PB81-213U23
Impact of Organic Solvents on the Integrity of
Clay Liners for Industrial Waste Disposal Pits
Implications for Groundwater Contamination
Colorado State Uni?., Fort Collins
Prepared for
Rooert S. Kerr Environmental Research Lab.
Ada, OK
Jun 79
j
Service
-------
FINAL REPORT ON
IMPACT OF ORGANIC SOLVENTS ON THF INTEGPITY OF
CLAY LINERS FOR INDUSTRIAL WASTE DISPOSAL PITS:
IMPLICATIONS FOR GROUNDWATER CONTAMINATION
BY
William J. Green, G. Fred Lee and R. Anne Jones
Department of Civil Engineering
Environmental Engineering Program
Colorado State University
Fort Collins, Colorado 8P|>23
Project No. R-SOU SM9 010-02-0
June 1979
Project Officer
D. Craig Shew
Robert S. Kerr Environmental Laboratory
US Environmental Protection Agency
Ada, Oklahoma 74820
-------
TECHNICAL REPORT DATA
(Pleasf nod Instructions on the reverse before completing)
1. REPORT NO.
IIENT'S ACCESSION NO
4. T.TLE AND SUBTITLE Tmpact of Organic 'Solvents on the
Integrity of Clay Liners for Industrial Waste Disposal
Pits: Implications for Groundwater Contamination
(Final Report)
5. REPOH1 DATE
June 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
William J. Green; G. Fred Lee; and R. Anne Jones
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Colorado State University
Department of Civil Engineering
Environmental Engineering Program
Fort Collins, CO 80523
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AT.f NCY NAME AND AC-DRESS
U.S. Environmental K-itection Agency
Robert S. Kerr Environmental Laboratory
Ada, OK 74820
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
16. SUPPLEMENTARY NOTES
16. ABSTRACT
This research project was aimed toward development of procedures by which
to evaluate the potential for groundwater contamination by hazardous industrial
wastes disposed of on land. Studies were conducted to describe the attenuatlve
properties of the soil, soil permeability, and the effects of a number of common
Components of hazardous wastes on \.h* characteristics of so^ls and standard clays.
This study represents the first systematic effort to determine the effects of bulk
organic solvents and solvent/water mixtures on the characteristics of clays. An
attempt was also made to develop a model to predict the potential impact on ground-
water quality that may result when a particular type of hazardous waste is disposed
of onto a particular type of land or remolded clay barrier (1-ner).
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b IDENTIFIERS/OPEN ENO£D TERMS c. COSATI Kwld/Group
IB. DISTRIBUTION STATEMENT
19. SECURITY CLASS
21. NO. OF PAGES
157
2O. SECURITY CLASS (ThltfHgf/
22. PRICE
tPA Form 2220-1 (••?!)
-------
CONTENTS
Figure iv
Tables ix
Acknowledgement • xi
1. Introduction 1
2. Summary and Conclusions 3
3. Recommendations 7
U. Literature Peview . . 9
Introduction ..... 9
Studies of Movement of Industrial Wastes to
Groundwater 10
Experimental On-Land Disposal Practices ..... 12
Sorption-Desorpticn and Ion Exchange iu
Permeability of Soils 18
Atterberg Limits 32
Clay-Organic Studies: X-Ray Investigation of
Swelling 33
Octanol/Water Partition Coefficient ...... -. 35
5. Methods 38
Introduction 38
Clay Mir.svalogy ................. 38
Particle Size Analysis 38
Moisture Density Relationships .... 39
Atterberg Limits UO
i
-------
CONTENTS (continued)
Specific Gravity .... ............ 41
Cation Exchange Capacity ............ Ul
Total Carbon and Carbonate Analysis ....... 42
Swell Properties ................ ^2
Coefficient of Permeability for Liquids and
Liquid Mixtures in Contact with Clays ...... »»3
X-Ray Diffraction Studies of Clays in Contact
with Solvents and Solvent Mixtures ....... >*6
Determination of Major Ions
Determination of Sorption Isotherms for
Volatile Organics ................ u6
5. Results and Discussion .... ........... 51
Introduction .................. 51
Characteristics of Clays Used in This
Investigation .... .......... .... 51
Effects of Organic Solvents on Clay Lattice
Spacing ..................... 61
Sorption Isotherm Studies ............ 67
Swelling of Clays in Contact with Water,
Organic Solvents, and Solvent Mixtures ..... 81
Effects of Organic Solvents on Permeability . . . 106
Concluding Remarks ............... 129
7. Survey of Current Practices for Land Disposal of
Hazardous Wastes .......... . . . ..... 13u
Introduction .................. 13U
Summary of Responses ........ , ..... 13U
References ......... . .............. 136
ii
-------
CONTENTS (continued)
Appendices
A. Data on Swell Properties of Clays Al
B. Data on Permeability of Clays to Organic
Solvents Bl
C. Suntnary of Responses of State Pollution Control
Agencies to Requests for Information on Current
Regulations and Practices for Land Disposal of
Hazardous Wastes Cl
D. Proposed EPA Hazardous Waste Guidelines Dl
-------
FIGURES
Number Page
1 Constant-head permeameter ............ 23
2 . Falling-head permeameter ............ 25
3 Effect of pumping on-line groundwater level ... 27
i* Atrerb*rg limits and indices . . , ........ 3H
5 Pressure permeability column .......... "*"*
6 Change in peak height (concentration) with time
for benzene sorbed onto Ranger Shale ...... **9
7 Moisture-density relationship for Kosse-Kaoline . 5?
8 Moisture-density relationship for Ranger Shale . 55
9 Moisture-density relationship for Fire Clay ... 55
10 Moisture-density relationship for Parker Soil . . 55
11 Fr^undlich isotherm for benzene sorption on
Ranger Shale ................... 72
12 Freundlich isotherm for m-xylene sorption on
Ranger Shale .................. 72
13 Freundlich isotherm /or carbon tetrachloride
sorption on Ranger Shale ............ 72
in FreundiJ.ch isotherm for acetone sorptior. on Ranger
Shale ...................... 72
15 Freundlich isothonr for benzene sorption on Kosse
Kaoline ..................... 73
IE Freundlich isotherm for m-xylene sorption on
Kcsse Kaoline .................. 73
17 Freundlich isotherm for carbon tetrachloride
sorption on Kosse Kaoline ............ 73
iv
-------
Figures (continued)
Number Page
18 Freundlich isotherm for benzene sorption on
Fire Clay 74
19 Freundlich isotherm for m-xylene sorption on
Fire' Clay 74
20 Freundlich isotherm for carbon tetrachloride
sorption on Fire Clay 7U
21 Chromatogram of m-xylene standard solution ... 78
22 Chromatograjn of m-xylene standard solution after
24-hr contact with Ranger Shale 78
23 Chromatogram of standard solution containing
a mixture of xylenes 79
24 Chromatogram of standard solution of xylenes
after three week exposure to Ranger Shale .... 79
25 Swell properties for Ranger Shale in benzene . . 82
26 Swell properties fcr Fire Clay in benzene .... 82
27 Swell properties for Ranger Shale in xylene ... 82
28 Swell properties for Fire Clay in xylene .... 8?
29 Swell properties for Kosse Kaoline in xylene . . 33
30 Swell properties for Ranger Shale in carbon
tetrachloride 83
31 Swell properties for Kosse Kaoline in carbon
tetrachlorid* 83
32 Swell properties for Fire Clay in carbon
tetrachloride <33
*
33 Swell properties for Ranger Shale in trichloro-
ethylane 85
34 Swell properties for Kosse Kaoline in trichloro-
ethylene 85
35 Swell properties for Fire Clay in trichloro-
ethylene 85
-------
Figures (continued)
N .unber Page
35 Swell properties i:or Ranger Shale in acetone
and in 50 volume percent acetone-water mixture . . 85
37 Swell properties for Kosse Kaoline 86
38 Swell properties for Fire Clay in acetone .... 86
3? Swell properties for Ranger Shale in deionized
water 86
UO Swell properties for Ranger Shale in deionized
water (duplicate) 86
Ul Swell properties for Kosse Kaoline in dsionized
water 87
42 Swell properties for Fire Clay in deionized
water 87
U3 Swell properties for Ranger Shale in methanol . . 87
uu Swell properties for Ranger Shale in glycerol . . 87
U5 Swell properties for Ranger Shale in mixcure of
acetone (25 mole t)and benzene (75 mole %)....' 88
U6 Swell properties for Ranger Shale in mixture of
acetone (50 mole Dand benzene v'50 mole %).... 88
u? Swell properties for Ranger Shale in mixture of
acetone (75 mole %) antl benzene (25 mole %) . . . 88
U8 Swell properties for Ranger Shale in 50 volume
percent mixture of xylene and deionized water . . 88
•*8a Comparison of swell properties of Fire Clay in
CC1U, acetone, and wat».r: effects of low,
intermediate, and high dielectric solvents .... 9U
U8b Comparison of swell properties for Kosse Kaoline
in CClttt acetone, and water: affects of low,
intermediate, and higl: dielectric solvents .... 95
<*8c Comparison of swell properties of Rangar Shale
in CClu, acetone, and water: effects of low,
intermediate, and high dielectric solvents .... 96
>»9 Relationship between solvent dielectric constant
and percent swell of Ranger Shale 97
vi
-------
Figures (continued)
Number Page
50 Relationship between solvent dielectric constant
and percent swell of Kosse Kaoline ........ 98
51 Relationship between solvent dielectric constant
and percent swell of Fire Clay .......... 99
52 Coefficient of permeability of Ranger Shale to
benzene ..................... 107
53 Coefficient of permeability of Ranger Shale to
xylene ..................... 107
Sk Coefficient of permeability of Ranger Shale
to carbon tetrachloride ............ : 10?
55 Coefficient of permeability of Ranger Shale
to trichloroethylene .............. 107
56 Coefficient of permeability of Ranger Shale .
to acetone ................... 108
57 Coefficient of permeability of Ranger Shale
to methanol ................... 108
58 Coefficient of permeability of Rang»r Shale
to glycerol ................... 106
59 Coefficient of permeability of Ranger Shale
to deionized water ............... 108
60 Coefficient of permeability of KOBSC JCsolj'ne
to xylene ................... 109
61 Coefficient of permeability of Kosse Kaoline
to acetone ...................
62 Coefficient of permeability of Kosse Kaoline
to deionized water ............... 109
63 Coefficient of permeability of Fire Clay to
xylene ..................... Ill
6W Coefficient of permeability of Fire Clay to
acetone ..................... Ill
65 Coefficient of permeability of Fire Clay to
deionized water ................. Ill
vii
-------
Figures (continued)
Page
so Coefficient of permeability of Fire Clay to
trichloroethyiene Ill
67 Coefficient of permeability of Ranker Shale to a
mixture of acetone (75 mole%) and oenzene
(25 mole%) 112
68 Coefficient of permeability of Ranger Shale to a
mixture of acetone (50 DoleI) and benzene
(50 mole^) 112
69 Coefficient of permeability of Ranger Shale to
a mixture of acetone (25 mole%) and benzene
(75 mole%) 112
70 Coefficient of permeability of Ranger Shale to
a mixture of acetone (75 mole*.) and carbon
tetrachloride (25 mole!) 112
71 Representation of water-containing clay 116
72 Relationship between coefficient of permeability
and octanol/water partition coefficient 118
73 Relationship between coefficient of permeability
and dielectric constant for solvents in Ranger
Shale 121
7u Relationship between coefficient of permeability
and dielectric constant for solvents on Kosse
Kaoline 122
75 Relationship between permeability and dielectric
constant for solvents on Fire Clay 123
75 Coefficient of permeability aa a function of
packed bulk clay density 126
77 K vs. e Vd for organic solvents and water on
three clays 132
viii
-------
TABLES
Number Page
1 Conditions for Organic Solvent Analysis ..... 50
2 Particle Size Distribution ........... 52
3 Clay Mineralogy ................. 5U
u Optimum Moisture Contents ............ 56
5 Atterberg Limits ................ 56
6 Cation Exchange Capacity ............ 58
6A Exchangeable Na, K, Ca, and Mg ......... 59
7 Major Ion Content of Clays Studied ....... 60
8 . Total Carbon and Carbonate Content ....... 60
9 Results of X-Ray Diffraction Measurements of Air-
Dried Ranger Shale after Exposure to Various
Solvents .................... 62
10 Results of X-Ray Diffraction Measurements for
Air-Dried Kosse Kaoline in Contact with
Various Solvents ................ 63
11 Results of X-Ray Diffraction Measurements for
Air- Dried Fire Clay in Contact with Various
Solvents .................... 6"
12 Results of X-Ray Diffraction Measurements for
Air-Dried Parker Soil in Contact with Various
Solvents .................... 65
13 Results of X-Ray Diffraction Measurements of
Air-Dried Montmori3 \onite After Exposure to
Various Solvent-Water Mixtures ......... 66
lu d(001) Spacing for Clays in Contact with Pure
Solvents .................... 67
15 Sorption of Organic Solvents on Ranger Shale . . 69
16 Sorption of Organic Solvents on Kosse Kaoline . . 70
ix
-------
Tables (continued)
Number Page
17 Sorption of Organic Solvents on Fire Clay .... 71
18 Freundlich Equation Constants for Sorption of
Organic Solvents Onto Clays 75
19 Sorption of Organic Coir.pou/ids Onto Clays At
Equilibrium Concentrations of 1 mM 76
20 Percent Swell of Test Clays in Contact with
Liquids Compared to Dielectric Constants and
Dipole Moments of Liquid 93
21 Percent Swell of Clays Due To Hydrophobia Solvents
In Relation To Organic Carbon Content of Clay . . 100
22 Classification of Clay-Organic Solvent Systems
According to Swell Properties 101
23 Swell Properties of Ranger Shale in Five Solvent
Mixtures 103
2U Equilibrium Permeabilities of Clays to Organic
Solvents and Water 113
25 Density-Viscosity Ratios for Organic Solvents
and Water II1*
26 Permeabilities and Percent Swell for All Single
Solvent Systems Tested ....'. 115
27 Permeabilities and Octanol/Water Partition
Coefficients for Solvents on Three Clays .... 119
28 Clay-Soil Characteristics 125
29 Coefficients of Permeability for Four Binary
Mixtures I/*/
30 Coefficients of Permeability for Typical Soils . 129
31 Calculated and Observed Permeability Values for
Eleven Single Solvent Systems ... 133
Al - A25 A1-A27
Bl - D19 B1-B23
-------
ACKNOWLEDGEMENT
The majority of this study was conducted at the University
of Texas at Dallas (UTD); the latter phases were completed at
Colorado State University (CSU). Support of this project was
provided by the US EPA Robert S. Kerr Environmental Research
Laboratory, Ada, OK; the Department of Civil Engineering, Colo-
rado State University, Fort Collins, CO; the University of Texas
at Dallas, Richardson, TX; and EnviroQual Consultants and
Laboratories, Fort Collins, CO. The authors wish to acknowledge
Dr. Craig Shew and Jack Keeley of the US EPA-Ada for their
assistance during this study. Ted Palit, graduate student at
UTD, provided assistance in conducting laboratory tests for de-
termining clay characteristics, permeability and swelling, and
in preparation of draft material for the literature review of
this report. Mike Day made significant contributions to the
X-ray analysis and sorption sections of this study and was
involved in major cation determination of the clays. His assis-
tance in this work is greatly appreciated. We acknowledge, too,
the help which Lyle Wynette provided during the permeability
studies. Dr. Carter of the University of Texas at Dallas pro-
vided assistance in the major ion content of clay analyses made
in this s* idy. K. Horstman, a graduate student in the Environ-
mental Engineering Program at Colorado State University provided
assistance in summarizing State regulations and procedures
governing on-land disposal of hazardous wastes.
The authors also wish to acknowledge the assistance of
G. Max and R. Leydon of the UTD and CSU Offices of Sponsored
Research, respectively, and the secretarial assistance of M.
Jaye, P. Wernsing and T. Smith.
xi
-------
SECTION 1
INTRODUCTION
During the past several years, increasing emphasis has been
placed on control of air and water pollution at the federal,
state, and local levels. With stricter controls on air and water
pollution and the increasing costs of associated pollution con-
trol equipment and fuel, interest in the land disposal of munic-
ipal and industrial wastes has increased. This, coupled with
the essentially complete prohibition of ocean dumping has re-
sulted in the on-land disposal of l*rge amounts of processing-
manufacturing chemical wastes. While land disposal of waste is
often relatively easy and relatively inexpensive where inexpen-
sive land is readily available, it is not without significant
potential environmental problems, the most important of which is
groundwater contamination.
Protection of the quality of groundwater resources of the US
has been a grossly neglected area which needs immediate atten-
tion. Although the passage of the 197U "Drinking Water Act" pro-
vides the legislation necessary to protect the US groundwater
resources from contamination from wastes disposed of on land, in
many cases the techniques and methodologies have not been avail-
able to property implement or enforce these regulations. For ex-
ample, this act makes provisions to prohibit on-land disposal
of municipal and industrial waste where there is a significant
potential for groundwater contamination. However, in order to
properly implement this provision, procedures are needed which
can be used to predict the transport of municipal and industrial
waste components fron their on-land disposal site to groundwaters.
When developed, these tests should be employed routinely to pro-
vide the regulatory agencies and the individuals responsible for
disposal of wastes, with "go" or "no go" information for a par-
ticular type of waste in a particular land disposal system.
Although wastes may contain one or more hazardous substances,
whether they pose a threat of groundwater pollution or not will
depend on a variety of factors including: quantity of waste,
leachability, rainfall, permeability, attenuative properties of
the soil, and the distance to and quality of the groundwater.
In most cases, it is not known beforehand whether disposal of a
given waste at a given on-land site will result in groundwater
-------
contamination. In fact, due to limited or non-existirg ground-
water monitoring, pollution incident;, are not usually predicted,
nor are they realized, until actua.'. damage has been experienced.
To protect public health and the environment, it is desir-
able to develop a procedure or procedures that can be used for
any given sice to 1) evaluate the potential for groundwater
degradation from potentially hazardous wastes, and 2) determine
if a potentially harmful quantity of a waste is to be involved
in a given disposal situation. These procedures should be in-
corporated into the standard site evaluation process used by
regulatory agencies.
This research project was aimed toward development of proce-
dures by which to evaluate the potential for groundwater con-
tamination by hazardous industrial wastes disposed of on land.
Studies were conducted to describe the attenuative properties of
the soil, soil permeability, and th? effects of a number of com-
mon components of hazardous wastes on the characteristics of
soils and standard clays. This study represents the first sys-
tematic effort to determine the effects of bulk organic solvent!
and solvent/water mixtures on the characteristics of clays. Ar
attempt was also made to develop a model to predict the poten-
tial impact on groundwater quality that may result when a
particular type of hazardous waste is disposed of onto a particu-
lar type of land or remolded clay barrier (liner). It should be
noted that while throughout this report, the clay-soils Kosse
KaoLinc, Fire Clay, and Ranger Shale are referred to as "clays,"
only portions of these materials were clay minerals.
-------
SECTION 2
SUMMARY AND CONCLUSIONS
Increasing amounts of industrial wastes are being disposed
of on land. This is a result of a variety of factors including
increasingly stringent regulations to protect against air and
water pollution. Land disposal of municipal and industrial waste
carries with it a significant potential for groundwater contami-
nation; there are numerous instances of groundwater pollution
from such disposal operations. As part of this study, a review
of state regulations governing land disposal has been conducted.
It has been found that only a few of the 50 states have defini-
tive regulations. Some states on the other hand, specify the
maximum permeability of the subsoil strata in the disposal area,
require the use of clay liners of a given thickness and permea-
bility if the natural permeability is greater than the specified
amount, and regulate the types of wastes that may be disposed of
in these pits. Many of the states that specify a maximum permea-
bility have not indicated how permeability tests should be con-
ducted and in particular th-s types of liquids that should be used
in the tests. It is generally recognized that there is need to
develop a more technically valid approach for the control of land
disposal of industrial wastes than exists today. Since many in-
dustrial wante disposal pits receive large amounts of organic
solvents, it is conceivable that these solvents could be readily
transported to groundwater.
The data obtained during this study have provided a basis
for predicting the movement of hazardous organic chemicals
through clays and soils. Models for the behavior of organic
substances in dilute aqueous solutions in contact with clays,
and for clays immersed in bulk liquids and liquid mixtures have
been developed. These will aid pollution control agencies in
assessing the potential environmental hazard associated with
land disposal of industrial wastes. The specific conclusions
reached are outlined below:
1. Sorption is an important parameter for assessing the
attenuation of contaminants by clay liners and by soil
and subsoil systems. Because of the large quantities
involved in some industrial waste disposal operations,
the crganic solvents frequently used in manufacturing
represent a significant potential for groundwater
-------
contamination when disposed of on land. It is there-
fore important to examine the sorption of organic sol-
vents on clays. Techniques have been developed during
this study for obtaining valid sorptior. data for highly
volatile organic compounds such as organic solvents.
These techniques involve taking precautions to minimize
the number of transfer steps, reducing the free volume
above the liquid (head space) to a few cubic centimeters
and preparing solution concentrations accurately using
methods described in this report.
Based on the systems studied, sorption of hydrophobic
solutes by the clays was extensive in most cases and
showed a marked increase with compounds of increasing
octanol/water partition coefficients. Of the three com-
pounds tested with each of three clays, xylene (highest
octanol/water partition coefficient of the three) con-
sistently showed the greatest amount of sorption. Those
clays richest in kaolinite tended to be the most effec-
tive sorbents.
It has been concluded that sorption by clays is an im-
portant mechanism by which the transport of low molec-
ular weight organics from waste disposal pits to ground-
water is minimized.
The shrink-swell properties of clays in contact with
solvent? is of importance in determining the ability of
clay liners in industrial waste disposal pits to retain
wastes. This study was the first systematic effort to
examine the effects of bulk organic solvents and sol-
vent mixtures on the swelling of clays. It can be con-
cluded that swelling probably occurred through inter-
particle separation rather than through intercalation
or complex formation. For a first approximation, the
solvent dielectric constant gave an indication of the
degree of swell of the clay with which it is in contact.
The montmorillonite (expandable layer clay) content of
the clay was of no consequence in predicting swell, but
there was a correlation between percent clay swell and
the percent organic carbon content of the clay.
In some pure solvent systems, notably those having low
dielectric constants, shrinking occurred probably due
to dehydration of the clay by the solvent. That is,
there was a transfer of water out of the clay with no
commensurate replacement by solvent molecules.
With respect to swell, each component of a bolvent mix-
ture appeared to have behaved independently of the
others, so that swelling of clays was caused solely by
that component having the greater (greatest) dielectric
-------
constant. In systems with a mixture of solvents* the
degree cf swelling of the clay will be roughly The same
as it is in a system of the clay with the pure solvent
of the greater (greatest) dielectric constant. Wastes
containing solvents with only low dielectric consxants
should be mixed with a high dielectric solvent such as
water during disposal in order to prevent cracking of
clay liners. Additional work must be done on the effect
of clay organic carbon content on the shrink-swell prop-
erties of a clay-solvent system in order to determine
the significance of thi? parameter in influencing the
ability of clay liners to retain industrial wastes.
The primary objective of t.Kis study was to determine the
effect of organic solvents on the permeability of clay
liners, soils, and subsoil systems to water-organic mix-
tures. Ar important contribution of this work has been
the development of permeability test techniques for
organic solvents. The conventionally-used techniques
suitable for water are not satisfactory for use with
many organic solvents because of the interaction between
the sealants used and the solvent. The new procedure
involved packing the clay directly into the column with-
out the use of sealants.
A somewhat surprising and important conclusion of this
study was that the permeability of packed clay columns
to organic solvents was less than to water. This can
b<* Attributed to the tendency of an organic molecule to
escape from the aqueous, inter-aggregate phase and to
thus be strongly sorbed on the clays, and to various
oth«r factors including microbial decomposition of the
or^inic solvent and subsequent liberation of gases which
clog clay pores. This biological interference was
likely a factor in the permeability test results ob-
tained for Ranger Shale.
The correlation between the permeability of solvents
through a clay, K, and their octanol/water partition
coefficients, P, was such that the larger the value of
P (greater escaping tendency from the aqueous phase),
the lower the value of K. Also, there is an apparent
relationship for a given clay between the solvent di-
electric constant and the coefficient of permeability;
the larger the dielectric constant, the greater the
flow permeability. As expected, the coefficient of
permeability is also markedly dependent upon the
packed, bulk clay density; th»a denser the clay the
lower the equilibrium permeability coefficient.
The behavior of solvent mixtures in clay column tests
was in general similar to the behavior of the most polar
-------
solvent component of the system, both in terms of the
equilibrium values of K, and the variation of K with
time.
During permeability tests with carbon tetrachloride,
xylene, and benzene, breakthrough was observed. How-
ever, when benzene and CCl^ were mixed individually
with acetone, breakthrough was not observed over the
36-day test period. This phenomenon of breakthrough
seemed to be related to the shrinking of the clays
noted above. It is concluded that designed co-disposal
of solvents may be an effective way of eliminating po-
tential cracking of clay liners.
Based on the single solvent systems examined, the fol-
lowing empirical relationship has been deriv*d for esti-
nating the coefficient of permeability, K, (in cm/sec):
log K = 1.17 (log
-------
SECTION 3
RECOMMENDATIONS
This study represented the first phase of a planned longer-
term investigation designed to ascertain the potential for siigra-
tion of organic solvents from industrial waste disposal pits to
groundwater. The first phase of this study was to be devoted to
pure solvent-defined clay systems in order to determine if any
generalized relationships could be developed under controlled
laboratory, simplified conditions. Based on these studies, the
following recommendations have been made.
1. All liquid organic wastes should be tested prior to be-
ing disposed of on land, using the consolidometer, to
determine whether they cause the specific clay of inter-
est to shrink or swell. If shrinking is observed, as
may be the case for low dielectric solvents, a spent:
solvent of higher dielectric constant or water should be
added so that the mixture gives a smooth swelling pro-
file. It is felt th.it frequency and extent of testing
of this type can be significantly reduced once the gen-
eral characteristics of the waste from a particular
manufacturing process have been ascertained.
2. Using the relationship developed in this study an esti-
mate should be made of the permeability coefficient of a
waste material prior to disposal, to insure that the
value of permeability is within the limits set by state
regulations.
3. Because of the potential value of an equation which can
be used to estimate permeability coefficients for a
variety of diverse systems, further efforts should be
directed toward establishing such a relationship, appli-
cable to both pure solvents and to the most complicated
mixtures. The direction that such an effort might take
has been indicated ir. this study, but there is need for
more laboratory data, particularly on mixtures, and for
more sophisticated investigations to establish the mech-
anism of solvent movement through clay.
"4. Additional column tests shculd be run using "sterilized"
clays in order to avoid the complications which arise
-------
from micrcbial activity. Using such a column a greater
range and diversity cf solvents and mixtxires should be
studied. Also, clays of different bulk densities and
raoi ..ture contents should be tested in order to determine
the effects of the.se parameters on permeability.
5. The apparent importance of clay shrinkage as it relates
to shrinkage of clay liners used in waste disposal pits
warrants that lurther work be carried out in this area
in order that this phenomenon can be predicted given a
knowledge of clay and solvent properties.
6. Field tests should be conducted, both with substances
such as CClij, benzene, and xylene which show a tendency
to migrate in bulk through clays, and with these sub-
stances mixed with acetone, mixtures which are well
behaved in the laboratory. There is need to determine
whether the conclusions reached in the laboratory phase
of this study are equally valid under field conditions
for a variety of natural substrates, clay linen, and
waste mixtures.
-------
SECTION «*
LITERATURE REVIEW*
INTRODUCTION
The groundwater supply of tho United States is a major nat-
ural resource, which, until recently has been considered virtual-
ly immune to contamination by the activities of man. Yet du.-ing
the past decade, numerous incidents of groundwater pollution from
chemical spills and landfill seepage have been reported and have
shown that even our sequestered reservoirs of potable water are
vulnerable. The potential seriousness of the problem becomes
apparent when it is realized that groundwater provides water for
about 100 million people in the U.S. (Jeffrey, I97>Of nearly half
of the population.
