'i6
o
o
in
r-.
I '"1
o
in
I IAIUVS
dlOl
q P8UIUJJ813Q
uojionpsa %
CO
OJ
(O
<-•
c
0)
£
CL
X
LU
c
CO
o
o
e/j
o
u.
CO
£
3
CO
(0
•*-•
(0
Q
a.
CO
u
'5.
uojjonpgy
1 IrtUVS
uoiionpaa
1 1A1HVS
uoijonpay
luatuao puepjod
-I lAiaVS
dlOl Ag
paujuuajaQ sv
I/Bui anfe/t
Aacne|n68y
dIOl
•a
c
1
E
o
O
CT> op cri o CN K rs
[^ ^ 05 co CD r^ r^
A
§ § 9 8- S o o
CO *~ O CO ^" LO O
V M
05 en en en en en en
A A
O tO o tO to CM Tf
, ,*~: P *-. P o p o
V V
O CM o in m «-; o
co CN cd r*^ •"* c^ en
*- co en r-. r^ c0 .co
A A
9 o **• o o
O co ** co co ^ O
§ - ° 2 CM ^ »
V V
o
0 0 0 0 ... o r,
o" oo "> *• 5= r~ S
r- o ID oo ri ^ S
-h +, -H -H -H -H •-
8 g 8 S 8 S 2
r- p^ co in »-= . E 55 g
g E CN £ >- fc -5.
^ <-) ,-- £ cn |£ ^
i — oj (*j TJ- ifj (o p^
cc
r-
A
to
en
A
o
en
CO
A
en en o
cn cn cn
A
S s §
V
CO CO CO
cn cn cn
A A A
m in _*
p o £2
V V V
O O O
CO (N to
cri CD tri
cn cn
A
S g £
V
g | o
+1 --H HH
§ S R
J2 S "*
to
Anthracene
Bis|2 ethylhexyl)
phthalate
i Pentachlorophenol
S S S
in
to
cn
A
O
CO
OJ
A
o
r--
(N
to
to r*- co to r* cn
O co to co o cw cn
«i en to cn cn cn CT
A A A
S 5 o o S o o
o o o o ci o" c>
V V V V V
O ift o »-: CD r~ o
2 S S § g S g
A I A A
S 5 g -S 2 S 3
o o o o o o" o
V V V V
o o o o o o
Q U3 O CO to r-* <-;
H co o cd o pj SS
cn 01 cn . 01 °3
A A A
m 5- S co ^ -a- ro
<— o o o 21 o <^-
o o o o o o o
V V V V
q S S g g a g
0 0 0 t-' c-j o _:
-H -H -H -H +i -H -H
en r- to o t-i o r-.
q .to o in g S g
O O O CO ,-: o CN
Oop O
tn i-: in tri
-s J -I 5
< 0 6 (3 jj il r^
^ « £2 2 £ « ^
en
CN
ID
O>
iri
cn
-517-
-------
uoijonpaa
anpisaa
uoijonpaa
%
o
c o
o in
»= in
Q.
u uoijonpaa
Q %
S 0
-C 0
h- in
a, co
3
1
1 " %
|2
uoijonpaa %
d101 ^9
pajEjauag anpisaa
-uoijejauiou|-
l/Bui
1 I/UUVS
|!0£| pd}69J^Up|
UQ B1BQ dlOl
Table 4 Comparison of
Treatment Technologies for
SARMI Using TCLP
Compound
i< cri cri CM ^f co co
en oo en oo r*- en oo
A
g S o o g o o
T* d d co CM d d
v v -
in CD CD co i— tn CD
CD* CO CD CD CD CO CD
CD CD CD CD CD CD CD
A A A A A A
o o o o o o o
T-1 d d d d d d
V V V V V V
o g g g ° g g
CD CO CD CD CD CO CD
CD CD CD CD CD CD CD
A A A A A
CD O O O O O <-
•~ d d d d d d
V V V V V
co co CM P": q co '—.
§d 06* co cri d CM
co en en co co
8 g § 8 '$2 g sa
o CN o ^ co T-^ in
Ol Tj* "•'
^- o o o o 2 °
CD IO CD (O O " 1"^
CD CD CD CO CD CD CO
CD CD CD CD CD CD CD
A A A
•* •* ,-, *° ^- ** m
r- o ° CD ?: o o
odd d d d T-:
V
000 0 .-, 0 g
O i co in *d; i± c~- _;
f~- d in oo co ^ "
-H -H -+) -H -H -H -H
So o o o o o
co co r~ o CD o
r^ r* CM tri i-1 co t-1
o CM * •— oo
CM
01
to c:
CO "!T
03 s: £
C t- CO "^
03 ^ r- 01
~ 2 S o
u ^ c o
05 JD JZ m CD 7=
C 0 0 XI C "5 J2
§ 5 =5 J. S E 1
^ cj ^ i2 en i^ x
en
oo
CO
CM
en
en
CM
en
en
r-' cri co
en en r-
Q ^J-
M- o 2
d d ^
en en ^.
cri cri <-:
o> en en
A A
T- co in
o o co
odd
V V
o 0 o
CD ;=: co
in S 06
en en CM
CD co r-
CO J-; CM
d d in
o o o
en ^1- CD
cri cri d
en CD CD
A
CM CO *—
O ?7 CD
d o ^
V
< 0.01 > 99.90
<0.01 > 99.90
<0.02 > 99.50
19.40 ± 18.10
' 26. 50 ±27.00
4.09 ± 3.71
(8) Anthracene.
(9) Bis(2 ethylhexyDphthalate
(10) Pentachlorophenol
o
CD
o
CD
> 99.80
O O CM CO CD CO ^*
"Z. 'd* CD co co co r^
A
.,, o oo r- CM en in
iO ^ CM o CM o CM
d o' d d d d co
V V
10 O «3- 't O O O
^ en o o CM CD co
d d d d d d co
V
o o o o o
CM O CD CO O O
Q ^ CO CO CO O CM
•Z. co ji 'fr •* o en
in o co co o o S
T— jl r- T— CO »— O
d ^ o m CM "-1 •*
v ^
in O O O O O O
T- CO *= CO OO •* CO
o r- o CD CM ^— r^
V N
o o o o o o
. in p ^ CD r*; co
Q CO CO CD O CN CD
2 CD CO CD CD CD CD
A A A A A .
ir> •— *- CM .r, •* CO
£? o o o " q q
d d d d d d d
V V V V V V V
r^- co ^J- o O ^ O
O CM O CO CO CM 10
o o o *~ o o ^-"
-H -H -H -H -H -H. -H
CD r- CD O o ^ O
O CD o tn co L" m
O O O CO IT- O CN
•i I 1 i »
Q) O o "O _s; (j
*" m _c O tD o c
< O O O 2 ''Z. K3
^ CM co -^ in co N.
CO
t~-
o
1
CD
-518-
-------
Table 4 Comparison of
Treatment Technologies for
SARM I Using TCLP
.Compound
(1) Acetone
(2) Chlorobenzene
(3) 1,2-dichloroethane
(4) Ethylbenzene
(5) Styrene
(6) Tetrachloroethylene
(7) Xylene
(8) Anthracene
(9) Bis(2 ethyihexyOphthalate
(10) Pentachlorophenol
(11) Arsenic
•(12) Cadmium
(13) Chromium
(14) Copper
(15) Lead
(16) Nickel
(17) Zinc
Soil Washing
2 mm to 250 um
Surfactant
(0 03
207.00 ±70.00
7.30 ±0.80
22.30 ±5.50
45.70 ±8.40
11.00 ±3.10
6.60 ±1.70
81.00 ±10.50
Solidification
Kiln Dust
19.40 ±.18.10 I <0.02
>99.7
(phthalate
10!
26.50 ±27.00
4.09 ±'3.71
0.10
0.41
0.09 ± 0.07 : I <0.15.
0.67 ±0.28
0.06 ± 0.04
3.50 ±1.80 '
0.26
<0.01
0.18
1.60 .±0.80 I 0.15
0.55 ±0.22
12.50 ±4.50
<0.06
1.90
99.6
90.0
96.5
ND
61.Z
>83.3
94.9
90.6
89.1
84.8
83.4
< 0.05
.0.13
>99.8
96.8
98.8
<0.15
<0.01
0.09
0.03
<0.15
<0.04
0.62
ND
> 98.5
-50.0
99.1
>90.6
>92.7
95.0
95.1
-519-
-------
TCLP values after treatment compared to before, hence the negative removal
efficiency values. This trend suggests that low temperature desorption may
not be appropriate for soils containing both organic and metallic contamina-
tion. Chemical treatment by KPEG reduced the chlorinated volatile compounds
(dichloroethane and tetrachloroethylene)by >98% and the semivolatiles by
about >90%. KPEG treatment was also effective in reducing metals; TCLP
values for metals in the residuals were overall 76.5% lower than for the
the untreated SARM. For soils washing, treatment of 2 mm to 250 m
fraction with surfactant reduced the volatiles in the TCLP by >98%, semi-
volatiles by >96%, and the metals by >83%. Stabilization utilizing kiln
dust reduced the metals by >95%. The apparent high removal rates for
organics (overall about 98%) following stabilization is thought to be the
result of offgasing during mixing rather than the results of chemical
reaction of the organics within the matrix.
The TCLP data collected around the five technologies, and four SARMs
have been summarized in order of decreasing treatment effectiveness as
shown in Table 5.
DISCUSSION
The TCLP system for evaluating the potential of a waste to release
hazardous contaminants was based on manipulating laboratory extraction
conditions until the results matched those from a pilot-scale system of
lysimeters containing 90% municipal waste and 10% industrial waste.
Its application to evaluating residues from different soil treatment
options, or judging treatment efficiency of CERCLA soils should be
approached with caution, and done on a case by case basis.
Each CERCLA site's soil will be different in some form and to this
extent the degree to which the TCLP will extract each compound from the
soil will also change. The attenuation of the individual compounds in
the untreated SARMs by 70 to 95% places a terrific burden on the analytical
techniques used to analyze the TCLP extract for two reasons. First, in
many cases, even though the concentration of a given contaminant may be
hundreds or thousands of parts per million in the untreated soil, the con-
centration produced in the leachate may lie at the fringes of the analytical
detection limit; when compared to a leachate value derived from the treated
residue, the two numbers may be virtually indistinguishable from each
other. Thus the limits of analytical detection can prevent a true, picture
from being formed regarding the effectiveness of a particular treatment.
This type of condition can be clearly seen in the data for arsenic and
chromium. Second and perhaps more importantly in many cases the treatment
efficiency will be based on the difference between two small numbers. This
was very evident in the metals data in Table 4 as well as all the metals
data collected on all the synthetic soils. The solubility of inorganic
compounds in an inorganic (i.e. aqueous) solvent system is the question
that must be dealt with when using TCLP, and this data indicates that
using this approach for evaluating treatment effectiveness could be risky.
However, when approaching TCLP from a health and safety standpoint
with impacts on the environment being quantitated, the use of TCLP would
-520-
-------
c
a)
4->
CC
0)
I -
+-> (/)
5 w
5 «"
15 c
> g>
in _>
+3
a. o
in
uouonpay
cc
uoijonpsa
%
cc
uojjonpaa
5
cc
uoijonpay
cc
w
r-m__cotoo
Q.W.O ^ « w "S "§ "S "Jjj ' — ~^
co to to •* ^T"o rl
c*>o4cor>'^"(Dor»tN
A
(1) Soils Washing +2 mm chelate
(2) Soils Washing +2 mm surfactant
(3) Solidification Lime/Fly Ash 28 day
(41 Soils Washing *2 mm water -•:•
(5) Chemical Treat KPEG Test 11
(6) Soils Washing 2 mm to 250 um chelate
(7) Soils Washing 2 mm to 250 um surfactant
(81 Soils Washing 2 mm to 250 um water
(9) Solidification Kiln Dust 28 days
•* CN co in in og
CD co 10 m \n n
(1) Soils Washing *2 mm chelate
(21 Solidification Kiln Dust - 28 days
(3) Solidification Lime/Fly Ash - 28 days
(41 Soil Washing +2 mm - water
(5) Chemical Treatment • KPEG Test 11}
(6) Soils Washing 250 um chelate
to tri -s-" •<* csi rvi ,-• 05 T-: r<
A A A
(1) Incineration
(2) Soils Washing »2 mm water
(3) Soils Washing *2 mm chelate
(4) Soils Washing +2 mm surfactant
(5) Soils Washing 2 mm to 250 um water
(61 Low Temp Thermal Desorb g 550° f
(7) Chem Treatment - KPEG - Test #1
(8) Low Temp Thermal Desorb g 150° F
(9) Soils Washing 2 mm to 250 ug water
(101 Soils Washing 2 mm to 250 ug chelate
O> Oi O) Ol O> O O O) CO CO
A A
Chemical Treatment - Test J1)
Solidification - Kiln Dust - 28 days
Incineration
Soils Washing *2 mm water
Soils Washing 2 mm to 250 um - surfactant
Soils Washing »2 mm - surfactant
Low Temp Thermal Desorb g 550° F
Solidification - Lime/Fly Ash - 28 days
Low Temp Thermal Desorp g 150° F
1 Soils Washing 2 mm to 250 um - water
I 1
SS£S£SS£$$
(1) Soils Washing *2 mm chelate
(21 Soils Washing +2 mm surfactant
(3) Soils Washing *2 mm water
(4) Solidification Port Cement 28 day
(5) Soils Washing 2 mm to 250 um chelate
(6) Soils Washing 2 mm to 250 um surfactant
(71 Solidification Lime/Fly Ash 28 day
(8) Soils Washing 2 mm to 250 um water
(9) Solidification Kiln Dust 28 day
(10) Chemical Treat KPEG Test *1
*u>r-«-,-h,mo
cnmcooococomn
(1) Solidification Port Cement 28 days
(2) Soils Washing 2 mm to 250 um chelate
(3) Soils Washing *2 mm chelate
(4) Soils Washing »2 mm water
(5) Soils Washing 2 mm to 250 um water
(6) Solidification Kiln Dust 28 days
(7) Solidification Lime/Fly Ash 28 days
(8) Chemical Treat KPEG - Test *1
A A A
(1) Soils Washing <2 mm - surfactant
12) Soils Washing *2 mm chelate
(3) Soils Washing *2 mm water
(4) Soils Washing 2 mm to 250 um - chelate .
151 Solidification - Lime/Flv Ash 28 day
16) Solidification, Portland Cement 28 day
(7) Low Temp Thermal Desorp g 150° F
(8) Soils Washing 2 mm to 250 um surfactant
(9) Solidification - Kiln Dust - 28 day
;10) Incineration
iTixj-'oioor-'Cor^o
A A A A
C
en
O
S. | „
™ co c E •§
•S 01 ~ __ 3 TO
' ' o at en » IS
1 s 1 1 1 1 1 *
1 1 If if 1 1
"o ^ "o '5 "o '6 "o .c
= £S«±!2<£ri2
S3|!JB|0A
S3|l)e|OAjUJ9S
-521-
-------
be very useful. Table 3 shows the proposed TCLP Regulatory values
for eight of the compounds listed (the other compounds havejno current
regulatory value proposed). The data in Table 3 indicate that the
solidification binders of Portland cement and lime/fly ash when used on
SARM I could not bring the TCLP values for chlorobenzene, 1,2-dichloro-
ethane, tetrachloroethylene and pentachlorophenol below the proposed
regulatory limits implying that these compounds could still be released
from the landfilled stabilized mixture in sufficiently high concentrations
to have potential adverse impacts on the surrounding environment. The
data in Table 4 also show that for SARM I low temperature thermal desorp-
tion at 150°F could release harmful concentrations of chlorobenzene,
dichloroethane, tetrachloroethylene, and excessive cadmium .could be re-
leased from the residuals regardless of the treatment temperature if land
disposed after treatment. None of the residues from the other technologies
had releases that would be flagged by TCLP standards.
Figure 3 expresses the TCLP data from Table 4 in bar charts by techno-
logy This data can also be rearranged to study the treatment by compound
class, and this is shown in Figure 4. When displayed in this manner,
specific CERCLA problems can be isolated and the best treatment options
selected.
A logical extension of the TCLP work is to study how it compares with
the data determined by total waste analysis (TWA) of the untreated SARMs
and the tested residues. A series of parallel data sheets were developed
for TWA that corresponded to the same sets of treatment options outlined
for TCLP. Table 6 is a summary chart of the treatment efficiencies as
measured by TWA for the different technologies in in decreasing order of
effectiveness. Generally speaking, the thermal technologies did well
against the organic fractions, chemical treatment and soils washing did
well on the semivolatile fraction, and soils washing and solidification
did well against the metals.
Figure 5 compares the effectiveness results of both TCLP and TWA for
SARM I, for each technology by class of compound. For the volatile com-
pounds, the TWA and TCLP data for treatment effectiveness were very close,
and it didn't appear to make much of a difference which method was used
for measuring treatment effectiveness. The TCLP/TWA percent effectiveness
values for semivolatile organic compounds and metals appeared to be mixed,
with generally higher effectiveness values associated with the TCLP data.
Overall, TCLP as a measure of effectiveness gives at least equal and often
higher results than TWA despite the fact that the initial concentration
found in the TCLP leachate from the untreated SARM were more dilute than
the TWA data by a range of 2 to 200 and some of the metals and semi volatile
compounds were near the quantitation limit of the analytical equipment.
These findings, of course, are based on the treatment and analysis of
residues from a freshly prepared synthetic soil. Comparable studies uti-
lizing aged and weathered soils from actual Superfund sites are necessary
to put these results into proper perspective. Such studies are currently
in progress and results are expected to be available in late 1988.
-522-
-------
cc
CO
c
JO
Q.
O
c „
CO CO
£ £
CO
o
c
o
_«
™
CO
Q.
o
o
Q_
3
G)
iZ
o *,
-P
-------
% OL"Z.Z. -
duuei MCT]
etals
09'9Z.
% O^'SS
S|ios
O0't6
% OL'26
cc
<
to
&
(A
W
JS
O
T3
§«
o o>
a. t
E i>
,9 E
JQ
C
O
.J2
1
o
o
0)
13
0)
J2=
a.
O
09'179
MCrI
CO
0}
r=
*j
JS
I
(U
CO
% OS'06 D3d>!
02'96
08'86
08'66
.2
-rr
% 06'68 JBSJ1 IBOiuusqQ
% 09"Z.6 uoijB3Uipi|OS
% 0£'86 BUJLJSB/\/\ s|jog
% LZ'66 UOJIBJSUIOUI
% £3'66 IBUiJaqj. doisi MO]
1 1 1 1 I
> O O O O C
3 00 CO •sf CM
' 1
3 CM
uoijonpaa %
-524-
-------
o
c
o
o o
•4=(/3
£"*-
(0 O
= +-
< c
«_ ®
"S £
c«
O 0
w i±;
•c !^
3s
LO
1
O)
o
o
% 01/26 - dlDl - isna u|!» -
- VM1 -
% 01/Z.Z.- - d"101 - uoj
ajniBJ8dai8j_
% 06' II- - VMJL - uoiidjosag |eujJ8qj_
00'176 -
- UOjiBJ8UIOU|
% OZ.'Z9 - VM1 -
% Of 88 - dlOl - 6uWSBM s,,oS
'Ml
% OL'9L - VM1 - SUILISBM s|jog
% 02 "9Z. - dlOl - JU9UUiB8JJ_ |BO!UI9t|0
- % OL'2L iU9UUJ.B9JJ_ |BOIUU9l)0 - 93d>|
- 93d>l
o
oo
o
CO
I
o
o
CM
-525-
-------
097.6 d1Ol -
92'86 VM1 -
"o
I
cu
1-5
CO O
= 4- —
< c w
•si!
