United States
                    Environmental Protection
                    Agency
Robert S. Kerr Environmental
Research Laboratory
Ada, OK  74820
                    Research and Development
EPA/600/S2-90/004 Apr. 1990
&EPA          Project Summary
                    Adsorption  of  Organic Cations  to
                    Soils  and  Subsurface Materials
                    John C. Westall, Bruce J. Brownawell, Hua Chen, John M. Collier,
                    and Julia Hatfield
                      A study  of the fundamentals  of
                   adsorption  of  amphiphilic organic
                   cations on  natural and  pristine sur-
                   faces was conducted to elucidate  (i)
                   the factors  that influence the extent
                   of adsorption and (ii) indirect effects
                   of adsorption of organic cations: the
                   competitive adsorption  of organic
                   cations  and  metal cations  and
                   increased adsorption of hydrophobic
                   organic  compounds  on  mineral
                   surfaces coated with organic cations.
                      The sorbents studied were kaolin-
                   ite, montmorillonite, silica, alumina, a
                   soil, and two aquifer materials. The
                   distribution ratio of  the alkylpyrid-
                   iriium depended  strongly on the
                   nature and  concentration of inorgan-
                   ic: cations in solution, but solution pH
                   had little effect The  adsorption iso-
                   therms  were distinctly nonlinear,
                   even  at very low surface concen-
                   trations  of  organic cations. The ad-
                   sorption of  dodecylpyridinium in the
                   presence of various  concentrations
                   of salts  was described quantitatively
                   with a multisite  competitive  ion-
                   exchange  model;  two  adsorption
                   reactions were significant: exchange
                   of pyridinium with a metal cation, and
                   adsorption  of pyridinium with chlor-
                   ide counter-ion.
                      The indirect effects of adsorption
                   of organic cations were investigated
                   as well. Dodecylpyridinium was found
                   to displace Cu(ll) effectively from
                   Lula  aquifer  material when the
                   surface  concentration of dodecylpy-
                   ridinium is  within  10 -  100% of the
                   adsorption  maximum.  Dodecylpy-
                   ridinium  on Lula  aquifer  material
                   increased the sorption of chloro-
                   benzenes; organic-carbon normalized
                   distribution  ratios of chlorobenzenes
(Koc's) were greater for  DP-loaded
sorbents than  for natural sorbents,
and the values of K^ were greater for
greater amounts  of dodecylpyrid-
inium on the surface.
    This Project  Summary  was
developed by ERA'S Robert S.  Kerr
Environmental Research Laboratory,
Ada, OK, to announce key findings of
the research project that is  fully
documented in  a separate report of
the same title  (see Project Report
ordering information at back).


Introduction
   Amphiphilic organic cations  are used
in a  wide  range of  consumer  and
industrial products such as detergents,
fabric softeners, dyes, herbicides, emulsi-
fying agents, and additives to drilling
muds.  Adsorption of organic  cations
influences their bioavailability, transport,
and fate. Adsorption can have important
indirect effects as well: organic cations
can effectively displace adsorbed metal
ions, and adsorbed  organic cations can
increase the affinity  of  sorbents for
hydrophobic  compounds.  Despite the
importance of these  compounds, there
has not been extensive systematic inves-
tigation of their  adsorption to natural
materials.
   Direct effects. Organic  cations
generally adsorb strongly due to the
electrostatic  attraction to negatively
charged surfaces and  the favorable
hydrophobic interaction. In the environ-
ment they  are likely to accumulate on
surfaces until they are degraded, for
example, in sediments downstream from
a  wastewater discharge.  The environ-
mental concentrations of organic cations
have not been surveyed as intensively as
            /7TV
            ^gA) Printed on Recycled Paper

