if'

1'

le.

~;' \

,-~ J,::,

\
1

- ..
KINETICS OF SOIL CHEMICAL ~~----
\;- .- ------PROCESSES
..-..:

-\

~ ----

_,...., .

Sparks, D. L. 1995. Environmental soil Chemistry,

Academic Press, San Diego, CA, USA.

Many soil chemical processes are rime-dependent. To fullv undersr:md rh e

dvnamic inreractions of metals, radionuclides. pesticides, industrial chernical s.
and planr nutrients with soils and ro predict rheir fate with rime, '' knowledge
of the kinetics of rhese reactions is imporrant. This chaprer will rro\ide an
overview of this tapie, with applicarions ro environmenrally importam re3 crions. The reader is referred ro se"eral sources for more definitive discu ssion s
on rhe tapie (Sparks. 1989: Sparks and Su .uez. 1991).

RATE-LIMITING STEPS ANO TIME SCALES OF SOIL CHEMICAL REACTIONS

Four main processes can affecr rhe rate 0f soil chemical reactions. These c~ n
be bro;:J[y classified as rransport and chcmical reaction processes (Fig. 7. 1).
The slowest of rhese will limit rhe rate of .1 particular reaction. Bulk tr:mspon
\1 in Fig. 7.1 ). which occurs in rhe solu rion phase, is 11ery rapid ami is not
normally rare-limiring. In rhe laborarory, r can he eliminared by rapid mi xing.
The acrual chemical reacrion (CR) at rhe surface (4), e.g., adsorption, is also
rapid and usually not rate-limiring. The rwo remaining transpon or rna ss
rr:111sfer processes, either singly or in combinaran, normal!v are rate-lirniring.
Film diffusion (FD) involves transpon of 111 ion or molecule rhrough a bound :lf)' laver or film (water molecules) thar surrounds rhe parricle sttrface (21.
l'anicle diffusion (PD), somerimes refened ro as intraparricle diffusion. involves transpon of an ion or molecule along pore-wall surfaces (3b) and/or
wirhin the pares of rhe panicle surface (3,1).

159

- - - - - - - - - - - ---- ----- ------ --

160

KIHETICI Of IOIL CHEHICAL PROCESIEI

CHAPTER

Rate-determining steps in soil chemical reactions. From Weber, W. J., Jr. (1984)
Evolution of a technology. J. Environ. Eng, Div. (Am. Soc. Civ. Eng.) 110, 899-917. Reproduced with

FIGURE 7.1.

permission of ASCE.

Soil chemical reactions occur over a wide time scale (Fig. 7.2), ranging from
microseconds and milliseconds for ion association (ion pairing, complexation,
and chelation rype reactions in solurion), ion exchange, and sorne sorption
reactions to years for mineral solution (precipitationldissolution reactions
including discrete mineral phases) and mineral crystallization reactions
(Amacher, 1991). These reactions can occur simultaneously and consecutively.
Certainly an important factor in controlling the rate of many soil chemical
reactions is the type and quantity of soil cornponems. For example, ion
exchange reactions are usually more rapid on clay minerals such as kaolinite
and oxides than on c_lay surfaces such as vermiculite and mica. Ths is attributed to the externa! exchange sites on kaolin!te versus the multiple types of
exchange sites with vermiculite and micas. Externa! planar, edge, and interlayer sites exist on the surfaces of vermiculite and micas with sorne of the latter
partially or rorally collaps.!d. High rates of reaction are often observed for
externa! si tes, intermediare rates on edge sites, and low rates on interlayer si tes
(Jardine and Sparks, 1984a).
A number of investigators have found that adsorption reactions of certain
metal cations such as Cu 1 + and anions such as borate, arsenate, molybdate,

161

RATE LAWI

selenire, selenate, and chromate occur on goethite surfaces on millist'-!ond tim;

scales (Zhang and Sparks, 1989, 1990b; Grossl et al., 1994). Sorption of
metals on humics is also rapid. Half-lives tor Pb2+, Cu1 +, and Zn2+ sorption.on
peat ranged from 5 to 15 s (Bunzl et al., 1976).
However, there are many adsorption, ;tnd panicularly desorption, reactions
involving organic chemicals such as pesticides, where the reaction rates are
very low. It appears that an important factor affecting the rates of organic
chemical reactions in soils is the ritne period over which the organic compound
has been in contact with che soil. For example, 1,2-dibromoethane (EDB)
release from soils reacted in the laboratory over a short period of time was
much more rapid than EDB release from field soils rhat had been contaminated
wirh EDB for many years. This difference in release was related to greater PD
into micropores of clay minerals and humic componems that occurred at
longer times (Steinberg et al., 1987).
lt would be instructive at this point to define two important termschemical kinetics and kinetics. Chemical kinetics can be defined as "the investigation of chemical . reaction rates ancl the molecular processes by which
reactions occur where transpon is not limiting" (Gardiner, 1969). Transpon
phenomena, as mentioned earlier, include transpon in the solution phase, film
diffusion, and particle diffusion. Kinetic> is the study of tirne-dependent processes.
The study of chemical kinetics in homogeneous solutions is difficult, and
when one studies heterogeneous systems such as soil components and, particularly, soils, the difficulties are magnified. It is extremely difficult ro eliminare
transport processes in soils because they are mixtures of severa! inorganic and
organic components that are often imimately associated with each other and
beca use soils have multiple types of sites with varying reactivities for inorganic
and organic adsorbates. Additionally, there are an array of different panicle
sizes and porosities in soils that enhance their heterogeneity. Thus, when
dealing with soils and soil componems, one usually studies the kinetics, simply
defined as the study of tirne-dependent processes, of these reactions.

RATE lAWS
There are two important reasons for investigating the rates of soil chemical
processes (Sparks, 1989): (1) to determine how rapidly reactions attain equilibrium, and (2) ro infer inforrnation on reaction mechanisms. One of che rnost
important aspects of chemical kinetics is the establishment of a rate law . Bv
definition, a rate law is a differential ~quation. For the following reactio~
(Bunnett, 1986),

Ion Assoclation
Multlvalent Ion Hydrolysis
Gas-Water

Ion Exchange
Sorptlon
Mlnerai-Solutlon

aA + bB-yY + zZ,

Mineral Crystalllzatlon _.,__.

S

min

day

mo

yr

mil

Time Scale

FIGURE 7 .2.

Time ranges required to attain equilibrium by different types of reactions in soil

environments. From Amacher ( 1991 ), with permission.

(7.1)

the rate of the reaction is proporcional ro sorne power of the concentrations of

reactants A and B and/or other species (C, D, etc.) in the system. The terrns a,
b, y, and z are stoichiometric coefficient>, and are assumed to be equal to one
in the following discussion. The power ro which theconcemration is raised
rnay equal zero (i.e ., che rate is independent of that concentration ), e ven for

l-62

,,

CHAPTER 7

KIHETICI OF IOIL CHEHICAl PROCEIIEI

reactant A or B., Rates are expressed as a decrease in reactant concentration or

an increase in product concemration per unir time. Thus, the rate of reactant A
abo ve, which has a concentration [A) at any time t, is ( -d[A)/(dt)) while rhe
rate with regard to product Y having a concenrration [Y] at time t is (d[Y]/
(dt)).
The rate expression for Eq. (7.1) is

d[Y]Idt = -d[A]Idt = k[ A]" [B]il

... '

(7.2)

where k is the rate constant, a is the arder of the reaction with respect to
reactant A and can be referred toas a partial order, and f3 is the order with
respect to reactant B. These orders are experimentally determined and not
necessarily integral numbers. The sum of all the partial orders (a, /3, etc.) is the
overall order (n) and may be expressed as
n=a+f3+.