Concern over the quality of the nation's groundwater is re-
flecto- in the clarity, specificity and stringency of recent
legislation aimed at protecting this resource. The Safe Drinking
Water Act of 197<* was the first major pisce of water management
legislation to specifically recognize groundwater as an indis-
pensibi? segment of the total national water resource. This act
was int'jndec to protect underground source1* of drinking water
by regulating both deep-well injection of chemicals and on-lana
disposal of municipal and industrial wastes.
Concern for groundwater quality is further evidenced by the
increased allotment of federal research funds for studying the
movement and fate of hazardous chemicals in the aquatic ar.d soil
environments. One of the focal points of recently orocos'•:
federal regulations governing on-land disposal of municipal and
industrial wastes, the December 18, 1978 Federal Register, is
the protection of groundwater. In this section of this report,
the results of these and related investigations which are per-
tinent to the present study are summarized.
Ted Palit co-authored drafts of parts of this section of this
report
-------
STUDIES OF MOVEMENT OF INDUSTRIAL WASTES TO GROUMDWATER
The variety of industrial chemicals that have been and may
be disposed of or spilled on land is enormous and includes toxic
heavy metals, mineral acids, volatile organic solvents, pesti-
cides and highly refractory compounds such as PCBs. Members of
all of these classes are potential threats to the quality of the
nation's groundwater supplies. There h*ve been several labora-
tory studies conducted to evaluate the potential for migration of
industrial wastes through soil systems. For example, Houle e_t al^
(1976) studied the migration potential of industrial, wastes con-
taining cadmium (from nickel-cadmium batteries, electroplating,
water-based pairt and irorganic pigment wastes) in laboratory
test columns. Ihe migration potential of cadmium was found to
be largely dependent on *he physical and chemical composition of
the soil upon which the warte was placed. However, differences
in waste composition and leachability caused large differences
in migration of specific elements or compounds through the soil.
Houle et al. (1976) also found that specific metals leach dif-
ferently Trom waste mixtures, depending on the type of solid or
solution complex in wnich the metal exists. The composition of
the leachate and its solubilizing and exchanging characteristics
determined how a specific iretal migrated through a given soil.
Streng (1976) attempted to assess the impact of co-disposal
of industrial hazardous waste materials containing Hg, Cr, Cu,
Zn, Pb, etj. with municipal solid waste. In general, there was
a reduction in the moisture adsorption capacity resulting from
the combining of th« municipal solid waste with semi-solid indus-
trial wastes containing high c-ncentrations of heavy metal con-
taminants. Metallic ions became more soluble due to acid and
reducing conditions. Mercury was converted either chemically or
microbially, to a form acre amenable to transport. Municipal
solid waste and refinery sludge mixtures underwent rapid decom-
position which reduced the quantity of organic matter in the
leachate.
The transport of the chelatirg agent NTA to groundwater un-
der various conditions was studied by Dunlap e_t al. (1971). They
concluded that when NTA moves through unsaturateH~"soil, rapiu
and complete degradation to inorganic nitrogen compounds and
carbonate results. However, if NTA i * transported through satu-
rated soil, degradation is quite limited and a major portion of
the compound reaches th* groundwater intact. Under these cir-
cumstances, higher concentrations of complexed metals soch as
iron, zinc, chromium, lead, cadmium, and mercury, would b^ like-
ly to b«i transported to the groundwater.
Walker (1973) investigated several hazardous material dis-
posal sites for their potential for transmitting ccntamih&nts to
gro>indwdter. He found that contaminants from disposal sites may
10
-------
move through aquifers in a bulb-like mass at rates of a few feet
per year. Therefore, a fairly small quantity of a toxic substance
may enter a well as a hazardous slug several years after it has
been deposited on land.
Pettyjohn (197U) examined groundwater pollution problems
from the disposal of neutralized pickling liquors (wastes) into
an abandoned strip mine in eastern Ohio. He monitored all of
the surrour.ding and nearby waterbodies and found that water qual-
ity was adversely affected by polluted groundwater drainage from
the'disposal area. The samples collected from disposal pit seep-
age were characterized by low pH and high concentrations of dis-
solved solids, hardness, sulfate, chlorv.de, nitrate, iron,
fluoride, aluminum, chromium, nickel and .zinc. In addition to
contamination from the disposed industrial waste, acid mine
drainage from surrounding areas was found to degrade both sur-
face and groundwater quality. It was concluded thit the prob-
lem of acid mine drainage would not have occurred in the aban-
doned strip mine had the mine not been used for waste disposal.
Two landfills, one 27 years old and the other ul years old,
in the upper glacial aquifer on Long Island, New York were
studied by Kimmel and Beirds (197U). The plumes of leacheate
fron these landfills were found to sink to the bottom of the
aquifer and extend out about 3,200 and 1,5CO m, respectively.
The pH in most of the leacheate-contaminated groundwater ranged
from about 7 near the landfills to about 5 near the perimeter
of the plume. Heavy metals other than iron, manganese, and zinc
were not found in concentrations greater than several tens of
micrograms per liter in the plume. The authors concluded that
lining the landfill and collecting the leacheate to eliminate
a potentially dangerous source of contamination, may be worth
the added cost.
Groundwater pollution problems caused by a leaking gasoline
pipeline near the City of Los Angeles were the subject of a paper
by Williams and Wilder (1971). It was estimated that 250,000
gallons (9SO m3) of gasoline had seeped into the groundwater,
polluting the groundwater supply of the area. Skiiuning opera-
tions to drain off gasoline resulted in recovery of 50,000 gal-
lons (190 m3). The area has bean closely monitore-1 since these
remedial measures were taken. Most of the wells in the area
now have only a taste or odor of gasoline with two of the wells
showing detectable levels of free gasoline. Zt is hoped that
bacteria will eventually break down the residual chemicals oo
that the aquifer may be restored to service.
In this connection, Golwer e_t al. (1971) investigated the
process of self-purification of groundwater near disposal sites.
The results of the study indicated that many pollutants are, j.n
11
-------
fact, eliminated by geocheraical and biochemical processes in
groundwater. Organic compounds such as gasoline residue may be
decomposed by bacteria.
Osgood (197U) reviewed the literature on over 200 on-land
spills of petroleum and petroleum products and concluded that
regardless of volume discharged, all were potentially dangerous.
The critical factor in determining the fate of the material, ac-
cording to him, was the hydrogeologic character of the area, par-
ticularly the airection of groundwater flow.
A specific incident involving spillage of a petroleum pro-
duct was reported and studied in detail by Delfino and Dube
(1.176). They monitored nine domestic and industrial wells in the
vicinity of an accidental phenol spill and found that groundwater
in the =»rea (up to UOO m from the affected site) contained
neat;urablc concentrations of the compound during the 19-month
SXUQV period During this time there was also an increase in the
incidence of gastrointestinal disorders and mouth sores of the
peoplt ">"*n% thi.3 water for drinking. The temporal and spatial
clustering of these sympton.s strongly suggested that an epidemic-
logical episode related to the phenol spill had occurred.
Lewallen (1971) investigated a farm well which was polluted
with pesticides derived from contaminated soil placed near the
well casing. Water and soil in and around the well were moni-
toreti for more than four years and it was found that both DDT and
toxaphene persisted over this period. Levels in the sediment
were much higher than those in the water, probably because of
the low aqueous solubility of these substances and their ten-
dency to sorb onto clays. Vertical soil sampling in the area
indicated that the pesticides had moved down ground water only
slightly.
Gibb C1976) evaluated the extent of groundwater pollution
from surficial toxic waste disposal sites in humid regions.
He found coring to be an effective tool for mapping the migra-
tion patterns of chemical pollutants. He also provided field
data to Aerify the effectiveness of various types of soil to
adsorb or retain different chemical pollutants.
•
EXPERIMENTAL DN-LAND DISPOSAL PRACTICES
Land disposal of industrial and hazardous wastes has been
commonly practiced all over the U.S. In most cases the poten-
tial impart of such operations on groundwatei" quality was not
investigated. It is only recently that the need for selectivity
in siting, and strict control of disposal practices has become
recognized. As discussed below, a variety of factors is
-------
involved in deciding where and with what precautions a particular
waste should be disposed.
Snyder e_t al. (1976) studied the land disposal of oil refin-
ery wastes and evaluated methods of applying ar.d treating these.
The investigation was conducted in a semi-arid area and supple-
mental fertilizer was applied. Adequate drainage control, soil
characteristics which allow good drainage, and microbial activity
were the major factors considered during site selection. Soil
samples were taken from established land plots and analyzed to
determine soil characteristics, nutrient levels, heavy metals,
pH, oil content, dehydrogenase activity, soil respiration,
aicrobial activity, and microbial composition and density. The
chemical analysis of the soil showed that the heavy metal content
did not change over this one-year study period. The pH values
also did not change appreciably. On fertilized plots, oil
degradation was found to be about 80 percent complete within
one year. On a study plot that had not been fertilized, the oil
had degraded only 55 percent.
Wiles and Lubowitz (1976) explored the possibility of dis-
posing some solid or semi-solid hazardous industrial wastes con-
taining arsenic, lead, mercury, selenium, beryllium, cadmium,
zinc and chromium by encapsulating them with a plastic jacket.
They first used a binder resin to agglomerate the waste and
then encapsulated it with a 0.25 inch (0.6U cm) thick jacket
of high density polyethylene and fused it with powdered poly-
ethylene, in situ. The test specimens were subjected to leach-
ing with simulated ocean water fo^' 120 days. The leaching data
indicated that the system had great!ability to prevent or at
least limit pollution by leaching to acceptable levels. The
specimens were also tested for other properties such as compres-
sive strength, freeze-thaw resistance, impact strengths, punc-
turability, etc. The overall performance of the specimens was
foun-i to be highly satisfactory. The investigators therefore
recommended that the polymeric encapsulation process be used for
disposal of hazardous wastes containing constituents such as
sodium mcta-arsenate and arsenic trisulfide, etc., which may not
be adequately manageable by other techniques.
The effect of different clay liners in attenuating selected
trace elements such as As, Cd, Cr, Cu, Hg, Ni, Pb, Se, V, and
Zn from landfill leachate was discussed by Fuller (1977).
They observed different migration rates of various trace ele-
ments through soils and found that organic constituen:.. in the
waste had an effect on Hg attenuation. The kinds of clay mater-
ials they recommended for liners were: agricultural limestone,
hydrous oxides of iron (ferrous sulfate mine waste), lime-sulfur
oxide (stack-gas waste), certain organic wastes and soil sealants
(natural clay material).
13
-------
Liskowitz et_ al. (1976) examined the efficacy of ten natural
and synthetic materials (bo*.tom ash, fly ash, vermiculite, il-
lite, Ottawa sand, activated carbon, kaolinite, natural zeolites,
activated alumina, and cullite) for removal of contaminants in
the leachate and liquid portion of three different industrial
sludges, by sorption. They found no single sorbent of those ex-
amined that could significantly reduce the concentration of all
the contaminants in a leachate to acceptable levels. They did
find, however, that depending upon the type of sludge disposed,
a combination of sorbents can be used to reduce the contaminants
in the leachate to acceptable levels.
SORPTION-DESORPTION AND ION EXCHANGE
Among the most important soil properties influencing the
ability of a soil to retard the movement of hazardous chemicals
are sorption-desorption and ion exchange. Considerable effort
has been expended to understand these complex processes and while
there appears to be much literature dealing with the uptake
of metals and polar organic molecules by clays, there has beer.
very little written on the sorptive behavior of volatile, weakly
polar organics. This review focuses on the interactions of or-
gan ics with clays.
Sanks et al. (1975) investigated cation exchange capacity
(CEO and carbonate content of different clay materials with re-
spect to their clay mineralogy. Their data showed that CEC of
clays increased both with increasing percent montmorillonite and
decreasing percent carbonate. Also, if two clays with the same
percent montmorillonite were evaluated for CEC, tne one with the
lower percent carbonate would exhibit the highest CEf!. They
concluded that the decreasing values for CEC resulting from in-
creasing percent carbonate may be due in part to the calcium
carbonate fraction exchanging Ca** with the clay mineral fraction.
McAfee (1959) studied the replacement of inorganic cations
on montmorillonite by base exchange with organic compounds. Re-
placement of sodium from the homoionic Na-bentonite and the re-
placement of divalent Ca** and Mg** from Ca-Mg bentonite were
studied for comparison. The results of the experiments showed
that sodium is replaced relatively easily by the large organic
cation on the Na-bentonite compared with Ca** and Mg replace-
ment on Ca-Mg bentonite.
Adsorption of anionic, cationic and nonionic surfactants
onto montmorillonitic soils and kaolinite was the subject of an
investigation by Law and Xunze (1966). They observed that
anionic surfactants were not appreciably adsorbed and that anion
exchange capacity of the kaolinite appeared to be equal to or some-
what greater than that of mcntmorillonite. Cationic surfactants
-------
were strongly adsorbed to all the clay samples in amounts either
equal to or greater than the cation exchemge capacity of the
soils. The excess, it was speculated, was held by Van der Waals
forces betwean the alkyl radicals. Mpntmorillonitic clays ad-
sorbed considerably greater amounts of nonionic surfactants than
kaolinite. The amount of adsorption by the clays was highly
dependent on the surfactant concentration.
A radiochemical-tracer technique was used by Wayman ejt al.
(1963) to study the adsorption of ABS (alkylbenzenesulfonateT"
on montmorillonite. They found that the adsorption of ABS on
montmorillonite depended on the length of the alkyl chain and on
pH. Montmorillonite clay adsorbed more pcntadecyl ABS per gram
than dodecyl ABS and more ABS from acid or neutral solutions than
fron alkaline solutions. For ABS adsorption, they found the
optimum condition to be pH k.
Aly and Faust (196U) examined the uptake of the herbicide
2,"*-D and its ester derivatives on kaolinite, Wyoming bentonite
(a montmorillonite clay), and Fithian illite. Results were
analyzed using the Freundlich equation and it was found that
sorption of these hydrogen bonding compounds from aqueous solu-
tion was insignificant (0.02 to O.iu mg/g).
Kuang (1971) looked at a variety of factors of possible sig-
nificance in the sorption behavior of pesticides in aquatic
systems. In the pH range of 6 to 10, decreasing the pH had the
effect of slightly increasing the sorption of dieldrin on mont-
morillo.'iite, whereas varying the temperature (between 10° and
30°C) caused no appreciable change in this system. Likewise,
addition of salt, various organic pollutants such as glucose and
alanine, and soluble organic matter from domestic wastewater had
little or no effect OP dieldrin uptahe by montmorillonite.
Huggenberger fit al. (1973) discussed adsorption and mobility
of pesticides in soils and noted that pesticide molecules in
general *re strongly adsorbed by soils having high organic con-
tent and/or high clay content, but only weakly adsorbed by sandy
soils. The latter observation is consistent with an earlier
study by Boucher and Lee (1972) who fiund that lindane was sorbed
to only a small extent by natural aqvifer sand. Boucher and Lee
also found that dieldrin was taken u^ in greater quantity than
lindane in this sand system. As a rule, pesticides with a low
water solubility ere sorbed to a greater degree than those with
higher aqueous solubility.
Hoffman and Brindley (1960) investigated the sorption of non-
ionic aliphatic molecules from aqueous solutions onto montmoril-
lonite. It was concluded that a chain length of from five to six
units is necessary for these molecules to show any appreciable
15
-------
sorption. For larger molecules (up to 10 units) the effect of in-
creased chain length is to increase sorption. Also important is
the electron withdrawing capacity of the functional group at-
tached to the hydrocarbon chain. Molecules with stronger elec-
tron withdrawing groups, it was found, tended to be sorbed to a
greater extent.
As part of their effort to determine the suitability of clay
beds for storing industrial wastes, Sanks et al. (1975) studied
the sorption of acetone, acetaldehyde and phenol onto five Texas
clays. Sorption at an equilibrium concentration of 0.1 moles/
liter, ranged from 1 to 90 mmoles/kg of clay, with acetaldehyde
showing greater uptake than either phenol or aceto: e.
The only reported attempt to measure so ption of highly
volatile orgenics from dilute aqueous solutionr was discussed by
Sanks e_t al. (1975) whc remarked on the extreme experimental
difficulties encountered in working wilh benzene. Because of
errors associated with the transfer, centrifugation and analysis
cf aqueous benzene solutions, no data were published.
Lee and Jones (1978) have recently reviewed the state-of-the-
art in assessing the significance of sorption-desorption for pre-
dicting the environmental chemistry-fate of organic contaminants.
While their review wa.s primarily directed toward dilute, aqueous
slurry systems such as might be associated with suspended sedi-
ments in surface waters, it contains some information that may
be useful in assessing the significance of sorption in ground-
water systems. The following section has been extracted from
their review.
Considerable progress has been made in the past few years
in developing techniques for assessing the extent that sorption
reactions rafty occur for contaminants in the aquatic envii'on-
ments. From relatively simple laboratory batch-type sorption
experiments, it is now possible to derive information that can
be used to predict, in a general way, whether sorption-desorption
processes will likely be of major signific«ince in influencing
the environmental chemistry-fate of a chemical contaminant in
aquatic environments.
To a large extent, environmentally relevant sorption research
has been restricted to pesticide-soil systems (Karickhoff - US EPA
Athens, GA, personal communication). According to Karzckhoff, for
these types of systems, the extent of sorption has been related to
pesticide properties (such as pKa, pKj,, water solubility, and
polarity) and soil properties (such as particle size distribu-
tion, organic matter content, and cation exchange capacity (CEO).
16
-------
The Freundlich equation,
x/m * KCe1/n (1)
is frequently used to describe the adsorption equilibrium, where
x/m is the mass of contaminant sorbed per unit mass of adsorbant,
Ce is the equilibrium solution concentration, and K and n are
constants. Sone investigators choose to use this for environ-
nental chemistry-fate modeling. Deviation of the value of r. from
1, a common observation (AIBS, 1974), reflects the non-linearity
of the process. If n were 1, then K would equal the partition
coefficient, Kp defined in Equation 2. According to Karickhoff
(personal communication), sorption cf many organic compounds on
sediment and suspended particulates in dilute aqueous systems is
approximately linear and can be described in terms of the linear
partition coefficient. This has been described by Baughman and
Lassiter (1978) by the equation:
Kp * EP]A/[P] (2)
where Kp is the partition coefficient, and CP]\ and [P] are
equilibrium concentrations of the contaminant on the sorbent and
water, respectively.
One of the most important factors affecting the partitioning
cf rn*ny organic contaminants between sediment and water is the
organic car-son content of the sediment. It has been found that it
is frequently useful to normalize sorption partition coefficients
by the organic carbon content of the sediment or soil. An ad-
sorption constant (KoC) is defined as the amount cf chemical ad-
sorbed per unit of organic carbon (in ppm) divided by the concen-
tration of the chemical dissolved in the water (in ppm). While
this normalization has been found to be applicable to a large
number of organic compounds especially non-ionics, it is not ap-
plicable to some of the other groups of organic compounds. Fur-
ther work is needed to define the major factors controlling the
sorption-desorption of these compounds in natural water sedi-
ments.
During the past several years, the US EPA Athens, GA, labora-
tory has undertaken a substantial research effort devoted to en-
vironmental chemistry-fate modeling. A significant part of this
effort has been devoted to sorption-desorption of contaminants
on natural water particulate matter and soils. Sor.e of this work
has been published (e.g.,Baughman and Lassiter, 1978). Substan-
tial parts of the results of tnis effort are in preparation as
reports and papers or are in press at this time. The reader is
referred to publications of this group as well as the review
paper by Hamaker (1972) for further information on the state-of-
the-art of modeling sorption-desorption processes in natural water
17
-------
systems and soils. The work conducted by this laboratory';, staff
and their contractors (e.g., Baughraan and Lassiter, 1978; Karickhoff
e_t al. , 1979) as well as others has resulted in the following
general conclusions:
1. The sorption of contaminants occurs to ftbout the same
extent on sediments and soils.
2 When normalized by sediment organic carbon content, -he
sorption of neutral organic compounds on sediments from
different areas varies by a factor of 2 to 3 (Baughnan
*.nd Lassiter, 1978). This is an extremely important
conclusion which, if substantiated based on additional
studies, greatly simplifies the prediction of the en-
vironmental behavior of many organics that tend to sorb,
since the results of sorrcion tests conducted on one
or a few sediments wouJj in general be applicable to
natural water system? located throughout the world.
3. While much of the sorption on natural water particulate
matter is rapid (essentially complete within a few hours)
and reversible, there are some compounds which apparent-
ly do not desorb or do so very slowly.
u. In general sorption on natural water particulate matter
has a low temperature coefficient (activation energy).
5. Ond|jof the most promising recent developments is the
correlation between the distribution coefficient for the
uptake 'of dilute neutral organics by natural water
soJ.ids and the octanol/water partition coefficient. A
relationship of this type is to be expected for a large
group of organic chemicals, especially the nrn-ionic
type. Results of studies at the US EPA-Athens, GA
laboratory provide the information needed to estimate
sorption distribution coefficients from octanol/water
partition coefficients (Karickhoff e_t a_l. , 1979). Also
Kenaga and Goring (1978) have recently completed a re-
view of this topic.
PERMEABILITY OF SOILS
Permeability is the ability of a material to transmit water
or other fluids, A material Is said to be permeable if it con-
tains continuous voids. Since such voids are contained in all
soils including the stiffeat clay, and all non-metallic construc-
tion materials including sound granite and neat cement, all of
these materials show some degree of permeability. Furthermore,
the flow of water through all void-containing solids obeys the
same general laws.
18
-------
Soil permeability is one of the most important soil
characteristics in the study of subsurface pollution problems
since migration of the pollutant itself is likely to be dependent
on the permeability cf soil. In the design of a confinement
area for hazardous chemical disposal, permeability characteris-
tics of the liner and strata in the disposal area should be
determined ir. order to prevent pollution of groundwater by chemi-
cal contaminants. The permeability tests should be run with in-
dividual potential contaminants or mixtures of potential con-
taminant 3 rather than with water only. Presented below 5s a
summary of the current general knowledge related to permeability
of both soils and subsoil systems as it relates to the movement
of hazardous materials.
Coefficient of Permeability
Taylor (1948) discussed in detail the flow of water through
the pores of soil. The pores of most soils are sufficiently
small so that flow of water through them is laminar. In laminar
flow, the water flows along definite paths which never intersect
other flow paths. However, in coarse soils, the flow may some-
times be turbulent where the flow paths are irregular and twist-
ing, crossing and recrossing at random.
Darcy (1855) demonstrated that the rate of flow of water
through porous media is proportional to the hydraulic gradient.
According to Darcy's Law,
where
(3)
3 velocity of flow Cm/sec)
= hydraulic gradient (m/m)
v * Q/A
where Q - flow through the aquifer per unit time
A - cross-sectional area of the aquifer
(•4)
From Equations (3) and (U)
Q «
Q =
Ai
KAi
/. Q = KA (Ah/AL)
(5)
19
-------
where K = coefficient of permeability
Ah/AL = hydraulic gradient
The coefficient of permeability, K, has the units of velocity.
It can be interpreted as the velocity of flow under unit hydrau-
lic gradient.
Permeability, Hydraulic Conductivity and_Transmissibility
There is considerable variation in the definition of the
terms permeability, hydraulic conductivity, and transmissibility
by different authors and investigators. Walton (1970) and Black
(1965) defined hydraulic conductivity (K) in the same way as the
coefficient of permeability has been defined, i.e., the velocity
of flow under unit hydraulic gradient. They have, however, de-
fined permeability as a function of hydraulic conductivity, as
follows:
P = Kn/Pg
where K
P
n
P
g
(6)
hydraulic conductivity
permeability
fluid viscosity
fluid density
acceleration due to gravity
Thus, according to these authors, hydraulic conductivity has
the units of velocity while permeability has the units of length
squared.
Walton (1970) defined transmissivity as follows:
T s KD
where T
K
D
(7)
transmissivity
hydraulic conductivity
depth of aquifer
Transmissivity is thus defined as the rate of flow of water
Through a vertical strip of the aquifer of unit width, extending
through the total thickness of the aquifer under unit hydraulic
gradient.
From an overall point of view, the literature seems to in-
Lcate that authors and investigators with geoscience backgrounds
20
-------
prefer the term "hydraulic conductivity" and very seldom use
term "coefficient of permeability". The National Water Well
Association (JIWWA) usually uses the term hydraulic conductivity
in most of their publications. But, Hajilberg and Martinell (1976)
used the term permeability in one of the recent publications of
NWWA. Todd (1959) defines both of these terms, coefficient of
permeability and hydraulic conductivity, as being analogous.
Walton (1970) also refers to the terms as being analogous,, but
confused them in later applications with permeability. While in
general no strict principle has been followed in any area fcr the
use of these terms in this study, coefficient of permeability
and hydraulic conductivity are considered to be analogous terms;
but, permeability is ideally considered to be a property of the
porous medium alone.
Measurement of Coefficient of Permeability
The coefficient of permeability is one of the most important
characteristics of clay material to be considered in assessing
the suitability of the material for use as liners in hazardous
waste disposal sites. Coefficients of permeability are deter-
mined under the assumption that Darcy's Law is applicable. The
fundamental assumption in the application of Darcy's Law is
that flow in a porous medium is laminar. The Reynold's number
(N^) has been developed a.'< a criterion to distinguish between
laminar and turbulent flow, and is expressed as:
NR s pvD/n (8)
0 = density of fluid
v a velocity
D « diameter of the fluid patn
n s viscosity of the fluid
To adapt this criterion to flow in a porous medLom, the ap-
parent velocity defined by Darcy's Law is used for v and an
average grain diameter, d, is substituted for D in the equation
for Reynold's number. According to Todd (1959), when various
investigators have attempted to estimate the value of Reynold's
number to determine the condition of flow in groundwater, they
have found the value of Reynold's number to vary widely in the
range of 60 to 700, below which laminar flow could b« possible
and above which the flow condition could be termed turbulent.
This implies that the Reynold's number is not a good criterion
for judging the condition of laminar or turbulent flow in ground-
water. For almost all natural groundwater movement, however,
NR < 1 and Darcy's Law should be applicable. Moreover, materials
with low coefficients of permeability (silt, clay, etc,) have
21
-------
.'.'$ values much less than one. Burmister (195**) presented the
realm of validity of Darcy's Law and showed that in all materials
finer than silt, flow will always be laminar.
The coefficient of permeability can be determined in three
ways:
1. By laboratory tests on soil samples.
2. By field tests.
3. By use of empirical formulae.
Each of these approaches is discussed below.
Laboratory Determination--
The most convenient and moderately reliable way to determine
the coefficient of permeability is by laboratory tests on soil
samples. It is based on the T.easurement of the quantity of water
that flows under a given gradient through a sjil sample of known
length and cross-sectional araa in a given period of time.
Terzaghi and Peck (19u8) were the original .investigators of the
characterization of soil permeabilities and different methods
of laboratory analysis. Kezdi (197u) and Black (1965) also
presented detailed procedures to determine the coefficient of
perneability. M discussed by them, laboratory determination
of coefficients of permeability can be made by two methods.
1. Constant-head permeability tes'C.
2. Falling-Lead peimeability test.
Constant-head permeameters are particularly suitable for testing
highly pervious, coarse-grained soilrj. For soils with medium to
low permeability, the falling-head permeameters are used.