-------
the  concentrations of  many  other
compounds. This situation is largely due
to difficulties in separation and analysis of
organic cations in natural water samples.
   Organic cations  appear  to  be  less
toxic  to  aquatic organisms  in  systems
containing suspended particles,  presum-
ably due to binding of the organic cations
(Lewis and Wee,  1983).  One of the
primary  problems  in  assessing  the
environmental safety of organic cations is
the characterization  of their physical and
chemical form  (speciation) in test  sys-
tems.  If meaningful  interpretation of
laboratory toxicity experiments  is to be
made, the  speciation of the compounds
in solution must be known.
   Indirect  effects.  In addition  to these
direct  effects of  organic cations  in the
environment, the indirect effects  must be
considered. While these effects will occur
naturally  as  organic  cations  are
introduced  into  the  environment, they
might  also be  the  basis for remedial
action  at sites already contaminated with
metals or  hydrophobic  organic com-
pounds.  The  observation that adsorbed
organic cations  increase  the  tendency of
mineral surfaces to sorb  hydrophobic
organic  compounds  has  led  to
consideration of applying organic cations
in the subsurface environment (e.g.,  clay
liners)  to slow  the  migration of  organic
compounds. However, such  a treatment
might  also lead  to desorption  and
increased mobility of toxic metal  ions. On
the other hand,  if  a pump and treat
process were being considered, such a
displacement  of metals  with nontoxic,
biodegradable organic cations would be
beneficial. This  study   addresses
fundamental questions that  need to be
resolved  before  these applications are
pursued.
   Scope of study. The objectives of this
study were: (i) to characterize the factors
that  influence  the  adsorption  of
amphiphilic organic  cations  to  natural
sorbents; (ii) to  describe  the observed
adsorption isotherms in terms of mechan-
istically based adsorption models, which
can be used for understanding  environ-
mental distributions;  and (iii) to assess
the effect of organic  cation adsorption on
the adsorption of metals and hydrophobic
organic compounds.
   The organic  cations selected  for this
study  were alkylpyridinium  cations;
selection  criteria  include environmental
relevance,  representativeness  of  the
class, and ease  of analysis. The  sorbents
were  aquifer materials,  a  soil,  kaolinite
and  montmorillonite,  and  silica  and
alumina.  Efforts  were concentrated  on
well  characterized,  low-organic-carbon
sorbents. Cu(ll) and chlorobenzenes were
chosen as  representative  metal  and
hydrophobic organic  compounds  for
examination of indirect effects.
Methods
   Materials 14C-labeled dodecylpyridin-
ium  bromide (DP)  was obtained  from
Pathfinder  Laboratories  Inc. (St.  Louis,
MO). Natural sorbents were low- organic-
carbon (< 0.03%  organic carbon) aquifer
materials Borden sand and Lula N6, and
a  relatively  high-organic-carbon  (2.3%)
soil,  EPA-12. Georgia kaolin (KGa-1) and
Wyoming  Na-montmorillonite (SWy-1)
were obtained  from  Source   Clays
Repository  at the  University of Missouri.
Berkeley 5 nm  silica and  10 tun MCA-
540  alumina  were obtained   from
Pennsylvania Glass Sand Corporation
and  Alcoa  Laboratories, respectively.  A
74 - 210 tim fraction of  Borden sand,
isolated by dry-sieving, was used  in this
work. The  organic carbon  contents and
cation  exchange  capacities  (CEC) of
these materials are given in Table  1.
   Equipment.  For  equilibration  and
phase separation,  a thermostatted shaker
(New  Brunswick Model G-24)  and
thermostatted centrifuge  (Beckman  Mod-
el  J2-21) were  used. Concentrations of
unlabeled pyridinium  compounds  were
determined spectrophotometrically  with
an HP 8452A UV-Visible  spectrophoto-
meter.  Radioactivity  of the labeled  com-
pounds was determined  with a Beckman
Model  LS  7800  scintillation  counter.
Labeled compounds  on sorbents  were
converted to 14C02 with  a  Harvey  Model
300 Oxidizer.
   Determination of distributions.    Dis-
tributions were  determined from  batch
equilibration experiments.  The  electro-
lytes were NaCI, CaCI2, and NaN3. NaN3
was used in place  of NaCI in experiments
conducted at low  concentration of DP to
inhibit  microbial degradation.  The  con-
tents of the tubes,  which were sealed with
PTFE-lined caps, were mixed on the
shaker  at a speed of 500  RPM at 25°C
for 4 hours, after which the aqueous and
sorbent  phases  were  separated  by
centrifugation at 11,000 G at 25°C for 30
-  60  minutes. The time  for equilibration
was  established  in  independent  experi-
ments  on  adsorption  and desorption
kinetics.
   After centrifugation,  the  amount of
alkylpyridinium in the aqueous phase, on
the sorbent, and adsorbed  to the walls of
the vessel was determined. The concen-
tration  distribution ratio,  Dc,  was
calculated from:
                                   (D
              C.(s)  (nmol/kg)
          D  = —	
           c   C.(w)    (pM)
where  Cj(s)  and  Cj(w)   were  the
concentrations  of the  component  i
determined directly on the sorbent and in
the solution,  respectively. The recovery,
or fraction of alkylpyridinium added to the
tube that could  be accounted for by
analysis, was also calculated  from these
data.
   Adsorption of Cu(ll).  Two  procedures
were  developed  for  investigating  the
effect  of dodecylpyridinium  on  the
adsorption of  Cu(ll)  on Lula  aquifer
material: (i) conventional batch adsorption
experiments with Cu(ll), DP, and Lula N6,
in which concentrations of Cu(ll) and DP,
in solution  and  on the sorbent,  are
determined by our standard  procedures;
and (ii) discrete additions of DP to Cu(ll)
and Lula  N6 in a pH-stat, in which the
change  in   Cu(ll)  concentration  is
monitored with  a cupric ion-selective
electrode. The advantage of  the  first
procedure is that the concentration of DP
in solution   and  on  the  surface  is
determined  simultaneously  with  the
concentration of Cu(ll);  the disadvantage
of this  procedure is  that  there is no
control over  pH during the time of batch
equilibration.  While the  adsorption of DP
is relatively  independent  of  pH,  the
adsorption of Cu(ll) depends strongly on
pH. In the second procedure,  the pH can
be controlled extremely well (± 0.03 pH
units),  but the  amount of Cu(ll) on the
sorbent  must  be  determined  by
difference rather than by  direct  analysis,
and the distribution of DP is found from a
separate set  of experiments.
   Adsorption of  chlorobenzenes.  The
adsorption of chlorobenzenes  to Lula N6
aquifer  material,  with  varying  surface
concentrations of  DP, was determined  in
conventional  batch experiments.  The
chlorobenzenes were  1,2-dichloroben-
zene,   1,2,3-trichlorobenzene,  1,2,3,4-
tetrachlorobenzene,  and  pentachloroben-
zene.  Slurries were   prepared   that
contained  25 g L-1 of  Lula N6,  total DP
concentrations between  60 nm and 40
mM, and  the chlorobenzenes at total
concentrations  approximately 1/100 of
their solubility.  The  experiments  were
carried out in  centrifuge tubes  with PTFE-
lined caps. Chlorobenzene concentrations
in the  aqueous phase were  determined
by  extraction  into  hexane and  gas
chromatography with  electron capture
detection, and amount sorbed   was
determined by difference.