(7.3)

Once the values of a, {3, etc., are determined experimenrally, the rate law is
defined. Reaction order provides only information about the manner in which
rate depends on concentration. Order does not mean the same as "niolecularity" which concerns the number of reactant particles {atoms, molecules, free
radicals, or ions) entering into an elementary reaction. One can define an
elementary reaction as one in which no reaction intermediares have been
detected or need to be postulated to describe the chemical reaction on a
molecular scale. An elementary reaction is assumed to occur in a single step
and to pass through a single transition state {Bunnett, 1986).
To prove that a reacrion is elemenrary, one can use experimental conditions
rhat are different from those employed in determining the law. For example, if
one conducted a kinetic study using a flow technique (see later discussion on
this technique) and the rate of influent solution {flow rate) was 1 mi min- 1,
one could study severa! other flow rates ro see if reaction rate and rate
constants change. If they do, one is not determining mechanistic rate laws.
Rate laws serve three purposes: rhey assist one in predicting the reacton
rate, mechanisms can be proposed, and reacrion orders can be ascerrained.
There are four types of rate laws that can be determined for soil chemical
processes (Skopp, 1986): mechanistic, apparent, transpon with apparent, and
transport with mechanistic. Mechanistic rate laws assume that only chemical
kinetics are operational and transport phenomena are not occurring. Consequently, it is difficult to determine mechanistic rate laws for most soil chemical
systems due to the heterogeneiry of the sysrem caused by different particle
sizes, porosities, and rypes of retention sites. There is evidence that with sorne
kinetic studies using relaxation techniques .(see later discussion) mechanistic
rate laws are determined since the agreement berween equilibrium constanrs
calculated from both kinetics and equilibrium studies are comparable {Tang
and Sparks, 1993). This would ind!cate that transpon processes in the kinetics
studies are severely limited {see Chaprer 5). Apparent rate laws include both
chemical kinetics and transporr-controlled processes. Apparent rate laws and
rate coefficients indicare that diffusion and other microscopic transport processes affect the reaction rate. Thus, soil strucrure, stirring, mixing, and How
rate all would affect the kinetics. Transpon with apparent rare laws emphasize

163

OHERMINATION Of REACTION OROER ANO RATE CONITANTI

transport phenomena. One often assumes first-order or zero-order reactions

(see discussion below on reaction arder). In determining transport with mechanistic rate laws one attempts to describe ; imultaneously transport-controlled
and chemical kinetics phenomena. One is thus trying to accurately explain
both the chemistry and the physics of the ;ystem.

OETERMINATION OF REACTION OROER ANO RATE CONSTANTS

There are three basic ways ro determine rate laws and rate constants
(Bunnett, 1986; Skopp, 1986; Sparks, 1989): (1) using initial rates, {2) directly
using integra red equations and graphing tite data, and (3) using nonlinear least
square analysis.
Let us assume the following elementarv reaction between species A, B, and

Y,
k,

A+ B:;::==Y.
k_,

(7.4)

A forward reaction rate law can be written as

d[A]Idt

-k 1[A][B],

where k 1 is the forward rate constant and a and f3 (see Eq. 7.2) are e:1ch
assumed to be l.
The reverse reaction rate law for Eq. (7.4) is

d[A]Idt

+ k_ 1[Y],

(7.6)

where k_ 1 is the reverse rate constant.

Equations {7.5) and (7.6) are only ap plicable far from equilibrium where
back or reverse reactions are insignificant. If both these reactions are occurring, Eqs. {7.5) and (7.6) must be combined such that,

d[A]Idt = - k,[A][B] + k_,[Y].

{7.7)

Equation {7.7) applies the principie that the net reacrion rate is the difference between the sum of all reverse reaction rates and the sum of all forward
reaction rares.
One way to ensure that back reactions are not important is to measure
inicial rates. The initial race is the limit of the reaccion rateas time reaches zero.
\'V'ith an initial rate method, one plots rhe concentration of a reacranr or
produce over a short reaction time period during which the concentrations of
the reactants change so litde that the instantaneous rate is hardly affecred.
Thus, by measuring initial rates, one could assume that only the forward
reaction in Eq. {7.4) predominares. This would simplify the rate l:.lw co that
given in Eq. {7.5) which as wrirten would be a second-order reaction, firstorder in reactant A and first-order in reactant B. Equation {7.4), under these
conditions, would represem a second-order irreversible elementary reaction.
To measure initial rates, one must have a:ailable a technique chat can mensure
r::~pid reactions such as a relaxation me rhod (see derailed discussion on rhis

CHAPTE~

164

7 KIHETICS Of SOIL CHEHICAL PROCESIES

later) and an ,a ccurate analytical detection system to determine product concenrrarions.

Inregrated rate equations can also be used ro determine rate constants. If
one assumesthat reactant B in Eq. (7.5) is in large excess of reactant A, which
is an example of rhe "method of isolation" to analyze kinetic data, and Y0 = O,
where Y0 is the initial concentraran of product Y, Eq. (7.5) can be simplified
to

d[A]Idt = - k 1 [A) .

(7.8)

not affected by species whose concentrarions do not change considerably

during an experiment; these may be substances not consumed in the reaction
(i.e., catalysts) or present in large excess (Bunnett, 1986; Sparks, 1989).
Least squares analysis can also be used ro determine rate constants. With
this method, one fits the best straight line ro a set of points that are linearly
relared as y = mx + b, where y is the ordinate and x is the abscissa datum
point, respectively. The slope, m, and the inrercept, b, can be calculated by
least squares analysis using Eqs. (7.10) and (7.11), respectively (Sparks, 1989),

n l:xy- Ix 2:y
m= n ,x-~ ,
" )2
(~x

The first-order dependence of [A) can be evaluated using rhe integrared

form of Eq. (7.8) using the inicial conditions at t = O, A = A 0 ,
log [A), = log [A) 0

kt

2.303'

1.6.----------------..,

(Mn2] 0 = 40 11 M
y= 1.61 . (8.65 x 1oJx, R2 = 0.998

::!; 1.4

b = l:y 2:x1

(7.9)

The half-rime (t 112 ) for the above reaction is equal to 0.6931k 1 and is the rime
required for half of reactant A to be consumed.
If a reaction is first-order, a plot of lag [A), vs t should result in a straight
line with a slope = - k12.303 and an inrercept of log [A) 0 An example of
first-order plots for Mn2 + sorption on 5-Mn0 2 at two initial Mn 2 + concentrations, [Mn 2 + ) 0 , 25 and 40 p.M, is shown in Fig. 7.3. One sees that the plots are
linear at borh concentrations, which would indicare that the sorption process
is first arder. The [Mn 2 + ) 0 values, obrained from rhe intercept of Fig. 7.3, were
24 and 41 p.M, in good agreemenr with the two (Mn2+) 0 val u es. The rare
constants were 3.73 X 10- 3 and 3.75 X 10- 3 s- 1 at [Mn 2 +] 0 of 25 and
40 p.M, respectively. The findings that the rate constants are not significantly
changed with concentration is a very good indicarion that the reacrion in Eq.
(7.8) is first arder under the experimental conditions that were imposed.
!t is dangerous ro conclude that a particular reacrion order is correct, based
simply on the conformity of data to an inregrated equation. As illustrated
above, multiple inicial concentrations that vary considerably should be employed to see rhat the rate is independent of concentrarion. One should also
test multiple integrated equations. lt may be useful to show rhat reaction rate is

165

KINETIC HOOELS

n 2:x 1

(7.10)

~xl:xy

Cix) 1

(7.11)
'

where n is the number of data points and the summarions are for all data
points in the set.
Curvature may result when kineric data are plottetl. This may be due roan
incorrect assumption of reaction order. If firsr-order kinerics is assumed and
the reaction is really second arder, downward curvature is observed. If second-order kinetics is assumed but the reacrion is first-order, upward curvature
is observed. Curvarure can also be due to fractional, third, higher, or mixed
reaction order. Nonattainment of equilibrium often resulrs in downward curvature. Temperature changes during the study can also cause curvarure; rhus,
it is important that temperature be accurarely conrrolled during a kineric
experiment.

KINETIC MODELS
While first-order models have been used widely to describe the kinetics of
soil chemical processes, a number of orher models ha ve been employed. These
include various ordered equations such as zero-order, second-order, and fractonal-order, and Elovich, power function or fracrional power, and parabolic
diffusion models. A brief discussion of sorne of these will be given; rhe final
forms of the equarions are given in Table 7.1. For more complete details and
applications of these models one should consult Sparks ( 1989).

:1.

Elovich Equation

1:

1.2

:E
Ol

.2

1.0

0.8

L , _ _ J _ _ _ J _ _ _ . . L __

_t___

_l__

_l__J__....J
~

00

Time, ms

FIGURE 7.3. lnitial reaction rates depicting the first-order dependenc of Mn 1 ' sorption as a
function of time for inicial Mn 1 ' concentrations ([Mn'-] 0 ) of 25 and 40 p.M. From Fendorf et al. (1993),
with permission. .