— t; JS
§8.9
§
O)
% S3'66 dlOl - uojidjosaa
% 6Z.'66 VM1 -
0£'86 dlOl - ui
% L8'66 - VM1 - UI
duiai MOTJ
% 06'68 dlDl lu8uiiB8.il
96'66 VM1 - 1U8LU1B8J1
- dlDJL - uo!iBJ8uiou|
66'66<- VM1 - uo!iBJ8Uiou|
o
00
o
CO
o
o
CM
%
-526-
-------
% 08'86 - d101 uor
% OZ.7.8 - VM1 -
O
o
O
+•«
O
ol
"55 :=
o o
••PC/)
— **- ^-,
JS ° w
= +- CD
< c«=
c fi
% 09*179 - dlOl -
% 09117 VM1 -
02'96 -
017'L8 - VM1 - 6ui
|BLUJ9L|x
i
I
% OC"06 dlOl
IBOIUIBUJQ -
% 09'S6 VM1 - IU8LU1B9JL
08'66
-------
|jl
if1
!
<
1
JT =ii
T3 S§>S
22 io|
3 w 55>
OT 3*
CO
CO
2
«
w
•«••
c
OJ ""
0
ij=
LU
+•» g w
1 i|i
^•J iff >
CO « 5 1
CO jj.3
£
^
>.
CO
E
to
(O £
CD
& 0
.Q
f- .2
COOOOOOOoOQ
encncncnencncn*cncnco
A A A A A A
1 2 I
^ £ « S
c T: o> — co
tjjiU ti
to S T; = a in co"
^5-ggool^«3=0
HESEooEuJdS
ST???™^**^"
tJjOJOiroPlOSro O)
2 .e .S .S .5 .c g £ c .=
P^X:J=J=J=-£|— 5^=
s|ilii|i|f
c=J2-2J2.<2.£_ §:9J2
S *§ '6 '5 'o *6 'S £ T5 "o
0tOC/3C/5COtOtOOCrtCO
OJOOOg0OOgO
OJ OT O] O) O) O} O) O? U> UU
A
O3
s s
I JS a-
lea- € ^ -
| 1 I S S l.!l
i.8 MtssNElE
*^ S23°"=J.=
i ECcc°Soco
|ESll,iSs|s
'^g««sf a Bl „
S?ra***csc3)co)cicc:
•J .g s c e § l^ 2 .0 2
'S 's .c "o * "o "S ^^ "o *o "o
U)l/)OC/>C/>WO«tOt/>
to O O o o O g*g r-
cria>coo5cc>raKcDct5
OT DJ OJ O) OT O* U' W> U>
A A A A
» I
to S i ^;
1 i 1 s-l
1 • • "° CB "-
| | | S -S | 0
1 •§ •« « s * * a g
l55D'£v:' .£
_g-JJ o^S OS1"
§-S-§-5= cxg-g g g.
Ifl!!^!^
S « 9« *£ ^ 5 °* — * 5
SOOOOQ-^OJ^Q
CCAC/>C/>CA±lOCOCJ_d
^cNC3«3-mcor^coojS
cnococNcNOcnooo
o> 01 05 o O) o> en oo co co
03 en en en o 01 en en en en
A A A
.2 § o o
^ $ o o >.
' 3 If3 UJ (O
S § a au si co
2 J o g « « 7
I g S « 8 g «
l^pSS^SacM
e3ESill^S
^g^^c^il^S
o>^ esrarajSiE c^
_cr*-ercch^'^o*-
g2 g2EE o.o.-g «
ll^illlll1^
|f §f?±^ti 1
O O
A
2 mm surfactant
+2 mm chelate
fl
S- jn
S "
o R
S §
Treat KPEG 11
Treat KPEG #2
1 Chemical
2) Chemical
5 g
A
a
E
CM
cr I
*(~ J2
"o 'o
B E
en en
A A
Incineration
Soils Washing +2 mm surfact
o o o
s-
•0
s
a s 1
a a E
fe s 1
S 1 "c
*= N= .0
| | 1
S 6 1
§ CO
en
+2 mm surfactan
+2 mm chelate
0) C?
in J2
1 w
— CM
0 0
CO CO
2
•i
•s
hing 2 mm to 250 um
shing +2 mm chelate
s s
« »
cl S
?=•
Soils Washing +2 mm w
Soils Washing +2 mm c
;= CM
en en
"c
Soils Washing *2 mm water
Soils Washing +2 mm surfact
T— CN
o
r-
en
+2 mm water
O)
1
B
CN
•*•
_w
'6
C/3
m
s
I
1
5
CN
'5
en
CO
s
CO
5
m
Soils Washing 2 to 250 uin v
CO
S
en
1
-P
2 to 250 um sui
01
.E
JZ
1
r3
'6
CO
0
CD
$
shing' 2 mm to 250 um
S
s
o
CO
??
g
*—
£
CO
1
i
Soils Washing 2 to 250
*r
o
to
£
c
u
•g
I Soils Washing 2 to 250 um £
^
o o o
r* en tn
§ s §
s
2 to 250 um wa
Jme/Fly Ash
Ciln Dust
ra — —
C c C
S ° o
III
=5 o =5
to to to
000
oi co' co'
tion Lime/Fly Ash
shing 250 um chelate
tion Kiln Dust 28 day
1 S 1
1^1
tn co 03
s ° 3
ui i"j "j
"c
0 u.
£ 3
1 1
1 1
Soils Washing 2 to 250
Soils Washing 2 to 250
Incineration
in co r»
So o
CM r*.
CO O CO
to **• co
&
T3
) Solidification Lime/Fly Ash 28
} Solidification Kiln Oust 28 da
1 Incineration
£ S £.
§
en
Sfc
C
h-
03
S
CO
9
CJJ
Chemical Treat KPEG 11
S.
-528-
-------
APPLICATION OF LOW-TEMPERATURE THERMAL
TREATMENT TECHNOLOGY TO CERCLA SOTT.R
U.S.
Michael F. Szabo
PEI Associates, Inc.
11499 Chester Road
Cincinnati, Ohio 45246
Robert D. Fox
IT Corporation
312 Directors Drive
Knoxville, Tennessee 37923
Robert C. Thurnau
Environmental Protection Agency
Hazardous Waste Engineering
Research Laboratory
26 Martin Luther King Drive
Cincinnati, Ohio 45268
ABSTRACT
The U.S. Environmental Protection Agency (EPA) is evaluating best demon-
strated available technology (BOAT) treatment levels for contaminated
Superfund soils and debris for purposes of CERCLA/SARA compliance with the
proposed 1988 Land Ban. The purpose of this project was to investigate the
capability of a laboratory-scale low-temperature thermal desorption technolo-
gy to remove volatile and semivolatile contaminants from a synthetically-
prepared soil spiked with predetermined quantities of these contaminants.
A laboratory testing program was initiated, consisting of 15 separate
bench scale tests (10 in a tray furnace and 5 in a tube furnace). The EPA
synthetic soil of volatile and semi-volatile compounds (high and low, and a
low concentration of metals in each soil) was tested at 150°, 350°, and 550°F
tor 30 minutes at temperature to determine the effect of this range of tem-
peratures on removal of the contaminants. The tray furnace was used to
fro^ir *? ef^ctivenes.s °f thermal desorption in removing contaminants
from the soil. The tube furnace was used to provide data on the concentra-
tions of contaminants in the off-gas for comparison with soil concentrations
in an attempt to establish a material balance. This paper describes the
thermal desorption test equipment and operating procedures, sampling and
analysis procedures, analytical results, quality assurance and quality con-
trol, and conclusions drawn from the results of the study.
-529-
-------
INTRODUCTION
INTRODUCTION
sponse (EPA-OSWER), is collecting data on "J™**^ ln the anticipation
levels instead of being exhumed,
landfill. One of. the technologies being
Low-Temperature Thermal Desorption
this paper was to vestigate
ties of contaminants
for the regulations is
to 3* & ^
pability of low-
o'and semivolatile
spiked With predetermined .uanti-
SOIL DESCRIPTION
The contaminated soil that « .tested
gate Superfund soil containing a wide range
ly occurring at Superfund sites. The fl™^ . After considerable re
an acronym for synthetic ^^^"^^^8 nationwide, EPA
search into the types of soils fo ^ £ £*« volume clay (a matrix of
decided on a soil 'omP0*f *°^%f ±^ent silt. 20 percent sand, 20 percent
montmorillinite and olinxte) , « Percent^ . ^P air_dried to mlnlmize
bitches Ja standard truc.-mounted cement
mixer.
Aple8clibea groUp of c^icals found
rinB at SuperEund sites «« then added to
SMller-Scale
^
acetone, chloroben-
Salts or oxides of the
Mum, «B«^
tamination in
Due to the wide range of con
diffrent SARM formulas containing either
for use in various
-530-
-------
TABLE 1. SARM SOIL TARGET CONTAMINANT CONCENTRATIONS
SARM-I
high,
SARM-II
low,
Volatiles
Ethylbenzene
Xylene
Tetrachloroethylene
Chlorobenzene
Styrene
1,2-Dichloroethane
Acetone
Semivolatiles
Anthracene
Bis(2-ethylhexyl)phthalate
Pentachlorophenol
Metals
Lead
Zinc
Cadmium
Arsenic
Copper
Nickel
Chromium
3200
8200
600
400
1000
600
6800
6500
2500
1000
280
450
20
10
190
20
30
320
820
60
40
100
60
680
650
250
100
280
450
20
10
190
20
30
Experimental
The principal test equipment was a Linberg Furnace, Model 51848, with an
electronic temperature controller and 1600-watt heater system. The oven is
of double-shell construction with interior surfaces made of Moldatherm, a
molded aluminum-silicate insulation material. This oven, which is capable of
operating up to 1100°C, has a relatively fast heatup rate because of its low
mass.
The untreated sample materials (SARM) were analyzed by EPA Methods 8240,
8250/70, and 6010/7000 for the indicator compounds listed in Table 1. The
characterized soil was sampled, weighed, and evenly distributed on an Incolay
tray. The tray was inserted into the oven and connected to a nitrogen purge
of 90 cm /min. Two thermocouples were used to measure temperature at 3 cm
above the soil surface at the center of the tray.
The soils were treated at temperatures of 150°F, 350°F, and 550°F for
thirty minutes, removed from the oven, cooled for one hour, and sampled. The
-531-
-------
asr
temperature desorption.
The conditions under which the experiments were conducted are summarized
in Table 2.
Results —
operating condition on each soil.
Volatiles-With the exception of acetone, the data generated by low-tern-
slowly levels off as temperature increases.
*. vogues in U» II
-------
TABLE 2. PEI SOIL THERMAL TREATMENT EXPERIMENTAL SUMMARY
lies at teap = 2$ iinutes
EXP i
PEI-3
PEI-4
PEI-5
PEI-7
(b)
PEI-15
PEI-24
PEI-8
PEI-9
(c)
PEI-18
PEI-11
PEI-14
(d)
PEI-37
ONOLYTICOL
SOMPLE *
555-12-20-5416
555-14-20-5416
555-16-20-5416
555-20-20-5416
555-28-30-5416
555-38-20-5416
555-77-20-5416
555-22-20-5415
555-24-20-5415
555-24-1 R-5415
555-26-20-5415
555-28-20-5415
555-36-20-5415
555-36-1 R-5415
561-7-20-5415
SOHPLE
TYPE
Soil
Soil
Soil
Soil
Soil
Soil
Soil
Soil
Soil
Soil
Soil
Soil
Soil
Soil
Soil
IT
SOIL!
5416 f
5416
5416
5416
5416
5416
5416
5415 g
5415
5415
5415
5415
5415
5415
5415
EXP
TYPE
Tray
Tray
Tray
Tray
Tray
Tray
Tray
Tray
Tray
Tray
Tray
Tray
PURGE
ROTE
cc/ain
98
98
98
98
98 a
98
98
98
98
98
98 a
98 a
TOR6ET
TEMP
deg F
558
558
558
358
158
358
558
558
558
358
158
158
KOX
SOIL TEMP
deg F
548
551
551
341
161
344
558
558
552
358
152
147
STORTINS
SOIL WT
g
72.98
73.27
74.15
75.74
74.26
185.68
41.62
41.22
41.74
41.97
41.44
(e)
TREOTED
SOIL HT
g
68.59
68.28
61.18
63.88
66.99
87.86
39.13
38.76
39.25
39.85
39.54
* NT
LOSS
17.4
17.8
17.5
16.7
9.8
17.6
6.8
6.8
6.8
5.8
4.6
CONDENSED
VOLUME
Hi
nc
nc
nc
nc
nc
nc
nc
nc
nc
nc
nc
a gas volume exchange rate equivalent of larger furnace to smaller unit
b duplicate sample from PEI-7
c duplicate sample from PEI-9
d duplicate sample from PEI-14
e sample from composite runs for TCLP
f IT soil identification number for PEI soil ID SflRM-I-1 (high organics/low metals)
g IT soil identification number for PEI soil ID SORM-II-1 (low organics/low ratals)
nc not collected
rev 81/27/88
-533-
-------
S
o
IT)
CO
fc
UJ
01
co
UJ
a
s - a-
01 S. ro
«= °- ^:
™ * a * ^ §
vi gj *-« OJ •* «**
£-,„»
' -" S S3 " eZ
™ -ri """_ *Q . r— •—- " •*
e
S 2 - S? SS" S " " S i S 3
ij
12 «* i^_ r^ flo cr> irt to
g --.I". *<*. - =-K - cor-^
3 v ft S 8 s I a
."? *~1 .
i~"s~~5 « 2SSSSS5S
| i i » -^~~SR~S
in »H-^- r*3CDOJf*^<-D'^HCP
S ^3 u5u3«5j ""oaSPs
s 5 3 g. s - *.
I
•s
s
= s. ^s Sgid S'os^iBS^s K
S8 a --.-. -JSJ • g -•"" li
•? =si$ . .
s s
B 5 S •
0 1 1-8
II "-B
gl 3 "3. o
b^ I S3 •§ „
s K I! a «,£
.534-
-------
g
ea
en
1
i--
ro
1
5
U_ -tr
ITS ^ ~^*
^ *
-tf
S i
U_ CH
IS S
£ ~i~
«c .
^- j£ £j£
= 1
00 _ .5
f^ """ II
§ §S
.- i: s
OO 5c ^ ,1.
LU 8 | tf
>- -I
g § _T
1- g -=£
en en
En u. cri
LU ^S « ^ ^
_J >. ^ R - y
CO S g
| —
i
LU
0.
c £^ '
'£ "I
m — i Q.
"S
cl
II
1
II
!i
in !i
if
= •
I ii ^
T ii P
a !i I
eg a B sss aasss-'-
- <=• " - s ~ ** • "^ ~ rf si g ~ ^
CD — 2 '^ co<5>S^ K3§? "^I^SB01^
-~ =" ~-a SS?O~~»K
^^S^^g^ ^Sr^ , ^^Sgg^S
— « *^ cu D^ in m r-- 9 cu c^ S *»--*- CQ u^
S-'-g.* ^^§ ^^^c^^P]ri
^_ ^ g -a- •* -a-
§Ii3gIi in SfflSag^S
8 •» = '
Sg^g^gg 2g* SlSSffiS0^^' '
~""
III 5i^g,,
z s z I
igoj-a-So, cotnto |j
S ggg — ro «r r-i ro P? r.
II
o ""* "^ "^' *-> cxj ro cu cu in
n
§5S§Ssi § i i ' • 2 s s S i ^ R f
""* ^ !!
i !!
UJ J
i— i ii
f^ i
£ ii
I 1 5d_ '
u g u g . • a S a - ii
ggSWzi.Tg |i
. S5--tS!iis-Siisffi SilseiSsI 1
i
|
!
j
|S
i R
11 -
-------
§
S
O
s
CO
CO.
s~
o
2
CD
Q.
I
on Technolog
CD
CO
E
CD
JZ
U-
b
en
O>
o
o
in
I &
<
oc
01
§ e=
eo
P
o
o
S- •<->
0)
CU =5
M- S-
O OJ
O-
E o
3 r—
4-
cn
fC E
•r~
C/5 (/)
fd 3
>>*-*
O
O (/">
O)
O
S-
Q I/)
•r- CU
O 5-
OJ -t->
c— (O
Q.
S-
o
O 0)
^> "I3J
o
o
o
c>
(SaiLLVlOA TW JO 30VU3AV) NOIlOnCBH %
-536-
-------
to
CD
CD
'E
CD
CO
- *
& e-
£ 8
ctg
E CD
o 2
.2 £3
O CD
C Q.
a E
m CD
CO
CO
c
CD
>
J
CC
E
c\
(S31I1V1OA nv do 30V33AV) Nononaau %
O)
-537-
-------
soil of SUm I. This trend was not as evident in Table 4
the elevated temperatures.
The TCLP analyses were performed on the residues from six tray-furnace
M ™«K 4o°F and 550°F for both SARMs I and II) . These results for
Kr£ sf-ssT-j-n^j^. SL-L? ;
the values are much higher than for the 550°F and 350°F analyses.
For the
SARM
run, chlorobenzene, 1,2-dichloroethane, and tetra-
is the only metal that was higher than its regulatory limit.
Overall, these TCLP results indicate that the synthetic-treated residue,
with the exception of a few contaminants in residues treated at the 150 F
level! is no?Pvery leachable; i.e., most of the contaminants remain bound up
with the soil particles.
DISCUSSION AND CONCLUSIONS
It appears fro. the data on volatile* and ...iwlatil.a ««t_lo.-t.,era
iffirss sjs
ture runs.
It is unknown at this time whether the treatment difficulties experi-
lid used for the tray after heating could have had residual acetone after
the cycle.
b
with the exception of a few contaminants at 150°F, which were much higher
than at the 550° and 350°F levels.
-533-
-------
or
ac
G"
o
co
o
LU
H™*
LU >— i
^s:
z cc:
co cc:
*-" O
CO U_
•
CO i—
O£ CO
^
CO
CO
^-*
^^
e
£
^^
to
u
•^
c
ro
cn
L_
0
CO
^^
r—
O
O O O CM O O LO
^J- ^-1 t— 1 CO «—l t-H LO
O O
u
CO T-
co c c
c: co to
•C ^>*"— ** &_
4-> jr •— C3
C O C CO 4J CO
CO I- CO O O i —
isl O N I- 4J -r-
C r— C O 4->
CO .C CO r— fO
CO J3 U J3 O) .C CO i—
C O -r- C U CO O
o s- ca i — corec >
4-> O 1 >> i- S- O) •>-
CO i— CM -C >>*J i— E
«:O.-ILUCOH-X co
CM CO CM
o o o
o o o
5* VO t-«
o o o
• * •
o o o
o o o
C> 0 0
O O Cft
LO LO CM
CO t-H
LO
1 1 •
CO
*
f—
(C
c-
x:
o.
^v^
r«— r1™"
>> 0
X C
OJ CO
i— O.