-------
 Table 1.  The Cation Exchange Capacities and Organic Carbon Contents of the Sorbents
         Used in this Study. Organic Carbon Contents were Determined with a LECO
         Carbon Analyzer. Procedures for Cation Exchange Capacity Determination are
         Given in the References Cited.
Sorbent material
Kaolinite (KGa-1)
Na-Montmorillonite (SWy-1)
Borden sand
Lula N6 aquifer material
EPA- 12 soil
Organic carbon
%
0.022
0.020
0.026, 0.0200
0.02 r, 0.033b
2.;'2, 2.33d
Cation exchange
capacity
meq/kg
20"
764"
7.0C, 23b
90"
735<*
a van Olphen, H.; Fripiat, J.J. Eds. Data Handbook tor Clay Materials and  Other Non-Metallic
 Minerals; Pergamon: New York, 1979.
b Bouchard,  D.C.; Powell, R.M.; Clark, D.A. J. Environ. Sci. and Health  1988, A23, 585-601.
c Fuller, C.C.; Davis, J.A. Geochim. Cosmochim. Acta 1987, 51, 1491-1502.
d Hassett, J.J.; Means, J.C.; Banwart, W.L; Wood, S.G. Sorption Properties of Sediments
 and Energy-Related Pollutants; U.S.Environmental Protection Agency. National Technical
 Information Service: Springfield, VA., 1980; EPA-600/3-80-041.
Results and Discussion

   Two  factors  complicate  this study.
F:irst,  the nonlinearity  of  the isotherms
precludes a simple interpretation of the
effects of given properties on the energy
of adsorption. We have ascribed much of
the nonlinearity  to  the  heterogeneity of
adsorption sites of the sorbents, and have
used  a  distribution of  adsorption-site
energies  in  describing  experimental
results.  Second, it is  difficult, if not
impossible, when studying aqueous sus-
pensions  of solid sorbents  to  vary one
experimental  parameter and  maintain all
others at  constant values. For example,
variation  in the solution concentration of
one  component  causes  a  significant
variation  in the solution concentration of
another component, through interaction
with the sorbent.
   Adsorption  isotherms.   Adsorption
isotherms of  dodecylpyridinium (DP) at
relatively high concentrations, on several
different sorbents, are shown  in Figure 1.
The  equilibrium  concentrations  of
dodecylpyridinium in  solution were  well
below  the  critical micelle  concentration
(CMC)  in all  cases.   The  surface
concentrations approach or reach plateau
levels, except in the case of EPA-12. The
plateau  concentrations correspond
closely, but not  strictly,  to  the cation
exchange capacities reported for the five
sorbents (Table  1).  From these  data,  it
appears  that adsorption  could be
explained as  a  simple,  virtually quan-
titative, ion-exchange reaction.  However,
if the data are considered over a wider
range of  concentrations, as shown in
Figure 2,  it is apparent that  the situation
is more complicated.
   The logarithmic  adsorption  isotherms
of  DP on  five different materials (Borden
sand, Lula N6 aquifer material, EPA-12
soil, kaolin,  and  montmorillonite)  are
shown in  Figure 2. The  affinities of the
sorbents  for  dodecylpyridinium, per  unit
mass of  sorbent,  follows the  order
montmorillonite  >  EPA-12  soil *•  Lula
aquifer material >  kaolinite > Borden
sand. This order corresponds roughly to
the CEC's of the materials.  A  judgment
about the presence  or  absence  of  a
correlation with organic carbon content of
the materials  is hardly warranted, since
the organic carbon content  of  most of
these materials is so very low.  However,
the adsorption affinity  of Lula aquifer
material  is  similar to that  of EPA-12
despite the 100-fold difference in organic
carbon contents of these sorbents.
   At  concentrations of  DP  in  solution
below 125 nM for  EPA-12  and Borden
sand, and below 20 pM  for  Lula aquifer
material,  the  logarithmic  isotherms
(Figure 2) are nearly linear and conform
to  the logarithmic form of the Freundlich
equation:

        log Cs = n log Cw + log K       (2)
where K and  n are constants. The values
of  n determined from the slopes of the
isotherms  in the low concentration region
are approximately 0.6. These  low values
of n indicate nonlinearity in the isotherms.
(A linear  isotherm  is one  in  which  the
surface concentration is  directly  propor-
tional to the solution concentration, i.e., n
= 1  in  Equation 2.) The  nonlinearity
extends  to  surface  concentrations  that
are as low as 0.02  % of the maximum
surface coverage measured in the case
of EPA-12 soil. As will be discussed, this
nonlinearity  is attributed to  heterogeneity
of adsorption sites.
   The effect of pH on adsorption of DP.
The  effect of  [H*] on the  adsorption of
DP by EPA-12, Lula aquifer material, and
kaolin is shown in Figure 3. The logarithm
of Dc is plotted against  log [H*]. A very
small  dependence  of  Dc on pH was
observed  for  the  natural sorbents;  the
slopes (A log Dc / A log [H*]) determined
from  linear regression  of the  data in
Figure 3  were -0.04 and -0.01  for Lula
and  EPA-12,  respectively. The  depen-
dence of  log Dc on  pH for kaolinite was
greater than for  the other two materials;
at pH values < 6.06 the slope A log Dc /
A log [H + ] was -0.23. The effects of [H + ]
on  adsorption of  hexadecylpyridinium
(HP)  by  the two  oxides,  silica  and
alumina  (not shown), were also  greater
than that shown for Lula and EPA-12.
   Two mechanisms through  which  pH
can influence adsorption are (i) a change
in surface  potential from protonation-
deprotonation  of  pH-dependent surface
functional groups, or (ii) ion-exchange of
H* for pyridinium. The weak effect of pH
on the two heterogeneous  sorbents, in
contrast to the clay and oxides,  can be
ascribed  to the  relative  abundances of
pH-dependent  surface groups and
traditional  ion-exchange  sites on   the
sorbents.
   Nonlinearity of the isotherms.  All of
the adsorption isotherms determined in
this study were  nonlinear,  even  at very
low surface coverages  of dodecylpyrid-
inium.  While the Freundlich isotherm can
be used  to represent  these  data
empirically,  it provides  no mechanistic
basis for  evaluating competition  effects,
saturation  effects,  and  comparison of
adsorption energies for  a homologous
series of organic  cations.  Thus   the
usefulness of  the Freundlich  isotherm is
limited.
   In  order to derive  a more  useful
description  of adsorption  data,  it is
necessary to consider  the   causes of
nonlinearity: (i) non-uniformity of energies
of adsorption  sites,  with saturation of
high-energy  sites;  (ii)  sorbate-sorbate
interactions  such  as  repulsive elec-
trostatic interactions or cooperative chain-

-------
         200
       o
       1  700
       CO
      o
                                                        •  EPA-12

                                                        m  Lula N6

                                                        A  Kaolinite
                                     C(w)
Figure 1. Adsorption isotherms; dodecylpyridinium in 0.01 M Na f.
    s
    O)
        -r
        -2
        -3
        -4
  O Montmorillonite

  • EPA-12
  m Lula N6
  A Kaolinite
- A Borden Sand
                                                             1  o  oo
                                                        A***
         -9
                  -8
                          -7
                                   -6       -5

                                   log C(w) [Ml

Figure 2. Adsorption isotherms; dodecylpyridinium in 001 M Na
                                                    -4
                                                                     -2
chain interactions, and (iii) experimental
artifacts that stem  from covariation of
experimental conditions, such as changes
of pH,  major  ion  concentrations,  or
aggregation  of particles, that accompany
changes in organic cation concentrations.
   For reasons  to be discussed,  we
interpret the nonlinearity  of isotherms in
these experiments primarily as the result
of heterogeneity of sites. Particularly in
the  low-concentration regions  of  the
isotherms, the changes  in pH, concen-
trations  of major ions in solution,  and
surface charge that accompanied adsorp-
tion of the organic cation were too small
to account for the isotherm nonlinearity.
   Thus, at low concentrations of organic
cation,  heterogeneity of sites  is  the
preferred  explanation. At higher concen-
trations it  is  more difficult to exclude the
influence of  other processes on isotherm
nonlinearity.
   The effect  of  electrolyte  on adsorp-
tion.  The effect of concentration and type
of electrolyte on  adsorption of DP  was
examined  with two complementary types
of data: (i)  adsorption isotherms  of DP
from  solutions with  different  concen-
trations  of salts;  and  (ii) concentration
distribution ratio of  DP as a function of
concentration of NaCI or CaCI2,  with  a
constant amount of  DP in the system. In
view of the nonlinearity of the isotherms,
the  first  method  is  preferable  for
development  of  mechanistic  models,
while the second  provides  a  quick
estimate of the magnitude of the effect.
   A  number  of  experiments  were
performed with different sorbents,  and
different  inorganic  salts  at different
concentrations. Here we present data for
one  representative  system.  Figure  4
shows the  isotherms of DP  on  Lula
aquifer material in the presence of 0.01
and  0.1  M NaCI. The effect of the  salt
concentration was  generally  much
greater than  that of pH.
   At low concentrations  of organic
cation,  an  increase   in  the  salt
concentration decreases  the adsorption
by a nearly constant factor. This behavior
is consistent  with   an  ion-exchange
mechanism.  At high concentrations of the
organic cation, the isotherms merge  and
cross at  surface  concentrations  that
corresponds roughly  to  the CEC.  This
behavior  is consistent  with  the  pre-
dominance of  adsorption of pyridinium-
chloride at  high  concentrations of  DP.
Similar  isotherms  (not shown) were
obtained  for EPA-12,  Borden sand,  and
kaolinite.
   The effect of Ca2+  on the adsorption
of DP  on  Lula aquifer  material is