7h« principle of the constant-head permeam«ter is shewn with
sketches in Figure 1. Upward and downward flow type? can be
constructed. In both, water levels are kept constant by means
of overflows. The hydraulic gradient, i, can be coi^utc.d as
the difference between the constant water levels, h, divided
by th'» depth of the sample, 1. Operation of both the upward-
flow and downward-flow constant-head permnameters depends on the
ability to obtain a mnasurable quantity of le^chate. The upward-
flow type with piezometer -tubes is mere applicable for samples
with less pervious, fine sands and ailty sands. More reliable
results can be obtained by measuring the actual water pressure
in the sample by means of piezometer tubes. The discharge, Q,
during a given time, t, is collected in a graduated cylinder to
determine the volume discharged.
22
-------
upward flow type
n n n
downward flow type
Figure 1. Constant-head Permeametcr
1 - overflows; 2 - measuring cylinder; 3 - soli sample;
4 - screen and filter lay»r; 5 - base plate
(See text for explanation of symbols)
23
-------
For soils of low permeability, measurable flow through a
sample cannot easily be obtained by the constant-head test.
Moreover, test results wculd be unreliable due to evaporation.
In such cases, the falling-nead perr.eameter is more suitable.
Figure 2 displays in principle the operation of such a permea-
meter. A cylinder containing the soil sample is placed into a
base fitted with a fine gauze screen. On top, the cylinder is
plugged with a rubber stepper into which is inserted a graduated
standpipe or burette. During the test, che fluid level in the
pipe will continuously drop. The quantity of fluid that flowed
through the pipe need net be collected and measured in this test,
since it can be Calculated from the fall of the fluid level.
If, as shown in Figure 2,
hi = height of the water in the standpipe from trough
at initial time, t]_
h2 * height of the water in tre standpipe from trough
at the end of the period of observation, t-
h = difference of height of the water column
in th« intermediate time, t
- cross-sectional area of the standpipe
a
A
I
cross-sectional area of the sample
length of the soil sample
then, the velocity of flow in the standpipe = -(dh/dt)
(- is used to indicate that the water level de-
creases as time increases)
According to Darcy's Law
Q « A (h/i)
or Q = -« (dh/dt) = KA (h/t) where K is the coef-
ficient of permeability.
(9)
(10)
Separating the variables and integrating both sides, within the
limits,
-a
h2
/ (dh/h) =
K (A/1) / dT
(11)
Hence K = (ai/A) • (l/C t-
in
(h1/hj)
(12)
Sanks e_t aJ . (1975) attempted to conduct laboratory permea-
bility tests with Doth undisturbed and remolded samples. Their
experiments with undisturbed samples proved unsuccessful.
-------
Figure 2. Falling-head Permeameter
1 - stopper; 2 - cylindrical column; 3 • Sv>11 sample;
4 - screen ar.'d filter; 5 - overflow
(See text for explanation of symbols)
25
-------
Further, they encountered lack of reproducibility of results in
their studies with remolded samples. They concluded that labora-
tory permeability tests on cores collected in the field, even
from closely spaced t»;st borings in the same horizon, can pro-
duce results which vary by factors of five or more.
Field Measurement. "K" values obtained from laboratory
tests often differ considerably from the. true coefficients of
permeability of the soil obtained by field measurements. There
are many reasons for this discrepancy, the most important being
thar the permeability of natural soil masses is greatly in-
fluenced by the heterogeneity and delicate, or even undetectable,
stratification of the soil and by the uneven distribution of the
fine-grained fractions within it. Such effects can rarely be
simulatp.d in normal laboratory testing.
A simple approach to determination of field coefficients of
permeability has been described by Black (1965). This method,
the Auger-Hole method, involves construction of a cavity with a
minimum of soil disturbance. Flow of water into the soil or
out of the soil into the cavity is then measured, depending on
whether the position of the cavity is above or below the water
table. Coefficient of permeability can be determined by direct
substitution into Darcy's equation.
A more sophisticated method for the field determination of
coefficients of permeability is the pumping test. To conduct a
pumping test it is necessary that the well yield be in a steady
state. Figure 3 shows a section through a well and surrounding
aquifer. In order to determine permeability of th'j Aquifer in
the region of the well, the following assumption are made:
1. The water table is of infinite extent in the horizontal
direction.
2. The water hearing stratum is homogeneous, horizontal
and is of constant thickness.
3. The test well is a perfect well. This means that it
extends to the bottom of the permeable stratum and is
perforated over the section which is below the water
table.
u. In its original state, the groundwater is at rest and
there is no flow into or out of the system during the
test.
Pumping generates a radial flow of water forward to the
center of the well and as a result the water table assumes a
curved surface called the draw-down surface. If the pumping is
continued for a sufficient time, a steady-state flow is even-
tually attained.
26
-------
Figure 3. Effect of pumping on-line groundwater level
1 - To pump; 2 - original groundwater table; 3 - lower groundwater
surface; 4 - permeable layer; 5 - Impervious layer
(See text for explanation of symbols)
-------
For computation of the coefficient of permeability, Darcy's
Law is applicable at the steady-state condition. The rate of
radial flow through a cylindrical surface of radius r is propor-
tional to the cross-sectional area of the aquifer multiplied by
tr.e hydraulic gradient.
q « Ay i (13)
where q = flow through the well
V = cross-sectional ar
i - hydraulic gradient
A = cross-sectional area of the aquifer
Therefore q = KA i = KAy (dz/dx) = K 2IIxz (dz/dx) (lu)
where K = coefficient of permeability
(z and x are shown in Figure 3)
Separating the variables and integrating both sides,
/ZdZ = /(qdx/2KJIx)
Z2 =
-------
K = (q/n) (£n(x1/x2)} I'CL-Z\) (18)
Naney e_t al. (1976) studied hydraulic conductivity of
aquifers and attempted to formulate predictive models for the
evaluation of groundwater paths. They used two approaches in
the study. One was the parallel steamtubes concept and the
other was the converging steamtube concept with flow converging
toward the center. The latter provided a hydraulic conductivity
distribution more nearly like that expected.
Although the pumping test is the best approach by which to
determine the value of the true coefficient of permeability in
the field, it cannot be used to determine "K" values for materials
such as silt and clay. First, with clcy of low permeability, it
is not possible to obtain a steady state condition of pumping.
Second, the above list of assumptions is not applicable for any
type of clay material. Therefore, for fine-grained soils with
very low coefficients of permeability, there are very few re-
liable field methods for the determination of "K" values.
Use of Empirical Formulae. The determination of coefficients
of permeability by the use of empirical formulae is tased on
Poiseuille's Law which expresses the relationship governing the
flow of water through round capillary tubes. The law says that
the average velocity through a capillary tube is proportional to
the square of the diameter of the tube. Therefore, it is rea-
sonable to expect that the seepage velocity through a given soil
and the coefficient of permeability of the soil are proportional
to the square of the average pore dimensions. Since grain size
has the greatest effect on the size of the pore, it may be con-
cluded that the coefficient of permeability is proportional to
the square of'the grain size.
Hazen (1911) found that the permeability of filter sands
could be roughly expressed by
K = 100 (D1C)2 (19)
where D^« is the grain size diameter, such that the aggregate
weight or all smaller size grains id 10 percent of the total
weight of the sample. This is also called the effective size or
Hazen's Effective Size. Hazen's observations were made on sands
for which the effective sizes were between 0.1 and 3 nut diameter
and for which the uniformity coefficient (the ratio of DC|1 to
D10) did not exceed 5. DU
29
-------
Porosity and void ratios are also measures of pore spaces
in a soil sample. They are defined as;
Porosity = e/(l+e) (20)
where e = void ratio
e = Vy/
-------
conductivity). The modified equations adequately predicted the
experimentally measured values and provided satisfactory con-
ductivity data for many applications.
Johnson (1963) used laboratory permeability and geological
data to predict the transmissibility of water over large areas
where obtaining actual data from well testing was not economically
feasible. Based on the relationship of particle size to permea-
bility, he prepared graphs and tables for predicting permeability
where data were not available.
Zanker (1972) developed a nomograph for the determination of
hydraulic conductivity and intrinsic permeability in a labora-
tory water-soil system, based on conunon parameters such as length
of the soil sample, volume of leachate, the differential head,
the area of the sample, the duration of observation and the
temperature of the water.
Although th* above empirical formulae and relationships pro-
vide a means to estimate the coefficient of permeability, in
general they arc not applicable to clay-soils. The bound water
associated with clays is thought to be the primary factor re-
sponsible for the inapplicability of these relationships. Be-
cause of this bound water on clays, seepage occurs only through
a part of the core space.
The effect of s&rption on the movement of ions through ex-
change columns and on the mobility of pesticides in soils has
been reported in the. literature. Hashimoto e_t al. (196U), for
example, defined a retardation factor, R, as:
R = 1 * P K N CN"1/8 (24)
where P = bulk density of the column,
KSN = constants in the Freundlich equation
(S - KC ), (S - amount sorbed per unit
mass of solid)
C = equilibrium concentration, and
9 s column water content.
Since R appears in the denominator of a general transport equa-
tion (Kishimoto et a_l. , 1961*), the larger its value, the more
slowly the solute moves.
That a given organic solvent moves through soil at a rate
which depends upon its distribution between an aqueous phase and
the soil itself has been demonstrated by Briggs (1973), who re-
ported a linear relationship between log (1/Rp-l) and log Q, where
RF is the retention factor for an unionized pesticide (obtained
from soil thin layer chromatography), and Q is the soil organic
matter-rwater partition coefficient. Briggs has taken the
31
-------
analysis a step further and developed an equation relating Rf to
the octanol/water partition coefficient, P (see subsequent dis-
cussion).
Changes in Permeability With Time
Several investigators have reported on changes in the mea-
sured coefficient of permeability as a function of time. Bodman
(1937), for example, observed in laboratory column ter.ts on
California soils that the permeability "constant" decreased with
tine, precipitously for the first 20 days and gradually there-
after. Rates of decrease were found to correlate closely with
silt content, whereas final permeabilities were related to the
percentage of clay-size material in the sample. Based on a study
of the conductivity of the percolate, Bodman concluded that the
permeability decrease could be attributed to leaching of electro-
lytes from the soil, and consequent arrangement of particles
into conducting pores.
In a later, more detailed study of this phenomenon,
Poulovassilis (1972) concluded that changes in pore volume and
geometry, while important, were responsible for only 15 percent
of the decrease in permeability. Of greater significance was
the evolution and entrapment of gases of microbial origin which
had the effect of clogging pores and thus of reducing permeabili-
ty by as much as 83 percent of the original value. Earlier,
Allison (191*") had recognized that microorganisms were respon-
sible for decreased permeability in submerged soils and attri- j:l I
buted this to the natural clogging of soil pores by cells, slime ,
and products of microbial growth.
ATTERBERG LIMITS
Physical properties of clay are critically dependent on
moisture content; a clay may exhibit 3iquid or plastic behavior
depending on how much water it contains. The water contents which
characterize the onset of liquid and plastic behavior and which
characterize soil or clay saturation are referred to as the Atter-
b*rg limits. The significance of these limits was noted by
Terzaghi (1926) wf.en he wrote, "The results of the simplified
soil tests (Atterberg limits) depend precisely on the same physical
factors which determine the resistance and the permeability of
soils (shape of particles, effective size, uniformity) only in
s. fcir more complex manner."
The liquid limit (W^) is that water content above which the
particles of a soil are no longer held together by attractive,
cohesive forces, that is at this limit or water content, the soil
behaves like a viscous fluid or paste. Yong and Warentin (1975)
defined the plastic limit (W ) as, "...that water content below
which the soil is no longer plastic when it is worked, and
32
-------
crumbles on application of pressure". The plastic limit can be
interpreted as the lowest water content where cohesion of parti-
cles is low enough to allow movement but high enough to permit
particles to maintain their positions after remolding. The shrink-
age limit (Ws) is the smallest amount of Water that can completely
saturate a soil.
Kezdi (197U) indicated that typical values of W4 ranged from
UO to 150 and values of W from 25 to 50 for clays. The range
of water content between the liquid and plastic limits is called
the plasticity index, I , and for clays this extends between 10
and 100 (Kezdi, 197W). pln general, the greater the plasticity
index, the greater the swelling potential of a clay (Mitchell,
1976). A graphic summary of Atterberg limits is given in Figure U,
CLAY-ORGANIC STUDIES: X-RAY INVESTIGATION OF SWELLING
It is commonly known that expandable layer silicatffs are
capable of forming interlayer complexes with water and with or-
ganic molecules and that the characteristic spacing between the
layers is dependent upon the nature of the intercalated molecule.
Greene-Kelly (1955) has observed, for example, that the d(001)
spacing jn raontmorillcnite can vary from 9.5 A when unexpended,
to 23.3 A in the presence of pyridine. In his orientation
studies with aromatic compounds, Greene-Kelly (1955) concluded
that molecules could orient themselves either parallel or per-
pendicular to the silicate sheet and that the preferred orienta-
tion depended upon the external concentration.
Olejnik e_t al. (197U), who investigated the swelling of
.tiontmorillonite Tn polar organic liquids, obtained much larger
interlayer spacings, particularly for* substances which have high
dipole moments and high dielectric constants such as formamide.
In general, the degree of swelling observed .'or a variety of
homoionic, alkali metal tnontmorillonites, was not in keeping with
expectations based on Norrish's (Norrish, 195U) swelling index.
According to Norrish (195H), swelling should be related to
UE/v?, where U is the solvation energy (i.e., the amount of hea':
given off when water or an organic molecule interacts with an
interlayer cation), C is the dielectric constant of the inter-
layer liquid, And v is the valence of the interlayer cation. The
failure of this relationship has led Olejnik e_t aJL (197U) to
conclude that the measured bulk dielectric constant of a liquid,
and the actual dielectric constant of that sar.e liquid in the
interlayer region of a clay, might bt radically different. Still,
as Barshad has observed (Barshad, 19S2>, tns dipole moment and
the dielectric constant can be effective predictors of swelling
behavior. This point will be returned \o in the discussion of
swelling presented in this report.
33
-------
u*
Liquid state
Plastic state
Seal-solid
state
Solid state
Liquid Unit («,}
Plastic Halt (w )
P
Shrinkage 1i«it (MS)
Plasticity Index,
V "1"W
Shrinkage Inde.< ,
V "s
Figure 4. Atterberg Linits and Indices
-------
An extensive review of clay-organic interactions can be
found in Theng's book Ths Chemistry of Clay- Organic Reactions
(Theng, 197U). The nature cf clay-organic complexes, including
the effect of organic molecules on interlayer spacings, is dis-
cussed for alcohols, polyhydroalcohols, ketones, aldehydes,
ethers, nitritas, amines, aliphatic and aromatic hydrocarbons,
organic pesticides and positively charged organic species. Of
interest here is Theng's observation that even in dehydrated
clays, intercalation of non-polar organic molecules either does
not occur at all, or proceeds only with great difficulty. This
is reasonable since substances such as benzene would be only
weakly adsorbed and would therefore be incapable of expanding
the silicate layers.
In this connection, Barshad (195?^ has observed, using X-ray
diffraction, that benzene and n-hexane intercalated with some
montmoril"onite samples after these had been dehydrated at 293°K,
but did not intercalate when the samples had been dried at S23°K.
McEwan (19U8) observed no intercalation with n-hexane and
n-heptane, even after the air-dried samples of mcntmorillonite
had been boiled in the hydrocarbon liquids. Benzene, however,
did form a double-layer complex after this treatment. This is
in contrast to the results of Greene-Kelly (1955) who used a
similar technique (where montmorillonite was dried at 353°K)
but failed to achieve intercalation. The record is somewhat
miy:d 3:1 efforts to form clay-organic complexes with molecules
having l.-jw polarity.
It should be pointed out that while a considerable litera-
ture exists on interlayer swelling of clays in organic liquids,
no one has reported on the shrinkage or swelling of bulk quanti-
ties of clay in contact with organic solvents, using the ccnsoli-
dometer method employed in the current study. For the most part
consolidometer studies have been confined to aqueous clay sys-
tems in connection with ground swell potential, failure of
structures, highway maintenance, etc. To the knowledge of the
authors, this investigation is the. first systematic study of
mechanical shrink-swell behavior using organic liquids.
OCTANOL/WATER PARTITION COEFFICIENT
Early studies on the partitioning of organic substances be-
tveen wtter and an apolar phase wer« conducted independently by
Mayer and Overton who were interested in the effect of organic
compounds in producing anesthesia and narcosis (Kansch, 197U).
They were able to show that the potency of a narcotic is direct-
ly related to the extent to which the compound partitions itself
between olive oil and water, that is, to its olive oil/water
partition coefficient. Mayer and Overton reasoned that s\nce
35
-------
nerves are •unrounded by fatty membranes, the more fat soluble
compounds would partition into these and disrupt nerve operation
more effectively than drugs with an affinity for the aqueous
phase.
Hansch and Elkins (1971) found a more satisfactory relation-
ship to exist between medicinal properties and the octanol/water
partition coefficient, P, defined as
p _ concentraticn of solute in n-octanol (25)
concentration of solute in water
Hansch e_t al. (1968) also reported that an excellent linear
correlation exists between the logarithm of *he aqueous solubili-
ty of organic liquids and log P as given in the equation:
log (1/S) - a log P * b (26)
where S is the molal solubility of the organic liquid in water;
P is the partition coefficient of the liquid between 1-octanol
and water; and "a" and "b" are constants. In general, the lower
the solubility of a substance in water, the greater its parti-
tion coefficient, P.
Chiou «t al. (1977) reported an empirical correlation be-
tween the aqueous solubilities of a variety of potential environ-
mental contaminants (including aromatic hydrocarbons; organo-
phosphate «nd organochlorine pesticides; and polychlorinated
biphenyls (PCB's))and the n-octanol/water partition coefficients
of those contaminants. More importantly, they also showed that
a good correlation exists between the bioconcentration factor of
an organic pollutant in rainbow trout and the aqueous solubility
of the pollutant, according to the equation:
log (B.F.) • 3.HI • 0.508 log S (27)
where B.F. is the bioconcentration factor and S is the aqueous
solubility (y moles/1). Thus, it is now possible to uee widely
available solubility or partition coefficient data to predict the
uptake of certain organic pollutants by certain .iquatic organisms.
More pertinent to the present study is the finding by Brings
(1973) that there i* a relationship between the octanol/water
partition coefficient for unionized organic compounds and their
adsorption on and movement through soils. Using Rp (retention
36
-------
factor for unionized pesticide) as an index for the mobility of
organic chemicals through soils, Br:'.g&s has classified compounds
into five types, ranging from injnob:.le to very mobile, and has
shown that in general irnmobile compounds are those which have
high values for log P (>3.78), and very mobile compounds are
those with low values (Log P < O.C8). This suggests that the
rate of movement of an unionized organic substance through water
containing soils can be related tj the tendency of the substance
to escape from the interparticia aqueous phase and to become
sorbed onto the immobile soil particle itself. The greater this
escaping tendency as measured by Log P, th»» lower the mobility
of the organic substance. (See discussion of sorption ard
permeability).
37
-------
SECTION 5
METHODS
INTRODUCTION
This section of the report presents the experimental proce-
dures for studies on those properties of clays which might in-
fluence their ability to augment or alter the transmission of
fluids.
Three clay-soils and one soil sample were initially selected
for study, bat, for a variety of reasons, only the three clay-
soils (without deflocculent) were used extensively in tests with
organic solvents (ACS grade or equivalent). Samples of the three
clay-soils, Ranger Shale, Kosse Kaoline, and Fire Clay, were ob-
tained from a commercial clay supplier. The soil sample was col-
lected from Parker, Texas (Collin County). The Parker sample was
obtained using an auger, from a depth of about 8 reet (^2.5 m),
which was well below the third soil horizon.
CLAY MINERALOGY
Clay mineralogy for Ranger Shale, Kosse Kaoline, arvj Fire
Clay was determined at the Colondo Schoo?. of Mines Research
Institute. Oriented jr.ounts were prepared, scanned over 2 - 21°
of 29, glycolated and then rescanned over the same range to deter-
mine expandable clays. Parker soil was examined in the UTO labo-
ratory using glycolation as discussed in Pierce and Siegel (19F9)
and in Biscayne (1965).
PARTICLE SIZE ANALYSIS
Fifteen to 25 grams of sample clay were measured into a 250
ru Erlenmeyer flask containing 100 to 150 ml or" distilled water.
7h» flasks were stoppered with Parafilm, placed on a gyroshaker
and shaken at 250 rpm for 12 to 16 hours. The rotating motion of
the table gently disintegrated the sample without significantly
reducing the individual particle sizes. The sample was then
quartiratively transferred to a one liter glass cylinder. One
grcvm cf Calgon was added as a deflocculant. Some samples were
run without Calgon to compare results of flocculated ard de-
flocculated samples. After thorough stirring, the mixture was
38
-------
allowed to stand for 2U hours to be sure it showed no signs of
f jocculation. If rlccculation was observed, the amount of Calgon
was increased. If no signs of f laccula-cion wera observed the mix
ture was again thoroughly stirred; when stirring was ceased, tii.'-e
was recorded. The settling ve^oci.*y of a particle 2 p iii diam-
eter was determined according to Stokes' Law (Black, 196S). Time
required for particles < 2 u diameter to settle a distance of 10
en was then calculated." After settling for the calculated time,
a 20 ml sample was d^-awn witv a pipette from a dept* of 10 cm so
that it contained particles about 2 u in diameter. The subsample
was transferred into a beaker of *nown weight, evaporatac at 130°
C for 24 hours and reweighed.
The clay size fraction was determined from the following
relationship:
W/WQ = P/100
(28)
where w = weight of particles present in the volume at
time, t ;
w = weight of particles in the volume initially;
P = percentage of particles by weight, of size 2
micron and smaller.
The sand fraction was determined by pouring the rest of the
sample through a number **0 sieve, washing it with distilled
wa':er and collecting the .residual material in a preweighed
beaker. This material was! dried at 110°C for four hours and re-
wiighed. The weight of clay was taken vo be of si
-------
Each sample was prepared by air drying two to three pounds
of the clay sample to a slightly damp condition. Lumps were
broken and then homogenized so that the material passed through
a number UO sieve. The sample was then divided into 6 to.8 por-
tions, such that each portion contained slightly more than enough
material for one test. To each portion, the approximate re-
quired amount of water was added to obtain the desired range of
moisture content. Aftar thoroughly mixing, each portion was
placed in a small glass jar tightly fitted with Parafilm and
stored overnight or until ready for testing. For samples that
mixed readily with water and had low dry strength, it was satis-
factory to add water and mix the specimen immediately prior to
testing. A compacted specimen should rot be remixed and used
over again.
Prepared clay was put in the mold. The surface was leveled
using a wood plunger. The tamper was inserted in the mold until
it was in contact with the surface of the soil, and pressed down
until the spring began to compress. After the force was released,
the tamp was shifted to another position. Each of the four tamps
was applied in a separate quadrant, adjacent to the mold. The
fifth tamp was in the center, making one complete coverage. This
cycle was then repeated until the appropriate number of tamps had
been applied. Tan tamps with a 20 Ib.(9 kg) spring were needed
for the stand.ird low compaction test. The tamps were applied at
the approximate rate of 10 tamps per 15 seconds.
The next layer of sample was added and the procedure repeated
until the required number of compacted layers had been placed
in the mold. The top layer extended at least 1.3 cm into the ex-
tension collar. .
The mold assembly was then transferred to the collar re-
mover. The mold was removed from the base and the excess soil
from the top of the mold was trimmed away. The sample from the
mold was removed with the ejector and was placed in a preweighed
container and reweighed.
The sample wa,- then dried in an oven at 110°C for 24 to U8
hours. Drying was continued to constant weight. If the speci-
men was usad for a permeability test it was either done at
known moisture content or the moisture content was determined
from the excess material removed from the collar. Additional'
specimens were compacted until points had been established on
both sides of the optimum moisture content curve.
ATTERBERG LIMITS
Atterberg Limits were determined using methods for finding
liquid limits and plastic limits for soils by the American
Association of State Highway Officials designations T-89 and
UO
-------
T-90, respectively (Black, 1965). For the liquid limit test,
Standard Liquid Limit Device (CL-207) manufactured by Soiltest,
Inc., Evanston, Illinois, was used. For the plastic limit test,
Plastic Limit Set (CL-251) by Soiltest, Inc., Evanston, Illinois,
was used. The standard liquid limit device was made of hard
rubber to maintain the uniformity of hardness, size and density
of all devices. The brass cup was die-formed to specified di-
mentions. The cup and drop adjustment parts and aluminum hand
crank were mounted on an alur.inum housing.
Determination of liquid limit was based on the number of
drops of the standard tamp needed to close the gap made in the
clay with the standard grooving tool and the corresponding mois-
t-irt content of the clay. The number of blows needed at various
ranges of moisture content were determired and the values were
plotted as the log of the number of blows versus moisture content,
This relationship was found to be linear. The liquid limit was
determined from the graph, as the moisture content at 25 blows.
Determination of plastic limit was based on the moisture
content of the clay which crumbled at the surface as the clay
sample was rolled narrower than l/8'; (3.2 mm) diameter. It was
determined by the use of the standard plastic limit plate.
SPECIFIC GRAVITY
Specific gravity was obtained by a standard procedure using
a pycnometer. It was calculated from the mass volume of the
sample. The mass was determined by weighing and the volume by
calculation from the mass and density of water displaced by the
sample.
CATION EXCHANGE CAPACITY
Cation exchange capacity was determined from the procedure
of Busenberg and Clemency (1973) using an ammonia specific ion
electrode. The procedure, which requires saturation of the ex-
change sites with ammonium ion, utilizes the electrode in the de-
termination of ajpnonia released by treatment of the ammonium-
saturated clay with strong base. The procedure is as follows:
Ammonium-saturated clay samples were prepared by placing the
sample in an excess amount of a 1 N ammonium acetate solution and
shaking on a gyroshaker for 16 tc 20 hours. The solution was ad-
justed to pH 7 and was transferred to a 50 ml centrifuge tube.
It was centrifuged at 2U,000 g for 10 minutes. The supernatant
was then decanted. A 1 N ammonium chloride solution was added
and the mixture was shaken for 5 minutes. After recentrifuging,
the supernatant was decanted. This procedure was repeated four
times.
Excess ammonium salts were removed by washing with isopropyl
alcohol. Isopropyl alcohol was added, the mixture was shaken,
•41
-------
centrifuged and the supernatant decanted. This procedure was re-
peated four times. After the clay was air dried, it was ready
for CEC determination. One gram of air dried (treated) clay was
placed in a ISO ml beaker and 100 ml of deionized water were
added and the mixture was stirred well. Then 1 ml of a 10 M
sodium hydroxide solution was added. After about a minute, the
released ammonia was measured by the Orion Specific ion ammonia
electrode method.
Calculation of CEC
The CEC of the sample was determined by the following
equation:
CEC = (c)(v)/(w)(f) (29)
where CEC = Cation Exchange Capacity of the sample in
meq/100 g clay
c = concentration of ammonia in moles/liter
v = volume of water added in ml
w = weight of the sample in mg
f = conversion factor, in this case a constant
eciual to 10" .
Exchangeable Cations
Exchangeable cations were identified by treating 10 g of
distilled watar washed clay with 1 N NHU Cl overnight. The re-
sulting solution was analyzed for Ca and Mg by titration (APHA
et al., 1976) and for Na and K by flame photometry using a
C"o~rning Model U30 Flame Photometer.
TOTAL CARBON AND CARBONATE ANALYSIS
Total carbon content of the clay sample was determined in
the IR-12 Carbon Determinator (Leco Corporation, St. Joseph,
Michigan). For carbonate analysis, total carbon was determined
on acidified samples. About 5 grams of sample were acidified
with dilute HC1 (20%) in a porcelain crucible until no further
bubbling of CO? occurred. The operation was repeated several
times until bubbling completely stopped. The sample was then air
dried and analyzed for total carbon on the IR-12 carbon analyzer.
The difference between total carbon content of unacidified and
acidified samples was equal to the carbonate carbon content.