The Elovich equation was originally developed to describe the kinerics of

hererogeneous chemisorption of gases on salid surfaces (Low, 1960). lt seems
ro describe a number of reaction mechanisms including bulk and surface
diffusion and activation and deacrivation of catalytic surfaces.
In soil chemistry, the Elovich equation has been used ro describe the kinerics
of sorption and desorption of various inorganic materials on soils (see Sparks,
1989). lt can he expressed as (Chien and Clayton, 1980)

q, = (113) In (a/{3) + (1/3) In t,

(7.12)

. i

166

CHAPTER 7 KIHETICI Of IOIL CHEHICAL PROCEIIEI

,,
-

TABLE -7.1. linear Forms of Kinetic

Equations Commonly Used in Environmental
Soil Chemistry"

[AJ, = [AJ 0

..,
i:
-e

1
1

Zero orde.r"
-

i1

kt

:::1.

6
'o

First order"

167

KINETIC HOOELI

kt

log [AJ, = log [AJo - - 2.303'

60
40

Second ordeib
1
1

..

Porirua Soll .

20
-2

A.-1

...

r2 = 0.990

..............
4

In t, h

-=-+kt
[Al, [AJo

Elovich
q, = (11/3) In (a//3) + (1//3) In t

.!

Parabolic diffusion

FIGURE 7 .4. Plot of Elovlch equation for phosphat<! sorption on two soils where ( 0 is the initial
phosphorus concentration added at time O and C is the phosphorus concentration in the soif solution at
time t The quantity (C.-Q can be equared to qr the amount sorbed at time t. From Chien and Claycon
( 1980). with permission.

~= R0 t 11!
q~

Power funcrion

radial diffusion in a cylinder where the ion concentration on che cylindrical

surface is constant, and initially the ion concentration throughout the cylinder
is uniform. lt is also assumed that ion diffusion chrough che upper and lower
faces of the cylinder is negligible. Following Crank (1976), the parabolic
diffusion equation, as applied to soils can be expressed as

In q = In k + v In t
.~

Terms are defined ln the texr.

Describing rhe reacrion A--+Y.
' In x = 2 .303 log x, is rhe conversion from natural
logarithms (in) ro base 10 logarirhms (log).

(7. 13)

where q, is the amount of sorbate per unir mass of sorbenr ar time tanda and
{3 are constants during any one experiment. A plot of q, vs In t should give a
linear relationship if rhe Elovich equation is applicable wirh a slope of (1/{3)
and an incercepc of ( 1/{3) In (a{3).
An applicarion of Eq. (7.12) ro phosphate sorption on soils is shown in Fig.
7.4.
Sorne invescigacors have used che a and {3 parameters from the Elovich
equation ro estimare reacrion rates. For example, ir has been suggested thar a
decrease in {3 and/or an increase in a would increase reacrion rate. However,
this is quesrionable. The slope of plots using Eq. (7. 12) changes wich the
concentration of the adsorpcive and with rhe solution to soil ratio (Sharpley,
1983). Therefore, che slopes are not always characceriscic of the soil but mav
depend on various experimental conditions.
'
Sorne researchers have also suggested thac "breaks" or multiple linear
segments in Elovich plots could indicare
changeover from one rype of
binding site ro .another (Atkinson et al., 1970). However, such mechanisric
suggescions may not be correcc (Sparks, 1989).

Parabolic Diffusion Equation

The parabolic diffusion equacion is often used to indicare thar diffusioncontrolled phenomena are rate-limiting. lt was originallv derived based on

where r is the average radius of the soil parricle, q, was defined earlier, qx is the
corresponding quantity of sorbate at equilibrium, and D is the diffusion
coefficient.
Equation (7.13) can be simply expressed as
q,lq~ =

R0 t 112 + constant,

(7.14)

where R0 is the overall diffusion coefficient. If che parabolic diffusion law is

valid, a plot of q/q~ versus t"z should yield a linear relationship.
The parabolic diffusion equation has successfully described metal reacrions
on soils and soil constituents (Chute and Quirk, 1967; Jardine and Sparks,
1984a), feldspar weathering (Wollast, 1967), and pesticide reactions (Weber
and Gould, 1966).

Fractional Power or Power Function Equation

This equation can be expressed as
q = kt'',

(7.15)

where q is the amount of sorbate per unit mass of sorbent, k and v are
constants, and vis positive and < l. Equation (7.15) is empirical, except for thr
case where v = 0.5, when Eq. (7.15) is >imilar to the parabolic diffusion
equation.

l
1.68

CHAPTER 7 KINETICI OF IOil CHEMICAl PROCEIIEI

..

KINETIC HETHODOlOGIEI
200

Equation (7.15) and various modified forms have been used by a number of
researchers to describe the kinetics of soil chemical processes (Kuo and Lotse,
1974; Havlin and Wesfall, 1985).

169

r------------------.

Comparison of Kinetic Models

In a number of srudies it has been shown that severa! kinetic models

describe the rate data well, based on correlation coefficients and standard
errors of the estimare (Chien and Clayton, 1980; Onken and Matheson, 1982;
Sparks and Jardine, 1984). Despite this, there often is nota consistent relation
between the equation that gives the best fit and the physicochemical and
mineralogical properties of the adsorbent(s) being studied. Another problem
with sorne of the kinetic equations is that they are emprica! and no meaningful
rate parameters can be obtained.
Aharoni and Ungarish (1976) and Aharoni (1984) noted that sorne kinetic
equations are approximations ro which more general expressions reduce in
cerrain limited time ranges. They suggested a generalized emprica! equation
by examining the applicability of power function, Elovich, and first-order
equations to experimental data. By writing these as the explicit functions of the
reciproca! of the rate Z, which is (dq/dt)- 1, one can show that a plot of Z vs t
should be convex if the power function equation is operacional (1 in Fig. 7.5),
linear if the Elovich equation is appropriate (2 in Fig. 7.5), and concave if the
first-order equation is appropriate (3 in Fig. 7.5). However, Z vs t plots for soil
systems (Fig. 7.6) :lre usually S-shaped, convex at small t, concave at large t,
and linear at sorne intermediare t. These findings suggest rhat the reacrion rare
can best be described by the power function equation at small t, by the Elovich
equation atan intermediare t, and by a firsr-order equation at large t. Thus, the
S-shaped curve indicares that the above equations may be applicable, each at
sorne limited time range.
One of the reasons a particular kinetic model appears to be applicable may
be that the study is conducted during the time range when the model is most
appropriate. While sorption, for example, decreases over many orders of

Time, h

FIGURE 7.6.

Sorption of phosphate by a Typic Dystrochrept soil plotted as Z vs time. The cirdes

represent the experimental data of Polnopoulos et al. (19.86). The solid line is a curve calculated
according to a homogeneous diffusion model. From Ahamni and Sparks ( 1991 ), with permission.

magnirude befare equilibrium is approached, with most methods and experiments, only a portion of the entire reaction is measured and over this rime
range the assumprions associated with a particular equation are valid. Aharoni
and Suzin (1982a,b) showed that the S-shaped curves could be well described
using homogeneous and heterogeneous diffusion models. In homogeneous
diffusion situations, the final and initial portions of the S-shaped curves (conforming to the power function and first-order equations, respectively) predominated (see Fig. 7.6 showing data conformity toa homogeneous diffusion
model), whereas in insrances where rhe heterogeneous diffusion model was
operacional, the linear portian of the S-sh<1ped curve, which conformed ro the
Elovich equation, predominated.
The fact that diffusion mode!s describe a number of sbil chemical processes
is not surprising since in most cases, mass transfer and chemical kinetics
phenomena are occurring simultaneously and it is difficult to separare them.
Therefore, the overall kinetics of many ;oil chemical reactions may often be
better described by mass transfer and diffusion-based mode!s than with simple
models such as a first-order model. This is particularly true for slower soil
chemical reactions where a fast reaction is followed by a much slower reacrion
(biphasic kinetics). This is ofren observed for many reactions in soils involving
organic and inorganic chemicals. Table 7.2 lists severa! types of alternative
kinetic models including one-site, two-site, and diffusion models that have
successfully been employed to describe sorption kinetics.

KINETIC METHODOLOGIES

0 0~~~2~~3--4--~5--6~~~--~9~10

lime, arbitrary unils

FIGURE 7.5.