CO ^ ^
C 4-5 O
CO CO i— i
U 1 -C
IO CM U 1/1
S_»— co i—
_f t \ —•
4-> CO C +->
C -i- CO CO
<: co o. s:
to co o co co r-» LD
»•"• r^* r™i co r^ LO •
O t-l O tv. CO »-« CO
•* CO VO CM CO O O
CO CM O CM O Ol CO
O «-l O CO CM O 1^-
CM
to «s- o cri r>« «d- o
«— i co «-H »— i LO o> vo
O i-H C3 CM •-! C3 VO
CM
CO «— 1 LO CO CO O ^~
<3- CO CO CM CM LO
•-< CM CO
o o o o
• • • 1 • 1 1
LO t— 1 LO LO
B
cj E 3
•i- 3 •!- $_ r—
C -i- E CO CO
CO E 0 0.-O J* 0
to X» J- CL (0 o C
s-
-------
*£
LU •—<
cc
II
CO 02
H-
LU
CO
8 en co o co op o
oo
- o o
co
i in
in
CVJ 10 »H
o 1-1 vo
• • •
O O •-(
ir> oo o o ^ vo o
r-l CvJ «-H OO OO CO CO
o vo
cu
CT» t-l t—< i—!<-*«—' »~J
T-H O O O O O O
O O O O O O
O «-l *-l T-H r-l «-< *-<
t-H O O O O O O
vo oo r~
oo r-« c\j
• • •
o o m
t-l CO in
o o co
• • •
o o o
O CM
' CO
to _ - - -
fli CO CSJ *i^ f"i i*' ^~x ' ~*
t. cococo«- -^
ns O>
Q) f—
a: a>
a;
i-l O
i i • i
o
vo
•
CO
o o o
• • •
in t-t in
0)
-t->
CO
u
IO
o
a>
it)
O) QJ
C O
MO
c: i—
a> -E
po
»-•!-»
o —
O «— ' LU CO I — X
o
(C
en
o
O)
o
>
O)
co
>> o
>? c
QJ O)
J= f
i— O.
>> p
a) .E s-
c -4-> o
O) Q> i—
O I f
(O esj o
S_ (O
£. -t->
+j to c
c ••- a>
<: co a.
O)
C T-
QJ E -
w -o t- a. 10
i_ «o ^: o a>
-------
The volatiles distilled off by the low-temperature desorption process
would need to be treated, reclaimed, or recycled. If the residu'e soil con-
tained metals, the soil would also need additional treatment, such as solid-
ification. Both of these factors will add significant costs to a project,
but these costs were not factored into the evaluation of the technology for
low-temperature thermal desorption.
REFERENCES
1. Federal Register, Vol 51, No. 114, Friday, June 13, 1986/Proposed Rules.
-541-
-------
EVALUATION OF SOLIDIFICATION/STABILIZATION AS A
BEST DEMONSTRATED AVAILABLE TECHNOLOGY
by
Leo Weitzman and Lawrence E. Harael
Acurex Corporation
Research Triangle Park, NC
and
Edwin Barth
Hazardous Waste Engineering Research Laboratory, U.S. EPA
Cincinnati, OH
' ABSTRACT
This project evaluated the performance of solidification as a means of
treating soil from "Superfund" sites. Tests were conducted on four different
types of artificially contaminated soil which are representative of the types
of contaminated soils found at Superfund sites. The soils were solidified
using the following three commonly used solidification agents or binders: (1)
Portland cement, (2) lime kiln dust, and (3) a mixture of lime and flyash.
At 7, 14, 21, and 28 days after soil and binders were mixed, samples of
the solidified material were subjected to Unconfined Compressibility (UCS)
testing. Samples of those mixes that had a UCS minimally greater than 50 psi
(pounds per square inch), or which showed the highest UCS below 50 psi, after
14 and 28 days were subjected to Toxic Contaminants Leaching Procedure (TCLP)
and Total Waste analysis (TWA). The 50-psi level was chosen based on
guidance from the Office of Solid Waste and Emergency Response (OSWER
directive No. 9437.00-2A). The principal goal of this program was
development of screening data in support of the Office of Environmental
Emergency Response (OEER). The schedule and experimental protocol were
therefore geared to satisfying these regulatory demands.
As an ancillary goal, the results were analyzed to determine if any
correlations could be obtained between the degree of toxicant immobilization,
as determined by the TCLP, and the following other parameters: (1) UCS
results, (2) curing time, (3) contaminant level, (4) binder type, and (5)
water concentration.
The impact of these parameters on the leachability of the contaminants
is complex. It was recognized at the start of the testing that this
-542-
-------
abbreviated set of experiments was not likely to adequately address these
interactions. For completeness, however, these correlations were attempted.
The following results were observed:
1.
2.
3.
4.
The water-to-total-solids ratio appears to be a better measure of
the amount of water needed to solidify a given mix than the
water-to-binder ratio that is commonly used. This was clearly the
case for the "Synthetic Analytical Reference Matrix" (SARM) with
these binders. This needs to be confirmed on other systems.
Solidifying the SARM resulted in significant reductions in the
amount of metal salt contaminants released, as measured by the TCLP.
Solidification did not appear to result in a similar reduction in
the amount of organic emissions from the SARM. Because of the large
losses of organics during the mixing process, the effect of
solidification on the organic leachate via the TCLP could not be
determined. The volatile and semivolatile organic contaminants did
appear to decrease during the solidification process; however, this
decrease can be attributed to their release to the air during
processing and curing.
No correlation between UCS and the results of the leaching tests was
observed. .
Mo
No.
9R~u Under Contract No- 68-03-3241, Work Assignment
2-18 from the EPA. This paper has not been subjected to agency peer
contents do not necessarily reflect the views and policies of
the U.S. Environmental Protection Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation of use
-543-
-------
The Hazardous Solid Waste Amendment Act (HSWA) of 1984 requires the U.S.
Environmental Protection Agency (EPA) to develop treatment standards or
treatment methods (called "Best Demonstrated Available Technology" or "BDAT")
for listed hazardous waste before it is land disposed. Treatment methods are
to be developed which reduce the toxicity or the likelihood of the migration
of the hazardous constituents in the waste. The Superfund Amendment and
Reauthorization Act (SARA or Superfund) requires that remedial actions meet
all applicable, relevant, and appropriate public health and environmental
standards. Therefore, the Superfund program must be consistent with the BDAT
approach when disposing of contaminated soils and debris from Superfund
sites.
In order to satisfy this requirement, it is necessary to establish the
level of performance (as determined by the above criteria) that different
technologies can achieve on materials from Superfund sites. The Office of
Research and Development (ORD) assisted the Office of Environmental and
Emergency Response (OEER) in evaluating five technologies which could be the
BDAT for soil and debris from Superfund sites. The five technologies were
1. incineration
2. low temperature thermal desorption
3'. KPEG reagent for dechlorination
4. aqueous soil washing
5. solidification/stabilization
This project evaluated the performance of solidification/stabilization
as a "BDAT" for treating soil from "Superfund" sites. Prior to starting the
discussion, it is necessary to define solidification and stabilization.
Solidification is a treatment process whereby a waste is converted into a
solid material that does not flow and which will support a load.
Stabilization is a process that results in the reduced mobility of the
contaminants in a waste or other material. Stabilization may be accompanied
by the solidification of a waste and, as a result, the terms are often
interchangeable; however, they do denote different phenomena.
For this program, four different types of artificially contaminated
soil, which are representative of the types of contaminated soils found at
Superfund sites, were solidified using three commonly used solidification
agents or binders. The products were subjected to Unconfined Compressibility
(UCS) tests, and each blend of soil and binder which had a UCS minimally
greater than 50 psi (pounds per square inch), or which showed the highest UCS
below 50 psi after curing, were subjected to Toxicity Characteristic Leaching
Procedure (TCLP) and Total Waste Analysis (TWA) to determine the performance
of this type of solidification as a possible pretreatment method for
contaminated soils from Superfund sites. The 50-psi criterion is consistent
with Resource Conservation and Recovery Act (RCRA) listed waste regulations.
The binders used were commonly available generic agents which are
readily available. Other binders, both proprietary and generic, are
available and could enhance the stabilization process. There is, at present,
no set protocol for evaluating the efficacy of stabilization technologies.
-544-
-------
ssr.-s
EXPERIMENTAL PROCEDURE
^
SARM I — low metals, high organics concentration
SARM II — low metals, low organics concentration
SARM III — high metals, low organics concentration
SARM IV — high metals, high organics concentration
for the^pelSiS'.T? ST ^ SUbJected to a Wai waste analysis (TWA)
Tor the specified metal and organic contaminants that had been added to it- i
its preparation Table 1 gives the results of this analysis By wt of
comparison Table 2 shows the results of the multiple Twlanalyses on
actua l SARM'S ^ the Proce<^es used to prepare
Table 3 Th^fh t / S "*** tO conta™i«ate the SARM's are listed in
iabie J. The three binders used were Portland cement TVDP 1 CPfV n
dust (KD); and equal weights of technical grade Hmf and^lyash (LF)
TABLE 3. CHEMICAL IDENTIFICATION AND SOLUBILITY
OF SARM METAL CONTAMINANTS
Chemical Type
Lead sulfate
Zinc oxide
Cadmium sulfate (3CdS04 8EzO)
Arsenic trioxide (As203)
Copper sulfate (CuS04 5H20)
Chromic oxide (Cr203)
Nickel nitrate
Solubility in Water
Slightly soluble
Insoluble
Soluble
Slightly soluble
Soluble
Insoluble
Soluble to very soluble
the
P°^lafrcement wf ^andard Type 1 portland cement manufactured by
local
*"
of Tenn Luttrell ompany
. f lyash obtained 'from a
-545-
-------
r
TABLE 1. RESULTS OF TWA FOR SARM SAMPLES
RECEIVED FOR THIS PROGRAM
(rag/kg)
Analyte
SARM 1
High organic,
low metal
SARM 2
Low organic,
low metal
SARM 3
Low organic,
high metal
SARM 4
High organic,
high metals
Volatiles
Acetone 3,150
Chlorobenzene 330
1,2-dichloroethane 380
Ethylbenzene 3,350
Styrene 1,000
Tetrachloroethylene 710
Xylene
4,150
230
9.2
3.9
74
26
16
210
220
8.9
3.1
100
24
13
150
13,000
270
830
2,500
540
540
3,700
Semivolatiles
Anthracene
Bis(2-ethylhexyl)
phthalate
Pentachlorphenol
Inorganics
940
600
135
275
34
62
265
140
15
775
500
78
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
18
17
27
193
190
27
392
18
23
37
260
240
32
544
904
1,280
1,190
9,650
15,200
1,140
53,400
810
. 1,430
1,650
13,300
16,900
1,380
28,900
-546-
-------
TABLE 2. RESULTS OF TWA FOR SARMS IMMEDIATELY AFTER MIXING<1>
(rag/kg)
by PEI from main blend Prior to shipping samples)
Analyte
Volatiles
SARM 1
High organic,
low metal
SARM 2
Low organic,
low metal
SARM 3
Low organic,
high metal
SARM 4
High organic,
high metals
Acetone
Chlorobenzene
1,2-dichloroethane
Ethylbenzene
Styrene
Tetrachloroethylene
Xylene
4,500 (7)
320 (7)
380 (7)
3,460 (7)
730 (7)
480 (6)
5,720 (7)
360 (8)
13 (6)
7 (8)
120 (8)
48 (7)
19 (8)
210 (8)
360 (2)
11 (2)
5 (2)
140 (2)
32 (2)
20 (2)
320 (2)
8,030 (2)
330 (2)
490 (2)
2,710 (2)
630 (2)
900 (2)
5,580 (2)
Semivolatiles
Anthracene
Bis(2-ethylhexyl)
phthalate
Pentachlorphenol
Inorganics
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
4,390 (7)
1,830 (7)
270 (5)
21 (8)
27 (6)
30 (6)
260 (8)
270 (8)
39 (8)
570 (6)
350 (7)
140 (6)
40 (4)
18 (7)
32 (6)
32 (6)
280 (8)
320 (8)
40 (8)
680 (8)
270 (3)
270 (3)
30 (3)
690 (4)
2,380 (2)
1,260 (4)
9,550 (4)
15,100 (4)
1,540 (4)
34,450 (4)
1,920 (3)
1,920 (3)
80 (3)
540 (4)
3,790 (2)
1,400 (4)
11,250 (4)
15,680 (4)
1,550 (4)
28,660 (4)
Note: Values in parentheses indicate number of samples analyzed
-547-
-------
In the first step, the water content of the SARM's was determined by
drying them to constant weight and attributing the weight loss to water
removed by evaporation. The results are given in Table 4. This
measurement includes volatile compounds other than water; however, the error
introduced by this is insignificant for the purpose of this program.
TABLE 4. WATER CONTENT OF SARM SOILS
SARM
Apparent
Water Content,
1
2
3
4
31.4
8.6
19.3
22.1
The next step of the program was to determine the amount of water
required to form a satisfactory product. This was done in two stages. The
first stage was a preliminary series of tests conducted to get the general
range and the second stage was a finer set of tests to identify the
approximate "optimum" amount of water needed to achieve solidification. The
"optimum" was defined for this test to be that mix which gave the product
most resistant to penetration by a U.S. Army Corps of Engineers, Cone
Penetrometer test after 24 h.
More than 600 samples were mixed as part of these tests to identify the
amount of water needed to achieve "optimum" conditions. Unfortunately,
because of regulatory time constraints, the "optimum" had to be determined on
the basis of only one day of curing rather than on longer cure times that
would be common in normal operations.
The tests were conducted by mixing each SARM and binder combination with
different amounts of water. Each sample was then molded into a plastic cup
and allowed to cure for 24 h. At that time, the sample was tested, using a
U.S. Army Corps of Engineers Cone Penetrometer, to determine how much it had
set up. The preliminary screening resulted in identifying the minimum amount
of each binder that was needed to solidify each SARM. The second set of
these tests identified the "optimum" amount of water that resulted in the
strongest product for each SARM-binder combination.
The tests showed two things. First, they showed that water-to-binder
ratios were not good indicators of the amount of water that should be used to
form the SARM's into a monolithic solid suitable for hardness testing. The
ability of the product to set up could be correlated reasonably well to the
water-to-total-solids ratio (W/TS) of the mix. This ratio is simply the mass
of water used versus the sum of the solid component of the SARM and of the
solidifying agent. In virtually all cases tested, a W/TS ratio of 0.4 to 0.5
resulted in an acceptable product. The second fact which became evident was
-548-
-------
that regardless of the soil or the binder used, a W/TS ratio of approximately
0.4 resulted in indications of solidification of the material.
The next phase of the program was intended to determine the minimum
binder-to-soil (B/S) ratio which would result in a sample of solidified soil
with an unconfined compressibility.greater than 50 psi. Actually, with some
binders this UCS level could not be achieved with any ratio tested within the
30-day curing time set under this program. In that case, the sample that
achieved the highest UCS level was used for subsequent testing.
The B/S ratio tests were performed by mixing each soil (4 soils) with
each binder (3 binders) at three B/S ratios. Each day, one SARM was mixed
with one binder at the three B/S ratios. Table 5 shows how this was
performed. Six samples of each mixture constituted one complete set. From
each set, four were molded into cubes. Three of these were used for one set
of UCS tests, conducted in triplicate, and the fourth as a spare. The fifth
was placed in several glass jars with Teflon lined lids and was sent to the
laboratories for TCLP and total waste analyses. The sixth set of samples was
stored for future reference.
On the seventh day after mixing the SARM with the binders, Set 1 of each
sample was subjected to UCS testing. On the 14th day, Set 2 was similarly
tested in triplicate. The same was done with Sets 3 and 4 on the 21st and
28th days, respectively. On the 14th and 28th days, the laboratories
conducting the UCS and TCLP tests were called and asked which of the samples
had either minimally satisfied the 50-psi UCS requirement or, if none had
achieved 50 psi, which one had the highest UCS reading. The program resulted
in a total of 648 samples, as shown below:
4 Soil types
x 3 Binders
x 3 B/S ratios
x 3 Triplicate samples
x 6 samples at each condition
648 Total samples
The samples that were used for UCS testing were molded according to the
specifications in ASTM C 109-86. The procedure calls for molding the
material in specially fabricated stainless steel or brass molds which result
in cubes two inches (5.08 cm) long on each side. The molds must meet strict
dimensional and stiffness requirements. The resulting samples were allowed
to harden for one to four days at 70 °F (+10 °F) and 90 to 100% humidity,
then unmolded, and each cube was placed in a scalable plastic bag to cure at
room temperature until it was tested. Table 5 also gives the ratios of each
component (SARM, binder, and water) in each sample and summarizes the results
of the UCS tests.
In a few cases, different B/S ratios satisfied the UCS criteria at the
14th and 28th days. For example, Sample 30 satisfied the UCS criteria after
14 days of curing while Sample 29 satisfied them after 28 days. This
happened with several mixtures.
-549-
-------
TABIE 5. MIX RATIOS AND RESULTS OF UCS TESTS FOR EACH SAMPLE SET, SUMMARY
Sample
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
SARM
Type
I
I
I
II
II
II
III
III
III
IV
IV
IV
I
I
I
II
II
II
III
III
III
IV
IV
IV
I
I
I
II
II
II
III
III
III
IV
IV
IV
Binder
Type
PORTLAND
CEMENT
TYPE 1
PORTLAND
CEMENT
TYPE 1
PORTLAND
CEMENT
TYPE 1
PORTLAND
CEMENT
TYPE 1
KILN
DUST
KILN
DUST
KILN
DUST
KILN
DUST
LIME/
FLYASH
LIME/
FLYASH
LIME/
FLYASH
LIME/
FLYASH
B/S
Ratio
0.7
1.2
2.3
0.7
1.2
2.3
0.7
1.2
2.3
0.7
1.2
2.3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
W/TS
Ratio
0.40
0.40
0.40
0.40
0.40
0.40
0.40
0.40
0.40
0.40
0.40
0.40
0.40
0.45
0.40
0.40
0.40
0.40
0.40
0.40
0.40
0.42
0.43
0.40
0.45
0.45
0.45
0.45
0.48
0.49
0.48
0.49
0.50
0.45
0.40
0.45
Days After Mixing
7
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
5
5
176
37.5
128
183
32.9
33
45.7
27.9
38.9
35.7
24.1
22.2
19.5
9.9
17.2
19.4
21.9
30.3
34.8
34.9
29.8
36.9
14
977
>1000
>1000
>1000
>1000
>1000
28
99
71
15.8
167
177
72.9
51.8
211
59.7
190
225
36.6
38.4
44.7
28.1
55.7
38.2
27.3
29
30.4
17.1
24.2
31.4
28.7
33
48.8
36
38.7
35.7
21
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
93
54
215
78.3
164
275
37.1
40.8
43.7
26.8
52.4
33
26
33.9
32.9
17.3
26.9
41.2
29.1
36.4
48.2
34.8
36.3
37.9
28
1093
>1000
>1000
>1000
>1000
>1000
>1000
>1000
>1000
16.2
160
300
113
241
81.1
85.1
216
252
38.5
39
79.8
32.2
52.2
40.1
32.3
40.4
46.6
28.8
62.4
73.4
30.7
36.5
50.9
37.9
40.5
42.2
-550-
-------
The organic volatile and seraivolatile emissions from the solidified
samples were qualitatively measured to track the loss of organic components
from the samples into the surrounding air. After curing overnight in the
molds, the SARM samples were placed into polyethylene bags and sealed. At 14
and 28 days, a five-miHiliter sample of the air in the bag of the SARM
sample was withdrawn for analysis before the bag was opened and the cube
measured.
It should be noted that this methodology only gave the concentration of
the compounds in the headspace. Because no gas flux measurements through the
plastic bags were made, these concentrations cannot be used to calculate the
emission rate of the organics. Because of this limitation of the experiment
they should be construed as qualitative in nature.
The TCLP analyses were performed, for both organics and metals; however
because of the significant losses of the organic constituents during mixing
and handling, the results of the TCLP analyses proved inconclusive. As a
result, they are not presented here. This observation is discussed further
below. The results of the TCLP analyses for metals are presented in Table 6.
Table 6 lists the SARM type (I through IV) and the sample number that
was tested in the first column. The second column identifies the type of
binder used. "RAW" is the contaminated SARM without stabilization and PC,
KD, and LF are the three binders. The numbers in parenthesis identify the
day the analyses were performed—14 or 28 days after mixing. The final
columns present the TCLP results for each metal: (a) giving the parts per
million (ppm) of that metal found in the extract, and (b) giving the percent
decrease that this represents over the raw SARM. It should be noted that the
values in the (b) column correct for the decrease in the concentration of
that metal that is due to dilution by the binders. For example, if the
binder-to-soil ratio was 1:1, then normal dilution would result in a halving
o± the TCLP. The percent reduction corrects for this dilution effect.