SWELL PROPERTIES
Swell properties were measured (using remolded clay) with a
6.<* cm diameter consolidometer manufactured by Karol-Warner, Inc.,
Highland Park, N.J. The dial indicator on this insturment was
capable of measuring displacements up to one ten thousandth of an
inch (2.5 x 10'^cm). Studies involving organic solvents were run
in a desiccator to prevent evaporation and, in the case of mix-
-------
tures, changes in solvent composition. Experiments were con-
ducted in a constant temperature room at 22 t 1°C. The detailed
procedure adapted from ASTM (1970) is given below.
A 500 to 600 g aliquot of well-ground (passed through a No.
8 sieve), air-dried sample was weighed and transferred to a mix-
ing pan. The amount of deionized water required to achieve the
optimum moisture content of the clay was added to the clay sam-
ple. After thorough mixing, the sample was placed in a glass
"jar tightly fitted with Farafilm and stored overnight or until
ready for testing. For samples that mixed readily v:ith water
and had low dry strength, it was satisfactory to add water and
mix the sample immediately prior to testing. It is important
that a compacted clay not be remixed and used over again. A
specially designed mold (7.6 cm long, 6.0 cm inner diameter) and
collar (7.6 cm long) were used to make the remolded clay sample.
With the nold and collar clamped to the base, about half of the
mixed clay was placed in the mold. T.c surface was leveled with
a wood plunger. The sampla wa., then compacted with a 5.5 Ib.
(2.5 kg) standard hammer to conform to the standard compaction
test procedures (12,375 ft. Ib. per cu. ft, (5.9 x 10s joules/™3)
of compacted volume). About 20 to 20 percent additional pressure
was applied depending on the amount of excess volume of compacted
clay. After removing the collar and shaping the sample with a
straight adge, the sample was ejected with a specially designed
plunger.
A 3/1* in. (1.9 cm) long sample was then cut and shaped
properly usinc a sharp knife and a straight edge. A 1.9 cm long
collar was used to create a fine and precise finish on the Sim-
ple. The molded sample was then transferred to a specially de-
signed consolidometer (3.8 cr long, 6.4 cm inner diameter) over
a 0.6 cm thick porous stone. Another piece of porous stone of
the same size was placed at the top of the sample. A load block
was then placed at the top. A dial gage indicator capable of
measuring within a range of 0.0001 to 0.5 inch (0.002 to 13 mm)
was mounted in proper position. To read any possible shrinkage
cf the sample the dial gage indicator was adjusted to read in-
itially at 0.01 to 0.02 in. (0.025 cm to 0.05 cm). If more
shrinkage was expected, the dial gage was initially adjusted
accordingly. Th» consolidometer was transferred to a desiccator
(without the desiccant). Fluid was then introduced up to a level
1.5 cm above the consolidometer, that is, samples were flooded
with the fluid. Wall friction in these experiments was assumed
to be negligible. Measurements of the swell or shrinkage of the
sample were made as a function of time.
COEFFICIENTS OF PERMEABILITY FOR LIQUIDS AND LIQUID MIXTURES
IN CONTACT WITH CLAYS
In this study the coefficients of permeability were deter-
mined on remolded clay using a laboratory apparatus with heavy-
duty permeameter columns (see Figure 5). These permeameters were
«
U3
-------
I",
High Pressure
Stainless Steel
Valve
99 cm
n
Cll
\
>r^
u
c?
s
/
To Nitrogen
. . . . . . Tan k
Liquid Level
8 mm Graduated Stanplpe
High Pressure Swagelock
fitting (Teflon)
Clamp
T.8 cm Thick X 10 cm ID
Pyex Glass
Teflon Gasket
High Pressure Joint
Whatman GF/A Fiberglass
Filter
40 Mesh Stainless Steel Screen
6 cm Pyrex Glass Support
NOT TO SCALE
To Collection
Figure 5. Pressure permeability column
-------
designed and built with Pyrex glass in the University of Texas-
Dallas machine shop especially for this study. All joints in
the columns were lined with Teflon to make them suitable for use
with organic solvents. The procedure followed for determining
the. coefficient of permeability, adapted from ASTM (1970), is
given below.
About 1000 g to 12CO g of well-ground (pass through a No. u
sieve) air-dried sample were weighed and transferrec to a mixing
pan. The required amount of deionized water corresponding to
the optimum moisture content, was added to the clay sample.
After mixing, The sample was placed in a glass jar, fitted with
Parafilm, and stored overnight. For samples that mixed readily
with water and had low dry strength, it was satisfactory to add
water and mix the specimen immediately prior to testing. It is
important that a compacted specimen not be remixed and used over
again. With the mold'and collar clamped to the base, about half
of the mixed clay was placed in the mold. The surface was level-
ed by pressing lightly with a wooden plunger. The sample was
then compacted with a 5.5 Ib. (2.5 kg) standard hammer to conform
to the standard compaction test procedures (12,375 ft. Ib. per
cu. ft. (5.9 x 10' joules/m^) of compacted volume). About 10 to
20 percent additional energy was applied depending on the amount
of excess volume of compacted clay. Actual moisture content of
the clay sample was determined separately, by oven drying a dupli-
cate sample at 110°C for 2U hours. The excess clay in the mold
was scraped and leveled carefully up to the 2 in. (5 cm) mark
derth. The molded clay wes ready for testing in the permeability
column, which was then assembled and test fluid introduced.
Final fluid level Li the graduated standpipe was adjusted very
precisely with ths help of a pipet bulb. Temperature measure-
ments of the fluid were taksn with a thermometer well immersed
in a reagent bottle filled with the fluid. The open end of the
standpipe was firmly covarcc with two layers of Parafilm and a
hypodermic needle was inserted through the Parafilm. Using this
technique, evaporation of so.'.vent was reduced to an insignificant
amount. If any clay-fluid system had a coefficient of permeabil-
ity less than 10~10 cm/sec, the open end of the standpipe could
be attached to a pressurized nitrogen tank and a constant pres-
sure in the range of 10 to 50 psi could be maintained. The co-
efficient of permeability 'K' in cm/sec was computed as:
K = QL/AH (30)
where Q is the flow of the percolate in ml per second, L is
the height of the sample in the column in cm, A is ths cross
sectional area of the sample in square cm, and H is the
average head of the fluid medium on the sample in cm.
The void ratio was calculated from the actual moisture content
of the compacted clay.
-------
X-RAY DIFFRACTION STUDIES OF CLAYS IN CONTACT WITH SOLVENTS
AND SOLVENT MIXTURES
Clays (< 2 y diameter) separated by settling according to
Stokes* Law were used to study the effect of organics on lattice
spacing. Approximately one gram of the separated clay was mea-
sured into a 250 ml Erlenm«syer flask containing 100 ml of treat-
ment solution.* The flask was then placed on a gyroshaker and
run for 12 to 18 hours at 250 rpm. The clay solution was sub-
sequently transferred to a polypropylene tube and centrifuged at
36,000 g for fifteen minutes, /.fter centrifuging, the super-
natant was decanted and the wet clay at the bottom of the tube
was smeared onto glass slides to form a thin, smooth coating.
The slides were then allowed to air dry for several minutes.
After drying, they were run in a Norelco X-ray diffractometer
and scanned at 2° 26 per minute from 2° to 30° 29 with copper
radiation. The 29 values of the peaks wera converted ~o clay
lattice spacing in A.
Initially, difficulties were encountered in obtaining X-ray
data for clays in contact with the neat organic liquids. Ranger
Shale treated with benzene, for example, formed a stiff mass
which did not adhere to glass. It was found, however, that add-
ing three drops of distilled, deionized water to the centrifuged
clay (as recommended by Hoffmann and Brindley, 1960) facilitated
adhesion. The clay was then dried in a desiccator over PjO, for
15 hours, and X-ray patterns obtained.
DETERMINATION OF MAJOR IONS
Three clay samples were sent to the University of Texas-
Dallas Applied Research Laboratory where they were examined for
major ions using the electron microscope. Intensities of char-
acteristic X-ray lines were compared with the intensities from
reference oxide standards for Si, Al, Ca, Fe and K<
DETERMINATION OF SORPTION ISOTHERMS FOR VOLATILE ORGANICS
Considerable time was spent in developing 'a method for ob-
taining reliable sorption isotherms, particularly for the highly
volatile hydrophobic compounds being used in this investigation.
A review of the literature revealed a few helpful references on
this subject, although Sanks et al. (1975) did point out some of
the difficulties which they encountered in their unsuccessful
efforts to plot Freundlich isotherms for benzene on several
clays. Although there are a number of papers which discuss the
sorption of organic compounds from dilute aquatic solutions, no
successful work, has been reported for highly volatile organic
solutes.
*Baker Analyzed Reagent Grade Solvents.
46
-------
Originally, it was proposed tc conduct sorption studies
with ?5 grams of clay mixed with 75 ml of solution in a 2SO ml
Erlenmeyer flask. The extent of sorption was to be determined
by measuring the difference between initial and equilibrium
organic carbon concentrations with the Beckman Total Carbon
Analyzer. However, in a series of experiments using benzene and
Rarger ShaJe, several very serious sources of error wer» encoun-
tered which necessitated extensive modifications of the technique.
Despite the fact that Ranger Shale had a low organic carbon
content, 0.28 percent, there was sufficient organic material
leached from the clay to interfere with analysis. The presence
of humic acids and other carbon-containing compounds in solution
gave apparent equilibrium concentrations which were erroneously
high. In some cases, where sorption was low, the equilibrium
concentration actually exceeded the concentration of the sti«.idard
solution. The problem was solved by usinf; gas chromatography
which separated the compound of interest '.'rom interfering sub-
stances.
A second, somewhat more subtle, problem in working with
volatile organics was loss of vapor to the air space above the
solution. Initially, it was thought that the quantity of benzene,
for example, lost to the air space would be insignificant when
compared to the amount sorbed by the clay. However, as the
following calculation shows, vaporization to the air space can
be appreciable:
Suppose that 50 ml of an aqueous benzene solution having a
concentration of 1.52 x 10 M is placed in 100 ml volumetric
flask capped with a ground glass stopper, and that the tempera-
ture is 20°C. The initial number of moles in the system is
76 x 10 moles. The molecules will distribute themselves ac-
cording to Henry's Law which can be written as:
P/c = k (31)
where P = partial pressure in atmospheres of benzene over
the solution,
c » concentration in moles/1, and
k = U.22 (Atm/mole) (at 20°C)
(Green and Frank, 1979)
If nt is "he number of mol*.j of benzens in the liquid and n is
the number of moles in the vapor phase, then the following *
equations must be satisfied:
-------
n£ + ng = 76 x 10"° (32)
and
P/C = (ngRT/Vg)/(n£/V£) = {(ng) ( 82 .02) ( 293.18)750} / (n£/0 . 05)
= U80..9 ng/(n£/0.05) = u.22 (33)
where V = volume occupied by the vapor in cm , and
O
V£ s volume of the solution in liters.
-6
'jolvinfi: for n£ gives 64.7 x 10 moles or a concentration of
1.29 x 10 moles/1 in the liquid phase. Under these conditions,
a concentration change of 0.23 mmoles/1 can be expected as a re-
sult of vaporizatior, alone. This is comparable to the amount
sorbed by the clays. Consequently, it is necessary to minimize
air space above the liquid.
While the escaping tendency for hydrophobic solutes is
higher than for acetone which hydrogen bonds with water, the
vapor pressure of sretone over aqueous solutions at low concen-
trations is also sufficiently high to warrant reduction of air
space above the liquid to a minimum. For example, at a concen-
tration of 2000 mg/1 (mole fraction = 6.2 x 10 ) the partial
pressure of acetone is 0.71U torr (Washburn, 1923).
A technique was developed for obtaining satisfactory iso-
therms for volatile solutes which involved taking a number of
precautions to prevent evaporative losses. Standard solutions
were prepared by using a variation of Marketos' method (Mar-
ketos, 1969) in which a small quantity of the solute (ranging
from 50 to 2000 mg) was weighed into a 5 ml capped vial. The vial
was immersed in a dry ice-acetone slurry (for acetone, liquid
N2 should be used) to freeze the contents. Once the liquid
was solidified, the cap was removed and the vial contents were
quickly transferred to a volumetric flask filled with one liter
of carbon free water and containing a stirring bar. Following
the practice of Bohun and Claussen (1951) in their studies of
hydrocarbon solubilities in water, the ground glass cap of the
flask was coated with silicone grease and securely taped. The
volumetric flask was then inverted a number of times to insure
mixing and the flask was placed on a magnetic stirrer for 2u
hours.
Chilled stock solutions were transferred in a cold room
to a re-pipette bottle and an accurately measured volume of the
solution was added to a flask containing a weighed amount of
clay. For both acetone and the hydrophobic organic compounds,
U8
-------
10 to 25 grams of clay per 100 ml of stock solution gave equi-
librium concentrations 5 co 50 percent lower than the ftock
solutions.
Flasks with the taped, silicone-coated stoppers which con-
rained the cljy, the solution, and several steel balls (to pre-
vent settling) were placed on a gyroshaker and swirled at 150
rpm for 2* to ^8 hours. The equilibration time used in this
work is consistent with the procedures, of Hoffman and Brindlev
(1960), Kuang (1-J71), and Aly and Faust (I960 and with obser-
vations in the current study that benzene sorption on Ranger
Shale is essentially complste within six hours (see Figure 6).
75
c
a
•a
O)
CL
< 3
O
451
35
75
Time (h-s)
Figure 6. Change In peak height, (concei.trst.1on) with time
for benzene sorbed o.ito Ranger Shale
An alternative procedure wh.icn also gave good results in
vr.lved the use of stoppered bottles with ground glass caps.
-------
Mixinp was accomplished by magnetic stirring in this case.
After equilibration, mixtures were transferred to polypropylene
centrifuge tubes, care being taken to avoid air space above the
mixture. Centrifugation ;.as carried out at 3°C for five minutes
at 1*3, COO g. At this temporsture, no evaporative loss of
benene or xylene was detected for solutions run in the absence
of clay. The time r«»^uired for essentially complete compaction
of the clay was about cr.e minute; during this period the temp-
erature of the solution fell 2 vo J°C. It is unlikely that a
significant shift in the equilibrium concentration, which is re-
ported for 2U - 3°C, took place under these conditions. The
clear supernatant was analyzed using a Varian gas chromatograph
with a flame ionization detector. Table 1 lists the conditions
for analysis.
TABLE 1. CONDITIONS FOR ORGANIC SOLVENT ANALYSIS
Compound
Acetone
Benzene
m-xylene
Carbon Tetrachloride
Packing
Chromasorb
Chromasorb
Chromasorb
Chromasorb
W
W
w
W
Gas Flow
Rate
(cc/min)
30
30
30
50
Oven
Temp
(°C)
110
110
110
140
Relay
Time
(Min)
.. 1.25
1.6
2.0
0.5
Throughout the course of this investigation, improvements
in the method which might allow for more expeditious collection
of data or for better analytical precision were constantly
sought. In the last months of the project, 50 ml stainless steel
centrifuge tubes equipped with 0-ring3 were used to equilibrate
and centrifuge the samples in the same container, thus eliminat-
ing one of the transfer steps.
A recommended procedure consistent with the above finings
can be summarized as follows:
Weigh out three to fifty grams of clay into eacv. of eight
50 ml stainless steel tubes. Place two 0.25 inch (0.6 cm)
diameter stainless steel balls into each of the tubes to insure
good contact between the clay and the organic solute. Add a
measured amount of solution sufficient to completely fill the
container; cap, and place on a gyroshaker at 225 rpm for a mini-
mum of 2u hours. Without transfer, centrifuge at room tempera-
ture for 20 minutes on a Beckman Model J-21B centrifuge (cooling
is not required with the stainless steel tubes). Then analyze
the supernatant for organic compounds of interest using a gas
chromatograph.
50
-------
SECTION 6
RESULTS AND DISCUSSION
INTRODUCTION
Presented in this section are the characteristics of the
clays used in this investigation; the results of the work on
sorption of organics from dilute aqueous solutions onto clays;
swelling of clays in water, organic solvents, and solvent mix-
tures; permeability of clays to water, organic solvents and sol-
vent mixtures; and the behavior of clay lattices in contact with
various liquids (X-ray diffraction studies). Also presented is
a discussion of the results obtained on the effects of solvents
and solvent mixtures on the characteristics of the clay and soil
samples examined.
CHARACTERISTICS OF CLAYS USED IN THIS INVESTIGATION
Particle Size Analysis
Particle size distribution is an important parameter in
characterizing the type of solid substrate. Size distributions
for materials used in this study are presented in Table 2. The
particle size distribution for Parker soil was obtained with
and without Calgon polyphosphate dispersant in order to deter-
mine the effects of this material on particle size results.
From the data in Table 2, it is evident that Calgon had a meas-
urable impact on the particle size distribution of the Parker
soil. As expected, the deflocculant caused an increased per-
centage of the clay particle size fraction.
Table 2 shows very high clay concent (particles of size
<2p without deflocculant) for all the samples except the Parker
djil. In terms of clay fractions, the materials used in this
study ranked in the following order: Kosse Kaoline > Fire Clay >
Ranger Shale > Parker Soil. It should be pointed out, however,
that the 2u limit is somewhat arbitrary and that some clay miner-
als may exceed this size (Mitchell, 1976, p. 2"O.
Clay Mineralogy
While it is generally not possible to predict the behavior
of clays and soils on the basis of their composition alone,
51
-------
Table 2
PARTICLE SIZE DISTRIBUTION*
Sample
Ranger Shale
A
B
Fire Clay
A
3
Kcsse Kaoline
A
B
Parker Soil
A
B
Parker Soil (with
def locc-ilant)
A
B
. Clay
UO
UO
US
•»3
57
U9
10
11
18
22
Silt
Weight Percent
59
59
53
56
U3
51
70
71
66
58
find
1
1
2
1
0
0
20
18
16
20
A and B are duplicates.
*'.r.i*:S3 otherwise noted, all samples were run without def locculant,
-------
clay mineralogy is of extrene importance in understanding the
physical and chemical behavior of soils. Mitchell (1976, p. 71)
has called attention to the dominating influence of the clay
phase by noting that "In general, the greater the quantity of
clay mineral in the soil, the higher the plastic:tv. the greater
the potential shrinkage and swell, the lower the permeability,
th<> higher the compressibility, the higher the true cohesion
and the lower the true angle of internal friction."
Each of the clay minerals is characterized by a particular
structure. Kaolinite consists of a silicate sheet (SiuO,Q)
bonded to an aluminum-hydrcxyl sheet in a two-layer structure in
which replacement of the aluminum and silicon t> other elements
is rare. In montmorillonite, the aluminum-hydroxyl sheet is
positioned between two silicate sheets stacked along the c-axis.
This arrangement allows for variable water content between the
layers; montmorillonite is known to expand in water and in other
polar liquids. Because substitution of other elements such -as
iron, magnesium and zinc, for alujr.ir.um is possible, the chemical
composition of montmorillonite is highly variable. Illite is a
general term referring to a mixture of muscovite (a potassium-
rich minera1.) and montmorilloiiito, both of which are three-
layered c'a^s (Mason, 1953).
Results of the X-ray diffraction determination of clay
mineralogy of ti.e samples evaluated i-. cms investigation are
presented in ."able 3. The clay fraction of the Parker soil was
10 percent by weight; approximately 50 percent of the clay was
montmorillonive, 35 percent iilite, and 15 percent kaolinite.
The clay fraction of the Ranger Shale (50 percent by weight) was
composed of a variety of clay types with about 10 percent being
montmorillonite. The clay fractions of the Kosse Kaoline and
Fire Clay were predominantly kaolinite with about 5 percent
illite-mica and 10 to 16 percent quartz. The amount of expand-
able layer clays was low in all clay-soils evaluated.
Moisture-Density Relationships
The moisture-density relationships for the solids used in
this investigation are plotted in Figures 7, 8, 9 and 10. These
figures display o well-defined relationship between moisture and
density. It has been found that the Harvard Miniature Compaction
unit is easy to handle and requires a much smaller sample to per-
form the test of this type than the larger units normally used.
It is sensitive to minor moisture changes. Its operation some-
what resembles methods used for compaction of clay liners for
industrial waste disposal pits. The clay-soils used in this
study for swelling and permeability measurements were picked at
the optimum moisture content determined from these figures.
Table U presents the optimum moisture content and corres-
ponding dry density for the solids under investigation. Parker'
53
-------
Table 3
CLAtf MINERALOGY
Percentage of Clay Fraction*
Sample Illite- Chlorite-
Kaolinite Quartz Mica Montmorillonite Amorphous
Parker Soil A 1<* - 35 51
B 14 - 39 U7
Kosse Kaoline 85 10 5
<£ Ranger Shale 2M 2B 2M 10 9
Fire Clay 78 16 6
Feldspar
_
-
-
5
-
•Values represent fraction of the clay portion of the sample. Parker Soil was 10.5
percent clay; Kosse Kaoline *• 53 percent; Ranger Shale, UO percent; and Fire Clay,
MU percent. (See Table 2). No deflocculant used.
A and B are duplicates.
Dash (-) indicates negligible amounts found.
-------
TO •
... j
*
I 0
55
-------
Table 4
OPTIMUM MOISTURE CONTENTS
Sample
Ranger Shale
Kosse Kaoline
Fire Clay
Parker Soil
Op cimum
Moisture Content
(%)
17.5
31
16
18
Corresponding
Dry Density
(lbs/ft3) (g/cc)
108
85
113
116
1.73
1.36
1.81
1.86
Table 5
ATTEFBEP.G LIMITS
Sample
Pang«:r Shale
Kosse Kaoline
Firs Clay
Pr.rker Soil
Plastic Limit
Moisture
(%)
36
38
31
22
Liquid Limit
Moisture
(%)
U6
50
32
26
Plasticity Index
Moisture
(%)
10
12
1
14
56
-------
Soil had the greatest dry density at optimum moisture content,
followed by Fire Clay, Ranger Shale and Kosse Kaoline.
Fcr a given soil sample, the maximum dry density represents
the minimum void ratio. That is, at maximum dry density the
ratio of the volume of voids to the volume of solid material is
a minimum. At maximum dry density, the porosity of a given
material is, therefore, at its. lowest point. The dry densities
reported here are sc mewhat higher than those determined by
Sanks e_t al. (1975) who measured values ranging from BO to
95 lb/ft3"To.96 to 1.52 g/cc) for six different Texas soils.
The moisture content and dry density were used to calculate the
void ratio. (See discussion of Permeability.)
Atterberg Limits
The values for the Atterberg Limits for the solid substrates
are presented in Table 5. The plastic and liquid limits were
in the order Kosse Kaoline > Ranger Shale > Fire Clay > Parker
Soil and the plasticity index followed the order Kosse Kaoline >
Ranger Shale > Parker Soil > Fire Clay.
Mitchell (1976, p. 173) has collected data on Atterberg
Limits for various clays and these indicate that montmorillonite
has the highest liquid and plastic limits (100 to 900 percent
and 50 to 100 percent, respectively) followed by illite (60 to
120 percent and 35 to 60 percent) and kaolinite (30 to 110 per-
cent and 25 to UO percent). These data are for various
ionic forms and the ranges of values reflect the dependency of
soil plasticity on the type of substituted cation. The Atterberg
Limits obtained in this study arc quite low when compared with
the ranges given by Mitchell for pure clays. This is probably
due to the high percentage of silt in the samples evaluated in
this study.
Odell et al. (1960) have shown that the Atterberg Limits
can be correlated with the percent clay \n a soil and have pre-
sented equations relating liquid and plastic limits and plastic-
ity index to the percentage of particles less than 0.002 mn
(clay). In general, liquid limits observed in this study were
smaller than those estimated on the basis of Odell et al.'s
equations, while observed plastic limits were larger. The
measured plasticity indices are thus considerably lower than
predicted. Other" factors besides percentage clay, such as per-
cent organic carbon and percent montmc>rillonitH clay also in-
fluence plasticity and these could in part be responsible for
the observed differences.
Seed et al. (196u) have related the plasticity index to the
tendency oT a clay to undergo swelling. In general, the higher
the plasticity index, the greater the total expansion of the
clay when exposed to water. The plasticity indices determined
57
-------
in this study suggest that Ranger Shale and Kosse Kaoline should
expand mere than Fire Clay. This trend was observed in this
study and. is commented on in a later section of this report.
(S»e section on Swelling.)
•L'aticn Exchange Capacity
The cation exchange capacities (CEO of the solid substrates
examined are presented in Table 6. The Kosse Kaoline, Fire Clay,
ar.d Parker Soil had CEC values ranging from 11 to 1U meq/100 g,
whereas the value for the Ranger Shale was about 5U meq/100 g.
The CFC of montmorillonite typically ranges from 80 to 150 meq/
100 g as compared with 3 to 5 meq/100 g and 10 to UO meq/100 g
values for kaolinite and illite, respectively (Grim, 1953). The
low CEC values obtained are in keeping with the generally low
montmorillonite content of these clavs.
TABLE 6. CATION EXCHANGE CAPACITY
Sample
Kosse Kaoline
Fire Clay
P.anger Shale
Parker Soil
Cation Exchange Capacity(CEC)j, (meq/100 g)
Test No. 1
13. U
11.2
5U.8
13. U
Test No. 2
13.3
11.2
54.0
1U.8
»• 'i Average
13.it
11.2
5U.U
14.1
Cation exchange capacity might be expected to have a
marked effect on the capacity of a soil to attenuate the irove-
ment of heavy metals from a landfill to groundwater supplies.
The effect of CEC on the movement of epolar organic solvents
has not been investigated.
Exchangeable Cations
Ranger Shale, Kosse Kaoline and Fire Clay were examined for
exchangeable- Na, K, Ca and Mg. The results which are presented
in Table 6A, indicate that for all these clay-soils, calcium was
the dominant exchangeable cation. Na and K were present in
relatively small amounts and magnesium could not be detected in
any of the samples. Ranger Shale, which had the highest CEC also
had the largest amount of exchangeable calcium.
58
-------
TABLE 6A. EXCHANGEABLE Na, K, Ca, AND Mg
(meq/100 g)
Sample Ca Mg Na
Fire Clay
Kosse Kaoline
Ranger bhale
9.7
•4.6
59
0,5
2.0
1.0
0.3
0.3
0.7
Dash (-) indicates none detectable.
Major Ions
The three clay-soils used most frequently in tests with or-
ganic solvents in this study, namely Ranger Shale, Kosse Kaoline,
and Fire Clay, were analyzed for major elements using the Applied
Research Laboratory Electron Kicroprobe. Five of the major ele-
ments in each clay were determined using the electron microprobe
and their values are expressed in weight percent of oxide in
Table 7. As expected, in all clays used, the dominant cation
determined was Si and the second most dominant was Al.
The relatively high percentage of potassium in Ranger Shale
was a reflection of its substantial illite content, illite being
a clay-sized muscovite (^Al^CSigA^^ (OH)i^), or a muscovite-
mcntmorillonite mixture. The iron present in the Ranger Shale
probably substituted for aluminum in the aluminum-hydroxyl unit,
as in the mineral nontronite.
Kosse Kaolir.e was the richest in calcium while Fire Clay
was characterized by a high aluminum content. Ther» are no
apparent correlations between major ion content and other proper-
ties reported herein.
Total Carbon and Carbonate Content
The percent total carbon and percent carbonate content of
the test soils are presented in Table 8. All four test samples
contained low percentages of both total carbon and carbonate
with the Ranger Shale, Kosse Kaoline and Fire Clay containing
less than one percent total carbon. The Parker soil contained
approximately three percent total carbon. The carbonate content
ranged from zero in the Kosse Kaoline and Fire Clay to approxi-
mately O.U percent in the Parker soil. The higher carbonate
content in the Parker soil is to be expected s ^e it is from a
calcareous area containing large amounts of dex. : i.al limestone.
59
-------
Lee and Jones (1978) have commented on the ability of organic
carbon to increase the sorptive capacity of soils for certain
types of organic pollutants. It also appears to be the case
(see section on Swelling) that clays having a high organic carbon
content swell more extensively in organic (waste) liquids than
clays with a low organic content. The significance of this is
discussed in a later section.