Plots of Z vs time implied by (1) power functlon model, (2) Elovich model. and (3)
first-order model. The equations for the models were differentiated and expressed as explicit functions
of the reciproca! of the rate, Z. From Aharoni and Sparks ( 1991 ). with permission.

A number of methodologies can be used to study the rares of soil chemical

processes. These can be broadly classified as merhods for slower reactions
(> 15 s), which include batch and flow techniques, and rapid techniques that
can measure reactions on millisecond ami microsecond time scales. Ir should
be recognized that none of these methods is a panacea for kinetic analyses.
They all have advantages and disadvantages. For comprehensive discussions

"

' 170
-

CHAPTER 7 KINETICS Of SOil CHEHICAl PROCESSES

TABLE 7.2.

KINETIC HETHOOOLOGIES

Comparison or Sorption Kinetic Models'b

Conceptual model

Fitting parameter(s)

Model llmitations

One-site model (Coates and Elzerman,

kd

Cannot describe biphasic sorption/

desorpcion

1986)

centrifugation .is necessary, and the solid: solution ratio may be al te red .1 1
.
. .
.
'
S t lt:
expenment proceeds. Too much mtxmg mav cause abraston of the adsorh,
may en hance m~ss tr ene,
t he sur f ace area, w h'l1 e too 1ttt 1e mtxmg
a 1tenng
-.t.
'"
. ns n
and transport processes. Anorher major problem with all batch techni ' , .
unless a resin or chelare material suc~ as Na-tetraphenylboron is usctl, isq~;~,~;
released speetes are not removed. Thts can cause mhtbt_tton in funhcr adsnrbate release and promotton of secondary prectpttatton m dissolution studics
Moreover, reverse reactions are not conrrolled, which makes the ca,.ulr
f.
~ ' ton o
rare coefficients difficult and perhaps inaccurate.
Manr of rhe disadvantages lisred above for traditional batch tcchniqucs ..
be eliminated by using a method like thar of Zasoski and Burau ( 1<J7u) .1 c.l!t

t) , S l0\\'11
m ftg. /.7. In thts method an adsorbent ts placed m a vessel comaining tht
adsorpttve, pH and suspenswn volume ar>: adtusted, and thc suspcnsio, .
vigorously mixed with a magnetic stirrer. At various times, suspension ,,, 1 1\
. hd
.
.
. .
,
Th
l"o"
are wtt rawn usmg a synng~ contatnmg \" gas. . e ~ 2 gas prevcnts COJ
and 0 2 from enrermg the reacnon vessel. The suspens1on ts rapidly filtcr , 1. 1
. hed an d ana lyze d . Wtt
. h rh'ts apparatus a con~tam
.lll(
rhe fi !trates are t h en wetg
)ll
can be maintained, reactions can be measured at 15-s intcrvals, c~cdll
mixing occurs, anda constant solid ro solu1ion ratio is maintaincd.
cnt

s~c
Two-site model (Coares and Elzerman,
1986)

kd,

K~, X 1

Radial diffusion: penetraran

retardarion (pore diffusion) model
(Wu and Gschwend, 1986)

Dual-resistance surface diffusion model

(Miller and Pedir, 1992)

s~c:~c

Cannot describe the "bleeding" or

slow, reversible, nonequilibrium
desorprion for residual sorbed
compounds lKarickhoff, 19801
Cannot describe instanraneous uptake
wirhout additional correction factor
(Ball, 1989); did not describe kinetic
data for times grearer than lO' min
(Wu and Gschwend, 1986)

s~~c
0, k.

Model calibrared with sorprion data

predicred more desorprion rhan
occurred ln the desorption
experiments (Miller and Pedir, 1992)

Reprinred wirh permission from Connaughton et al. (1993). Copyright 1993 American Chemical Socierj.
Abbreviarions used are as follows: S, concenrration of rhe bulk sorbed conraminant (g g 1); C, concenrration of
rhe bulk aqueous-phase contaminanr (g mJ- 1); kd, firsr-order desorprion rare coefficienr (min- 1); S,, concentration of the sorbed contaminant that is rate limir<d (g g'); 51 , concenrrarion of the contaminanr rhar is in
equilibrium wirh rhe bulk aqueous concenrrarion (g g- 1 ); X1, fracrion of che bulk sorbed coma mina m rhar is in
equilbrium with the aqueous concentration; KP, sorption equilbrium particion coeffienr (mi g-l); Deth
effective diffusiviry of sorba te molecules or ions in the parricles (cm 1 s -) S', concentration of conraminant in
immobile bound srare (mol g- 1); C', concemrarion of concaminanr free in che pare fluid (mol cm - 3 ); n,
porosiry of che sorbent (cm 3 of fluid cm-.l toral); O m, pore fluid diffusiviry of che sorbare (cm' s- 1 ); p, specific
graviry of rhe sorbent (g cm- 3 ); ((n,t), pore geometty factor; kh, boundary !ayer mass ttansfer coefficient (m
s- ); R, radius of rhe spherical so lid patticle, assumed consta m (m); p, macroscopic particle densiry of che so lid
phase (g m 3 ); e;, solurion-phase solure concentrarion corresponding roan equilibrium wirh rhe solid-phase
salute concenrrarion at rhe exterior of the parricle (g lirer- 1); D, surface diffusion coefficient (m s- 1).
' K, can be derermined independemly.
J K,, Dm, and p, can be determined independently.

Flow Methods
Flow methods can range from continuous flow techniqucs (Fig. 7.:\), wh .1
are similar to liquid-phase chromatography, ro srirred-flow merhods ( : 1 ~. -/~J;
thar combine aspects of borh batch and flow methods. lmportant :tttrih;ttl's ol
flow techmques are that one can conduct -;rudtes at reahst1c soil ro '"lut"
ratios that better simulare_ field conditions, rhe adsorbent is l'Xpo"d 1111
greater mass of tons than 111 a stanc batch system, and rhc flowin" ,. 11 1 .
,., '
llt\111!
removes desorbed and detached species.

::

on kinetic methodologies one should consult Sparks (1989), Amacher (1991),

Sparks and Zhang ( 1991), and Sparks et al. (1995).

~co,
b-=<""'
-J
mi!i!i]

Batch Methods
Batch methods have been the niost widely used kineric techniques. In the
simplest rraditional batch technique, an adsorbent is placed in a series of
vessels such as cenrrifuge tubes wirh a particular volume of adsorptive. The
tubes are then mixed by shaking or stirring. At various times a rube is sacrificed for analysis, i.e., the suspension is either centrifuged or filtered to obrain
a clear supernatant for analysis. A number of variations of batch methods exisr
and rhese are discussed in Amacher ( 1991 ).
There are a number of disadvantages ro rraditional batch methods. Often
che reaction is complete before a measuremenr can be made, particularly if

171

1,1) ~

...,

0.1 14 NnOft

Dig1tal Buret

Heat Shietd

Magnetic Stlrring Unit

FIGURE 7.7. Schematic diagrain of equipment used in batch technique of l.IHI'\kl ,11111 1\tu.u,
( 1978), with permission.

172

CHAPTH 7

KINETICI OF IOIL CHEHICAl PROCEIIEI

With continuous flow methods, samples can be injected as suspensions or

spread dry on a membrane filter. The filter is attached to its holder by securely
capping it, and che filter holder is connected to a fraction collector and
peristaltic pump, the latter maintaining a constant flow rate. Influent solution
then passes through the filter, reacts with the adsorbent, and at various times,
effluents are collected for analysis. Depending on flow rate and the amount of
effluent needed for analysis, samples can be collected about every 30-60 s.
One of the majar problems with this method is that the colloidal particles may
not be dispersed, i.e., che time necessary for an adsorptive ro travel through a
thin !ayer of colloidal particles is not equal at all locations of the !ayer. This
plus mnima! mixing promotes significant transport effects. Thus, apparent
rate laws and rate coefficients are measured, with the rate coefficients changing with flow rate. There can also be dilution of the incoming adsorptive
solurion by che liquid used ro load the adsorbenr on the filter, parricularly if
the adsorbent is placed on the filter as a suspension, or if there is washing out
of remaining adsorptive solution during desorption. This can cause concentration changes not due to adsorption or desorption.
A more preferred method for measuring soil chemical reacrion rates is che
stirred-flow method . The experimental setup is similar to the continuous flow
method (Fig. 7.8) except there is a stirred-flow reacrion chamber rather than a
membrane filter. A schematic of this method is shown in Fig. 7.9. The sorbent
is placed into the reaction chamber where a magnetic stir bar or a overhead
stirrer (Fig. 7.9) keeps it suspended during the experimem. There is a filter
placed in the top of the chamber which keeps the solids in the reaction
chamber. A peristalric pump maintains a constant flow rate and a fraction
collector is used to collect the leachates. The stirrer effects perfect mixing, i.e.,
che concentration of che adsorptive in che chamber is equal ro the effluent
concentration.
This method has severa! advantages over the continuous flow technique and
other kinetic methods. Reaction rates are independent of che physical properties of che porous media, the same apparatus can be used for adsorption and
desorption experiments, desorbed species are removed, continuous measuremems allow for monitoring reaction progress, experimental factors such as
flow rate and adsorbent mass can be easily altered, a variety of solids can be

I~IH
Reservoir

FIGURE 7.8. Thindisk flow (continuous flow) method experimental setup. Background solution
and salute are pumped from the reservoir through the thin disk and are collected as aliquots by the
fraction collection. From Amacher ( 1991 ), with permission.