Table 7 gives the results of the TWA analyses for the metals in both the
raw and solidified SARM's. The data is presented in a similar format as in
Table 6, except that the results are only presented as parts per million of
each metal in the product. Because TWA is only an indication of the amount
of each component present in the sample, the results are not corrected for
dilution in Table 7. The final report does present the results corrected for
dilution.
DISCUSSION OF RESULTS AND CONCLUSIONS
The results indicated that the Portland cement formed a much stronger
?™han the °ther tW° binders- Typically, the Portland cement resulted
in a UCS exceeding 1000 psi (the upper limit of measurement with the
available equipment) for three out of the four SARM's. Further, it achieved
these levels at far lower soil-to-binder ratios than the other two binders,
resulting in a smaller volume of waste requiring ultimate disposal The
strength of the product solidified with Portland cement was significantly
lower for SARM IV than for the other three SARM's. The SARM IV had been
contaminated with very high levels of both organic compounds and metal salts
-551-
-------
II
II
II
II
II
II
II
II
II
II
II
[I
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
ja
u
5
r-i
1
^•^j fg
t ^
CO
CD
f
romium
b
0
Cadmium
a b
rsenic
b
< CO
k
CD f^
Q CO
•H a
pq ^
-s <3" CO CO CO CO
CD 0 0 0 0 0 0
o o o o o o o
CD CD CM CD CD CM
g O O O O O O
d o d d d o
o o o o o o
o o o o o o
r-H i— 1 i— 1 i-H r-H rH
CO
O
1 1 1 1 1 1
^4 t^< ^* CO CO OC
r-H rH r-H CM CM CM
- CD CD CO
CO O O g O O O
o o o o o o
+ + , 1. -t- ..-r. i
co co co m
C3 O O g O O g
o o o o
o o o o o o
o o o o o o
i-H r-l i-H rH rH rH
CO
o
1 1 1 1 1
"^ CM CO t> CM t>
Q O CM O O rH O
o o o o o o
ooooo
o o o o o
rH r-H rH rH rH
CM CM
rH g g O O g O
CO
CM CO rH
I i m i CD m
CD rH rH CD
co g a co g CM c~
CD 0 0 O
«d* *Np ^ 00 CO OC
rH rH rH CM CM CM
(Zi P-, 3 t-3 Q-i 3 i—
t-H t> rH CO C- rH C»~
M CM CO CM C*1
• O C- CD O C- CD
O CD CD O CD CD
rH i— 1
cn c- CM >3< CM =*
co m CM c- c- co
CD O •* CO O CO CO
j:
O 0 O O 0 0
0 0 O O 0 0
r-l ,_( r-| ,— | r— 1 rH
CD
CM
CD CO CD CD
CD «3< + CD +
CD C-
<^}< CO <3< CO CO •*
0 0 CM rH 0 rH in
t- r-l CD CM CD
o o- CD o t- in
O CD CD O CD CD
rH r-H
O C
Ii CO O
II -P
II 0) CO
ii "d co
l! f.
n o
O
a
-H
I I
g
•rl
•s
rH
•H
"C
CO
4->
O
O
s"
•H HJ
W 3
-P T3
rH CO
3 iU
W
CO -P
a g
O CO
EH (O,
-552-
-------
TABLE 7. SUMMARY OF TWA RESULTS, METALS
(SARM)
Sample
No.
I
1
14
27
1
15
27
II
4
16
30
4
16
29
III
7
21
33
7
21
33
IV
10
23
35
10
23
36
Binder
(Day)
RAW
PC(14)
KD(14)
LF(14)
PC.(28)
KD(28)
LF(28)
RAW
PC(14)
KD(14)
LF(14)
PC(28)
KD(28)
LF.(28)
RAW
PC (14)
KD(14)
IF(14)
PC(28)
KD(28)
LF(28)
RAW
PC(14)
KD(14)
LF(14)
PC(28)
KD(28)
LF(28)
As
18
18
15
29
15
12
30
18
15
14
28
23
15
32
904
528
223
196
584
233
180
810
506
290
281
563
271
225
Cd
17
18
12
8
17
12
9
23
18
17
10
24
20
11
1,280
797
315
258
934
326
251
1,430
858
541
448
952
490
306
TWA
Cr
27
49
31
14
56
22
19
37
47
51
15
45
27
19
1,190
1,010
391
299
1,060
432
279
1,650
1,060
550
461
1,020
516
386
Results
Cu
193
195
113
78
164
101
62
260
125
133
85
218
153
106
9,650
6,390
2,420
1,810
7,960
2,660
1,660
13,300
7,040
4,230
4,440
10,100
4,860
3,430
(ppm)
Pb
190
453
183
89
189
119
113
240
149
280
97
294
193
134
15,200
11,600
4,710
3,830
12,100
4,390
2,780
19,900
12,100
6,320
6,590
8,680
5,190
4,590
Ni
27
37
65
19
32
69
16
32
34
50
21
39
53
20
1,140
625
300
216
724
300
169
1,380
616
418
374
753
449
255
Zn
392
393
299
182
320
232
151
544
351
383
161
479
404
276
53,400
14,800
7,600
5,850
22,200
7,690
4,830
28,900
17,500
11,200
9,890
21,000
12,300
7,020
-553-
-------
and it appears that this combination resulted in a large amount of
interference to the solidification process.
The lime kiln dust and the lime/flyash mixtures used for these tests did
not result in values of the UCS as high as those observed with Portland
cement—even for SARM IV. The samples solidified with these binders were
generally very weak. The strength (UCS), however, continued to increase
during the course of these tests. The trend in the data was very clear and
confirmed the general impression that lime-based binders will continue to
harden over time.
The SARM samples stabilized with lime and lime kiln dust/flyash
continued to cure over time. The UCS values for these samples started very
low but as time progressed, they increased. The test suggests that the
curing time for these binders should have been extended to determine their
ultimate strength. The trend in the data suggested that these samples would
continue to show increases in their UCS beyond the 30-day period.
An observation made during the initial screening tests of this program
appears to be useful for further work. These tests showed that a
water/solids ratio of approximately 0.4 would result in a solidified product
regardless of the binder used—within the overall context of the experiment.
This observation, if confirmed with other systems, could result in a
significant reduction in the number of experiments required to test a given
waste/binder ratio.
The results of the TCLP for the metals on the treated SARM's was very
encouraging. The data show that the metals leaching from the SARM's are
reduced significantly by the solidification process. In fact, the reduction
approached 100& for many of the compounds. Many of the metal salts appeared
to be attenuated by the SARM itself.
The TCLP results on the raw (unsolidified) contaminated SARM's were
lower than the expected values for almost all of the metals. This made the
data difficult to interpret as many of the analyses were being made at or
near the detection limits. Nevertheless, the results clearly indicate a
significant reduction in the TCLP of each of the metals in almost all cases.
It should be emphasized that the unsolidified SARM's themselves proved
to be reasonably good stabilizing agents. Examination of the TCLP results of
the raw SARM's showed that the matrix itself prevented a large portion of the
metals from being released to the TCLP. Many analyses of the raw,
contaminated material approached the minimum level of sensitivity of the
analyses.
All of the binders reduced the leachability of the cadmium, copper,
nickel, and zinc. In all of these cases, the solidified SARM's resulted in
only trace amounts or less of these metal salts in the leachate. The TCLP
results for lead was less consistent. All of the samples that were
solidified with Portland cement showed very large reductions in the
leachability of this ion. The kiln dust showed some reduction in the
-554-
-------
alth°Ugh the Deduction was not as great as for the
nt th bi»<*er showed no apparent reduction in the leachability
i ?\ f In faCt> the results actually showed an increase in the
leachability after correcting for dilution. The increase is most likely due
to the error introduced by the large amount of binder used for these samples
and the resultant large dilution factor, rather than being an actual
increase However, the results clearly showed no apparent reduction in the
leachability of lead with this binder.
af?enic ^stabilized well by the Portland cement and the kiln
reduced the arsenic's
The solidification process had virtually no effect on the chromium ion
in this program In fact, as with the lead solidified by the lime/fly^
mix, the concentration of chromium in the leachate appeared to increase after
SSTS1?* the results of the TCLP on the solidified material for dilution.
Once again, this, increase is most likely an artifact of the experiment-
did not reduce the
The^reason for the negative results on the chromium and (to a lesser
extent) the lead with lime/flyash is uncertain. Table 7 lists the" SuS
* K W6re USS? t0 Splke the SARM'S and their solubilities in
As can be seen, the solubility of the salts does not appear to be a
nH fh01"^ °X-de 1S insoluble «* so is zi*c oxide, yet the zinc was
fixed and the chromium was not. Similarly, both the lead and arsenic salts
used are slightly soluble, yet they also behaved differently? arSeni° SaltS
raetal°Xii5S
-------
again, the problem is most likely in the "leveraging" of any error by the
dilution correction.
The variability due to the analytical method can be estimated by
examining the difference between the TWA for the metals for each mix after 14
and 28 days. For example, SARM III solidified with portland cement (PC)
after 14 days (Sample 7) showed 528-ppm arsenic. The same sample at 28 days
showed 584 ppm. Generally, comparison of the metals analyses for each sample
at 14 and 28 days showed a similar consistency. This type of variability is
quite small, indicating that the mixing procedures used in this program
resulted in a homogeneous product and that the analytical protocol appeared
to give reasonably consistent results.
The analysis of the volatile and semivolatile organic compounds in the
headspace by gas chromatography/flarae ion detector (GC/FID) seemed to
indicate that the emissions dropped only slightly from 14 days to 28 days.
This is consistent with other research done by Acurex (EPA Contract No.
68-02-3993, W.A. 32 and 37) which showed that volatile organic emissions
occur mostly during mixing, and then continue at a steady rate after curing
in a stabilized sample, dropping as the organic content of the solidified
material is reduced.
The solidified SARM's generally showed a lower TCLP value for the
volatile organic contaminants than the original SARM's. This cannot,
however, be attributed to the solidification process binding the volatile
compounds so that they are not accessible. Rather, this is most likely due
to a simple release of the volatile compounds during the mixing process and
during the sample preparation prior to extraction. This was corroborated by
other research (2). Because of this fact, the TCLP results for the volatile
organics are not considered to be significant and are not presented in this
paper. They are presented in the final report for this program.
The TCLP for the semivolatile organics, in general, showed a significant
decrease because of solidification. The results show that the percent
decrease in the TCLP analyses for the semivolatile organics is greatest for
SARM's I and IV (those contaminated with high levels or organics) and least
for II and III. SARM's II and III also show a greater variability for the
semivolatile reduction, but this is most likely due to analytical errors
caused by the low concentration of the semivolatile compounds.
The TWA analyses for the volatile organics showed the same pattern as
the TCLP. The TWA analyses, however, showed the results magnified. That is,
the solidified SARM's contained on the order of 80 to Q0% less volatile
organics than the original material. This is consistent with the hypothesis
that the volatile organics were released to the air rather than trapped in
the solid. Had the volatile organics truly been solidified, then the TWA
would have shown a constant value for these materials while the TCLP would
have shown a decrease.
The TWA results for the semivolatile organics was unexpected.
Stabilization appeared to result in an apparent increase in almost all of the
values. This is most likely an artifact of the analytical method. The TWA
-556-
-------
results appear to have a very wide variation in them. The reason for this is
unclear, but it may be due to the physical nature of the sendvolatile
compounds They are heavy solids that go into solution slowly. As a result,
the amount of each constituent in the liquid after the extraction for
analysis may be more of a function of how much of the material actually
dissolves than of the total amount of that compound in the waste. Under
normal conditions, this error is not significant; however, in this case, the
TWA values are corrected for dilution. This results in a "leveraging" of any
error and a much higher degree of uncertainty for the TWA results
In conclusion, it appears that solidification can reduce the
leachability of many metals to near zero. Even in this case, when no effort
was made to match the solidification process to the contaminant, most of the
contaminants were effectively immobilized as determined by the TCLP. It is
likely that with a proper choice of binder, it should be possible to
immobilize other inorganic contaminants and, possibly, even some of the
i.
2.
REFERENCES
Locke, B.B., Esposito, P.M., Furman, C., and Traver, R. P. CERCLA BDAT
standard analytical reference matrix (SARM) preparation and results of
physical soils washing experiment. Draft Paper.
Weitzman L Hamel, L. and Cadmus, S. Volatile organic emissions from
stabilized hazardous wastes. Final Report, EPA Contract #68-02-3994, WA
6£ and 37, August, 1987.
-557-
-------
SITE DEMONSTRATION OF SHIRCO INFRARED INCINERATOR
by
Howard 0. Wall
Thermal Destruction Branch
Hazardous Waste Engineering'Research Laboratory, EPA
and
Seymour Rosenthal
Technology Manager
Enviresponse, Incorporated
ABSTRACT
A Shirco Infrared System used for a removal action at a PCB-contain-
ing oil refining waste site in Brandon, Florida (a suburb of Tampa) was
evaluated. The evaluation included a determination of toxics in the
material being decontaminated as well as all the effluent streams such as
ash, air emissions and wastewater. These streams were analyzed for heavy
metals, organics, dioxins, furans as well as NOX, and inorganic acids.
Leaching tests were performed on the ash. The results indicated that the
PCB was reduced from 5 to 100 ppm to less than 1 ppm in the ash, which
was the purpose of.the removal action. Although research had indicated
that the lead compounds in the ash would become insoluble because they
would be complexed with carbon, the ash could not be considered non-leach-
able based on the EP toxicity tests.
-558-
-------
SUMMARY
In 1987, the USEPA started the Superfund Innovative Technology Evalua-
tion (SITE) program to demonstrate new or emerging technology for cleanup
of Superfund locations. One of the first technologies planned for testing
was the Shirco Infrared System, a furnace which heated a waste on a travel-
ing belt using infrared heaters. The temperature and combustion air levels
were sufficient to pyrolyze, crack and then combust the organics present in
the waste and convert them to carbon dioxide and water. If products of
incomplete combustion (PICs) were formed, they were destroyed in the unit's
secondary section (which is a fossil-fuel-fired afterburner)*
As the program started, plans were made to characterize a pilot-scale
unit. However, the investigation of which locations were available for
testing indicated that a full-scale unit having a 100 ton/day capacity was
already being used for a cleanup action by EPA's Region IV who asked the
SITE program to test it. This full-scale unit was therefore characterized
as it was being used and the pilot-scale work was shifted to another loca-
tion. ,
The full-scale test described in this paper was at Brandon, Florida
(near Tampa) at a former recyle-oil refinery. The material being treated
had been placed in a lagoon; it was a thick, black, liquid-appearing waste
having 5 to 100 ppm of PCB in it. Although the material appeared to be
liquid, it could not be pumped and had to be mixed with the sand or soil
surrounding the lagoon before it could be handled by a conveying system.
This mixing included neutralization with lime to reduce corrosion of the
equipment and the Shirco furnace; the material was sized to less than
one-inch diameter before it was introduced to the conveying system and the
Shirco unit.
The results and data in this paper are preliminary; however, a full
report is being processed and will be issued in about mid-1988. Based on
four sets of data, the following results were obtained:
1. The Shirco Infrared System sucessfully achieved the reduction of
the PCB content of the feed (3-5 ppm) to less than I ppm which was
the objective of the removal action.
2. A destruction and removal efficiency for PCB in the gaseous exhaust
stream was greater than 99.999 percent (The original PCB content
of the lagoon was 5 to 100 ppm).
3. Essentially all of the PCB was destroyed by the burning process,
rather than by removal processes, because PCB was not present in
the scrubber water.
-559-
-------
9.
Based on past determinations that any emissions from the stack will
be diluted by a factor of 10+8, there were no emissions-of any
metals or organics which would exceed any existing regulations at
ground level. (The air pollution abatement system used for this
location, however, was not suitable for the operation. Emissions
exceeded the converted grain loading of 0.08 grains/DSCF for two
of the four samples taken. The materials of construction for the
air pollution abatement system as well as the design of the system
should be changed before this particular Shirco unit is used for
another study.)
The determination of the soluble chromium in the air emissions was
not successful. The total amount of chromium, if considered
soluble, did not exceed any established or proposed regulations.
Lead emissions at ground level were not sufficient to be a signifi-
cant environmental contaminant.
The toxicity characteristic leaching procedure (EP Tox) indicated the
material was leachable. These tests will be repeated to assure the
results are correct.
Metals and organics in the feed, the ash, and the emissions to the
atmosphere were determined and will be added to the collective
data base. This data base is being used to record information on
all hazardous waste systems.
More research is necessary to determine how the Shirco unit per-
forms on other hazardous wastes and to get a variation of operat-
ing points. A realistic operating cost for a Shirco unit is about
$425 per ton of material processed.
BACKGROUND
During the 1950's, an oil refinery, Peak Oil, reclaimed used oil by
distilling it at Brandon, Florida (a community near the southeastern edge
of Tampa, Florida). The wastes from this processing operation were placed
in natural lagoons. These wastes appeared to be a dark thick oil and were
found to contain PCBs and lead which were seeping into the ground water.
Although the PCBs and lead were known to be contaminants, it was suspected
that many other organic materials we now know can be poisonous, toxic or
hazardous were also present in the lagoon. The lagoon was located in a
sandy soil in a shallow aquifer typical of Florida where the lead and PCBs
had started contaminating local drinking water supplies. The Peak Oil site
was ranked on the National Priorities List primarily due to the contamina-
tion of the local water by PCBs.
-560-
-------
EPA Region IV initiated and supervised a removal action at this site
A contract for the removal action work was given to HazTech Inc an emer-
gency cleanup contractor. HazTech purchased and operated a Shirco Infrared
.System capable of 100 tons per day operation to perform the cleanup.
PROCESS DESCRIPTION
This Shirco Infrared System was the first full-scale, truck-transport-
able unit to be used for thermal decontamination of hazardous wastes. The
components consisted of a primary unit, secondary unit, an emergency stack
an air pollution abatement system, an exhaust stack and a control van.
The primary unit had a feed distribution system for putting waste onto
the woven metal belt which carried the material being treated under the
infrared heating rods. The belt travel rate was adjustable. At Peak Oil
the rate of travel was sufficient to keep the feed in the primary section
18 to 19 minutes.
Rotary finger rakes at several locations stirred the material on the
belt as it passed under them to assure all the material was exposed to the
infrared heat. A blower provided air at selected locations along the belt
and could be used to control burning rate of the feed and the location of
the feed burning on the belt. The air flow was countercurrent to the feed
as it progressed down and off the belt.
As the material (ash) dropped off the belt, it was quenched with water
sprays. This quench was scrubber water effluent and was used to reduce the
amount of scrubber water effluent for disposal or reprocessing. The ash
was screw-conveyed out of the unit into the ash hopper. Then it was removed
to a holding area until it was analyzed for PCB content. After it was
determined that the PCB content was less than 1 ppm, it was piled in a new
location away from the processing area.
The exhaust gases from the primary combustion unit were ducted to the
secondary combustion unit. The secondary unit provided about three seconds
of retention time at up to 2300°F. It contained bars to increase the tur-
bulence so that any remaining gaseous organics were mixed with air and
destroyed. Propane was used as the energy source to supply supplemental
heat. Gases from the secondary unit were ducted to the pollution abatement
equipment via an induced draft blower located after the scrubber system
prior to the stack. '
The air pollution abatement system consisted of a spray quench for the
gases followed by a venturi system which had 15-inch minimum pressure drop.
During these tests, only one venturi was used. A horizontal packed scrubber
system followed the venturi and the water used contained sodium carbonate
and sodium hydroxide to neutralize any acid vapors. An emergency stack prior
to the pollution abatement system was installed so that if the temperature
control system and its interlocks were to fail the air pollution abatement
system would not be melted by the hot .gases.