TABLE 7. MAJOR ION CONTENT OF CLAYS STUDIED
(all values in weight percent of oxide)
Clay
Ranger Shale
Kosse Kaoline
Fire Clay
Fe2°3 K2°
6.10 2.35
0.3U 0.21
1.37 0.7U
A1203 CaO
15.13 3.11
27.U 9.36
35.29 U.03
SiOj Total
U2.88 69.57
U1.28 78.59
50.67 S.2.10
TABLE 8.
TOTAL CARBON
AND CARBONATE
CONTENT
Sample
Ranger Shale A
B
Average
Kosse Kaoline A
B
Average
Fire Clay A
B
Average
Parker Soil A
B
Average
% Total Carbon
0
0
0
0
0
0
0
0
0
3
3
3
.61
.60
.60
•11
.12
.12
.03
.03
.03
.16
,19
.18
% Carbonate
0.32
0.33
0.32
0
0
0
0
0
0
O.U1
O.U2
O.U2
A and E are duplicate tests
60
-------
EFFECTS OF ORGANIC SOLVENTS ON CLAY LATTICE SPACING
X-ray diffraction data were collected for the clay-soils
which had been in contact with solvents and with solvent mix-
tures. Values of 26 were converted to lattice spacings by using
the Bragg relationship. All data for pure solvents are presented
in Tables 9 through 12. The purpose of this part of the investi-
gation was to examine the effect of solvents on clay lattice spac-
ings in order to determine whether or not a correlation could "be
found between lattice spacing and permeability. The existence cf
such a correlation would enable permeabilities to be predicted
from easily-obtained X-ray diffraction measurements.
Initial work focused on the effect of water, acetone, carbon
tetrachloride, xylene, trichloroethylene (TCE) and aqueous mix-
tures of these organic solvents on Ranger Shale. Data for the
nsat liquids are presented in Table 9. Examination of these
values indicates that there is little, if any, relationship be-
tween the lattice spacing and solvent type. The d(001) values
ranged from 13.95 A" for TCE to 11.31* & for acetone, with no ob-
vious correlation between the spacings and molecular properties.
Furthermore, the spacings were all sufficiently close to the
d(001) value for Ranger Shale in contact with water that, given
the errors to be expected in these measurements (± 0.2 to 0.3 A),
it was concluded that what «as being observed was the spacing for
the Ranger Shale-water interlayer complex. This indicates that
in the presence of atmospheric moisture, or in the presence of
even the small amounts of water dissolved in the nominally pure
orga.iic liquids, Ranger Shsle will preferentially form a charac-
teristic water complex with possibly complete exclusion of the
organics from the interlaver region. The data in Table 13 show
that this is true for the mixtures as well. (It should be pointed
out here that the mixtures are, with the exception of acetone-
water, two-phase systems where the aqueous phase is saturated with
organic component and the organic phase is saturated with water.
Solvent systems of this type would be in contact with clay in a
typical disposal pit). This behavior can be explained if inter-
calation is recognized to be a competitive process wherein mole-
cules in the treatment solution compete for sites around the
interlayer cation. Water, because it is a small, highly polar
molecule, can effectively out-compete any of the organics studied
here for solvation sites in the clay, even when it is present ir
relatively minute amounts.
Barshad (1952) discussed the problems involved in forming
characteristic clay-organic complexes and showed that for a num-
ber of weakly polar molecules, such a complex can only be formed,
if at all, by dehydration of the clay at temperatures up to 250°C.
Benzene, even under the extreme conditions used by Barshad, gave
no complex. Olejnik et al. (197U) in their study of the mont-
morillonite swelling Tn polar liquids, found it necessary to pre-
treat the clay by heating at 150°C aru u x 10"u mm of mercury to
remove traces of water. They also took care to scrupulously dry
the immersion liquid over a molecular sieve for several days.
61
-------
TABU:
M':;uub 01 X-RAY OHTKATIUN MLASUKI:MI:NT^ i't MK
KANULK SIIAU' ATTLH I XHOSUKL Tu VAKluUli SOLVLNTS
en
Solv^nl
Aceione
Xylene
T«-r4Chlorid
TricMoro-
ethy ler.e
Water*
A
B
A
R
« A
B
A
B
A
B
A
B
A
B
1
i
6
6
6
o
5
D
3
3
6
>
3
3
Si
.2
-
. 3
.7
. 3
.2
.3
.x
. 1
.7
.3
•?
.7
.7
?nd
H
8
8
8
8
8
a
a
•
8
. a
. 7
.8
.9
.8
.6
.9
.9
.»
.8
^
-
_
~
Valuer, ot ?9
3rd
LJ . X
17. 3
17. S
17.x
17.X
17. 3
17.X
17.X
17. H
17.X
-
_
"
xth
17. B
17.7
17.8
17.8
17. H
17. 7
17.8
17.8
17. «
17.8
, _
-
—
~
Sth
^0.9
70.9
21.0
70.9
70.4
70.8
71.0
70.8
70.9
20.9
-
_
•
6th
/X.9
7X.8
2S.O
7X.9
7S.O
7X.9
7X.S
7b.O
7S.O
7X. 9
-
_
•
7tl
/b
76
76
76
76
76
76
26
76
76
h
_ ,
.6
.7
.7
.7
.6
.7
.7
.7
.6
-
_
-
Calculated Valu
1 bt
ix . 3
-
IX. 0
lu . 1
IX. U
IX . 7
IX. 1
13.8
IX. X
ix.x
IX. 1
IX.?
IX. 7
1M . 3
7.".ri
10. U
lu.2
10.0
9.9
10. U
10.0
10.0
10.0
10.0
10. C
-
_
-
}rd
/. 1
7.2
7. 1
7.1
7. 1
7.7
7. 1
7.1
7.1
7.J
-
_
-
Xth
S.O
S.O
S.O
S.O
S.O
S.O
S.O
b.O
S.O
S.O
-
_
-
«.-s of d(A)
Sth
» . J
H. 3
X. 2
X. ?
X. i
X. 3
X. 7
X. 3
X. 3
X.3
-
-
6th
3.0
3.6
3.6
3.6
3.t.
3. t
3.6
3.h
3.b
3.6
-
-
7th
}
3
3
3
3
3
3
3
3
3
. 3
.3
. J
. 3
. 3
. 1
. 3
. 3
. 1
. 3
-
-
A and B dre r?plicat» sampler..
•J sets if duplicates run.
Dash (-) indicates not Measured.
-------
10. FlibULTS OF X-RAY DI »TKACTIv!M ML'ASUKCMtNl S fOK AIR-DRIED
K03SC KAOLlNf IM CuNTACT WITH VARIOUS SOLVLNTS
Sol vent
Vd.ues of 76
Cal".;:,»ted V.ilues of d(A)
i ;t
TnJ
3rJ
"4th
Sth
6th
7th
2nd
3r.j
"4th
Sih
f.fh
7th
Acetone
Fen iene
Xy lene
Carbon
Te trichloride
Trichloro-
ethy lene
Water
A
B
A
B
A
B
A
B
A
B
A
B
S
S
s
">
s
s
s
s
s
s
s
s
. b
.6
.2
.1
.«
.«
.6
.6
. 5
.6
. 3
.1
8
8
8
8
8
8
8
a
8
8
B
B
.9
.•}
.9
.8
.8
.8
.8
.8
.9
.8
.8
.9
12
12
12
12
12
12
12
12
12
»?
12
12
. 14
•»
.„
. 14
.H
••»
.11
.«•
,
-------
TAB;.L
KL:,ULTS ui X-KAY PUTNAM TION Mi>:;i'Hi.Mi.urj I
CIRC CLAY IN IVNTACT WITH VARIuUS b"l.Vi:»«TS
AIP-DKK.L
CT>
Values ot 24
So Ivent
Ace tune
Benzene
Xy lene
Teuach.or-:
TrK-h lor---
ethy lene
Water
A
B
A
B
A
B
.« A
B
A
B
A
B
1st
5,
t
b
b
's
b
,,
b
b
b
i
b
.A
•<
.7
.7
.0
. 3
.b
•»
.S
.b
.6
'*
2nd
8
8
8
a
8
9
8
8
a
a
8
8
.9
.8
.9
r
. ^
.9
.0
.9
.8
.8
. 7
.9
.8
3rd
12.
12.
12.
12.
12.
12.
12.
12.
U.
12.
12.
17.
2
?
7
1
3
2
2
3
3
2
b
3
>»th
17. 7
17.7
17.8
17.7
17. 7
17.7
17.9
17.9
17.8
17.6
17.9
17.8
Jth
72
22
22
2?
22
22
72
22
22
22
22
22
. 7
.S
_,
.H
. b
.6
.b
.b
.H
.3
.8
.b
6th
2
bth 6Lh
3
3
u
14
.1
J
3
3
'!
M
J
i
.9
.9
. J
.0
. .,
5
,,
.9
.0
.0
i
.9
3.b
3.6
3.6
3.6
3.f,
?.6
3.6
3.6
J.6
3.6
3.b
3.6
7th
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3
3
3
»
3
3
3
J
.
'•
1
1
B ar«? replicate
-------
TABLE 12.
RESULTS Of )!-RAY DMTKACTIOM MEASUREMENT" FOR AIR-URIED
PARKER SOIL IN LJNTACT WITH VARIOUS SOLVENTS
O>
in
Solvent
Acetone
Benzene
Xylene
Carbon
Tetrjchloride
Tr icMoro-
ethy lene
Water
Values of 2«
A
B
A
B
A
B
A
B
A
B
A
B
1st
6.S
6.3
6.3
6.7
6.4
6.4
7.0
6.7
7.6
7.i
6.4
6.6
7nJ
8. 7
8.9
8. 3
8. 3
8.4
B.S
8.8
9.1
8.S
8.S
1.1
8.8
3rd
17.6
17.0
17. S
17.4
17.9
12. S
17.7
17.2
12.4
17.4
17-1
17.3
4th
17.0
16.8
16.9
16.9
16.6
16. 3
16.9
16. S
16. &
16.4
16.6
16. 7
ith
',3.1
i3. 1
71.0
23. 1
73.0
73.0
23.1
73.1
23.1
73.7
73. 1
bth
76.7
76.7
76.7
26.7
76.8
26.6
76.6
76.7
76.7
76.6
?6.8
76.7
7th
-9.4
79. S
79.4
79. S
79. S
79.4
79.4
29. S
29.4
29. S
79.5
29. S
1st
13. b
14.0
14.0
14. J
13.8
13.8
17.6
14.3
11.6
11.6
13.8
13.3
Calculated Values of
7nd
10.1
9.9
10.7
10.7
10. &
10.4
10.0
9.7
10.4
10.4
11.0
10.0
3rd
7.0
7.4
7.1
7.1
6.9
7.1
7.3
7.7
7.7
7.1
7. J
7.7
4th
S.7
S. J
S. i
S.7
S. 3
S.4
i.7
S.4
S.4
S.4
S. 3
S.S
Sth
3.8
3.8
3.9
3.8
3.8
3.9
3.9
3.8
j . *
3.8
3.8
3.8
d(A)
Sth
3.3
3. 3
3.3
3 3
3. 3
3.9
3.9
3.3
3.3
3. 3
3. 3
3.3
7th
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
A and B *r« replicate st
-------
TABLE 13. RESULTS OF X-RAY DIFFRACTION MEASUREMENTS OF AIR
DRIED MCNT:-:CRILLON:TE AFTER EXPOSURE TO VARIOUS
SOLVENT-WATER MIXTURES
Treatment SolutionMean Value of d (S)
(Volume % Organic) 1st 2nd 3rd
C% Acetone
5
25
50
75
100
0% Carbon Tetra-
chloride
5
25
53
100
5* Xylene
25
50
100
5% Trichloroethylene
25
50
100
14.36
} 4 . ^4
14.48
14.38
14.60
1U.34
*
(14.23) 14.26
14.13
(14.03) 14.15
(14.22) 14.21
14.12
(14.31) 13.88
(14.04) 14.02
(14.23) 14.13
14.08
14.13
14.26
13.94
13.95
10.04
10.00
10.04
10.10
3.0.21
10.10
10. OC
9.95
9.y4
9.96
10.02
9.94
9.96
9.97
9.96
9.96
10.03
10.02
9.94
7.13
7.15
7.16
7.19
7.1i
7.15
7.14
7.12
7.13
7.13
7.13
7.12
7.10
7.12
7.13
7.13
7.12
7.12
7.13
o
* Values for 1st d(A)for additional run.
Since this study centers on how clays behave when in contact
with organic solvents under conditionr approximating those in the
field, such measures as dehydrating tne clay and the immersion
liquid were considered unnecessary and indeed, inappropriate.
From the work with Ranger Shale, it appears that when an
organic liquid conteininj small quantities of moisture (as would
be tne case in the fi^ld) is in contact with air-dried clay, *-he
Lattice spacing (dOOl)) will be that corresponding to the cl^iy-
water complex (^lu.2A) regardless of the organ:*? solvent involvf-d,
That is, there is no penetration of organic molecules into the
inter-layer spaces.
After the s«:-j'jy of Ranger Shale, it was decided that only
pure solvents should be looked at since mixtures appeared to
66
-------
yie.td little interpretable information. Samples of Kosse Kaoline,
Fire Clay and Parker Soil in contact with six solvents were X-
rayed and the d(QOl) values were computed. These are summarized in
Table lu. Similarities in lattice spacir.gs regardless of solvent
were ?gain apparent. For example, the d(001) value for Kosse
Kaoline in acetone, xylene, CC14 and TCL' were all in the vicinity
of 16 A. The similarity of effects of water and benzene, two
highly dissimilar solvents, appeared to be anomalous. Values for
Fire Clby in acetone, CCli,, TCI. and water are also about 16 A,
with benzene and xylene being slightly greater. For Parker Soil,
d(001) values were all close to lu A with the exception of TCE
which pave a spacing a full 2.5 A lower.
A rationale has been presented for the constancy of the lat-
tice spacing for Ranger Shale in a variety of solvents running
the gamut of molecular properties. The sajne explanation applies
to the other clays. The exceptions noted above are more difficult
to account for and no attempt will be made to do so in this report.
TABLE 1U. d(001) SPACING FOR CLAYS IN CONTACT
WITH PURE SOLVENTS
^-»^Clay Fire
-olvent^"^-^^^ Clay
Acetone
Benzene
Xyio:ie
cciu
V;L
Water
16.
16.
17.
16.
16.
16.
08
9P
18
26
09
07
Xosse
Kaoline
15.
17.
16.
15.
15.
17.
78
16
36
78
69
07
Ranger
Shale
lu.
-
lu.
lu.
13.
lu.
3u
03
12
96
26
Parker
Soil
13
lu
13
lu
11
13
.79
.07
.79
.26
.62
.57
Dash (-) indicates not measured; see previous discussion in
Methods Section.
SORPTION ISOTHERM STUDIES
In wante disposal areas, sorptJon onto the native soil or
c.lay liners is one of the most important mechanisms relied upon
for preventing the movement of hazardous materials from the dis-
posal aroc to groundwater. This characteristic is crucial in
explaining and predicting the permeability of a soil or clav to
wat',-r and organic solvents.
67
-------
Transport of contaminants is primarily determined by the
permeability of the material through which this solution is
passing. As previously discussed, the permeability of clays
c'uid be markedly affected by organic solvents. However, even
if the clays are sufficiently permeable to allow transport of
tn= liquid consisting of a water-organic mixture, the organic
is or.ly ~f consequence in affecting groundwater quality if the
clay sorption capacity for that organic is exceeded. Therefore,
it becomes important to learn something about the extent of
sorptior. of organics on clays. This is an area that has received
attention in this study.
Tables 15,15, and 1"" present the data collected for the
sorpticn of various organic compounds on Ranger Shale, Kosse
Kaoline, and Fire Clay. These data were used to define the
Freundlich isotherms according to: (Glasstone, 1955)
log(x/m)= (log K)+((l/n) leg C )
(3U>
where x = amount scrbcd in millimoies,
m = mass of clay in kg, and
C0 = equilibrium concentration of the solute in
millimole/1.
K and n = empirical constants
The ordinate, x/m, was calculated as follows:
::/m =
where
Ce)V ) /m
(35)
V =
initial concentration in millimole/1 and
volume of solution in contact with clay, in 1.
The Freundlich isotherms for the organic solvents on the three _
clays are presented in Figures li through 2'j. L;.n<»s of best fit
were determined visually. It appears that sorption in these
./r^anic-clay systens generally obeyed the Freundlich relation-
ship as evidenced by the linear relationships seen in Figures
II through 20. The iTeundliih equation constants, X -ind n, were
o^rainec rrom tnc plots ard are reported in Table 18. The values
01 these constants arc useful in esti mating the amounts of sorp-
tion of the solvents on clays under various conditions and solute
and clay concentration ranges.
66
-------
TABLE 15. SORPTION OF ORGANIC SOLVENTS ON RANGER SHALE
(25 i 3°C)
Initial Cone.
(mM)
18.5
9.22
1.56
1.55
1.55
0.899
3.50U
0.5CU
39.16
15 .9U
3.87
Equilibrium
(mMoles)
17.1
8.00
1.U6
0.7«*9
0.767
0.528
0.388
0.379
37.07
1U.75
3.06
Cone. Wt. of Clay
( grams )
Benzene
5.0U
10.01
5. US
m-xylene
10.07
10.03
10.12
10.03
10.06
Acetone
25
25
25
Amount Sorbed
(mMoles/kg)
29.8
13. u
1.8U
3.75
3.09
1.83
0.53
0.62
8.36
U.76
3. 24
Carbon Tetrachloride
9. OS
3.08
2.29
2.29
5.50
1 .83
] .60
l.UO
20
20
20
20
14.36
5.00
2.76
3.56
69
-------
TABLE 16. SORFTICN' OF ORGANIC SOLVENTS ON KOSSE KAOLINE
(25 - 3°C)
T « -°
irf .
9.
1 .
i
j. •
1.
1
J. •
T
J. .
1.
1.
U .
•* ,
U.
4 .
U .
H .
U .
tidi Coac,
(.TuM)
6
7 P
56
56
00
00
00
oc
00
92
32
92
32
•3 **
92
92
Equilibrium
(niMoles)
15.3
0
0
0
0
0
1
1
1
1
1
0
0
3.6e
1.53
1.U6
.875
.812
.686
.638
.559
Carbon
.23
.39
. 39
.33
.12
.83
.66
Cone. Wt. of Clay
(grams)
Benzene
5.03
1.65
1.02
5.03
m-xylene
1.08
2.13
6.03
8.02
15.09
Tetrachloride
3
u
5
6
12
15
18
Amount Sorbed
(mMoles/kg)
65
36
2
T
.*.
28
22
12
11
7
U3
29
21
17
6
U
2
.5
.3
.94
.99
.9
.1
.9
.3
.3
.0
.1
.9
.U
.33
.OU
.60
70
-------
TABLE 17. SORPTICN OF ORGANIC SOLVENTS ">N FIRE CLAY
(25 - 3°C)
Initial Cone.
(mM>
22.8
22.3
22.8
22.8
22.8
22.8
22.8
1.85
1.85
1.95
1.85
1.85
1.55
1.55
2.68
2.68
3.70
3.70
5.12
5.12
Equilibrium Cone. wt . of Clay
(mMoles)
12.55
9.85
8.98
8.61
7.86
7.88
6.25
0.348
0.333
0.290
0.2U9
0.185
Carbon
1.18
0.98
2.11
1.84
2.63
2.48
3.33
3.22
( grams )
Benzene
3
6
12
18
24
30
42
m-xylene
12
18
24
?0
36
Tetrachloride
20
30
20
30
20
30
20
30
Amount oorbed
(mrtoles/kg)
157.2
99.8
46.1
27.6
18.7
12.4
6.7
5.00
2.95
1.95
1.33
1.06
4.62
4.75
7.12
7.00
13. «*0
10.16
22.37
15.83
71
-------
1.3
> J
«
-------
10 •
'0 •
to
to
M
10
0.5
"f«*« 'i. '••••«<«l» '!•«•••• *tr »irl*1k« I
-------
I ) 4 1 t 'I»O «}
"•t f«f
c—
(r
•»•
i
I
1.
]
> > 4 I •> I til
7U
-------
TABLE 18. FREUNDLICH EQUATION CONSTANTS
FOR SORPTION OF ORGANIC SOLVENTS
ONTO CLAYS
System r. K
m-xylene/R.S.
CClu/R.b.
:3enzene/R.S.
Acetone/R. S.
m-xylene/K . K.
CC1U/K.K.
Benzene/K.K.
m-xylene/F.C.
CC1U/F.C.
3enzene/F. C.
0. 36
0. 77
0.93
2.72
0. 3U
0.22
0.63
0.36
0.93
0.18
7.5
1.9
1.2
2.0
42
7.2
1.3
72
4'° -u
2.2x10
R.S. = Panger Shale
K.K. = Kosse Kaoline
F.C. = Fire Cl.ay
One way the results of the sorption tests of the various
solvents on The three clays can be compared is if the amount of
organic sorted is computed at a single equilibrium concentration,
Comparable sorption values across solvents or sorbents can be
obtained Ly extrapolation or interpolation of the x/m value
which corresponds to an equilibrium concentration, Ce, of 1 mil-
limolar. These x/m values for the systems tested are presented
in Table 13.
These values can be interpreted in the following manner:
If one kilogram of Kosse Kaoline, for example, is placed in an
effectively infinite reservoir of a 1 mM aqueous solution of
m-xylene, then the clay would sorb 42 mill:'.moles of the organic
compound. Examination of this table indicates that a greater
amount of m-xylane tends to sorb per unit mass of clay than
either CCl^ or benzene. In general it appears that a greater
amount of carbon tetrachloride than benzene would sorb onto
clays. It is important to emphasize that these conclusions are
based on the 1 mmolar equilibrium concentration of the solvent
in contact with the clay and do nor apply to substantially dif-
ferent concentration ranges. However, for Ranger Shale, where
n increases from m-xylene through acetone, it can be supported
using equation 3'4 that relative sorption is In the order
m-xylene > CClt^ > benzene > acetone for all equilibrium concen-
trations, C .
e
75
-------
TABLE 19. SORi-TlCN OF CRGA:,'i: COMPOUNDS ONTO CLAYS
AT EQ'JILIBr.I'JM CONCENTRATION'S OF 1 mM
( values ir. .Tjr.oies/kg)
-"'a ' ,-i^-r Water rartitic
I- J. a v „ ...... ,
'..oe: f vcier.t"
Ranger Kosse Fire •
Shale Kaoline Clay
r- yler.e
Cnrbon
tetrachloride
o-inner.e
Acetone
3.15
2.6U
2.13
-C . 2u
7 _
1.
1.
2.
5 U2 72
3 7.2 u.O
— u
2 1.3 2.2x10
0
I .is I; '-) indicates not measured.
-After Hansch et al. (1963).
The x/m value obtained for acetone at Ce=l mM (2 mmoles/kg)
falls within the range observed by Sanks e_t al. (1375) who
fo-ind that acetone was sorbed by Texas clays to the extent of
0.8 to ID mmoles/kg. They observed similar values for phenols.
The extent of xylene uptake measured in the current study was
cms iderably greater than the scrption of apolar DDT determined
i>y Sanks e_t al. (1975). This is surprising since xylene is a
more hydrophobia, less soluble substance than DDT and its, sorp-
tion would be expected to have been less.
In light of the difficulties in determining sorption ten-
dencies of highly volatile organics and in light of the goal of
developing a model to predict the behavior of solvents and
Ivent mixtures based on mc.i.-«> easily measured characteristics
: the solvent(s) and the clays, it is of value to determine if
he sorption tendency of a sclvent can oe related to one of its
ore readily measureable characteri*tics. Baughman and
,i33iter (1978) suggested th^t the physical driving force for
-,rptiori of molecules from aqueous solution is the fugacity, or
':.3ndency of the solute from solution. One readily
-f.iinabl-e measure of the escaping tendency of a substance is
~.-, oct-inol/water partition coefficient, P, defined as the
'.uilibriun ccnc< ntration of a substance, i, in octanol, divided
•• its concentration in water when a small amount of i is
^uiliLrated with an immiscible mixture of water and octanol,
.e.,
-------
P = [C^Coct)] / CC^HjO)] (36)
Thus, it might be expected that the greater the octanol/water
partition coefficient, the greater the extent of scrption from
aqueous solution onto a given clay-soil. The octanol/water par-
tition coefficients for the organics tested are presented in
Table 19. For Ranger Shale, Kosse Kuoline, and Fire Clay the
order of sorption from the 1 miilimolar solution paralleled the
order of the log of the octanol/water partition coefficients,
i.e., m-xylene, with the highest log P value, showed tne greatest
sorption, and benzene, with the Icwest log P value, showed the
least sorption at a C of 1 millimole/i. An exception was the
sorption of acetone oH Ranger Shale which was somewhat greater
than for benzene at 1 mM. It wculd be expected, therefore,-that
substances such as n-xylene which have large P values or high
escaping tendencies from aqueous solution would be taken up and
immobilized by clays to a greater extent than benzene, for ex-
ample, which has a smaller P value.
Based on the limited data collected it also appears that
clay-soils having a high total weight percentage of kaolinita,
such as Kosse Kaoline (u5 percent) end Fire Clay (3<* percent),
have a larger sorptive capacity for apolar or weakly polar com-
po^nds than clay-soils with xow kaolinite content such as Ranger
Shalo (9.6 percent). Based on thefe results it appears that soils
rich in kaolinite will tend to reduce the concentration of hazard-
ous, hydrophobic compounds in the aquatic environment to a greater
decree than those clay-soils which are relatively poor in this
mineral.
In the permeability section of this report it will be argued
thac the large and unexpected difference in the permeability of
clay-soils to water on the one hand, * nd to several organic sol-
vents covering a range of bulk and molecular properties on the
otner, is due to the ability of the clay-soils to strongly sorb
and therefore to impede the motion of organic molecules through
the column. It is at this juncture that the sorptive properties
of clay-soils - organic systems becomes crucial in explaining and
predicting the results of columr. tests and of field experiments.
As a final note on the sorption studies, it was observed
that meta-xylene appeared to undergo iaomeric conversion to ^ara-
xylene in the presence of Ranger Shale. A comparison of chroma-
tograms for the m-xylene standard solution (Figure 21) and tl-.e
same standard solution after 2U hours' contact with Ranger Shale
(Figure 22) shows the appearance of a distinct new peak having
the same retention time as p-xylene after exposure of the standard
solution to Ranger Shale. Vhen a solution containing a mixture of
ortho, meta, and para isome-s of xylene was left in contact with
Ranger Shale for a three we»k period, the height of the peak asso-
ciated with p-xylene incre?.>ed relative to t>e others. Figure 23
is the chromatograin of the initial /ylene solution; Figure 2U
77
-------
P-«X If»t
Atr Pt<«
'• \
TV
•!. C^ro«ct*onl)
•—•-•/ Ifn
$«•»)( 8
of "-tf'«nt itl»dir«
contict
tott S««p1*l A iff I >«rt tli«" fro* 4lfftrtflt
ttt t**t tr»itntnt. Only A inovl conitMlo".
r»c»'»'i9
-------
rt«t
Fl9ur«
Ckro«*ta«ri« a' HtndiriJ talutlan coit»lnin9
« •titurt Of ix'«n«i (Trlpllcitt I n j «c t • on l )
n.t
figure 24. Chro*«togr«» of tttndtrd sjlytlon of >y<«n«i
»fttr lhr«t •««k eipo»u'« to R«n9«r Sh«l«
(OjpMc*t< injections)
-------
shows the increase in the height of the para-xylene peak after
Icr.per ter.T. exposure tc Range:' Shale. This phenomenon was net
investigated furc.-.ev, but it suggests that possible mici-o-
bi:/logi;al activity cf the clay-soils might play a role in the
:r-ir.sforrr.ation cr c-2-:rr.pc5iti.in of orr.anic molecules in the
:i:;vircr.nent, ar.c! r.ight have a pronounced effect on parameters
?u:.:h as perir.eab: li *•/ 'see r.er.r.oability discussion).
