-------
oo
LU
o
o;
o.
<=c
:c
o
O
o.
o
O
=c
•a:
UJ
CQ
•=C
c
o
i.
ro
Q
£
o
cu
4-
o
S-
Q_
cu
ro
o
ro
cu
— i
,
ro
E
o
E
,
0
CL.
4-
0
j_
CU
~O
S-
o
Q
ro
ro
Q
cr
E
"l~
0
ro
cu
—
o
•*-*
l ^
^
T~
'o
^:
o
4_>
ro
CO
oo cu
cj
CD E
E ro
E S-
•
cu
ro
o =c
ro
c:
cu
Q-
3 3 O CJ <
o-o-
^HCDt-HCO^HO OOOOCDC3CDCD
oo o o o
«—< cr> cn ufi CNJ
O C3 O O CNJ
O O O O O
O O O O O
co CM i—i to oo ^r
*-* I*-- O O T-I »-«
^j- i—i f-H O
S-
(O O
S- 0
CO -r-
-E CU
cu ro
O -E
S- •(->
o cu
i— CU O
-E E i.
o cu o
I CU O
CM O •!-
•> S- S.
•-I O I—
I 1— I
in -c oj
C U "
f- * =tt=
[a * =tfe * =«t=
153
-------
USE OF BATCH ADSORPTION PROCEDURES FOR
DESIGNING EARTHEN LINERS FOR LANDFILLS
C. C. Ainsworth, R. A. Griffin, I. G. Krapac and W. R. Roy
Illinois State Geological Survey
615 East Peabody Drive
Champaign, Illinois 61820
ABSTRACT
An investigation is being undertaken to develop a "Batch Sorption" procedure and
scientific documentation for hazardous wastes on soils and soil constituents. The
procedure will help in evaluating various earthen materials with respect to their
ability to attenuate hazardous waste. A very simple procedure has been adopted and is
now being tested as to its flexibility, reproducibility, and usefulness. The procedure
will be revised and tested using various soils from different parts of the country. Soil
CO solution ratio, most effective mixing method, separation techniques, and sample size
are just some of the variables that are being addressed in this investigation. As each
question is studied the procedure is revised and upgraded to reflect the data obtained.
In using this systematic approach, the procedure will be well documented by laboratory
experimentation and data and will hopefully be of use in critical evaluations of an
earthen material's ability to attenuate pollutants.
INTRODUCTION
The retention of chemical species in
various soils and soil constituents have
been studied for many years. The adsorp-
tion of inorganic cations and anions by
clays and soils (Bingham et al., 1964;
Griffin and Shimp, 1976; Frost and Griffin,
1977; Harter, 1979 and others) has been
investigated; as has organic compounds
(Griffin et al., 1978; Chou and Griffin,
1981; Hassett, 'Banwart, and Griffin, 1983).
Although there are many variations, the
"batch" adsorption technique has been
widely used. This consists of mixing a
solution of known composition and concen-
tration with a mass of soil for a given
period of time. The solution is then
filtered and analyzed for concentration
changes to determine adsorption by the
soil. This has been called the standard or
traditional "batch" technique. In point of
fact there is nothing truely standard about
the procedure. In reviewing just a portion
' of the literature, 80 batch adsorption
studies were found using various adsor-
bates and adsorbents. Of these, few were
procedurally comparable.
Apart from the inherent variability
of different soils there are a considerable
number of factors that affect the sorption
of a given constituent. For inorganic
species some of these are temperature,
method of mixing, composition and concen-
tration of the solution, and moisture con-
tent (Barrow, 1978). Organic species are
much more complicated; this is due to the
need to identify and control various re-
actions that would reduce or in some way
affect the precision and accuracy of the
results. Photodegradation, polymerization,
microbial degradation, hydrolysis, and
volatilization are reactions that could
effect the outcome of an organic sorption
study.
154
-------
The development of uniform batch-
test sorption procedures would enable
investigators to obtain sorption-retention
data of known reproducibility under
standard conditions. This would help in
comparing and evaluating various materials
for their suitability as landfill liners.
It would also allow calculation of minimum
thicknesses of clay liners to remove pol-
lutants from landfill leachates (Griffin
et al., 1976).
To facilitate the comparisons of
various systems (adsorbent/adsorbate)
sorption data has been fitted to adsorption
curves (Sibbesen, 1981; Barrow, 1978).
However, if the techniques used to obtain
the data are not comparable then the com-
parisons that are made may be erroneous.
It is the hope of this investigation that
a procedure and guidelines can be devel-
oped to allow for useful and reproducible
collection of adsorption data for evalu-
ation of landfill liner materials.
PURPOSE
The purpose of this research project
is to develop a Technical Resource Docu-
ment and a Background Document to be used
for performing batch type sorption studies
and for using this data to estimate pol-
lutant retention by earth material liners
at hazardous waste disposal facilities.
The specific objectives are:
1. To develop standardized procedures for
collection of sorption data.
2. To prepare a document detailing the
routine use of the procedures.
3. To prepare a background document
describing the scientific basis and
rationale behind the procedures
developed and recommended for use.
In using a batch sorption procedure
to develop a sorption curve for a given
system, the soil to solution ratio used
for an investigation may be predicted upon
the ability to detect the steady-state
concentration of the specie of interest.
Hassett et al. (1980), using a radio-
isotope labelling as a means of detection,
needed three different soil to solution
ratios to determine sorption properties of
14 organic compounds onto 14 soils and
sediments. Without the use of isotopic
labelling the soil to solution ratios
would probably have been much wider and
more varied. In this respect a single pre-
determined soil to solution ratio is not
feasible.
A single sorption curve (equilibrium
concentration versus amount of sorption),
should, in theory., describe a given system
at equilibrium no matter what the soil to
solution ratio. Whether one holds the con-
centration (initial) constant and varies
the soil to solution ratio or vice versa, the
same adsorption curve should describe both
techniques. This of course is true provided
all other parameters are equal or accounted
for in the adsorption curve. Griffin and Au
(1977) found Pb sorption on Ca-montmoril-
lonite to be apparently dependent upon the
soil to solution ratio. However, once the
Ca ion activity in solution was accounted
for, a single adsorption curve described
the system regardless of the soil to solu-
tion ratio.
In more complex systems, such as soils
which contain many materials in addition to
clay minerals, results of soil to solution
ratio studies have been less straight-
forward. Using Atrazine (2-chloro-4-
(ethylamino)~6-isopropylamino-5-triazine)
as the adsorbate, Dao and Lavy (1978) and
Anderson et al. (1980) found that soil to
solution ratio had no effect on the amount
sorbed by soil. Nearpass (1967), and Green
and Obien (1969) found a 1.4 fold decrease
and 1.6 fold increase, respectively, in
Atrazine sorption with an increase in soil
to solution ratios. However, both felt
these were statistically insignificant.
Grover and Hance (1970) saw a decrease in
the Freundlich constant (Kf), by a factor
of 2.6 when the soil:solution ratio was
changed from 1:10 to 1:0.25. Koskinen and
Cheng (1983) also found differences in the
Kf for 2,4,5-T sorption on soils which
were ascribed to soil:solution ratio.
Their results indicated the variation was
only present at higher soil to solution
ratios.
While a great number of systems reach
steady-state conditions rapidly others re-
quire days and weeks. Bar-Yosef (1979)
studied Zn sorption on soils and determined
the steady-state conditions to be reached
only after 16 to 168 hours. Hope and Syers
(1976) found that phosphate sorption in
soils had not reached equilibrium after 146
hours. R. E. White (1966) found equilibrium
between phosphate and certain soils to be
attained only after 120 days. W. R. Roy
155
-------
(unpublished data) found Mo, As, and P to
take up to 5 days to reach steady-state
conditions. Besides the dependence on the
soil used or constituent of interest the
rate at which sorption occurs (and hence
attainment of steady-state) also has been
shown to be dependent upon soil:solution
ratio, vigor of shaking, and method of
shaking (Barrow and Shaw, 1979). Unfor-
tunately, the dependent variables mentioned
above are not predictable. Therefore, it
would be presumptuous to arbitrarily set a
time interval to carry out the procedure.
However, beyond about a 24 hour period
the investigator must be conscious of con-
comitant reactions that may affect the
adsorption study. Of primary importance is
raicrobial assimilation or degradation. This
is particularly true of organics.
Although the literature does suggest
that low soil to solution ratios are less
likely to cause problems, it can be seen
from the previous discussion that work re-
mains to be done on this topic. Even if the
rationale has not been worked out for a
procedure which accounts for this phenome-
non, it does not preclude the practical
application of an empirical soil to solu-
tion ratio given certain constraints. This
is one of the areas that is actively being
pursued via laboratory investigations in
this project.
While the procedure for organics may
be similar to inorganics, the work to
develop the laboratory basis for such a
procedure is much slower. This is primarily
due to analytical considerations. However,
tests must be designed to allow for the
determination of various side reactions
that may bias the batch sorption study re-
sults (i.e.: photodegradation, volatili-
zation, etc.). Therefore, organics are not
as easily characterized as inorganic pol-
lutants with respect to developing docu-
mentation of a sorption procedure.
APPROACH
The approach to developing a useful
batch adsorption procedure is based on the
premise that to be useful it must be flex-
ible enough to accomodate those that would
use it and have a well founded scientific
rationale. The flexibility of a procedure
is seen by the investigator's ability to
use various techniques and equipment with-
out altering the results of a given study.
Each step and option (point of flexibility)
within the procedure must be shown to have
a rational and scientific basis. The latter
is to be determined through the use of the
published literature and laboratory
studies.
Figure 1 is a flow chart of the
initial proposed procedure that was based
on a draft procedure developed by ASTM
Committee D34.02.03. From this, one may
identify the points at which the procedure
may be unnecessarily confining. These are:
1) mass of soil needed, 2) soil:solution
ratio, 3) the type of container, 4) type of
mixing apparatus, 5) time of shaking and,
6) mode of separation. Other points that
may need to be studied are temperature and
ionic strength at which the investigations
are to be carried out.
As the procedure is run and points in-
vestigated, a new generation of the proce-
dure is developed. This is then tested for
reproducibility through an ASTM sensitivity
analysis using several volunteer labora-
tories. The cooperating laboratories run
the procedure on a given soil using pre-
scribed solutions. Results and comments on
the procedure are then correlated. This is
very useful in identifying problems and
appropriate changes. From the first test
the major problem identified was that of
shaking vigor.
Concomitant to the actual testing of
the procedure itself, laboratory studies
are being carried out to investigate
various improvements in the procedures such
as methods of separations, variations in
sample size, etc. Also, attempts are being
made to determime how to best delineate the
range of solutions used, soil to solution
ratio, and time of mixing. At the present
time only one soil (Catlin silt loam) and
two adsorbates (arsenate and cadmium) are
being used for inorganic adsorption tests.
Once a procedure is completely worked out,
it will be verified on five additional
soils from various parts of the United
States .
RESULTS
Most of the laboratory work to date
has been the collection and characteriza-
tion of the seven soils to be used in the
investigation. This includes, mineralogy,
surface area (N2 B.E.T.), particle size
distribution, cation exchange, native
cation and anion determinations (both H20,
KC1, and NHijCl extracts) and various other
tests. Some work has been done on the
points of interest in the procedure. A
156
-------
ISGS 1984
Figure 1. Flow chart for sensitivity test.
157
-------
Cat 1in AI horizon (fine silty, mixed mesio,
Typic Argiudoll) has been used as the
sorbent in these initial studies. This
soil has 2-3% organic carbon, 24% clay,
65% silt, 11% sand and has a pH of 7.1
(1:1; H20).
Using 200 ppm of Cd (as CdCl2) dis-
solved in deionized HgO as the sorbate, a
kinetics study was carried out using 350 g
soil in a 1:4 soil:solution ratio. Samples
were allowed to shake on a reciprocating
shaker (100 strokes/min; 1.25 inch throw)
for periods of 1 to 32 hours. Samples were
filtered and then analyzed using an in-
ductively coupled argon plasma (ICAP). In
conjunction with this study, triplicate
samples of the above prescription were al-
lowed to shake or stir (paddle stirrer) at
various speeds and displacements. Also,
using a 200 ppm arsenate solution (as
NaaHAsOi,), a sample size-separation study
was conducted. This consisted of triplicate
samples of Catlin soil at a 1:4 ratio of
the As solution using either 350 g or 50 g
of soil. The samples and 2 separation
techniques (12 runs total) were shaken at
100 strokes/min and 1.25 inch displacement
for 24 hours on a reciprocating shaker„
Samples were then separated using either
filtration or centrifugation. Filtration
was accomplished using four 25 mL aliquots
passed through a 0.45 yra pore size filter;
the initial aliquot was discarded. The
centrifugation used either 50 mL or 150 mL
of the suspension in nalgene tubes at 1.6 x
101* or 5.0 x 103 rpm, respectively, for 30
minutes at constant temperature.
Figure 2 illustrates that the Catlin
soil/Cd solution reached a. steady-state Cd
concentration in approximately 2 hours.
However, after shaking similar samples for
24 hours at various speeds, none of the
other suspensions had attained the final
concentration that set C (Table 1)" had
after 2 hours. Barrow (1978) showed that
various shaking vigors could affect the
speed at which a steady-state condition was
reached. But, he also demonstrated that
either ped destruction or abrading of
surfaces could also cause sorption to take
place at an increased rate. This latter
point does not happen to every soil. It may
well be that the differences observed here
are a function of one or both of the above
phenomena. Since this experiment was run,
an end-over-end tumbler has also been
employed in the investigations, which
promises to provide superior results.
The results of the sample size-
separation study demonstrated that sample
size need not be as voluminous as original-
ly proposed (Table 2). Also, it would ap-
pear that the separation can be obtained
through filtration or centrifugation with-
out sacrificing reproducibility.
While these studies are preliminary
and the breadth of samples narrow, the
results do help in identifying problems. As
a consequence of these studies the maximum
mass of soil has been decreased from 350 g
to 70 g, filtration or centrifugation has
been adopted and, due to other work simi-
lar to the above, the soil/solution ratio
for the Catlin soil has been revised to
1:20. Now that the seven soils to be used
in the investigation have been collected
and characterized, the breadth of experi-
ments involving both adsorbent and adsor-
bate will be extended. Through this type of
systematic investigation it is felt that a
reliable, well documented procedure can be
produced that will be useful in assisting
in the selection of earthen materials for
landfill liners.
ACKNOWLEDGEMNTS
Financial assistance from the U. S.
Environmental Protection Agency, Solid and
Hazardous Waste Research Division, Award
No. R-810245-01 and the assistance of our
project officer, Dr. Michael Roulier, are
gratefully acknowledged. Also, we would
like to thank ASTM and ASTM D34.02.03
participating laboratories for giving of
their time and effort so freely.
LITERATURE CITED
Anderson, J. R., G. R. Stephenson, and
C. T. Corke. 1980. Atrazine and
cyanazine activity in Ontario and Mani-
toba soils. Can. J. Soil Sci. 60:773-781.
Barrow, N. J. 1978. The description of
phosphate adsorption curves. J. Soil Sci.
29:447-462.
Barrow, N. J., and T. C. Shaw. 1979.
Effects of solution:soil ratio and vigor
of shaking on the rate of phosphate ad-
sorption by soil. J. Soil Sci. 30:67-76.
Bar-Yosef, B. 1979. pH-dependent zinc ad-
sorption by soils. Soil Sci. Soc. Amer.
J. 43:1095-1099.
Bingham, F. T., A. L. Page, and J. R.
Sims. 1964. Retention of Cu and Zn by
H-montmorillonite. Soil Sci. Soc. Amer.
Proc. 28:351-354.
158
-------
c
'E
£ S
I *
O
-------
TABLE 1. RESULTS OF 24 HOUR SORPTION STUDY USING
VARIOUS VIGOR OF SHAKING AND AN INITIAL
Cd CONCENTRATION OF 200 ppm
Cd
Set
A
B
C
D
in Solution Shaker Rate '
ppm strokes/min throw (inch)
13.82* 59 3"
19.3* 70 1.25"
1.78* 100 1.25"
3.1* paddle stirrer
*Each concentration is the average of 3 separate
runs .
Sample
1
2
3
4
X
S
TABLE 2.
350
1400 mL
Filter
128.0
131.0
127.0
128.6
2.08
RESULTS OF THE SAMPLE SIZE-SEPARATION STUDY
USING THE CATLIN SOIL AND AN INITIAL As
CONCENTRATION OF 200 ppm
200 ppm initial As solution
g soil 50 g soil
solution 200 mL solution
Centrifuge* Filter Centrifuge** Centrifuge*
•
129.0 126.5 124.5 129.5
128.0 128.0 126.5 127.0
127.0 131.0 126.0 127.0
136.0 126.0 127.5
128.0 130.4 126.0 127.8
1.0 4.19 1.08 1.10
* - 50 mL of the suspension was centrifuged at 1.6 x 101* rpm for
30 min.
** - 150 mL of the suspension was centrifuged at 5.0 x 103 rpm for
30 min.
160
-------
Chou, S. F. J., and R. A. Griffin. 1981.
Adsorption of PCBs by cellulose fiber
filteraids and carbonaceous adsorbents
used for water treatment. Fourth Annual
Conference of Applied Research and
Practice on Municipal and Industrial
Waste. University of Wisconsin-Extension,
432 N. Lake St., Madison, WI 53706,
p. 238-249.
Dao, T. H., and T. L. Lavy. 1978. Atrazine
adsorption on soil as influenced by
temperature, moisture content and
electrolyte concentration. Weed Sci.
26:303-308.
Frost, R. R., and R. A. Griffin. 1977.
Effect of pH on adsorption of arsenic
and selenium from landfill leachate by
clay minerals. Soil Sci. Soc. Amer. J.
41:53-57.
Green, R. E., and S. R. Obien. 1969.
Herbicide equilibrium in soils in re-
lation to soil water content. Weed Sci.
17:514-519.
Griffin, R. A., and A. K. Au. 1977. Lead
adsorption by montmorillonite using a
competitive Langmuir equation. Soil Sci.
Soc. Amer. J. 41:880-882.
Griffin, R. A., R. Clark, M. Lee, and E.
Chian. 1978. Disposal and removal of
polychlorinated biphenyls in soils. U. S.
Environmental Protection Agency,
Cincinnati, OH 45268. EPA-600/9-78-016,
p. 169-181.
Griffin, R. A., R. R. Frost, and N. F.
Shimp. 1976. Effect of pH on removal of
heavy metals from leachates by clay
minerals. U. S. Environmental Protection
Agency, Cincinnati, OH 45268. EPA-600/9-
76-015, p. 259-268.
Griffin, R. A., and N. F. Shimp. 1976.
Effect of pH on exchange-adsorption or
precipitation of lead from landfill
leachates by,clay minerals. Envir. Sci.
& Tech. 10:1256-1261.
Grover, R. and R. J. Hance. 1970. Effect
of ratio of soil to water on adsorption
of linuron and atrazine. Soil Sci.
109:136-138.
Hassett, J. J., J. C. Means, W. L. Banwart
and S. G. Wood. 1980. Sorption properties
of sediments and energy-related pol-
lutants. U. S. Environmental Protection
Agency, Athens, GA 30605. EPA-600/3-80-
041.
Hassett, J. J., W. C. Banwart, and R. A.
Griffin. 1983. Correlation of compound
properties with sorption characteristics
of nonpolar organic compounds by soils
and sediments: Concepts and limitations.
In "The Environment and Solid Wastes:
Characterization, Treatment, and Dis-
posal," C. W. Francis and S. I. Auerbach,
Eds., Butterworth Publishers, Woburn, MA,
p. 161-176.
Barter, R. D. 1979. Adsorption of copper
and lead by Ap and B2 horizons of several
northeastern United States soils. Soil
Sci. Soc. Amer. J. 43:679-683.
Hope, G. D., and J. K. Syers. 1976. Effects
•of solution:soil ratio on phosphate sorp-
tion by soils. J. Soil Sci. 27:301-306.
Koskinen, W. C., and H. H. Cheng. 1983.
Effects of experimental variables on 2,4,
5-T adsorption-desorption in soil. J.
Environ. Qual. 12:325-330.
Nearpass, D. C. 1967. Effect of the
predominating cation on the adsorption
of simazine and atrazine by Bayboro clay
soil. Soil Sci. 103:77-182.
Sibbesen, E. 1981. Some new equations to
describe phosphate sorption by soils.
J. Soil Sci. 32:67-74.
White, R. E. 1966. Studies on the phosphate
potentials of soils. IV. The mechanism of
the 'soil/solution ratio effect'. Aust.
J. Soil Res. 4:77-85.
161
-------
ORGANIC FLUID EFFECTS ON THE STRUCTURAL STABILITY OF COMPACTED KAOLINITE
Yalcin B. Acar, Ivan Olivier!, Stephen D. Field
Department of Civil Engineering, Louisiana State University
Baton Rouge, Louisiana 70803 -
ABSTRACT
The effects of small particle sizes and the surface force interactions in fine
grained soils are manifested by a variety of interparticle attractive and repulsive
forces which strongly influence the macroscopic engineering behavior and structural
stability of compacted soils. This paper presents the results of studies to explain the
post-constructional behavior of compacted soils in terms of the properties of the indi-
vidual components of the system. Since organic leachates are of immediate concern, the
static dielectric properties of the pore fluid and the activity of the soil are empha-
sized.
The results of permeability and pore size distribution studies indicate, that
changes in hydraulic conductivity with organic fluids are due to variations in the sur-
face forces of interaction. These changes in the interaction forces are well reflected
with changes in the liquid limit and free swell of kaolinite. It is suggested that such
simple classification tests with the leachate be used to obtain a qualitative estimation
of the expected changes in hydraulic conductivity of a soil.
INTRODUCTION
Containment of wastes with different
organic and inorganic constituents in
shallow land waste disposal facilities
has stimulated the study of leachate
effects on the engineering properties of
compacted soil liners. Presently, highly
active (activity is defined as the plas-
ticity index divided by the percent finer
than -2 pm by weight) soils are chosen in
the design of these liners. Although
such a choice would render lower permea-
bilities (12) , post-constructional
changes in the chemistry of the pore
fluid would alter the physico-chemical
surface interactions of the fine grained
fraction with high surface charge den-
sity, leading to changes in the engineer-
Ing behavior.- It then becomes necessary
to assess whether these changes and the
structural stability of the compacted
soil could be estimated from a knowledge
of the properties of the soil and the
pore fluid.
This article presents the results of
the ongoing studies on the volume change
behavior and the permeability of com-
pacted clay with reference to the present
state of knowledge on the effect of vari-
ations in the forces of interaction
between the minerals and the pore fluid
to the engineering characteristics.
Since a majority of leachates con-
tain a wide variety of organic constit-
uents and recent studies indicated dra-
matic increases in hydraulic conductivity
of compacted soils (3,6) with organic
fluids, static dielectric properties of
the pore fluid and the activity of the
soil are emphasized.
FORCES OF INTERACTION IN FINE GRAINED
SOILS
It is well established that both
attractive and repulsive electrical
forces exist between clay particles.
Repulsive forces are primarily attributed
to the interaction between diffuse double
layers existing on the surfaces (10,15).
The system variables that control these
forces are defined in the development of
the Gouy-Chapman theory of diffuse double
layers. An approximate quantitative
indication of the thickness of the dif-
fuse double layer is given by (13),
162
-------
H -
DT
\2 2J
h v
o
(1)
where H = relative thickness of double
layer, D = dielectric constant of the
medium, T = temperature, h = electrolyte
concentration, v = cation valence, and
n = a constant (s%). Hence, changes in
the thickness of diffuse double layer is
directly related to the forces of repul-
sion between the clay minerals.
The principal contribution to
attractive forces arise from the van der
Waal's forces. The van der Waal's forces
could either be ion-dipole or dipole-
dipole interactions. Other factors that
contribute to attractive forces are
Coulombic attractions between negative
surfaces and positive edges, cation
linkages and hydrogen bondings (10,11).
It should be noted that attractive forces
are insensitive to slight variations in
the chemistry of the pore fluid.
Moore and Mitchell (14) developed
analytical expressions for both electro-
static and electrodynamic forces of
interaction between particles. The
expressions developed indicate these
forces of interaction are functions of
the static dielectric constant of the
pore fluid. Figure 1 presents the
o
z <
2
cr
UJ
o
UJ
ATTRACTION___
40 6O 8O
DIELECTRIC CONSTANT
100
^REPULSION
Figure 1.
Forces of Interaction on Clay
Surfaces (14).
changes in these forces with variations
in the dielectric constant of the medium.
As it is observed the net forces of
interaction closely follow the Van der
Waal's attraction curve with minimum
attraction at approximately a static
dielectric constant of 4.0 which repre-
sents the dielectric constant of the clay
particles. In order to evaluate the
applicability of their theory, Moore and
Mitchell (14) conducted vane shear and
triaxial tests on resedimented Georgia
kaolinite after permeating miscible
organic fluids of variable dielectric.
constants. The sample volume x^as kept
constant during this permeation. The
relationships between the strength and
the dielectric constant confirmed the
hypothesis exhibiting shapes corres-
ponding to the net interaction curve in
Figure 1.
The above discussion indicates that
permeating an organic fluid through com-
pacted soils would change the forces of
repulsion and reestablish the net forces
of interaction in the fine grained frac-
tion leading to changes in the micro-
fabric and the macroscopic engineering
behavior. Since forces of repulsion are
also a function of the surface charge
density, minerals with a higher surface
charge density (smectitic minerals) will
be affected to a higher degree with these
changes. In order to assess and quantify
the changes in volume change behavior and
permeability of compacted soils, an
experimental procedure is implemented
where the surface charge density and
dielectric constant of the pore fluid
were taken as the variables.
EXPERIMENTAL PROCEDURE
Kaolinite and Ca-montmorillonite are
chosen for permeation studies. For index
properties and free swell tests Na-mont-
morillonite is also used. The mineralo-
gical composition, index properties and
compaction parameters of kaolinite,
Ca-montmorillonite and Na-montmorillonite
are presented in Table 1. Organic fluids
representing a wide range of dielectric
constants were used. Table 2 presents
the characteristics of various compounds
used in this study together with their
hazard classification and maximum
reported concentration in leachates
collected from waste disposal facilities
-(16).
163
-------
TABLE 1. COMPOSITION AND CHARACTERISTICS
OF CLAYS USED IN THIS STUDY
Characteristic
Ca-M
Na-M
Mineralogical Composition (% by weight)
Kaolinite
Illite
Ca-Montmorillonite
Na-Montmorillonite
Index Properties (%)
Liquid Limit
Plastic Limit
Plasticity Index
Specific Gravity
% Finer than 2 m size
Activity
Proctor Compaction Parameters
2
Maximum Dry Density (t/m )
Optimum Water Content (%)
98
2
92
100
64
34
30
2.65
90
0.32
88
54
34
2.70
12
2.8
425
58
367
2.70
80
4.5
1.37 1.15 1.15
31.0 25 25
K - Georgia Kaolinite
Ca-M - Ca-Montmorillonite
Na-M - Na-Montmorillonite
T - Trace Quantity
164
-------
y~^
rH
s^
CO
H
13
W
! 1
0
CO
0
M
55
co cd
rH
CO
V /
• ,— X
0 rH
X CJ --.
CO O 00
*s^
CO
CO
CO
rH
O
p-{
ft
g
«
>-
a.
to
to
rH
i
M
O
Fn
13
3
O
ft
B
0
o
1 8
1 1
1 1
O -*
r^» \o
CO
CO CM
. .
rH CM
CO rH
0\ rH
rH
O O
r-l rH
CM
VO
•%
4-1 .
CO 4-1
Is w
O
C O ft
CU | CU
PQ PM HI
I—I
vo
rH
i — i
C
0
•H
4-1
CO
O
•H
H |
•H
CO
CO
CO
rH
O
TS
M
CO
N
&
i
.
CO
CO
CO
rH
O
/-*N
O
o
O
CN
4-1
co
4J
CO
(3
o
O
O
•H
M
4-1
O
CU
rH
CU
•H
""O
1
CO
J-J
C
co
4->
rH
t__ |
0
f*t.
>,
4J
•H
O M
•H 0
!*! -H
0 f-i
•M Oi
i i
EH PM
TJ
p
O
ft
8
0
0
rH
rH
CO CO
3
o a
•o o
!-l -H
CO 4-1
CTJ CU
& CO
1 1
pa co
^— s
u
o
O
CM
x_ s
S-+*
CO
CJ
•"•^
a
\^s
4-1
£1
60
•H
CU
4J
•H
id
1
*"
/r-^
rH
••^^
oo
e
CO
CU
4-J
CO
O
^J
CU
rH
a
•H
T3
CU
4-1
I-l
0
CU
M
ta
•H
4J
CO
4J
a
CU
O
B
o >,
O 4J
•H
B rH
i -H
•H D
ctt O
B ^
1 1-4
CU
4J
0 CO
C !3
0
0 1
^ H 1
CO P3
s w
^^
o
O
C5
CM
y^X
CD ^-s
0) CO
w d)
'o ^
p^
-------
Permeabilitv Tests
Free Swell Tests
Permeability tests are conducted in
flexible wall permeameters. The results
are also compared with tests in rigid
wall permeameters. Mercury intrusion
tests were performed on the samples
before and after permeation with the
organic fluids. The details of the test
setup and testing procedure are presented
elsewhere (1,2). Presently, tests with
kaolinite are completed.
Atterberg Limits
Free swell tests is another method
of determining the repulsive forces
between the clay particles. The free
swell test is performed by pouring 10 cc
of dry soil, passing No. 40 sieve into a
100 cc graduate cylinder filled with the
liquid. The final swelled volume is
recorded after the suspension came to
rest. Free swell (F.S.) is given by
V, - V
F.S. = -2- x 100%
(2)
In order to determine the effect of
organic fluids on the liquid limit and
hence the repulsive forces in the clays,
Casagrande apparatus was first used. It
was then realized that this apparatus
gave inconsistent results with organic
fluids of low dielectric constants. This
was mainly due to the fact that as the
plasticity index of the clay mineral
decreased, the shear failure was
mobilized at the interface between the
soil and the brass cup resulting in
uncomparable liquid limit values. In
order to avoid this discrepancy, Swedish
fall cone tests were used (6).
where V, = soil volume after swelling and
V = soil volume before swelling. This
procedure is often used in the classifi-
cation of soils for swelling potential.
ANALYSIS OF RESULTS
The results of permeability tests
with four pure organic fluids and com-
pacted kaolinite are presented in
Figure 2. Dramatic increases experienced
in previous studies with rigid wall per-
meameters (3,6) were attributed to side
leakages (1). Tests with nitrobenzene
> 3.0
O
Z3
O
O
o
2.0
tc
o
X
UI
>
1.0
tu
tc
T
BREAKTHROUGH(50%)
PV « 1.6
T
EFFECTIVE STRESS. kPo
PHENOL_._
ACETONE
ACETONE
NITROBENZENE AND BENZENE
(O.OOI TO 0.003)
3456
PORE VOLUME, PV
9
Figure 2. Hydraulic Conductivity of Compacted Kaolinite to Pure Fluids (1) ,
166
-------
0"
lor' ' '" o" To'1
LIMITING PORE DIRHETER. MICRONS
ItT
Figure 3.
lor8 ID"' '•"Iff ' ' ?cr
LIMITING PORE DIHMETER. MICRONS
Pore Size Distributions of
Compacted Koalinite Before and
After Permeation with Benzene.
and benzene resulted in three orders of
magnitude decrease in relative permea-
bilities, while acetone and phenol result
in up to one-fold increase. Table 3 com-
pares the final mean relative and abso-
lute permeabilities with pure organic
solvents and the 0.01 N CaSO, solution.
Figure 3 compares the pore size
distribution of compacted samples before
and after permeation with benzene. Cumu-
lative porosity distributions indicate
that major changes are not observed. The
mercury intrusion method scans pores of
80 Ac or greater and the modal pore dia-
meter for these samples is 600 A°. It is
then expected that the majority of flow
is through pores of 600 A°.
Figure 4 plots the pore size para-
meters obtained from the distribution
analysis with respect to absolute permea-
bility (4). It is observed that pore
size parameters do not respond to varia-
tions in absolute permeability. Conse-
quently, two explanations could be given
for the change in hydraulic conductivity;
plugging or unplugging of pores due to
particle migrations or unusual properties
of the pore fluid close to clay surfaces.