-561-
-------
The exhaust stack had sample ports and a deck for sampling. A trailer
van with thp control unit was also part of the installation.
When the unit was fully operational, the feed stream and all effluent
streams were evaluated for a 100 tori per day operational point. Over a
four-day period, three replicate samples and a duplicate of one of the
replicates were taken from the stack gas to be analyzed for PCB, dioxins,
furans, semivolatile priority pollutants, volatile orgamcs and heavy
metals. Continuous emission monitors were used for 02, COg, CO, THC and NUX
and downwind of the Shirco unit to determine if any air contamination was
being created by the operation. Corresponding operating data were collected
from the control van while the system was being tested.
Figure 1 shows the sampling locations and the schematic layout of the
Shirco Infrared System at Peak Oil.
All samples and analyses were performed at EPA QA level II. This level
required an approved sampling and analysis procedure and audits of the pro-
cedures. It was a high level of control, but not sufficient for use in a
court case.
These tests were conducted for the primary purpose of determining the
performance of the Shirco Infrared System. Table 1, shows the properties
of the feed material. Although the feed material used during the four days
of testing was from the same general pile, it changed the third and fourth
day. An estimated 300 tons of feed had been sized and stockpiled for the
test, however the estimated 300 tons turned out to be less than 200 tons
and a second batch had to be stockpiled. Since the high heating value was
about 1640 RTU/lb for the second batch and 2065 BTU/lb for the first batch,
the primary unit operated with less air and more infrared heat during the
last two days. This also accounts for the wide range of orgamcs and metals
shown in Table 1 for the feed material. It appeared that the second batch
contained far more inerts (soil and sand) and less of the lagoon contami-
nants.
RESULTS
PCB
The results of the work indicate the PCB content of the feed was
reduced to less than 1 ppm, which was the purpose of the removal action.
Destruction and Removal Efficiency (ORE) for PCB was greater than
99 99 for all tests. Table 2 shows the average of the four samples was
99.99931. Because of the test methods for PCB, it would .have taken about
three days of sampling for each test to determine if the unit was doing
99 9999 or better. Essentially all of the PCB was destroyed by the burning
processes, rather than by removal processes, because PCB was not present in
-562-
-------
FIGURE 1
Sampling Locations
Schematic/Layout Diagram
Peak Oil Incineration Unit
© STACK OASES
*""\
2) SOUO WASTE FEED
FURNACE ASH
SCRUBBER UQUIO
1PFLUENT
SCRUBBER SOUOS
-563-
WATER INLET
©AMBIENT AIR
(DOWNWIND)
©AMBIENT AIR
(UPWIND)
-------
Table 1
Waste Feed Solid Analysis
nanograms per gram
PCB (Total)
heptachlorobiphenyl
hexachlorobiphenyl
pentachlorobiphenyl
tetrachlorobiphenyl
trichlorobiphenyl
dichlorobiphenyl
ethyl benzene
methylene chloride
toluene
xylene
3480 to
940 to
1100 to
200 to
400 to
570 to
120 to
40 to
80 to
130 to
260 to
5850
2200
1700
490
830
820
190
140
120
300
770
antimony
arsenic
cadmi urn
chromium
copper
strontium
lead
vanadium
zinc
moisture
carbon
sulfur
chlorine
ash
micrograms per gram
2.1 to 3.6
2.0 to 2.9
3.9 to 4.6
20 to 24
44 to 55
50 to 62
4400 to 5000
7 to 11
950 to 1100
Percent
14.2 to 16.6
7.0 to 7.8
1.8 to 2.5
less than 0.1
70 to 75
BTU value (HHV)
1640 to 2065 BTU/lb
-564-
-------
ash' A reduct1°" '" sample size resulting from an
Particulates
J*?^1cul;*e emissions averaged 0.1015 grains per DSCF, corrected to 7
Tahi o .2' wh^cn was somewhat above the 0.08 regulation set by RCRA.
Table 2 indicates the first two samples were above the 0.08 value and the
last two samples were less than 0.08 showing the scrubber may not have been
suited for the waste being decontaminated at this location. Since the
operation was a one-time operation, this does not necessarily indicate that
the scrubber system is not suitable for other operations on different
materials. The scrubber ducting and stack were of fiberglass construction
which was not suitable for this type of operation. eryidss construction
Chlorine
f A Th! Concentration of chlorine was below the detection limit in the
feed and therefore an efficiency of removal (of HC1) could not be determined
The gases were caustic-scrubbed to remove hydrochloric acid and sulffrous
acids. Since S02 is much harder to remove than HC1, the S02 levels in Table 2
2 indicates the HC1 and SO? emission
Sulfur
The sulfur in the feed was
emissions was less than 2.5 Ibs
of in excess of 99 percent.
Chromium
300 Ibs per hour and the sulfur in the
per hour. This indicates a removal rate
i M A ^ th? ar^nt °f Chrom1um emitted in the various compounds
luble and insoluble) were not successful. However, if all the chromium
in the feed were emitted through the stack, it would be of such a low concen-
tration that it Would not be significant. concen
Lead
* A ur t0 the test> tnere was an unconfirmed report that samples of the
feed had been roasted and that solubility tests had been run for lead
Results of this bench-scale work indicated that the lead formed an insoluble
comply rendering the lead inert. Other bench-scale tests made for lead-
containing wastes have indicated that lead remains in the ash and is not
-565-
-------
Table 2
Emission Data
Date of
Run
l\ U 1 1
8/1/87
8/2/87
8/3/87
8/4/87
8/4/87
ORE for PCB
%
99.99967
99.9988
99.99972
99.99905
Parti cul ate
corrected to
7%02, grains/DSCF
0.1590
0.0939
0.0768
0.0761
HC1
g/hr
<0.051
0.600
0.220
0.200
S02
g/hr
27.6
1070.0
22,0
20.6
Ave rage
99.99931
0.1015
0.2678 285.05
-566-
-------
the
Cadmium
Tahla o?adT!; (the el"sive metal) was very low in the leach tests (see
Other Metals
Semi -volatile Organics
small
f°P
^mnnhn W!re detected' b«t these were probably from the plastic
sample bottles and not the process since they were also detected n
IS" '»' -"s '
Dioxins and Furans
ana1^ses are required whenever PCB is in a waste
recovery
-567-
-------
Table 3
Leaching Test Results
TCLP Analysis
tr lUAH-iky — — : — •
Parameter Average Regulatory Average
parameter ^y Level , mg/L Level, mg/L
Arsenic
Barium
Cadmi un
Chromium
Lead
Mercury
Selenium
Silver
Only
Acrylonitrile
Methyl ene chloride
0.020
1.35
.099
.037
31
.0015
ND*
.031
TCLP compounds
—
5.0
100
1.0
5.0
5.0
0.2
1.0
5.0
detected
Toluene
1,1,1, Trichloroethane
Trichloroethane
.007
.25
.008
.037
.011
NO*
.031
.059
are listed below
013
020
.0020
.0006
0006
Regulatory
Level
5.0
100
1.0
5.0
5.0
0.2
1.0
5.0
5.0
8.8
14.4
30
.07
*Not detected
-568-
-------
Table 4
Metals Analysis
Parameter
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Lithium
Magnesium
Manganese
Mercury
Molybdenum
Nickel
Phosphorus
Selenium
Sil icon
Sil ver
Sodium
Strontium
Sul fur
Thai luim
Titanium
Vanadium
Zinc
Solid Waste Feed
1625
2.15
2.55
505
.168
NA
4.15
37500
22
.75
49
2050
4800
ND
850
47
• ND
ND
8
790
ND
NA
2.0
5550
57
20500
ND
41
9
1030
mi crog rams /gram,
ug/q
Ash
2500
3.3
2.6
757
3
• *-•
NA
4.1
50000
27
2
64
2600
6400
ND
1050
55
ND
ND
10
770
ND
NA
4.0
5600
76
24000
ND
115
13
1060
Stack gas
<210
38
675
625
1920
1680
270
420
~ f— \j
440
58000
01
C.L
180
in
CO
DU
42
o
3.2
780
10
18600
1 n
1U
160000
630
<50
<25
^t—\j
9400
Not determined
NA - Not analyzed
The stack gas contained 0.1015 grains/DSCF of particulate (One grain = 64.8 mg)
-569-
-------
Since this test collected data for one operating point for one material,
more research should be done on the Shirco system to collect data for more
operating conditions on different feed materials. To do this, we have per-
formed a SITE test using the Shirco Infrared pilot plant system at the Rose
Township, Michigan waste location.
Cost and Economics
An economic analysis concluded that the cost of using a Shirco unit
could be as low as $200 per ton of contaminated material or as high as $800
per ton depending on the equipment usage factor considered realistic for
the project to be evaluated. A projected realistic average cost per ton of
contaminated material is $425 per ton.
-570-
-------
DEMONSTRATION OF THE PYRETRON" ENHANCED OXYGEN BURNER
AT THE U.S. EPA COMBUSTION RESEARCH FACILITY
by
Grigory Gitman, Mark Zwecker and Tom Wechsler
American Combustion, Inc.
Norcross, Georgia,
Larry Waterland and Johannes Lee
Acurex Corporation
Mountain View, California
and
Robert E. Mournighan and Laurel J. Staley
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The Pyretron* developed by American Combustion Inc. is an enhanced
oxygen burner designed to be retrofitted to an existing incinerator. As
part of the Superfund Innovative Technology Evaluation (SITE) program, a
Pyretron was installed on the rotary kiln system at the U.S. EPA's Combus-
tion Research Facility (CRF) where it was used to treat a mixture of soil
and decanter tank tar sludge from coking operations, listed waste K087.
This paper presents a preliminary evaluation of the Pyretron's performance
during this test.
Introduction
The Superfund Amendments and Reauthorization Act (SARA) of 1986
established the Superfund Innovative Technology Evaluation (SITE) Program
to promote the development and use of innovative technologies to treat
Superfund and hazardous wastes. As part of the SITE program, the Pyretron™
Thermal Destruction System, developed by American Combustion, Inc., was
demonstrated at the U.S. EPA Combustion Research Facility (CRF) in Jefferson
Arkansas. '
The purpose of the demonstration was to evaluate three claims made by
American Combustion which represent improvements in hazardous waste
incinerator burner design. These claims are as follows:
1. That the Pyretron, with its oxygen enhancement and proprietary
process control system, reduces transient particulate emissions that occur
when solid material is intermittently fed to a rotary kiln.
-571-
-------
2. That the improved heat transfer and process control provided by
the Pyretron make it possible to achieve significantly higher waste throughput
rates.
3 That the cost savings resulting from the higher throughput rates
offset'the cost of the required oxygen resulting in an economical process.
Although the demonstration was completed early this year, all of the
data have not been thoroughly evaluated yet and as ^result ^he final
report on this project will not be available until later this year. This
paper will report on the data available at this time. After a more detailed
description oTthe Pyretron and how it was installed at the CRF,a more
detailed description will be given of the results obtained thus far.
Description of the Pyretron and of the CRF's Rotary Kiln
Figure 1 is a schematic diagram of the Pyretron enhanced oxygen
burner. The Pyretron combustion system utilized for^this dej°?8!"t
S exieed'el iK^
chamber. Oxygen flowing through the center of the burner partially oxidizes
and thermally pyrolyzes the propane mixing concentrically around it result-
in! in!™mong other things, the formation of radiant soot microparticles.
Radiant heat transfer from these microparticles is the main mechanism for
energy transfer for the combustion of the solid material being treated in
the rotary kiln. Although this also happens with conventional burners, the
higher Spetature of soot microparticles inside the flame envelope resulting
from oxygen enhancement intensifies this effect. These soot microparticles
are further burned out by contact with the oxygen and air streams flowing
concentrically around the pyrolytic flame zone.
The Pyretron Thermal Destruction System installed at the CRF included
two enhancS oxygen burners like the one described above accompanying gas
trains and a process control system which dynamically controls the air/
oxygen/fuel ratios based on readings of CO, 02 and pressure in the rotary
kiln and afterburner.
This computerized process controller for dynamic incineration process
control is able to maintain process temperature and excess oxygen levels
and to respond to rapid process deviations by Dynamically changing the
relative amounts of combustion air and oxygen fed to the incinerator. These
changes are based on measurements of CO, 02, and pressure between the
kUn^and afterburners. By varying the air-to-oxygen ratio, the amount of
nitrogen fed to the incinerator is closely controlled. Nitrogen occupies
a 5j£ fraction of a conventional incineration combustion chamber volume
-572--
-------
g
01
U
9",
o,
0)
u-i
O
O
CO
60
-573-
-------
and is a major heat sink. Figure 2 shows how the Pyretron was incorporated
into the CRF's rotary kiln system.
The design characteristics of the CRF rotary kiln system are provided
in Table 1. For testing, the kiln was operated at approximately 1800 F and
the afterburner was operated at 2050°F. The kiln rotation was such as to
provide a solids residence time of approximately one hour (.2 rpra}.
Description of Demonstration Tests
As mentioned previously, the objective of the demonstration tests was
to evaluate the performance of the Pyretron Thermal Destruction System in
comparison to a conventional air based incineration system utilizing the
CRF rotary kiln system (RKS). For this comparison, the only modification
to the CRF RKS was the substitution of the conventional air burner system
by the Pyretron Thermal Destruction System.
A waste stream was selected for this test which was a mixture of
waste material from the Stringfellow Superfund site and decanter tank tar
sludge, listed waste K087. The resulting waste stream which is character-
ized in Table 2 has a heat content of approximately 8600 BTUs/lb. The
POHCs designated for this waste are several hazardous polynuclear aromatic
hydrocarbons (i.e., the principal organic hazardous constituents).
Eight tests were conducted as part of this demonstration. There are
two reasons why so many tests were required. First, the claims made about
the Pyretron were comparative in nature and so required that tests be
conducted without oxygen enhancement to establish a basis for later
comparisons. Second, it was necessary to determine the maximum feed rate
for both air and oxygen enhanced operation in order to evaluate one of
the claims. This required two separate tests. First, the maximum amount
fed per charge must be determined. Second the maximum frequency at which
those charges can be fed must be determined. Table 3 lists the tests
conducted as part of this SITE demonstration.
For each test, all feed and effluent streams were sampled using standard
methods.
Test Results/Conventional System
The optimum feed charge mass and the optimum mass feedrate attainable
utilizing a conventional air-based incineration system on the CRF rotary
kiln consisted of a mass charge of 21 pounds at a charging interval of 12
minutes. This level was determined after a series of optimization trials
and was based on emission levels as measured by continuous emission monitors,
combustion chamber temperatures and the ability to control process upsets.
Higher feed rates destabilized the process too much. This is shown by test
data from a trial run using a mass charge size and charging interval that
were an increment above the levels determined to be optimum (i.e., at
-574-
-------
60
w
to
a
u
o
(U
4-1
05
a
o
CO
O
r*
w
s
4)
£
e
o
4J
V
£
Oi
Cv)
Ol
bO
-575-
-------
I
TABLE 1. Design Characteristics of the CRF Rotary Kiln System
Characteristics of the Kiln Main Chamber
Length, outside
Diameter, outside
Length, inside
Diameter, inside
Chamber volume
Construction
Refractory
Rotation
Solids retention time
Feed system
Liquids
Semiliquids
Solids
Containerized
Temperature (max)
2.61 m (8 ft, 7 in)
1.22 m (4 ft, 0 in)
2.13 m (7 ft, 0 in)
0.95 m (3 ft, 1-1/2 in)
1.74 m3 (61.36 ft3)
0.63 cm (0.25 in) cold rolled steel
12.7 cm (5 in) thick high alumina
castable refractory, variable
depth to produce a frust-
roconical effect for moving solids
Clockwise or counterclockwise 0.2 to 1.5 rpm
1 hr (at 0.2 rpm)
Pos. displacement pump via water-cooled lance
Moyno pump via front face, water-cooled lance
Metered twin-auger screw feeder
5.7 L (1.5 gal) fiberpack ram feeder
1000°C (1850°F)
Characteristics of the Afterburner Chamber
Length, outside
Diameter, outside
Length, inside
Diameter, inside
Chamber volume
Construction
Refractory
Gas residence time
Temperature (max)
3.05 m (10 ft, 0 in)
1.22 m (4 ft, 0 in)
2.74 m (9 ft, 0 in)
0.91 m (3 ft, 0 in)
1.80 m3 (63.6 ft3)
0.63cm (0.25 in) thick cold rolled steel
15.24 cm (6 in) thick high alumina castable
refractory
Depends on temperature and excess air,
1.2 to 2.5 sec
1,200°C (2,200°F)
Characteristics of the Air Pollution Control System
System capacity
Inlet gas flow
Pressure drop
Venturi scrubber
Packed column
Liquid Flow
Venturi scrubber
Packed column
pH control
1.07 m3/min
1,200°C
101 kPa
7.5 kPa
1.0 kPa
(3,773 acfm) at
(2,200°F)
(14.7 psia)
(30 in. we)
(4 in. we)
77.2 L/min (20.4 gpm) at 69 kPa (10 psig)
116 L/min. (30.6 gpm) at 69 kPa (10 psig)
Feedback control by NaOH solution addition
-576-
-------
TABLE 2. Waste Characterization a»b
Parts Per Million
Naphthalene 49772
Acenaphthalene 11722
Fluorene 6397
Phenanthrene 22569
Anthracene 6764
Fluoranthene 11529
Pyrene 11036
2-Methylnaphthalene 3457
Dibenzofuran 3926
Miscellaneous PNA 5017
Miscellaneous Unknowns 1031
a60% by weight K087
40% Stringfellow Soil
Average over 9 composite samples
-577-
-------
TABLE 3. Summary of 8-Week Pyretron Demonstration Program Test Conditions
Test/Purpose
Charge Feed -Afterburner-
Charge entered, Rate, Kiln T, Kiln Temp., 02,
size, Ibs min. Ib/hr °F 02, % °F
1 Same as test 2 except 8
using air-only burner
2 Test burn Stringfellow 8
waste with A500 ppm
C2Clg and 4500 ppm
C6H3C13 with °xysen
enhancement (per
Region IX request)
3 Determine max feed rate 21
with air-only burner
3A Duplicate of test 3 21
A Assess oxygen-enhanced 21
operation at feed rate
determined in tests 3
and 3A
5 Assess oxygen-enhanced 34
performance at max feed
mass/charge at feed rate
determined in test 3
6 Determine maximum mass/ 21
charge with oxygen en-
hancement
6A Duplicate test 6 21
12
120 1870 8.9 2017 11.2
120 1866 13.7 2017 13.0
105 1906 11.3 2036 7.7
12 105 1781 11.8 2030 7.4
12 105 1772 17.1 2022 15.2
19.5 105 1778 14.5 2042 14.6
6 210 1794 14.8 2010 13.5
6 210 1814 15.4 2017 15.5
-578-
-------
24 lbs/10 minutes). These data are provided in Figure 3. In this case,
emissions or "puffs" were high enough to impede operation of the kiln.
Depletion of oxygen in the kiln resulted in flame outs and in excessively
high.CO levels exiting the kiln and CO breakthroughs from the afterburner.
This depletion of oxygen typically occurred immediately following batch
charges as waste combustion consumed all the available oxygen.
When attempts were made to increase air flows to provide the additional
oxygen required, residence times were reduced below levels necessary for com-
plete combustion and CO levels were therefore increased. Several flameouts
also occurred at that time. In addition, gaseous emissions were observed
on several occasions due to the loss of negative pressures in the kiln.
With the high BTU waste stream selected for this demonstration, the
kiln was unable to dissipate the heat build-up at the maximum mass charge
size and feed interval selected for this test. This build-up is partly a
function of the limited capability of the conventional air-based flame to
be turned down to reduce auxiliary fuel input while maintaining sufficient
flows of oxidizer into the kiln as well as auxiliary flame stability to
conduct proper incineration.