-------
SWELLING T CLAYS IN CONTACT WITH WATER, ORGANIC SOLVENTS, AND
SOLVENT MIXT'.RES
Another characteristic of clays which must be considered in
evaluating the potential of a given system to transport organic
solvent wastes is the solid's tendency to snrink or swell upon
exposure to a liquid such as water and wastes disposed of on
land. It is important then to detemine the effect that organic
rolvents and solvent mixtures may have 01 this parameter. If a
solvent causes ar expandable layer clay to sw-=ll, this action
could aid in waste retention: however, if it caused the latti.e
to shrink, additional transport ol" contaminants through a clay
liner isay result.
Results
Data on shrink and swell characteristics of the clays in
contact with the various solvents and solvent mixtures examined
are tabulated in Appendix "ables A 1 through A 25. The percent
swell of the clays during contact with each solvent and solvent
mixture tested is plotted as a function of time in Figures 25
through US; smooth curves were fitted to the data visually.
Benzen"" -.aused the two clay-soils tested (Ranger Shale and
Fire Clay) to shrink prior to causing them to swell 'Figures 25
and 26). In the case of Ranger Shale, after shrinking by 0.07 per-
cent, the clay swelled rapidly, reaching an equilibrium of about
0.05 percent swell after a total of abcut three d,-ys' exposure.
For the Fire Clay exposed to benzene there was a three-day lag
period prior to the shrinking (by 0.03 percent) during which
swell to about 0.03 percent was noted. Overall, benzene caused
the Fire Clay to swell to a greater degree than Ranger Shaln;
equilibrium at about 1.3 percent swell WAS reached after 23 -Jays.
Xylene caused both Ranger Shale and Fire Clay to shrink
(Figures 27 and 28). For both clays equilibrium was reached at
about 28 days; maximum shrinkage of Ranker Sh-jle was about 0.11
percent and of Fire Clay, about 0.25 p«-c»nt. Xylene caused the
Kosse Kaoline to swell C.2 percent in about two days (Figure 29).
The clay then appeared to gradually shrink by about O.OU percent
over the next 20 days of exposure.
Carbon tetrachloride caused Ranger Shale to swell after an
approxi.-nately s-ix day lag period (Figure 30). By Che lut!i Jay
of exposure the clay had swelled by 1.1 percent where it rem.iin. duration of monitoring. The carbon te*rachloridc causer!
both ths Kosse Kaoline and Fire Clay to shrink (Figures 31 c*nd
32). The Kosse Kaoline shrank oy about 0.013 percent in :he
first c'.ay and maintained that level for about two weeks when th>-
apparent equilibrium percent shrink increased to ^bout 0.0?6
percent. Fire Clay showed a linear decrease in size until day 7
when it reached an equilibrium of about 0.63 percent shrink.
-------
1"
t
it n
tt '•* "- cut <• mini
ItM ?MI> XI
" -08»
•
• -00*
-aio
•an
• I It '• B M It
82
-------
H IP
J
j
t .
> 4 • n
J
i
i •••
II. h»ll !!!• f IMM IMIlM !• Ufttm
83
-------
Each of the three clay-soils exposed to trichloroethylene
showed a different shrink-swell response. Ranger Shale (Figure
33) swelled by 1 percent over a several day period. Kosse Kao-
lir.e (Figure 3u) showed a gradual shrinking by about 0.1 percent
during the first 32 days of exposure. During the following 12
days, it showed a net swell of about 0.6 percent. Fir-i Clay
(Figure 35) showed a swell of about 0.6 percent by day 2; the clay
then shrank by about 0.53 percent during the following two weeks'
exposure. Its net equilibrium result was about 0.07 percent
swell.
Acetone caused the Ranger Shale, Kosse Kaoline and Fire Clay
to swell, each reaching equilibrium in about one day (Figures 36,
37, and 38). Ranger Shale swelled by about 5 percent in the first
day and equilibrated at about 4 percent swell by about the second
day of exposure. Kosse Kaoline swelled by 8.5 percent and Fire
Clay by 3.5 percent.
The last solvent tested on each clay was deionized water.
For all clay-soils, it caused greater swelling than any of the
other pure solvents tested. Duplicate swell measurements were
made using Ranger Shale in deionized water. The data in Appendix
Tables A16 and A17 show that in the first run, at equilibrium, the
clay-soil had swelled by 11.55 percent and in the second run, by
11.83 percent. The rates of swell and swelling profiles (Figures
39 and UO) are the same within experimental error. The difference
in the two values indicates that the relative error to be expected
for these measurements is on the order of 2 percent. Kosse Kaoline
in water also swelled by about 11 percent in abo-it S days (Figure
Ul); Fire Clay swelled by about 8 percent in 12 days (Figure U2)..
For the same solvents tested, ncne of the three clays consistently
showed greatest amcunt of swell; Ranger Shale and Kosse Kaoline
tended to shrink to a greater extent than Fire Clay, however.
For the pure organic solvents tested witl. Fire Clay, bsnzene,
acatone, and trichloroethylene caused a net swell (in order of
decreasing amount of swell) and xylene and CCl^ caused net shrink-
ing. Exposure of Kosse Kaoline to acetone, Trichloroethylene, and
xylene resulted in n=t clay swelling (in order of decreasing per-
cent swell) over the period monitored. Carbon tetrachloride
caused Kosse Kacline to shrink.
Exposure of Ranger Shale to pure metharol (Figure 43),
glycerol (Figure 44), carbon te/trachloride (Figure 30) and tri-
chloroethylene (Figure 33) resulted in clay swell (in order of
decreasing amount of net swell) whereas xylene caused Ranger Shale
to shrink.
Percent swell of Ranger Shale was tested using three acetcne
benzene mixtures; 25/75 mcle % acetone/benzene (Figure 45);
50/50 mole % acetone/benzene (Figure 46); and 75/25 mole %
acetone/benzene (Figure 47). In the 25/75 acetone/benzene solu-
tion Ranger Shale showed an initial shrinkage as was seen with
84
-------
CO
tn
I II. toll »f«Mr«tM »•» •Hii> •>!• I*
trli«IOT*l»fl«*
(to* !«•!• Ml
ii i« n ii it n M
1«t (•» P*r« CUr l
jilt
I I U U
M M M
IW. 1^1. lit)
llr M HI.
M. W«ll «r«*«rtt«« r«y •«•«•» Ucl« !•
•llt^ •'•(•f*
UM r«ilm 'II tfi ml
-------
- I
CD
O>
llM 'Ml* •'«!
4
. I
KM II M
I*. (Mil I IIIHtM «^ »l
ll« ><•!• till
r, Cllr '•
-------
II
!•
rt«« ioni
M. IMM monlM »•» !•»«• IMIIX <• MIMIH4
(M> !«•!• III)
I
J
I II It
<• (Wrii
_ I
J
: •
\ I i ' '
It II II 14 II
lM '•!• •••«••• IMK <• till Ult I IMM Mft»r«l ••
'•wl mM> Mifo* O^.MMI IM imii IM IM
flty. r«ti |p«r» »M*«*tMir •"•ftM (*• •••••*««
TIM IMril
87
-------
ftf«r« 41. W«M »r*««riU> f«r •••ffcr IMU t«
«•« »••!••« Ill Ml«l)
IM« IMU Wtl
UM (It •*!•«!
rtfuv* «4. u«it iMMtiM r«» u*t«r t**i* >• •(•u^t
• I M«UM IM f>KI) IM M«IM« IM Mltll
(SM IMU XII
HOT r«u in)
88
-------
the pure benzene solution (Figure 25). The clay-soil swelled
after the first day by about 5.5 percent durir.$ the following
two weeks' exposure. That percent swell was maintained for the
following three weeks whan the test was terminated. The Ranger
Shale swelled to a greater extent when exposed to the 25/75
acetone/benzene mixture than when exposed to either pure benzene
(about 0.05 percent) or pure acetone (net U percent in two days)
(Figure 36).
Tne 50/50 mole percent acetone/benzene mixture caused
Panger Shale to swell by about 2.2 percent during the first two
days' exposure, shrink by about 0.4 percent by day 6, then
to steadily swell by a net of U percent when the test was ter-
minated (day 36) (Figure 46). Equilibrium was not apparently
reached during that time. The 75/25 mole percent acetone/ben-
zene mixture caused the Ranger Shale to behave in a manner
similar to when it was exposed to pure acetone (Figures 47 and
36). Response to the mixture was not rapid as it had been in
the pure solvent, however,
The 50/50 volume percent xylene/deionized water immiscible
mixture (Figure 48) caused the Ranger Shale to shrink by about
0.15 percent during the first four days (as seen for pure xylene
(Figure 27)), and then increase essentially linearly to a maximum
of about 0.9 percent swell by day 34 when the test was termi-
nated. Pure deionized water had caused the clay to equilibrate
at a percent swell loif about 11 percent after several days ex-
posure. !
The 50/50 volume percent acetone/water mixture caused
Ranger Shale to equilibrate at about 11 percent swell after
about one week of exposure (Figure 36). This response was
essentially the same as the response to exposure to pure water
(Figure 39).
Discussion
One of the clay-soils (Ranger Shale) contained measurable
amounts of montmorillonite. Montmorillonite is an expandable
layered silicate mineral, each layer consisting of an aluminum
octahedral sheet between two tetre.hedral silicate sheets. The
forces holding the negatively charged montmorillonite layers
together are weak electrostatic interactions between the ex-
changeable interlayer cations and either the surface oxygen (if
one adopts the Hoffman-Endell-Wilson structure) or the surface
hydroxy groups (if the Edleman-Favejce structure is accepted)
(Theng, 1974).
X-ray investigations have shown that water or other polar
liquids can penetrate into the interlamellar regions of montmo-
rillonite and cause expansion along the c-axis. Olejnik e_t al.
89
-------
(197U) have pointed out th^-t the degree of swelling (as measured
by X-ray diffraction) of montmorillonite in water depends, on a
variety of factors including the polarizing power of the exchange-
able cation, the hydrostatic pressure, and the salt concentration
of the swelling liquid solution. .For the case of non-aqueous sol-
vents, Barshad (1952) examined the effect of the dielectric con-
stant of the immersion liquid on interlayer spacing and found
that the degree of swelling of dehydrated montmorillonite and
verniculite increased with increasing dielectric constant when
the clays were immersed in polar liquids (such as alcohols and
ketones) of similar dipole moments. However, low dielectric
liquids such as benzene, pentane, hexane and octane caused no
interlayer swelling as detected by X-ray diffraction. Since me-
chanical swelling of clays in low dielectric media was observed
in the current study using the consolidometer, an explanation is
required.
While the purpose of this study was neither to provide a
detailed mechanism for the swelling process, nor to sort out the
significance of each of the numerous molecular and ionic proper-
ties affecting that process, an attempt has been made to corre-
late the swelling data with certain bulk or molecular charac-
teristics of the clays and solvents under investigation.
Since swelling experiments were done not with the pure clay
but rather with a sample of the clay-soil containing water (optimum
moisture content), the interlayer cations of the montmorillonite
were probably hydrated to an appreciable degree before the clay
was immersed in the test liquid. Walker (1961) demonstrated
that this water of hydration is of two types: the first, Water
I, is directly coordinated and thus strongly bound to the inter-
layer cation; the second, Water II, is indirectly linked to the
cation through the primary hydration shell. This second type
of water is more labile than the first and hence more susceptible
to replacement. Tnterlayer swelling of Ranger Shale in a given
liquid can, therefore, be attributed to any or all of the follow-
ing:
1. Direct solvation of an unsolvated or partially solvated
interlayer cation (Type I interaction).
2. Indirect solvation of interlayer cations through bridg-
ing wa'ier molecules (Type II interaction).
3. Replacement of either Type I or Type II water, or
both, by molecules of the immersion liquid. This
would only be expected for highly polar molecules.
Evidence that water and polai' organics generally interact
with exchangeable cations to cause interlayer swelling, rather
than with the negatively charged silicate sheets, has been pro-
90
-------
vided by infra-red spectroscopy (Mortland, 197C). For fi given
clay sample, it seems reasonable that properties such as the
dipole moment and the polarizability of the molecules of the
immersion liquid, or the bulk dielectric constant of the liquid
itself, should be related to the degree of swelling.
It is known that the interlayer swelling discussed above
can cause at most a doubling of the volume of the dry clay. The
greater degree of swelling o^ten observed for montmorillonite
and other clays requires that a second mechanism be postulated.
The surface of a clay particle ca-. be thought of as a nega-
tively charged area on which water molecules are adsorbed and
to which cations are attracted by coulombic forces. Because of
thermal agitation, cations are not rigidly bound to the surface
but instead form a diffuse ion layer (diffuse double layer)
whose average distance from the negatively charged surface de-
pends upon such factors as ionic charge and dielectric con-
stant of the medium in which the particle is immersed (typically
water). The interaction of the diffuse ion layers associated
with two adjacent particles is important in determining the ex-
tent of swelling.
Interparticle swelling can be envisioned to occur in the
following manner. If a clay is immersed in water, there will be
a tendency for the water to be adsorbed at negative particle
surfaces, as long as the free energy of the adsorbed water is
lower than that of the free water. This adsorbed watei' will
force adjacent particles apart causing expansion. Van Olphen..-'
(1963) refers to this stage as short range particle interaction,
or swelling due to surface hydration. It has been estimated
(Yong and Warentin, 1975) that when the inter-particle distance
exceeds about 15 A, diffuse ion layers — ones associated with
each particle surface -- begin to form. Since the concentration
of ions is higher in the region between particles than in the
pore water itself, water molecules will move in response to
this concentration gradient into the interlayer region. Addi-
tionally, so-called "osmotic swelling" results. Van Olphen
(1963) calls this longer-range particle interaction either
osmotic swelling or electrical double layer repulsion.
X-ray data obtained in the current study show that the
various solvents used did not cause interlayer swelling to occur.
It can be concluded, therefore, that the expansion of clays
observed in consolidometer tests was the result of some combina-
tion of short-range and long-range repulsion between particles,
that is, interparticle swelling.
For double layer repulsion, those factors which tend to
increase the extent of the double layer near a particle surface
a.lso tend to increase the swelling potential of a clay or soil
(Mitchell, 1976, p. 118). According to the Gouy theory, double
layer thickness, 1/K, is determined by the following relationship:
91
-------
(1/K) = [(ekT) / (8tmoe2z2)]* (37)
where e is the dielectric constant, no is the electrolyte con-
centration, 2 is cation valence, e is the unit electrical charge,
k is the Boltzmann constant and T is absolute temperature. From
this it can be seen that an increase in dielectric constant of
the immersion fluid should cause an increase in the thickness of
the diffuse double layer near a particle and that this shruld
lead to enhanced repulsion and swelling.
Table 20 presents the dipole moments and dielectric con-
stants for the solvents tested and the resulting swell for Ranger
Shale, Kosse Kacline and Fire Clay. First, it should be noted
that even the two substances having zero dipole moment, carbon-
tetrachloride and benzene, caused some expansion of clay exposed
to it. Second, and more importantly, it can be seen that there
is a rough relationship between the degree, of expansion and the
magnitude of the dielectric constant, as predicted by the Gouy
Theory and as shown in Figures "40 a, b, and c. (Plots of per-
cent swe]l vs c (dielectric constant) showed better linearity
than plots using e1* as might be suggested in equation 37.)
In Figures 49 through 51, percent swell caused by each sol-
vent is plotted as a function of the solvent's dielectric con-
stant for each of the three clays studied. In all cases, the
solvents with high and intermediate dielectric constants (water
and acetone, respectively) caused the clays to swell much more
than those solvents in the low dielectric range (benzene, xylene,
carbon tetrachloride and trichloroethylene). The deviation from
linearity suggests that during swelling experiments some of the
adsorbed water remained attached to the clay surface. This
would cause the dielectric constant to be higher than expected.
Interestingly, the montmorillonite content of the clays
appeared to have little effect on the degree of swelling. That
is, Ranger Shale, which was the only clay-soil to contain mont-
morillonite, swelled 11.7 percent in water, the same as Kosse
Kaoline which contained no measurable montmorillonite. The sol-
vents with low dielectric constants caused about the same degree
of swelling ii. all three clays. The fact that there is no cor-
relation between the percent expandable layer silicate (mont-
morillonite) and the degree of swelling reaffirms the conclusion
that for these systems, interlayer expansion is of relatively
little importance.
Not shown in Figure U9 is the point for methanol. Ranger
Shale swelled more than 11 percent in this solvent but the re-
sult was vitiatad by the fact that a dense bacterial or fungal
growth occurred throughout the liouid and the clay. It seems
92
-------
TABLE 20. PERCENT SWELL OF 1EST CLAYS IN CONTACT WITH LIQUIDS COMPARED TO
DIELECTRIC CONSTANTS AND DIPOLE MOMENTS OF LIQUID
t£>
OJ
Immersion
Liquid
Water
Glycerol
Methanol
Acetone
Trichloroethylene
Carbon
tetrachloricl-3
Benzene
Xylene
Dielectric*
Constant
79.38
42.5
32.7
20.70
3.42
2.24
2.28
2.32
Dipole*
Moment
(Debyes)
1.82
-
2.87
2.69
0.80
0.0
0.0
0.3
Ranger
11.
5.
11.
4.
1.
1.
0.
-U.
Equilibrium
Shale Kosse
Percent
Kaoline
7 11.7
3
14**
0
0
1 -0
05
11 0
-
-
8.7
0.6
.03
-
.16
Swell
Fire
8.
-
-
:<.
0.
-0.
l.
-0.
Clay
2
0
1
6
3
25
*After Riddick and Bunger (1970).
**See text discussion.
Dash (-) indicated not tested.
-------
10
9
8
7
6
1 5
«-• 4
c
I 3
2
I
0
-I
-2
-O-o —
) 0-0
y
/
*'
.° /
/
Acetone
Illllllllllllllf
Fi
2 4 6 8 10 12 14 16 18 2O 22 24 26 28 30 32
Time (days)
gure 48a. Comparison of Swell Properties of Fire Clay In CC1.,
Acetone, and Water: Effects of Low, Intermediate,
and High Dielectric Solvents
-------
to
en
—O-—O-O-O- —O
Water
Acetone
O-O
-00-0 CC1
8 10 12 14 16 18 20 22 24 26 28 30 32 34
Time (days)
Figure 48b. Comparison of Swell Properties for Kosse Kaollne In CC1.,
Acetone, and Water: Effects of Low. Intermediate, and
High Dielectric Solvents
-------
-------
Water
12
I I
10
9
8
7
"a* 6
CO
+* 5
c
a*
£ 4
o»
Q.
3
2
I
0
-I
Glycerol
20
40 60
€ (dielectric constant)
80
Figure A9. Relationship between solvent dielectric constant and
percent swell of Ranger Shale
-------
ID
00
-------
ID
(O
10
9
6
7
6
V
s
00
01
u
i.
4*
a.
5
4
3
2
I
0
-I
B«utn«
,Ac«ton«
20
40 60
€ (dielectric constant)
80
Figure 51. Relationship between solvent dielectric constant and
percent swell of Fire Clay
-------
clear that the extent of swell was influenced by this growth.
The phenomenon should provide an interesting future investiga-
tion.
While the dielectric constant of the solvent plays the major
role in determining swell properties, a more subtle factor for
the hydi-cphobic solvents is the percent organic carbon content
of the clay. In T^ble 21, degree of swelling and milligrams
organic carbon per gram clay are presented. For all those three
hydrophotic solvents studied on three clay-soils, an increase
in percent swell was observed with an increase in organic car-
bon content from Fire Clay with 0.3 mg C/g, to Kosse Kaolins.
with 1.2 mg C/g, to Ranger Shale with 2.8 mg C/g. This is
likely the result of Van der Waal's interactions between sol-
vent molecules and the organic substances contained in the clay-
soils.
TABLE 21. PERCENT SWELL OF CLAYS DUE TO HYPROPHOBIC SOLVENTS
IM RELATION TO ORGANIC CARBON CONTENT OF CLAY
(values are percent swell)
Solvent
Xylene
CCl^
TCE
Fire Clay
(0.3 mg C/g)
-0.25
-0.6
0.1
Kosse Kaoline
(1.2 mg C/g)
0.16
-0.03
0.6
Ranger Shale
(2.8 mg C/g)
-0.11
1.1
1.0
For a number of clay-solvent mixtures, the phenomenon of
shrinkage has been observed. The clay-organic systems have
therefore been classified into four categories, depending on
whether they exhibited only swelling, only shrinkage, or some
conbination of the two (Table 22). Typically, those solvents
which caused swelling only were polar liquids while those which
caused only shrinkage were apolar. However, there is no one
generalization that can be applied to Table 22.
The behavior of a clay in a given organ?c solvent depends
upon a complex array of factors. Therraodynamic equilibrium
will be attained once the following two conditions are satisfied
(where y is chemical potential)i
£. H.O H.O
w = ]t i (38a)
clay soln.
100
-------
b. org org.
W = P
clay soln,
(38b)
Condition 38a states that the chemical potential of the
water in the clay is equal to the chemical potential of wa~er
in the organic solvent. Condition 38b states that the chenical
potential of the organic in the clay is equal to the chemical
potential in the solution (traces of.water dissolved in organic
liquid).
TABLE 22. CLASSIFICATION OF CLAY-ORGANIC SOLVENT SYSTEMS
ACCORDING TO SWELL PROPERTIES
Swelling Swelling Then
Only Shrinking
R.S./HjO R.S. /Acetone
R.S./Gylcerol K.K./Xylene
R.S./Methanol F.C. /Acetone
R.S./CC1U F.C./TCE
R.S./TCE F.C./Xylene(N.S.
K.K./ Water
K.K. /Acetone
F.C. /Water
R.S. = Ranger Shale
F.C. = Fire Clay
N.S. = Net Shrinkage (Net Swell
observed unless indicated
otherwise)
Shrinking Then Shrinking
Swelling Only
R.S. /Benzene R.S. /Xylene(N. S . )
K.K./TCE K.K./CC14(N.S.)
F.C. /Benzene F. C./CC14(N.S. )
)
K.K. = Kosse Kaoline
TCE = Trichloroethylene
Two transfer processes occurred during the experiment,
namely, the transfer of some water from the clay to the sol-
vent, that is dehydration of the clay; and transfer of solvent
from the liquid into the clay. Depending on which one of these
processes predominated at equilibrium., the clay would either
swell or shrink. For those clay/solvent systems which undergo
swelling, there is a net transfer of liquid into the clay; for
those systems which undergo shrinkage, there is net transfer
in the opposite direction.
The factors which favor movement of water out of the clay
are a) interlc.yer cations with small charge/radius ratio, and
b) relatively high water solubility in the organic liquid.
101
-------
Factors which tend to favor movement of the organic into the clay
are a) weak intermolecular forces in the liquid itself, and b)
strong interactions of the organic molecules either with clay
cations or with organic materials found in the clay.
The situation as outlined above is sufficiently complex at
the molecular level tc preclude predicting at this time the ef-
fect of a given solvent on a giver, clay. Even an analysis of
the elemental composition of the clays has not provided the basis
for a rationalization of the observed behavior in terms of inter-
actions at the molecular level.
While the degree of swelling depends upon thermodynamic
factors, the rate is a matter of kinetics. Whether swelling
precedes shrinking or vice-versa is determined by the relative
rates of transfer of water out of the clay and for organic
solvent into the clay matrix. Table 22 shows that, in most
cases, the latter process occurs more rapidly causing swelling
of the clay.
The behavior of carbon tetrachloride and xylene deserves
additional comment. Of all of the solvents studied these
were the only ones which caused a net shrinkage to occur in
some of the clays (see Table 22). Carbon tetrachloride
caused Kcsse Kaoline and Fire Clay to shrink while xylene
brought about shrinkage of Fanger Shale and a net shrinkage
(after initial swelling) of Fire Clay. In all experiments in-
volving these solvents there was shrini^ge at sor.te time during
the run. In column tests discussed in! the next section,
xylene, benzene, and carbon tetrachloride tended to "break
through" tne clay plug in the perme
-------
TABLE 23. SWELL PROPERTIES OF RANGER SHALE IN
FIVE SOLVENT MIXTURES
System
Percent
Swell
Comment
25 mole% acetone/
75 mole% benzene
50 mole% acetone/
50 mole% benzene
75 mole% acetone/
25 mole% benzene
50 volume% xylene/
50 volume% water
50 volume% acetone/
50 volume% water
benzene
acetone
xylene
water
5.75
5.7
U.6
0.8
11
0.05
it
•0.11
11.7
System not at equili-
brium
System far from equili-
brium
Net swell
Net shrink
The conclusion drawn from these data is that in mixtures of
solvents, a clay will tend to swell as though it were immersed
in the component of higher dielectric constant only. That is,
Ranger Shale swelled about 11 percent both in pure water and in
the 50 percent acetone/water mixture. In the three acetone/
benzene mixtures, the percent swell ranged from t.6 to 5'. 75,
which compares closely to the t percent swell observed for Ranger
Shale in pure acetone. Thus it appears thct a clay selectively
sorbs that component of a miscible solvent mixture with which it
can interact more (most) strongly. Since the driving force
for swelling is the energy lost when surface or interlayer ions
are solvated by molecules of the swelling liquid, it seems
reasonable that the more polar component of the mixture (water
in the case of acetone/water mixtures; or acetone in the case
of acetone/benzene) should be preferentially imbibed by the clay.
The significance of this observation is that it suggests
that in cases where a disposed solvent might cause a clay liner
to shrink and crack (e.g., xylene or CCl^), that solvent should
be co-disposed of with a liquid such as acetone or water, which
103
-------
would prevent this from occurring. This possibility will be
discussed in the next section.
Concluding Remarks
The swelling and shrinkage behavior of clays and soils in
contact with water has been a subject of major interest to the
civil engineer for many years. Swelling of soils can have a
profound impact on the integrity of building foundations and
pavement; cracking, often associated with shrinkage, can have
adverse effects on the stability of dams and embankments.
In general, shrink-swell behavior has received little at-
tention in the literalrure from those concerned with the storage
of waste materials in landfills. For example, Sanks et al.
(1975) investigated the suitability of clay beds for Industrial
waste storage and reported in some detail on the characteristics
of the clays used. Fuller (1977), who studied the movement of
various metals in soils also discussed a number of important
factors to be considered in selecting disposal sites. In
neither report, however, was shrink-swell behavior discussed.
Also virtually absent from the literature are reports
dealing with the effects of organic liquids on the behavior of
soil. As Mitchell (1976, p. 125) noted in his discussion of
the effects of the low dielectric solvent ethanol on the extent
of the diffuse double layer, "detailed consideration of the in-
fluence of dielectric constant may seem academic because the pore
fluid in soils usually is water." Mitchell went on to say that
there may be certain cases where oil or warte chimioals could be
the pore fluid. These cases, however, have apparently not re-
ceived wide attention.
The clays and soils used in this study all had low plasticity
indices according to the criteria of Holtz and Gibbs (195R)
and it is not surprising that the measured volume increases
were generally small (maximum -\, 12 percent), even for immersion in
water. The greater degree of swelling observed for the Kosse
Kaoline and Ranger Shale in water (11.7 percent) when compared
with Fire Clay (8'percent) can be explained in terms of the
higher plauticity indices for the former clays. The results
here agrse with expectations based on the work of Seed et al.
(196U).
The observation that organic liquids caused swelling to occur
in consolido:neter tests but did not change the characteristic
interlayer distance, d(001), for the clays, indicates that the
mechanism for volume change is interparticle repulsion. This
conclusion is supported by the fact that the high dielectric
liquids cause greater swelling than low or intermediate di-
electrics, in keeping with the general theory of clay particle
interactions.