173 ,,

KIHETIC HETHOOOLOGIEI

I~IH
Reservo ir

FIGURE 7.9.

Stirred-flow reactor method experimental setup. Background solution and solute are

pumped from the reservoir through the stirred reactor containing the solld

phase nd are collected as

aliquots by the fraction collector. Separation of salid and liquid phases ls accomplished by a membrane
filter at the outlet end 6f the stirred reactor. From Amacher ( 199 1}. with permission.

used (however, sometimes fine particles can clog che filter, causing a buildup in
pressure which results in a nonconstant flow rate) with the technique, the
adsorbent is dispersed, and dilution errors can be measured. With this method,
one can also use stopped-flow tests and vary influent concentrations and flow
rates to elucidare possible reaction mechanisms (Bar-Tal et al., 1990).

Relaxation Techniques
As noted earlier, many soil chemical reactions are very rapid, occurring on
millisecond and microsecond time scales. These include metal and organic
sorption-desorption reactions, ion exchange proesses, and ion associarion
reactions. Batch and flow techniques, which meas u re reaction rates of > 15 s,
cannot be employed to measure these reacrions. Chemical relaxation methods
must be used to measure very rapid reacrions. These include pressure-jump
(p-jump), electric field pulse, remperaturejump (t-jump), and concentrationjump (c-jump) methods. These methods are fully oudined in other sources
(Sparks, 1989; Zhang and Sparks, 1993). Only a brief discussion of the theory
of chemical relaxation and a description of p-jump merhods will be given here.
The theory of chemical relaxation can be found in a number of sources (Eigen,
1954; Takahashi and Alberty, 1969; Bernasconi, 1976). lt should be noted
that relaxation techniques are best used with soil components such as oxides
and clay minerals and not whole soils. Soils are heterogeneous, which complicares the analyses of the relaxation data.
All chemical relaxation merhods are based on the theory that the equilibrium of a system can be rapidly perrurbed by sorne externa! factor such as
pressure, temperature, or electric field strength. Rate information can then be
obtained by measuring the approach from che perturbed equilibrium ro the
final equilibrium by measuring the relaxation time, T (the time that it takes for
che system to relax from one equilibrium state to anorher, after the perturbation pulse) by using a detection system such as conductivity. The relaxation
time is related to the specific rates of the elementary reactions involved. Since
the perturbation is small, al! rate expressions reduce to, first-order equations
regardless of reaction arder or molecularity (Bernasconi, 1976). The rare

'174

CHAPTER 1

KIHETICI Of IOll CHEMICAL PROCEIIEI

equations are then linearized such that

T- 1

= k,(CA +

C8 )

+ k_ 1,

175

EFFECT Of TEMPERATURE OH REACTIOH RATEI

r--

{~__o_lg.:..;_uz_e_r_

_,

(7.16)

where k, and k_, are the forward and backward rate constants and C_... and C8
are the concenrrations of reactants A and B at equilibrium. From a linear plot
of ,-, vs (CA + C8 ) one could calculare k, and k_, from the slope and
intercept, respectively. Pressure-jump relaxation is based on the principie that
chemical equilibria depend on pressure as shown below (Bernasconi, 1976),

a InKo)
(= -tl.V!RT'
a In p T

(7.17)

where Ko is the equilibrium constant, ll Vis rhe standard molar vol u me change
ot the reaction, p is pressure, and R and T were defined earlier. For a small
perturbation,

FIGURE 7.1 O.

Schematic diagram of che electron paramagnetic resonance monitored stopped-

flow kinetic apparatus. From Fendorf et al. ( 1993), with permission.

(7.18)
Details on the experimental protocol for a p-jump study can be found in
severa! sources (Sparks, 1989; Zhang and Sparks, 1989; Grossl et al., 1994).
Fendorf et al. (1993) used an electron paramagneric resonance sroppedflow (EPR-SF) method (an example of a c-jump merhod) to study reactions in
colloidal suspensions in situ on millisecond time scales. lf one is srudying an
EPR active species (paramagnetic) such as Mn, this technique has severa!
advantages over other chemical relaxation methods. With m::my relaxation
merhods, the reacrions must be reversible and reacranr species are not directly
measured, Moreover, in sorne relaxarion studies, rhe rate constanrs are calculated from linearized rate equarions that are dependent on equilibrium parameters. Thus, rhe rate paramerers are not directlv measured.
With the EPR -SF method of Fendorf et al. ( i 99 3) the mixing can be done in
< 1O ms and EPR digitized wirhin a few microseconds. A diagram of rhe
EPR-SF instrumenr is shown in Fig. 7.10. Dual 2-ml in-porr syringes feed a
mixing cell that is located in the EPR spectrometer. This allows for EPR
detection of rhe cell contents. A single outflow port is firred with a 2-ml
effluenr collection syringe equipped with a triggering switch. The switch activares the data acquisition system. Each run consisrs of filling the in-port
syringes with the desired reactants, flushing rhe sysrem wirh rhe reacrants
severa! times, and initiating and moniroring rhe reacrion. Fendorf et al. ( 199 3)
used this system ro study the kinetics of Mnz+ sorption on y-MnO". The
sorprion reaction was complete in 200 ms. ,Data were raken every 50 fl-5 and
lOO points were averaged to give the time-dependent sorption of !v!n(II).

Choice of Kinetic Method

The method that one chooses to study the kinerics of soil chemical reacrions
depends on severa! facrors. The reaction rate will certainly dicta te the choice of
merhod. Wirh batch and flow merhods, rhe most rapid measuremenrs one can

make require about 15 s. Por more rapid reacrions, one musr use relaxarion
techniques where millisecond and microsecond time scales can be measured.
Anorher factor in deciding on a kinettc merhod is the objecrive of one's
experiments. If one wishes to measure rhe chemical kinetics of a reaction
where transpon is mnima!, mosr batch :md flow techniques are unsuirable
and a relaxarion technique should be employed. On rhe other hand, if une
wants ro simulare time-dependent reactions in the field, perhaps a flow technique would be more realisric than a batch merhod.

EFFECT OF TEMPERATURE ON REACTION RATES

Temperature has a marked effect on reacrion rare. Arrhenius noted rhe
following relationship between k and T;
(7.19)
where Ar is a frequency factor and E, is the energy of acrivation . Converting
Eq. (7.19) ro a linear form resulrs in
In k = In Ar - E,!RT.

(7.20)

A plot of In k vs 1/Twould yield a linear relationship with the slope equal to

- E,IR and rhe intercept equal to In Ar. Thus, by measuring k values ar severa!
temperatures, one could determine the E, value.
Low E, values (<42 kJ mol-'} usually indicare diffusion-controlled rransport processes, whereas higher E, values would indicare chemical reaction or
surface-controlled processes (Sparks, 1985). For example, E, values of 6.726.4 kJ mol- 1 were observed for pesticide sorption on soils and soil components (Haque et al., 1968; Leenheer and Ahlrichs,. 1971; Khan, 1973), while
gibbsire dissolution in acid solutions, which appeared to be a surface-controlled reaction, was characterzed by E., values ranging from 59 :2:: 4.3 to
67 ::!: 0.6 kJ mol-' (Bloom and Erich, 1%7).