Data from stripchart recordings of the series of tests at the optimum
mass charge and optimum feedrate for air-only operation are provided in
Figure 4. As can be seen from these graphs, oxygen content and temperature
levels in the kiln were maintained within the operating limitations of the
kiln. Carbon monoxide levels at the kiln exit reached moderate levels,
however, these levels were within the capacity of the afterburner system so
that stack emissions stayed within regulatory requirements.
Test Results/Pvretron Thermal Destruction System
Figure 5 presents the data from testing the Pyretron Thermal
Destruction System. The mass charge size was maintained at the 21 pounds
utilized with the conventional air-based incineration system, however, the
charge interval was reduced from 12 minutes to 6 minutes. This reduction
in the charge interval represents a doubling in the throughput rate with
the Pyretron system as compared to the conventional incineration system.
As can be seen from the graphs, the Pyretron system provided temperature
control in both the kiln and afterburner. Oxygen levels at the kiln exit
were maintained at sufficiently high levels and CO at the kiln exit was at
a minimum with no indication of CO at the stack.
This data depicts the ability of the the oxygen enhanced Pyretron to
process waste at a faster rate than air-only combustion. Dynamic water
injection was used to control temperature in the kiln in the oxygen-enhanced
mode.
-579-
-------
32OO
g. 2000
1
1800
Storting Time : 10:27
BOOO
JT
% WOO
w
J 4000
LL.
2000
0
too
Air
Propone
200
Minutes
300
400
t5
o
o
CM
O 5
I
C>
O
1500
1000
500
100
Storting Time : 10:27
200
Minutes
300
400
Figure 3. c.R.F. Kiln Data for 12-08-87
Conventional System - 24 lbs/10 oin.
-580-
-------
Figure 3. (Cont.)
C.R.F. Afterburner Data for
Convention.! System . „*
2200
f^'^'vwwvwwi^^
Storting Time : 10:27
400
C.R.F. Stack Data for 12-08-87
Conventional System - 24 lbs/10 min,
E
CL
Q.
O
O
«00
-581-
-------
Fiaure 4.
C.R.F. Kiln Data for 12-09-87
Conventional System - 21 lbs/12 Bin.
2200
.r 2000
E
V
g- icoo
18BI
BOOO
in
o
6000
4000
2000
Slortinj Time : 12:40
'
W*frtJ\rtufy+lS\/\/^^
Propone
' ' ' —•-
100
150
Minutes
200
250
300
20
-^ 15
C
V
I 10
CM
1500
1250
1000
.
- 750
500
250
0
C.R.F. Kiln Data for 12-09-87
Conventional System - 21 lbs/12 min.
Storting Time : 12.40
I L
50
100
ISO
Minutes
200
250
300
-582-
-------
Figure 4. (Cent.)
C.R.F. Afterburner Data for 12-09-87
Conventional System.- 21 lbs/12 «in.
2200
„- 2000
*••
o
g- 1800
!
aeee
1.6E4
.8 '-2"
w
IBOOO.O
4000.0
o.o
Storting Time : 12:40
100
150
Minutes
200
Propone x 10
250
300
C.R.Fo Stack Data for 12-09-87
Conventional System - 21 lbs/12 min.
E
0.
0.
O
O
95
75
55
35
15
-5
C
Storting Time : 12:40 Meosured ot Stock Exit
•
•
> •— ' • — «-• -*<«. -~ , L^ - - - 1 . J 1 .
50 '00 150 200 250 300
Minutes
-583-
-------
Figure 4. (Cont.)
GO
I*-* UJ
oo «—
I =3
en z:
CD — '
i E:
CVJ
o oo
u_ oa
CN4
ce:
LU s:
z: UJ
ce: f—
=> co
ca >-
C£ CO
UJ
C£
o
-584-
-------
Figure 5.
C.R.F. Kiln Data for 1-21-88
Pyretron System - 21 lbs/6 ain.
2200
100
150
Minutes
200
20
c
0)
u
'0
800
a
8 *°°
Starting Time : 12:00
100
150
Minutes
200
250
300
250
300
-585-
-------
Figure 5. (Cont.)
C.R.F. Afterburner Data for 1-21-88
Pyretron System - 21.1bs/6 min.
2200
u.
o
tf 2000
t_
=>
"o
ex
E
1800
U
w
to
o
1.8E4
1.6E4
1.4E4
1.2E4
1.0E4
80000
6000.0
4000.0
2000.0
0.0
Storting Time : 12:00
Air
Propon* x 10
Oxygen
=jtyiirrmfmf|rt^^
100
150
Minutes
200
250
300
C.R.F. Afterburner Data for 1-21-88
Pyretron System - 21 lbs/6 mln.
95
75
I »
a.
8 35
15
0
Storting Time : 12:QO
50
100
150
Minutes
200
250
300
-586-.
-------
A f inal set °f data is provided in Figure 6 representing an attempt to
challenge the Pyretron Thermal Destruction System by increasing the size of
the batch charge by 60%. For this test the mass of the batch charge was
increased to 34 pounds in an attempt to determine the capability of the
system to handle the so-called "puffs" which are experienced with many
incinerators immediately following a batch charge. As can be seen from the
data, even with a 60% increased batch charge size, the system was able
to maintain sufficient oxygen concentration in the kiln and CO levels at
the kiln exit well within the capacities of the afterburner system. Again
during this test using the Pyretron Thermal Destruction System, there was no
CO emitted from the afterburner.
Destruction and Removal Efficiencies and Particulate Emission
Destruction and Removal Efficiencies (ORE) exceeded 99.9999% for all
tests. No POHCs were detected in the stack gas.
™» These results show that, even with increased throughput rates, 99.99%
DRE is still achievable.
Particulate emissions for each of the 8 tests are summarized in Table
4. Data are corrected to 7% 02 only for the air-only tests. At this point
the best way to make this correction in cases where oxygen enhancement is
used has not been determined.
Conclusions
Although all of the data is not in yet, the data available so far
supports two of the claims made about the Pyretron. That is, for the waste
testedBand the conditions under which it was tested, transient emissions
i.e. puffs , were reduced, and waste feed rates could be increased while
still meeting 99.99% DRE.
Final evaluation of the claims made about the Pyretron1s perform-
ance will have to await final analysis of all the data. This data evalua-
tion should be completed by October, 1988.
-587-
-------
Figure 6.
C.R.F. Kiln Data for 1-14-88
Pyretron System - 34 lbs/19.5 min.
2200
20
— 15
O.
CM
O
1000
800
E 600
Q.
CL
R* 400
200
0
Storting Time : 13:1 A
SO
IOC
150
Minutes
200
300
250
300
-588-
-------
Figure 6. (Cont.)
C.R.F. Afterburner Data for 1-14-88
Pyretron System - 34 lbs/19.5 ain.
2200
250
300
-589-
-------
Test
TABLE 4. Particulate Emissions Data
Particulate3
Concentration
mg/dscm
Particulate
Concentration'1
mg/dscm
Particulate
Concentration
(uncorrected)
gr/dscf
1
J.
p
£»
3
.J
4
5
f.
\j
7
8
8.0
9.0
99.0
59.0
63.0
21.0
37.0
38.0
8
9
47
29
28
10
28
39
.0016
.0018
.0124
.0082
.0106
.0032
.0073
.0086
Corrected to 7% 02 for the air-only tests
Corrected to 7% 02 considering the effect of oxygen en-
hancement
-590-
-------
TERRA VAC IN SITU VACUUM EXTRACTION PROCESS SITE DEMONSTRATION
Peter A. Michaels
Enviresponse, Inc.
.Edison, NJ 08837
Mary K. Stinson
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Releases Control Branch
Edison, NJ 08837
ABSTRACT
nv.n demonstrati°n testing of the Terra Vac In Situ Vacuum Extraction
Among the factors to be considered in technology selection
-591-
-------
INTRODUCTION
This SITE program demonstration was planned to demonstrate the
effectiveness of vacuum extraction technology in removing contamination to a
depth of 25 feet. The Terra Vac In Situ Vacuum Extraction demonstration, the
process for which a U.S. patent has been issued, was conducted on the property
of an operating machine shop in Groveland, Massachusetts, jhe property is at
the southwest corner and the highest elevation of the Groveland Wells Superfund
site, which contains the municipal water supply system for the town or
Groveland. The Superfund site has three sources of contamination, one of them
being the Valley Manufactured Products Co., Inc. machine shop. A leaking
storage tank and former improper storage and handling practices of waste oils
and decreasing solvents have caused the subsurface soil contamination with
volatile organic compounds, mainly trichloroethylene.
As is the case with many field projects, modifications had to be made to
the original demonstration plan to accommodate the situations encountered in
the field involving drilling, sampling, equipment installation, and operation
during the worst months of a severe New England winter. The results obtained
from this project will be applied by EPA to predict success of the vacuum
extraction technology at other Superfund sites.
SITE CONTAMINATION AND CHARACTERISTICS
The vadose zone at the location of the vacuum extraction wells is
approximately twenty-five feet in depth and is divided into two segments by a
claj lens that is three to four feet thick. Contamination is present above and
below the clay lens. Contaminated water is present as perched water on this
clay lens and as groundwater, which flows in a northwesterly direction into a
body of water known as Mill Pond.
The contamination is present mainly as trichloroethylene (TCE);
trans-l,2-dichloroethylene; 1,1,1-trichloroethane; and chloroform. The only
one of the contaminants that was actually in use as a deceasing solvent at the
machine shop was trichloroethylene. The other compounds are thought to be
biodegradation products of TCE. Trichloroethylene is by far the contaminant
present in the greatest amount. During the pretest soil sampling, the highest
concentration detected in the soil was 1600 ppm of TCE, approximately 18 to 20
feet below the surface at the southeasternmost portion of the Valley site.
SITE PREPARATION AND LOGISTICS
The site was prepared prior to mobilization of equipment in order to
conduct the test in an orderly and economical manner. Minor cleaning and
grubbing was required to assure access to the extraction wells and monitoring
well area. Of most importance in the field preparations were
o Procurement, placement, and outfitting of the two field trailers required
at the job site
o Providing the utilities required, mainly electric power
-592-
-------
areas
o Providing for personnel and equipment decontamination
o oven, refrigerator, hood
0
0
0
0
0
0
o organic Vapor Analyzer (OVA)
o personal computer with printer
o gas cylinders for nitrogen, argon/methane mixture, hydrogen, and compressed
o electronic balance
o gas tight syringes - 50 ml , 5 ml , and 1 ml
o liquid syringes - ioo/i], 50 jul , 10/ul, and 5 /jl
o glass bulbs, Tedlar bags
o barometer, thermometers, and manometer
o solvents - methanol , hexane
, too,
The EPA trailer was equipped with the following:
portable OVA with FID
portable photoionizer
portable GC with photoionization detector (PID) and computer
calibration gases
sound level meter
-593-
-------
o portable rotameter, vacuum gauges
o two-way radios
o water cooler and coffee maker
The 120V/60Hz power supply was tapped off the street utility pole and
metered to the two trailers. The 480V/3-phase/60Hz power requirement was
5p£d off an existing 480V line in the Valley plant This line was run inside
the building through a circuit breaker and meter for the purpose of supplying
power to the vacuum pump skid.
Two decontamination areas were provided at.the site: one Jor personne;I and
one for equipment. The equipment decontamination area was used primarily for
drill rig augers. Steam cleaning was employed to cut down on the amount of
water collected in the tarpaulin-lined bermed area. Decon water collected was
pumped into 55-gallon drums and temporarily stored on site to await ^1
disposal" Water for the steam generator or "jenny" .came from a hose connection
on the Valley building.
EQUIPMENT LAYOUT AND SPECIFICATIONS
The equipment layout is shown in Figure 1, and specifications are given in
Table 1 for the equipment used in the initial phase of the demonstration. This
equpment was later modified when unforeseen circumstances required a shutdown
of the system. The vapor-liquid separator, activated carbon canisters, and
vacuum pump skid were inside the building, with the stack discharge outside the
balding The equipment was in an area of the machine shop where used cutting
oils and metal shavings had been stored.
Four extraction wells and four monitoring wells (MW1 - MW4) were drilled
south of the shop. Each well was installed in two sections one section to
inst above the clay lens and one section to just below the clay lens. The
ex1ract?on wefls were screened above the clay and below the clay. As showrj in
Figure 2, the well section below the clay lens was isolated from the section
above by a bentonite portland cement grout seal. Each Action operated
independently of the other. The wells were arranged in a tr1 angular
confiauration with three wells on the base of the triangle (VE2, VE3, Vh4) ana
one will at the apex (VED- The three wells on the base were called barrier
wells Their purpose was to intercept contamination, from underneath the
bui dingand tl the side of the demonstration area, before this contamination
reached the main extraction well (VE1). It was the area enclosed by the four
extraction wells that defined the area to be cleaned.
INSTALLATION OF EQUIPMENT
Well drilling and equipment setup were begun on December 1, 1987 A mobile
drill rig was brought in, equipped with hollow-stem augers, split spoons, and
Shelbv tubes The locations of the extraction wells and monitoring wells had
beln staked out prev?ously based on contaminant Concentration profiles from a
previously conducted remedial investigation and from bar punch probe soil gas
monitoring.
-594-
-------
Figure 1. Schematic diagram of equipment layout at Terra Vac
SITE Program demonstration site, Groveland, Mass.
-595-
-------
TABLE 1. EQUIPMENT LIST
Equipment
Extraction wells
Monitoring wells
Vapor-liquid
separator
Activated carbon
canisters
Vacuum pump
skid
Number required
4 (2 sections each)
4 (2 sections each)
1
Primary: 4 units in
parallel
Secondary: 4 units in
parallel
Description
2" SCH 40 PVC 24' total depth
2" SCH 40 PVC 24' total depth
192-gal capacity, steel
55-gal drums with 200 Ibs of
carbon in each drum
2" inlet and outlet nozzles
25 HP motor - positive
displacement lobe type blower
3250 rpm
-596-
-------
2//PVCPIPE
BENTONITE
SAND
SCREENING
GROUT
T^BENTONITE
SAND
SCREENING
Figure 2. Schematic diagram of an extraction well
-597-
-------
Each well drilled was sampled at 2-foot intervals with a split spoon
pounded into the subsurface by the drill rig in advance of the hollow stem
auqer. The hollow stem auger would then clear out the soil down to the depth
of the split spoon, and the cycle would continue in that manner to a depth of
24 feet The drilling tailings were shoveled into 55-gallon drums for eventual
disposal. After the holes were sampled, the wells were installed using 2-inch
PVC pipes screened at various depths depending upon the characteristics ot the
soil in the particular hole. The deep well was installed first, screened from
the bottom to various depths. A layer of sand followed by a layer of bentonite
and finally a thick layer of grout were required to seal off the section below
the clay lens from the section above the clay lens. The grout was allowed to
set overnight before the shallow well pipe was installed at the top of the
grout. A layer of sand bentonite and grout finished the installation.
While sampling and well installation was going on, the rest of the _ _
equipment was being assembled. Four-inch flexible PVC piping was installed to
a distribution header at the inlet of the activated carbon canisters. Two-inch
flexible PVC piping was used to each of the four primary activated carbon
canisters. The 2-inch outlets from each canister were piped with flexible PVC
to a 4-inch header. A 4-inch flexible line was then run to the vacuum pump
skid, which discharged through a silencer to the atmosphere through a 4-inch
line.
When the well installation was complete, the three wells in line were
connected to each other by a 4-inch line or header. Each well section was
equipped with a block ball valve. The main extraction well was connected to
the 4-inch header by a 4-inch PVC line. Each well head had a septum from which
to take gas syringe samples, a connection for a vacuum gauge, and coupling
connections to attach a portable rotameter for flow measurement.
SAMPLING AND ANALYTICAL PROGRAM
The sampling and analytical program is the tool by which to judge the
effectiveness of this technology. The program was conducted in five separate
periods: the pretreatment period, the commissioning period, the active
treatment period, the midtreatment period, .and the posttreatment period.
Soil borings were taken during the pretreatment, midtreatment, and
posttreatment periods. The midtreatment period occurred 28 days after the
start of the 56-day active treatment period. The posttreatment period started
after the end of the active treatment period.
Soil borings were analyzed for volatile organic compounds (VOCs) using
headspace screening techniques, purge and trap GC/MS techniques, and the
EP-TCLP procedure. Additional properties of the soil were determined by
sampling using a Shelby tube, which was pressed hydraulically by the drill rig
into the soil to a depth of 24 feet. These Shelby tube samples were analyzed
to determine physical characteristics of the subsurface stratigraphy, such as
moisture, bulk density, porosity, pH, and grain size. (See Table 2.) These
parameters, which may influence the bulk transport of volatile organics through
the subsoil, will be used to define the basic soil characteristics.
-598-
-------
TABLE 2. ANALYTICAL METHODS
Parameter
Grain size
pH
Moisture (110°C)
i
Particle density
Oil and grease
EP-TCLP
TOC
Headspace VOC
VOC
VOC
VOC
VOC
VOC
VOC
*Third Edition, November 1986.
Analytical method
~——~^——.^—.__
ASTM D422-63
SW846* 9040
ASTM D2216-80
ASTM D698-78
SW846* 9071
F.R. 11/7/86,
Vol. 51, No. 216,
SW846* 8240
SW846* 9060
SW846* 3810
GC/ECD
GC/ECD
SW846* 8010
SW846* 8010
Modified P&CAM 127
SW846* 8240
Matrix
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil gas
Process gas
Separator liquid
Groundwater
Activated carbon
Soil borings
-599-
-------
The active treatment period was redesigned to be a total of 56 Jays in
order to be able to project a cleanup time for the site with greater confidence
than was possible inThe original 14-day period During the active treatment
period, gas sampling was conducted according to the schedule in Table 3.
Active treatment analyses were done in the field in the mobile laboratory
m a
Gas sampls were collected by syringes, and for co"centra^°^tSb^e^Dppm a GC
with an FID detector was used. For low concentrations a GC with an ECD
detectoVwas used Computer-generated curves show TCE concentration versus
timffSr each will, both shallow and deep. This enabled close tracking of the
progress of the vacuum extraction process.
Other measurements were taken routinely during the course of the active
treatment period, including
o flowrates from each well section, with a portable rotameter
o vacuum at each well section
o moisture content of well-head gas, using Modified Method 4 sampling train
o temperature
o Wattmeter readings on vacuum pump skid
Activated carbon samples were taken from all spen *«rbon canisters to
k for VOC loadinq This was done both as a check on the calculations tor
rhprk or oanq
thfquaStity of VOC Extracted from the soil and to gather data on adsorption at
various inlet concentrations of VOCs.
During this project, blanks, splits, and duplicates were collected as
required to assess the accuracy and precision of the sampling and analytical
methods employed for all critical parameters.
STARTUP AND OPERATING EXPERIENCE
The initial startup and operation of the system proceeded according to
Dlan Startup commenced on December 16, 1987 with high extraction rates of TCE
and oerched water from the clay lens. After less than a week's operation it
was decided to brelk ?or the holiday season Upon arriving back at the job
site, the piping, valves, and fittings were found to be frozen.
The entire piping network was then electrically traced and insulated, and
the system IS started up during the week of January 11, 1988. Large
Quantities of water were being extracted with rates up to 600 gallons per day.
Seated carbon wL being usld up at the rate of four canisters °r 800 pounds
oer day The system was shutdown on January 15 in order to scale up the
equipment. This was necessitated by the much higher than expected recovery
rates of volatiles and water. After the new equipment was installed, the
system was restarted on February 10.