104
-------
The clay organic content, recently discussed by Lee and
Jones (1978) in relation to sorption, was established as an
important factor governing the percent of swell of clays in
nonionic hydrophobic solvents.
Shrinkage was observed for some clays in carbon tetrachlor-
ide and xylene, probably as a result of clay dehydration by
these solvents. In nature, shrinkage of clay soils also occurs
as the result of changes in moisture content and is often asso-
ciated with cracking. The cracks formed in dehydrated clay
soils can conduct wastes through an otherwise impermeable
material (Yong and Warentkin, 1975). The results of this study
suggest that certain solvents of low dielectric constant have
an ability to cause shrinkage and cracking and might, therefore,
cause damage to clay-lined pits. This implies that co-disposal
measures, which have been discussed by Streng (1976) for
hazardous inorganic wastes, might also be employed for spent
industrial solvents. The results of work with mixtures in the
current study suggest that low dielectric liquids be mixed
with those of higher dielectric constant in such proportions
as to give a smooth swelling curve.
105
-------
EFFECTS OF ORGANIC SOLVENTS ON PERMEABILITY
The coefficient of permeability is probably the single most
important laboratory-determined parameter for predicting the
movement of hazardous solvents through clay liners. However,
this :'.s one parameter which can be measured only over long
periods of time (typically on ths order of a month), using equip-
ment which is neither commercially available nor readily con-
struct?ble. One of the objectives of this research has been to
relate the coefficient of permeability to more readily measurable
parameters of clays and solvents in order that permeability might
be estimated under a variety of circumstances.
Results
Permeability data were collected on 15 clay/solvent systems.
These data are presented in Appendix Tables B 1 through B 19.
Figures 52 through 70 show the coefficients of permeability over
t ime.
For most of the pure organic solvents tested, Ranger Shale
showed patterns of decreasing permeability for the first five
to ten days, oefore a plateau was reached. After eight days'
exposure to benzene, Ranger Shale exhibited breakthrough (Figure
52). The permeability of Ranger Shale to xylene equilibrated
after about ten days at 4 x 10~9 cm/sec ;Figure 53); the per-
meability to carbon tetrachloride reached an apparent equilibrium
of 25 x 10~9 cm/sec after about four days (Figure 54). Equi-
librium permeability of Ranger Shale to trichloroethylene,
2 x 10~9 cm/sec, was reached after about two weeks whereas
that of acetone, 2.5 x 10~9 cm/sec, was reached after approxi-
mately one month's exposure (Figures 55 and 56). Equilibrium
permeability of Ranger Shale to methanol was about 15 x 10~9
cm/sec and was reached after about three weeks of exposure
(.Figure 57); for glycerol, permeability at equilibrium was
0.9 x 10~9 cm/sec after about a week's exposure (Figure 58).
The permeability of Ranger Shale to deionized water remained
mere or less constant at about 38 x 10~9 cm/sec during the one
month's monitoring period (Figure 59).
The permeability of Kosse Ka-3line to m-xylene (Figure 6C),
acetone (Figure 61) and deionized water (Figure 62) appeared to
decrease during the first week or so of exposure and then in-
crease to a plateau level. For xylene an equilibrium permea-
bility coefficient of about 50 x 10~9 cm/sec was reached after
about 2U days; for acetone, the equilibrium was about 65 x 10*9
cm/sec and was reached after about three weeks; and for de-
ionized water, equilibrium of 220 x 10"^ cm/sec was reached
after about two weeks.
As was found for permeability of Kosse Kaoline to xylene,
acetone, and water, the permeability of Fire Clay to xylene
106
-------
• <»
J
|k«lt It •••<•«•
IIM tUK II)
t
t
• 4
• •
II
110 IMfll
II.
IIM '«•!• Ml
14. CM'flilMt •' t»p«*M
-------
rl ri
"ttn M. 'ieMauu 2
IU. TOW til
•• «
• • .
•
. •
t« MIMIIM mUf
(!•• TMIt Ml
108
-------
"«w» M. c«t"
'XllWI •> MMMIIIt* •' ltlt< '
'»>• Ml
J 1
^ I
; •
i.
II It II II U II
'»• IMTII
I(M 'Mil 111!
109
-------
(Figure 63), acetone (Figure 6U), and water (Figure 65) de-
creased in the fiiat week of exposure, remained low for 1.5
to 2 weeks, and then increased. It appears that xylene broke
through after maintaining a permeability coefficient of about
1 x 10~3 cm/sec. Equilibrium levels for acetone and deionized
water were 7 x 10~9 and 13.5 x 10"9 cm/sec, respectively; both
reached equilibrium after about four weeks. The coefficient of
permeability of Fire Clay to trichloroethylene decreased over
the first month of monitoring to an equilibrium level of about
2.5 x 10~9 cm/sec (rigure 66).
There was considerable scatter in the coefficient of per-
meability values for Ranger Shale with mixtures of 75/25 mole
percent acetone/benzene and 50/50 mole percent acetone/benzene
such that patterns over time were difficult to determine (Fig-
ures 67 and 68). Equilibrium values of about 7.5 x 10~9 cm/sec
and on the order of 15 to 2-0 x 10~9 cm/sec, respectively, were
found for those systems. The permeability of Ranger Shale to
the 25/75 mole percent acetone/benzene mixture appeared fairly
constant at about 5 x 10"9 cm/sec over time (Figure 69). Per-
meability behavior over time did not appear to be related to
that found fop either the benzene or acetone pure systems.
Figure 70 shows the permeability over time of Ranger Shale
to a 75/25 mole percent acetone/carbon tetrachloride mixture.
Behavior appeared similar to that for acetone alone, although
equilibrium was reachad more rapidly and the equilibrium per-
meability was somewhat higher (9 x 10~9 cm/sec) than that for
acetone.
Table 2U summarizes the values of equilibrium permeabilities
of the clays to the pure solvents tested. Ranger Shale showed
greatest permeability to carbon tetrachloride and water (discount-
ing the benzene breakthrough), and least permeability to glycerol,
trichloroethylene and acetone. Both Kosse Kaoline and Fire Clay
showed greatest permeability to water (discounting xylene break-
through), followed by acetone. For all solvents tested, Kosse Kao-
line was much more permeable than either Ranger Shale or Fire Clay.
Initially, experiments were carried out in an air condi-
tioned room without careful temperature control. Large daily
fluctuations in K under these conditions were probably due to
changes .in room temperature, which had the effect of increasing
or decreasing the volume of fluid in the permeability chamber.
This volume change was then reflected as a rise or drop of the
meniscus in the 8 mm diameter graduated standpipe. After the
initial runs with water and Ranger Shale, experiments were con-
ducted in a temperature controlled room in order to minimize
scatter in the data.
The most common pattern observed in the permeability data
was an exponential decrease in K during the early stages of the
110
-------
: n
i
2 n
it II II
»if.rt II. CMmitnt if n>«Mtllt> •' >ln (lit u iflM*
(!«• '«•!• I'll
ft*
II
r• Ill)
111
-------
MUD >.< »•.!«. (It M
! ..
J "0 « • l« !• M
TIM (MMI
tt. CM>rict>» «» nrMWilllr '•• •••*' <••<•
U • •!•!•?• •/ t^MM (II MUII tM
M«IM« l» Ml«l)
ItM lMI« (It)
i ..
• 1C
I
• III)
••.. |MMl>l|
- H
•
s
« 4 • It
10 «4 21 M
II*. (Mf>l
*"*" M?IW III Mljtl ••' «tr£«« MlriolwIM (l» Ml.X
|l.< 1*1. lift
-------
TABLE 2«*. EQUILIBRIUM PERMEABILITIES OF CLAYS TO ORGANIC
SOLVENTS AND WATER
(x ID"9 cm/sec)
Benzene
Xylene
Carbon tetrachloride
Trichloroethylene
Acetone
N^thanol
Glycerol
Water
Ranger
Shale
2 (Break)
4
25
2
2.5
15
0.9
38
Kosse
Kaoline
—
SO
-
-
65
-
-
220
Fire
Clay
^
1 (Break)
-
2.5
7
_
-
13.5
Dash (-) indicates not measured.
experiment followed by a gradual leveling off as the system
approached equilibrium. Variations in K with time have been
reported in the literature by a number of investigators (Allison,
19U7; Bodman, 1937). Poulovassilis (1972) studied the factors
governing this phenomenon for water and aqueous solutions in
contact with natural clays and showed that by far the most pro-
nounced reductions in K were caused by microbial production of
gases, and consequent clogging of soil pores. Because of evi-
dence for microbial activity observed during the sorption and
swelling tests conducted during the current study, it is possible
that the decline in K with time was due in pa^t to microbial
production of either C02 or CHi^ during the early stages of the
experiment when organic concentrations in the column were still
sufficiently low to be non-toxic (see Alexander, 1961 for a
discussion of the effect of soil microorganisms on organic com-
pounds).
In addition to raicrobially induced changes in K, changes
arising from a reduction in the pore space geometry of the clay
also seem plausible. Poulovassilis (1972) has used this ex-
planation to account for the 15 percent decrease in hydraulic
conductivity (coefficient of permeability) observed with steri-
lized soils where microbial action was not a factor.
What is perhaps most striking about the permeability data
collected during the current study is the fact that each of the
organic solvents when in contact with a given clay had equili-
brium permeability values which were smaller than that of water
on the same clay, by an order of magnitude in most cases. Ac-
cording to Davis and De Wiest (1966), the coefficient of per-
meability K, can be determined by:
113
-------
K = ccT (Y/V)
(39)
where c = dimensionlesss constant or shape factor which
includes grain size, packing and porosity,
d = average pore size of the porous medium,
Y = density of the solvent, and
M = viscosity.
If the first two terms (c and d) are taken to be characteristic
of the clay or soil, the coefficient of permeability will depend
upon the ratio Y/U- Table 25 presents Y/W values for solvents
evaluated in this study. Based on the equation above, most of
the organic solvents woul^ have been expected to move through
Rangar Shale between I.H and 2.5 tiraes more rapidly than water.
The results of the current study, however, which showed that
water had the highest permeability of all solvents tested, con-
tradicted this prediction. It is interesting to note that it is
likely that Equation 39 (basically Y/u) which was developed
based on permeabilityof saline solutions would likely have been
used as a basis for estimating the permeability of organic wastes
if this study had not been conducted. It is evident from the
results of this study that this approach is not appropriate
for organic solvents.
TABLE 25. DENSITY-VISCOSITY RATIOS FOR ORGANIC SOLVENTS
AND WATER
Solvent
Y (g/cc)
Patio
Water
Glycerol
Methanol
Acetone
Trichloroethyiene
Carbon Tetrachloride
Benzene
m-xylene
1.0
1.26
0.796
0.79
0.79
1.59
0.979
0.60
1.0
mi2
0.551
0.32
0.3K
0.100
C.6U9
0.50
1.0
8.9 x 13-"
l.i*i*
2.H6
2.32
1.59
1.35
1.20
After Riddick and Bunger (1970).
It was mentioned earlier that the swelling of grains tends
to have a retarding effect on the transmission of fluids through
a clay by decreasing the pore sizes. Data have been examined,
therefore, for a possible correlation between the degree of
swelling rind the permeability. Table 26 lists the values for
swell properties and coefficients of permeability for all single
liu
-------
TABLE 26. PERMEABILITIES AND PERCENT SWELL FOR ALL SINGLE
SOLVENT SYSTEMS TESTED
Clay
Ranger Shale
Kosse Kaoline
Solvent
Benzene
Xylene
Carbon tetra-
chloride
Trichloro-
ethylene
Acetone
Methanol
Glycerol
Water
Xylene
Acetone
Water
Percent Coefficient pi Permeability
Swell (:t 10"' cm/sec)
0.05
-0.il
1.1
1
!»
11. U
•5.3
11.7
0.16
8.7
11.7
2 (Break)
u
25
2
2.5
15
0.9
38
50
65
220
Fire Clay
ill \
Xylene
Carbon tetra-
chloride
Acetone
Water
-0.25
-0.6
3.6
8.2
1
2.5
7
13.5
(Break)
solvent systems tested. It can be seen that the markedly smal-
ler permeabilities of the organic solvents cannot be attributed
to attenuation of flow through swelling. In fact, the solvent
which caused each of the c]*ys to undergo the greatest swelling,
namely water, had the largest coefficient of permeability - a
result which directly conflicts with expectations based on
earlier work with aqueous solutions.
To account for the permeability results, it is necessary
to consider how an organic molecule Cor water) might move
through a sorbing, porous medium containing interstitial water.
Such a medium has been described by van Genuchten and Wierenga
(1976) as consisting of the following five regions: (see
Figure 71)
1. Air space
2. Inter-aggregate water
3. Intra-aggregate water
115
-------
Figure 71. Representation of water-containing clay
After van Genuchten & Ulerenga (1976)
(See text for discussion of areas)
116
-------
4. A dynamic soil region, in contact with the inter-
aggregate water, where sorption can occur readily
5. A stagnant soil region where sorption can occur
only after organic molecules have diffused through
the surrounding intra-aggregate water.
Molecules of the permeant move by diffusion and convection
through the inter-aggregate water and at each stage along the
way are partitioned between the aqueous phase and the dynamic
soil region according to the Freundlich equation. Those mole-
cules which are weakly sorbed by soil particles tend to move
quickly through the aqueous channels. Hydrophobia substances
such as benzene, xylene, carbon tetrachloride and trichloro-
ethylene, which are highly partitioned at any instant onto the
soil or clay phase, would be expected to have lower permeabili-
ties than water and acetone. This model suggests a plausible
correlation between the coefficient of permeab:lity, whitih
measures the rate of movement of the permeant along the column,
and the octanol/water partition coefficient, which measures the
tendency of permeant molecules to escape from the aqueous phase.
Along these lines, Briggs (1973) has shown that the rate of
movement of unionized pesticides through soils correlates with
their octanol/water partition coefficients.
It can be seen in Table 27 that, in general, for « single
clay-soil, the permeability of a liquid decreases as ';he log
of its octanol/water partition coefficient, P, increases. Since
P is a measure of escaping tendency of the material.from water
(Baughman and Lassiter, 1978), it seems reasonable that those
substances least compatible with water should move most slowly
through the column. This generalization seems to hold for sol-
vents on Kosse Kaoline and Fire Clay, as can be seen when the
coefficient of permeability is plotted against the log of the
octanol/water partition coefficient for solvents on these clays
(see Figure 72 and Table 27). In both cases, the more positive
the value of log P, the smaller the value of K, or in other
words, the more hydrophilic the organic, the more rapidly it
moves through the culay. This relationship did not apply as con-
sistently in the case of Ranger Shale, however. The three sol-
vents which exhibited anomalous behavior may have done so for
the following reasons:
1. Carbon tetrachloride. Despite a high value for log P,
carbon tetraohloride had a coefficient of permeability
much larger thai, that of the other hydrophobia solvents
studied (i.e., benzene, xylene and trichloroethylene).
This is possibly the result of i".s ability to cause
clay shrinkage. This is discussed below.
117
-------
cm/sec
o
X
r
Pormeabt
<*.
o
c
01
o
01
o
220
£10
200
190
ISO
170
160
150
140
130
120
MO
100
SO
80
70
60
50
40
30
20
10
Fire Cloy
•2-10 I 234
Log P
Figure 72. Relationship between coefficient of per-
meability and octanol/water partition
coefficient
118
-------
TABLE 27. PERMEABILITIES AND GCTANOL/WATER PARTITION
COEFFICIENTS FOR SOLVENTS ON THREE CLAYS
Clay
Solvent
Equilibrium
Coefficient of
Permeability
(x 10~9 cm/sec)
Log Octanol/Water
Partition
Coefficient
Ranger Shale
Benzene
Xylene
Carbon tetra-
chloride
Trichloro-
ethylene
Acetone
Methanol
Glycerol
Water
2
U
25
2
2.5
15
0.9
38
2.13
3.15
2.64
2.37
-0.24
-0.32
-2.56
-1.15
Kosse
Kaoline
Fire Clay
Xylene
Acetone
Water
Xylene
Carbon tetra
chloride
Acetone
Water
50
65
220
2.5
7.0
13.5
3.15
-0.2>+
-1.15
3.15
2.64
-0.24
-1.15
Acetone. This is a hydrophilic substance having a
negative value for log P; it would, have been expected
to move through the clay much core rapidly than benzene
and xylene. That it did not suggests the possibility
that microbial decomposition of acetone on the biologi-
cally active clay, Ranger Shale, resulted in CO? produc-
tion and clogging of pores, thus reducing the oSserved
permeability value.
G.iycerol. This has a more negative log P value than
waver itself and would hc've been expected to permeate
at a higher rate than water. Here, however, the extreme
viscosity of glycerol (1400 cp as opposed to 1.0 cp
for water) was clearly a factor in reducing flow. It
should be noted, however, that glycerol permeates much
more rapidly than would have been expected based on a
simple comparison of its viscosity with that of water.
Given this consideration alone, it should have moved at
119
-------
l/1000th the rate of water. That in fact it had a K
value of I/36th that of water can be explained in terms
of its high affinity for the aqueous phase. Work with
glycerol suggests that viscosity is a factor that should
be considered in any model of solvent flow through clays
but it is clearly not as important as the hydrophilic
or hydrophobic character (as measured by P or by e) of
the permeating substances.
There *lso appears to be a good relationship between the
permeability of clay to a solvent and the solvent's dielectric
constant, such that the pro.iter the dielectric constant, the
greater the value of K (Figures 73 through 75). This is not
surprising in view of the above model, since the dielectric con-
stant is another approximate measure of a liquid's hydrophobic
or hydrophilic character. Substances with high dielectric con-
stants tend to be hydrophilic and can therefore be expected to
move more quickly through the aqueous channels in the clay. Low
dielectric substances will be sorbed and thus retarded in their
movement. The relationship between this parameter and the coef-
ficient of permeability will be discussed at the end of this sec-
tion.
A number of the systems clearly showed anomalous behavior
with respect to the permeability - dielectric, constant relation-
ship; one of these was carbon tetrachloride/Ranger Shale. Here,
the value of K was too high when compared with other solvents
having similar dielectric constants and octanol/water partition
coefficients. As noted earlier, of all the solvents studied,
CCl^ had the greatest tendency to cause shrinkage of clays. As
a result of this shrinkage, channels were formed which allowed
the solvent to flow through the system more readily. Measured
permeability val les were therefore higher than expected and this
caused the point for CCl^ to be far from the line of best fit
in Figure 73. Glycerol, as discussed previously, had an unusual-
ly low permeability probably because of its high viscosity. This
is reflected in its position in Figure 73.
Although the benzene/Ranger Shale system showed solvent
breakthrough due to clay shrinkage during several permeability
tests, it appeared to fit the general relationship between per-
meability and dielectric cons cant shown in Figure 73. Fire Clay
in xylene also shewed solvent breakthrough eventually. The
point for this system in Figure 75 represents equilibrium
position seen in Figure 63.
Solvent Breakthrough
An important aspect of the question of the effect of organic
solvents on clay permeability is the breakthrough phenomenon
which was observed in a number of solvent/clay systems. These
120
-------
a;
in
*^
E
U
o\
o>
E
fc.
40r
36
32
28
24
20
16
I 12
u
8
CCI
Xylene
Water
'Methano!
.Acetone
Glycero:
0 10 20 30 40 50 60 70 80
€ (dielectric constant)
Flgjre 73. Relationship between coefficient of
permeability and dielectric constant for
solvents in Ranger Shale
121
-------
Water
200
180
E
u
160
i
o
*—
* 140
- 120
J3
-------
o
01
U)
C7>
I
O
20-
18
X
$ 16
I 14
-------
cases were related to the fact that certain solvents caused the
clay to shrink, thus forming channels through which the solvent
could flow, in addition to migrating by convection in two unsuc-
cessful runs, with the permeability testing system, CClij broke
through Ranger Shale in six and eight days. After one break-
through, fresh CCli4 was poured onto the column and was observed
to flow along the glass walls of the permeameter.
Another extreme case of solvent breakthrough caused by
shrinkage was observed with benzene on Ranger Shale (see Figure
52). There, the permeability vs. time curve appeared to be nor-
mal for the first seven days of the run, but by the eighth day
all of the solvent had drained out of the permeameter chamber. A
less dramatic case of breakthrough occurred with xylene on Fire
Clay (see Figure 63) when after 2U days of normal behavior,
the coefficient of permeability rose precipitously over a four
day period from an equilibrium value of 1 x 10~9 cm/sec to
US x 10~9 cm/sec. Examination of Figure 28 shows that Fire Clay
underwent slow shrinkege in xylene, and it is likely that this
was responsible for the breakthrough. While initial shrinkage
was seen in the Shrink-Swell tests conducted on Ranger Shale
with benzene, the Ranger Shale - CCl^ system showed about 1.1 per-
cent swell during the first 10 days of exposure.
It can be concluded that any solvent which causes a clay to
shrink appreciably during any stage of a shrink-sveil experiment
is a potential hazard when stored in a clay-lined disposal
area. Even though such a solvent may move slowly through a
wet clay by diffusion and convection, there is in principle a
danger that it will eventually cause shrinking and cracking
and thereby allow fluid transmission in bulk. Because of its
obvious implication for the storage of commercial organic sol-
vents in clay-lined pits, this is one of the signal observations
to be derived from this study.
Properties of Clays of Potential
Importance in Governing Permeability
Table 28 summarizes several properties of the clays inves-
tigated which might be influential in governing their
permeability to solvents.
Montmorillonite Content; Montraorillonite is an expandable
layer clay; closure to liquids could cause swelling of indivi-
dual montmorillonite grains and clogging of soil pores and
channels. Ranger Shale was the only clay-soil evaluated that
had measurable montmorillonite. While it might have been ex-
pected to have the lowest permeabilities, this was not observed.
Void Ratio: The "oid ratio, or volume of voids divided by
volume of solids, (e = VV/VS) is generally acknowledged to be an
important factor in determining permeability. (See, for example,
12U
-------
TABLE 28. CLAY-SOIL CHARACTERISTICS
Packed, Bulk PercentPercent K (x 1C"5
Clay-Soil Density Organic Montmoril- Void cm/sec)
lb/ft3g/cc Carbon linite Ratio For H20
Ranger Shale
Kosse Kaoline
Fire Clay
108
85
113
1.73
1.36
1.81
0.28
0.12
0.03
"* O.i+ -0.53 38
C.37-0.7U 220
0.19 13.5
Ter^aghi and Peck (19U8) and Araer and Awad (197U)). Based on
void ratios, Ranger Shale should have been the most permeable
clay, followed by Kosse Kaoline and then Fire Clay. The ob-
served order was Kosse Kaoline > Ranger Shale > Fire Clay.
Therefore, for the limited number of clays evaluated in this
study, there was no correlation between void ratio and perme-
ability.
Clay Density: Clay density, which is a function of mineral
composition and packing, seems like a less fundamental parameter
than void ratio and no correlation between this and permeability
was anticipated prior to the experimental work. Yet, of all the
properties considered, density appeared to be the one which
correlated best with the observed data. The observed, relation-
ship was that as packed, bulk density increases, the permeability
of the clay decreases. This is shown in Figure 76. (Later in
this section of the report, an empirical equation is derived
which relates permeability to clay density).
Permeability to Solvent Mixtures
Permeability data were obtained on four binary solutions
containing acetone mixed with either benzene or carbon tetra-
chloride., The choice of solvent systems was based on the need
for acquiring information on the behavior of mixtures, particu-
larly those containing one component that causes clay shrinkage
and another that causes swelling. Shrinkage can in principle
bring about a rapid transmission of fluid through the clay and
should be prevented at all costs, possibly by judicious mixing
of solvents. This set of experiments was designed, in part, to
test the effectiveness of such mixing in preventing breakthrough.
Solution compositions and permeability coefficients for the sol-
vent mixtures evaluated are listed in Table 29.
The acetone-benzene mixtures all had higher permeability
coefficients than either of the pure components, but there
appeared to be no correlation between permeability and solvent
mixture composition. One likely explanation for the higher
125
-------
220
200
80
u
01
160
2 140
X
>»
Z 120
01
E
u
0)
0.
100
r. 80
01
O
60
40
20
80
90
100
110
I 1
120
1.3 1.4 1.5 1.6 17 1.8 1.9
—lb/ft3
g/cc
Figure 76. Coefficient of permeability as a function
of packed bul^ clay density
126
-------
TABLE 29. COEFFICIENTS OF PERMEABILITY FOR FOUR
BINARY MIXTURES
(with Ranger Shale)
System
K (x 10"9 cir./sec)
25 mole% acetone/
75 mole% benzene
50 mole% acetone/
50 mole% benzene
75 mole% acetone/
25 mole% benzene
75 mole% acetone/
25 mole% carbon tetrachloride
acetone
benzene
carbon tatrachloride
5.0
17.5
7.5
9.1
2.5
2.0 (before break-
through )
25
permeability values follows from the earlier discussion of swell-
ing. There it was noted that clays tend to behave selectively
in their interactions with mixtures. Such selectivity probably
comes into play during permeation as well, with one component
of the mixture moving more rapidly than, and independently of,
the other. Since acetone has the lower octanol/water partition
. iefficient it should migrate more quickly through the saturated
pores of the clay. Thus the measured coefficients for the
mixtures probably represent, to a large degree, the movement of
acetone alone. This is entirely consistent with the view that
the clay plug acts as a chromatographic column and partitions
molecules according to their sorption behavior.
This, however, leaves unexplained the fact that mixtures
have a higher permeability than acetone itself. One possibility
is that the K value obtained for pure acetone is low because
of the aforementioned microbial activity. For the mixtures,
such activity would be prohibited by the bactericidal properties
of benzene and carbon tetrachloride.
The nost important observation, however, is that none of
the mixtures, after a period of 36 days, underwent breakthrough
as did pure xylene, benzene and, in two preliminary cases, pure
127
-------
carton tetrachloriae. This is significant, because it indicates
thaf mixing two solvents in such proportion as to cause swelling
(recall that the thr?e mixtures being discussed all caused Ranger
Shale to swell more than 5 percent) prevents breakthrough front
occurring. Even though the coefficients of permeability are
higher for the mixtures than the pure solvent, they are still
quite low as compared with water.
For the mixture of acetone and carbon tetrachloride, the
coefficient of permeability was markedly lower than for carbon
tetrachloride alone. The lower value .'an be attributed to the>
fact that in this mixture the clay die' not undergo shrinkage, so
there was no possibility of leakage &±;und the edges. Again,
the measured permeability was due largely to the movement of
acetone, and it was for this reason that the K value for the
acetone/carbon tetrachloride mixture (75 volume percent/25
volume percent) closely approximated that for acetone/benzene
(75 percent/25 percent).
Finally, it should be noted that the behavior of mixturts
was, in general, quite similar to that of pure solvents, both in
terms of the temporal variation in the permeability coefficient,
and in terms of its magnitude. Based on an understanding of the
behavior of the components, it is now possible to predict the co-
efficient of permeability for organic solvent mixtures. Based
on the dielectric constants as shown in Table 20 and the findings
shown in Table 29, a quaternary solution composed of eqva.l
amounts of benzene, xylene,I carbon tetrachloride and mettanoi,
would be expected to have a coefficient of permeability (using
Ranger Shale) of 15 to 30 x 10~9 cm/sect a somewhat higher
permeability value than for methane!.
While there is considerable work to be done in the area of
mixture permeability studies, this investigation h?.s shown that
an understanding of the basic properties of organic solvents and
clays can, in principle, be used '.o predict the behavior of com-
plex industrial wastes.
128
-------
CONCLUDING REMARKS
The coefficient of permeability, one of tha most important
soil characteristics affecting contaminant transport from land
disposal sites to groundwater, can have a wide range of values
extending over many orders of magnitude. Terzaghi and Peck
(1967) have summarized representative values which are presented
in Table 30.