-------
              •  EPA-12

              M  Lula N6

              A  Kaolinite
                                        -7

                                      log [H
                                                                       -3
           generally stronger  than that of Na* (not
           shown), consistent with ion exchange.  A
           solution  of 0.001  M Ca2 *  impedes
           adsorption  of  DP  more effectively than
           0.01  M Na*.
              Adsorption  model. The  mechanistic
           models that were  considered for these
           data  were  (i)  multisite   models,  (ii)
           electrostatic  models, and  (iii)  some
           combination  of the two.  Most  of  the
           phenomena observed in  this study can
           be explained readily through the multisite
           model, with electroneutral reactions and
           no  explicit  electrostatic  energy  term.
           While  this approach  is  obviously  an
           approximation, it  is  consistent  with  a
           variety  of observations,  and it   is
           conceptually  very  simple. The  role   of
           electrostatic energy will be examined in a
           future  study  involving  adsorption and
           electrophoretic mobility. The adsorption
           model that we use  is  based  on  the
           following reactions:

                      ~~                   (3)
                                                                                  R + + NaX = RX + Na
                                      RX
                                                                                  R + + Cl  + Y = RC1Y
                                             (4)
Figure 3. The effect of pH on distribution of a/kylpyndmium
                             Dodecylpyndm/um on Lula

                                      2.5 gL
     O)
     JC
     o   -2
     O
         -3
         -4
         -5
                n 0.07 M NaCl

                • 0. ^  M NaCf
          -9
                   -8
                           -7
                                    -6       -5

                                    log C(w) (Ml
                                                     -4
                                                             -3
-2
Figure 4. Dodecylpyridinium on Lula N6; 2.5 g/L.
                  2NaX = CaX2
                                                                                                            LCaX2
(5)
                                                                                 where  the  overbar  represents a species
                                                                                 associated  with  the  sorbent,  X  is  a
                                                                                 negatively  charged  cation-exchange  site,
                                                                                 Y  is an  uncharged site (that could be
                                                                                 thought of  as a site  for hydrophobic
                                                                                 adsorption),  and  R+  represents  the
                                                                                 alkylpyridinium.  The mass action  equa-
                                                                                 tions for Reactions 3-5  are:
            
-------
mmol kg'
C
Figure 5. Adsorption of Cu(ll) on Lula A/6; effect of DP.
                i	1
                    1,2-Di
           A	A  1,2,3-Tn
          -A	A  1,2,3,4-Tetra
           n—Penta
                           30     40    50     60

                                COp (s)immol kg'1
 Figure 6. The effect of dodecyclpyridinium on adsorption of chlorobenzenes.
 (X1( X2, ... Xn), each of which obeys its
 own mass action equations and material
 balance equations given above. Thus the
 model could be described as a multisite
 competitive Langmuir model. A similar
 ensemble  could be  invoked for  the  Y
 sites. The adjustable parameters  in the
 model  are the  equilibrium  constants
 given  by  Equations  6-8 and  the total
 number of sites, given in Equations 9 and
 10. For  the   X-sites,  the  discrete
 equilibrium-constant-spectrum approach
 was used: the values of KRXj were set to
 integral powers of ten (10°,  101, ...) and
 the total  concentration  of  each type of
 sites was determined, Tx,.  The objective
 of  this  approach is to  represent the
 experimental data in  terms of a regular
 distribution of  sites,  rather than the
 smallest number of  adjustable param-
 eters.  This approach  facilitates  com-
 parison of complex  isotherms obtained
 for different sorbents.
    All parameters were determined simul-
 taneously with  FITEQL 2.0,  a weighted
 nonlinear  least squares  adjustment
 procedure  for  multicomponent chemical
 equilibrium problems.
   Application of model. This model was
 applied first to the adsorption of DP on
 Lula aquifer  material.  In the  first  step,
 data at two  different concentrations of
 Na+ (Figure  4)  were  used to determine
 the TXj of  the equilibrium-constant
 spectrum; since only one type of sites Y
 was considered,  both KY and TY  were
 determined.  The constants are  given in
 Table 2,  and  the isotherms calculated
 from the constants are given by the lines
 in Figure 4. The agreement between the
 model and the data is excellent.
   This sort of model was applied to data
 for  adsorption  of DP on EPA-12  from
 NaCI  solutions,  and  to  DP  adsorption
 onto Lula and  EPA-12  from  CaCI2/NaCI
 solutions. In  all  cases  the  agreement
 between the  models  and the data was
 excellent.
   Certainly this  "discrete  equilibrium-
 constant-spectrum" approach  is subject
 to the criticism that a model can  be fit to
 any set  of  data if enough adjustable
 parameters are  used.  Furthermore, there
 is some covariance among  the values of
 the adjustable parameters; the values are
 not unique,  but just  one of an infinite
 number of sets  that can  be  used  to
 represent the data.
   However,  the  advantages  of  this
approach  are   significant,  too.  The
practical advantage is that the effects of
multiple interactions in complex systems
can be  represented   in  a  way  that is
completely compatible  with  chemical
equilibrium models used in environmental