Non-Darcian type of flow through
compacted soils have previously been
demonstrated to be caused by particle
E
=L
cvj
< -o
UJ
Q
o
S
COMPACTED SILTY CLAY
Water (Bengochea and Lovell, 1981)
COMPACTED KAOLINITE
.OIN CaS04
Nitrobenzene
Acetone
Phenol
Benzene
-------
TABLE 3. A COMPARISON OF RELATIVE AND ABSOLUTE PERMEABILITY
OF COMPACTED KAOLINITE (1)
Permeability
Relative
Permeant _
x 10 cm/sec
k.* k,, K.
i f i
Nitrobenzene (1) 5.60 0.025 560
(2) 5.20 0.015 520
Acetone (1) 5.60 6.00 560
(2)** 5.00 11.00 500
Phenol 6.00 14.00 600
Benzene 5.10 0.010 510
Absolute
.--15 2
x 10 cm
Kf Kf/K.
1.51 2.7 x 10~3
0.91 1.8 x 10~3
1470 2.63
2690 5.38
119.0 0.20
1.38 2.7 x 10~3
&
Initial permeabilities with 0.01 N CaSO,
**Test at effective stress of 13.8 kPa (2 psi)
K = kp/Y
Y = unit weight of permeant
p = viscosity of permeant
168
-------
migrations (8). It is also possible to
initiate such a mechanism by changes in
the forces of interaction. The increase
in dispersive forces might tend to dis-
locate particles. There exists contra-
dicting information on whether the
unusual pore fluid properties around clay
particles effect the fluid flow through
fine grained soils (13). However, if the
properties of the pore fluid in the dif-
fuse layer around the clay particles are
different than that of water, the changes
in the thickness of this layer will
result in variations in the' effective
pore space contributing to the total
flow.
Since the extent of repulsive forces
is related to the thickness of the dif-
fuse double layer and the liquid limit
values and free swell tests present a
good estimate of this thickness (12),
such tests were conducted on the clays
with the organic fluids used in this
study. As presented in Figures 5 and 6,
dielectric constant correlates with both
the liquid limit and free swell of the
smectitic minerals. However, kaolinite
did not follow this trend; the liquid
limit and free swell is increased with
nitrobenzene, and benzene slightly
increased with phenol and decreased with
acetone. Such observations are possible
in low active kaolinite minerals since
the unaccounted charges at the broken
edges of this mineral have a major effect
on the forces of interaction.
If the liquid limit and free swell
of a. soil with any pore fluid increases
when compared to that of water, this
implies a corresponding increase in the
thickness of the diffuse layer. Consi-
dering that this thickness might extend
to 400 A° with water and that the modal
pore diameter of 600 A° and the frequency
are not changed with organic fluid per-
meation indicate that the variations in
the forces of interaction on the clay
surface would be the major factor affect-
ing the flow. In fact the changes in
liquid limit and free swell are consis-
tent with absolute permeability values
presented in Table 3. An increase in the
liquid limit and free swell with nitro-
benzene and benzene shows that repulsive
forces and the thickness of diffuse layer
is increased resulting in a. decrease in
permeability (1). The reverse mechanism
would be true for acetone.
If the results of the permeability
and pore size distribution tests with
Ca-montmorillonite are consistent with
the conclusions reached using kaolinite
as reported herein, the case for using
simple classification tests with a
specific leachate to provide an initial
qualitative estimate of changes to be
expected in hydraulic conductivity will
be greatly strengthened.
CONCLUSIONS
The present results of ongoing
studies are summarized as:
1. Organic fluid permeation does
not affect the size and distri-
bution of pores in kaolinite
samples compacted at the wet of
optimum water content.
2. Three order of magnitude change
in hydraulic conductivity at
the same pore size distribution
indicate that these changes are
initiated by changes in the sur-
face forces of interaction on
clay surfaces.
3. The liquid limit and free swell
of sodium or calcium-saturated
montmorillonite appear to be
strongly related to the dielec-
tric constant of pure organic
fluids. These data indicate
that forces of repulsion in
this mineral vary with the
dielectric constant of the
liquid medium.
4. The direction of changes in
liquid limit and free swell of
kaolinite with organic fluids
is well reflected by the varia-
tions in hydraulic conductivity
of compacted samples.
ACKNOWLEDGEMENT
The information presented in this
study is funded by the United States
Environmental Protection Agency under
assistance agreement EPA Cooperative
Agreement CR 809714010 to the Hazardous
Waste Research Center, Louisiana State
University. However, it does not
necessarily reflect the views of the
Agency and no official endorsement is
inferred.
169
-------
IOOO
500
g
^100
H"
§ 50
O
o
-1
10
100
80
60
I 1 ' 1 1 I 1 1 l I 1 1 1 — | — i i i i
I i WYOMING BENTONITE (KINSKY, ET AL 1974) ;
• No-MONTMORILLONITE A = 4.5_, -
0 Co-MONTMORILLONITE yt -
A=ACTIVITY /& -
~ ^^' — — -*~~-
' ^--^'^ ~.£-"r"~~^'~~' ~~
•$^~-— — -—"6 -_
'
'
Phenol Nitrobenzene
B"jene 1 Acetone Woter
Lt 1 1 1 1 1 1 1 1 1 1 l ii 1 1 . ill
1 I — 1 — 1 1 1 1 1 1 1 1 — 1 — 1 1 1 1 1 1
- • GEORGIA KAOLIN ITE • -
~ m A= 0.32-1 —
1 1 1 1 1 1 1 1 1 * i , 1 i . i i~
REFERENCES
5 10 50
DIELECTRIC CONSTANT
Figure 5. The Effect of Organic Fluids
on the Index Properties (1).
IOOO
50O
UJ
(/I
UJ
u
a:
IOOO
BOO
600
• No- MONTUORILLONITE
OCo- MONTUORILLONITE
A«ACTIVITY
• A=2.8-
Hiplont Xyllnl
P-DioiccK Bmwi
Anlllnl
.III. ...I... I
Acitoni E.GIyeol
phtnol
Nfrobtn-
1 - 1 — i — i i i 1 1
GEORGIA KAOLINITE
,,
5 10 50
DIELECTRIC CONSTANT
Figure 6. The Effect of Organics Fluids
on Free Swell of Clays (1).
Acar, Y., Hamidon, A., Field, S. D.,
and Scott, L. (1984), "Organic
Leachate Effects of Hydraulic Con-
ductivity of Compacted Kaolinite,"
Submitted for publication at the
Symposium on Impermeable Barriers
for Soil and Rock, ASTM, Boulder,
CO.
Acar, Y., Olivieri, I., and Field,
S. (1984), "Organic Leachate Effects
on the Pore Size and Frequency Dis-
tribution of Compacted Kaolinite,"
Submitted for publication at the
Symposium on Impermeable Barriers on
Soil and Rock, ASTM, Boulder, CO.
Anderson, D., and Brown, K. (1981),
"Organic Leachate Effects on the
Permeability of Clay Liners," In:
D. W. Schultz (ed.), Land Disposal:
Hazardous Wastes, Seventh Annual
Research Symposium, EPA-600/9-81-
0026, USEPA, Cincinnati, OH 45268,
pp. 400.
Bengochea, J. G., and Lovell, C. W.
(1981), "Correlative Measurements of
Pore Size Distribution and Permea-
bility in Soils," ASTM Special Tech-
nical Publication 746, pp. 137-150.
Green, W. J., Lee, G. F., Jones, R.
A., and Pallt, T. (1983), "Interac-
tion of Clay Soils With Water and
Organic Solvents: Implications for
the Disposal of Hazardous Wastes,"
Environ. Sci. Technol., V. 17,
pp. 278-282.
Green, W. J., Lee, G. F., and Jones,
R. A. (1981), "Clay-Soils Permea-
bility and Hazardous Waste Storage,"
J. Water Pollut. Control Fed., V. 53,
pp. 1347-1354.
Hansbo, S. (1957), "A New Approach
to the Determination of the Shear
Strength of Clay by the Fall Cone
Test," Royal Swedish Institute,
Proceedings, No. 14.
Hardcastle, J. H., and Mitchell, J.
K. (1974) , "Electrolyte Concentra-
tion-Permeability Relationships in
Sodium Illite-Silt Mixtures," Clays
and Clay Minerals, Vol. 22,
pp. 143-154.
170
-------
9. Kinsky, J., Frydman, S., and Zaslav-
sky, D. (1974), "The Effect of
Different Dielectric Liquids on the
Engineering Properties of Clay,"
Proceedings of the 4th Asian Con-
ference on Soil Mechanics and Foun-
dation Engineering.
10. Lambe, T. W. (1954), "The Permea-
bility of Fine-Grained Soils," ASTM
Special Technical Publication 163,
pp. 56-67.
11. Lambe, T. W. (1960), "A Mechanistic
Picture of the Shear Strength of
Cohesive Soils," Proceedings of the
Conference on Shear Strength of
Soils," Boulder, CO.
12. Lamb, T. W., and Whitman, R. V.
(1969), Soil Mechanics, John Wiley
and Sons, Inc., New York.
13. Mitchell, J. K. (1976, Fundamentals
of Soil Behavior, John Wiley and
Sons, New York, 422 p.
14. Moore, C. A., and Mitchell, J. K.
(1974), "Electromagnetic Forces and
Soil Strength," Vol. 24, No. 4, pp.
627-640.
15. Seed, H. B., Mitchell, J. K. , and
Chan, C. K. (1960), ''The-Strength of
Compacted Cohesive Soils," Pro-
ceedings of the Conference on Shear
Strength of Soils, Boulder, CO,
pp. 877-961.
16. Shuckrow, A. L., Pajak, A. P., and
Touhill, C. J. (1980), "Management
of Hazardous Waste Leachate," USEPA,
Office of Water Waste Management,
Report No. SW871.
171
-------
AIR EMISSIONS OF VOLATILE ORGANIC
CHEMICALS FROM LANDFILLS: A PILOT-SCALE STUDY
Louis J. Thibodeaux, Charles Springer and Guadalupe Hildebrand
College of Engineering, University of Arkansas
Fayetteville, Arkansas 72701
ABSTRACT
Through a combination of vapor phase chemical transport models and a pilot-scale
simulator, volatile organic chemical emission through soils were examined. An experimen-
tal apparatus was built to simulate a landfill cell containing a liquid chemical in pure
form. Models were used to determine the flux of the volatile chemical under various con-
ditions. Two soil cover depths were investigated and'carbon dioxide was used as a land-
fill gas purge. The landfill gas purge and cover depths had varying effects on the
ability of the volatile chemical to diffuse through the soil layer, dependent upon the
volatile chemical used. This paper discusses the changes of chemical flux rates of
selected chemicals and compares the model prediction to experimental results.
INTRODUCTION
Burial pits and landfills have been
classified as near-surface impoundments.
Such land depositories have been and will
undoubtly continue to be disposal sites
for solid and chemical wastes. Hazardous
substances are known to have escaped the
sites in liquid form and as water leachate
and these emissions have been studied
extensively with the result that a broad
base of technical knowledge exists on this
subject. However, until recently, aspects
of vapor-phase emissions have been given
little study.
The loss of chemicals to air via
vapor-phase emission from landfills and
land burial has been documented. Two
noteworthy incidents are: the volatiliza-
tion and subsequent transport by moving
air currents of hexachlorobenzene (HCB)
about the environment in South Louisiana (1]
and the much publicized Love Canal area of
Niagra Falls, NY, where analysis of air
samples nearby the landfill site reveal the
presence of at least 26 volatile organic
compounds (3).
Purpose
The goal of this research project is to
provide verified methods of estimating vola-
tile chemical emissions into air from a
variety of near-surface impoundments. The
emphasis of this project is on landfills and
similar subterranean structures. This goal
can be achieved by a combination of model
development, laboratory simulation experi-
ments and field monitoring.
172
-------
Approach
The present approach to the research
project is along two parallel lines of
development. One line is the procurement
of data that reflect VOC emissions in a
realistic landfill-type environment. This
is being accomplished by use of a pilot-
scale simulator. The other line of
investigation is to review, evaluate, and
develop vapor phase chemical transport
models for the same landfill-type
environment.
Pilot-scale simulator. An experimen-
tal apparatus as shown in Figure 1 was
built to simulate the idealized single
cell landfill containing a chemical
(liquid) in pure form. The rectangular
box is made of galvanized sheet metal, has
an open top and dimensions of 45.7 cm (18
in.) x 45.7 cm (18 in.) x 91.4 cm (36
in.). A soil support structure consisting
of a metal frame, iron grating, screen and
cheesecloth, is in-place through the top
to retain the soil-clay test section. The
metal frame can be raised or lowered to
give various soil depths. The,bottom part
of the apparatus contains a recording pan-
evaporator. Other elements of the simula-
tor include a small vapor circulating fan,
thermocouples, gas inlet and distributor
pipe, gas sample ports and a transparent
access door. Another larger fan keeps the
top soil surface vapor free. A silicon
rubber cement is used around the metal
frame, door and other entry points to
assure that all vapors and decomposition
gas must exit through the soil layer.
A simulation run is performed by
placing a quantity of pure liquid test
chemical in the pan of the evaporator. If
decomposition gas purge is also being
studied carbon dioxide is fed in the bot-
tom. The entire simulator is placed in a
constant temperature and humidity room.
The room is vented with an exhaust fan to
avoid the build up a chemical vapor and
decomposition gas. Weight remaining in the
pan is continuously recorded with time.
The emission flux rate is computed from
the weight loss rate and the area of the
soil surface 2090 cm2 (2.25 ft2). Gas
samples are withdrawn from the chamber for
analysis by chromatography.
The soil presently being used has the
general characteristics shown in Table 1.
Three chemical are presently being
tested in the simulator. They are methanol
(MeOh), methylcyclohexane (Mch) and
1,2-dichloropropane (Dcp). Table 2 sum-
marizes properties of these chemicals.
Reagent grade chemical are used in all
tests.
Gas-phase transport models. Several
transport models with specific application
to emission of chemical vapors from land-
fill sources have appeared in the recent
literature. Farmer, Yang and Letey (1)
extend Fick's first law for the diffusion
of chemical through a porous media. The
effective diffusion coefficient is a func-
tion of the molecular diffusivity of the
chemical, the air filled porosity and the
total soil porosity. The model is tested
with hexachlorobenzene waste.
Thibodeaux (4) extends the Fickian
transport model to include the effect of
decomposition gas flowing through the cap
to enhance the vapor transport.
Mathematical simulations with benzene,
chloroform, vinyl chloride and PCB suggest
that bio-gas purge can enhance the volatile
chemical emission rate by a factor of
seven. An effective diffusion coefficient
based on field measurements of radon
(222Rn) and Farmers, et. al. hexach-
lorobenzene data was used to establish the
general creditability of the model.
The model was extended (5) to consider
the dynamics of the landfill environment.
Landfill decomposition gas dynamics and
flow were simulated to account for baro-
metric pressure effects on emission. Based
on in-situ gas permeability measurements in
a waste trench containing tritium, seasonal
cap permeabilities and seasonal emission
rates of benzene were computed. The fluc-
tuation of the barometric pressure was
found to increase the emission rate by
as compared to constant pressure con-
ditions. A sensativity analysis of the
effect of season, cover thickness, cell
porosity, cell depth, and decomposition gas
generation rate on emission of benzene was
also performed.
173
-------
(O
S-
o.
a.
S-
o
ra
Zi
•o
£T
ro
a>
i.
174
-------
Murphy (2) presented a model for
volatiles diffusing through the bulk
waste, through the soil and then into the
lower atmosphere boundary layer. A time
dependent rate equation is presented
inasmuch as the flux into the atmosphere
decreases the amount of vdlatile chemical
in the waste and increases the transport
distance. The model equations are not used"
to make any specific prediction of emission
rates.
Table 1. Soil Characteristics*
Sand Percent
Silt Percent
Clay Percent
Carbon Content
Porosity
Water Content
Bulk Density
30.4
47.3
22.3
1.42%
0.43
2.43%
1.29 g/cm3
*Analysis performed by University of Arkansas Soil Characterization Laboratory.
Table 2. Chemical Properties
Chemical
Molecular Weight Density (g/cm3)
Vapor 25-C Water*
Pressure (mmHg) Solubility
Methanol
MethyIcyclohexane
1,2-Dichloropropane
32.04 dls=0.79609 125.03
d25=0.78675
98.19 d20=0.76936 46.36
d2S=0.76506
112.99 dzo=1.5600 50.33
d2s=1.159
Used d = 1.16 for
flux calculations
* °°-soluble in all proportions
i-insoluble
6-slightly soluble (1,2-DCP soluble 2,700 mg/1 at 20-C)
175
-------
Continuing study of the diffusion
process in light of the volatile materials
used in this study suggests that the
following equation is most appropriate
for interpreting the experimental results.
b. f\ _ PA/PT-EXP[-(NA/MA + VP-r/RT) LRT/PTD]f
^ J
(1)
where NA = flux of volatile chemical, g/cm2-s,
V = apparent decomposition gas velocity, crnVs-cm2,
PA = vapor pressure of the volatile, atm.
M/\ .= molecular weight of the volatile, g/mol,
R = gas constant 82.1 atm-mol/cm3-K,
T = temperature of the gas-vapor mixture, K,
Pf = atmospheric pressure, atm,
L = thickness of the soil layer, cm, and
D = effective diffusion coefficient, cm2/s.
It should be noted that NA also appears on the right side of the Eq. 1. This is a
necessary inclussion for chemicals of high vapor pressure in the range P. > 0>05 P An
iterative scheme is required for solution of EQ. 1. For the case of no decomposition gas
flowing through soil (i.e., V=0) the appropriate equation is:
NA = (D PT MA/RTLjlnCl/1 -
(2)
The equation can also be used to commence the iterative solution of Eq. 1.
Probelms encountered. Early experi-
ments resulted in extremely high apparent
emission rates. Values of up to 4500
ng/cm2*s for Dcp, 4100 ng/cm2-s for Mch
and 3900 ng/cmz-s for Meoh were obtained
through a 38.1 cm (15 in.) soil layer.
Adsorption calculations suggested that the
soil could be retaining quantities of the
chemicals. If so this would give high
apparent emission rates. It was therefore
necessary to perform additional experi-
ments to determine the extent to which
adsorption occurs.
Three desiccators were used to per-
form chemical vapor adsorption experiments
with the soil. Quantities of varying
amounts of oven-dried soil were placed in
25 ml. vials. Into each desiccator was
placed a sufficient amount of liquid che-
mical. A perforated porcelain plate was
placed over the chemicl pool and the vials
containing the soil were placed on top of
the plate. The lids on the desiccators
were sealed in-place, and the soil vials
were allowed to absorb chemical vapors for
approximately sixty days.
The soil samples were re-weighed and
the weight change due to adsorption noted.
The adsorption results were: Meoh = 0.0469
+ .0016 g/g-soil; Dcp = 0.0649 +. .0026
g/g-soil; and Mch = 0.0440 _+ .0030
g/g-soil. These adsorption results indi-
cated that it would be necessary to pre-
treat the soil layer with the respective
chemical prior to performing emission simu-
lations. The soil had a mass of 95.5 kg
for a 38.1 cm layer and this amounted to
4.5 kg Meoh, 6.20 kg Dcp and 4.2 kg Mch
sorbed.
Elaborate means of soil pre-treatment
were undertaken prior to further simulation
experiments. This included covering the
soil surface, observing chemical and gas-
vapor temperatures inside the simulator and
recording the total quantity absorbed.
Soil pre-treatment took up to 9 days.
176
-------
A wet-test meter was used as an in-
line decomposition gas flow measuring
device. Moisture introduced through the
meter likely interfered with some of the
flux results. Water also adsorbs onto the
soil and hampers chemical vapor transport.
Water is soluble in Meoh. This practice
was stopped and a Dri-rite filters was
placed in the decomposition gas inlet
lines to adsorb water vapor.
Results
During this reporting period 36 flux
measurements have been completed. Three
chemicals: Meoh, Dcp and Mch, and two
soil layer depthst 38.1 cm and 7.62 cm
have been tested. Table 3 presents the
operating conditions and flux results
prior to installation of the Dri-rite
filter for the 38.1 cm soil depth.
liquid giving faulty readings. The experi-
ments were repeated after installation of a
moisture trap and these measurement results
appear in Table 4.
The new methanol results are still
inconsistent. The flux with and without
decomposition gas fluctuate highly. The
Dcp data show a slight increase (i.e. 6.3%)
with decomposition gas purge. The Mch data
show a substantial increase (i.e., 31%).
Further experiments were performed in a
room with temperature and humidity control
features. In addition 250 ml. aliquots of
the gas from the simulator were periodically
withdrawn with an evacuated gas sampling
bulb and analysis performed on a Tracor 565
gas chromatograph with a flame ionization
detector. The soil depth was reduced to
7.62 cm (3 in.). The results of the latest
flux experiments appear in Table 5.
Table 3: Chemical Fluxes Through Soil*
Chemical Run No.
gas
(*C)
Gas Rate cm3/s Flux ng/cm2-3
Meoh
Meoh
Meoh
Meoh
Dcp
Dcp
Dcp
Dcp
1
2
3
4
5
6
7
8
23.4
20.4
19.9
21.7
20.0
20.0
22.0
21.0
19.9
17.8
17.9
19.7
19.0
• 19.2
21.0
20.9
0.0
0.0
2.05
2.05
0.0
0.0
2.03
2.03
525
472
444
472
1445
1651
1541
1581
*Soil Depth: ~ 38.1 cm, w/o Dri-Rite
As is evident from the results in
Table 3, there is essentially no dif-
ference between the flux with and without
decomposition gas flow. The introduction
of high moisture levels appears to
have been influencing the process. Water
can adsorb on the soil and into the pan
Gas analysis results, available only on
runs 21 through 33 indicate that the gas
below the soil bottom surface in the simu-
lator is not saturated with chemical vapor.
Comparison with the pure component vapor
pressure at the temperature of the liquid
indicate the partial pressure of Meoh is
177
-------
Table 4. Chemical Fluxes Through Soil*
Chemical
Meoh
Meoh
Meoh
Meoh
Dcp
Dcp
Dcp
Dcp
Mch
Mch
Mch
Mch
Run No.
9
10
11
12
13
14
15
16
17
18
19
20
Tgas (*C)
23.8
25.5
24.5
23.8
22.3
23.1
23.4
26.0
24.4
23.6
23.6
25.0
Tliquid('C)
20.8
22.5
22.0
20.8
21.2
21.8
22.0
24.5
22.5
21.7
21.6
23.1
Gas Rate cm/s
0.0
0.0
1.90
1.90
0.0
0.0
1.90
1.90
0.0
0.0
1.90
1.90
Flux ng/cm2-s
1046
923
1328
701
2240
2340
2470
2400
1146
1167
1556
1474
*Dri-Rite Filter in Place, Soil Depth: 38.1 cm
Table 5: Chemical Fluxes Through Soil*
Chemical
Meoh
Meoh
Meoh
Meoh
Meoh
Meoh
Meoh
Meoh
Meoh
Meoh
Dcp
Dcp
Dcp
Dcp
Dcp
Dcp
Run No.
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Tgas C-C)
17.6
17.4
17.6
17.8
17.2
17.4
17.8
17.2
16.8
16.5
17.4
17.5
16.8
17.0
17.5
17.3
TliquidC'C
12.5
12.0
12.5
13.2
13.4
13.6
13.0
12.8
13.4
12.7
15.7
15.9
15.3
15.7
16.1
15.6
) Gas Rate cm/s
0.0
0.0
0.0
0.0
0.0
0.0
0.0
5.70
5.70
5.70
0.0
0.0
0.0
0.0
0.0
0.0
Flux ng/cm2-s
909
911
910
984
978
883
895
1192
1195
1203
1432
1436
1436
1441
1444
1453
*Dri-Rite Filter in Place, Soil Depth 7.62 cm
51.6% (cr=6.666%3 saturated for no decom-
position gas and 66.1% (cy=2.76%) saturated
with gas. Dcp with no gas was 61.6%
(cr=3.55%) of saturation. Measurements
with Mch are incomplete at this time.
The chemical flux rates of Meoh appear
to be significantly enhanced with purge gas
flow. The data in Table 5 indicate an
average of 29% increase in the flux.
Temperature variations preclude a com-
178
-------
parison of the effect of decreased soil
depth on flux in Table 5 from that in
Tables 3 and 4.
The extraction of the effective dif-
fusion coefficients from the raw data pro-
vides an additional parameter to interpret
the results. These values also provide a
means of translating the laboratory data
to field conditions. This has been per-
formed by solving Eq. 1 for D and using
the laboratory data for N/\, P/\, L, etc.
The effective diffusion coefficients
appear in Table 6.
The mean values for methanol are .253
(a=.105)cm2/s for the 38.1 cm. soil depth
and .115(0=.0169)cm2/s for the 7.62 cm.
soil depth. The dichloropropane values
are .445(a=.0807) and .085(a=.00488)cm2/s
respectively. The methylcyclohexane
apparent diffusion coefficient for the
38.1cm soil is .349(a=.0462)cm Vs. In
general these deep soil values are higher
than the molecular vapor diffusivities
(i.e., .051 to .17 cm2/s). The 7.62 cm
soil depth apparent diffusion coefficients
are in general slightly less than the mole-
cular vapor diffusivities. Diffusivities
were estimated using the Wilke-Lee modifi-
cation of the Hirschfelder-Bird-Spotz
method (6).
The results presented here are based on
limited data so no conclusions will be
drawn at this point.
Acknowledgement
Mr. Stephen C. James is the Project
Officer for this research endeavor. His
continuing interest in this subject and his
technical input create a stimulating
atmosphere for inquiry. Vicki Welytok and
Phil Lunney are acknowledged for helping
with the chemical analysis of the gas
samples.
Table 6. Effective Diffusion Coefficients*
Meoh
Run No. DCcmVs)
Dcp
Run No. D(cm2/s)
Mch*
Run No. D(cmVs)
1
2
3
4
9
10
11
12
21
22
23
24
25
26
27
28
29
30
.229
.230
.147
.142
.435
.351
.382
.207
.142
.146
.116
.107
.119
.0892
.112
.113
.0982
.104
5
6
7
8
13
14
15
16
31
32
33
34
35
36
.371
.420
.342
.352
.520
.558
.535
.462
.0923
.0861
.0855
.0886
.0776
.0805
17 .345
18 .259
19 .463
20 ,407
* Top half of table for 38.1 cm soil, lower for 7.62 cm soil depth
* Assumed 60% vapor saturation.
179
-------
Literature Cited
1. Farmer, W. J., M. S. Yang, and J. Letey, "Land Disposal of Hexachlorobenzene
Waste-Controlling Vapor Movement in Soil" EPA-600/2-80-11, August, 1980.
2. Murphy, B. L., "Air Modeling and Monitoring for Site Excavation", Proceeding Nat.
Conf. Management Uncontrolled Hazardous Waste Sites, Nov. 29-Dec. 1, 1982,
Washington, D.C.
3. Shen, E. T. and T. J. Tofflemire, "Air Pollution Aspects of Land Disposal of Toxic
Waste", Journal of Environmental Engineering, ASCE, Vol. 106, No. 5, pp. 211-225.
Feb. 1980.
4. Thibodeaux, L. J., "Estimating the Air Emissions of Chemicals from Hazardous Waste
Landfills", Jo. Haz. Mtl. 4 (1981) 235-244.
5. Thibodeaux, L. J., C. Springer and L. Riley, "Models of Mechanisms for the Vapor Phase
Emission of Hazardous Chemicals from Landfills", Jo. Haz. Mtl., 7 (1982) 63-74.
6. Treybal, R. E., Mass Transfer Operation. 3rd Ed., McGraw-Hill, N.Y. (1980), pp. 31.
180
-------
ASSESSMENT OF SURFACE IMPOUNDMENT TECHNOLOGY
By
Masood Ghassemi, Michael Haro, and. Linda Fargo
MEESA
18811 Crenshaw Place, Suite 2O1
Torrance, CA 9O5O4
ABSTRACT
Case studies of hazardous waste surface impoundments (SI) comparing
actual and projected performance can contribute significantly to improving
their design, construction, and operation. To this end, the available
data for 22 Si's in the southwestern United States were reviewed, and six
of them were assessed further for a case study. The six Si's represent a
range of industries, waste types and volumes, ages, environmental settings,
liner types and designs, and systems for leak detection and groundwater
monitoring.
Poor performance of several ponds studied was attributed to lack of
good project planning during the design and construction phase. Failure
to execute proper quality assurance and control (QA/QC) procedures during
the project also contributed to failure of these ponds. Deviations from
desired liner specifications and poor liner-waste compatibility testing
resulted in poor performance. At at least one site studied, poor impound-
ment performance resulted from the lack of rigorous site-specific investi-
gations to develop the basis for proper design and construction.
The successful performance of Si's at one facility is attributable to
(1) use of a very impermeable clay as liner material after extensive waste-
liner permeability studies, (2) use of competent design, construction, and
inspection contractors, (3) close scrutiny of all phases of design,
construction, and QA inspection by the owner/operator, (4) excellent QA/QC
and record keeping during all phases of the project, and (5) good communica-
tion between all parties involved in establishing the Si's.
181
-------
INTRODUCTION
The Solid and Hazardous Waste
Research Division (SHWRD) of EPA's
Municipal Environmental Research
Laboratory (MERL, Cincinnati, OH) is
investigating the use of surface
impoundments (Si's) for hazardous
waste management. The primary aims
of this research are to develop
criteria for improved SI design and
operation and to provide technical
support to EPA's Office of Solid
Waste in developing SI regulations
according to the provisions of the
Resource Conservation and Recovery
Act (RCRA).
Currently, the principal source
of EPA information on hazardous
waste surface impoundments is the
Surface Impoundment Assessment (SIA)
data base developed by the Office of
Drinking Water. Though the data
base contains some background survey
information on the numbers, types,
and uses of wastes handled by muni-
cipal, industrial, and agricultural
Si's in the United States, it con-
tains little or no information on
their engineering design and actual
performance. Comparison of pro-
jected and actual performances at
operating Si's and identification of
reasons for observed differences are
essential for designing better new
Si's and for formulating approp-
riate corrective actions at existing
sites.
PURPOSE
The purpose of the present
study is to develop a data base on
hazardous waste surface impoundment
design and operating practices and
to compare actual and projected
performance for a selected number of
facilities. The study will identify
gaps in the existing data base and
areas that most warrant research and
development.
APPROACH
Two complementary approaches
are being used to define the state-
of-the-art for hazardous waste
surface impoundment technology: (a)
case studies for a selected number
of Si's, and (b) interviews with
technical experts.
CASE STUDIES
The Si's for the case studies
were selected largely from those
operating in the southwestern
United States, primarily California
and Texas. In selecting specific
impoundments, the following data
sources were consulted:
• State data files for CA,
TX, UT, NM, and AZ.
• State Surface Impoundment
Assessment (SIA) surveys.
• Background information
collected and analyzed
recently for a number of
hazardous waste management
facilities in connection
with several recent projects
sponsored by the Land Dis-
posal Branch of EPA's Office
of Solid Waste, including
"Risk Assessment for RCRA
Subtitle C Land Disposal
Regulations and Regulatory
Impact Analysis," "Prepara-
tion of Background Informa-
tion on Hazardous Waste
Management Facilities,"
"Assessment of Technology
for Constructing and
Installing Cover and Bottom
Liner Systems for Hazardous
Waste Facilities," and
"Waste Stream Description
and Categorization."
Twenty-two facilities were
selected as potential candidates for
case studies based on consultations
with the above data sources. These
sites were further screened by means
of exploratory discussions with state
regulatory agencies and facility
owners/operators, and the following
factors:
• Amount of data available
• Representation of a range
of SI service types
182
-------
• Type of liner (including no
liner) used '
• Types of hazardous waste
handled
• Design and construction
practices used
• Facility performance
• Absence of litigation
• Anticipated level of
cooperation from owner/
operator
The prelimary screening elimi-
nated eight sites (Table 1) from
further consideration. These sites
were rejected because of multiple
potential sources of pollution
(three), ongoing litigation (two
sites), and lack of operating data
because of the newness of the sites
(two sites). The previously
collected data on the remaining 14
sites were then supplemented with
additional data obtained from state
agencies, owners and operators and
used to prepare detailed assessments
for these sites. As a result, eight
additional sites were eliminated
and detailed case study and assess-
ment summary reports were prepared
for the remaining six sites.
Table 2 presents background
information on eight facilities
that were subjected to detailed
evaluation and rejected as unsui-
table for the case studies, primary
reasons for rejection are included.