TABLE 30. COEFFICIENTS OF PERMEABILITY FOR
TYPICAL SOILS
Soil Type
Comment
Course Gravel
Sand or Fine Sand
Silty Sand
Silt, Fine Sandstone
Clay
> 1 x 1C"1
10'1 to 10'3
10'3 to 10'5
10'5 to 10"7
< 1 x 10
~7
Very Permeable
Medium Permeability
Low Permeability
Very Low Permeability
Impervious
After Terzaghi and Peck (1967).
K values as low as
clays.
2 x
en/sec have been observed for some
Based on this system of classification, Ranger Shale
(K = 38 x 10*9 cm/sec for water) and Fire Clay (K = 13.5 x 10"9
cm/sec for water) are "impervious" and Kosse Kaoline (K =
2.2 x 10"7 cm/sec for water) has a "very low permeability."
With respect to the organics studied, all of the soils are
"impervious" since the highest coefficient of permeability ob-
hain*>rt in f-iia unr
-------
K= [l/(k0T2s£)] [n3/(l-n)2] [r/n] 0*0)
K = coefficient of permeability or hydraulic
conductivity (cm/sec)
kQ= pore shape factor (^2.5)
T - tortuosity factor (^ /?")
So= specific surface area per unit volume
n = porosity
r = viscosity
n = unit weight of pore fluid
has been discussed in detail by Mitchell (1976, p. 2U8) and con-
siders largely the effect of unequal pore size in the clay.
Mitchell has also noted that the permeability of a clay can be
strongly influenced by molding water content and by method of
compaction and that, in general, "properties of any given soil
are dependent on structure to such an extent that analyses based
on properties determined from the same material but with a dif-.
ferent structure may be totally in error." The problem of
predicting permeability is thus related to the considerable
difficulties involved in adequately representing soil structure by
-± suitable parameter.
Ideally, the Kozeny-Carmen equation should be applicable
to all permeants since the fluid viscosity and density are in-
cluded as variables. However, the results of this study point
out an additional problem, namely, that the permeability of a
non-aqueous liquid does not depend on the ratio of density to
viscosity as widely supposed but rather on the hydrophobic and
hydrophilic character of the solvent. This hydrophobic/
hydrophilic nature is best represented either by the logarithm of
the octanol/w,ater partition coefficient or by the dielectric
constant. Any future efforts to develop a general permeability
equation for the movement of liquids through a water-containing
soil should take these into account.
The permeability data reported here were analyzed in the
following way: It was assumed that the coefficient of per-
meability, K, was related to the dielectric constant, e, by
the equation
K = Le° (1*1)
m was the slope obtained by plotting log K vs. log t for sol-
vents en a given clay; L was a constant. For all three clays,
m was estimated to be about 0.75.
Based on limited data, the density of the clay appeared to
be the other significant variable and the dependence of K on
density, d, was obtained from
130
-------
K = Rdn
From a plot of log K vs. leg d for each solvent, n, the slope,
was determined to be approximately 7; R was a constant.
Finally i combining Equaticns ul and U2, K was plotted
against eO.yS/d? on log-log paper (see Figure 77) and from
the measured slope and intercept the empirical dependence of
the coefficient of peraeability on solvent dielectric constant
and density was determined to be
log K = 1.17 [Jog (e°'75/d7)] - 7.23 (*3)
Table 31 presents calculated and observed values for the 11
single solvent systems which behaved in reasonable accord with
the model discussed in this section. In general, the estimates
of measured K values are quite good. The purpose of developing
this equation was twofold:
1. It provides an empirical relationship) between per-
meability and the two nujor variables found in the
study and therefore acts as a summary of the per-
meability res Mies for this work.
2. It calls attention to the important role which di-
electric constant plays in determining permeability.
In its present form, this relationship cannot be used to predict
pei-neabilities for any c3.ay-sol"ent system. It is expected,
for example, that it would not hold at moisture contents other
than opt ilium, nor would it apply to the same clays compacted in
a different way. In general, it summarizes too small a number
of data points to be of significance other than as a prototype
for equations incorporating dielectric constant.
A luore empirical approach to predicting permeabilities of
clay liners to hazardous organic wastes would be to obtain the
permeability of the liner to water and then, using the results
of this study, estimate the permeability of the clay to the*
waste. Based on the foregoing discussion, this could be ac-
complished using the equation
Ks " Kw «VEw>°'75 <*">
where K and K are the coefficients of permeability of the
solvent and water, respectively, and e end e ara the respective
dielectric constants. w
131
-------
lOOOr-
~ 500
o
4)
e
o
o
x
Ol
o
O)
5
L.
0.01 0.02 0.03 O.I 0.2 0.5 I 2 5 10
«or"/d7
Figure 77. K vs. E0<7^/d7 for organic solvents and water on
three clays
.RS • Rang»r Shale
FC - Fire Clay
KK • Kosse Kaollne
Ace • Acetone
Ben • Benzene
Met • rtethanol
Wat • wat»r
Xyl • .X
13;
-------
TABLE 31. CALCULATED AND OBSERVED PERMEABILITY
VALUES FOR ELEVEN SINGLE SOLVENT SYSTEMS
System KObS Kcalc
(cm/sec x 109) (cm/sec x 10*)
Ranger Shale
Benzene 2.0 1.38
Xylene u.O l.<*0
Trichloroethylene 2.0 1.96
Methanol 15.0 . 1U.O
Water 38.0 30.8
Kosse Kaolins
Acetone 65.0 67.5
Water 220 219
Fire Clay
Xylene 1.0 1.00
Carbon tetrachloride 2.5 0.98
Acetone 7 6.8
Water 13.5 22.2
The development of an equation which can be used to predict
the permeability of any soil to any permeant is a formidable
task and attempts to accomplish this have met with only limited
success. The contribution of this study, in this regard, has
been to point out the significance of permeant properties other
than viscosity and density and to show how these might be used
to estimate the permeability of soils to hazardous organic
wastes. In addition, it has been shown that the ability of lov;
dielectric liquids to dehydrate and shrink clays and to possibly
cause them to transmit fluids more rapidly, adds a significant
dimension to the problem of waste storage and containment. How-
ever, the results of mixture experiments indicate that the
potential for breakthrough is greatly diminished or eliminated
by the addition of a. high dielectric liquid such as water.
The results point out the need for more laboratory studies
to be undertaken on the effect of clay properties on the
transmission of organic solvents. There is also a crucial need
for field work to test the validity of the generalizations of-
fered in this report. Until this has been carried out it will
net be clear to what extant the conclusions drawn here actually
apply to a real waste storage problem.
133
-------
SECTION 7
SURVEY OF CURRENT PRACTICES FOR LAND DISPOSAL
OF HAZARDOUS WASTES
INTRODUCTION
Shortly after initiating this study, it was decided that it
would b« informative to determine current state regulations and
practices for land disposal 'of hazardous wastes. In late summer
1977, a letter was sent to each state's pollution control agency
requesting a copy of the regulations governing land disposal of
hazardous wastes. Specific information was requested on the
handling; of waste organic solvents, specifications for permeabil-
ity of soils and/or liners (maximum permeability coefficients)
for industrial waste disposal pits, methods for determination of
permeability cf soils and liners, and other information pertinent
to land disposal cf hazardous industrial wastes with a high or-
ganic solvent content.
Many of the states responded to the initial request during
the fall and winter, 1977. Two additional follow-up requests
were made to those agencies who had not responded to previous
requests. By the spring of 1978 all states had responded.
Several states indicated in their response that they were in the
process of developing regulations pertinent to this area. Further,
a number of states indicated that they were planning to review the
US EPA's forthcoming criteria for land disposal of hazardous chem-
icals for adoption. These criteria were published in the Decem-
ber 18, 1978 Federal Register.
SUMMARY OF RESPONSES
Over three-fourths of the states did not, at the time of
inquiry, have specific regulations or guidelines governing the
land disposal of potentially hazardous wastes containing organic
solvents. A number of states did not, at the time of inquiry,
permit the disposal of hazardous wastes of this type within the
state. Several states also responded that no regulations were
needed within the state. Many states indicated that they ap-
proached land disposal of hazardous waste on a case-by-case basis
but did not provide any guidelines or other information upon
which they based their decision about the suitability of a dis-
posal site or wastes permitted to be disposed. Many states
13U
-------
indicated that they required monitoring of disposal sites but
did not specify the characteristics of the monitoring program
that should be followed.
Of particular pertinence to this study, is that only 11
states specified a maximum permeability for clay liners or the
soils in the region of the disposal area. The specified values
ranged from 10 ^ cm/sec for soils in their natural state (Michi-
gan) to California's and Oklahoma's values of 10 ~8 en/sec.
Most frequently a maximum permeability of 10 "' cm/sec was
specified. None of the states specified the procedures that were
to be followed in determining the coefficient of permeability.
Texas was the only state that specified that some of the waste
material should be tested, however no specific guidance was given
or. how this should be done.
A wide variery of soil types and liner thickness were being
permitted fcr use in disposal areas. There seemed to be little
or no concensus about appropriate soil types or thickness of
soil liners.
From an overall point of view, it appears from the results
of this survey that most states have not developed specific
criteria for such parameters as soil type, liner permeability
and thickness, and methods of measurement of permeability as
they apply to hazardous waste disposal site selection. The US
EPA in the December 18, 1978 Federal Register proposed a soil
permeability of not less than 10 "7 cm/sec.They do not, how-
ever, specify the procedure for determination of permeability.
This is a significant omission in that many of the standard
permeability tests that are used can not be performed satis-
factorily in the presence of organic solvents. It is evident
from this survey that there is an immediate need for evaluating
the significance of organic solvents in affecting the ability of
industrial waste disposal pits to retain hazardous wastes.
Since the initiation of this study, there has been a
general public and legislative awakening to the potential and
real hazards of land disposal of some types of industrial wastes.
This increased awareness has further increased the need for work
in this area.
A synopsis of the information obtained from the states as
well as a synopsis of portions of the December 18, 1978 Federal
Register pertinent to this project have been submitted to the
US EPA-Ada as a separate set of Appendices to this report. Copies
are available from the Robert S. Kerr Environmental Research Labo-
ratory upon request.
135
-------
REFERENCES
AIBS, Report 68-01-2157, Analysis of Specialized Pesticide Prob-
lems, Vol. II (197U).
ASTM, Special Procedures for Testing Soil and Rock for Engineer-
ing Purposes, ASTM, Philadelphia. PA. p. 101-103 (1970).
Alexander, M. Soil Microbiology. John Wiley and Sons, Inc.,
New York. Chapter 11.(1961).
Allison, L. E. Effect of Microorganisms on Permeability of Soil
Under Prolonged Submergence. Soil Science. 6_3_:139 (1947).
Aly, 0. M. and S. D. Faust. Studies on the Fate of 2,1-D and
Its Derivatives in Natural Surface Waters. Agricultural and
Food Chemistry. No. 6 r2:512 (1961).
Aruer, A. M. and A. A. Awad. Permeability of Cohesionless Soils,
Proceedings of the American Society of Civil Engineers, Geo-
technical Engineering Division, 100, GT 12 (1971).
American Public Health Association (APHA), American Water Wcrks
Association (AWWA), and Water Pollution Control Federation
(W?CF), Standard Methods for the Examination of Water and
Wastewater, llth edition. APHA, Washington, D.C. (1976).
Barshad, I. Factors Affecting the Interlayer Expansion of
Verraiculite and Montmorillonite with Organic Substances.
Soil Science Society Proceedings, p. 176 (1952).
Baughman, G. L. and R. R. Lassiter. Prediction of Environmental
Pollutant Concentration, In: Cairns, J., Dickson, K., and
Maki, t\. (eds), Estimating the Hazard of Chemical Substances
to Aquatic Life,American Society for Testing and Materials
5peciar~Fublication STP 657, ASTM, Philadelphia, PA (1978).
Biscayne, P. E. Minerology and Sedimentation of Recent Deep
Sea Clay in the Atlanf'r Ocean and Adjacent Seas and Oceans,
Geol. Soc. of Amer. Bull. 7J1:303-829 (1965).
Black, C. A. Methods of Soil Analysis. Part I, American
Society of Agronomy, Inc.(1965).
136
-------
Eodman, C». B. The Variability of The Permeability "Constant"
at Low Hydraulic Gradients During Saturated Water Flow in
Soils. Soil Science Society America Proceedings, <2.:U5 (1937),
Bohun, R. L. and W. F. Claussen. The Solubility of Aromatic
Hydrocarbon in Water. J. Ar.er. Chem. Soc. 7jhl571 (1951).
Boucher, F. R. and G. F. Lee. Adsorption of Lindane and
Dieldrin Pesticides on Unconsolidated Aquifer Sands. Env.
Sci. £ Tech., 6_:53? (1972).
Briggs, 3. G. A Simple Relationship Between Soil Adsorption of
Organic Chemicals and Their Octanol-Water Partition Co-
efficients. Proceedings of the British Insecticide and
Fungicide Conference (1973).
Burmister, D. M. Principles of Permeability Testing of Soils,
Symposium on Permeability of Soils. ASTM Special Technical
Publication No. 163 (1954).
Busenberg, E. and C. V. Clemency. Determinat-en of the Cation
Exchange Capacity of Clays and Soils Using an Ammonia
Electrode. Clays and Clay Minerals, Pergamon Press, Great
Britain, 2.1:214-217 (1973).
Chiou, C. T., V. H. Freed, D. W. Schmedding, and R. L. Kohnert.
Partition Coefficient and Bioaccumulation of Selected
Organic Chemicals. Env. Sci. 6 Toch., n_:475 (1977).
Darcy, H. Les fontaines publiques de la ville de Dijon.
Paris: Dijon, (as cited in Taylor, 1948) (1856).
Davis, S. N. and R. J. M. DeWiest. Hydrogeology. John Wiley
and Sons, Inc., New York (1966).
Delfino, J. J. and D. J. Dube. Persistent Contamination of
Ground Water by Phenol. Journal Environmental Science
Health, 6_:3U5-355 (1976).
Dunlap, W. J., R. L. Cosby, J. F. McNabb, B. E. Bledsoe, and
M. R. Scalf. Probable Impact of NTA on Ground Water.
Proceedings of the National Ground Water Quality Symposium,
Denver, CO (1971).
Fuller, W. H. Movement of Selected Metals, Asbestos and
Cyanide in Soil: Applications to Waste Disposal Problems.
EPA-600/2-77-020, US EPA, Cincinnati, OH (1977).
Gibb, J. P. Field Verification of Hazardous Industrial Waste
Migration from Land Disposal Sites. Residual Management by
Land Disposal, Proceedings of the Hazardous Waste Research
Symposium, EPA-600/9-76-015, pp. 94-101 (1976).
137
-------
Glass-tone, S. Textbook of Physical Chemistry. 2nd Edition.
Van Nostrand, New York, pp. 1196-1216 (195*5).
Golwer, A., G. Matthess, and W. Schntider. Effects of Waste
Deposits on Grour.dwater Quality. Groundwater Pollution -
Symposium - Pollution des Eaus Souterraines, Proceedings of
the Moscow Symposium, August (1971).
Green, R. E. and J. E, Corley. Calculation of Hyuraulic Con-
ductivity: A Further Evaluation of Some Predictive Methods,
Proceedings of Soil Science Society of America, ^I'^-S (1971).
Green, W. J. and H. S. Frank. The State of Dissolved Benzene in
Aqueous Solution. Journ. Solution Chemistry, _8:187 (1579).
Greene-Kelly, R. Sorption of Aromatic Compounds by Montraorillo-
nite, Part I - Orientation Studies. Trans. Farad. Soc. 51:
U12 (1955).
Grim, R. E. Clay Mineralogy. McGraw-Hill, Inc., New York (1953).
Hallberg, R. 0. and R. Martinell. Vyredox - In Situ Purification
of Ground Water. Ground Water, Journal of the Technical
Division of NWWA, 1U No. 2, March-April (1976).
Hamaker, J. W. Adsorption. In: Goring, C. A. and Hamaker,
J. W. (eds), Organic Chemicals in the Soil Environment, Vol.
I, pp. U9-143, Marcel Dekker, Inc., New York (1972).
Hansch, C. The Luck of the Draw. J. Chem. Ed., !5_1:360 (197U).
Hansch, C. and D. Elkins. Partition Coefficients and Their
Uses. Chemical Reviews, 72.: 525 (1971).
Hansch, C., J. E. Quinlin, and G. L. Lawrence. The Linear Free
Energy Relationship Between Partition Coefficients and the
Aqueous Solubility of Organic Liquids. J. Organic Chem.,
3j3:3«»7 (1968).
Hashimoto, I., K. B. Deshpande, and H. C. Thomas. Peclet Numbers
and Retardation Factors for Ion Exchange Columns. Ind. Eng.
Chem. Fund., 3_:213 (196
-------
Hoffman, R. W. and G. W. Brindley. Adsorption of Non-Ionic
Aliphatic Molecules from Aqueous Solutions on Montmoril-
lonite. Clay Organic Studies - II. Geochim. et Cosmochim.
Acta. .20:15 (1960).
Holt?, W. G. and H. J. Gibbs. "Engineering Properties of Expan-
sive Clays," Trans. ASCE 120 (1356).
Heule, H. J., R. E. Bell, D. E. Long, and J. E. Soyland. Indus-
trial Hazardous Waste Migration Potential. Proceedings of the
Hazardous Waste Research Symposium, EPA-600/9-76-015, pp 76-85,
July (1376),
Huang, J. C. Effect of Selected Factors on Pesticide Sorption
and Desorption in the Aquatic System. Jour.. Wat. Poll. Cont.
Fed. u_3:1739 O971).
Huggenberger, F., J. Letey, and W. J. Farmer. Adsorption and
Mobility of Pesticides in Soil, California Agriculture, Feb.
(1973).
Jeffrey, E. A. C.-oundwater Use in Small Water Supplies, Ground
Water, Vol. 1T_ (197U).
Johnson, A. I. Application of Labor-.tory Permeability Data,
United States D«partment of Interior, Geological Survey, Water
Resources Division, Denver, CO (1963).
Karickhoff, S. W., D. S. Erown, and T. A. Scott. Sorption of
Hydrophobic Pollutants on Natural Sediments, Water Research
13_:2U1-2U8 (1979).
Kenega, E. G.-and C. Goring. Relationship between Water Solu-
bility, Soil Sorption, Octanol-Water Partitioning, and Bio-
concentration of Chemicals in Biota, Third Aquatic Toxicology
Symposium ASTH, New Orleans (October 1978).
Kezdi, A. Handbook of Soil Mechanics, Elsevier Scientific Publish-
ing Co., frew York, !_ (197i4).
Kimmel, G. E. and 0. C. Bairds. Leach-ite Plumes in a Highly
Permeable Aquifier, Proceedings of the Second National Ground
Water Quality Symposium, National Water Well Association,
Denver, CO (1974).
Law, Jr., J. P. and G. W. Kunze. Reaction of Surfactants with
Montmorillonite: Adsorption Mechanisms, Soil Science Society
of America Proceedings, 30 (1956).
139
-------
Lee, G. F. and R. A. Jones. Research Needs in the Area of the
Role of Sorption-Desorption in Predicting the Environmental
Chemistry-Fats of Chemical Contaminants in a Hazard Assess-
ment Program, Environmental Engineering, Colorado State
University, Fort Collins, CO, Occasional Paper No. 37
(December 1978).
Lewallen, M. J. Pesticide Contamination of a Shallow Bored Well
in the Southeastern Coastal Plains, Proceedings of the
National Groundwater Quality Symposium, Denver, CO (1971).
Liskowitz, J. W., P. C. Chan, R. B. Trattner, R. Dresnack, A. J.
Perna, M. J. Sheih, R. Traver, and F. Etterbusch. Evaluation
of Selected Sorbents for the Removal of Contaminants in Leach-
ate from Industrial Sludges. In: Residual Management by Land
Disposal, Proceedings of the Hazardous Waste Research Symposium,
EPA-600/9-76-015, July (1976).
McAfee, J. L., Jr. Inorganic-Organic Cation Exchange on
Montmorillonite, American Mineralogist, 4^:1230-1236 (1959).
McEwan, D. C. M. Complexes of Clays with Organic Compounds - I.
Complexes between Montmorillonite and Halloysite and Certain
Organic Liquids, Trans. Farad. Soc. 4_4:349-367 (1948).
Marketos, D. G. Direct Ultraviolet Determination of Benzene in
Aqueous Solution. Analytical Chemistry, 4^195 (1969).
Marshall, T. J. A Relation Between Permeability and Size Dis-
tribution of Pores. J. Soil Sci., 9:1-8 (1958). (Cited in
Green and Cor ley, 1971). ~~
Mason, B. Principles of Geochemistry, 2nd Ed., John Wiley and
Sons, New York, p. 152 (1958).
Millington, R. J. and R. J. Quirk. Transport in Porous Media.
Int. Congr. Soil Sci., Trans. 7th (Madison, Wis.) 1.3:97-106
(1960). (Cited in Green and CorJey, 1971).
Mitchell, J. K. Fundamentals of Soil Behavior. John Wiley and
Sons, New York, (1976~T!
Mortland, M. M. Clay-Organic Complexes and Interactions. Advan.
Agron. 2£:75 (1970).
Naney, J. W., D. C. Kent, and E. H. Seely. Evaluating Ground-
water Paths Using Hydraulic Conductivities. Ground Water,
14.:4, 205-213 (1976).
Norrish, K. The Swelling of Montmorillonite. Trans. Farad.
Soc. 18:120-134 (1954).
140
-------
3dell, R. T., T. H. Thornburn, and L. J. McKenzie. Relation-
ships of Atterberg Limits to Some Other Properties of Illinois
Soils. Proc. Soil Sci. Soc. Ame^., 2^:297-300 (1960).
Olejnik, S., A. M. Posner, and J. P. Quink. Swelling of
Montmcrillonite in Polar Organic Liquids. Clays and Clay
Minerals, 21:361 (1974).
Osgood, J. 0. Hydrocarbon Dispersion in Ground Water:
Significance and Characteristics. Proceedings of the Second
National Ground Water Quality Symposium, Denver, CO (197U).
Pettyjohn, W. A. Pickling Liquors, Strip Mines, and Ground-
Water Pollution. Proceedings of the Second National Ground
Water Quality Symposium, Denver, CO (197U).
Pierce, G. W. and F. R. Siegel. Quantification and Clay Mineral
Studies of Sediments and Sedimentary Rocks. J. Sedimentary
Petrology, 3^:187-193 (1969).
Poulovassilis, A. The Changeability of the Hydraulic Conductivity
of Saturated Soil Samples. Soil Science, 113:81 (1972).
iP.iddick, J. A. and W. B. Bunger. Organic Solvents. Wiley-
'" Interscience, New York (1970).
i
Banks, R. L., J. M. LaPiante, and E. F. Gloyna. Survey-Suit-
ability of Clay Beds for Storage of Industrial Wastes.
Publication by Center for Research in Water Resources, En-
vironmental Health Engineering, University of Texas-Austin
(1975).
Seed, H. B., R. J. Woodward, and R. Lundgren. Fundamental As-
pects of the Atterberg Limits. Journal of Soil Mechanic and
Foundations Division, A.S.C.E., ££:75-105 (196U).
Snyder, H. J., G. B. Rice, and J. J. Skujins. Disposal of
Waste Oil Refining Residues by Land Farming. In: "Residual
Management by Land Disposal," Proceedings of the Hazardous
Waste Research Symposium, EPA-600/9-76-015, July (1976).
Soiltest, Inc., Soiltesting Equipment. Evanston, IL (1976).
Streng, D. R. The Effects of the Disposal of Industrial Waste
Within a Sanitary Landfill Environment. Proceedings of the
Hazardous Waste Research Symposium, EPA-600/9-76-015, pp.
51-70 (1976).
Taylor, Donald W. Fundamantals of Soil Mechanics. John Wiley
and Sons, Inc., ifew'Vork (1948).
-------
Terzaghi, K. Simplified Soil Tests for Subprades and Their
Physical Significance. Public Roads, October (1926).
(Cited in Mitchell, J. K. (1976), Fundamentals of Soil
Behavior, John Wiley and Sons, New~Vork, o. 177)..
Terzaghi, K. and R. 9. Peck. Soil Mechanics in Engineering
Practice. John Wiley and Sons. Inc.(1948).
Terzaghi, K. and R. B. Peck. Soil Mechanics in Engineering
Practice. John Wiley and Sons,Inc., New York, 2nd Ed.(1967),
Theng, B. K. G. The Chemistry of Clay Organic Reactions. John
Wiley and Sons, Inc., New York (1974).
Todd, D. K. Groundwater Hydrology. John Wiley and Sons, Inc.,
New York (1959).
van Genuchten, M. T. and P. J. Wier.anga. Mass Transfer Studies
in Sorbing Porous Media: I - Analytical Solutions. J. Soil
Science Society of America, ^0:473 (1976).
Van Olphen, H. Clay Colloid Chemistry, Interscience, New York,
pp. 148-149 (1963).
Walker, G. F. Vermiculite Minerals. In: The X-Ray Identifi-
cation and Crystal Structures of Clay Minerals. G. Brown
(ed.). Mineral Society, London, pp. 297-32V (1961).
Walker, W. H. Where Have all the Toxic Chemicals Gene?
Journal Ground Water, 1_1:2, March-April, (1973).
Walton, C. W. Groundwater Resource Evaluation. McGraw-Hill,
New York (197571
Washburn, E. W. (ed.) International Critical Tables of
Numerical Data. VolTTIMcGraw Hill, New York, p., 101
(1923).
Wayraan, C. H., J. B. Robertson, and H. G. Page. Adsorption of
the Surfactant ABS 35 on Montraorillonite. Art. 59 in U.S.
Geological Survey Prof. Paper 475-B (1963).
Wiles, C. C. and H. R. Lubowitz. A Polymeric Cementing and
Encapsulating Process for Managing Hazardous Waste, int
"Residual Management by Land Disposal," Proceedings of the
Hazardous Waste Research Symposium, EPA-600/9-76-015, July
(1976).
Williams, D. E. and D. G. Wilder. Gasoline Pollution -f a
Ground-Water Reservoir - A Case History. Proceedings of the
National Ground Water Quality Symposium, Denver, CO (1971).
142
-------
Yor.g, R. N. and B. P. Warentin. Soil Properties e.-.d Behavior.
Elsevier Scientific Publishing Company, New York, p. 62
(1975).
Zanker, A. Nomograph fcr Hyjraulic Conductivity and Intrinsic
Permeability of Water-Soil Svstem. Soil Science, 113;5
(1972).
-------
Appendices are available from the
authors upon request.
Impact of organic solvents on the integrity of clay
liners for industrial waste disposal pits
implications for groundwater contamination
OC: 16517857
-------
-------
o .<
o • a> 5
•a-g-d S
Reproduced by NTIS
National Technical Information Service
U.S. Department of Commerce
Springfield, VA 22161
This report was printed specifically for your
order from our collection of more than 2 million
technical reports.
For economy and efficiency, NTIS does not maintain stock of its vast
collection of technical reports. Rather, most documents are printed for
each order. Your copy is the best possible reproduction available from
our master archive. If you have any questions concerning this document
or any order you placed with NTIS, please call our Customer Services
Department at (703)487-4660.
Always think of NTIS when you want:
• Access to the technical, scientific, and engineering results generated
by the ongoing multibillion dollar R&D program of the U.S. Government.
• R&D results from Japan, West Germany, Great Britain, and some 20
other countries, most of it reported in English.
NTIS also operates two centers that can provide you with valuable
information:
• The Federal Computer Products Center - offers software and
datafiles produced by Federal agencies.
• The Center for the Utilization of Federal Technology - gives you
access to the best of Federal technologies and laboratory resources.
For more information about NTIS, send for our FREE NTIS Products
and Services Catalog which describes how you can access this U.S. and
foreign Government technology. Call (703)487-4650 or send this
sheet to NTIS, U.S. Department of Commerce, Springfield, VA 22161.
Ask for catalog, PR-827.
Name
Telephone
- Your Source to U.S. and Foreign Government
Research and Technology.
-------
|