The new equipment consisted of
-600-
-------
TABLE 3. ACTIVE TREATMENT SAMPLING SCHEDULE
Sample
locations
VE1-S
VE1-D
VE2-S,D
VE3-S,D
VE4-S,D
MW1-S,D
MW2-S,D
MW3-S,D
MW4-S,D
Separator
outlet
Primary carbon
outlet
Secondary carbon
outlet
Week 1
Once per
day
Every other
day
Every other
day
Every other
day
Weeks.2 & 3 Weeks 4 & 5 Weeks 6, 7, & 8
Twice per
day
Once per
day
111
Once per
day
Every other
day
Every other
day
Every other
day
Every third
day
Every third
day
Every other Every other Every third
day day
Once per
week
Every day
Once per
week
Once per
week
Every day
Once per
week
Once per
week
Every other
day
Once per
week
-601-
-------
o 1000-pound capacity stainless steel carbon canisters, 45" diameter x 64"
height
o 1000-gallon capacity vapor-liquid separator
o 1-HP centrifugal pump to pump liquid from the separator
o 2000-gallon holding tank for contaminated water
The new activated carbon canisters were topped off with an additional 200
pounds each of carbon from unused 55-gallon drums. The system was then piped
up with two primary carbon canisters in parallel followed by a single secondary
unit.
The small centrifugal pump was installed to take suction from the drain on
the 1000-gallon vapor-liquid separator. A level switch activated the pump when
the level in the separator rose to the half-full mark. The pump would then
pump a 300-gallon batch of contaminated water approximately 80 feet to the
holding tank, shutting off automatically when activated by the low-level
switch.
The new carbon canisters had operating times that made operation of the
unit much easier. The first set of primary canisters lasted five days; the
second set, nine days; and the third set, fourteen days. At this point the
program had reached the midpoint of the demonstration test.
After a five-day break, including a weekend, during which time midtreatment
soil sampling was done by split spoons adjacent to the installed monitoring
wells, the unit was restarted and run for an additional 28-day period.
WASTE DISPOSAL
As mentioned before, decontamination water during drilling and sampling
operations was collected and pumped into 55-gallon drums. Contaminated
groundwater collected by the vacuum pump system was initially pumped into
SB-gallon drums. These drums froze and had to be stored on site until they
were thawed with the advent of warmer weather. When the equipment change was
implemented, most of the water collected was stored in the 2000:gallon capacity
holding tank.
A tank truck came periodically to the site to pump the water from the
holding tank. This contaminated water was manifested and sent to a biological
treatment facility that was fully permitted and in regulatory compliance. The
SB-gallon drums of spent carbon along with the drilling tailings were
overpacked in salvage drums, manifested, and shipped as flammable solids to a
permitted incineration facility. The thawed drums of contaminated water
eventually were pumped into the tank truck for biological treatment. The
frozen drums had bulged and so were not suitable to be recycled. They were air
dried and cut up as scrap metal.
POSTSCRIPT
The major objectives of the demonstration have been to determine the
-602-
-------
ACKNOWLEDGMENTS
special support and cooperation that helped make this project a successful one.
-603-
-------
WHITE ROT FUNGUS DEVELOPMENT PROGRAM
by: John A. Glaser
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The development of the white rot fungus Phanerochaete chrysosporium
as a degrader of hazardous waste shows some promising opportunities for the
development of new hazardous waste treatment technologies. The fungus
secretes a mixture of strongly oxidizing enzymes that are identified as a
major component in the ability of this microorganism to degrade xenobiotics.
A line of research that strives to uncover the activity relationships of the
components of this enzyme mixture is only one component of the EPA's overall
development scheme for this organism. The enzyme studies re designed to
determine optimal enzyme compositions necessary for the degradation of
targeted pollutants. The control technology development is devoted to the
degradation of organic wastes associated with the remediation of wood pre-
servation technology sites. A soil treatment technology based on the fungus
is currently being developed and evaluated at the bench scale. Testing this
soil technology at field scale may be possible in the next calendar year.
An associated water treatment is further advanced to pilot scale testing.
-604-
-------
THE IMPACT OF A MODEL ORGANIC LEACHATE
ON SLURRY WALL PERFORMANCE"
Richard M. McCandless and Andrew Bodocsi, Ph.D.
Department of Civil & Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
Soil-bentonite slurry cutoff walls are frequently constructed to control
the migration of contaminated groundwater at Superfund sites. Previous
research by these authors and others has focused on the hydraulic characteris-
tics of slurry walls and the potential for the development and/or in situ
remediation of hydraulic defects (windows). Nearly all of this work, how-
ever, has involved the use of a clean model groundwater and uncontaminated
slurry wall construction materials. Moreover, although numerous authors
have studied the effects of various chemicals on the hydraulic performance
of soil-bentonite backfill, most of these studies have involved the use of
permeant liquids having high contaminant concentrations and the testing of
small (laboratory size) backfill samples. The purpose of this study is to
assess the impact of, a more representative model organic leachate on the
long-term performance of a near pilot-scale model slurry wall. Data and
conclusions cannot be presented at this time.
A model organic leachate has been designed for this project based upon a
statistical formulation developed by SAIC, Inc. and published in an EPA
report entitled "Composition of Leachates From Actual Hazardous Waste Sites".
Primary criteria for the design of a three-component model leachate included
maximum possible representativeness in terms of level of occurrence in. the
statistical formulation, miscibility and stability in aqueous solution, com-
patibility and lack of synergistic effects, and manageable health and safety
requirements. With these factors in mind, a model organic leachate contain-
ing phenol, acetone and N,N-dimethylacetamide at a total net organic fraction
of 6.0% by volume is proposed for use.
The slurry wall test tank used by the authors in the conduct of previous
studies has been substantially modified for this project. A circular test
wall approximately 508 mm (20 inches) in diameter, 559 mm (22 inches) in
height and 102 mm (4 inches) thick will be surrounded by a partitioned granu-
lar leachate collection zone which, in turn, will be surrounded by a second
cutoff wall serving to contain the leachate within the leachate collection
-605-
-------
zone. Vertical standpipes will be used to periodically sample leachate which
has permeated through the test wall at each of three different elevations.
The leachate samples will then be analyzed for each of the three organic
components and the results will be used along with measured flow parameters
to generate chemical and hydraulic breakthrough curves for the model. Test-
ing is scheduled to commence in the spring of 1988 and may continue for a
time period of up to one year, depending upon the aggressiveness of the model
organic leachate.
-606-
-------
BIOLOGICAL TREATMENT OF AQUEOUS ORGANIC HAZARDOUS HASTE
by: Lisa M. Brown
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Margaret K. Koczwara and Richard J. Lesiecki
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221-49TE
ABSTRACT
As part of a program to perform hazardous waste treatment technology
assessments, the USEPA has been directed to evaluate technologies applicable
to the treatment of organic wastes. Biological treatment is one of several
technologies being evaluated for treatment of hazardous waste at the USEPA
Test & Evaluation Facility.
While biological treatment is a potential technology for aqueous
organic waste containing contaminants in concentrations up to 10,000 ppm
these contaminants may be present in the waste at concentrations toxic to
the microorganisms, or they may be recalcitrant and require a different
treatment approach. Also, some compounds are more readily degraded in
aerobic systems, while others may require an anaerobic system for degrada-
tion.
At the USEPA Test & Evaluation Facility, bench-scale studies are being
conducted to determine toxicity/biodegradability of hazardous waste constit-
uents in both aerobic and anaerobic systems. Pilot-scale activated sludge
systems are being used to evaluate the fates of the waste constituents in an
aerobic system.
-607-
-------
DECONTAMINATION OF BUILDINGS, EQUIPMENT AND DEBRIS AT SUPERFUND SITES
by: Michael L. Taylor, Majid A. Dosani, John A. Wentz, Roxanne B.
Sukol, William R. Parker, and Jack S. Greber, PEI Associates,
Inc., Cincinnati, Ohio.
Naomi P. Barkley, EPA Hazardous Wastes Engineering Research
Laboratory (HWERL), Cincinnati, Ohio.
John Woodyard, IT Corporation, Knoxville, Tennessee.
Pat Esposito, BHE, Inc., Cincinnati, Ohio
ABSTRACT
It is well known that a large number (>1000) of sites in the United
States are contaminated with hazardous chemical residues. Many of these
hazardous waste sites include structures (houses, office buildings, manufac-
turing facilities) which are contaminated with hazardous organic chemicals
such as PCBs. In addition, many hazardous waste sites are littered with
debris (e.g., scrap metal, masonry materials, pieces of wood, equipment or
furniture) which is also contaminated.
Implementation of effective, non-destructive methods for on-site decon-
tamination of structures and debris would reduce the distribution of con-
tamination to off-site locations and could facilitate reoccupation of the
contaminated site including the structure. Several methods for decontami-
nating structures are available and have been implemented as described in the
"Guide for Decontaminating Buildings, Structures and Equipment at Superfund
Sites" (EPA 600/2-85/028); however at this time data are not available which
provide reliable indications of the relative efficiencies of these various
decontamination procedures.
The goal of this project is to generate reliable data which are indica-
tive of the relative efficiencies of various debris and building decontamina-
tion methods. Principle tasks include: 1) location of actual structures
contaminated with hazardous wastes which afford suitable sites for compara-
tive evaluations of building and debris decontamination procedures, and 2)
performance evaluations of promising building and debris decontamination
technologies at the selected field locations.
The first task of this project has now been completed. Several hazard-
ous waste sites were visited and ultimately two PCB-contaminated facilities
-608-
-------
were selected which met the established criteria for acceptability. Both
facilities are located in Detroit, Michigan and are less than one-half mile
apart. A concrete floor of a building located at one of these sites contains
levels of PCBs ranging up to 10,000 ug/g, while the second site contains
metal and masonry debris which is contaminated with PCBs.
In the second task, which is currently underway, efforts are being
focused on evaluation of two emerging technologies for removing PCBs embedded
in concrete structures. These two technologies are: 1) A method for in situ
degradation of PCB's entailing application of an alkali metal/polyethylene
glycolate mixture directly to the concrete surface, and 2) A hydroblasting
technique which entails use of a high pressure water jet (30,000 psi) to cut
away concrete surfaces. In order to evaluate these PCB removal techniques,
concentrations of PCBs in the top one-half inch of the contaminated concrete
floor will be determined prior to treatment by analyzing cores obtained from
selected locations on the concrete floor. Subsequent to the implementation
of each of the decontamination technologies, cores of the treated concrete
will again be obtained and the post-treatment PCB concentrations will be
assessed. The efficacy of the two decontamination techniques will be judged
on the basis of the percent reduction of PCBs achieved with each technique.
In addition, the costs of large-scale implementation of each of the two tech-
niques will be calculated and compared.
In addition to the evaluation of concrete decontamination procedures
described above, an innovative approach for decontaminating debris will also
be evaluated. A prototype of a portable module for on-site decontamination
of debris will be designed, built, and tested. The module will consist of an
enclosure into which debris can be placed and subsequently solvent-cleaned
using a closed loop solvent delivery system. During testing of the proto-
type, the efficacy of various solvent/additives combinations for removing
PCBs from debris will be evaluated.
In this presentation, the status of the work in progress will be thor-
oughly reviewed.
-609-
-------
CHARACTERIZATION OF 2.4,5-T DEGRADATIVE GENES IN PSEUDOMONAS CEPACIA
•—'STRAIN AC1100 AND SPONTANEOUS MUTANTS'
by: R. A. Haugland
J. Johnson
U.M.X. Sangodkar
A. M. Chakrabarty
University of Illinois at Chicago
Chicago, Illinois 60612
P. R. Sferra
A. Kornel
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Pseudomonas cepacia strain AC1100 is currently being evaluated in the
treatment of aqueous solutions different chlorinated phenoxyacetates and
phenols. The introduction of a naturally-occurring plasmid into this
organism was shown to extend the range of compounds that it efficiently
degrades. Other studies have shown that the ability of AC1100 to degrade
2,4,5-trichlorophenoxyacetic acid (2,4,5-T) and 2,4,5-trichlorophenol Is
often lost due to instability within certain regions of this organism s
genome. Two such regions have been isolated and identified is containing
2 4 5-T degradative genes using molecular cloning and phenotypic complemen-
tation procedures. Work is presently in progress to determine the physical
basis for the instability of these DNA regions.
-610-
-------
DEMONSTRATE COMPUTER ASSISTED ENGINEERING (CAE)
TECHNIQUES OR REMEDIAL ACTION ASSESSMENT—
P. R. Cluxton, W. G. Harrar, and L. C. Murdoch
University of Cincinnati
Department of Civil & Environmental Engineering
Cincinnati, Ohio 45221
ABSTRACT
A computer workstation dedicated to remedial action assessment is being
developed. The system is composed of several off-the-shelf software and
hardware modules, with software development limited to the integration of
these modules. The completed system will be an example of a Computer Assisted
Engineering (CAE) type system. _
The purpose of the project is to demonstrate how the remedial action
evaluation process can be improved and expedited thru use of the CAE system.
The CAE system capabilities are being demonstrated in two ways. First,
several small projects have been undertaken for the regional offices which
make use of a limited part of the CAE system, e.g., the contaminant mapping
capabilities of the system. Second, a Superfund site is the object of a
complete demonstration of the CAE system capabilities.
-611-
-------
ACTIVITIES OF THE LOUISIANA STATE UNIVERSITY
HAZARDOUS WASTE RESEARCH CENTER
Danny D. Reible
Department of Chemical Engineering
Louisiana State University
Baton Rouge, LA 70803
ABSTRACT
The Hazardous Waste Research Center at Louisiana State University (LSU) is
conducting fundamental and exploratory research in these general areas: en-
vironmental media/waste interaction, incineration, and alternative methods of
treatment/destruction. Individual research projects are being conducted by
multidisciplinary groups representing a number of academic departments. One
such project currently underway is entitled "Modeling Transport of Multiphase
Subsurface Contaminants", and deals with the transport and fate processes of a
nonaqueous liquid phase.
The presence of a separated non-aqueous liquid phase in the subsurface
often controls the rate and magnitude of ground-water contamination. In
addition, remediation efforts that do not directly address the non-aqueous
phase material are unlikely to provide cost effective or timely solutions to
the ground-water contamination risk. Recent research defining the transport
and fate processes of a non-aqueous liquid phase is described in this poster.
Process modeling efforts that hold promise of providing practical guidance for
regulatory development and site assessment and remediation planning are also
described. To-date, a model to predict unsaturated zone infiltration of
non-aqueous phases has been developed and preliminary models of saturated zone
aquifer/contaminant interactions have been proposed. Equilibrium and mass
transfer models between the separate fluid phases and the soil matrix have also
been proposed. Preliminary experiments testing modeling ideas for both
unsaturated and saturated flow have been completed. Additional validating
experiments are underway as well as comparisons to field data. The results
attained in each of these areas will be described in the presentation.
-612-
-------
INNOVATIVE DELIVERY AND RECOVERY SYSTEMS:
HYDRAULIC FRACTURING
Larry Murdoch
University of Cincinnati
Department of Civil and Environmental Engineering
Cincinnati, Ohio 45221
ABSTRACT
A hydraulic fracture will form adjacent to a borehole when injection
pressures exceed the sum of the confining stress and the resistance to
fracture propagation. Once the fracture is formed, sand is pumped in to
hold it open and provide a high permeability channelway that can be used
to increase both the recovery of contaminants and the delivery of reme-
diating substances.
Preliminary investigations, which consist of theoretical calculations
and analogies to applications reported by the energy industry, suggest
that hydraulic fractures can be created in many sites of contamination.
The orientation of the fractures is expected to be either horizontal in
rock and over-consolidated soil, or vertical in normally consolidated
soil. Lengths of hydraulic fractures are expected to be similar to their
depths of origin, and widths are expected to range from several mm to
several cm. Some horizontal fractures lift their overburden and they
could be one or more dm in thickness.
Possible application of hydraulic fracturing include increasing the
efficiency of pump and treat systems, stimulating the extraction of vapor
phases from tight soils, or forming a horizontal sheet-like drain to
capture the permeant in a leaching operation.
Currently, we are conducting lab experiments to measure the fracture
toughness—or resistance to fracture—of soils as a function of water
content or composition. Field experiments into the mechanics of hydraulic
fracturing in glacial till are planned for early spring 1988. Preliminary
results will be available at the Symposium.
-613-
-------
INCINERATION OF NITRATED PESTICIDES_1N__A
LOW NOx PRECOMBUSTOR/PACKAGE BOILER SIMULATOR
William P. Linak, Joseph A. McSorley
Combustion Research Branch, MD-65
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Ravi K. Srivastava, Jeffrey V. Ryan
Acurex Corporation
4915 Prospectus Dr.
Research Triangle Park, NC 27713
ABSTRACT
An investigation examining the incineration characteristics of the
nitrated pesticide, dinoseb (2-sec-butyl-4,6 dinitrophenol), was conducted
by EPA's Office of Research and Development/Air and Energy Engineering
Research Laboratory (ORD/AEERL) and Acurex personnel on EPA's 879 kW
(3,000,000 Btu/hr) low NOx precombustor/package boiler simulator. The
purpose of this study was to provide EPA's Office of Pesticides and Toxic
Substances (OPTS) with technical information regarding incineration disposal
options if these types of pesticides are banned from further use. Specific
information desired included the destruction and removal efficiencies
(DREs), and measurement of nitrogen oxide (NOx) emissions under different
incineration conditions, some of which included combustion strategies for
NOx control. Additionally, qualitative, measurement of products of incomplete
combustion (PICs), and quantitative particulate emissions were sought. In
conjunction with the Hazardous Waste Engineering Research Laboratory (HWERL),
the AEERL study examined a class of dinoseb product consisting of dinoseb
in organic solvent. Future tests are intended to examine a second class of
dinoseb in alcohol, alkanolamine, and water.
Results to date indicate that, although spiked dinoseb blank recovery
from XAD-2 resin was low (approximately 10 percent), no dinoseb was detected
in any incineration sample taken by modified glass SASS technique and
analyzed by GC/MS. Based on instrument sensitivity and dinoseb recovery,
DREs greater than 99.99 percent were achieved. PIC emissions, measured by
SASS and VOST sampling in conjunction with ,GC/MS analysis, indicate several
common combustion PICs in concentrations typically less than 10 ppb.
Particulate emissions were below the RCRA limit of 180 mg/sm3, and consisted
primarily of calcium sulfate and calcium oxide. Nitric oxide (NO) emissions
without any form of combustion modification for NOx control exceeded 4400
ppm (corrected to 0 % oxygen). When NOx controls, in the forms of air
staging and natural gas reburning, were employed these emissions were
reduced to below'150 ppm. Dinoseb contains 11.76 percent nitrogen.
-614-
-------
ASSESSMENT OF INTERNATIONAL TECHNOLOGIES
FOR SUPERFUND APPLICATIONS
Thomas J. Nunno
Jennifer A. Hyman
Alliance Technologies Corp.
Bedford, MA
ABSTRACT
Site remediation is a pressing issue in European countries due to limited
availability of land. As a result, much progress is being made in the
development of effective technologies for remediating contaminated sites.
The purpose of this program was to investigate successful and innovative
foreign technologies for potential application to the United States market.
This EPA-sponsored project was based on a 6-month research effort which
identified 95 innovative technologies in use or being researched worldwide.
The most promising technologies from this group were studied in-depth
through personal interviews with the scientists who research and apply
these technologies and through tours of laboratory models and full-scale
installations. These technologies, developed in Holland, West Germany, and
Belgium, include vacuum extraction of hydrocarbons from soil, in situ washing
of cadmium-polluted soil, rotating biological contactors for treating
pesticides in ground water, high-temperature slagging incineration of
low-level radioactive wastes, in situ steam stripping, and a number of
composting and soil washing operations. The results of this program provide
a detailed description of 12 site remediation techniques that have shown
promise in lab studies or in full-scale practice.