As noted in Table 2, these sites
were generally found unsuitable for
performance analysis because of the
absence or lack of adequate moni-
toring data (Facility A), multiple
potential sources of contamination
(Facilities B and D), recent changes
in operating practice that eliminated
discharge of hazardous waste to pond
(Facility C), refusal of the owner or
operator to participate in the study
(Facilities E, F, and G), and dis-
covery of the unsuitability of the
site after more detailed analysis
(Facility H).
Tables 3, 4, and 5 present the
information on the six cases that
were evaluated in some detail and
for which case study and assessment
summary reports were prepared. Table
3 contains general information on
these sites, including the types and
qualities of wastes handled. As
noted in the table, these Si's serve
a variety of industries, handle
different waste types and volumes,
are used for disposal or treatment
purposes, and range in age from less
than 2 to more than 3O years. Some
general information on the environ-
mental setting (surrounding land use,
geology, and hydrology) for the Si's
is presented in Table 4 to indicate
the spectrum of conditions repres-
ented by the sites considered. The
SI technology information on the
sites appears in Table 5. The case
studies cover a range of liner types
and designs (single versus double
liner, clay versus Flexible Membrane
Liners (FML) versus clay and FML)
and leak detection systems. The
extent of groundwater monitoring
also varies from use of no observa-
tion wells to use of multiple wells
at strategic locations.
The case study report prepared
for a site is essentially a compila-
tion of all available relevant data
for that site and is the basis for
performance assessment. Depending
on the number of data available, the
report contains some or all of the
information categories shown in
Table 6. For each facility, the
draft case study report was sub-
mitted to the site owners/operators
for review to assure accuracy and
completeness and to provide them
with an opportunity to expand or
supply additional information or
clarification on the various topics
covered. In transmitting the draft
reports, the owners/operators were
requested to provide any quantita-
tive engineering data (e.g., on the
original site design, actual
construction, liner inspection/
maintenance programs, etc.) that
might support some of the quali-
tative statements and assertions.
Comments received from the
reviewers were incorporated in the
case study reports as appropriate.
183
-------
I
H
CO
w
H
H
H
>4
H
w
I
E
0
•r—
4J
ra
E
«p—
E
•r—
p—
LU
S-
,0
4-
00
•1—
CO
ra
CO
S-
•P E*
3 T
XI O
E 4-*
3 -p-
0 E
i- a
C3 E
E
•p-
ca <->
cu cu
H _i -P
cu
a
cu
a.
'
t.
cu
E
'JJ
CU
CL
J—
CU
00
>3
•M
£1
•p*
CJ
s:
•fj
3
E
O
•4-*
ro
O)
•r-
•4-*
•r—
i—
E
X)
OJ
>
I—
O
>
E
•p- CU
4-3
^•j ca
E 4-*
ca co
CL
E CU
O .E
O 4->
CO
cu
CO
cu
>-
*
CO
_J
E:
LL.
00
O
•1—
^
to
CO
•P- CO
•P 4-^
00 CO
3 ra
•0 3
E
•p- CO
00 O
3 XJ
0 S-
•p- ro
S- N
co ra
> -E
CO
3
O
XI
S-
ra
N p—
CO CO
.E 00
o
r— CL
CO CO
O XJ 4-*
Q) rl) r—
§4™' *p~
co o
O ro ro
O34-
^>
E
O
CU
o
•p-
>
^.
cu
00
e~
•p-
X)
cu
o
ca
• "a.
E
CU
4-^ ^)
00 1 —
>* 4-*
CO E
CU
CU O
r- S-
00
CU
0
z:
cu
CU
CU ro
a- 3
"io >>
O 4°
•r~~ *P™
^ P—
Q} *r~
JZ O
u ca
4-
r~
CO E
i- 0
3 «p-
p- ra
o 'zj
IP— F^
i- S—
cn o
"^ H—
^~
O
•i —
-[->
fO
o>
•t—
4_)
,(—
^—
XI
ZJ
4-
S_
CU
a.
Z3
to
cu
I—"
_a
•i —
CO
CO
o
a.
CO
0
•z.
o
s»
D-
00
3
O
T3
^
ro
N
ro
C"
CO
3 00
T 00
ra ca
> 3
Dl
•r-
i.
*^
4^
O
CO
4-
E
ra
4-1 >,
4- -E?
ra -P-
S- r—
O T-
i- 0
•p- ra
=C 4-
CU 1
ra E
3 ra
ro ^— ^ CU
^> 4J
CU 00 'p-
CU 00
O -P-
4J CD E
0 0
4-> p-
i— o ff~"^
3 E a>
O -E E
•p- CJ -p-
t-sg
•p- ro
XJ S- 4-
0)
CU -E X)
(J 4-> E
E O ro
rO p—1
E 4—
S- O XI
0 E
4- CU ro
S- CO
CU 3 oo
Q. CO r—
O i—
1— 1 CU •!-
00 J3 4-
00
CU
Q
^
>^
CO
p—
O
3
4-)
•p-
CO
1
E
»— i
j_
•fJ
00
XI 3
E XI
ra c
•p-
r—
cu ra
p- O CO
£!$
4_> ru co
cu .E ro
a. o 3
CO
3
O
-o
s-
ro
N p-
ro ra
^: co
o
p- a.
ra oo
•p- -p- >,
0 XI 4->
§CU P™
4^ •p'
00 O
O CO CO
O34-
^
f-~
40
E
CU
O
cu
S-
^^
S-
cu
ra
CD
cu
^-^,
E ro
O CO
•p- CT1
4-> p-
ra
S- CD
CU E
Q.T-
0 S-
Q.
to • —
CO
CU
00
0)
>-
t
SI
"•
00
3
XJ
c_
CO
tsl
CO
-C
00
3 CO
'£ 00
ro ro
> 3
CU
o
ca
a.
CO
O CD
i. E
CU -p-
=c s-
4- 0
(t5 fO
S- 4-
O 3
S- E
•p- ra
=C E
E
3
O
XI
XJ
CU
00
O
13
-l_^
E
ro
CL
CO
ra
XJ
O
E:
o
_j
E:
LJL.
CO
4->
00
ra
3
'rO
CU
SI
XI
i- s-
ra cu
0 t-
JD 3
4-> U
•p- ro
3 4-
U 3
S- C
•p- CO
0 E
XJ
ca
00 CU
p~ *i —
•p- 00
4-
XJ E
E O
ro
r- XJ
- — CU
4J
00 CO
CU O
•ri O
CDp-
O
*O CD
E E
.E 'P~
fQ
4-
S-
CU XI
-[__> frt
0 i—
CO
cu
o
^J
rO
P—
O
ra
•p-
i-
•fj
00
3
XI
•p-
00
3 CO
^ CO
ra co
> 3
CO
3
O
XI
CO
N r-
ca ca
x: co
o
i — a.
ro co
•p- -p- >,
O XI 4->
CU -p-
E oo cj
O ro ro
r_>34-
1
E ra
0 S-
s- cu
4- CL
O
00
CU CD
r— E
'5.43
CU CU
4_9 g
CO 00
ro
'A
o a.
o
P_ •
•p- XJ
3 CU
O
j * p^
•i— tO
p- .a
•p- ra E
u o
CO E 'P-
Uu ro 4J
00
CU
o
l: _l
s
u"
00
cu
4J
CO
ca
3
CJ
•p-
XJ
G
E
CO
'o.
XJ
o
ro
CJ
•p-
3
4-
3
to
Ul
M
0)
•H
(D
G
(0
t?
f**
6
fl)
s
0)
•H
X
fi)
Ux
rH
pq
•K
184
-------
CO
W
EH
H
CO
STUDY i
W
CO
<
U
CO
Q
W
E-i
CJ
1-3
w
«
E-i
D
ra
Q
H
(_>
QJ
QJ
ce:
s_
<£
C
o
l/l
ro
QJ
c£
1"
'C
d.
QJ CD
ro •(-
Z3 -r-
°i
c
(O U
OJ
0
QJ
Q.
s-
QJ
_J
QJ
>,
t-
in
.g
QJ
Q.
^
' 5-
O
ro
u_
c
o
•ij1*-^
•r- ro
i — d
•r- CD
O -i-
ro (/i
Ll_ QJ
CD
TJ
QJ
4->
rog.
11
OJ 4->
QJ E
-Q O
E "O
4_> QJ
0 S- •—
C OJ *—
•t-> ro
-o ro 4-»
i— 3 in
ZJ -Q C
O SZ T-
O ZJ
0 4->
QJ 1- in
a en ^
c ••->
ro QJ
ESS
£ cu2
S- V) E-
QJ rJ QJ
Q. ro 4J
(J to
K-. QJ >>
on JD in
in
QJ
O
•a
QJ
I
c:
QJ
r— O
E ro^
n re
w
*->
i
•a
o
en c
S- ro
i — Q.
ro cn
•»-» C
E 10
i (/)
SOJ
O
XJ O
^
t/) *>~ "O ' "*"* E ^> * ' ' 5 ni "Jj 2 ^ e Pi f—
-o 4-1 4-> aj c in •!— ro S- ui > — -i-Jaj +jc et_ t_ s_
^ ro^-^-^8^ ^^B.^i^tfl^ row rowQJ rowaj
+J OO>flj"S3->°'"*J:^S'r-'-S*f- O -E Sn° -Sd)°
ro2 'o"ct«'f~U'5 laroaj-c-^ajirt TJC: 'ucr— 'oc-—
Q '£ -o "~ _<^- 1- 1— i T- "O i — o ro-r-tn *Oi — ~o ro ~o ro
4l§ s_-o>o O oi-C i— S^S rotnro roinro
S *~~ o.^ in -1^ ^ ro ui Hi 'r^1^ 4- 2 ro Q.JZ o.^: •*-> Q-J= •<-»
•^cn o +J =3 r— ro N -r--Mt>iajo M- o-u^a o-i-'^ \-§
rtjQj SQJN cnc mroro>^(/)v)QJro SQJ-CQJ ^SSo'o-SSS
.^,H °fdJc;oTc: vj aj < — ojQ.cn rozis- rocrs- rocrs-
^_,.^ | — OCuiE-^ 1 — H- Oi ro •— < O in d. H- O QJ tn 1 — Orotn I — (jrai/i
ui tn w w
0 QJ 0 QJ OJ QJ
^ >- 21 >->->-
t/i i/) _ z: >- >- ^ 2-
= 1- E
54! ^ °S CSJ M-
•r- >, ^
+j t- «i— fi c ro m
•OOJQJ>iin< — i^ O- — QJ *^
S^.= -°e^^ £ -a°o «
•i-oi— -OOC^JD QJM-C: J=:
r-o-Qjro^ oo-^ Q-
i .--i->-o c s- ro c jg
oji — i/irodino r~"m *
^3 E OJ ro to a.~o >, JT ~a
rs i £ a. o c ~ +•* Luu-ac ex ca_ •»-
ocbjzOJM-. — roi3 Q_ccro >> Q o
QI — minoinincQ co-i-roin 3; ^t v*
.E:
•r- OJ tn Ol
toi — 'roro rotn •<-* I+J
ajro -i-^S 4->ro ro '?
4-> 4-3 QJ QJ 5 ' — *~
i/) QJ ' E en E O- *^
^ -o^^ ^ro !5'^ tj-a1 ^-S1
'El 'oSo *oS? OQJ r—.— rf "" '
ca r- E «r- D. Q- ^J '*~
.^ro Ss- a- a. ^^^
O "o. 53 (_> O UJ M- O
03 O a LU U- CD
o
C U
DOM
H C O
>i O C
.C H -H
185
-------
CO
u
EH
to
H
K
W
EH
O
/3
p£
2
o
$
i
Q
Z
CO
K
g
w
Pn
J
K
Z
W
O
1
CO
w
t-l
EH
H
H
0
fa
SH
Q
to
H
to
o
n
H
m
EH
,|5*
QJ •*—
4-> P
w C
o-
(U
VI
m
o,^
N*IA
•r- >
•a
3
•P
to •
I/I
r**cn
OJ *r- Ol O)
"§•*= ^"3 -^
0 rjj C ^ ^ C
•cd" i —
O r—
CV,
CT» OI
rd "O
to c:
0 O
CL a.
•5 *(O
to
Si 1
So *(—
t/J CL O
|
5
C
CJ rd
'-P *CL
>>
r- 0>
2 =
-P C
O T-
Cj L!
^-
it
ra c:
1-0 =
°&|
SI'5 =
CQ^ — - |=
i_
•P
§ =
QJ
to
£=
S_
*
*u
«
QJ
r- r-
? 9
cn
2 2 ,-
c
C 1 O
10 OJ-r-
,rz -i— t/i -P
to -P -o c: rd -o
rd rd c: •.— I- c:
S I- O S- O O
O CL CL CL
25« 51-S
O OJ to (J to
•r- O •*— ^- O
-P C CL -P Q) CL
to S w u to P to
-P
c c:
O rd
-p 'o.
id
3 O
£ -p'
O 3
tu'tl
•a -P
•r- to
O '1—
•t- X>
•p
QJ c:
Cu rd
OJ
CvJ
CO
01
r~"
00"-
CD
5°J
QJ
« 0)
QJ 3 C
« +J ^— n n3
•r- E OJ QJ l/l
O Q.-5 O
"O (d QJ 'I— r—
c: « w -P (d
c £ E $ .,--
•P QJ OJ -P « -£Z
(O C E O QJ O
^^ ^^"E aj
illl&f
«
1 , r^~
™ - ^
i n rd i
X) QJ 10 (O
C en o t-
o s- £ w CL
E 4- -P "O > ra
•»- to OJ ui
~o ai o
CO QJ « O> 1 — CL
(/) O1"O i/l
.. 3 C 3 .. -i—
«=C ••— i — ca Xi
to — to to
OJ -P -P QJ C TJ
•P C -P *O -P O C
•r- OJ OJ C -i- -.- O
-P*
O O
•a rd
s"Z
rd rd
r- CL
(d to
U T3
S-
QJ QJ
1 1
CO
>> >^
-o -a
o c: o c:
O 0 CJ O
«r— O r—
o rd ro ro
c\j 01 r^ en
§p
U "O C
10 o ^ "i -p t?.".?
o CL v o rd P zi
M-ozcto 'cLat.— cxj
x: CL QJ i- 3 v •• — »
QJ ^^. j= -i- o a: QJ
4J QJ 4-> ,— S- 1- TD CLT3
3 Q> QJ C rtJ
£3 •*- Dl (/> -i— "O -r-
•r- u -r- -1-03
0 P E
CD o ra -P
CO) J- 03
;= *- as
P O ro -P
•p is e > ^
l/J "O V) "O
QJ CL OJ C
r— to >i CO
i — 3 O> O Q-
J_
QJ
N
i — P
•r- C
•P rd
S- r-
QJ CL
H-
cn
r— E
2r
3 3
-P P
r^~ O
0 4^
I1 1
^f
>j
ro
O C
O O
O r—
r- OI
E
to
rd
E
ra
CO
ro
«D
CTl
I
o>- — -
C P
-i— C
i — QJ
0 B
U fO
QJ
C P
O •— '
186
-------
r— *> VI T3
•P O O O E
CD O
8r=
C* S-
•r- O
O O
•i— HI
" 01 ro :
ro o
4-> -o
^V>
O.S-
4-» 4J O -O
T3 E E "O 13
c tu cu 5 E
ro E E o
u ^ .^ ,_ a,
vi ro ~o JD c:
•o oj QJ T-
£= ^ T-
i — t- o O *-
rtj-a
•o
c
•H
-P
G
O
U
t/i to
S_ E
ro
W
o
CL
o CL
O-VI
ro-i-
"O "O
(OErdC
cuoajo
_c:ci._c:o-
Cn o
£ •£
-a -a -i- i—
ro to
s- o
o o-
JZ C O- V)
cn o ro •<-
: n 4-J uj -—
CL >,-*->
i — "3
>— +-> QJ
s- o
ao.
t-i
ro •<—
OS
Q.S
&>,B
r— 4J OJ
ro E SI
ss^
Q. t, T3
187
-------
CO
El
U
FEATU:
1
&
VIRONM
M
1
a
E-i
H
H
U
<
S
W
w
o
t
^*
S
9 '
OJ
L. O •
m t/j
*5i
oi 4J c
S oj **-
0 ••-* 0
o
cu
>,
o>
o
S
1
S:
0
i
1"
4->
*O
t.
S
1
V) •
sg
U
to
1 "O U C
S g^3.°.B
i- O.CM ra 4J
c rx. 3; ro -M
Sotoi S i. c
Cr— 43 0 QJ
C -r-"— » in O- O
Ol i— ro id rd •
J3 O "O 3: > -o QJ
8Qj UJ T3 4->
SE-0 rd-r-
Q rjj in
f- *0 -0 C • M
in QJ C •«- T3 "O V) QJ
1 C rd •— C C -i- C
S^SiRRS S
• 4->
>> 1 QJ -O QJ QJ Ot
r- M , 3: o
ro (J •£? 5 J-
« t- 1 *i— i — OJ QJ
t/i E ai ••— i — u 4->
ocuH-£.r-Qji-«d
j= -M *.- a> u j= 3 3:
U 10 3 W ,rd -M O
C >> O- »t- in r—
•»- tn rd rd tn rd
OJ QJ r— *r—
cot-Qj-o^rcnros.
• QJ r— C +3 J_ o. •*->
CO -M .O rd rd •<— in
—. flj O CT3
QJ -6 Of -r- Ol T- C=
0»Ct-4Jr— J- I- -r-
23 QJ f- L. CL
p -M in a> t_ -a
o» C «o -o QJ o> c
>cn3:aJc:|+-J=rd
rd J= =J ••- -M
X ro +3 3 _
r— QJ « CT in rd •
r— ^-**-J=t-rdt-a. rd
rd Q. o •*-* OJ -r- -M
**-g ro^-t-JSU • nj
COO)OJ*r-a}(J*r-rd "D
•*- U C C 3 O.-I- C QJ
S^c^g-g-iiJs S
5 =
SI JZ
ro 0-W
c c QJ tn
ra QJ gr- 3:
*J3 T3 O •—
01 ro t= i— OJ -r-
c ra jr to
f- CD *t- £3
i— C QJ O "O
^^•w +5 Jl .3
r- C tn +-> rd rd
(O QJ *r- QJ >, TD
C X .g -^ QJ r— O
• - ^ g
3 OJ QJ *O. iT JD *3
O >i > Ol 4J J_ O
tO ,— QJ QJ rd -r-
^«^(/j^OJ.E "oZOj"?
^ ^- § S o? • Q) -
OJ-O3C IdZD^-'^
Ct-QJOQJ > O (J
O*ON"UO CETU*
•r- *r- U rd i — S- >r- QJ i —
Ol JZ 'rd M *'5> •? -S in E -2
>- t— ;r- rd rd >tnajo+->
f -S"11 v> ^ 2 S ° S
u • tn i- t- <— jr *«"a
a»*->3OO rd43T30J.r-
Q 3;-r- g J- >to
pll
8 S e"1
•=c-g"
o^i ^.;
4-> C 1 i —
n o *-* *rd S- -r- cr
i- t- > a) rd ro
S °^E?ra£r^
•r- •OJ^rdTDQjtt-'o
>< QJ jz rd c c s-
Ott-4JjrECD rdO
^ _^ 5 M «2 . -"^ Q.^
cci— aj-M^in^-
ep CO +-» in jz in
!— ^ i — T— XJ
en CTJ QJ c: aj tz
QJ -U T- jz; T3 rd
•f- r— 00 'r- i- QJ C:
i — aj +J QJ m >, rd
«^3-l-»<:-(-> > 3 rd>—
QJ t- • -t- T3 O QJ -C
o 4— •+-* m rd 3 S- i.
£rd rd S- o f) rd rd
LO ^ QJ Oi E mr~ i — E
00 >> O
Sit 4-> 4-> E
rd ro i — O
§QJ ^£ tn C3 >»4-
SO C 0 [--•*->
rjj ^_> .,_ O 4_ , — w
Q. S- S- o >>•<- Q>
i — tno J^QJ'O'Ord •
>> rd" J!? OJ ^ 3 rd "u*
Tr> ^~ E r— O >^ QJ
C t-QJJZ r^-O-i-^C.
(d *r- O +^ **— rd *i— (J
^~ C in ^— T3 tn -Q
S- ra QJ i — c « — in rdCO
XJS-C4- lOCoEO
crord Qjrdrd s-i —
3 s- «c:.E:in-i-»aj
**" *^S QJ a. o
4-* QJ in QJ 4—
QJ -P T- in TD 0^ >, >SU3
-0^1
c: i/i c
QJ "J S o
tn .Q -r- o
13 S- r- O
C rd O ' —
rd OJ in «t-
r~ QJ XJ
it! -3
3 >i-i- in
O -C O
10 E 3XJ
1
•i— rd aj in c -i- •
rd"~QJL? ^XlV-O
s_ l_ in rd o o o o.
•r-4->cnaJ M-i— ^fc
lll!IIs'51
aJrdJ-COini — QJO
^"ss^^^sls.
^_
"nt ^ nt -r/i "? 4-)
'ojaj" •— xiT^^
CU O • C 4-> CU
<1J i — 4- QJ QJ QJ •*-*
CD O_. — O~'i- to O O
CO 4-> rd 3 -i- '« -r- i —
in rd *O fd t- o CL
tn-i— «cc:iDli-4J3cu
X 03 Q) M— JZ Zi Cn C fd
QJ O H- - C (J O"-^ o O
J-H-ZJCUOCU S- , —
QJ S- O~i — CX"O •(— i —
4->3rdJDrd QJ rdtn
20 s-J-Jai oui — "a o i-
i- QjrdaJt-crQjci^— o
m -t-> H—
•P rd C 0 >,
tn cu -H U LO -M
rd -r- -Q r—
C i — "O O
cu o fd cu tn
•r- oil— e c:
CU *rd
c: cu -t->
4_ .,_ i_ ,_
111 I
r— i- rd
rd 3 4->
> co rd i-
r- ££
rd *
•i— •*-> CU >> •
> to S- r— in
«— jz in rd o
"* 3 ~C -^ £" V
cCtyir- 0.0.
*
Q3
c
o
s:
£ CD
QJ 0
QJ tO 3
>Q 4-* O
'rjj O
rd fd o
oil — tn
-O •*- 4J
rz ro rd
to iz S-
o cu
rz u -P
rd (d
u {/} (^ ,
-C JZ S- QJ
\— is om-
rd
* J-
i — CU CU
> QJ 4->
.g 2^^ S_
. rjj
•^ *O rd* C
*~ r— "O
in in u a)
4J 4-> 4->
•i— in "O rd
tn -r- c O
O in rd O
O. C r—
T3 O •»-> CU
•— i- •
'i — >v* rd >>
zi ••- -a r— C^
• — •>- rd rd rd
rd
t°- QJ
OJ
Is
C -M
0 3 .
H-
CU M-
in cu
•o i —
QJ
LT>
CU
C
o
^
S- O
QJ CD C
QJ 'aj C QJ
O rd OJ •(- JZ
S- (Z to ra
QJ 3 Q) in C
ro .a in
3 *•!- QJ
•a cu 4-> to
3 'r- QJ QJ
O tn 4- 4-> i —
S- rd fd
01 c o 3: T-
S~ 01 -0 O
Scu c: -r-
4-» TD 3 M-
S" C OJ J= 4^-
'
o >, q r
rO -P rtj
rd c: u in p o
>, -M -o -^ cu
c In aj' o
'ra o 5 "^ 4J ' —
i— (J rd T— QJ E
QJ T3 1: "" l£ £
•0 QJ C >-, 4-
C -Q 3 -O O
3 IZ O to
QJ « fd r— CU •
S- rd C ^ 4- C QJ
rd i — -i— -o O rd in
4-> C i- \
.,_ QJ 4-> o t!
i — -O QJ
CU «=C T3
-------
s
W
EH
>H
C3
DNITORINi
S
Q
Z
K
W
2
H
W
H
EH
H
1 — |
CJ
fa
X
Q
D
EH
W
CJ
in
W
J
1
S-
CU Ol
rd •!-
|s
11
o s:
c
o
4->
CJ
CU E
S3
O in
jw: LO
ra
QJ
i
o
4J QJ
rd •—
i — rd
ro i —
C >
i — i <£
C rd
01 -P
•i- ro
to Q
OJ
. QJ
0.
I-
CU
c
t/1
4-
QJ
Q.
•o
to
CU O
vi z:
fd
QJ j —
.c E ~o rd ,H- TO fO
fO i- C QJ
" ii^^S
,— X
QJ .C QJ in Ol
> O VI C
fd C QJ O •«-
s- •.- CL i- -P
01 1 -i- (J O
ro o. (d QJ
-P C
4- QJ CJ >, C
1 C >• fd O
. — o n- 3 0
c
OJ O
fd c: o
> vi -a
•d c
O X)
V) •«- C
•i— U *
siEl
-O O -r-
^^-d
•r- (/) *> QJ
rd -a -P
-*-» O C V)
cu c o >^
-o
QJ
•i- O
fd
-0 S- S- QJ
C ra QJ >•)
M- O CD
0 c? it1 ra
vi rd
QJ **— "O to
a. o c •<-
°^s~
"%X? .
•o >, c cj
C fO O >
id < — * D_
o a. +-»•!-
4_> 5 i O 1
(g'5'icBg
-C
in I i —
g2g
cu a. o.
TJ fd v)
•r~ > •>-
U QJ "O
'•P C C TJ
VI 3 0 C
QJ 0 T- 0
a. -o -P o.
OJ
QJ
3
c
o
fd
>
QJ
c
s
1 O- ro
0- S-
ra CU
CJ V)
•" > JD
S- 0- 0
|-o c
^ 2S
CO
rd 4- -O r—
VI J- QJ ' —
Q) 4-> QJ
•P a. o 3
M— QJ
1 .C C C
i— U C O
c o •«-
=
=
=
CJ
• •=*•
LU a_
D_
"*i
•r- 1
EO
O *"~
ra vi
II
cu •«—
>^ [rt
CTD
fd c
E 0
•i- U
S-- CU
Q. to
•a
1- = 8
cu o S-
QJ 4-J «i—
•O rd -P ^~
•r- 3 fd rd
O 1 S- V)
•r- CU O O
•P vi a. a.
to c: rd vi
QJ -r- > •!-
a. s- QJ -a
• o
VI S-
i — fd
QJ V)
* ^
QJ QJ
QJ 3 •
S- \f^
jE ^ to
r— QJ
«t fd -C
13 U +J
CQJ QJ -r-
rd i. OJ -p- -P o
QJ JZ -r- 01 fd -P
^r 3 -P to c E C
o •<- o •«—
C Q. O>4— O--*J
cu E c: o fd 3 ^i
•P tO i — QJ V) (d
rO OILtJ QJ o
C C TD S-
•r- O "O QJ fO "O
rtJ -r- QJ • CU
L. 4-> r— C r- VI D-
4-> O i — S- f^ T3 E
^j cu rd cu aS T- i3
n r- *J 4-> r- 3 0. - QJ
t/1 i — (/) V) O" Vt C
O C ro C -i- >-,^ O
«c p -i- cu -r- r— i — to z:
QJ
1 C +J
C QJ -P
Scf^
o c
-a •<- i3 -
rd rd o, vi
0» 0 "~ C cu
•o "" o •— =
C CU «i—
*o o ' — 4J
Q V) rd i — "O
>, >>
CJ CJ
QJ (O IV)
t vi •» t/i fd "O
E = OITIJ o > c
•i— c c a. QJ o
v) -i— fd in a.
00 -P r- T- r-
^ +J n -Q C f—
• •OJ-PQJ "Oro
Qj CQ-r-in
OJ C S- TD QJ id d.
+->ZJS_OI3 4->S_v)
•r- O O -P — T- O *i-
tO CL.1*- VI Ul tO Q-Ta
CO
QJ
z
QJ
C
O
^
fd
V) -O
•r- rd
•S4S
ra
Q) TD
+J C
J -P
t •
•J— VI
I •«- C
C J3 O
'S ^ QJ
fd t OJ -M
•P °"JZ
•r- \ C
E fd E QJ
Ij -P
jr i -P
. -M O VI
"O i — •«—
Q) QJ- — X
C +J (U
•i— rd i —
,— CJ fO —
C -r- QJ >>
iz> -a tn -P
E
O 3
4-> V)
01 a.
C T3 >,
•r- QJ p*
-P QJ
1/1 XJ O
C CJ
i— O QJ
r— Q. S-
W
QJ
^
CJ C
vi > fd
C Q-
S^. -i— d. VI
1 O., —
•r- JT rO Ol
T3 -P S- C
1 -r- "O T-
^» QJ S-
^r o
OJ T3 U -P
QJ QJ C -r-
S- C -i- C
J= -r- | O
(— r— ro E
T3
O c in
O- O C
U O
-o -,-
QJ -a -t-1
•*-> C 4^ S
QJ Ol-i— fd
-a .,-{_),—
C T3 V) >
rd to
S- C
QJ QJ O V)
4J C -r- 01
"a.""" g '3
CJ rO VI X>
CU
^- , •
QJ C O Q- TJ i —
O ' O
rd 0 O CU QJ QJ
C -P !- O
rd O S- O rd
.O C tn E
50J JD ••- QJ O
^ Z3 QJ Q. CJ
'g ^ | ^ 4J
C oJJ C V) M-
~O O O 1
QJ QJ S- QJ i —
c cj s- cj -a
••—>•!- ra T- m
_J ti. I— Q VI ••-
^_^
C HP
O C
fO -P
S- rd
O CU
!£
«1
c o
o a.
QJ
c
s
=
=
s
=
=
=
=
•p
c.
oiB
^s
0 -P
O-— •
(J
T)
QJ C
C O
o a.
189
-------
Sg1
!*
II
Continued
to
w
4-» O)
ro i—
r- J3
r— «
43 £
= 1
oe
J- (O
C7>r—
S- 0>
3 E
O -i-
LL- TJ
E QJ (O
O -O (0
U (d O
Q) £- -P rd
It) C CD «r-
S- OJ CJ» >
SOI— ro C
E >> T- E
O -r- .^ E 4-> -.-
(U >>-P T- O W
^ "u >>-O C JD
SQJ — QJ E
> O U O Cn
O O. rd O E
> O 3 U O
OJ U E E -O
QJ >r_ .,_ |
-
SQJ O Ul
c: -r- o>
(U -r- 4-> C
S u
•i- QJ
E C£
(U
-
mo.ra
o o. t/} at to
•— O J= E O
t- a. 100.
tu *o wi JT •*- m
>•— -r- 014-*.!-
•r- O *O -r- rd TD
;a^_
) E 'ra
j o w
:£g.
: rd wi
n i_ T-
;§.2
4-> E -0
•i— ^6
E O E O
O Q. O Q.
•r- I/) -r- to
CX >^
l QJ fd
S- r—-
190
-------
TABLE 6. TYPES OF INFORMATION SOUGHT FOR THE CASE STUDIES
• Basis for site selection:
Geotechnical and hydrogeological considerations/studies, including results
of any soil boring tests, modeling effort, and environmental impact
analysis.
• Basis for selection of the design firm (and construction/installation contrac-
tors).
t Criteria used and basis for design:
Design objective and purpose of impoundment (disposal, storage, treatment);
Waste characteristics data (flow, composition, etc.);
Hydrogeological considerations (e.g., distance to groundwater, precipita-
tion, evaporation, temperature, etc.);
Waste-liner compatibility considerations;
Soil permeability and leachate attenuation test results used;
Regulatory/permit agencies requirements;
Cost considerations;
Etc.
• Description of design and design specifications:
Unit size, dimensions, and configuration;
Liner type, characteristics, and recommendations for installation;
Specifications for foundation preparation, leak detection system, and
liner backfill placement;
Inlet/outlet, aeration unit, and other appurtenances including systems for
diversion of surface runoff, prevention of vandalism and animal damage;
Observation wells and monitoring systems;
QA/QC specifications;
Etc.
• Description of as-built facility and related matters:
Modifications to and deviations from the recommended design and specifi-
cations, and the basis for such modifications and deviations;
Problems encountered which had not been anticipated or adequately
addressed in design, and mitigation measures used;
Level of QA/QC and competent inspection actually employed.
• Performance data and operating experience:
Evidence for liner/system inadequacies or failures (groundwater monitor-
ing data and leak detection information);
Operating difficulties encountered and mitigation measures used;
The adequacy and reliability of the monitoring system to detect or fore-
warn problems;
Inspection and maintenance programs;
- Etc.