176

CHAPTER 7 KINETICS Of SOil CHEMICAL P~OCESSES

KI'NETICS OF IMPORTANT SOIL CHEMICAL PROCESSES

KIHETICS Of IMPOHANT lOil CHEMICAL

ters, a first-order reaction was derived (Zhang and Sparks, 1990b)

Adsorption-Desorption Reactions

_ 1
T

Heary Metals

A number of studies have been conducted on the adsorption/desorption

kinetics of heavy metals on soils and soil components. Bruemmer et al. (1988)
hypothesized that adsorption of Ni 2 +, Zn 2 +, and Cd 2 + on goethite occurred on
both externa! and interna! surface sites. As reaction time increased from 2 hr
to 42 days at 293K and pH 6, the adsorbed Ni2+ increased from 12 ro 70% of
total adsorption, and total increases in Zn2+ and Cd 2 + adsorption over rhis
point increased 33 and 21%, respectively. The irtcreased adsorprion wirh rime
is consistent with the assumption of continued adsorption on interna! sites
within the porous structure of goethite, which could be a diffusion-controlled
process.
Zhang and Sparks (1990b) studied the kinetics of selenate adsorption on
goethite using pressure-jump relaxation and found that adsorption occurred
mainly under acidic conditions. The dominant species was (Se0 4 ) 2 - . As pH
increased (Se0 4 ) 2 - adsorption decreased. Selenate was described using the
modified triple-layer model (see Chapter 5). A single relaxation was observed
and the mechanism proposed was:
(7.21)
where XOH is 1 mol of reactive surface hydroxyl bound ro a Fe 10n m
goethite.
A linearized rate equation given below was developed and tested,
T-

1 = k 1 ([XOH][Se0.- 2 ] + [XOH][W] + [Seo-][W]) +k_, ,

(7.22)

where the terms in the brackets are the concentrations of species at equilibrium . Since the reaction was conducted at the solid/liquid interface, the electrostatic effect has to be considered ro calcula te the intrinsic cate constants (k't"'
:.md k~\). Using the modified triple-layer model to obtain dectrostatic parame200

::::a:
~

150

"'.:..

.:..

Lt

*
X

Cll

100

50

.. r2 = 0 .9973

~..

0.2

0.4

0.6

0.8

1.0

exp( -F (1J1 a -2'V~)/RT) ([XOH][Seo~] +

[XOHJlH'J+(SeoaJlw])x1o7

177

P~OCEllEl

FIGURE 7.11. Plot of relatlonship between T " 1 with exponencial and concentration terms in Eq.
(7.23). Reprinted wich permission from Zhang and Sparks (1990a). Copyright 1990 American Chemical
Sociecy.

exp

(-F(t/Ja -2t/J{3))-kinr[ 2 RT

((XOH][SeO~-) +

exp

(-f(t{!n -2t/J{3))
RT

[XOH][W] + [Seo - ][W])

k~'1

(7.23)

A plot of the left si de of Eq. (7.23) vs the terms in brackets ori the right si de of
Eq. (7.23) was linear and the k'" and k'~'1 values were calculated from the slope
and intercept, respectively (Fig. 7.11) . The linear relationship would indicare
that the outer-sphere complexation mecha nism proposed in Eq. (7.21) was
plausible. Of course, one would need to use spectroscopic approaches to
definitively determine the mechanism. This was done earlier with x-ray absorption fine structure spectroscopy (XAFS) ro prove that selenate is adsorbed
as an outer-sphere complex on goethite (Ha ves et al., 1987).
Organic Contaminants
There have been a number of srudies on rhe kinetics of organic chemical
sorption/desorption with soils and soil components. Many of these investigarions ha ve shown that sorptionldesorption is characterized by a rapid, reversible stage followed by a much slower, nonreversible stage (Karickhoff et al.,
1979; DiToro and Horzempa, 1982; Karickhoff and Morris, 1985) or biphasic
kinetics. The rapid phse has been ascribed ro retention of the organic chemical in a labile form rhat is easily desorbed. However, the much slower reaction
phase involves the entrapment of the chemical in a nonlabile form that is
difficult ro desorb . This slower sorptionldeso rption reacti6n has been ascribed
ro diffusion of the chemical into micropores of organic matter and inorganic
soil components (Wu and Gschwend, 1986; Sreinberg et al., 1987; Ball and
Roberts, 1991). The labile form of the chemical is available for microbial
attack while the nonlabile portion is resisra nt to biodegradation .
An example of the biphasic kinetics that is observed for many organ:c
chernical reactions in soils/sedirnents is shown in Fig. 7.12. In this srudy 55%
of the labile polychlorinated biphenyls (PCBs) was desorbed from sediments in
a 24-hr period, while little of the remaining 45% nonlabile fraction was
desorbed in 170 hr (Fig. 7.12a) . Over anorher 1-year period about 50% of rhe
remaining nonlabile frattion desorbed (Fig. 7.12b).
In another study wirh volatile organic compounds (VOCS), Pavlostathis
and Mathavan (1992) observed a biphasic desorption process for field soils
contaminated with trichloroerhylene (TCE l, tetrachloroethylene (PCE), roluene (TOL), and xylene (XYL). A fast desorption reaction occurred in 24 hr,
followed by a much slower desorption reacr.ion beyond 24 hr. In 24 he, 9-29,
14-48, 9-40, and 4-37% of the TCE, PCE, TOL, and XYL, respectively,
were released.
A number of srudies have. also shown thar with " aging" the nonlabile
portian of the organic chemical in the soil!sediment becomes more resistant ro
release (McCall and Agin, 1985; Steinberg et al., 1987; Pavlosrathis and
Mathavan, 1992; Scribner et al., 1992; Pignare!lo et al., 1993). However,
Connaughton et al. (1993) did not observe rhe nonlabile fraction increasing
with age for naphthalene-contaminared soi ls.

- - - - - - --

- - - -- - -

178

CHAPTER

-a

.Q

'
~
u.

Sorption Distribution
Coefficients for Herbicides in "Freshly
Aged" and "Aged" SoiiS"

Herbicide

Soil

~-!erolachlor

CVa
CVb
Wl

:g

t 0.4
~

u.
0.2

0.2

W2

O+-~r-~~-,~-r---r~,-~~~r4

20

40

60

80

100

120

Desorption Time. h

K:

....

K'

e:

0.4

co
0.6
a.

co
~ '\)6:

TABLE 7.3.

0.8

179

KINETICI Of IHPORTANT IOIL CHEHICAL PROCEISEI

0.8

KINETICS Of IOIL CHEHICAL HOCEIIEI

140 160

6
8
10
Desorption Time. mo

:\trazine

12

FIGURE 7. 12.

(a) Short-term PCB desorption in hours (h) from Hudson River sediment contaminated with 25 mg kg- 1 PCB. Distribution of the PCB between che sediment () and XAD-4 resin (O) is
shown, as well as the overall mass balance (t.) . The resin acts as a sink to retain the PCB that is
desorbed. (b) Long-term PCB desorption in months (me) from Hudson River sedime~t contaminated
with 25 mg kg' 1 PCB. Distribution of che PCB between the sediment () and XAD-4 re<ln (e) is shown.
The line represencs a nonlinear regression of the data by the two-box (site) model. [Reprinted with
permission from Carroll et al. ( 1994). Copyright 1994 American Chemical Society. J

One way to gauge the effect of time on organic contaminant retention in

soils is to compare Kd (sorption distriburion coefficient) values for "freshly
aged" and "aged" soil samples (see Chapter 5 for discussion of rhese coefficients). In most studies, Kd values are measured based on a 24-hr equilibration
berween the soil and rhe organic chemical. When these values are compared to
K, values for field soils previously reacted with the c~ganic chemical (aged
samples) rhe latter have much higher K values, indicating rhat much more of
the organic chemical is in a sorbed state. For example, Pignatello and Huang
(1991) measured K.1 values in freshly aged (Kd) and aged soils (K, 00 , apparent
sorption distriburion coefficient) reacted with atrazine and metolachlor, two
widely used herbicides. The aged soils had been treated wirh the herbicides
15-62 months befo re sampling. The K,PP values were 2.3-42 times higher
than the K values (Table 7.3) .
Scribner et al. (1992), studying simazine (a widely used triazine herbicide
for broadleaf and grass control in crops) desorption and bioavailability in aged
soils found rhat K,PP values were 15 times higher than KJ values. Scribner et al.
( 1992) also showed that 48% of the simazine added ro the freshly aged soils
was biodegradable over a 34-day incubation period while none of the simazine
in the aged soil was biodegraded.
One of the implications of these results is that while many transpon and
degradation models for organic contaminants in soils and waters assume rhar
the sorption process is an equilibrium process, the above studies clearly show
that kinetic reactions must be considered when making predictions about rhe
mobility and fate of organic chemicals. Moreover, K, values that are determined based on a 24-hr equilibration period and which are commonly used in
fate and risk assessment models, can be inaccurate since 24-hr Kd values often
overestimare the amounr of organic chemical in the solution phase.
The finding that many organic chemicals are quite persistent in the soil
environmenr has borh good and bad features . The beneficia! aspect is that the
organic chemicals are less mobile and may nor be readily transponed in