-------
  Table 2. Equilibrium Constants for Adsorption of Dodecylpyridinium on Lula Aquifer
  Type of site      Equilibrium constants"-*3
               log KRX   log KCaX2  log Kync/
                          Total site density
                             mol kg'1
                 5.00
3.00
2.89 x  TO'4
X2 3.00 0.57
Y 4.95
3.21 x JO'2
7.99 X 70'2
 a Reactions described in text, Reactions 3, 4, 5. Equilibrium constants valid for 25 ± 2 °C,
  with aqueous concentrations in mol L1 and concentrations on sorbent in mol kg-1.
  Properties of sorbents listed in Table 1.
 b Constants were determined by the following procedure: values of Kpx were set according
  to the equilibrium constant spectrum approach, and values of Tx, TY, and KYRCI were
  determined from data in Figure 4a with FITEQL. Then values of K"caX2 were determined
  from data not shown here.
chemistry.  An adsorption  isotherm  that
conforms  closely  to the  Freundlich
equation over more than  six  orders  of
magnitude  range  in  C(w)  can  be
described  very  well by  conventional
mass  action and  material  balance
equations;  these  equations  have  a
mechanistic  basis, which  potentially
could be used to understand the factors
that influence adsorption.
   Another  advantage of such a model is
that  it  provides  a  framework for
evaluating quantitatively mechanisms that
could be responsible for the phenomena
that  are observed experimentally,  such
as, (i) reversal of effect  of  electrolyte
concentration on  isotherms near the CEC;
(ii)  lateral  shift  of  isotherms with  salt
concentration  far below  the CEC; (iii)
absence of pH  effect; and  (iv) relative
effect of Na vs. Ca on the isotherms.
   Adsorption of Cu(ll) on Lula A/6. The
adsorption  isotherms of Cu(ll) on Lula N6
in  0.01  M  NaCI, in the  absence  and
presence of various concentrations of DP,
are shown  in  Figure 5. (The isotherms in
the   figure   are  from   the  batch
experiments;  the isotherms from the pH-
stat experiments were quite similar, but
are not shown.)
   In  the absence of DP, the  apparent
distribution ratio (Dc) of  Cu(ll) varies
between 2000 and 100 L kg-1 for solution
concentrations between  5  and  100 nM.
Over this range of solution  concentrations
the surface concentrations range between
I  and 10  mmol  kg-1, which is still  far
below an  apparent  cation  exchange
capacity   of  the  sorbent.  Solution
concentrations were held  below 200 pM
and pH  below 5.5 in  these experiments to
avoid problems  with  precipitation   of
Cu(OH)2(s).
   Adsorption  isotherms  of  Cu(ll) with
              various total concentrations  of  DP  are
              also  shown in Figure  5.  (The total
              concentration refers  to the total number
              of moles of DP added per liter of solution;
              this  is  not the  same  as  the final
              equilibrium concentration  in solution. The
              approximate relation between total and
              adsorbed DP is given  in  Table 3.) The
              effect of DP is to depress the adsorption
              of Cu(ll) significantly, particularly above
              100 iiM total DP (or 17 mmol kg-1 DP on
              the sorbent)  At 400 and  1000 nM total
              DP, the DP concentrations on the surface
              were  close  to  the apparent "CEC"  for
              adsorption  of  DP;  only under these
              conditions did the moles of DP desorbed
              per mole of Cu(ll) adsorbed approach the
              stoichiometric  2:1  limit characteristic of
              an electroneutral ion exchange. It is clear
              that the amphiphilic organic cations could
              be  used to promote  desorption of metal
              cations. For example, with 1000 nM total
              DP in the system, the Dc  of  Cu(ll) is
              about one tenth of its value with no  DP,
              with the same total  amount of Cu(ll) in the
              system; at 400  pM total DP, the drop in
              Dc of  Cu(ll) is to about one fourth. Indica-
              tions  are that this  drop depends  on  the
              surface concentration of  DP relative to
              the CEC of the sorbent. Since Cu(ll) does
              seem to cause some desorption of DP at
              concentrations near  the  CEC,  it  is
              possible that another organic cation that
              adsorbs  more  strongly  (such  as
              tetradecyl- or hexadecylpyridinium) could
              displace Cu(ll) more effectively.
                 Effect of  DP  on adsorption  of
              chlorobenzenes  by  Lula A/6. The
              concentration distribution  ratio (Dc) of the
              chlorobenzenes, as  a function  of  the
              concentration of DP on  the surface, is
              shown in Figure 6. Clearly  the presence
              of  DP  on  the  surface  promotes
              adsorption. (However, at sufficiently high
concentrations  of DP, micelles  are
formed  in solution and the Dc  falls off
again. All of these data are not shown in
the figure.)
   It  is  significant that the value  of Dc
climbs not  linearly with concentration of
DP on the surface (and fractional  organic-
carbon  content of the sorbent,  f^)  but
almost quadratically, particularly for  the
chlorobenzenes with greater numbers of
Cl- substituents. Thus the  "quality" of
organic  carbon for adsorbing  hydropho-
bic compounds  increases  with  surface
coverage. One could associate this beha-
vior  with the  formation of  hydrophobic
domains on the surface; the higher  the
surface  coverage the greater the likeli-
hood that aggregates  of DP are  situated
to interact  with  a chlorobenzene  mole-
cule.
   The   range  of   experimentally
determined  values   of   Dc  can  be
compared to  those predicted from  the
Kow'foc paradigm,
                              log Dc = a log Kow + log foc + b
                                 (11)
                              For this comparison we use the values
                           of  a,  b, and  Kow  values  from  the
                           correlation  of  Schwarzenbach  and
                           Westall (1981),  which was  derived with
                           chlorobenzenes  and alkylbenzenes. The
                           foe of the sorbent is found from the mass
                           of carbon in a mole  of  DP (204 g/mol)
                           and the concentration  of  DP on  the
                           surface  (mol/kg).  The  values of  Dc
                           predicted from Equation 11 are all less
                           than those found experimentally for the
                           DP-treated Lula N6.
                              It is clear that organic cations could be
                           used to increase  the organic carbon
                           content and  the ability of an  inorganic
                           substrate  to  retain a  hydrophobic
                           compound, with  as little as 10  mmol kg-1
                           (0.2% organic carbon  for the sorbent). At
                           this  level the  distribution  of  Cu(ll) is
                           affected, but not severely. Thus by fine
                           tuning the hydrophobicity of the organic
                           cation to the  specific  environmental
                           problem  (properties  of the  sorbent,
                           properties of  the  hydrophobic solutes,
                           etc.), it should be possible to immobilize
                           hydrophobic organic compounds without
                           significantly  mobilizing metal cations.
                           Conclusions  and
                           Recommendations
                              Large  organic  cations,   such  as
                           dodecylpyridinium, are strongly adsorbed
                           to  sediment,   soil,   and  subsurface
                           materials. Adsorption  depends primarily
                           on the CEC of the  sorbent  material, the
                           nature  and  concentration  of  the
                           electrolyte, and  the concentration and
                           alkyl chain length  of  the organic cation.
                           Solution  pH has  little  effect on  the