191
-------
The identity of the case study
sites was of little consequence to
the project objective. Thus to
promote cooperation of the owners/
operators, all sites were kept
anonymous and are referred to only
by letter or number designations.
INTERVIEW WITH EXPERTS
This phase of the effort began
in September 1983 to seek expert
perspective on factors affecting SI
performance and deviation from
design predictions. Information
was sought from those most
intimately involved in the design,
construction, operation and regula-
tion of hazardous waste Si's. To
date, a number of individuals and
organizations representing many
viewpoints have been identified and
contacted as possible interviewees.
These interviews will compile
information on real-world problems,
mitigation measures, and research
and development needs. As with the
case studies, summary reports will
be reviewed and approved by inter-
viewees before the collected
material is used to supplement the
results from the case studies.
PROBLEMS ENCOUNTERED
One major difficulty with the
case studies has been the identifi-
cation of an adequate number of Si's
suitable for the case studies. As
noted, many of the sites had to be
eliminated from further considera-
tion, because of one or more of the
following reasons:
• Absence of rigorous data
and documentation on the
original design and con-
struction and on any
subsequent problems and
solutions.
• Lack of adequate monitoring
data for meaningful perfor-
mance assessment.
• Many potential sources of
pollution at or adjacent
to a site (e.g., landfills,
waste piles, other Si's,
etc.) that make it difficult
to attribute any observed
incident of contamination
to a specific source.
RESULTS
Detailed evaluation of the data
compiled for the six cases studied
has resulted in the following
assessments.
CASE STUDY NO. 1
The two FML-lined hazardous
waste Si's at this site were
designed and constructed by the
owner in 1972 and 1979. Since no
regulatory requirements governed
the design and construction of such
facilities at the time, no geo-
technical and hydrogeological
studies, environmental impact
analysis, or laboratory or field
investigations preceded the actual
design and construction to develop
sound basis for siting, liner
material selection, foundation
preparation, liner installation, QA/
QC inspection, etc. Onsite avai-
lability of land was the primary
consideration for SI site selection.
Apparently, there was no rigorous
QA/QC program or inspection of the
completed Si's by professionals
trained in FML design and installa-
tion.
Because of shortcomings, the
Si's at this site have failed to
provide satisfactory service.
Limited water quality data from
monitoring and production wells
indicate contamination of both the
upper and the lower groundwater
aquifers, as evidenced by increases
in the total dissolved solids,
sulfate, and nitrate concentrations.
Because cracks appeared along the
exposed sides of the liner in one
of the ponds, the liner in this pond
was replaced once. Possibly, the
liners in both ponds are now leaking.
Because of the relatively high
permeability of the geological strata
underlying the plant (O.44 x 1O~2
to 1.4 x 1O~2 cm/sec) and the strong
acidic nature of the heavy metal-
bearing waste disposed to the Si's,
any liner failure could result in a
substantial underground waste
192
-------
release and hence presents major
potential for groundwater contamina-
tion.
CASE STUDY NO. 2
The very satisfactory opera-
tion at this site indicates that
(given proper site and wastewater
characteristics) designing and
constructing small ponds to provide
satisfactory performance can be
simple tasks.
Two very small (48 x 28 x 5 ft.
and 3O x 2O x 3 ft.), relatively
new, onsite impoundments serve this
pesticide formulation and packaging
plant which generates intermittent
discharges of wash-down and rinse
wastewater. The waste volume is
very small, seldom exceeding 4OO
gallons per discharge, two to three
times per month. Groundwater in the
area is at 215 ft. The wash-down
pond is lined with two layers of
polyvinyl chloride (PVC) sheeting
as the primary liner on both the
bottom and the side slopes, and a
3O-mil PVC as the secondary liner
on the bottom only. The primary
and secondary liners are separated
by 1 ft. of gravel. The rinse-
water pond is lined with 2O-mil
chlorinated polyethylene (CPE)
underlain with 1 ft. of sand and a
lO-mil PVC liner, with both liners
extending along the bottom and the
side slopes. The leak detection
system for each pond is merely a
single, perforated PVC pipe (a 3-
inch pipe for the wash-down pond
and a 1-inch pipe for the rinse-
water ponds) extending halfway
across the pond bottom and
connecting to an observation well.
A 1/4-in. fiberglass cover was
recently placed over the primary
liner in each pond for better
protection against liner deteriora-
tion and damage during pond
cleaning. A 2O,OOO-gallon storage/
equalization tank was also recently
installed to control liquid level
in the wash-down tank.
During the 4-year operation of
the wash-down pond, no liquid was
observed in the observation well.
Since the rinse-water pond was
placed in operation only in late
1982, similar results from the leak
detection have not yet been reported.
CASE STUDY NO. 3
This facility, which has
impoundments located in two separate
areas, is an example of the diffe-
rences in performance between poorly
planned and designed ponds and well-
planned ponds . Problems resulting
from poor planning cannot always be
fully and permanently corrected
through piecemeal remedies applied
as the problems surface. This fact
is illustrated at Site A, where
nothing indicates that detailed site
selection investigations or pond
design took place. When the
facility was investigated in 1971,
wastes were seeping through pond
levees that had been built on top of
old waste fill. Work was performed
to correct the problem at that time,
but seepage was reported again
during investigations in 1978. At
this time, specific levee permeabi-
lity and thickness requirements were
imposed, and the levees were rebuilt
to conform to these requirements
(5 ft. of clay with 1O~8 cm/sec
permeability or the equivalent). But
leachate was discovered again in
198O, indicating that even the
improved dikes were not able to
prevent seepage.
In contrast to Site A, Site B
was developed in 1971 with some
effort to design ponds that would
prevent waste migration. A site
investigation was performed, and
soil compaction and other design
criteria were specified before con-
struction. When the site was inves-
tigated in 1978, no seepage was
reported, even though the levees did
not all conform to the new permea-
bility requirements and had to be
modified.
Site A may not have been
explicitly sited and designed to
prevent waste seepage, though Site
B was. This possibility appears to
have been a significant factor in
the performance of the facility with
respect to seepage. Building Site
A pond levees on a garbage foundation
193
-------
undoubtedly contributed to the
seepage problem, and the displace-
ment fill method of improving the
dikes apparently was unable to
solve the problem. Trench key work
apparently did not provide a com-
plete solution either, unless the
leachate originated (as the
facility owners contended) from an
adjacent commercial landfill.
CASE STUDY NO. 4
The operating experience at
this site illustrates (1) how the
materials contained in the waste can
provide an adequate barrier against
further waste seepage under certain
circumstances, and (2) how liner
failure and poor performance can
result from deviations from desired
liner specifications, reliance on
liner-waste compatibility tests
that poorly simulate actual field
conditions, inadequate attention to
geotechnical factors, and poor
design and operating practices.
The major surface impoundment
systems at this fertilizer manu-
facturing facility are 14-acre
"gypsum" ponds, an 8-acre evapora-
tion pond, and a 38-acre cooling
pond. The gypsum ponds are unlined
sedimentation ponds that have been
used to recover gypsum for almost
2O years. The natural build up and
solidification of gypsum in those
ponds have rendered them impervious.
This fact has been verified by
actual examination and permeability
testing of the core specimens from
the bottom, which have indicated
the presence of a very hard, low-
permeability material.
The FML-lined evaporation and
cooling ponds have failed in the
past and are currently leaking.
The liner specifications, which were
written by the facility owner,
called for a material that would not
deteriorate, when exposed to a waste
with the following characteristics:
pH of 2, maximum temperature of
11O°F, O.5% sulfuric acid, 1.63%
phosphoric acid, O.O5% chlorides,
O.5% fluorides, and l.O% organics.
Though the available data indicate
that the liner met the alkali and
acid (pH 3.O) resistance tests, the
material may not have been tested
with a waste simulating the above
characteristics. The actual liner
manufacturer warranty also provides
for an acid resistance of 1% weight
change versus an original specifica-
tion of O.3% (both at pH 3.O) and
limits the definition of organics
to only five specific compounds, the
total concentration of which is not
to exceed 1%. Geotechnical and soil
investigations had indicated that
the native clay had a high content
of calcium carbonate. This soil
characteristic (which would lead to
gas formation underneath the liner
in the event of acidic waste leakage
through the liner) was also not
addressed in the site and liner
design.
Liner bubbles have been
observed in the FML-lined ponds.
Inspections during repairs suggested
seaming failure as the main initial
cause of the liner leak. Seepage
of the acidic waste into the under-
lying carbonate-bearing clay resulted
in the production of large volumes
of gas. Since no provisions had
been made for venting, gas accumula-
tion helped spread the seams and
further aggravated the leaking. An
overhead discharge pipe that diverted
wastewater from the gypsum ponds to
the evaporation pond was also an
apparent contributor to the seam
failure problem in the splash area.
This problem was eliminated by
installing a splash pad after the
liner area was repaired.
CASE STUDY NO. 5
This facility containing eight
clay-lined impoundments demonstrates
what can be achieved when the im-
poundments are well designed, con-
structed, and operated. The major
reasons for the highly successful
performance of these clay-lined
ponds include: (1) the use of a
very impermeable clay (which was
available onsite) as liner material,
(2) the use of competent design,
construction, and inspection con-
tractors, (3) a conscientious owner/
194
-------
operator that closely scrutinized
all phases of impoundment design,
construction, and operation (from
site selection to QA inspection),
(4) extensive waste-liner permea-
bility studies, (5) excellent QA/
QC and recordkeeping during all
phases of the project, and (6) good
communication (input and feedback)
between two different state regu-
latory agencies and all parties
involved in the establishment of
the ponds.
The performance of the impound-
ments is documented by many years
of inspection reports and observa-
tion of the leak detection systems
located beneath each pond. These
leak detection systems were field-
tested by state inspectors during
construction. The good perfor-
mance of the impoundment is not
surprising, since the clay used to
line these ponds has held the borate
deposits mined at this facility for
centuries.
CASE STUDY NO. 6
The operating experience at
this site illustrates how poor im-
poundment design and inadequate con-
struction, inspection and record
keeping can lead to leakage and poor
performance. Wastewaters began
seeping laterally out of one impound-
ment because (1) sand lenses within
the natural clay were not identified
and removed as specified in the
design, and (2) at least one of the
embankments was not adequately keyed
into unweathered clay to prevent
lateral waste migration.
This commercial hazardous
waste disposal facility contains
three 5-acre Si's lined with in-situ
clay (on the bottom) and recompacted
clay embankments. Pond 8 receives
geothermal and petroleum industry
sludges and wastewaters, and the
other two ponds are currently used
for land treatment of organic
sludges.
A soils investigation con-
ducted before impoundment design
indicated the presence of sand and
silt lenses within the natural clay
beneath the site. The design
specifications called for (1) at
least 2 ft. of natural clay (with
a permeability of 1O~^ cm/sec)
beneath the ponds in which no sand
or silt lenses were discovered
during pond excavation and con-
struction and (2) subsequent place-
ment of recompacted clay. The
wastes began to migrate laterally
along a sand lens underneath the
embankment and surfaced outside the
impoundment because: (1) all sand
and silt lenses were not detected by
boring tests or during excavation
and construction, and (2) at least
one embankment of Pond 8 was not
keyed into the natural unweathered
clay as specified (and this non-
compliance with specifications was
apparently not discovered and docu-
mented during QC inspection).
Though no seepage has been
detected outside the other ponds,
the other embankments at the site
may not have been keyed into un-
weathered clay. The inspection and
engineering certification reports
for the site are very poor and
unclear on this point.
ACKNOWLEDGMENTS
This work has been carried out
by MEESA for the Solid and Hazardous
Waste Research Division 'of EPA's
Municipal Environmental Research
Laboratory (Cincinnati, Ohio)
through Contract No. 68-O2-3174, Work
Assignment Nos. 97 and 123, Subcon-
tract Nos. M98967JQ3E and M98968JQ3E.
The authors wish to express their
sincere gratitude to the EPA Project
Officer, Mr. Carlton Wiles, for his
advice and guidance during the course
of the project. Thanks are also due
to the many individuals and organiza-
tions who fully cooperated with this
study and provided data and input to
the program.
195
-------
TECHNICAL CONSIDERATIONS IN LOCATING SURFACE IMPOUNDMENTS AND
LANDFILLS IN LOW PERMEABILITY SOILS
Dr. C. Clark Allen and Dr. Ashok S. Damle
Research Triangle Institute
Research Triangle Park, North Carolina 27709
ABSTRACT
The question of the anticipated life
of waste disposal facilities located in
either saturated or unsaturated porous
media of low hydraulic conductivity needs
further consideration due to the reduced
pore velocities through such materials and
the increased travel time this implies.
Technical issues have been identified
which influence the degree of protection
provided in saturated versus unsaturated
soils with low hydraulic conductivity.
Permit applications for landfills located
wholly or partially in the saturated zone
of soil require demonstration of effective
containment on a case-by-case basis. A
better understanding is needed of leachate
movement in saturated soils with low hy-
draulic conductivity to accurately assess
the potential environmental endangerment
from the operation of landfills under these
conditions.
Previous advantages which have been
identified for locating surface impound-
ments and landfills in low permeability
soils include pollutant adsorption/attenu-
ation capability and containment of con-
taminants in the event of liner failure.
Many naturally occurring low permeability
clay soils are located with a water table
near the surface of the clay. In addition,
surface streams are often nearby. These
clay soils often have heterogeneous and
anisotropic hydraulic conductivities which
permit surface seepage of contaminants.
The environmental implications associ-
ated with surface impoundments or landfills
located in saturated soils with low
hydraulic conductivity were identified and
examined. A more detailed assessment of
the potential environmental implications
of landfill management was obtained than
had been provided by previous studies,
particularly the impact that heterogeneous
and anisotropic saturated low permeability
soils will have in influencing the movement
of subsurface leachate releases. In addi-
tion, the influence of soil saturation due
to capillarity above the water table on
leachate movement was examined.
A number of numerical models are avail-
able to describe the groundwater flow with
solute transport. Analytical solutions are
also available for a limited number of sit-
uations and usually involve simplifying
assumptions such as isotropy and homogeneity
of soils. Numerical models generally employ
either finite difference or finite element
techniques to solve the governing flow and
transport equations. Finite element schemes
offer more flexibility over finite differ-
ence schemes but are also highly involved
comutationally. In order to fully assess
the environmental impact of landfills loca-
ted in saturated soils with low hydraulic
conductivity, the predictive models should
incorporate the effects of soil heterogene-
ity and anisotropy and should be able to
handle low hydraulic conductivities. With
these considerations, the applicability of
existing groundwater flow and solute trans-
port models to the present case was inves-
tigated. Models typical of different
numerical and analytical methods were com-
pared for accuracy, cost, time requirement,
and ease of use for a typical set of input
parameters.
196
-------
FIELD VERIFICATION OF AIR EMISSION MODELS FOR HAZARDOUS
WASTE DISPOSAL FACILITIES
W. D. Balfour, C. E. Schmidt, R. G. Wetherold,
D. L. Lewis, J. I. Steinmetz, and R. C. Hanisch
Radian Corporation
Austin, Texas 78766
ABSTRACT
Field tests were conducted at four
hazardous waste treatment, storage, and
disposal facilities (TSDFs) in order to 1)
measure volatile organic compound air emis-
sion rates and 2) collect the necessary
data to estimate the air emission rates
using available predictive models. The
activities for which air emission rates
were measured included various surface
impoundments, active and inactive land-
fills, a landfarm, drum storage, and a
solvent recovery operation. Due to the
limitations of the existing air emission
models, calculation of predicted emissions
were limited to the surface impoundments
and landfarm.
A direct comparison of emission rates
measured using the concentration-profile
technique and the emission isolation flux
chamber was made based upon measurements at
a surface impoundment. For most compounds,
the flux chamber technique resulted in a
statistically significant greater air
emission rate than as measured by the
concentration-profile technique.
The Thibodeaux, Parker, and Heck air
emission model was used to predict
emissions from three surface impoundments.
Predicted emissions were compared to
emission rates for ponds 1 and 2 measured
using the flux chamber. No statistically
significant differences were determined
between predicted and measured emission
rates in half of the cases examined for
pond 1. In all other cases, the predicted
rate was greater than the measured rate
for pond 1. For pond 2, the predicted
rate was orders of magnitude greater than
the measured rate in all cases. This
discrepancy is attributed to problems in
modeling the sludge/oil/aqueous surface
encountered for this lagoon. Predicted
emissions were compared to emission rates
for pond 3 measured using both
concentration-profile and flux chamber
techniques. In general, the predicted
rates are statistically greater than those
measured fay the concentration-profile
technique and less than those measured by
the flux chamber. A predictive model was
not used to estimate emissions from the
inactive chemical landfills in light of
the heterogeneous nature of the waste and
inability of the model to account for
vented emissions.
197
-------
SOLIDIFICATION/STABILIZATION ALTERNATIVES FOR REMEDIAL
ACTION AT UNCONTROLLED HAZARDOUS WASTE SITES
Murdock John Cullinane, Jr.
USAE Waterways Experiment Station
Vicksburg, MS 39180
ABSTRACT
Typically, the solidification/
stabilization process involves the addition
of solid and/or liquid reagents to produce
a solid matrix with handling properties
similar to soil matrices.
Planning for the application of
solidification/stabilization technology to
a particular remedial action site is
divided into two distinct phases: process
selection and scenario selection,, The
process selection phase is primarily
concerned with the chemistry of the
solidification/stabilization process as
applied to the specific waste type; whereas
scenario selection is primarily concerned
with the development of equipment require-
ments, construction sequencing and cost
estimating for accomplishing the selected
solidification/stabilization process. The
selection of an appropriate scenario must
consider: safety and environmental
concerns, available equipment, purchase
and handling of solidification/
stabilization chemicals, and handling of
waste materials.
A review of current field practices
indicates that four basic field scenarios
have been successfully applied to the
solidification/stabilization of waste
liquids and sludges. These scenarios
include in-drum mixing, in situ mixing,
plant mixing, and area mixing. Various
modifications to each of the above basic
scenarios have also been developed.
Each scenario is documented and
information is presented on equipment
requirements, operating characteristics
and capacities. A methodology for
estimating the unit cost associated with
implementing each scenario is also
developed. A flow chart for selecting and
implementing an appropriate process and
scenario is developed and illustrated.
The development and selection of the
solidification/stabilization operations
plan for a particular remedial action site
is dependent on several factors such as the
nature of the waste material, the quantity
of the waste materials to be treated, -the
location of the site, the physical
characteristics of the site, and the
solidification/stabilization process to be
utilized. As a result, no conclusions are
drawn as to the selection of a particular
scenario for a specific remedial action
project; however, the attributes of each
scenario are developed and presented,,
198
-------
PREPARE TECHNICAL RESOURCE DOCUMENT ON DESIGN, CONSTRUCTION, MAINTENANCE,
AND EVALUATION OF CLAY LINERS FOR HAZARDOUS WASTE FACILITIES
Leonard J. Goldman, Garrie L. Kingsbury, C. Clark Allen, Robert S. Truesdale,
Coleen M. Northeim, Scott M. Harkins, Ashok S. Damle,
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
and Mike H. Roulier
U.S. Environmental Protection Agency
26 W. St. Clair Street
Mail Stop 472
Cincinnait, OH 45268
ABSTRACT
The purpose of this project is to
develop a Technical Resource Document (TRD)
on clay liners for landfills, surface
impoundments, and waste piles. The TRD
will be a summary of the current
technologies and evaluation techniques
applicable to these facilities. Critical
factors in the design, construction, and
maintenance of compacted clay liners will
be investigated and addressed.
Major research and development
projects related to clay liner performance
will be identified and their results
compiled on an ongoing basis. Relevant
information that is obtained through this
data-gathering effort will be presented in
the final TRD.
Another aspect of this project will be
the presentation of the state-of-the-art
knowledge and techniques currently used for
clay liner design and installation. The
project team will contact selected
engineering firms, waste management
companies, and professional associations to
collect information on current practices
for designing, installing, testing, and
maintaining clay liners. Problems
encountered during installation will be
highlighted.
A data base documenting the
performance of installed clay liner
systems will be developed. Relevant
information will be obtained from both
governmental and private sources.
The influence of waste liquid
composition on infiltration rate into
partially saturated clay liners will also
be studied. This analysis will, in part,
be based on results of laboratory tests
performed with various clay and liquid
waste systems. In a closely related
area, an effort will be made to document
the accuracy achievable and the effort
required to use available methods for
predicting transit time of leachate in
partially saturated compacted clay
liners.
Research reports and the TRD Draft
for public comment are scheduled to be
completed in May 1985.
199
-------
APPLICATIONS OF ENGINEERING FABRICS
AT LAND WASTE DISPOSAL SITES
Raymond C. Horz
U. S. Army Engineer Waterways Experiment Station
Vicksburg, Mississippi 39180
ABSTRACT
Engineering fabrics have proven to be
cost effective materials for solving
numerous drainage and stability problems
in geotechnical engineering, and they can
potentially be utilized to solve similar
problems in the containment and disposal
of solid and hazardous waste. A variety
of applications of engineering fabrics is
presented which utilizes the filtration,
drainage, separation, and reinforcement
capabilities of these materials. The
applications addressed include
(1) construction of leachate collection,
and groundwater control systems,
(2) protection of waste covers from surface
erosion, (3) drainage blankets .in landfill
covers and waste disposal lagoons, (4) gas
venting and puncture protection beneath
synthetic membranes, (5) permanent capping
of soft waste sediments, and
(6) construction of haul roads and
temporary working platforms on soft
sediments. In each of these applications,
design considerations arid techniques,
fabric requirements, and construction
techniques are discussed. Model
specifications for fabrics in the various
applications are given. For drainage
systems and erosion control, criteria for
the selection of fabrics based on the
fabric's piping resistance and clogging
resistance are presented. Strength
requirements based on the severity of
the construction environment and long-term
chemical/biological degradation are
addressed. Fabric selection criteria are
given for fabrics used for haul roads, for
reinforcing dikes, and as a component of
caps for soft sediments. Construction
techniques are described for each of these
applications.
Important physical, hydraulic, and
endurance properties of fabrics are
discussed. Tensile strength, modulus,
and puncture resistance, as measured by
the grab tensile test, the wide width
tensile test, and the Corps of Engineers
puncture resistance test, are emphasized
as being the most practically applicable
physical properties. Other physical
properties such as creep susceptibility,
tear resistance, fractional and pull-out
resistance with soil, and seam strength
are also discussed.
200
-------
LEACHATE PLUME MANAGEMENT
Charles Kufs, and Edward Repa
JRB Associates
McLean, VA 22102
To successfully control leachate
plumes from uncontrolled hazardous waste
sites, a thorough understanding of plume
dynamics and aquifer restoration tech-
nologies is essential. To address this
need, a handbook has been developed which
provides an overview of the fundamental
concepts, procedures, and technologies
used in leachate plume management. The
handbook consists of eight chapters:
o Chapter 1 — Introduction: Overview
of plume generation and summary of
the handbook's contents.
Chapter 2 — Plume Dynamics: ^^^^.^
of groundwater flow patterns, leach
characteristics, and plume/geologic
characteristics,
media interactions
Effects
leachate
Chapter 3 — Plume Delineation: Data
sources and procedures for estimating
plume boundaries, characterizing plume
chemistry, and extrapolating future
plume movement
Chapter 4 — Control Technology Selec-
tion: Overview of plume control tech-
nologies and evaluation and selection
criteria for site applications.
Chapter 5 — Groundwater Pumping:
Well hydraulics, design, installation,
operation, and maintenance.
Chapter 6 — Subsurface Drains: Drain
hydraulics, design, installation, and
maintenance.
Chapter 7 — Low Permeability Barri-
ers : Barrier materials, placement,
design, installation, and maintenance.
Chapter 8 — Innovative Technologies:
Bioreclamation and in situ chemical
treatment technologies.
Groundwater pumping technologies
involve the extraction or injection of
water, through wells, to alter a plume's
movement. Extraction wells create a
cone of depression causing groundwater
to flow toward the well. A cone of im-
pression is created around an injection
well causing groundwater to flow away
from the well. Groundwater injection
and extractions wells can be used sepa-
rately, or in combination, to change the
flow of groundwater in order to contain
a plume.
Subsurface drains are continuous,
permeable barriers designed to intercept
groundwater flow. They can be used much
like wells, especially in low permeabili-
ty aquifers. Groundwater collected in
drains or wells generally must be pumped
to a treatment system. This additional
water can increase the treatment facili-
ties operation costs substantially.
Low permeability barriers consist
of a wall or other appropriate configura-
tion of low-permeability material con-
structed underground for the purpose of
isolating contaminants or diverting
groundwater flow. Slurry trench barriers
are used to contain contaminant plumes
by completely surrounding a waste site,
capture insoluble contaminants migrating
down gradient, or lower a groundwater
table to prevent contact with wastes.
Grout barriers can be constructed simi-
larly to slurry trenches but high costs
and technical difficulties generally re-
stricts the use of grouts to sealing
fractures and other large openings in
aquifers.
In addition to these more establish-
ed techniques, innovative technologies
are being developed to control leachate
plume movement. These include in situ
biological and chemical treatments. "
201
-------
TECHNICAL HANDBOOK FOR DESIGN AND CONSTRUCTION OF COVERS
FOR UNCONTROLLED HAZARDOUS WASTE SITES
Colin C. McAneny
U.S. Army Engineer Waterways Experiment Station
Vicksburg, Mississippi 39180
ABSTRACT
This technical handbook provides
guidance to the cover-system designer and
technical information for regulatory per-
sonnel. A properly designed and main-
tained cover system reduces percolation
and leachate formation, prevents contam-
inated dust emissions and contaminated
surface runoff, and results in rapid
aesthetic improvement to a site.
Most materials used in any cover
system are likely to be soils, which are
best classified by the Unified Soil Class-
ification System (USCS). Soils may be
stabilized or modified by the use of soil
additives such as lime, bentonite, portland
cement, fly ash, or bituminous materials.
Other applicable materials in a cover sys-
tem may include impermeable substances such
as asphalt and synthetic "geomembranes,"
and permeable substances such as various
waste or residual materials, pipes and
tiles, and synthetic "geotextiles." Sig-
nificant properties of all these materials
are discussed in the handbook, and tests
and criteria for them are described or
referenced.
Proper cover-system design requires
decisions regarding design life of the
system, tolerable quantities of leachate,
and future use of the site. To estimate
the amount of leachate expected from a
given design, it is necessary to perform
a water-balance analysis.
To provide for various functions and
attributes, the most efficient cover system
is one composed of layers. From bottom to
top, the layers in a cover system may be
as follows: a gas-permeable layer to con-
trol gases produced by organic decay within
the wastes; one or more filter layers where
materials of contrasting gradation are
placed in direct contact with one another;
a foundation or buffer layer to serve as a
platform on which to build the rest of the
system; a hydraulic-barrier layer, composed
of either impervious natural soils, amended
soils, or a geomembrane; a drainage layer,
to intercept percolating water and provide
a path for it to a disposal outlet; a
biotic barrier layer to stop animals and
plant roots from penetrating and disrupting
the hydraulic barrier; and a surface layer
to provide for vegetative support, stabil-
ize the surface against erosion, aid in de-
watering the cover through evapotranspira-
tion, and provide an aesthetically pleasing
appearance. An effective surface water
management plan is essential to control
surface runoff.
Good construction is as important as
good design. The handbook discusses methods
and equipment used in cover construction,
and organization and practices important
for construction quality control.
Two case histories of existing cover
installations at hazardous waste sites are
presented to demonstrate some of the cover
design concepts described in the handbook.
202
-------
TESTING AND EVALUATION OF PERMEABLE MATERIALS FOR THE REMOVAL OF POLLUTANTS FROM
LEACHATES AT REMEDIAL ACTION SITES - BENCH AND LABORATORY TESTING
James E. Park
University of Cincinnati, T&E Facility
Cincinnati, OH 45268
ABSTRACT
Conventional practices for controlling
pollutant migration at remedial action
sites seek to restrict ground water flow.
Impermeable barriers such as slurry walls
and grout curtains, ground water pumping
wells and/or surface drains can restrict
contaminant movement. These pollutant
migration management techniques, however,
do not mitigate their toxicity.
"Permeable barriers," using materials
that can reduce the concentration of con-
taminants in solution by sorption, precip-
itation or other mechanisms may be a viable
addition to the alternatives available for
remediating a hazardous waste site. Per-
meable materials would allow the ground
water flow to be maintained while the
pollutant load is decreased.
There are many materials that might be
used for this purpose. Activated Carbon
and man-made zeolites have been extensively
tested in pollution control and industrial
treatment operations. These materials,
along with many synthetic, commercial
sorbents would seem appropriate, however
the quantities required at most sites, may
make them cost prohibitive.
Natural materials such as limestone,
coal and clay, and waste products such
as fly ash may offer some capability in
removing contaminants from leachates.
They also are comparatively low in cost
and are generally available throughout the
country. It is unlikely that any one of
the four materials mentioned above has the
ability to retain all of the possible con-
taminants that may be encountered at a
remedial action site. This project will
examine multiple layers of available,
inexpensive materials to evaluate the
effectiveness of various layer orderings,
appropriate layer*thicknesses and the
individual materials' abilities to.retain
a leachate's contaminants.
If the results are sufficiently posi-
tive, they will be used to develop a
protocol for testing these and other suit-
able materials with actual leachates. The
intention of such a protocol will be to
provide site-specific data and design in-
formation for the use of a "permeable"
barrier at a remedial action site.
The Technical Project Monitor for this
work is Mr. Jonathan G. Herrmann, U.S.
Environmental Protection Agency, Municipal
Environmental Research Laboratory, Solid
and Hazardous Waste Research Division,
Cincinnati, Ohio.
203
-------
IN SITU TREATMENT TECHNIQUES APPLICABLE TO LARGE QUANTITIES
OF HAZARDOUS WASTE CONTAMINATED SOILS
Ronald C. Sims
Utah Water Resource Laboratory
Utah State University, Logan, Utah
Kathleen Wagner
JRB Associates
McLean, VA
ABSTRACT
In situ treatment may be a practical and
economical approach to the treatment of
large volumes of hazardous waste con-
taminated soil. A literature study and
evaluation was conducted to determine
in situ treatment techniques which may
be used as alternatives to excavation,
hauling, and disposal of contaminated
soil. Specific site scenarios addressed
for in situ treatment included low level
contamination around the periphery of
a site, in the hot or contaminated zone
after the bulk of contaminant removal,
or across a broad terrestrial system.
Three alternative treatment schemes
were considered for in situ treatment.
These include: 1) management of the soil
system to utilize natural soil reactions
for treatment (degradation, detoxifica-
tion, or immobilization of waste constit-
uents); 2) addition of treatment agents
to the soil; and 3) washing or flushing
the soil to remove contaminants. Treat-
ment techniques evaluated and selected
for the first two schemes involve sorp-
tion processes to control mobilization;
biological processes to control degrada-
tion and detoxification; and chemical
and photochemical processes to control
detoxification, degradation, and immobi-
lization. Soil flushing, involving the
elutriation of contaminants from soil,
followed by treatment of the elutriate
may also be used.
A total of forty in situ treatment
techniques were identified. Fourteen
techniques for treatment of hazardous
inorganic constituents were evaluated
including precipitation, organic matter
addition, chelation, reduction and oxida-
tion, sorption by soil, activated carbon,
zeolites and soil flushing. Twenty five
techniques for treatment of hazardous
organic constituents were evaluated in-
cluding: 1) immobilization through soil
moisture control, soil incorporation,
and sorbent addition; 2) chemical treat-
ment through soil catalyzed reactions,
addition of oxidizing and reducing
agents, and soil pH adjustment; 3) bio-
logical treatment including, management
of aerobic and anaerobic environments,
analog enrichment, and addition of exoge-
'nous acclimated or mutant microorganisms
and cell-free enzymes; 4) soil flushing;
and 5) control of volatilization through
physical and chemical barriers.
A methodology was developed for
selection and implementation of in situ
treatment methods based on: 1) character-
ization of the site, soil, and waste;
2) assessment of the environmental impact
of the waste; 3) treatment technique
selection; and 4) monitoring to evaluate
treatment success.
Results of the evaluation of in situ
treatment techniques that have been used
to treat hazardous waste contaminated
soil indicate that, although several
techniques appear promising, very few
have been field tested. It is therefore
necessary to conduct bench and site spe-
cific pilot scale testing of potential
treatment techniques to determine the
extent of applicability at CERCLA sites.
204
-------
CONTAMINANT FIXATION: PRACTICE AND THEORY
H. R. Lubowltz, R. W. Telles, B. H. Ellash, S. L. Unger
Environmental Protection Polymers, Inc.