CVa
CVb
WJ

2.96
1.46
1.28
0.77

39

2. 17

28

1.32

29

1.75

27

49
33

' Adapted from Pignarello and Huang (1991)

with permission; Herbicides had been added to
soils 31 months prior to sampling for CVa and
CVb soils, 15 monrhs for rhe Wl and W2 soils,
and 62 months for the W3 soil.
Sorption disrribution coefficient (liter kg - 1 ) of
" freshly aged" soil based on a 24-hr equilibration period.
' Apparenr sorprion distribution coefficient (iirer
kg- 1) in contaminated soil ("aged" soil) determined using a 24-hr equilibration penod.

groundwater supplies. The negative aspect is that their persisre:1ce and inaccessibility ro microbes may make decontarnination more difficult, parttcul a rhif in situ remediation techniques such as bi )degradation are employed.

Ion Exchange Kinetics

Ion exchange kinetics are greatly dependent on the adsorbent and the io n .
Figure 7.13 clearly shows how rhe type of clay mineral affects K-Ca ex change. Reaction rates are much more rapid on bolinite where the exchange
sites are externa! than ort vermiculite that contains both externa! and interna!
exchange sites. The interna! sites may be fully expanded or partiall y collapsed.
1400

'i:n 1200
~
e

1000

.2

800

600

'O

400

200

c.
1/)

<l:

20 40 60 80 100 120 140 160 180 200 220 240 260

Time, min

FIGURE 7 . 13.

Potassium adsorption versus time for clay minerals. O . Kaolinice;

ite; .l. vermiculice. From Sparks and Jardine ( 1984). witi1 permission.

e. montmorillon-

180

CHAPm 1 KINETICS Of IOil CHEHICAl PROCEHES

The type of ion also has a pronounced effect on the rate of exchange.
Exchange of ions like K., NHt, and Cs+ is often slower than that of ions such
as Ca2+ and Mg2+. This is related to the smaller hydrated radius of the former
ions. The smaller ions fit well in the interlayer spaces of clay minerals, which
causes parcial or total inrerlayer space collapse. The exchange is chus slow and
particle diffusion-conrrolled. However, with the exception of K, NHt, and
Cs+ exchange on 2:1 clay minerals like vermiculite and mica, ion exchange
kinetics are usually very rapid, occurring on millisecond time scales (Tang and
Sparks, 1993). Figure 7.14 shows that Ca-Na exchange on monrmorillonite
was complete in < 100 ms.

is the reaction rate constant. Inregrating,

Dissolution of minerals involves severa! sreps (Stumm and Wollast, 1990):

(1) mass transfer of dissolved reactants from the bulk solution to the mineral
surface, (2) adsorption of salutes, (3) interlattice transfer of reacting species,
(4) surface chemical reactions, (5) removal of reactams from rhe surface, and
(6) mass transfer of products inro the bulk solution. Under field conditions
mineral dissolution is slow and mass transfer of reactants or produces in che
aqueous phase (Steps 1 and 6) is not rate-limiting. Thus, the rate-limiting steps
are either transport of reactants and produces in the salid phase (Step 3) or
surface chemical reactions (Step 4) and removal of reactants from the surface
(Step 5).
Transpon-controlled dissolution reactions or those controlled by mass
transfer or diffusion can be described using the parabolc rate law given below
(Stumm and Wollast, 1990);

de

r = - = ktdt
where r is the reaction rate,

tt2

(7.24)

'

e is the concenrration in solution, t is time, and k

esurl~b

1.4

Cl)

"'en
e e:
o.,
U .e
.. u

1.0

Qi

0.4
0.2
0.025

FIGURE 7.14.

Surface controllcd

ebulk

distance

bulk

distance

el ''
0.075

0.125

0.175

0.225

0.275

Time,s
-

0.8
0.6

"'
a:

Montmorillonite = 10.2 g L1
lonic Strength = 0.01 M
pH =6.8

1.2

(7.26)

and r is proporcional to the mineral's surface area, A. Thus, for a .surface~con

trolled reaction the relationship between time and e should be lmear< Ftgure
7.15 compares transport- and surface-controlled dissolution mechanisms.
Incense arguments ha ve ensued over che years concerning the mechanism for
mineral dissolution. Those that supported a transport-controlled mechanism
believed that a leached !ayer formed as mineral dissolution proceeded and that
subsequent dissolution took place via diffusion through the leached layer
(Petrovic et al., 1976). Advocates of this theory found that dissolution wa ~
described by the parabolic rate law [Eq. 7.24)]. However, the "apparenc"
transport-controlled kinetics may be an artifact caused by dissolution of hyperfine particles, formed on the mineral sur faces after grinding, that are highly
reactive sites or bv use of batch methods that cause reaction produces to
accumulate, causing precipitation of secondary minerals. These experimental
artifacts can cause incongruent reactions and pseudoparabolic kinetics. Recent
studies employing surface spectroscopies such as x-ray photoelectron spectroscopy and nuclear resonance profiling (Schort and Petit, 1987; Casey et al.,

Transport controlled

o::

(7.25)

.
where e0 is the initial concentration in solution.
If the surface reactions are slow compared to the transport reactwns,
dissolution is surface-controlled, which is the case for most dissolution reactions of silicates and oxides. In surface-conrrolled reactions the<concentrations
of salutes next to the surface are equal to thc bulk solution concenrrations and
the dissolution kinetics are zero-order if steady state conditions are operacional
on the surface. Thus, the dissolurion rate, r, is

dC
r =-=kA
dt
'

Rate-limiting Steps

:~

e increases with .t112 ,

e = e0 + 2kt 112

Kinetics of Mineral Dissolution

181

KINETICS Of IHPO~TANT IOil CHEHICAl PROCESm

Typical pressurejump relaxation curve for Ca-Na exchange on montmorillonite

showing relative change in conductivity vs. time. From Tang and Sparks (1993), with permission.

e= e

Rate = kt',.,
0 +2kl 1,.,

FIGURE 7.15.

~lme

e=

Rate = k (suriace a real

e 0 ~kt

Transport- vs surface-controlled dissolution. Schematic represenution of concencra.tion in solution, C. as a function of distance from thf~ surface of the dissolving mineral. In the lower
part of the figure. the cha~ge in concentration is given as a function of time. From Scumm. W. ( 1992).
Chemistry of the Salid-Water Interface. Copyright 1992 John Wiley & Sons. lnc. Reprinted by permtS
sion of John Wiley & Sons. lnc.

T
" 1.82

CHAPTEk 1 KINETICS Of IOIL CHEMICAL PkOCEIIEI

1989) have dernonstrated that although sorne incongruency rnay occur in the
initial dissolution process, which rnay be diffusion-controlled, the overall
reaction is surface-controlled. An illustration of the surface-controlled dissolution of y-Al 2 0 3 resulting in a linear release of A[3+ with time is shown in Fig.
7.16. The dissolution rate, r, can be obtained frorn the slope of Fig. 7.16.

'

Surface sites + reactants (H +, OH-, or ligands) ~ Surface species (7.27)

<low
Je.,chmen of

M (aq).