-------
  Table 3.  Range of Concentration of DP in Solution and on the Surface over the Course
          of Cu(ll) Adsorption Isotherms.
 Total concentration in solution   Solution concentration at
         (initial) jjM              equilibrium fiM
            Surface concentration at
              equilibrium mmol kg'1
1
10
50
100
400
1000
0.016
0.355
3.19
12.2
206
778
- 0.020
- 0.424
-4.26
- 15.0
-229
-805
0.180
1.83
8.46
17.0
34.0
39.9
-0.183
- 1.87
-9.03
- 17.5
- 36.6
- 42.4
adsorption of  alkylpyridinium  on the
natural materials studied.
   The adsorption isotherms are distinctly
nonlinear,  even  at  very low surface
concentrations  of organic cation.  The
data conform to the  Freundlich equation
and can be interpreted mechanistically as
a summation of Langmuir isotherms.
   Maximum or plateau adsorption values
of adsorbed DP corresponded roughly  to
the  reported  CEC's  of the  sorbent
materials.  Adsorption  was  somewhat
greater than the CEC for EPA-12 soil; this
effect may be related to an increased role
of organic matter for high foc materials.
   The  effects   of   electrolyte
concentration  on adsorption  could  be
explained with two types of electroneutral
reactions:  cation  exchange and
adsorption  of the organic cation  with an
inorganic anion. Ca2+ had a larger effect
on adsorption than did Na + , as expected
for ion exchange.
   The adsorption of DP to Lula  aquifer
material  displaces  Cu(ll).  The
displacement occurs when the  surface
concentration of DP is within  10  - 100%
of the adsorption maximum  of  DP (an
effective  CEC) in 0.01 M  NaCI. However,
the reaction cannot  be described as a
simple  ion exchange  reaction;  pH-
dependent surface complexation of Cu(ll)
and  electrostatic competition with DP
seem to be involved.
   The adsorption of DP to Lula  aquifer
material  promotes  sorption  of
chlorobenzenes. Over the range  of DP
equivalent to foc  =  0.0001 - 0.005, the
values of Koc were not constant, but were
greater for greater loadings of DP. This
result is  attributed to cooperative  effects
of DP  molecules  at  higher  surface
concentrations.
   This work  is relevant to two potential
site-remediation  methodologies: (i) the
use  of  amphiphilic organic  cations  to
increase  the tendency  of  inorganic
substrates to retain hydrophobic organic
compounds, and  (ii)  the  use   of
amphiphilic  organic cations to  promote
desorption  ot  metal  ions  in  pump-and-
treat type  operations.  However, further
study and site-specific  information is still
necessary  for  the development  of  a
treatment program.
   For further study we recommend that
efforts be concentrated on: (i) the nature
of competition  in  the  system  organic
cation/NaCI, CaC^transition metal cation,
i.e., clarification of the role  of electrostatic
and  surface-complexation  interactions  in
the sorption of organic  cations and metal
cations  on  heterogeneous sorbents;  (ii)
the role  of the "quality" of organic carbon
in the  sorption of hydrophobic  solutes
such as  chlorobenzenes,   and  the
dependence of apparent Dc on coverage
of the  surface by organic cations; (iii)
adsorption-desorption kinetics  of organic
cations  in soil-column  experiments,  and
the concomitant effect  on  metal  ion and
hydrophobic organic compound  adsorp-
tion-desorption;  and (iv)  the effect  of
organic cations  on   the   hydraulic
conductivity of soil columns.
References
Lewis, M.  A. and  V. T.  Wee.  1983.
Aquatic Safety Assessment for Cationic
Surfactants   Environ.  Toxicol.  Chem.
2:105-118.
Schwarzenbach,  R.  P. and  J.  Westall.
1981. Transport  of  Nonpolar  Organic
Compounds  from  Surface  Water to
Ground  Water.  Environ. Sci.  Technol.
15:1360-1367
Figures.
1.  Adsorption  isotherms of dodecylpyri-
   dinium on different materials; data are
   from  the high  concentration  range of
   Figure  2,  plotted  here on  a linear
   scale. Note  the  correspondence  be-
   tween the  plateaus of the adsorption
   isotherms  and the cation exchange
   capacities listed in Table 1. Ranges of
   pH were:  kaolinite, pH =  5.39  - 4.72;
   soil  EPA-12, pH  =  7.43  -  7.00; and
   Lula  aquifer material, pH =  6.81  -
   4.49. Solution contained 0.01  M Na*;
   ratio  of solids  to liquids are given in
   Figure 2.
2.  Logarithmic  adsorption isotherms  of
   dodecylpyridinium  on  different
   materials:  montmorillonite,  ratio  of
   solids to  liquids  (Cs(w)) was 0.0050
   kg/L; kaolinite, Cs(w) =  0.025; soil
   EPA-12, Cs(w) = 0.025;  Lula  aquifer
   material, Cs(w) =  0.0025; and Borden
   sand, Cs(w)  =  0.025 or  0.050 kg/L.
   Solutions contained 0.01 M Na+.
3.  Effect of log  [H + ] on the concentration
   distribution  ratio  Dc  of dodecyl-
   pyridinium between aqueous  solutions
   and  selected sorbents:  EPA-12, Cs(w)
   = 0.015 kg/L; Lula, Cs(w) =  0.0025;
   and  kaolinite, Cs(w) = 0 0125. Solution
   contained 0.01  M Na*. The amounts of
   dodecylpyridinium adsorbed  to  the
   surface were nearly constant for EPA-
   12 and Lula sorbents.  CR(s)  =  0.20
   and  29  pmol/g,  respectively.  CR(s)
   ranged from  7.6  -  15  nmol/g  for
   kaolinite.
4.  Logarithmic  adsorption isotherms  of
   dodecylpyridinium on  Lula  aquifer
   material from 0.1  M Na*  and 0.01  M
   Ma*.  The ratio of solids to liquid was
   0.0025 kg/L.  The  solid lines in Figure 4
   were  computed from the   model
   defined in Table 2.
5.  Adsorption of Cu(ll) on Lula N6 (0.005
   kg L-1) in 0.01 NaCI at pH  in the range
   about 4.9 -  5.6, with various total
   concentrations  of DP (given  on  the
   figure), shown on  linear  and  log-
   arithmic scales. The approximate rela-
   tion  between total and adsorbed DP is
   given in Table 3. Total Cu(ll)  was in
   the  range   10 - 200 ^M  for the
   isotherms.  Isotherms were obtained
   from  the batch method, in which both
   solution and adsorbed Cu(ll) and  DP
   were determined  directly.
6.  Concentration distribution ratios (Dc) of
   chlorobenzenes  between  Lula  N6
   (0.025 kg L-i) and 0.01  M NaCI, as a
   function of  amount  of DP  on  the
   sorbent.