Hawthorne, California 90250
ABSTRACT
Data from a literature review are used to describe and evaluate elements of hazardous
waste fixation. Factors are then Identified for producing high-performance stabilized
hazardous wastes, and some conclusions are drawn about the performance of current
stabilization techniques. Some suggestions are made for future research.
INTRODUCTION
Though the current trend 1n managing
hazardous waste 1s either to destroy or
reuse toxic residue,d) the formation of
some unmanageable waste Is usually
unavoidable.(5) such materials are
generally disposed of on the land by
broadcasting them 1n small concentrations
or by fixing them 1n large concentrations
within a binder that prevents or restricts
contaminant release Into the environment.
U.S. Environmental Protection Agency (EPA)
toxldty tests have shown that fixation
processes 1n current use may not, 1n fact,
achieve the high performance levels
claimed for themJ7) In addition,
future fixation products will be even more
difficult to manage with binders as
fixatives because Improved waste treatment
processes will yield wastes with higher
concentrations of contaminants. The
purpose of this paper 1s therefore to
examine contaminant fixation practices and
establish criteria for producing high-
performance fixation products.
Examination of these criteria and an
understanding of their scientific
principles can provide a basis for
predicting performance of stabilization
processes and products.
ELEMENTS OF HAZARDOUS WASTE FIXATION
There are several factors to consider
1n judging the expected cost and
performance or merit of a contaminant
fixation process using binders. These
Include (!) the waste to be fixed, (2) the
binder (fixative) used, (3) the type of
product formed (solid solution, chemically
fixed product, or mlcroencapsulate),
(4) processing techniques, and
(5) stability of the product under
expected environmental stress. The most
Important factors 1n judging performance
are Hems 3 and 5, the type of product
formed and Us stability. By carefully
examining these two factors, one can judge
the relative merits of any proposed
fixation process for managing any type of
waste. Information about each of these
was gathered 1n a literature review and 1s
discussed here briefly.
Wastes to be Fixed
The compositions of waste materials
are highly variable and often defy tight
definition. Wastes of concern, however,
are primarily hazardous Industrial
wastes. Those most amenable to
contaminant fixation by current
stabilization processes used are normally
produced 1n large amounts. They consist
mainly of Inorganic materials 1n aqueous
solution or suspension and contain large
amounts of heavy metals or Inorganic
salts. Organic wastes and waste streams
containing organic constituents are less
amenable to contaminant fixation by
state-of-the-art processes.
205
-------
Binders Used
A great variety of binders are used
for hazardous waste stabilization, Includ-
ing naturally-occurring clays, lime-based
cements, scrap plastics, and fly ash.
Such materials are economical, but their
performance properties under waste dis-
posal conditions may be highly variable.
These uncertainties along with those of
the highly variable waste feeds produce
fixation products whose performances may
not be predictable 1n the field. A key
factor 1n determining product performance
1s the stability of the binder. Even
though the stability of the binder alone
1s less than optimal, 1t may still be
acceptable depending on the type of
product 1t forms when combined with waste.
Binders (fixatives) are of two
types—organic and Inorganic. They may be
classified as follows:
1. Organic fixatives:
o Asphaltlc and bitumens
o Ureas and phenollcs
o Epoxldes and polyesters
o Vinyls
2. Inorganic fixatives:
o Cements and concretes
o Calcium compositions
o Silicates
o Sulfur compositions
Inorganic cements are the most
commonly used fixatives and are employed
mainly for fixing heavy metal contaminants
eventhough organic fixatives are generally
more chemically stable than Inorganic
ones. They are also more physically
stable, but as nonrelnforced materials
their load bearing properties may not be
comparable. Raw material cost
considerations and familiarity presently
favor the use of Inorganic fixatives.
Types of Products Formed by Fixation
Because waste and binder compositions
vary considerably, 1t Is difficult to use
these factors to predict expected
performance of the products 1n the field.
A better approach 1s to examine carefully
the types of end products formed with
hazardous waste fixation processes—solid
solutions, chemically fixed products, and
mlcroencapsulates. These products and
their merits with regard to contaminant
fixation are defined briefly as follows:
1. Solid solutions consist of
submlcroscoplc contaminant particles (most
likely 1n the molecular and colloidal size
range) dissolved 1n a solidified fixative
and characterized by weak chemical
Interactions such as Van der Waals
forces. Solid solutions should generally
be avoided because the weak chemical bonds
could permit contaminant loss through
diffusion. They may be acceptable 1n some
cases, however, If the permeates and the
fixative are such that they retard
contaminant diffusion. For examples,
solid solutions may contain-high molecular
weight contaminants and binders having
macromolecular structures characterized by
chain cross-links and entanglements.
2. Chemically fixed products consist
of submlcroscoplc contaminant particles
(most likely 1n the molecular and
colloidal size range) Incorporated 1n a
fixative and characterized by strong
chemical Interactions with the fixative
such as chemical bonds. Hazardous waste
products that are truly chemically fixed
are difficult to achieve but are highly
desirable 1n all cases since contaminant
molecules are expected to resist diffusion
because of their strong chemical bonds to
the fixative. Strong chemical bonds refer
to bonds with high dissociation energies
and Indicate, but not Invariably, high
bond resistivities to chemical
dissociation (e.g. hydrolysis). The great
merit of chemical fixation 1s that the
fixative can hold contaminants even though
the fixation product may fragment under
service conditions.
3. Hlcroencapsulates consist of
microscopic and macroscopic contaminant
particles Individually encased 1n a
fixative that does not significantly
affect the contaminant molecules because
they are unavailable for Interaction.
Mlcroencapsulatlon can yield
high-performance contaminant fixation If
the products maintain their Integrity
under environmental stress.
206
-------
The merit of mlcroencapsulates under
service conditions can be estimated by
monitoring their surface areas or (at
least) of their dimensions. An Ideal
mlcroencapsulate will exhibit no change 1n
surface area, whereas that of a less
acceptable mlcroencapsulate will degrade
and may provide a path for contaminant
release.
Processing
Processing techniques for fixing
hazardous wastes Involve mixing 1n an
attempt to homogeneously disperse the
contaminants throughout the fixative.
Thus the nature of the fixative largely
determines the type of processing
required for blending and Isolating the
product.
Inorganic fixatives are processed
mainly at ambient temperatures and use
water to effect their consolidation. In
some cases, water Is Immobilized by the
fixative without being chemically
Incorporated.
The processing of organic fixatives
may require greater than ambient
temperatures and may or may not
require water. The presence or
absence of water In processing Is
Important to determining the
performance levels of the final
product, since water Is the vehicle by
which all contaminants but gaseous ones
are released and dispersed 1n the
environment. Without water contact,
contaminants remain fixed.
Stability and Performance of Fixation
Products In the Field
Fixation product performance Is
assessed by considering the following:
1. The size of the contaminant
particles 1n the fixative
(submlcroscoplc versus microscopic
and macroscopic),
2. Orientation and strength of the
Interfadal forces acting on
submlcroscoplc particles
(nondlrectlonal and weak, as 1n
solid solutions, versus
nond!rect1onal/dlrect1onal and
strong, as 1n chemically fixed
products), and
3. Accessibility of leachate to the
contaminant particles (limited or
constant surface area of the
product versus increasing surface
area and eventual fragmentation of
the product under service
conditions).
Contaminant Particle Size
The particle size of the contaminant
and how 1t is dispersed In the fixative
greatly affects the ability of a process
to fix that contaminant. Submlcroscoplc
particles are affected much more than
microscopic or macroscopic particles by
the chemical forces at the contaminant
particle interfaces. For example,
submicroscopic contaminants are released
by leachate at the wetted surface of the
fixed product. More contaminant particles
then diffuse from the interior to the
surface in response to the decreased
concentration gradient leading to the
surface. The diffusion is enhanced by
high concentrations of submlcroscoplc
particles.
Microscopic and macroscopic particles
are also released by contact with
eachate at the wetted surface of the
product. But diffusion of these
contaminants through the product is
Impeded because the particles are large
and can be insoluble in the fixative.
Thus, current fixation products seem
functional, but eventually the fixative
moiety itself breaks down, exposing more
particles to leachate and accelerating the
loss of contaminants.
Our work Indicates that most fixatives
in current use stabilize contaminants
through m1croencapsulat!on rather than
through chemical fixation or solid
solution. Microencapsulatlon is the main
vehicle for contaminant fixation even
though chemical reactions may occur
between contaminants and fixatives. A
truly chemically fixed product Is not
likely to be produced 1n current practice
because the required dissolution of the
reaction products by the fixative 1s 1n
reality not probable under the processing
conditions generally employed. Instead,
the processing produces microscopic and
macroscopic particles that are entrapped
(encapsulated) as the fixative
solidifies. Such products should be
considered mlcroencapsulates.
207
-------
Orientation and Strength of Interfaclal
Chemical Forces
Weak, nonchrectlonal chemical
Interfaclal forces characterize solid
solutions, and strong directional or
nondlrectlonal forces characterize
chemically fixed products. Thus the
submlcroscoplc contaminants 1n solid
solutions are much more readily released
from the fixative to the leachate than are
those 1n chemically fixed products. These
forces affect the stability of microscopic
and macroscopic contaminant particles only
minimally. They are Important, however,
In estimating the mechanical performance
of the products.
Accessibility of Water to Contaminant
Particles
Accessibility of water to fixation
products depends on the product's surface
area and the extent to which that area 1s
wetted. A preferred product 1s one that
maintains a microscopic surface area that
approximately equals the geometric area of
the monolith, or at the very least, one
that has a constant surface area or
becomes constant before the mechanical
properties of the product are
significantly affected.
Thus the effectiveness of a preferred
product depends on minimizing the surface
area, which Is the site for contaminant
release and transfer to the environment.
Some preferred products perform better
than others. In a solid solution, for
example, the contaminant 1s dissolved 1n
the fixative and thus 1s available at all
the wetted surfaces of the product. But
1n a preferred mlcroencapsulate the
contaminant Is physically encased and not
dissolved. Thus only those contaminants
In particles actually at the surface are
released. The contaminant In a preferred
mlcroencapsulate Is not available at the
surface except for a relatively few
particles that remain exposed after
processing. Further release of
contaminants would not occur as long as
the particles remained Insoluble 1n the
fixative and the product remains Intact.
A nonpreferred product, on the other
hand, shows continual Increase In Us
surface area under service conditions and
may finally disintegrate. Processes that
yield nonpreferred products can be used
for waste stabilization 1f contaminants are
chemically fixed to the binder. In such
cases, even though the surface area may not
remain constant, the product can be useful
because of the appreciable time needed for
the monolith to disintegrate Into water
entralnable particles. But 1f the product
Is not a chemically fixed one, a nonprefer-
red product exposes contaminant particles to
leaching In a relatively short time through
the ready formation of flaws, cleavages,
capillaries, etc. In the monolith.
The type of water action on the product
1s also Important—static, flowing, etc.
Static action can activate release of
contaminants because of salting In effects,
or 1t can retard It by common Ion effects.
Even under leachate flow, static water
exists 1n flaws, cleavages, and
capillaries. This static water can also
Induce chemical and physical changes 1n the
product's surface, which may then perhaps
transpire contaminants more readily than
possible with contact by virgin leachate
only.
PROBLEMS IDENTIFIED IN PRODUCING STABILIZED
PRODUCTS
As mentioned, our observations Indicate
that state-of-the-art fixation products 1n
current use are really mlcroencapsulates.
As such, contaminant release occurs mainly
at the interface of microscopic and
macroscopic particles with water
(leachate). Water normally contacts the
contaminant only at the product surface and
In Its flaws. However, enough water may be
used In the process of consolidating some
fixatives so that 1t characterizes the
matrix as 1t does In a gelJ8' 13) This
water may solublUze contaminants and thus
promote contaminant diffusion and
additional contaminant loss.
Solid solutions and chemical fixation
products were not readily available for
study because state-of-the-art fixation
products generally hold microscopic) and
macroscopic Insoluble heavy metal contam-
inants. Exceptions occurred only 1n high
temperature sintered products. In the
future, however, organic contaminants may
be dissolved In organic fixatives to
produce solid solutions. This procedure
will not be easy to carry out. An
appreciably more difficult task would be to
attach the contaminant chemically to the
fixative.
208
-------
CRITERIA FOR PRODUCING HIGH-PERFORMANCE
STABILIZED PRODUCTS
Based on our work and the IHerature
review conducted for this study, we have
established the following criteria for
producing high-performance stabilized
products:
1. Fixatives (binders) selected for
use with hazardous wastes must be
chemically and d1mens1onally
stable under expected service
conditions, both before and after
the waste has been Incorporated.
For example, asphalt binders would
not be expected to provide
dimensional stability to a
solidified waste even though such
materials are chemically stable.
2. For a chemically fixed product,
processing should disperse
contaminants throughout the
fixative 1n the molecular or
colloidal state, and H should
Incorporate contaminants 1n that
state by strong chemical forces
within the matrix rather than by
weak chemical bonding forces.
That 1s, the contaminant must be
chemically grafted to the
molecular structure of the
fixative.
3. For a mlcroencapsulate, encasement
of the microscopic and macroscopic
contaminant particles by the
fixative must be as complete as
possible. Ideally, 100% of the
particles would be completely
encapsulated. Careful processing
conditions are required to achieve
this goal, however.
4. Stabilization of mlcroencapsulates
should be enhanced by minimizing
the surface area of the product
potentially exposed to leaching
solutions.
5. Formation of leachate Insoluble
particles would enhance the
stability of mlcroencapsulates.
CONCLUSIONS
Our review of the IHerature Indicates
that the mechanism predominantly used for
fixing contaminants 1n available processes
1s actually microencapsulatlon and not
chemical fixation^6), as claimed by
several manufacturers. Though some
fixatives (e.g., silicates, lime, and
cement) were believed to cause molecular
dispersion of contaminants and to effect
chemical fixation,(3« 4« 10« 12) our
work Indicates that only minimal amounts
of contaminants are actually fixed 1n
molecular or colloidal dispersion.
In fact, fixing agents react with
heavy metal contaminants and cause
precipitation of microscopic and
macroscopic particles that are presumed to
resist dissolution by water
(leachate). (2) Waste 1s sometimes
pretreated to condition contaminants to be
reactive with the fixative.(9) When the
product 1s consolidated, the particles are
locked Into the fixative (I.e.,
encapsulated). Thus even 1f the
consolidated products are water permeable,
the contaminants will not be released
because they are In an Insoluble state.
There are means other than
precipitation to form water-Insoluble
compounds. Wastes can be treated with
additives to produce contaminant particles
that resist water action. The preferred
approach Is to cause chemlsorptlon of
contaminants onto leachate-stable
particles.HI) This approach 1s,
nevertheless, Inferior to chemical
fixation because the chemlsorbed particles
may be small enough to be subjected to
water action.
We believe that fixation products
holding heavy metals, In general, are
acceptably stable only when contaminant
particles are all completely encased In
water-tight compartments (I.e.,
mlcroencapsulated). The reason for this
conclusion 1s that leachates specified 1n
the regulatory guidelines for assaying
contaminant stabilization dissolve the
heavy metal oxides and silicates that are
believed to resist dissolution. We found
1t nearly Impossible 1n most cases to form
heavy metal compounds that were stable
when contacted with the specified
leachates.
We are currently Investigating means
for stabilizing mlcroencapsulated
contaminants 1n organic binders so that
total encasement of particles would be
unnecessary and future processing demands
209
-------
would be less rigorous. One approach Is
to reduce the soluble Ionic character of
heavy metal compounds by Introducing
dissolution-resisting covalency Into the
composition. Thus, we view polysulfldes
and organo-sulfldes as being more suitable
than sulfldes, oxides, and silicates In
producing leachate-stable heavy metal
contaminant particles.
Acknowledgements
This work has been conducted by EPP,
Inc. for the Solid and Hazardous Waste
Research Division of EPA's Municipal
Environmental Research Laboratory
(Cincinnati, Ohio) under Contract
68-03-2993. Mr. Carlton C. Wiles Is the
EPA Project Officer. The authors wish to
thank Ms. Anne Hamilton for her editorial
assistance.
References
1. Anon., "Land Disposal Hay be Grinding
to a Legislative Halt," Chemical Week,
Vol. 132, No. 25, June 22, 1983,
pp 10-11
2. Bishop, Paul L., Stephen B. Ransom,
"Fixation Mechanisms 1n
Solidification/Stabilization of
Inorganic Hazardous Waste," 38th
Annual Purdue Industrial Waste Conf.,
May 10-12, 1983, 15 pp.
3. Conner, Jesse R., "Ultimate Disposal
of Liquid Wastes by Chemical
Fixation", Purdue Univ. Engineering
Extension Series, Engineering Bulletin
1974, Pt. 2, pp 906-922.
4. Halladay, D. W. (Oak Ridge National
Lab., Oak Ridge, TN) Report 1978,
ORNL/TM-5757, 105 pp., Avail NTIS
Energy Res. Abstr. 1978 3(16), Abst.
No. 36949.
5. Josephson, Julian, "Immobilization and
LeachabHlty of Hazardous Wastes,"
Environ. Sc1. Techno!., Vol. 16,
No. 4, 1982, p. 219A.
6. Lubowltz, H. R., R. W. Telles, "Waste
Management by Fixation Processes,"
Unpublished Report submitted to EPA
Cincinnati, Contract 68-03-2993,
March 1981, 71 pp.
10.
11,
12.
13.
Malone, Phillip G., Richard B.
Mercer, Douglas W. Thompson, "The
Effective- ness of Fixation
Techniques 1n Preventing the Loss of
Contaminants from Electroplating
Wastes," First Annual Conf. on Adv.
Pollut. Control for Metal Fin.
Indus., EPA-600/8-78-010, May 1978,
pp. 130-144.
Moffatt, William G., George W.
Pearsall, John Wulff, The Structure
and Properties of Materials, Vol. 1,
John Wiley & Sons, NY, 1964,
pp. 116-121.
Rousseau, J. M., and A. B. Craig,
Jr., "Stabilization of Heavy Metal
Wastes by the Sollroc Process," Proc.
3rd Annual Conf. on Adv. Pollu.
Control for Metal Finish Ind, NTIS,
PB81-191900, February 1981, pp. 70-75.
Rysman de Lockerente, Sege, and
Norbert Van de Voorde, "A Procedure
for the Treatment of Waste and
By-Products," Request for Letters of
Patent, "Soclete Internationale de
Publlcite et d'Agences Commerlcales,"
No. 11234, 1978, 18 pp.
Sakuma, S., et al, "Treatment of
Waste Polychlorlnated Blphenyl
Solution," Japan Kokal 78 101 847,
1978.
Schofleld, John T., Toxic and
Hazardous Waste Disposal, Vol. 1,
Ann Arbor Science, Ann Arbor, HI,
1979, p. 306.
Sneed, M. C., J. L. Maynard, General
Inorganic Chemistry, D. Van Nostrand
Co., NY, 8th ed, pp. 779-780.
210
-------
MULTIPLE SOIL LAYER
HAZARDOUS WASTE LANDFILL COVER:
DESIGN, CONSTRUCTION, INSTRUMENTATION,
AND MONITORING
Richard C. Warner, James E. Wilson, Nathaniel Peters
Agricultural Engineering Department
Harry J. Sterling
Civil Engineering Department
University of Kentucky
Lexington, Kentucky
40546-0075
Walter E. Grube
Municipal Environmental Research Laboratory
United States Environmental Protection Agency
Cincinnati, Ohio
45268
ABSTRACT
The design, construction, instrumentation, and monitoring of three multiple soil layer
landfill covers are described in detail. The major scope of this research is to ascertain the
hydrologic balance of such a cover with emphasis placed on soil moisture movement through-
out the layered soil profile. The experimental design allows for the evaluation of numerous
alternative design options. These include alternative pervious drainage spacing and leachate
collection systems. . The construction phase of this project is presented in detail, encompas-
sing a comprehensive analysis of the clay, sand, and topsoil layers, as well as drainage and
subsidence provisions of the experiment. A multi-layer, multi-sensor soil moisture system
employing tensiometers, resistance blocks, thermocouples, and the neutron probe is presented
with emphasis placed on installation without disturbing the effects of construction equipment
or the integrity of the multi-layered cover. A tipping bucket system was designed to measure
surface and subsurface drainage to within ±0.5% of the actual discharge. Calibration of re-
sistance blocks yields a log/linear relationship between resistance and soil moisture content.
The multi-sensor approach provides a comprehensive data base for multi-layer cover model-
ing used to evaluate alternative designs.
INTRODUCTION
Project Scope
The overall scope of this project is
to design, construct, instrument and monitor
a multiple soil layered hazardous waste
landfill cover. The objectives are twofold:
(1) delineate the procedure used to con-
struct a multi- layered cover that meets
current regulatory requirements, and (2)
ascertain the hydrologic balance of such a
cover with emphasis placed on soil mois-
ture movement throughout the layered soil
profile. Phase II of this project includes a
more extensive period of data gathering and
model applications to extend results to
evaluate alternative multi-layered cover
operations/designs.
Site Description
Tyrone, Kentucky is the location of
this project. The research site is located
22 miles southwest of the University of
Kentucky at Lexington, Kentucky. Selection
of the site was based upon: (1) a required
natural land slope of less than four percent
in order to minimize construction costs, (2)
an acceptable source of clay and topsoil
materials, and (3) convenient administrative
and logistic aspects.
211
-------
PURPOSE
Synopsis of Multi-Layered Design
The purpose of this section is to
provide a capsule overview of the multi-
layer landfill design and in-place instru-
mentation. A thorough discussion of design,
construction, and instrumentation procedures
is presented later.
Dimensions
Three multiple soil layered landfill
covers were constructed in order to moni-
tor the hydrologic balance of such a sys-
tem. Each cell can be viewed as a slice
from a large hazardous waste site. An
overview schematic is shown in Figure 1
illustrating the major features of the cells.
Each cell was constructed on a three per-
cent grade, 27 m in length and 6 m wide.
Overall cell height is 3 m. The layers from
bottom upward are: (1) protective sand
layer (0.3 m), (2) PVC 20 mil liner to
prevent moisture movement in or out of
cell, (3) sand to be augered out during the
subsequent hydrologic effect of subsidence
studies (0.9 m), (4) compacted clay liner
(0.6 m), and (5) sand (0.6 m), and topsoil
(0.6 m) (Figure 2)
Drainage
Perforated FVC pipe was installed at
six locations in each cell. Lengthwise,
drainage pipe was positioned 9, 18 and 27
m from the upslope point. Pipes were situ-
ated at the bottom sand/plastic liner
interface and at the sand/clay interface
(Figures 2 and 3). Placement at these
positions enables determination of the
influence of either 9 or 27 m spacing as
well as leachate collection and/or reduction
of head on the clay layer due to lateral
removal of water along the top sand layer.
Both of these techniques have been em-
ployed in proposed landfill cover designs.
Also, surface drainage was provided by
three half-round PVC pipes located at
the 9, 18, and 27 m marks. All drains
were placed on a one percent slope toward
the observation pits.
Instrumentation
Flow from all twenty-seven drains is
directed to tipping buckets installed in the
observation pits. (Figure 3). The tipping buckets
are used to measure flow from the drainage
tiles. Since emphasis is being placed on
the movement (both upward and downward)
of soil moisture, several redundant instru-
mentation systems are used. Soil moisture
is being determined via: (1) tensiometers,
(2) gypsum (resistance) blocks, (3) thermo-
couple wire, and (4) the neutron probe. Six
each of tensiometers, resistance blocks
and thermocouples are installed near the
center of the cell above and below all soil
interfaces. Also, three access tubes are
located as shown in Figures 2 and 3.
APPROACH
Construction
Site Requirements
The site for the multi-layered landfill
project is at Tyrone , Kentucky. Prior to
actual landfill construction a gravel road
with appropriate subbase and drainage was
built. Two hundred-twenty volt, single
phase 200 amp service power was installed
at the field fabrication building and adjac-
ent research locations. A1.6 m high ten-
sil strength fence was provided under con-
tract to exclude cattle. The field fabri-
cation building was constructed to provide a
work area to construct and assemble re-
quired materials and instrumentation for the
landfill cells.
Preliminary Requirements
Excavation for the 3 m deep cells
required the removal of topsoil and 2.3 m
of limestone. - The limestone was removed
through use of a hoeram. This operation
required five weeks of continuous hoeram
usage and produced a rough grade of
approximately three percent. The hoeram
was also used to smooth the cell walls to
prevent 'highly irregular surfaces. Approxi-
mately 10,000 polyethylene sandbags were
filled and placed along cell walls to protect
the liner from sharp rocks and provide the
27 m by 6 m cell dimension. Also, these
sandbags were used to support the upper-
most sections of the cells that were located
slightly above the natural land grade.
--Next, the placement of Fibretex
fabric along the cell sides was installed to
provide additional protection from punctur-
ing the 20 mil PVC liner. Sand was placed
212
-------
Jrnulti-level-
multi-sensor
instrumental
sand bogs I 20 mil
PVC liner
Figure 1. Plan view of multi-layered cover cells
to protect the bottom of the liner from
rock fragments. Also, the sand afforded a
uniform three percent slope along the cell
base.
A single sheet of 20 mil PVC liner,
15.24 m by 30.48 m, was placed to com-
pletely encapsulate the sides, ends and
bottom of the landfill cell. The liner was
continuously checked to determine that its
water-tight integrity remained intact. On
the few occasions when small punctures or
tears were discovered patches were immed-
iately installed and visually inspected.
Subsidence Provisions
In order to provide the option of de-
termining the hydrologic consequences of
subsidence in a multi-layered cover access
for an augering system was provided. Three
15.24 cm diameter PVC pipes, 4 m in length,
were placed above sheets of marine plywood
located on a 2 cm cover of sand above the
plastic liner. Auger access pipes, shown in
Figure 1 , were placed to allow later inser-
tion of a guide sleeve and auger device.
This auger access system provides the op-
portunity to remove a known quantity of
sand, thus developing a given void space at
three desired locations along a cell. The
created space could simulate any size void
expected to occur in hazardous waste mater-
ials. Hence, studies of subsidence, clay
layer fracturing, etc., could be conducted to
measure associated water movement through-
out the multi-layered landfill cover.
Leachate Collection
As stated in the synopsis section of
this paper alternative leachate collection
facilities were installed across the cell at a
9 m spacing along the length of the cell
(Figure 1). Leachate collection was pro-
vided by PVC pipe which was drilled with
0.8 cm diameter holes spaced 2.5 cm apart.
Approximately 120 holes were drilled
per meter length, yielding the surface area
typical of tile drainage. Fibretex fabric
was wrapped around the drain pipe to pre-
vent -entrance of sand. The leachate
collection pipes were placed on a one percent
slope across the width of the cell. Compacted
213
-------
yfiX-Tr-^v-rtfAXl-oWrc-O^w^i-v. Snrroeo ».. u _ *£*:.
e^^PfPh7rr--^^
^?&w-5Xt>^v»« 33v! eiti>f»A. _t_ .. --vsvSt
I
5 cm diameter '
1 PVC leachate
collectors
sand bags
20 mil PVC liner
Figure 2. Cross section of multi-layer landfill cover (A - A')
Fibretex
/fabric
20 mil PVC liner
INSTRUMENTATION
PIT
J~~| "instrumentation
"-1 circuits
•pressure gages
resistance
(gypsum)
thermo- block
couple
Figure 3. Cross section of multi-layer landfill cover (B - B')
214
-------
clay and plastic were placed beneath
the collector pipe to prevent any down-
slope flow under the pipe. The pipes lo-
cated 9 m and 18 m along the 3% slope
may be plugged by pipe inserts such that
both 9 m and 27 m drain spacing can be
evaluated. Flow exiting the leachate col-
lection pipes is measured by automatic
recording tipping buckets.
Simulated Waste Layer
Nine-tenths meter of sand was placed
by a 0.44 m3 bucket skid-loader. A total
of 1 ,635 metric tons of sand was used
for sand layers on all cells. Ninety tri-
axial truck loads were delivered to the
research site. The depth of sand was used
to simulate hazardous waste such that the
material could be augered for the subsi-
dence study. The sand was covered with
10 cm of clay provided as an upper driv-
ing surface.
Instrumentation was located approxi-
mately 10 cm deep in the sand layer.
Note that soil moisture instrumentation
was placed above and below all soil layer
interfaces at the time of construction in
order to achieve a true measure of the
influence of construction equipment upon
soil moisture movement throughout the
layered soil profile. -Tensiometers, gypsum
blocks, and thermocouple sensors were
placed above and below each interface.
Lines from all instruments were placed
in conduit to prevent crimping from con-
struction equipment loads.
Description of Compacted Clay Liner
An excellent clay source was located
within 0.25 km of the cells. Extensive soil
testing was conducted: (1) Procter density,
(2) particle size analysis, (3) liquid limit,
(4) plastic limit, (5) plasticity, (6) per-
meability, and (7) soil moisture-tension.
Numerous trial tests were conducted using
a wheel-roller, sheeps-foot and a Dynapac
compactor to determine the ability to blend
respective clay lifts and produce a relatively
impervious clay liner. Documentation of soil
moisture and density for each lift was
determined by nuclear density, sand cone
test, and gravimetric samples taken through-
out each lift. Additionally, cores were
removed from each cell for in-lab analysis
of permeability and subsidence testing.
Description of Clay Layers for Cover
The two foot thick clay layers for
the landfill covers for cells 1 and 2 were
constructed using five lifts of a brown, silty
clay. The clay used, from a local borrow
pit, is a Type CH, based on the Unified Soil
Classification System, with a plastic limit
of 30, a liquid limit of 57, and a plas-
ticity index of 27. It has a specific gravity
of 2.76.
Construction initially involved placing
the clay across the whole surface of the
plot, tilling the soil to break up any large
clods, removal of small rocks which were
present in it, adding water when below the
optimum moisture content, and then using a
vibratory compactor to achieve the desired
density.
Field tests were conducted to deter-
mine lift height and the number of com-
pactor passes necessary to achieve 90%
compaction. Initial soil moisture ranged
from 22 to 27%. The maximum density
from the standard Procter test was found to
be 1525 kg/m3 at an optimum moisture
content of 25.7%. The average density from
Cell 1 was 1458 kg/m3 and for Cell 2
it was 1442 kg/m3 , and the average mois-
ture content was 29% for Cell 1 and 28%
for Cell 2 (Figure 4).
A soil moisture of approximately 28%
was selected for actual compaction. This
above optimal soil moisture represents a
compromise between density and plasticity.
A higher than optimum soil moisture allows
for greater plasticity which may be needed
for later cover subsidence. It was found
to be difficult to incorporate the desired
soil moisture throughout the lift profile.
Of special concern was the ability to
blend or mesh lift layers together. It was
found that no matter how many -equipment
passes were made, the two lifts oftentimes
contained a slight separation. This problem
was resolved by rototilling the following
clay lift prior to compaction. The roto-
tilling operation roughened the prior com-
pacted layer and eliminated or greatly
reduced the separation problem. Also, this
procedure assisted in thoroughly incorpor-
ating soil moisture throughout the lift pro-
file.
215
-------
no
too
z
uj
Q
80
cc
a
70
T
Maximum
Density =
_95,2pcf
Optimum Moisture =
25.7 %
15
4O
20 25 30 35
MOISTURE CONTENT (%)
Figure 4. Procter Moisture Density Curve
Clay placement was accomplished by
five lifts approximately 15 cm in height. A
Dynapac CA12PD self-propelled, vibratory
compactor, which has an operating weight
of 3254 kg, was used for compaction. It
has 72 drum pads, each with a contact area
of 87.1 cm2, arranged in a chevron pattern.
The vibration system is capable of producing
103,356 - 207/H4 kg/mz of contact pres-
sure. Approximately 60 - 70 passes per lift
were required to produce the desired
compaction.
In order to reduce the possibility of
damage to the PVC liner the skid-loader
was filled with sand and used to compact
nearby all side and end walls. Also, a
plastic strip was fastened along all sides to •
act as an anti-seep collar. The strip,
extended approximately 15 cm into the cell,
was used to preclude side wall seepage.
A mobile soil testing lab was re-
ceived on loan from the Engineering Trans-
portation Research Laboratory at the Uni-
versity of Kentucky. This facility enables
rapid determination of soil moisture. The
standard oven drying procedure was replaced
by a microwave oven to obtain a rapid re-
sponse. Six nuclear density tests were
taken per lift at both the 10 and 15 cm
depths. Ten cylinder tests and two sand
cone tests were also, employed for each lift.