-615-
-------
PREDICTION OF THE FATE OF TOXIC METALS IN HAZARDOUS WASTE
~~ INCINERATORS "
by: R.G. Barton, P.M. Maly, W.D. Clark, and W.R. Seeker
Energy and Environmental Research Corporation
Irvine, California 92718
C.C. Lee
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
Emission of toxic metals during the incineration of metal bearing wastes
presents a potential health hazard of increasing interest to federal and
state regulatory agencies. A model which predicts the emission of metals
from combustion devices is being developed to aid in the formulation of
effective regulations and control strategies. Before such a model could be
developed, it was necessary to identify the phenomena that control the
behavior of metals during the incineration of waste materials. Examination
of the characteristics of emissions from a wide variety of incinerators led
to the identification of the following potentially important phenomena:
- Reactor thermal behavior
- Particle entrainment
- Metals reactions
- Metals vaporization
- Vapor condensation
- Particle coagulation
- Particle removal by air cleaning equipment.
Submodels addressing each of these phenomena were adapted from existing
models or were developed as new models. These submodels were assembled into
a composite model for metals partitioning in hazardous waste incineration.
The model was used to assess the impacts of various parameters on metal
emissions. Parameters considered include:
- Combustor temperature
- Waste chlorine content
- Saturation ratio
- Entrained particle size distribution
- Gas residence time
-616-
-------
- Waste sulfur content
- Gas oxygen concentration
- Metal
Temperature of the burning waste, waste chlorine content, gas oxygen
concentration and the metals present were found to have the strongest effects
on the predicted emissions.
Data which can be used to verify the model is scarce. However,
comparisons of the model's predictions with the experimental data which is
available indicate that the model correctly accounts for key phenomena.
-617-
-------
LOADING POINT PUNCTURABILITY ANALYSIS OF GEOSYNTHETIC LINER MATERIALS
by: Daren L. Laine
Michael P. Miklas, Jr.
Charles H. Parr
Southwest Research Institute
San Antonio, Texas 78284
ABSTRACT
This study examined geomembrane liner performance in laboratory tests
designed to subject several different liner materials to varying pressures,
temperatures, and point loads. Point loads were induced by placing the geo-
membrane material over truncated rigid epoxy cones arranged in three-cone
clusters in a sand subgrade and applying a hydrostatic load to the top side
of the liner. Cone heights of 9.5, 19.0, and 25.4 millimeters (mm) above the
subgrade were used in this study. Preliminary tests resulted in the selec-
tion of a cast epoxy resin cone having a 35-degree apex angle and truncated
2.8 mm from the apex at 45 degrees to the cone axis. Polyvinyl chloride
(PVC), chlorosulfonated polyethylene (CSPE), and high density polyethylene
(HOPE) materials, in two thicknesses each, were subjected to a constant
hydrostatic load of 17.93 kilovoltamperes (kPa) at 23°C and 50°C over a 365-
day period. HOPE material measuring 1.5 mm thick failed for the loading
point height of 25.4 mm above the subgrade. After 365 days, the loading
pressure was increased to 60.03 kPa for an additional 30 days. Failures were
induced in 1.5-mm HOPE for loading point heights of 19.0 and 25.4 mm and in
2.5-mm HOPE for loading point heights of 25.4 mm. Nonwoven geotextile fabric
material of 1.5, 3.8, and 5.3 mm was placed between the liner and the loading
points and selected tests were run. HOPE with a thickness of 1.5 mm failed
for a loading point height of 19.0 mm with a 1.5-mm geotextile placed between
the HOPE and the loading point. These tests were conducted at 17.94-kPa
pressure and ambient temperature. HOPE with a thickness of 2.5 mm overlaying
a 3.8-mm geotextile failed under 60.03-kPa pressure for a 25.4-mm loading
height at the high temperature test condition. No materials failed when
overlaid upon a 5.3-mm geotextile for any of the test temperatures or loading
heights. Transient pressure loading tests of the membrane material without
geotextile support showed failures when the pressure rate of change exceeded
55.20 kPa per hour. The maximum pressure load on the membrane proved to be
the failure stress factor rather than the rate at which pressure was applied.
-618-
-------
The test results indicate that moderate economic benefit may be gained
by allowing larger particles than current engineering practice deems accept-
able to remain in a subgrade surface. The loading point heights at which
the respective materials failed are correlated with the related subgrade con-
struction and geosynthetic material costs. With PVC and CSPE, without a geo-
textile underlay, particles as large as 25.4 mm above the subgrade would not
be expected to cause liner failure under conditions analogous to the study
parameters. Finished installation cost savings of up to 28 percent might
result if the largest particles in the subgrade were comparable with the
tested sizes. Performance of all of the liner materials was improved with
the addition of geotextiles, indicating a positive cost-benefit advantage
when a geotextile underlay is used with a geomembrane liner.
-619-
-------
RESULTS OF A LABORATORY CHARACTERIZATION OF PULP AND PAPER
MILL SLUDGE AND FLY ASH OF POTENTIAL UTILIZATION AS HYDRAULIC
BARRIER CONSTRUCTION MATERIAL IN LANDFILLS
by: Van Maltby
Jay P. Unwin
National Council of the Paper Industry
for Air and Stream Improvement, Inc.
Kalamazoo, Michigan 49008-3844
ABSTRACT
A laboratory investigation of the physical characterization
of sludge and fly ash produced by the pulp and paper industry
was conducted. Sludge samples were collected from both primary
and secondary wastewater treatment operations, and fly ash
samples were collected from wood, coal, and wood/coal fired
boilers. Hydraulic conductivity was the chief parameter of
interest. Special procedures were developed for the design and
operation of both rigid-wall and flexible-wall permeameters due
to the potential for biological activity and the low strength/
high compressibility nature of sludges. To minimize the effects
of consolidation, permeameters were operated at very low
hydraulic gradients (4 to 6). Hydraulic systems were converted
from falling-head to constant-head to minimize variations in
stress caused by a variable head. Back pressure-was utilized in
both types of permeameters to ensure a greater degree of
saturation.
Results suggest that there are a variety of industry wastes
that are sufficiently impermeable to serve as barrier materials
in landfill covers (having hydraulic conductivities of 10 6 to
ID"7 cm/sec or less). Fifteen sludges tested had hydraulic
conductivities ranging from 10~5 to 10~8 cm/sec, which generally
decreased with time. Eight fly ashes tested had hydraulic
conductivities between 10~3 and 10~7 cm/sec.
-620-
-------
THE U.S. EPA COMBUSTION RESEARCH FACILITY
by: Larry R. Water!and, Robert W. Ross, and Johannes W. Lee
Acurex Corporation
Jefferson, Arkansas 72079
ABSTRACT
During FY'87, 25 weeks of incineration testing took place in which 53
individual incineration tests were completed under the operations and
research program at the EPA Combustion Research Facility in Jefferson,
Arkansas. Test programs completed included: an extended evaluation of the
fate of volatile trace metals fed to the liquid injection incineration
system (LIS) and testing to evaluate the valence state of chromium
discharges from the LIS; testing to support evaluation of the composition
of residual discharges from the incineration of five listed hazardous
wastes from specific sources; and preliminary evaluation testing of the
American Combustion Pyretron Thermal Destruction System as an innovative
Superfund site waste treatment technology. Results from these tests -re
abstracted in this poster. Detailed results from the American Combustion
Pyretron system tests are discussed in a separate paper in this symposium.
-621-
-------
WASTE REDUCTION INNOVATIVE TECHNOLOGY EVALUATIONS (WRITE) PROGRAM
by: Michael F. Szabo
PEI Associates, Inc.
Cincinnati, OH
Harry M. Freeman
HWERL, U.S. EPA
Cincinnati, OH
ABSTRACT
The purpose of this poster session will be to familiarize conference
participants with the WRITE program that HWERL is initiating as a part of
its Waste Minimization Research Program. The WRITE program is designed to
involve EPA with private industry to encourage the development and/or
demonstration of effective techniques and technology for hazardous waste
minimization. Initial plans are to structure the program with private
industry to be similar to the Agency's Superfund Innovative Technology
Evaluation (SITE) program where EPA generally provides funds to support
only the evaluation of the demonstration. The WRITE program will there-
fore provide credible technical information on new waste minimization
processes. This paper will provide background on the WRITE program and
discuss program activities planned for FY88 and 89.
-622-.
-------
CASE HISTORIES OF
: UNDERGROUND STORAGE TANK CORRECTIVE ACTIONS
by: Joyce K. Hargrove and William M. Kaschak
COM Federal Programs Corporation
Fairfax, Virginia
Robert W. Hillger and Richard A. Griffiths
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Edison, New Jersey
ABSTRACT
The Environmental Protection Agency's (EPA) proposed regulations for
?±SrOU? ?fc°uage tankS (UST) require ****• Corrective action be taken in
response to leakxng UST. However, the experience of personnel in the EPA
regions, the states, and the local environmental agencies in this new field
varies widely As a result, what constitutes appropriate correlation
for leaking underground storage tanks is not well defined. The EPA is
expanding its Case History File (File) database to facilitate technology
transfer among the new personnel involved in underground storage tank
corrective actions and those involved in hazardous waste site cleanup.
pom JhenFile.contains reports submitted by On-Scene Coordinators (OSCs) and
Remedial Project Managers (RPMs) about technical, administrative, financial,
hSSSS* '£*,??** aS?eCtS °f the spi11 and waste-site cleanups they have
sSShf; tn h/U5 C0ns^sts of a Database section, which allows computerized
searches to be made, and a narrative section, which is a detailed report on
the response. The database section offers menu-controlled searches in any
of 27 categories. The narrative section is organized into the following 10
subsections: General Information, Chemical Information, Effects of the
Incident, Site Characteristics, Containment Actions, Removal/Cleanup
Actions,_Treatment Actions, Disposal Actions, Operational Considerations,
and Termination of the Response. The File is being modified to incorporate
additional data relevant to UST, such as methods of detection, causes of the
leak, tank/piping construction, etc. New reports are being added as the EPA
receives them from the states and regions.
The File may be accessed with the use of a computer with modem; a
telecommunications program; and a user's guide which provides detailed
instruction on the system's operation. The File is managed by the Technical
information Exchange (TIX), Releases Control Branch of the EpI's HazSctois
Waste Engineering Research Laboratory in Edison, New Jersey.
This poster describes the Case History File and presents synopses of
underground storage tank case histories. P
-623-
-------
HOSPITAL WASTE INCINERATION
by: Teri L. Shearer
U. S. Environmental Protection Agency
Cincinnati, OH 45268
ABSTRACT
The popularity of infectious waste incineration has grown steadily
since the 1984 amendments to the Resource Conservation and Recovery Act of
1976 were passed; these amendments banned the landfill of such untreated
wastes. Generators of hospital waste are now responsible for treating and
disposing of an estimated 5,900 tons of waste/day, of which a significant
portion consists of plastics, pathogens and RCRA regulated components.
Emissions range from particulates and PCBs to potentially viable viruses.
Currently there are no federal regulations in force that mandate the
control of emissions from hospital waste incinerators. Until clear-cut data
is presented that corroborates harmful health and environmental effects, the
need for such regulations will, in all likelihood, remain unwarranted. The
majority of states have non-specific requirements for the disposal of
hospital wastes. However, most are complicated and rarely address the key
issue of emission control (although this aspect is changing due to increased
public awareness and concern). Combined with the innate incapability of some
existing hospital waste incinerators to adequately handle the complex waste
streams, it is apparent that the need for concerted effort to investigate,
rectify and regulate hospital waste incinerators now exists.
-624-
-------
ASSESSMENT OF SOLIDIFICATION/STABILIZATION TECHNOLOGIES
FOR SUPERFUND CONTAMINATED SOILS
Richard M. McCandless, P.E., C.P.G.
Department of Civil & Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
Solidification/Stabilization treatment processes are currently being
evaluated as possible cost effective alternatives for high volume, low con-
taminant concentration soils at Superfund sites. Work is now in progress at
the Center Hill Solid and Hazardous Waste Research Facility (HWERL, Cincinna-
ti) to assess the feasibility and efficacy of various solidification/stabili-
zation technologies. General objectives of this program are to:
° determine technical justification for current standard test methods
used to evaluate process performance and recommend modifications or
, new methods as appropriate;
technical assistance services to the EPA Regions in the form of
field and/or laboratory treatability studies; and
« contribute to the development of a broad technical data base character-
iz ing the performance of various treatment processes applied to a
range of soil/contaminant systems.
Since initiation of the project in October 1986, a capability to perform
most of the standard physical and leaching test procedures has been estab-
lished. Service activities to date have included field and laboratory stud-
ies to evaluate the performance of generic or vendor-proposed solidification/
stabilization processes for use at the United Chrome, Sand Springs and White-
house Oil Pits NPL sites. Current research tasks include the testing of
standard analytical reference matrix (SARM) soils in support of Agency
of™"8 ,° establish BDAT Performance levels and the testing of modified
SARM soils to more fully evaluate various leaching test methods currently
considered to be the primary measures of performance.
-625-
-------
SMALL QUANTITY GENERATOR RESEARCH PROGRAM;
MINNESOTA TECHNICAL ASSISTANCE PROGRAM
by: Cindy A. McComas
Minnesota Technical Assistance Program
University of Minnesota
Minneapolis, MN 55455
James Bridges
U.S. EPA
' Cincinnati, OH
ABSTRACT
The U.S. Environmental Protection Agency (EPA) has established a Small
Quantity Generator Research Program with the special hazardous waste needs
of small businesses in mind. This program is developing and promoting the
use of innovative technologies and management practices that reduce or
prevent the generation of hazardous waste, through a number of research,
development, and demonstration projects. Results of this program are
providing practical and economical solutions to assist small businesses
with proper hazardous waste management, reduction, and regulatory compliance,
S a non-regulatory framework. In order to achieve the objectives of this
program, a two-year cooperative agreement was established between the EPA
and the Minnesota Technical Assistance Program (MnTAP). This cooperative
agreement established a small quantity generator research awards program in
Minnesota. The research program is designed to provide small quantity
generators of hazardous waste with new methods or technologies for waste
reduction, or new applications of existing technologies. A total of nine
research awards have been established with industry, associations, and
academic institutions in Minnesota. Topic areas for these research awards
include cyanide detoxification using ozonation and blue-green algae, -
pesticide rinsate reuse and recycle, treatment of chromic acid laboratory
waste, optimizing the use and reuse of water-soluble coolant, organic
solvent removal from wastewater using photooxidation, metal removal and
recovery from radiator repair wastewaters, copper and lead recovery from
metal finishing wastewaters using aluminum displacement, redesign ot
chemistry lab experiments to minimize waste, and treatment of caustic_
degreaser waste from engine rebuilding shops. Results from these projects
will be disseminated through publications and presentations to benefit
small quantity generators nationwide.
-626-
-------
PILOT-SCALE PHYSICAL SEPARATIONS FOR TREATMENT OF
ORGANIC HAZARDOUS WASTES
Jeffrey P. Herrin and Sardar Q. Hassan
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221-49TE
Douglas W. Grosse
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
With the increasing demand for elimination and/or reduction of land
disposal of hazardous waste, it has become necessary to look for viable
alternative waste treatment technologies. As a part of the USEPA "Hazardous
Waste Treatment Technologies Assessment" program, various physical separation
processes will be tested for treatment of organic hazardous wastes. These
processes include steam stripping of volatile organics, batch distillation
for solvent recovery, solvent extraction and activated .carbon adsorption.
Test plans for different technologies are at different stages of development
with steam stripping and batch distillation as active projects for current
fiscal year and solvent extraction and carbon adsorption as future projects
After thorough review of EPA listed hazardous wastes and existing technolo-
gies a number of wastes have been identified as probable test wastes for
steam stripping and batch distillation. A steam stripping unit with a two
inch column and variable packing height has been fabricated for removal of
volatile organics. A batch distillation unit with a six inch column and
fifty theoretical plates has been installed for solvent recovery testing.
Results of the tests and future plans in this area will be presented.
-627-
-------
APPLICATIONS OF SUPERCRITICAL EXTRACTION TO ENVIRONMENTAL CONTROL
by: Gregory W. Leman, Steven R. Alferi, David L. Tomasko & Charles A. Eckert
Department of Chemical Engineering
University of Illinois, Urbana, IL. 61801
ABSTRACT
Preliminary economic analyses have already demonstrated that the
extraction of organic toxins by supercritical fluids may have substantial
cost advantages over the conventional methods currently being practiced.
Specific cases studied were extraction of contaminants from soil, and the
supercritical fluid regeneration of granular activated carbons (GAG) used
in the cleanup of wastewaters and leachates.
We are now in the process of gathering additional experimental data
for the refinement of the preliminary designs. We are taking specific data
for the solubility of contaminants in supercritical C02, and for C02 con-
taiminated with water. Other studies are aimed at the equilibrium between
adsorbed contaminants on GAG and supercritical C02. Further we have con-
structed and begun operating a pilot plant for the regeneration of contam-
inated GAG and the separation of toxins, and we report on these preliminary
results as well.
Finally we are using the new data available with process design, sim-
ulation, and cost estimation processes such as ChemCAD. For this package
we have modified the thermodynamics part in the supercritical and near-
supercritical region to reflect the new experimental results. This has
permitted much more precise process calculations and economic feasibility
studies. The new designs for both mobile and stationary GAG regeneration
units are presented.
.-628-
-------
Results of a Trial Burn of EPA's Mobile Incineration System
by: Gopal Gupta, Robert Sawyer, James Stumbar
Enviresponse, Inc.
8 Peach Tree Hill Road
Livingston, N.J. 07039
Frank Freestone and Joyce Perdek
U.S. E.P.A.
Hazardous Waste Engineering Research Laboratory
Releases Control Branch (Edison, N.J. 08837)
Cincinnati, Ohio 45268
Minda Ho
Linde Division, Union Carbide Corp.
Old Saw Mill River Road
Tarrytown, NY 10591
ABSTRACT
Under the sponsorship of the Office of Research and Development of the
?MKV ^Tm-nta] Pr°tect1on A9ency (EPA), the Mobile Incineration System
MIS) was designed and constructed to demonstrate the high-temperature
incineration of toxic and hazardous wastes. The system consists of a
^nnlry-11ned :°tary ki1n' a seco^ry combustion chamber (SCC)f and an
air pollution control system mounted on three heavy-duty semi-tra ers Flue
gas and stack gas monitoring equipment is contained in a fourth trailer.
In a series of tests and trial burns conducted between 1982 and 1985 th?
orqaSicVl?au?d,abi ityi?° de^y PCB-co"^minated and other chlorin ?ed
organic liquids, as well as dioxin-contaminated liquids and soils The
results of these tests and of the field demonstration, duHngwhch the MIS
kg °f ^"^ing materials,^ been
uc ? s.ries of suPPlemental tests were conducted on the MIS
RCRA npm?t°fgh S^m^r^ 19S7 ' Th^ work culminated with the issuance of
a RCRA permit for pesticide wastes. The MIS was able to achieve the required
destruction and removal efficiencies (DREs) for all principal orqanfc
hazardous constituents (POHCs) tested! POHCs consisted of- 1 J9?
iastes°rcarSonnteptartE?ly-H10^na^d fiPhe^ls' in both the liquid and solid
the solid wa?te' in qUid W3Ste; and he>
-------
Preparatory tests conducted prior to the trial burn showed that high
rates of chlorine input increased particulate emissions despite satisfactory
acid gas removal. Consequently, the air pollution equipment was modified
prior to the trial burn.
Supplementary tests were conducted to check the feasibility of feeding
pesticide-contaminated vermiculite and brominated sludges. Vermiculite
(0.0648 g/cm3 bulk density) feedrates up to 365 kg/hr were achieved, and
measured particulate emissions were well below regulated limits. The tests
of solids containing 30% brominated sludge demonstrated a feedrate of 1820
kg/hr.
-630-
* U.S. GOVERNMENT PRINTING OFFICE: 1989- 648-163' 87057
-------