-------
John  C.  Westall, Bruce J. Brownawell Hua  Chen, John  M. Collier,  and Julia
      Hatfield are with Oregon State University Corvallis, OR  97331-4003
Robert Puls is the EPA Project Officer (see below).
The  complete report,  entitled "Adsorption of Organic  Cations  to  Soils  and
      Subsurface  Materials,"  (Order No.  PB 90-171 927/AS;  Cost: $17.00,
      subject to change) will be available only from:
          National Technical Information Service
          5285 Port Royal Road
          Springfield, VA22161
          Telephone: 703-487-4650
The EPA Project Officers can be contacted at:
          Robert S. Kerr Environmental Research Laboratory
          U.S. Environmental Protection Agency
          Ada,  OK 74820
                                                                                    *U.S. Government Printing Office: 1993— 750-071/60263

-------

-------

-------
m
T)
£
O)
o
o
C/)
N>
ci>
o
                                                                                              3  —•
                                                                                              0)  Q.
                                                                                              s: 05.


                                                                                              O?

                                                                                              -oi
                                                                                              =- CD
                                                                                              <  M

                                                                                              CD

                                                                                              C

                                                                                              CD
                                                                                              O
                                                                                              O
CO  3  D
 CD  <  =V
 3  =;• CD
 O  O  Q.
    O  0)

    =  CD
    0)  (/)


    TJ

    O

    CD
    O
                                                                                                          ^' 2* CD
                                                                                                          o 2 =
                                                                                                          5' 3 ®
                                                                                                          3 w
                                                                                                          O
                                                                                                          o
                                                                                                          03
                                                                                                                m
                                                                                                                3
                                                                                                                3J
                                                                                                                CD

-------