A drill rig was supplied by the Civil
Engineering Department to remove cores from
each cell for in-lab permeability analysis
and subsidence studies. The permeability of
a Type CH clay under optimum_gCompact[og
is expected to range from 10 to 10
cm/sec. Thus, the perspectives for a highly
impervious clay liner are excellent.
An apparatus designed specifically for
the measurement of permeabilities of fine-
grained soils (clays) has been constructed
(1983) in the Daniel V. Terrell Soils Lab-
oratory of the Department of Civil En-
gineering at the University of Kentucky.
The apparatus is constructed primarily of
stainless steel components, is capable of
operating at pressures up to 100 psi, uses a
triaxial type of chamber with upflow of the
permeant, and has computerized data
collection and analysis. These features have
been incorporated due to some of the in-
herent difficulties with performing per-
meability tests on clays. Some typical
problems are air entrapment in the soil
sample, long test times, and having high
back pressure. Automated data collection
help to alleviate these problems. Three
samples can be tested simultaneously.
The permeabilities of the clays taken
from the covers will be tested using this
apparatus. In addition, samples have been
obtained from which horizontal permeabil-
ities can be determined. Recompacted
samples will also be tested to ascertain the
effect of different moisture contents and
densities on the permeability. Some field
tests are planned and will be compared to
the laboratory results. This information
will be provided for input to the models
being used to describe the hydrologjc ef-
fectiveness of the landfill cover.
The Pervious Layer
Perforated PVC drains, wrapped in
Fibretex fabric, were installed on top of
the clay layer. These drains are employed
to increase water movement along the
pervious sand layer. As with the leachate
collection system, the first two drains,
located at the 9 and 18 m marks, can be
plugged and the efficiency of a 27 m drain
spacing can then be evaluated. Another
alternative would be to plug all drains,
which would allow evaluation of saturated
conditions.
A 0.6 m sand layer was placed above
the three drains. This pervious layer is
used to relieve pressure above the clay
layer. As before, multi- sensor instrumen-
tation was located above and below the
clay/sand interface. All sand was placed by
the skid-loader.
216
-------
Surface Layer
Instrumentation was located above
and below the sand/topsoil interface. The
topsoil was placed by the skid-loader in
approximately one 0.6 m lift and tilled to
prepare a seedbed. Care was taken not to
compact the topsoil layer, in order to ob-
tain a bulk density of 1 .4 -1.5 gm/cm3.
Such a bulk density is representative of
topsoil conditions commonly found in the
field.
Surface drains of either half-round
pipe or fabricated metal troughs were placed
across the cell width at 9, 18, and 27 m
from the farthest upslope position (Figure
2). Flow from these surface drains is
directed towards the instrumentation pits
and measured by tipping buckets. Rainfall
shields were placed above all surface
drainage pipes to intercept rainfall falling
directly into the pipes and to direct this
precipitation off the cell.
INSTRUMENTATION
Overview
Standard soil characterization equip-
ment was used for initial soil testing. In-
place density was obtained from the nuclear
density device, sand cone, and sand cylinder
apparatus. Permeability is being measured
by a high pressure triaxial chamber with
permeant upflow provisions.
In-field instrumentation includes soil
moisture measuring devices, temperature
probe, and flow measuring equipment.
Soil Moisture Devices
Soil moisture is being determined
from: (1) tensiometers, (2) gypsum (resi-
stance) blocks, and (3) neutron probe. The
tensiometers are specially constructed to
yield relatively rapid response. Standard
porous cups (1.9 cm diameter , 7 cm length)
were attached to 0.32 cm diameter nylon
tubing. The narrow tubing significantly
reduced water volume and thus yielded
a more sensitive response.
Tensiometers were placed along with
gypsum blocks and thermocouple wire at six
locations in each cell (Figures 1 through 3).
It is essential to determine soil moisture at
each interface. Hence, sensors were placed
10 to 15 cm below and above each inter-
face: (1) lower sand-clay, (2) clay-upper
sand (pervious layer), and (3) sand-topsoil.
A soil moisture/tension relationship
was determined for each soil layer through
pressure-plate analysis. Thus the moisture
percentage for each soil layer can be deter-
mined from tensiometer readings. Pressure
gauges were mounted in the instrumentation
pits for subsequent data recording.
Tensiometers have a relatively limited
range of sensitivity, approximately 3,100 to
8,266 kg/m2. Beyond about one bar (approx-
imately 10,000 kg/m2) the soil becomes too
dry and air cavitation occurs, rendering the
device useless. If actual soil moisture is
above this level, the tensiometers operate
reliably; otherwise, continuous priming,
which moistens the area around the porous
cup, is needed and data will reflect mois-
ture forced into the soil via priming and
not actual soil moisture movement through
the multi-layered profile. It is also recom-
mended that the tensiometer be drained
during anticipated freezing conditions.
Hence, if only tensiometers are used, winter
data gathering would cease.
Gypsum, electrical resistance, blocks
are highly absorbent. Two stainless steel
probes are molded in the gypsum blocks and
spaced such that a large resistance exists
between probes. A low level AC voltage is
applied to the probes; as moisture pene-
trates the gypsum material the resistance
decreases, changing the voltage drop across
the probes. This voltage drop is related
to soil moisture content during block
calibration, and is recorded by the data
gathering electronics.
Due to the anticipated problems with
tensiometers, a multi-sensor approach of
soil moisture data gathering was initiated.
Gypsum blocks were installed at each inter-
face about 15 cm from the tensiometers at
the identical horizontal level. A calibration
curve was developed for each soil type
which relates percent soil moisture to resis-
tance. Such a curve is shown in Figure 5
for the clay layer. The effective range is
3,100 to 15,500 kg/m2, that is, complete
coverage of expected soil moisture levels in
the multi-layered cover system. Gypsum
217
-------
100,000
10,000
UJ
o
z
en
-------
buckets were found to have an accuracy of
+ 0.5 percent. This contrasts with an
expected accuracy of ± 10 percent for
flume type devices.
Installation of Sensors
Soil moisture and temperature meas-
uring devices were installed approximately
15 cm above and below each soil interface
during construction of each layer. Also,
sensors were placed approximately at mid-
depth in the bottom sand layer in cells 2
and 3.
Installation of access tubes for the
neutron probe was accomplished with
minimal disturbance of the soil profile.
Access tube placement commenced after
compaction of the clay layer. The clay
was augered using a bit with a diameter
slightly less than that of the access tube.
Following clay extraction, a pipe of the
same diameter as that of the auger was
driven to the sand layer. Water was
added to the sand to assist in removal
and then the pipe was driven to within
approximately 45 cm above the liner.
The pipe, containing compacted sand, was
removed and the access tube of a slight-
ly larger diameter was driven through the
hole.
A pipe was inserted into the access
tube and a machined steel point placed
at the lower end of the pipe. Also, in
order to drive the access tube without
damage, an impact plate was machined
and attached to the top of the access
tube. The complete assembly was driven
into the previous augered hole. Then the
pipe assembly was removed for later use.
An anti-seep collar booth was attached
to the access tube slightly below the clay
surface to prevent any piping effects.
Finally, the remaining sand and topsoil
layers were backfilled around the access
tube.
A great deal of effort was exerted
to place sensors during construction and
to install access tubes without altering
the influence of construction equipment
upon the multi-layered cover. It is felt
that the additional time and monies util-
ized to install such a multi-sensor system
without any significant disturbance of the
soil profile are worth the effort and
should ensure an excellent data base.
Data Acquisition and Processing System
Compound Description
The major data acquisition and pro-
cessing components consist of: (1) Campbell
CR-5 data logger, (2) Tecmar Lab Master
multiplexer, and (3) the IBM Personal Com-
puter. The Campbell unit receives and
stores temperature data from the 20 loca-
tions in the cells at hourly intervals and the
Tecmar/lBM PC units receive, process, store,
and analyze information from the tipping
buckets and gypsum blocks.
The IBM Personal Computer is equip-
ped with two dual-sided, dual density, Si-
inch disk drives, and has 256K of Random
Access Memory (RAM). It contains several
expansion boards related to data acquisi-
tion. One is a Quadboard, manufactured by
Quadram Corporation. It provides an
additional 256K of RAM. The Quadboard
also has a battery powered clock/calendar,
which is used in conjunction with data
analysis.
The PC-Mate Lab Master, manufac-
tured by Tecmar, Inc., consists of a "moth-
er board" located inside. The computer
chassis and the "daughter board" are
externally connected by a 50 ribbon cable.
The Lab Master converts analog signals, in
the form of voltages, to digital signals.
The Lab Master only has pins for 16 analog
to digital conversion channels to be read;
with a multiplexer, it can handle up to 256
channels. Sixty are required for current
applications.
The PC-Mate IEEE488 Interface Board,
also by Tecmar, is connected to the Lab
Master daughter board. The version of the
IEEE488 used can multiplex 48 channels of
information and sent it to the Lab Master.
With the 16 channels the Lab Master can
process, a total of 64 channels exist.
A Screw Terminal/Signal Conditioning
Panel, manufactured by Data Translation, is
connected to the IEEE488 by a 50 ribbon
cable. The wires leading from the tipping
buckets and gypsum blocks are connected to
these screw terminals.
Application
The program driving these devices
219
-------
scans all 27 tipping buckets in less than
1.5 sec. Minimum tipping rate of the
buckets under the most intense design storm
conditions is 1.7 sec. Thus, a complete
scan is completed prior to subsequent buck-
et tipping. The resistance blocks are scan-
ned at the same frequency as tipping buck-
ets for efficient programming. Resistance
block voltages are written to a disk with
the associated time at five minute intervals.
The Campbell CR-5 data logger stores
temperature, in °C, on a cassette tape. All
systems have hard copy back-ups and are
protected from voltage spikes by a power
supply filter.
Rainfall Simulator
The rainfall simulator is a new,
highly portable nozzle type simulator de-
signed for erosion, infiltration, and runoff
experiments. The simulator is being modi-
fied, retrofitted, and calibrated prior to
application on this multi-layered cover
Study. Major characteristics of the rain-
fall simulator are: (1) near natural rainfall
drop-size distribution and drop inpact velo-
city, (2) spatially uniform rainfall intensity
and drop-size distribution, (3) near vertical
angle of drop impact, and (4) reproducible
storm patterns, durations, and intensities.
The rainfall simulator will be used to aug-
ment natural rainfall as required to stress
the multi-layered cover system.
Lab Subsidence Study
The proposal for the EPA-sponsored
study of the hydrologic response of hazard-
ous waste landfill covers included an eval-
uation of these clay covers under subsidence
conditions. In order to study this type of
behavior under controlled conditions, a
laboratory model of subsidence and cracking
in a compacted clay layer was proposed.
Several approaches used to physically model
soil behavior have since been evaluated.
The method of centrifugal modeling was
chosen as the most appropriate for this
study. Centrifugal modeling has been de-
termined to be appropriate Jor simulating
certain types of soil systems . Centrifugal
modeling consists of building models of the
prototype structure at a reduced scale.
The models are accelerated by rotating
them in a centrifuge. This acceleration is
expressed as a multiple of gravitational
acceleration that represents a scaling factor
which multiplies the reduced dimensions of
the scale model and allows it to respond to
loading conditions in the same way as the
prototype. Centrifugal model behavior can
accurately reproduce and predict prototype
behavior if correct experimental procedures
are used. Some advantages of this method
include: the relatively small amounts of
undisturbed soil that are required for model
construction, the minimal effect of sampling
on the prototype clay cover; and the acceler-
ation of time effects, which allows for
more experimentation in a given period of
time.
Once this method of study was decided
upon, the design and construction of a
centrifuge capable of producing the required
accelerations was begun. During the
summer of 1983, a 1.83 m diameter centri-
fuge was assembled that is capable of
accelerating a 27 kg sample to 100 G. It
is constructed of steel tubing and is secured
to a 0.6 m thick reinforced concrete floor
for stability. The centrifuge is powered by
a 2 HP electric DC motor with a controller
capable of precise rotational speed selec-
tion. As a safety measure, a 0.9 m high
sandbag wall has been constructed around
the centrifuge.
The 0.6 m thick clay cover will be
modeled using 2.5 cm thick layers of the
same clay that was used to construct the
full sized test covers. The models will be
accelerated to 24 G, in the centrifuge.
Thus, the 2.5 cm layer should behave in a
similar manner to the 0.6 m layer. In
order to accommodate the behavior of the
full width of the test plot, the soil layers
in the model will be circular with a dia-
meter of 0.3 m. Soil samples will be sup-
ported on a rigid aluminum disk while being
accelerated and mounted to the centrifuge
in a cylindrical container made of plexiglas
and aluminum. To simulate the existence of
cavities under the clay layer, holes will be
cut in the aluminum base plate. These
holes will vary from 2.5 cm to 20 cm
diameter and will simulate cavity diameters
of 0.6 m to 4.88 m in the prototype.
This centrifugal modeling study has
several objectives. The primary goal is to
predict how the test plot covers will behave
when cavities are created beneath them.
This prediction will probably consist of an
220
-------
estimate of the maximum cavity diameter
that can form beneath the surface without
causing subsidence of the cover layer, and a
qualitative estimate of the degree to which
cracking vs. plastic deformation of the clay
layer occurs when the critical cavity dia-
meter is exceeded. In addition, recom-
pacted clay samples will be accelerated to
study the relationship between soil mois-
ture content, compactive effort and the
soil's ability to deform with cracking over a
cavity.
RESULTS
The design and construction phases . of
this project have provided a valuable oppor-
tunity to describe and document a layered
soil system, and the difficulties encountered
in properly building this system. The multi-
layered, multi-sensor soil moisture monitor-
ing approach, utilizing tensiometers, resis-
tance blocks, thermocouple wire, and
the neutron probe, will ensure an accurate
measure of soil moisture movement through-
out the soil profile. These sensors should
provide a comprehensive data base for model-
ing efforts. Use of highly accurate tipping
buckets can also assist in determining the
economic feasibility of alternative drainage
and leachate collection systems.
Although the initial construction phase
of this project was not complete at the
submittal time of this article, a limited
amount of data collected from two storms
monitored on the initially constructed cover
appear to be extremely promising. Temper-
ature readings from the thermocouples
preceded increased ' moisture readings on
the resistance blocks in the uppermost
sensors. Using the resistance block, a
rapid response of the passage of the wetting
front by the thermocouples was detected,
and accurate moisture changes were re-
corded.
ACKNOWLEDGEMENTS
The authors acknowledge the support
of the Kentucky Utilities Company, upon
whose property the site is located and
for operation of construction equipment;
numerous staff of the University of Ken-
tucky Agricultural Engineering Department,
Civil Engineering Department, and Trans-
portation Research Center, for advice, test-
ing analyses, and office support. Dr. Wayne
Skaggs, North Carolina State University,
provided valuable hydrologic consultations.
Whayne Supply Company, Lexington, supplied
the loan of a Dynapac Compactor.
The information in this report has re-
sulted from research funded in part by
the United States Environmental Protec-
tion Agency, under Cooperative Agreement
CR-810431, to the University of Kentucky.
It has been subjected to the Agency's
peer and Administrative review and has
been approved for publication. Mention
of trade names or commercial products
does not constitute endorsement or recom-
mendation for use.
Investigations in this study are published
by the permission of the Director of the
Kentucky Agricultural Experiment Station
and this study is designated as Article
No. 84-2-17.
REFERENCES
I.Moore, I. D., M. C. Hirschi, and B. J.
Barfield. 1983. Kentucky Rainfall
Simulator. Transaction of the ASAE,
Vol. 26, No. 4, pp. 1085-1089.
2.Bassett, R. H., M. C. R. Davies, M.
J. Gunn, and R. H. G. Barry. 1981.
Centrifugal Models to Evaluate Numer-
ical Methods. Proceedings, 10th Inter-
national Conference on Soil Mechanics
and Foundation Engineering, Stockholm.
221
-------
ESTIMATION OF MAXIMUM COVER SUBSIDENCE EXPECTED
IN HAZARDOUS WASTE LANDFILLS
William L. Murphy and Paul A. Gilbert
Geotechnical Laboratory
USAE Waterways Experiment Station
Vicksburg, Mississippi 39180
ABSTRACT
Numerical models were developed to simulate horizontally layered hazardous waste
landfills. The models were analyzed using methods of linear elasticity to estimate the
maximum amount of subsidence of the final cover to be expected after landfill closure and
after deterioration of the waste containers. The landfill consisted of alternating layers
of intermediate inert cover soils and steel drums filled with simulated waste materials.
Waste drums, waste materials, and intermediate cover soils were assigned values of density,
Young's modulus and Poisson's ratio for the analysis. Landfill geometry, layer thick-
nesses, waste drum placement, steel drum stiffnesses, and laboratory consolidation tests
on the soils and simulated wastes were also considered. To simulate postclosure waste
layer deterioration, compression of the fill was calculated for decreasing values of the
Young's modulus of the waste layers. The analyses indicate that as much as 92 percent of
the expected subsidence of the cover is caused by closure of cavities (void space) inherent
in landfill ing. The maximum expected subsidence was calculated to be approximately
12 percent of the total landfill thickness.
INTRODUCTION
Settlement of sanitary and low-level
nuclear waste landfills with subsequent
damage to the covers has been documented.
The U. S. Environmental Protection Agency
(EPA) is concerned that settlement/
subsidence of hazardous waste landfills may
produce similar problems with cover systems.
Little or no documentation exists of the
potential for settlement in and subsidence
of the covers of hazardous waste landfills
operating in accordance with EPA interim
guidelines under the Resource Conservation
and Recovery Act (RCRA). This paper pre-
sents the results of analyses of maximum
subsidence expected to occur after deteri-
oration of waste drums in a representative
hazardous waste landfill. The analysis is
part of a broader study conducted for EPA.
PURPOSE
The purpose of this study was to
(a) determine the maximum amount of post-
closure subsidence of a landfill cover that
will occur after the waste containers
(steel drums) have deteriorated and exposed
their contents; (b) estimate the proportion
of the total subsidence contributed by var-
ious landfill components; and (c) develop
analysis and predictive techniques to esti-
mate the ultimate postclosure cover subsi-
dence to allow EPA to develop guidance for
landfill operators to design final covers
to withstand the stresses and the loss of
runoff slope associated with subsidence.
222
-------
APPROACH
Subsidence mechanisms postulated for
similar and related geotechnical activities
were assessed to develop the appropriate
analytical approach. Several active and
proposed hazardous waste landfills were ob-
served so that a representative landfill
could be developed for model analysis.
Analysis of the representative landfill
consisted of applying gravity loading to a
filled and closed, horizontally layered,
below-grade landfill. Total subsidence of
the cover was determined by calculating the
settlement of each fill layer for several
decreasing values of elastic modulus ap-
plied to the layer materials. Waste drums,
waste materials within the drums, and inert
fill (soil) comprising intermediate layers
between layers of drums were assigned val-
ues of density, Young's modulus, and Pois-
son's ratio in the analyses. Landfill
geometry and layer thicknesses were also
considered. Estimates of steel drum stiff-
ness were made from existing data. Moduli
of intermediate cover soils and of simu-
lated drum contents were derived from
existing data and from laboratory one-
dimensional consolidometer tests,
respectively.
THE LANDFILL MODEL
The landfill configuration chosen for
analysis is shown in Figure 1 (the configu-
ration for drums placed on their sides is
shown). The excavation is 50 ft deep,
200 ft wide at the base, and has side
slopes of 1 vertical on 3 horizontal. The
final cover has approximately 5 percent (1
vertical on 20 horizontal) slope from the
crown to the original ground surface. The
total thickness at the center of the land-
fill is 62 or 62.5 ft, depending on the
drum configuration. The landfill is lined
with compacted clay 8 ft thick at the base
and 5 ft thick on the sides and in the
cover. The 55-gallon drums containing the
waste material are placed across the fill
in continuous horizontal layers, which are
2 ft thick for drums on their sides and
3 ft thick for drums on-end. Waste layers
alternate with layers of intermediate soil
cover 1.5 ft thick. Cavities inherent from
the drum placement geometries exist within
the fill.
The liner, final cover, and interme-
diate cover are a clay with known
laboratory-tested properties determined for
a representative model material, the
"Vicksburg buckshot clay," with Atterberg
limits similar to liner clays of actual
landfills (see Table 1). The liner and
final cover clays are compacted; the in-
ternal intermediate cover soil is not
compacted. The simulated waste in
the drums was a commercial absorbent, non-
calcined crushed calcium bentonite clay-
stone, which was soaked in water for
48 hours to better simulate in-place waste
materials. (Clays fired at 1200°F are said
to be "calcined"; clays fired at lower tem-
peratures are said to be "noncalcined.")
ANALYSIS OF POSTCLOSURE SUBSIDENCE
Assumptions
In order to estimate the maximum or
worst case subsidence to be expected,
it was assumed that no filling of cavities
inherent from drum placement occurred by
downward sifting of intermediate cover
soils, and that the cavities ultimately
close completely. The analyses show that
as much as 92 percent of the ultimate sub-
sidence can be attributed to closing of the
inherent cavities. The remaining subsi-
dence is attributed to compression of land-
fill contents.
Other assumptions are: (a) all set-
tlement occurring in the liner and interme-
diate cover layers is complete by the time
the landfill is filled and closed, at which
time the cover is rebuilt to accommodate
any subsidence that has occurred; (b) the
stress condition in the landfill at the end
of filling is determined and does not
change after closure of the landfill;
(c) all postclosure subsidence occurs
through closing of cavities and through
compression of only the waste layers by
degradation of the waste layers' elastic
moduli; (d) maximum subsidence will, be-
cause the landfill is symmetrical about the
central vertical axis, occur at the center
of the landfill and can he determined by
analyzing the settlement of a central,
layered column; (e) stress is assumed to
vary linearly with depth; (f) one-
dimensional deformation occurs in the
223
-------
*
o
w
01
•H
in
•H
01
O
0)
o
n)
M
o
o
^3
10
a
o
•H
4-1
60
•H
M-l
a
O
CJ
13
.3
to
aS
ctf
N
(S
T3
1
(fl
•3
£>0
•H
224
-------
TABLE 1. ATTERBERG LIMITS OF TYPICAL HAZARDOUS WASTE INTERIM COVER MATERIALS
Liquid Limit,
percent
Site Location
AT abama
South Carolina
New York
Model material
Range
32-48
65-75
46-54
56
Average
38
68
50
Plasticity Index,
percent
Range
12-30
20-30
26-33
34
Average
18
24
30
(Vicksburg buck-
shot clay)
central column; (g) each landfill layer is
homogeneous and isotropic and can be repre-
sented by a single material with composite
properties; and (h) the fill is not
saturated.
In equation form, total subsidence,
may be represented as
5T =
6LV + 6DV +
(1)
where
= subsidence due to consolidation
of the intermediate cover
layers
6, v = subsidence due to closing of
the inherent geometric layer
void space
DV
subsidence attributed to void
space inside the drums. This
component is assumed to be zero,
but should be quantified and
accounted for in a general
approach
5AE = subsidence due to change in
stiffness of the waste layers
Settlement by Closing of Cavities
The maximum volume of void, or cavity,
that can be included in the landfill by
drum placement may be calculated geometri-
cally for drums placed on their sides and
for drums placed on end. The maximum cav-
ity volume for the former is 10.73 percent
of the total landfill volume and for the
latter is 9.31 percent. If the cavities
close completely, the maximum total subsi-
dence must be the sum of cavity closure
plus subsidence by compression of the land-
fill contents. However, actual landfill
operations will reduce maximum cavity vol-
ume because the operations do not achieve
perfect drum placement and because filling
of interdrum space by sifting of cover
soils during filling is generally practiced
to some extent. Conversely, any unexpected
void within the drums would contribute to
the volume-of inherent cavity.
Equations for Settlement
by Compression of Haste Layers
An equation was developed using the
methods of linear elasticity to calculate
settlement of the waste layers after clo-
sure for decreasing values of the Young's
modulus of the waste layers. The subsi-
dence in the cover is assumed to be the cu-
mulative settlement of all the waste layers
within the landfill. The equation is
AL =
(2)
225
-------
where
AL = the change in length of the ana-
lyzed column of waste
Y s the density of the composite
waste layers
L = the cumulative length of the
waste layers, i.e., the waste
column
v and E = Poisson's ratio and
Young's modulus, respec-
tively, assigned the
composite waste layers.
The analysis proceeds by applying gravity
to the column and calculating AL for suc-
cessively lower values of E and for sev-
eral values of v . Families of curves
were plotted for the two drum placement
configurations shown in Figure 2. Two
families of curves are needed because the
value of L in Equation 2 is different for
the two drum configurations (i.e., 14 waste
layers or 28 ft total waste layer thickness
for drums placed on their sides; 11 layers
or 33 ft total waste layer thickness for
drums placed on end). Figure 2 shows that
deformation increases rapidly as the modu-
lus falls below about 100,000 psf. Ulti-
mate subsidence can be estimated using the
curves of Figure 2 if values of E are
known for the initial condition (immediate-
ly after landfill closure) and final condi-
tion (after complete waste drum
deterioration). The ultimate subsidence is
the difference between the subsidences for
the initial (intact drum) and the final
(deteriorated drum) conditions. Use of the
curves in this way simulates the deforma-
tion or settlement of the landfill contents
after closure.
Estimation of Subsidence
for Representative Waste Layers
The waste layer for drums placed on
their sides is a composite of steel drum,
drum contents (the soaked calcium bentonite
claystone), and that portion of the space
between the drums that is filled with in-
termediate cover soils during landfilling.
The composite Young's modulus for the in-
tact drums placed on their sides (the ini-
tial condition) must be comprised of the
YOUNG'S MODULUS, PSF
A. Drums Placed on Their Sides
'O6 10'
YOUNG'SMODULUS, PSF
B. Drums Placed on End
Figure 2, Postclosure subsidence prediction
curves for compression of waste
layers by decrease in elastic .
modulus.
226
_
-------
modulus of the steel drum (a steel cylinder
on its side) and that portion of the inter-
mediate cover soil in the space between the
drums. The composite intact-drum waste
layer modulus can be determined using the
relationship
4EDEs
'Ti
4ED (1/2-TT/8)
(3)
where
ET. = elastic modulus of the initial
composite waste layers, psf
ED = modulus of the steel drum
E = elastic modulus of the interme-
diate cover soil, psf
The E was determined from consolidation
tests on the model intermediate cover soil.
The composite intact-drum waste layer modu-
lus for an E fiof 1.04 x 10D psf and an
En of(-9.7 x TO psf is, from Equation 3,
9 x 10 psf. This modulus corresponds to
an initial subsidence of approximately
0.37 in. for a Poisson's ratio of 0.35 in
Figure 2a.
If the waste drums are allowed to cor-
rode completely and expose the contents,
the wastes are then subjected to compres-
sion, which is controlled by the modulus of
the waste material. One-dimensional con-
solidation tests were performed on the non-
calcined crushed claystone simulated wastes
and on another commercial product, oil-
spill absorbent (calcined crushed clay-
stone), after allowing the materials to
soak in water for 48 hours. Figure 3 pre-
sents the results of the consolidation
tests in the form of constrained elastic
moduli. The secant modulus of 40,000 psf
(275 psi) of the noncalcined material was
chosen to represent the model waste mate-
rial (for final conditions) because it is
the lower of the two and because it is com-
monly used in the disposal industry. As in
the initial condition, the final composite
waste layer modulus must be calculated, in
this case from
-Tf
4E,
(4)
Calcined crushe
claystone
N on calcined
crushed
clayst one
4 8 12 16 20
PERCENT STRAIN
24
Figure 3. Constrained elastic moduli of two
simulated waste materials from one-
d.imens.i.onal consolidation tests.
where
E,..p = elastic modulus of the composite
system in the final condition
(steel drums corroded entirely
away)
E = modulus of drum contents (the
noncalcined material)
ES = modulus of intermediate clay
cover soil, as before
Note that Equations 3 and 4 are the same
except that En of Equation 3 has been re-
placed by E in Equation 4, representing
deterioration of the drum and exposure of
the drum contents in the final condition.
Substituting the modulus values for E
(40,000 psf) and E (104,000 psf) into
Equation 4 yields a corresponding composite
modulus of 48,000 psf, which represents a
^inal condition subsidence on Figure 2a of
7 in. (with v = 0.35). Subtracting the
initial value of 0.37 in. from the final
value of 7.02 in. yields a corresponding
227
-------
value of 6.65 in. or 0.89 percent for sub-
sidence by compression of cell contents for
drums placed on their sides in a landfill
750 in. (62.5 ft) thick. The subsidence
possible from closing of inherent cavities
is 10.73 percent, and the ultimate maximum
expected subsidence is the sum or
11.62 percent (approximately 87 in. for the
modeled landfill).
A similar analysis can be performed
for a landfill with drums placed on end.
The effective elastic modulus of a hollow
steel cylinder carrying all of the axial
load can be shown to be En = 3.86 x 10
psf , which is an initial modulus corre-
sponding in Figure 2b to essentially zero
subsidence. When the drums have completely
corroded, the final modulus (drum content
alone) is 40,000 psf, which corresponds to
a final subsidence value of 11.70 in.
(Figure 2b), or 1.57 percent strain in a
height of 744 in. (62 ft). The subsidence
possible from the closing of inherent cavi-
ties is 9.31 percent, and the ultimate
maximum expected subsidence for drums
placed on end is 1.57 + 9.31 or
10.88 percent (approximately 81 in.).
These subsidence values would result in a
final average cover slope of about
2 percent, which would be unacceptable
under current RCRA guidelines, assuming the
original slope was 5 percent.
will be reduced below the recommended
value. The settlement/subsidence operation
from which the prediction curves for this
study were produced can be applied to
representative hazardous waste landfills
having any depth, material densities, and
elastic moduli.
CONCLUSIONS
Analysis of simulated hazardous waste
landfills indicates that most (approxi-
mately 86 to 92 percent) of the fill set-
tlement and corresponding cover subsidence
can be attributed to the closing of inher-
ent cavities in the fill. The remaining
deformation is attributed to compression of
v/astes after deterioration of steel drum
containers. The source of the cavity vol-
ume within the landfill is presumed to be
unfilled space between waste containers
(primarily steel drums). The landfill
operator can substantially reduce the ulti-
mate expected subsidence by striving to
fill the interdrum space during operations.
There app.ears to be no appreciable advan-
tage of one drum placement configuration
over another. If the maximum predicted
subsidence should occur, the cover slope
228
-------
REFERENCES
Daniel, D. E. 1983. "Shallow Land Burial of Low-Level Radioactive Waste,"
ASCE Journal of Geotechnical Engineering, Vol 109, No. 1, pp 40-55.
Kahle, R., and Rowlands, V. 1981. "Evaluation of Trench Subsidence and
Stabilization at Sheffield Low-Level Radioactive Waste Disposal Facility,"
NUREG/CR-2101, U. S. Nuclear Regulatory Commission.
Luttoti, R. J., Regan, G. L., and Jones, L. W. ,,1979. "Design and Construction
of Covers for Solid Waste Landfills," Report No. EPA-600/2-79-165, Municipal
Environmental Research Lab, U. S. Environmental Protection Agency, Cincinnati,
OH.
Withiam, J. L., and Kulhawy, F. H. 1978. "Analytical Modeling of the Uplift
Behavior of Drilled Shaft Foundations," Geotechnical Engineering Report 78-1,
School of Civil and Environmental Engineering, Cornell University, Ithaca, NY.
229
-------
PERMEABILITY CHARACTERISTICS OF FLEXIBLE MEMBRANE
LINERS MEASURED IN POUCH TESTS
Henry E. Haxo, Jr., Nancy A. Nelson
Matrecon, Inc.
Oakland, California 94623
ABSTRACT
Permeability characteristics of a variety of thermoplastic and partially crystalline
membrane materials to several waste and test liquids have been explored in pouch experi-
ments. These liquids include aqueous salt solutions (such as sodium chloride and lithium
chloride), organic solvents (acetone and xylene), mixtures of acetone and water, several
waste liquids, and water- and organic-soluble dyes that have been either suggested for use
or used for detecting leaks in membrane liners in service. The pouch test involves seal-
ing a quantity of waste or test liquid in containers that are fabricated of the membrane
under test. The pouches are immersed in deionized water or other liquid. Transmission
through the walls of constituents either inside or outside the pouch is monitored by
changes in the weight of the filled pouch, chemical analysis, pH, and electrical conduc-
tivity of the outer liquid, and the appearance of the dyes in the outer liquid or on the
pouch surface. Transmission rates for various components of the liquids through the wall
can be calculated from concentration data.
Results of the pouch experiments show the permeability of polymeric membranes to
water, some organic liquids, e.g. oils, acetone, and xylene, organic dyes, and hydrogen
and hydroxyl ions, but not to other inorganic cations and anions. The results also
indicate that the principal driving force in the transmission of the various species is
the concentration gradient across the pouch wall membrane.
With additional development the pouch procedure appears to be a feasible and prac-
tical method for assessing the permeability of membrane liners to different molecular
species under conditions that simulate some aspects of actual service, and for assessing
the compatibility of a liner with a specific waste.
INTRODUCTION
The low permeability of flexible
polymeric membranes to gases, liquids, and
vapors makes this type of liner the best
approach to meeting the requirements of
the U.S. Environmental Protection Agency's
regulations for the lining of waste
storage, treatment, and disposal faci-
lities to prevent the migration of waste
liquids into the groundwater. This low
permeability makes it impossible to test
polymeric membranes by conventional
methods used for soils and other porous
materials. In contrast to the porous
structures of soils and concretes, poly-
meric membrane liners are nonporous.
Moisture vapor transmission, for
instance, as measured in accordance with
American Society for Testing and Mater-
ials (ASTM) Method E96, has been used as
a test for the permeability of membranes.
The conditions encountered by the mem-
brane in such a test differ significantly
from those encountered in actual service;
thus, the permeability of membrane liners
in service to waste liquids and dissolved
components probably is not reflected by
the water vapor transmission data. Other
methods of assessing the permeability of
liner materials were investigated. One
of the most promising methods was a
procedure in which a waste or test liquid
is sealed into a pouch fabricated from
230
-------
the liner membrane under test and then
placed in a container filled with deionized
water or other standard liquid. Such an
arrangement simulates some aspects of a
liner in a surface impoundment containing a
waste liquid. The permeation of dissolved
constituents of the waste liquid through
the wall can be followed by pH and elec-
trical conductivity measurements of the
water in the outer container; the perme-
ation of water into the pouch containing
waste liquid can be determined by moni-
toring the weight of the loaded pouch.
Also, the organic constituents of the
wastes that pass through the wall into the
outer liquid can be followed by analysis of
the outer liquid by gas chromatography.
listed in Table 1. All of the materials
are thermoplastic or partially crystalline
thermoplastics. Full descriptions of the
individual materials and data on their
properties are presented in Haxo et al
(1982). Only one membrane in this study
was fabric-reinforced, i.e. a chlorosul-
fonated polyethylene [CSPE (6R)]. Fabric-
reinforced membranes were excluded to
avoid possible pinholes or leaks caused by
the threads. The principal objective of
the program was to study the permeability
characteristics of different polymeric
compositions. Fabric-reinforced materials
are included in later studies now in
progress.
This paper describes the pouch test
and presents the results of using this type
of test to assess some permeability char-
acteristics of polymeric membranes.
Results of some long-term tests of a
variety of lining materials and several
different waste liquids and test fluids are
presented. Included are the results of
tests with actual wastes used in our past
liner research programs, salts, two organic
solvents (acetone and xyl ene) , and in-
organic and organic tracers.
OBJECTIVES
The two main objectives of this study
TABLE 1. POLYMERIC MEMBRANES3 IN POUCH EXPERIMENTS
were:
1. To determine under conditions that
simulate actual liners in service
some of the permeability charac-
teristics of polymeric membrane
lining materials to constituents
found in waste liquids that may be
stored, treated, or disposed of on
land.
2. To assess the permeability charac-
teristics of polymeric membranes
to a .variety of test liquids and
selected wastes using a pouch
test.
Polymer
Abbreviation Liner serial
number
Thickness, Type of
mil composition0
Chlorinated
polyethylene
Chlorosulfonated
polyethylene
Elasticized
polyolefin
Polybutyl ene
Polyester
CPE
ELPO
PB
77
86
55
85
98
221
30
22
31
35
33
8
30
TP
TP
TP
TP
TP
CX
CX
el astomer
Polyethyl ene,
low-density
Polyethylene,
high-density
high-density PE-
EPDM alloy
Polyvinyl chloride
PEEL
LOPE
HOPE
HOPE-A
PVC
PVC-ORd
75
21
184
180
11
17
19
59
88
93
137
146
8
10
30
20
30
20
22
33
20
11
30
30
CX
CX
CX
CX
TP
TP
TP
TP
TP
TP
TP
TP
details given in Haxo et al, in draft, and SW870, Revised (Matrecon,
1983).
bTP = thermoplastic, CX = partially crystalline thermoplastic
cFabric-reinforced with 6x6 nylon scrim.
dOR - Oil-resistant.
Waste and Test Liquids
METHODS AND MATERIALS
Polymeric Membrane Liners,
The polymeric membranes that were
fabricated into pouches and subjected to
permeability studies in this work are
A wide variety of test liquids were
included in this study. They included
actual waste liquids, several salt solu-
tions, e.g. NaCl and LiCl, organic sol-
vents, and various solutions containing
dyes. Table 2 lists the wastes and test
liquids in the study and presents some
information regarding each.
231
-------
TABLE 2. WASTE AND TEST LIQUIDS IN POUCH EXPERIMENTS
Type
Name Matrecon
waste
number
PH
Electrical
conductivity
pmho/cm
Reference
Haste liquids
MSW leachate ... W-l 5.2 1,500
Acidic waste "HN03-HF-HOAc" W-9 1.1 155,000
Alkaline waste "Slopwater" W-4 12.0 129,000
Haxo et al, 1982, p. 98
Haxo et al, in press
Haxo et al, in press
Test liquids
Brine
LiCl solution
Lid solution
Organic-ketone
Organic-water
mixture
Organic-hydro-
carbon
Tracer dye in
DI water
Tracer dye in
DI water
5% Nad solution W-19
5% ... 4.8
10% ... 4.6
Acetone with dyea
50:50 acetone:
water with
Xylene wi th dyea
1% Fluorescein
solution ... 7.9
1% Sevron Red
solution ... 2.4
62,000
68,000
106,000
6,400
660
solution with Automate Red or with Fluorescent Yellow.
solution of methyl violet.
Description of the Pouch Test
In the pouch test, which simulates
some conditions of liners in service
in surface impoundments, small containers
are fabricated of the membrane to be
tested and are filled with waste liquid or
other test liquid, sealed, and wholly
immersed in a liquid of known composition,
e.g. deionized water. Permeabilities of
the membrane to water and to the dissolved
constituents can be determined by observing
with time the change in the weight of the
pouch and changes in the pH, electrical
conductivity, and the composition of
the outer (originally deionized) water.
Because of differences in the concen-
trations of different species across
the membrane wall, osmosis will cause the
outer water, which is higher in water con-
centration than the pouch liquid, to enter
the pouch and also cause ions and dissolved
constituents of the pouch liquid in which
the concentrations are higher to tend to
leave the pouch and enter the outer water.
These effects are illustrated schematically
in Figure 1.
During monitoring the pouch is
weighed; however, an increase in weight of
the pouch may only reflect absorption of
the outer water by the wall and not perme-
ation of water into the pouch. When the
test is completed the pouch and its con-
tents are weighed separately to determine
the actual amounts that have been absorbed
232
-------
by the wall and the amounts that have
permeated the pouch wall. The pouches are
dismantled and the walls are analyzed for
changes in composition, e.g. volatiles and
extractables; physical properties, e.g.
tensile and tear properties, are measured
to determine the effects on the liner of
one-sided exposure of the membrane to the
waste.
CONDUCTIVITY
Figure 1. Pouch assembly with an aqueous
waste or test liquid to show the
movement of waste constituents
during the pouch test. The
filled pouch is immersed in
deionized water. Arrows
indicate the flow of specific
constituents.
Results reported in Haxo et al (1982)
show the feasibility of a pouch-type test
for assessing simultaneously the perme-
ability of polymeric membranes to wastes
and to some waste constituents and the
effects of one-sided exposure to a waste
liquid on the physical properties of the
membrane.
Mechanism of Flow Through
Polymeric Membranes'
In contrast to soils and other porous
materials, polymeric membrane liners are
nonporous. In soils the flow of liquids
containing dissolved constituents is
through the pores with hydraulic gradient
the principal driving force. In the case
of the membranes that are nonporous, flow
of the fluid, be it gas, liquid, or vapor,
proceeds essentially in three steps:
1. Dissolution of the fluid in the
membrane.
2. Diffusion of the fluid through the
membrane.
3. Evaporation of the fluid on the
downstream side of the membrane.
Step 1 depends upon the solubility of
the fluid in the membrane and relative
"activity"/ of the fluid on both sides of
the interface. The diffusion through the
membrane in Step 2 involves a variety of
parameters which include the molecular size
of the permeating fluid and various charac-
teristics of the membrane compound, e.g.
presence of fillers, crystalline zones,
crosslinks, etc. Step 3 is similar to the
first step and depends upon the relative
"activity" of individual constituents on
both sides of the interface.
Overall, the major driving force for
the movement of a given constituent is the
relative concentration on both sides of
the membrane. Each constituent in a
mixture will tend to move from a higher
concentration of the specific species to a
lower concentration of that species. The
placing of a waste in a pouch immersed in
deionized (DI) water creates a significant
concentration difference which will cause
the movement of water and other waste or
test liquid constituents.
Fabrication and Filling of Pouches
For the pouch test to be effective,
the seams must be liquid-tight so that
whatever passes in or out of the pouch
passes through the walls. The thermo-
plastic and partially crystalline membranes
can be heat-sealed which is the most
effective method of making small seams;
such seams require the least area and, if
well made, have held up for the duration of
exposure. A portable heat-sealing unit,
designed primarily for sealing films less
than 10 mil in thickness, was used in this
study. Consequently, some problems were
encountered in seaming the thicker mem-
branes used for liners. Longer dwell time,
greater pressure, and higher temperatures
were needed to make seams of the thicker
materials. Some seams of thermoplastics
were made with adhesives and solvents.
233
-------
Adequate pouch seams of cross!inked mem-
branes could not be made with the two-part
room temperature curing adhesives supplied
by the manufacturers; thus, only pouches of
thermoplastic and crystalline membranes
were included in this study. Nevertheless,
the general permeability characteristics
determined for polymeric membranes are
valid unless crosslinking is a factor.
A variety of pouch shapes were used
during the course of this study. Figure 2
is a schematic of the pouch assembly used
for most pouches. The dimensions of the
individual pouches varied somewhat; some of
the first pouches were 18 x 14 cm, which
gave an effective wall area of approxi-
mately 500 cm^ in contact with the pouch
liquid. For the sodium chloride solution,
the pouches were 17 x 12 cm between seams
which yielded an effective surface of ap-
proximately 400 cm2. For the experiments
with solvents and the studies of tracer
chemicals smaller pouches were used which
could be placed in closed quart jars.
Also, smaller pouches made of high density
polyethylene (HOPE) and other crystalline
polymers were used because of the stiffness
of these sheetings.
As is shown in Figure 2, most of the
pouches included a neck through which the
test liquid could be introduced and which
could then be sealed. Figure 3 shows a
pattern used for cutting and seaming the
pouches made of the more flexible sheet-
ings. The seams were tested by filling
the pouches with water and checking for
leaks after several days. To simulate
extended exposure in a moist environment
some of the pouches, i.e. those that
were filled with the highly acidic waste
(W9), were first immersed for two weeks
in water at 70°C; for some pouches it
was necessary at 70°C to put water in
the pouch to prevent the pouch walls
from sticking together. The soaking of
the pouches at 70°C further tested the
integrity of the seams. After a pouch was
filled with the test liquid, as much air as
possible was removed from the pouch and the
neck was sealed; care was taken to keep the
inside of the neck dry to insure a good
seal. The volume of waste liquid added to
the individual pouches varied from 30 to
200 ml, depending on the size and configu-
ration of the pouch. Most of the pouches
were heat-sealed and the rest were sealed
with either an adhesive or a solvent.
OUTER BAG-
POLYBUTYLENE
DEIONIZED WATER.
IN OUTER BAG
INNER POUCH-
MEMBRANE UNDER TEST
WASTE OR TEST LIQUID
IN THE INNER POUCH
Figure 2. Schematic of pouch assembly
showing inner pouch made of
membrane material under test.
The pouch is filled with waste
liquid and sealed at the neck.
The outer polybutylene bag,
which can be easily opened, is
filled with deionized water.
For this configuration the water
' in the outer bag is monitored
for pH, conductivity, and odor;
the pouch is monitored for
weight change.
Monitoring the Pouches During Exposure
To monitor the pouch assemblies, the
following tests were performed periodically
during the exposure:
1. The water in each of the outer
bags was tested for pH, electrical
conductivity, and odor, e.g.
butyric acid, as in the case of
tests with MSW leachate.
2. The pouches containing the test
liquids were removed from the
outer water, wiped dry, inspected
for possible leaks, and weighed.
Figure 4 shows the pouches, the tray
assembly for holding the pouches, and the
auxiliary equipment for monitoring the
pouches.
During the exposure, evaporation
caused the water level in some of the outer
bags to drop and expose the top part of
234
-------
the pouch to air; the loss was made up with
deionized water. Also, if the concentra-
tion of ions in the water in the outer bag
(as shown by electrical conductivity)
appeared to be too high because of migra-
tion or possibly a leak, the outer water
was replaced with deionized water. This
would return the concentration gradient
across the membrane essentially to its
original value and increase the trans-
mission rates.
\+- --•- 7" - -- -*-
T
Open for waste 2"
to be added ]
i
t
P
-it
,_^_ _— ™ ~ _- {-.
|| ||
_JI .„ „ i ii
ii H! T !!
II GRAIN ||
II DIRECTION ||
!i * i
i
i! II
_j— „__—- j^,
'
7
,
Figure 3. Pattern for cutting pieces of
membranes for making some of the
pouches. Dotted line indicates
the heat-seal of the pouch. The
inside dimensions of the pouch
are 4.5 in. x 5.75 in. (11.43 x
14.6 cm), which yields an
effective area (2 sides) of
about 52 in.2 (335 cm2).
Dismantling the Pouches After Exposure
Most of the pouches were allowed to
continue their exposure for compara-
tively long times in order to assess the
effects of exposure to water and wastes.
If a leak developed in the pouch, as
indicated by abrupt changes in the moni-
toring data, the pouch was withdrawn from
exposure and emptied, and the liner and
liquids were tested.
Dismantling most of the pouches
involved the following .steps:
1. Weigh the filled pouch.
2. Determine pH and electrical
conductivity of the water in
the outer bag.
3. Measure length and width between
seams of pouch.
4. Empty pouch and weigh contents.
5. Determine pH and electrical
conductivity of waste removed
from pouch.
6. Dismantle pouch at seams, leaving
bottom seam intact for measurement.
7. Blot the pouch dry and weigh to
determine weight increase.
Figure 4. Pouch and auxiliary equipment
for monitoring the pouches of
polymeric membrane 1iners to
assess permeability to water and
constituents of waste liquids.
Monitoring equipment for mea-
suring pH and electrical con-
ductivity of the water in the
outer bags is shown.
Reporting the Results
for
The
each
1.
fol1owing results were
pouch assembly tested:
reported
Electrical conductivity and pH of
water in outer bag as a function
of time.
235
-------
2. Weight of filled pouch as a
function of exposure time.
3. Measurement of amount of waste
liquid in pouch at start and at
time of dismantling.
4. Analysis of waste in pouch, i.e.
pH, electrical conductivity, and
percent total solids, at beginning
and end of exposure.
5. Analysis for volatiles and ex-
tractables of unexposed pouch
membrane and pouch wall at time of
dismantling.
6. Calculation of water transmission
rate based on the early portion of
the pouch weight-time curves.
(Probably a better method for this
calculation is to base it on the
actual weight increase of the
liquid in the pouch that is ob-
tained when the pouch is disas-
sembled at the end of the test.)
Table 3 shows the measurements that
were regularly made initially, during
monitoring, and at the time the pouches
were dismantled. A few additional measure-
ments were made such as chlorides content
of the outer liquid in the case of the test
of NaCl solutions.
TABLE 3. MEASUREMENTS MADE OH POUCHES AND CONTENTS
Parameter
Initial Monitoring Final
Weight of pouch (empty) X
Height of pouch (filled) X
Height of test or waste Xa
liquid in pouch
Pouch wall
Extractables
Volatiles
Haste liquid in pouch
Electrical conductivity
pH
Sol ids
Hater in outer bag
Electrical conductivity
PH
Sol Ids
X
X
X
X
X
X
X
X
X
X
X
X
X X
X X
X
Problems Encountered
Difficulty in Making Pouches
Cross-linked Sheetings
with
^Height by difference in weights of filled and empty pouches.
"Weight by difference, as in footnote a, and by direct mea-
surement of liquid at the time the pouch was disassembled.
Unsuccessful attempts were made to
fabricate pouches of crosslinked rubber
sheetings, e.g. butyl rubber and neoprene,
using two-part, room-temperature-curing
adhesives supplied by the manufacturers.
Seams 1.5 to 2 in. wide were made and
appeared to cure; but, when the pouches
were filled, the seams gradually opened and
the pouches failed. Using the appropriate
raw compound and high temperature and
pressure, adequate pouches for testing can
probably be fabricated from crosslinked
membranes. For the initial purpose of this
study, pouches made of thermoplastic
sheeting were sufficient to demonstrate
some of the permeability characteristics of
polymeric membrane lining materials.
Failure of Pouches During Exposure
In spite of the testing of the pouches
before filling with wastes, several of the
pouches failed prematurely, after one to
872 days. All of the failures, except that
of one polybutylene (PB) pouch, were in or
at the seams. Those that failed at one and
nine days were replaced. Also replaced
after approximately 790 days was P30A which
was made of elasticized polyolefin (ELPO)
and filled with slopwater (W-4), the
highly alkaline waste.
Seam failure appeared to predominate
among the pouches made of thick membranes,
seamed with solvents, or filled with slop-
water. No seam failures occurred among
the heat-sealed PB and polyethylene (PE)
pouches. Seam failures indicate problems
in the making of pouches; they do not
reflect on the seaming by manufacturers of
the liners, either in the factory or in the
field. Since all of the heat-sealed
pouches that failed were fabricated with a
device that was not intended for seaming
such thick materials, these failures can
only reflect the inadequacy of the equip--
ment.
RESULTS
Aqueous Sodium Chloride Solution
The first test solution explored in
the pouch experiments was a 5% aqueous
236
-------
solution of sodium chloride which was
loaded into pouches of chlorinated poly-
ethylene [CPE (77)], CSPE (6R), ELPO (36),
polyester elastomer [PEEL (75)], and two
polyvinyl chlorides [PVC (11 and 59)].
Later, a second set of pouches was prepared
and put into test. The second set con-
sisted of membranes of CPE (86), CSPE (55),
low-density polyethylene [LDPE (21)], PB
(98), and two additional PVC's (17 and 19).
In the first set, three of the pouches
failed when their seams opened; the remain-
ing three were continued in test until they
were dismantled after approximately 1152
days of exposure along with the second set
which had been in exposure from 310 to 316
days. These pouch assemblies were sub-
jected to the measurements described above.
The results are presented in Table 4.
All of the filled pouches increased in
weight during exposure though the amounts
differed. In all cases but one, the
liquids contained by the respective pouches
increased in weight during the exposure.
In several cases, the increases were caused
by the absorption of water in the pouch
TABLE 4. POUCH TEST OF MEMBRANES WITH 5% SODIUM CHLORIDE SOLUTION AS THE TEST LIQUID
Parameter
Matrecon liner
Liner thickness
Pouch number
Days exposed
serial number
, mil
CPE
86
20
77
316
CSPE
6R
34
2
1151
CSPE
55
35
76
315
ELPO
36
22
4
1151
LDPE
21
10
75
315
PB
98
7
78
316
PVC
17
20
73
310
PVC
19
20
74
315
PVC
59
30
5
11153
Area of pouch exposed to waste.
(original dimensions), m2 0.0392 0.0408 0.0365 0.0408
Weight of empty pouch
Unexposed, g 40.00 45.33 62.90
After exposure, g 43.60 52.30 68.25
Change from unexposed, g 3.60 6.97 5.35
Weight of filled pouch
At beginning of exposure, g 184.30 388.52 163.10
After exposure, g 188.32 398.75 167.80
Change in weight, g 4.02 10.23 4.70
Test liquid in pouch
Weight at beginning of
exposure, g 144.30 343.19 100.20
Weight after exposure, g 144.72 346.45 99.55
Change in weight, g 0.42 3.26 -0.65
pH after exposure 3.6 6.9 7.3
Electrical conductivity
after exposure, umho/cm
Liquid in outer bag
pH at conclusion of test 3.6 7.4
Electrical conductivity at
conclusion of test, pmho/cm 86 585
Calculated concentration
of NaClc, % 0.005 0.035
Chloride contentd, ppm 4 29
Water transmission through pouch
wall6, g/m2/d 0.291
Water transmission into pouch*,
g/m2/day 0.315
23.39
23.40
-0.01
326.50
328.82
2.32
303.11
305.42
2.31
5.0
0.0350
12.41
12.45
0.04
163.90
164.75
0.85
151.49
152.30
0.81
4.5
0.0375
7.48
7.62
0.14
186.00
186.65
0.65
178.52
179.03
0.51
5.9
0.0357
35.30
34.90
-0.4
158.60
165.05
6.45
123.30
130.15
6.85
3.5
0.0351
32.80
32.20
-0.6
134.70
136.47
1.77
101.90
104.27
2.37
6.6
Electrical conductivity9 of DI water
exposed to pieces of sheeting, 325
umho/cm
Calculated concentration of NaClc in
DI water exposed to sheeting, % 0.019
0.069
0.203
87
0.005
6.7
66
0.004
3
-0.002
0.384
110
0.006
4.8
34
0.002
' 3
0.049
0.053
61
0.003
4.4
31
0.002
3
0.073
0.033
93
0.005
5.1
27
0.001
190
0.043
0.051
47
0.003
3.1
371
0.022
27
0.619
0.184
220
0.013
6.4
16
0.001
2
0.214
0.181
33
0.002
0.0408
47.24
47.43
0.19
537.05
548.05
11.00
489.81
500.62
10.81
6.4
64,000 80,000(7) 63,000 60,000 67,000 63,000 64,000 62,000 62,000
5.8°
57b
0.003b
(b)
0.238
0.238
26
0.001
apouch failed between monitoring at day 1115 and dismantling at day 1152. Data presented for the exposed pouch are for
day 1115 except for the emptied pouch weight, weight of test liquid removed from pouch, and pH and electrical conductivity ot
bAfterefa11ure,1at1day 1152 the outer water had pH = 5.4, electrical conductivity = 4500 umho/cm, NaCl concentration = 0.280 %,
and chlorides content = >1000 ppm. ,,„„„,
cBased on data in Handbook of Chemistry and Physics. 63rd edition, p. D-261 (1982).
dMethod of chlorides analysis: FHWA-RD-77-85 (Clear and Harrigan, 1977)
eCalculated from change in weight of waste in pouch at time of dismantling.
fCalculated from weight of filled pouch based on early portion of weight-time curves or on entire exposure.
gconductivity of DI water in PB bags in which coupons of membrane liners had been immersed for 805 days.
237
-------
walls. The rate of transmission of water
into the pouches Is reported in Table 4.
The greatest transmissions were through the
CPE, CSPE, and the PVC's; the lowest were
through the partially crystalline membranes
(ELPO, LDPE, and PB). Judging by the
change in electrical conductivity of the
liquid in the outer bags, a small number of
ions may have passed through the walls of
the fabric-reinforced CSPE and two of the
PVC's. Most of the increases in conduct-
ivity were low and were probably the result
of ions, e.g. sodium and chloride, Teaching
out of the membranes. Also, analyses run
on the solutions indicated very small
amounts of chlorides passing through the
membrane walls.
Overall, these results show that all
of the membranes are permeable to moisture
to varying degrees and that the difference
in concentration between the outer liquid
and the salt solution in the pouches fur-
nishes the driving force that causes the
water to migrate into the pouches. The
movement of the ions from the test liquid
in the pouches into the outer liquid is
minimal. More detailed accurate analyses
of the outer liquid should be made.
Municipal Solid Waste (MSW) Leachate
The first waste liquid to be used in
the pouch-liner permeability study was the
MSW leachate generated in the landfill
simulators used in the study of lining
materials for MSW landfills (Haxo et al,
1982). The membrane liner materials in
these tests were based on CPE (77), ELPO
(36), PEEL (75), and PVC (11, 17, and 59).
After 500 days of exposure (Table 5)
it was apparent that both the water and
dissolved ingredients initially in the
leachate had moved through the pouch walls.
The electrical conductivity of the water in
the outer bag increased, indicating the
permeation of some ions from the leachate
in the pouch into the deionized water.
Also, the odor of leachate in the water in
the outer bags indicated that some organic
constituents in the leachate were permeat-
ing the pouch walls. An increase in the
weight of the pouches containing leachate
indicated movement of water through the
walls into the pouches. In that series,
the ELPO allowed the lowest transmission of
water and dissolved components, and the CPE
appeared to be the most permeable.
TABLE 5. CHANGES TO POUCHES CONTAINING MSW LEACHATE
AFTER IMMERSION IN DI WATER FOR 500 DAYS
Pouch3
Original values
Values at 500 days
Liner Pouch Electrical Electrical Weight
Polymer number number pHb conductivity"3, pHb conductivity13, increase0,
umho/cm u mho/cm g
CPE
ELPO
PEEL
PVC
PVC
PVC
Bl ankd
77
36
75
11
17
59
• * •
14
11
13
9
10
12
8
5.7
5.1
4.0
5.8
5.0
5.7
5.5
5.2
4.3
20.5
6.0
13.3
5.9
1.33
6.5
4.5
' 6.4
6.0
2.8
6.3
4.3
124.0
17.8
50.0
32.0
325.0
23.2
11.6
4.74
0.22
2.95
1.12
1.37
1.21
• • •
aAreas of the pouches exposed to the leachate ranged from 544 to 624
and electrical conductivity of (originally deionized) water outside the pouches
containing leachate.
°Weight increase of pouches containing leachate.
outer bag with DI water only.
238
-------
Pouches Containing a Highly Acidic Waste
The first set of pouches in this study
that contained a hazardous waste were
with the highly acidic waste
(W9) that was being used in
on liners for hazardous
in press). Six,pouches
a different membrane—
(36), PB (98),
fil led
"HN03-HF-HOAc"
a research project
wastes (Haxo et al,
were used, each of
(CPE) 86, CSPE (85), ELPO
PVC (19), and PVC (88).
Monitoring the Pouches During Exposure
Monitoring the pouch assemblies
included weighing the pouches containing
the waste liquid and measuring the pH and
electrical conductivity of the water in the
outer bags. In addition, the general
condition of the pouches and the waters was
observed. Deionized water was added when
too much water had evaporated from the
outer bag, and the outer liquid was
completely replaced with DI water when the
concentration of dissolved constituents
appeared to have become too high and was
slowing the transmission of these constit-
uents into the outer liquid.
Selected monitoring data for Pouch 19
are presented in Table 6. The concentra-
tion of ions as a function of time was
calculated from the electrical conductivity
data based on the assumption that the ions
were from NaCl. Monitoring data for the
six pouches (including weights, pH, and
electrical conductivity) are presented in
Figure 5.
Dismantling of Pouches
Condition of the Pouch Assemblies—In
this set of six pouches only the PB failed
prematurely at 609 days and was dismantled
and tested. The failure was caused by
embrittlement and cracking resulting from
exposure of the top of the pouch to light.
The PB sheeting did not contain carbon
black. The remaining five of the original
six pouches survived to 1.887 days of
exposure, when they were disassembled so
that the empty pouches and the contents
could be weighed separately, the contents
analyzed, and the properties of the pouch
walls measured.
All of the pouches had more or less
retained their original flexibility
or had swollen and softened during the
exposure period. The water in the outer
bags had a chemical or acrid odor indicat-
ing migration of constituents from the
waste liquid in the pouches to the water in
the outer bag.
The CPE pouch (P19) appeared to be in
good shape at the time of dismantling,
although the seams were relatively easy to
pull apart. The surface on the inside of
the pouch was somewhat pitted and the mem-
brane had swollen.
The CSPE pouch (P18) had blistered
badly on one side even though the sheeting
was not fabric-reinforced. The liquid
contents of the blisters were removed with
a hypodermic needle and analyzed for pH and
electrical conductivity. The pH of the
liquid in the blister was 1.35, and the
electrical conductivity of a 1:1. dilution
with deionized water was 12,400 pmho. The
waste fluid had permeated the sheeting
which, when pulled by hand, delaminated
relatively easily in the area of the
blisters.
TABLE 6. MONITORING OF POUCH 19 (CPE 86) CONTAINING
THE HIGHLY ACIDIC WASTE3
Days exposed
1
5
40
93
143
386
500
552
625
790
900
949
1,160
1,169
1,174
1,199
1,427
1,887
a"HN03-HF-HOAc",
^Based upon the
sodium chloride
p. 0-261 (1983)
cPouch rinsed, a
Pouch
Change in
weight, g
-0.70
-0.52
0.01
0.65
1.05
6.78
15.78
20.28
26.58
38.27
46.38
52.34
62.98
(c)
65.43
79.68
100.00
Outer
Electrical
conductivity
pH iimho/cm
5.9 S.5
6.0 14.8
5.9 19.0
5.8 23.4
5.6 32.2
2.7 420
2.0 3,550
1.6 5,500
2.2 6,300
1.8 8,500
1.8 9,900
1.8 12,300
2.1 14,600
(c) (c)
3.1 350
2.6 900
2.0 4,000
1.8 5,900
water
Concentration0, %
0.0005
0.0008
0.0010
0.0013
0.0018
0.026
0.22
0.34
0.40
0.54
0.65
0.80
0.95
(c)
0.022
0.054
0.25
0.37
Matrecon waste W-9.
electrical conductivity of aqueous solutions of
in Handbook of Chemistry and Physics, 63rd edition,
nd liquid
in outer bag replaced with deionized water.
239
-------
'MO 1200 1400 1800 1800
Days of exposure
a. Weight change in pouch.
b. pH of outer water.
c. Electrical conductivity of outer water.
Figures. Monitoring of pouches that
contained the highly acidic
waste, "HNOa-HF-HOAc" (W-9).
The pouch made of ELPO (P17) retained
its original condition well during ex-
posure. It appeared to have changed little
during the 1887 days of exposure.
The PB pouch (P15) that had failed at
609 days was dismantled at 625 days. It
had cracked badly at the upper seam due to
light degradation. The portion of the
pouch that was below the water level of the
outer bag appeared to have retained
its flexibility and original properties.
The PVC (19) pouch (P16) was flexible
but the inside was sticky. Also, the seams
which had been made with a 50:50 solvent
mixture of tetrahydrofuran and trichloro-
ethane were easy to peel apart. The PVC
(88) pouch (P20) had stiffened somewhat
during exposure. The outside surface had
become rough and discolored compared with
the unexposed material. The inside was
considerably more discolored and pitted,
indicating some interaction between the
membrane and the acidic waste liquid.
Weights—The various weights of the
pouches and their contents are presented in
Table 7. These data confirm the monitoring
data; they show substantial increases in
the contents of the pouches of CPE (P19 ,
CSPE (P18), PVC (P16), and PVC (P20),
further demonstrating the movement of water
from the outer bag into the pouch due to
the higher concentration of dissolved
constituents in the pouch liquid.
The pH and electrical conductivity
measurements of the water in the outer bags
also show the movement of the hydrogen ions
out of the pouches into the water in which
the pouches were immersed. This result is
confirmed by the lower conductivity and
higher pH of the waste removed from the
pouches at the end of the exposure compared
with the waste placed in the pouches at the
beginning of the exposure. The ELPO pouch
had a particularly low permeability to
water and the dissolved constituents of the
waste.
Analysis of the Pouch and Outer Bag
Contents—After the contents of the res-
pective pouches were collected and weighed,
240
-------
TABLE 7. POUCH TEST OF THERMOPLASTIC AND PARTIALLY CRYSTALLINE POLYMERIC MEMBRANES WITH ACIDIC WASTE, "HNOa-HF-HOAc" (W-9) —
CHLOROSULFONATED POLYETHYLENE, CHLORINATED POLYETHYLENE, ELASTICIZED POLYOLEFIN, POLYBUTYLENE, AND POLYVINYL CHLORIDE
Data on Pouch and Waste Before and After Dismantling
Parameter
Matrecon liner serial number
Liner thickness, mil
Pouch number
Number of days exposed
Area of pouch exposed to waste
(original dimensions), of
CPE
86
22
P19
1887
0.0431
CSPE
85
33
P18
1887
0.0484
ELPO
36
22
P17
1887
0.0446
PB
98
8
P15
PVC
19
22
P16
1887
0.0443
PVC
88
20
P20
1887
0.0523
Weight of empty pouch
Unexposed, g 39.17
Conditioned5, g 57.44
After exposure, g 53.45
Change from unexposed, g 14.28
Change from conditioned, g -3.99
Weight of filled pouch
At beginning of exposure, g 250.22
After exposure, g 350.22
Change in weight, g 100.00
Waste contents of pouch
Weight at beginning of exposure, g 192.79
Weight after exposure, g 291.49
Change in weight, g 98.70
pH after exposure0 1.10
Electrical conductivity^ after
exposure, wnho/cm 110,000
Liquid in outer bag
Volume at conclusion of test, mL , 375
pH at conclusion of test 1.80
Electrical conductivity
Before water change*, umho/cm 14,600
At conclusion of test, jimho/cm 5,900
Water transmission through pouch wall9,
g/m2/day 1.213
59.03
79.46
87.49
28.46
8.03
261.93
318.80
58.87
182.47
229.04
46.57
1.00
170,000
395
1.95
2,500
3,600
0.510
27.32
26.70
28.35
1.03
1.65
128.31
131.89
3.58
101.61
103.01
1.40
0.95
170,000
520
3.25
92
86
0.017
122.98
114.16
33.24
34.09
35.18
1.94
1.09
176.60
199.10
22.50
142.51
163.54
21.03
0.90
3.0
440
463
2.0
4,800
3,200
0.252
31.68
33.45
43.14
11.46
9.69
189.47
234.90
45.43
156.02
191.86
35.84
1.05
170,000 140,000
400
1.85
9,000
4,600
0.363
aPouch failed due to cracking and emb'ri ttl ement at upper seam caused by light degradation. Film was unpigmented, i.e. it con-
tained no carbon black to absorb ultraviolet light.
bEmpty pouches conditioned for two weeks in water at 70°C before waste added. Water added to CPE pouch (P19) after one week
to prevent sides from sticking together.
cpH at beginning of test = 0.9.
^Electrical conductivity of waste at beginning of test = 200,000 umho.
eThe volume of liquid in the outer bag varied in the course of exposure due to evaporation. Deionized water was added period-
ically to cover the pouches.
ffluter liquid replaced with deionized water after 1169 days of exposure.
9Calculated from change in weight of waste in pouch at time of dismantling.
they were analyzed for pH and electrical
conductivity. The results, also pre-
sented in Table 7, show a substantial
increase in the volume and a significant
decrease in the electrical conductivity of
the waste in the pouch, indicating a
dilution by the water that had entered
through the pouch walls. Rates of trans-
mission of the water into the pouches were
calculated (Table 7). The calculations
show the CPE to be the most permeable of
the group and the ELPO the least permeable.
Although the PB bag failed prematurely, the
transmission value based on the weight
gains was 1 ow.
These transmission values correlate
well with the values calculated from the
weight increases of the pouches during the
exposure, but the fact that part of the
weight increase is a result of the ab-
sorption of water by the pouch walls
requires that the earlier values be
corrected.
Highly Alkaline Waste Liquid
Pouches fabricated from six different
membranes, CPE (77), CSPE (85), ELPO (36),
PB (98), and two PVC's (19 and 88) were
241
-------
loaded vn'th the highly alkaline waste,
slopwater (W-4). In this series two
pouches of ELPO and one each of the other
membranes were placed in test. Premature
seam opening occurred in four of the
pouches including an ELPO pouch which,
because of its unusual behavior during
monitoring, was replaced with a second
pouch. The CSPE and the PB pouches remain
in test at this writing at more than 2220
days.
Measurements on the pouches taken
prior to leaking or dismantling, and those
still in test at 2044 days, are presented
in Table 8. The results overall show
significant weight increases by the pouches
and increases in pH and electrical conduc-
tivity by the outer liquids. The ELPO
pouches showed low permeability early in
the exposure but much higher permeability
after 300 days. The behavior of these
pouches is of particular interest and is
described below.
ELPO Pouches
The first of two ELPO pouches was
filled with the highly alkaline waste and
put into test in February 1977. After 790
days of exposure it failed at the seam and
was dismantled and tested. Because of the
apparent change in permeability of the
pouch wall at approximately 300 days shown
in the monitoring (Figure 6), a second
pouch was placed in test in January 1979 to
verify behavior of the first pouch. The
second pouch (P30B) behaved similarly; it
showed no significant weight gain until
it was under test for about 300 days. The
first pouch (P30A) showed a greater rate
of weight increase than the second pouch
(P30B). Both pouches showed a similar
rise in the electrical conductivity of
the water outside the pouch, reaching
1,000 pmhos after approximately 300 days.
Again, Pouch 30A showed a steeper rate of
increase in electrical conductivity after
that point.
The second pouch (P30B) was dismantled
after 1718 days of test because it had
gained so much weight that bursting seemed
imminent. The pouch waste was weighed and
measured for pH and electrical conductivity
as was the outer water. The empty pouch
was weighed, measured, and tested for
changes in physical properties. Measure-
ments made on the two ELPO pouches, includ-
ing those made at the time of dismantling,
are presented in Table 9.
TABLE 8. POUCHES CONTAINING THE HIGHLY ALKALINE WASTED AT LAST MONITORING BEFORE DISMANTLING OR OBSERVATION OF LEAK
Pouch
Wall Effective
Liner thickness, areac, Pouch Elapsed
Polymer number0 mil cm2 number time, d
Outer liquid
PVC
88
20
pH
Electrical Pouch,
conductivity, change in
pmho/cm weight, g
Comments
CPE
CSPE
ELPO
ELPO
PB
PVC
86
85
36
36
98
19
22
33
22
22
8
22
486
329
486
374
486
425
P32
P31
P30A
P30B
P28
P29
65
2044d
753
1350
2044d
872
7.9
9.3
12.4
11.1
9.4
12.5
270
1800
10,800
5460
1950
14,700
2.7
36.5
213.6
227.8
16.8
152.4
Leak noted at 93 d;
dismantled at 625 d.
Pouch wrinkled and
blistered; odor in water.
Leaked at seam at 790 d;
blisters noted.
Pouch swollen; yellow crys-
tals on outer surface.
Pouch has small blisters.
Pouch very stiff; seam
486
P33
65
9.6
235
failed at 930 d.
Leak noted at 93 d;
dismantled at 625 d.
a"S1opwater," Matrecon waste W-4.
''Hatrecon liner serial number.
cArea of pouch contacting the waste liquid.
In test after more than 2,200 days.
242
-------
MEASUREMENTS ON THE TOO ELPOa POUCHES
FILLED WITH SLOPWATER WASTE (W-4)
Parameter
Exposure time, days
Pouch liquid
pH
Electrical conductivity, umho/cm
Outer water
pH
Electrical conductivity, ymho/cm
Filled pouch, original wt. , g
Final weight, g
Change in weight, g
Empty pouch, original wt., g
Final weight, g
Change in weight, g
Original area of pouch, cm^
Final area of pouch, cm^
Change in area, cm^
Rates of water transmission
into pouch, g/m^'d
Initial (0-300 d)
Intermediate (300-1200 d)
Final (1200-end)
Overall curve (0-end)
Calculated from increase
in pouch liquid
(0-end)
(300- end)
Analysis of pouch wall
Volatiles, *
Extractables, %
Original
waste and
ELPO
0
13.1
160,000
7.0
1.3
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.15
5.50
Exposed
P30A
790
11.9
105,000
12.0°
90,000b
215.13
428.72
+213.59
27.62
31.45
+3.83
486
521
+35
0.193
10.387
N/A
3.745
5.463,
7.288
10.85
2.09
pouch
P30B
1725
13.2
84,000
8.9C
3,100C
98.45
346.94
+248.49
26.25
28.98
+2.73
374
383
+9
0.220
5.129
2.702
4.113
3.812
4.530
8.47
2.85
aMatrecon liner number 36; 20 mil thickness.
bprior to seam leak the electrical conductivity of the outer liquid was
14,000 umho/cm and the pH was 12.5.
cThe outer water had a pH of 10.3 and an electrical conductivity of
8,500 umho/cm before it was replaced with new DI water on day 1682.
Discussion of Results of Testing
the ELPO Pouches
At the time the first pouch (P30A) was
dismantled, a seam had opened and a portion
of the contents had entered the outer
liquid in sufficient quantity to raise the
pH to 12 and the electrical conductivity to
90,000 ymho/cm, which approach the orig-
inal values for the waste (Table 9).
Analysis of the pouch wall showed that it
had absorbed a considerable amount of water
and had a decreased extractables content.
The large weight increase of the filled
pouch indicated that a high overall trans-
mission of water into the pouch had occur-
red after the first 300 days of exposure.
During the first 300 days, little change
in weight had occurred, as illustrated in
Figure 6.
In the case of the second pouch
(P30B), the outer liquid had become only
moderately alkaline, as shown by the pH
and electrical conductivity, indicating
that only a relatively small number of ions
had penetrated the pouch wall and entered
the outer liquid. On the other hand, the
0 200 400 600 800 1000 1200 1400 1600 1800
Time exposed,.days
a. Weight change in pouches.
0 200 400 600 800 1000 1200 1400 1600 1800
Time exposed, days
b. pH of outer water.
14,000 r
13,000
12,000
11.000
§ 10,000
•5
•5 9000
tJ 7000
| 6000
.| 5000
.2 4000
3000
2000
1000
0 200 400 600 800 1000 1200 1400 1600 1800
Time exposed, days
c. Electrical conductivity of outer water..
Figure 6. Monitoring data for ELPO pouches
P30A and P30B containing the
highly alkaline waste (W-4).
243
-------
waste liquid content of the pouch had
increased significantly, to the point that
it had affected the electrical conductivity
of the pouch liquid. The increased water
content of the pouch liquid reduced the
waste concentration and resulted in a lower
rate of water transmission into the pouch
because of the lower water concentration
gradient across the wall.
Thus, it appears that the highly
alkaline waste liquid interacted slowly
with the ELPO wall. The effect became
apparent during monitoring only after 300
days of exposure when the rates of trans-
mission of the water into the pouch and the
electrical conductivity of the outer water
increased dramatically. When the pouches
were dismantled, measurements of the
pouch wall confirmed that some losses had
occurred in the tensile strength, elonga-
tion, and tear resistance. These data
were confirmed by the results of the
immersion tests conducted in another study
on the same polymer on exposure to this
highly alkaline waste liquid (Haxo et al,
in press).
The similar results for both pouches
indicated the potential reproducibility of
the pouch procedure for assessing perme-
ability characteristics of membranes.
Additional Wastes in Pouch Experiments
Four essentially inorganic or dilute
wastes were included in our pouch experi-
ments. These were:
1. A mildly acidic waste,
2. A spent caustic waste that is
essentially a brine,
3. An industrial waste that contains
some minor amounts of organics,
and
4. A pesticide waste that has a low
concentration of organics.
The pouches in these experiments have
been in exposure for more than 1800 days.
The monitoring has indicated no unusual
behavior by any of the pouches. Pre-
liminary results of these tests are being
presented in the final report for Contract
68-03-2173 (Haxo et al, in press).
In addition to the above wastes, a
series of wastes containing oily con-
stituents, either in large or small
proportions, have been loaded into the
pouches and subjected to exposure. Only a
few of the pouches have failed, and those
remaining are being monitored at exposure
times greater than 1650 days. In most
cases, the weights of loaded pouches have
changed little, but they have changed in
hardness and, in several cases, are showing
permeation of the oily constituents to the
outside. The outside surfaces have become
sticky and, in some cases, they fluoresce
under UV light. Preliminary results are
also being reported in the final report of
Contract 68-03-2173 (Haxo et al, in press).
Water-Soluble Tracers
The use of tracers such as inorganic
salts and organic dyes has been suggested
as a means of detecting leaks in membrane
linings of impoundments that contain waste
liquids. The question arises whether a
membrane liner might allow a tracer to
permeate the membrane and thus falsely
indicate the presence of a hole. The use
of pouches appeared to be a means of
determining whether specific tracers could
permeate membrane liners.
Three water-soluble tracers were
tested in pouches of several lining
materials. These tracers included one
inorganic salt, lithium chloride, and
two water-soluble organic dyes, fluorescein
and Sevron Red. Lithium chloride is gen-
erally found only in trace amounts in soil
and has been used as a tracer in waste
impoundments to detect leaks. The organic
dyes have also been used for tracing water
flow. Information on the organic dyes
included in this study appears in Table 10.
The pouches that were fabricated and
the liquids loaded into them are listed in
Table 11. Two pouches each of HOPE and PB
were filled with 1% aqueous sodium fluores-
cein and Sevron Red solutions. Previously,
only one pouch was used for a test, but
two were used in this experiment to develop
information on the reproducibility of the
permeability observations. All the pouches
were placed in deionized water each in
individual containers.
244
-------
TABLE 10. DYES USED AS TRACERS IN POUCH EXPERIMENTS
Dye
Automate Red B
Fl uorescei n-sodi urn
Color index number
Solvent red 164
Acid yellow 73
Color
Red
Ye How- red
Solubility
Petroleum products
Water
Description
Proprietary AZO dye
C2()Hio05Na2-yel low-green
Fluorescent Yellow
Methyl violet
Sevron Red
Solvent red 175
680
Brown oil
Xylene, acetone
fluorescence in neutral or
alkaline solutions
Organic, proprietary
yellow-green fluorescence
Yellow at pH 2-3.1 Water, alcohol,
Violet at pH >3.1 chloroform
Red
Water
Cationic proprietary dye
TABLE 11. POUCHES WITH AQUEOUS TEST LIQUIDS CONTAINING WATER-SOLUBLE TRACERS
Liner mate
Polymer
PVC
PVC-OR
HDPE-A
PB
Number
137
146
180
221
rial Lithiun
Thickness, 5%
mil
30 106
30 108
20
30
Tracer
i chloride Fluoresceln
ice. n
107 110
109 111
112,113
116,117
Sevron Red
IX
114,115
118,119
Monitoring
The pouch assemblies were monitored
regularly by measuring the weight of the
pouches, measuring the pH and electrical
conductivity of the outer liquids, and by
visual observations with normal and UV
light for the permeation of the dyes. The
weight changes in the PVC pouches contain-
100
500
600
a.
200 300 400
Time exposed, days
Pouches of PVC 137. Pouch 106 contains
5% LiCl, and Pouch 107 contains 10%
Lid.
ing 5 and 10% solutions of LiCl are shown
in Figure 7.
At 573 days of exposure the pouches of
polyvinyl chloride with the LiCl solutions
increased in weight in differing amounts
depending upon the concentration of the
LiCl in the pouch and the specific PVC
membrane. The pouch with a 10% concentra-
tion of LiCl increased in weight at twice
the rate of the pouch with the 5% LiCl
solution. On the other hand, the elec-
trical conductivity of the outer water
exhibited almost no change during this
period. This result indicated that the
water passed through the pouch walls into
100
500
600
b.
200 300 400
Time exposed, days
Pouches of PVC-OR 146. Pouch 108
contains 5% LiCL, and Pouch 109
contains 10% LiCl.
Figure 7. Weight changes of PVC pouches containing 5 and 10% aqueous solutions of LiCl
during immersion in DI water.
245
-------
the pouch, but little if any lithium
chloride passed through the pouch walls
into the outer water.
All 6 pouches with 1% aqueous solu-
tion of sodium flourescein showed indi-
cations of transmission through the mem-
brane, particularly in the case of the
PVC-OR pouch. Under UV light, fluorescent
specks showed on the surface of some
pouches, in scratches, and at corners where
the membrane had been thinned during heat-
seal ing. Also, observations under UV
light indicate that a small amount of the
organic dye permeated the PVC-OR wall as
there was distinct fluorescence of the
outer water. When the pH of the outer
water was increased, traces of fluorescence
appeared under UV light for all pouches.
The gains in weight of the filled pouches
were very small.
In the case of the HDPE-A and PB
pouches that contained 1% aqueous solution
of Sevron Red, also water soluble, no signs
of dye appeared in the outer water or on
the outside of the pouches in 440 days.
The weight gains of the pouches were small,
i.e. 0.20 g for HDPE-A pouches and 0.32 g
"for the PB pouches. Based on the weights of
pouches that have been dismantled, it
appears that the weight gains were in the
pouch walls, presumably by absorption of
outer DI water. Overall, the results in-
dicate that Sevron Red probably does not
permeate the walls, but, if so, only at a
very slow rate.
Organics and the Use of Tracer Dyes
In preliminary experiments using an
organic vapor transmission test similar to
the moisture transmission E96 test, we
observed relatively high transmission rates
for some organics through polymeric mem-
brane materials. The pouch procedure
appeared to be an appropriate method to
assess the permeability of selected
materials to the organics and to assess the
permeability of organic tracer dyes through
the membranes. Consequently, a series of
experiments was performed with different
membranes and organics.
A series of pouches was prepared of
HOPE, HDPE-A, and PB (Table 12). All of
these materials were selected because of
their low solubility in organics and their
low extractables contents. Also, leakproof
pouches could be made relatively easily
with heat-sealed seams. The two solvents
that were selected were xylene and acetone
with Fluorescent Yellow and Automate Red in
1% solutions in the solvents (Table 10).
Duplicate pouches containing these solvents
were placed in glass jars of distilled
water and in the respective solvents. In
the latter configuration permeation of the
dye could be observed.
Monitoring of the pouches was per-
formed principally by measuring their
weights as a function of time. The odor of
the outer liquids in the jars was also
observed as well as the appearance of the
TABLE 12. SERIES OF POUCHES CONTAINING ORGANICS WITH DYES
Matrix of Liner Materials and Test Liquids
Liner material
Polymer Number Thickness
mil
HOPE 184 30
HDPE-A 180 20
Liquid in pouch
, Acetone3 50:50 Acetone :water Xylenec
In water •
132,133
156,157d
128,129
In acetone In water
134,135
158,159d 148,149
130,131
In acetone In water
136,137
150,151 152,1536
124,125
In xylene
138,139
154,1552
126,127
PB
221
30
140,141
142,143
144,145
146,147
Acetone + 1% Automate Red.
50:50 acetone:DI water with 1% methyl violet.
GXy1ene + 1% Fluorescent Yellow.
Acetone with 1% Fluorescent Yellow.
eXylene with 1% Automate Redl
246
-------
dyes either on the surface of the pouches
or in the outer liquid. The weight changes
are illustrated in Figures 8 and 9.
Outer xylene light
orange at 1 d
POUCHES IN XYLENE
Xylene odor, pouch surface
red, red on surface of outer
water at 1 d
POUCHES IN WATER
Pouch surface very
red at 7 d
"~—— A "'
-£ £,
.Q..
o
150
200
350
Time exposed, days
Pouches of 20
with xylene
Pouches 154
immersed in
movement of
Pouches 152
immersed
ment of
mil HDPE-A (180) filled
and 1% Automate Red.
(n) and 155 (o) were
xylene. Note lack of
xylene into the pouch.
(o) and 153 (A) were
in DI water.
xylene out
Note the move-
of the pouch.
Time exposed, days
b. Pouches of 20 mil HDPE-A (180) filled
with xylene and 1% Fluorescent Yellow.
Pouches 126 (a) and 127 (o) were
immersed in xyl ene. Note the movement
of xylene into the pouch. Pouches 124
(o) and 125 (A) were immersed in DI
water. Note the movement of xylene out
of the pouch.
Figure 8. Monitoring the weight change of
pouches filled with xylene
immersed in xylene or DI water.
1 "8
en -10
£
5-12
-14
-16
-18
-20
POUCHES IN ACETONE
Fluorescence oi
seams at 212 d
,/
P158.P159
POUCHES IN DI WATER
"^t-f£ — "Acetone odor at 64 d
N»%™*— z?^— Fluorescence on
^ seams at 1 50 d
'**••• P156 leak in seam
0 50 100 150 200 250 300 350 400
Time exposed, days
Pouches of 20 mil HDPE-A (180) filled
with acetone and 1% Fluorescent Yellow.
Pouches 158 (a) and 159 (o) were
immersed in acetone. Note lack of
movement of water or acetone. Pouches
156 (o) and 157 (A) were immersed in DI
water. Note the movement of acetone
out of the pouch.
•W 3
POUCHES IN ACETONE
POUCHES IN DI WATER
*~
50 100
150 200 250 300 350 400
Time exposed, days
Pouches of 20 mil HDPE-A (180) filled
with 50:50 acetone: DI water and 1%
methyl violet. Pouches 150 (a) and 151
(o) were immersed in acetone. Note the
movement of acetone into the pouch.
Pouches 148 (o) and 149 (A) were
immersed in DI water. Note the move-
ment of acetone out of the pouch.
Figure 9. Monitoring the weight change of
pouches filled with acetone or
50:50 acetone:DI water immersed
in acetone or DI water.
247
-------
The xylene and acetone with the
dissolved organic dyes migrated through
the HDPE-A pouch walls. The pouches that
contained the xylene with the dyes and
which were placed in pure xylene increased
in weight (Figure 8). The dye in the
xylene in the pouch permeated the liner
into the outer xylene as indicated by the
red color in the outer xylene in the case
of the Automate Red and by the fluorescence
of the outer xylene in ultraviolet light in
the case of the Fluorescent Yellow. Judg-
ing from the leveling off of the weight
increase and the magnitude, it appears that
the weight increase is due to absorption of
xylene by the pouch wall.
In the case of a similar pouch placed
in water, the xylene migrated out of the
pouch into the water but, because it is not
soluble in v/ater, the xylene floated to the
top of the outer water. The dye permeated
the pouch wall and, because it is a solid
and insoluble in water, it precipitated on
the outside surface of the pouch. The
outer water did not fluoresce in UV light.
In the case of the pouches (P156,
P157) that contained acetone with Fluore-
scent Yellow dye and placed in water, the
acetone also permeated the pouch wall; but,
because acetone is totally miscible with
the water, it dissolved .in the water to
form a dilute solution. The dye also
permeated the pouch wall and precipitated
on the outer surface causing it to fluo-
resce under UV light. When acetone was the
outer liquid, the dye permeated into the
outer acetone but the pouches did not
change in weight. This behavior in-
dicates that the pouch wall is not absorb-
ing acetone.
To assess the effects on concentra-
tion on transmission rates, another series
of 20 mil HDPE-A pouches containing a 50:50
mixture of acetone and water was placed in
acetone and in DI water in closed quart
jars. Methyl violet, which is soluble in
both acetone and water, was added as a
tracer to the mixture at 1% concentration.
Changes in weight of the pouches up to
to 300 days are shown in Figure 9b. The
pouches in water lost weight, leveling
off as the acetone concentration in the
pouch dropped and that in the outer water
increased. The pouches placed in the
acetone as the outer liquid gained weight
as the acetone permeated into the pouch.
Again, the rate of transmission dropped as
the concentration of acetone in the pouch
increased.
The initial rate at which the acetone
in the 50:50 mixture moved into the outer
water was less than half that of the
acetone in the pouch with the 100% acetone
(compare Figures 9a and 9b). Calculated
rates are, respectively, 1.68 vs 5.68
g/m^d for the losses of acetone from the
pouch. The methyl violet did not appear to
have permeated the HDPE-A walls.
DISCUSSION
The primary value of the pouch test is
its ability to assess the permeability of
liner materials in a condition that sim-
ulates some aspects of a lining in a pond.
The results indicate the extent of permea-
tion of several waste constituents sim-
ultaneously — water, dissolved ions (both
cations and anions), and organics.
Permeability to Mater
The pouches with the highly acidic
waste (W9) showed the transmission of water
from outside to inside the pouches.
Deionized water in the outer bag was
transmitted into the concentrated waste
inside the pouch by osmosis resulting from
the large concentration gradient across the
wall (Table 7). Even; more striking was the
behavior of the pouches containing the
highly alkaline waste (slopwater), parti-
cularly the pouch fabricated of ELPO
which gained major amounts of water.
Of the nine pouches filled with the 5%
NaCT aqueous solution, eight showed the
transmission of DI water from the outer bag
into the pouch, as judged by the increased
weight of the pouch contents (Table 4).
The ninth pouch, CSPE 55 (P76), gained in
overall weight but showed a weight loss for
the liquid inside the pouch, which in-
dicates absorption of the pouch liquid by
the wall and probably little or no trans-
mission through the wall:.
The effect of the concentration.
gradient across the pouch wall is cTearTy
demonstrated in the PVC pouches with LiCl
(Table 13). The pouches containing
5% LiCl show a rate of transmission that is
approximately half that of the pouches with
10% LiCl. The pouches with solvents
248
-------
Permeability to Ions
TRANSMISSION RATES3 THROUGH WALLS OF
POUCHES IMMERSED IN DEIONIZED WATER
Parameter
Liner number
Thickness, mil
Specific gravity
Extractables, %
HOPE
184
30
0.951
0.73b
HDPE-A
180
20
0.949
2.09b
Polymer
PB
221
30
0.907
3.68b
PVC
137
29.0
' 1.283
30.14C
PVC-OR
146
30.0
1.239
34.61C
Area of pouch, cm2 142
Pouch liquid
Acetone, g/m2d
50:50 acetone:
water, g/m2d
-0.866-A -6.53d-A -1.316-A
(P132,P133) (P128.P129) (P140.P141)
-5.68e-A
(P156.P157)
-1.68f-A
(P148.P149)
+2.09f.9-A
(P150.P151)
5% Lid water
solution, g/m^d
10* Lid water
solution, g/m^d
+0.143-W W.208-W
(P106) (P108)
+0.286-W +0.448-W
(P107) (P109)
Xylene, g/m2d -1.788-X -16.84e-X -4.40-X
(P136.P137) (P124,P125) (P144.P145)
-8.48
-------
wastes have permeated the pouch walls and
dissolved in the water in the outer bags,
but direct analyses of these waters were
not made.
pouches immersed in DI water; both dyes are
insoluble in water. These results again
show the permeability of polymeric mem-
branes to organic molecules.
Permeability to Organic Dyes
The dyes are also organic chemicals
and show varying permeation. Three water-
soluble and two organic-soluble dyes were
studied in pouch experiments at 1% con-
centrations. The water-soluble types
included sodium fluorescein, Sevron Red,
and methyl violet. The sodium fluorescein
showed positive transmission through
PVC-OR, and trace transmission through
HOPE, PB, and PVC. Sevron Red in water
showed no detectable transmission through
HDPE-A and PB. Methyl violet in the 50:50
acetone:water mixture showed no detectable
transmission through HDPE-A. On the other
hand, the two organic-soluble dyes, Auto-
mate Red and Fluorescent Yellow, in the
pouches containing acetone and xylene
quickly permeated the HDPE-A pouch walls.
These dyes dissolved in the outer liquid
of the pouches immersed in xylene or ace-
tone and precipitated on the outside of the
Comparison of Pouch Results
with Other Methods
The relative ordering of the perme-
abilities among the six polymeric membranes
is presented in Table 14 for water and
hydrogen ions as measured by pH and elec-
trical conductivity. In the same table the
relative permeabilities as measured by
moisture vapor transmission in the E-96
test are also presented.
The pouch test results indicate that
the ELPO has the lowest permeability to
water and to the hydrogen ions and the PB
has the second lowest. The highest perme-
abilities were shown by the CPE and the two
PVC's. The pouch of CPE showed a low
permeability during the early stages,
comparable to the values indicated by the
moisture vapor transmission; however, after
a year of exposure, during which the
changes in weight of the pouch, the elec-
trical conductivity, and pH were small,
TABLE 14. ORDERS 0F POLYMERIC MEMBRANE PERMEABILITY TO WATER AND HYDROGEN IONS, IN POUCH TEST
WITH ACIDIC WASTE, "HN03-HF-HOAc" (W-9), AND TO MOISTURE VAPOR IN ASTM E96 TEST
Pouch Status at 1887 d
Liner material
Waterb
Polymer
Number6 Thickness,
rail
Electrical
conductivity0,
pmhos/cm
Moisture vapor
transmission, E96
g/m2 dfOrder
Chlorinated poly-
ethyl ene 86 22
Chlorosulfonated
polyethylene 85 33
Elasticized poly-
0.643
0.438
olefin
Polybutyl ene
Polyvinyl chloride
36
98
19
88
22
8
22
20
1
19
2
3
. 1
29
4
5
1
...
4
3
0.142
• • •
2.78
2.94
1
• • •
4
5
aOrder of increasing permeation through polymeric membranes in test. 1 = Least permeable and 6 = most
permeable.
bBased on change in weight of filled pouch for 400 days of exposure and for 1800 days of exposure.
cElectrical conductivity of water in outer bags.
dpH of water in outer bags.
^atrecon liner number and thickness, in mils, in parentheses.
^Test values.
SExtrapolated.
250
-------
significant increases in weight and con-
ductivity took place. These effects
would indicate that the permeability of
the CPE increased, probably because of high
water absorption. Also, pinholes may have
developed in the membrane during the
exposure period as the inside of the pouch
had a pitted appearance. A failure such as
occurred with the PB pouch did not occur at
the seams of the CPE pouch to cause the
very abrupt increase in conductivity.
There was an indication that the
relative order of permeability is not
the same for all of the constituents of a
waste. In addition, maintaining a membrane
in a moist condition appears to be neces-
sary to assess the long-term permeability
of membranes that may slowly become af-
fected by the waste liquid. This does not
occur in the ASTM E96 moisture vapor
transmission test, in which one side of a
membrane is at 50% humidity, and the other
side is maintained at 100% humidity.
Nevertheless, even under these conditions,
some increases in moisture vapor trans-
mission have been observed in long-term
tests.
Expanding the Pouch Test Method
to all Polymeric Membranes
The pouch method for assessing the
permeability of membranes over long
periods of time and of the durability of
the materials in contact with wastes
appears to be a highly feasible method,
though extended exposures may be required.
This conclusion is based on the results
with the more flexible sheetings of
thermoplastic and partially crystalline
materials. Expanding this test method
to cross!inked polymers and elastomers, to
thicker and more rigid sheetings, and to
fabric-reinforced materials appears to be
desirable. Achieving good seams in the
pouches may require the development of
special adhesives and possibly new pouch
designs that can accommodate the more rigid
materials, e.g. 100 mil HOPE. It is desir-
able, of course, to use the commercial
sheetings to fabricate these pouches and
not to depend upon molded containers.
ACKNOWLEDGMENTS
The work reported in this paper was
performed under Contract 68-03-2134,
"Evaluation of Liner Materials Exposed to
Leachate", Contract 68-03-2173, "Evaluation
of Liner Materials Exposed to Hazardous and
Toxic Wastes", and Contract 68-03-2969,
"Long-term Testing of Liner Materials",
with the Municipal Environmental Research
Laboratory of the U.S. Environmental
Protection Agency, Cincinnati, Ohio.
The support and guidance of Mr. R. E.
Landreth, Project Officer, are gratefully
acknowledged.
REFERENCES
1. Clear, K. C. and E. T. Harrigan, 1977.
"Sampling and Testing for Chloride
Ion in Concrete11"! Report .No. FHWA-
RD-77-85, Federal Highway Administra-
tion, Washington, D.C. Interim Report.
2. CRC Handbook of Chemistry and Physics,
1982. 63rd edition. CRC Press, Inc.,
Boca Raton, Florida, p261.
3. Haxo, H. E., R. S. Haxo, N. A. Nelson,
P. D. Haxo, R. M. White, and S.
Dakessian, in draft.. Final Report:
Liner Materials Exposed to Hazardous
and Toxic Sludges.
ProtectionAgency,
(In press).
U.S. Environmental
Cincinnati. Ohio.
4. Haxo, H. E., R. M. White, P. D. Haxo,
and M. A. Fong, 1982. Final Report:
Evaluation of Liner Materials Exposed
to Municipal Solid Waste Leachate.
U.S. Environmental Protection Agency,
Cincinnati, Ohio. NTIS No. PB 83-147-
801.
5. Matrecon, Inc., 1983. Lining of Waste
Impoundment and Disposal Facilities.
SW-870 Revised. U.S. Environmental
Protection Agency, Washington, D.C.,
448 pp. 6PO #055-00000231-2.
251
U.S. GOVERNMENT PRINTING OFFICE: 1984 _ 759-10Z/0881
-------
-------