(7.28)

meui(Ml

Thus, the attachment of the reactants to the surface sites is fast and de~ach
rnent of metal species frorn the surface into solution is slow and rate-limiting.
Ligand-Promoted Dissolution

Figure 7.17 shows how the surface chernistry of the mineral affects dissolution. One sees rhat surface protonation of rhe surface ligand inci"eases dissolution by polarizing imeratornic bonds close to rhe central surface ions, which
prometes the release of a carion surface group into solution. Hydroxyls that
bind to surface groups ar higher pHs canease the release of an anionic surface
group into the solution phase.
Ligands rhat forrn surface cornplexes via ligand exchange wth a surface
hydroxyl add negative charge to the Lewis acid center coordinarion sphere,

pH= 2.5

_16
~~ 14

3.0
3.5

12

"' 10
b

4.0
4.3

~ 2
30
Time, h

20

40

FIGURE 7.16. linear dissolution kinetics observed for the dissolution of y-AI,O,. Representative
of processes whose rates are controlled by a surface reaction and not by transport. Reprinced from
Furrer, G .. and Stumm. W. ( 1986). The coordination chemistry uf weathering. Geochim. Cosmochim. Acta
SO, 1847-1860. Copyright 1986, with kind permission from Elsevier Science Ltd .. The Boulevard.
Langford Lane. Kidlingcon OXS 1GB, UK.

OH

M/

'

M/

'o 'o

)-

o
M/

'

o- ] 2 -

' o- '
/

'

M/

OH

Surfsce complex

Surface protonahon

rormation w1th bidentate (mononuclear) tigands.

e.q .. oxalate, sati

cylate. citrate. di
phenois. etc.

/ OH,

'

OH

' o' "

"
o/

"

Surfaee deprotonalion

Sur ace spectes

lnhibltlon of dlssolutlon

Enhancement ot dlssolutlon

Surface-Controlled Dissolution Mechanisms

Dissolution of oxide rninerals via a surface-controlled reaction by ligandprornoted and proton-promoted processes has been described bv Sturnrn and
co-workers (Furrer and Sturnrn, 1986; Zinder et al., 1986; Sturn~ and Furrer,
1987) using a surface coordination approach. The irnportant reactants in these
processes are H 1 0, H ... , OH-, ligands, and reductants and oxidants (see
definitions in ehapter 8). The reaction rnechanisrn occurs in rwo steps (Srurnm
and Wollast, 1990):

183

KIHETICS Of IHPOmNT IOIL CHEHICAl HOCEIIEI

OH

M/

'

/
M

OH

"

CH 1 ~(CH 1 1.,

COOH

COOH

OH

/
M

"o "
/

CH 3 ~(CH1)

OH

OCr*

M/

' o "o
/

Surtace comolex
lonnation tooi-or
multinuclear com
plexes or surtace

B!ockinq of surtace groups by hydrophobic moietles

of fatty ac1ds. humic acids or macromolecutes

mms. blockage ol
surface groups by
metal eations

FIGURE 7.17.

The dependence of surlace reactivity and of kinetic mechanisms on the coordinative

environment of che surface groups. From Stumm and Wollast (1990). Rev. Geophys. 28, 53 - 69.
Copyright by the American Geophysical Union.

and lower rhe Lewis acid acdity. This polarizes the M-oxygen bonds, causng
derachrnent of the metal caton into the ;olution phase. Thus, inner-sphere
surface complexation plays an importanr role in mineral dissolution. Ligands
such as oxalate, salicylate, F- , EDT A, and NTA increase dissolution but
orhers, e.g., so- , ero-, and benzoate, inhibit dissolution. Phosphate and
arsenate enhance dissolurion at low pH and dissolurion is inhbired at pH > 4
(Sturnm, 1992).
The reason for these differences may be that bidentate speces rhar are
rnononuclear promete dissolution while binuclear bidentate species inhibit
dissolution. With binuclear bidentate cornplexes, more energy rnay be needed
ro rernove two central atorns frorn the crystal structure. With phosphate and
:usenate, at low pH mononuclear species are formed, while at higher pH
(around pH 7) binuclear or trinuclear sur face cornplexes forrn. Mononuclear
bidentare cornplexes are forrned with oxa late while binuclear bidentate cornplexes forro with ero~-. Additionally, the electron donor properties of erOiand oxalate are also different- With ero- a high redox porennal ts rnamrained at the oxide surface, which restricrs reducrive dissolurion (Sturnrn and
Wollast, 1990; Srurnrn, 1992).
Dssolution can also be inhibited by cations such as Y0 2 , er(IIl) and
Al(Ill) that block surface functional groups.
One can express rhe rare of the ligand-prornoted dissolution, RL, as.
(7.29)

wherek~ is the rate constant for ligand-p romoted disso.lution (rirne- 1 ), ==ML

--------------~------------~----------------------~------------~-----1

r.
'

. 184

CHAPTER 7 KINETICS Of SOIL CHEMICAL PROCESIES

o

><~o

..
:r::

>~o

<:'E

"1

...

a:

KIHETICS Of IMPORTANT SOIL CHEMICAL HOCEISES

18,5 ,.

can be expressed as (Stumm, 1992)

RH

= k~ (=MOH!)i =

k~ (C'H)i,

(7.30)

where k~ is the rate constant for proron-promoted dissolution, =MOHi is

the metal-proron complex, CH is the concentration of the surface-adsorbed
proton complex (mol m- 2 ), and corresponds to the oxidation state of the
central metal ion in the oxide srructure (i.e., i = 3 for Al(III) and Fe(III) in
simple cases). If dissolution occurs by only one mechanism ; is an inreger.
Figure 7.19 shows an application of Eq. (7.30) for the proton-promored
dissolurion of y-Al 2 0 3
Overall Dissolution Mechanisms

e~. [10-6 mol m2]

FIGURE 7. 18.

The rate of ligand-catalyzed dissolution of y-AI 10 1 by the aliphatic ligands oxalate.

malonate. and succinate, Rl (nmol m 1 h " 1), can be inrerpreted as a linear dependence on the surface
concentrations of the ligand complexes, q. In each case the individual values for q were determined
experimentally. Reprinted from Furrer, G., and Stumm, W. ( 1986). The coordination chemistry of
weathering. Geoch;m. Cosmoch;m. Acta 50, 1847-1860. Copyright 1986, with kind permission from
Elsevier Science Ltd .. The Boulevard, langford lane. Kidlington OXS 1GB. UK.

is the metal-ligand complex, and q is the surface concentration of the ligand

complex (mol m- 2 ). Figure 7.18 shows that Eq. (7.29) adequately described
ligand-promoted dissolution of y-Al 20 3
Proton-Promoted Dissolution

Under acid conditions, prorons can promote mineral dissolution by binding

to surface oxide ions, causing bonds ro weaken. This is followed by detach
ment of metal species into solution. The proton-promoted dissolution rate, RH,

-8.0
-8.2
:.:8.4

a:

t:n

.S! -8.6
-8.8
log e~

9.0

-6.0, 5,8 (5,6 . 5.4.,

6 5

4 3

pH

FIGURE 7.19. The dependence of the rate of protonpromoted dissolution of y-AI 10 1, R,., (mol
m' h'). on the surface concentration of the proton complexes, C',.,(mol m'). Reprinted from Furrer, G..
and Stumm, W. ( 1986). The coordination chemistry of weathering. Geadtim. Cosmachim. Acta 50,
1847-1860. Copyright 1986. with kind permission from Elsevier Science Ltd .. The Boulevard. Langford
Lane. Kidlington OXS 1GB. UK.

The rate of mineral dissolution, which is the sum of the ligartd-promored,

proron-promoted, and deproronarion-promoted (or bonding of OH- ligands)
dissociation [RoH = k~H(CbH)i] rates, along with the pH-independent portion
of the dissolution rate (k~, 0 ), which is due to hydration, cart be expressed as
(Stumm, 1992)
(7.31)

Equation (7.31) is valid if dissolution occurs in padtllel at varying metal

centers (Furrer and Stumm, 1986).

Suggested Reading
Lasaga, A. C., and Kirkpatrick , R. J., eds. ( 1981 ). "Kinerics of Geochemical Processes." Mineralogical Sodety of America, Washington, DC.
Liberti, L., and Helfferich, F. G., eds. (1983) . "M.tss Transfer and Kinetics of Ion Exchange ."
NATO ASI Ser. E; No. 71 , Martinus Nijhoff, The HJgue, The Netherlands.
Sparks, D. L. (1985). Kinetics of ionic reacrions 10 clay minerals and soils. Adv. Agron. 38,
231-266.
Sparks, D. L. (1986). Kinetics of reactions in pure ctnd mixed systems. In " Soil Phvsical Chemistry." (D. L. Sparks, ed.), pp. 83-145. CRC Press, Boca Raton, FL.
Sparks. D. L. (1989). "Kinetics of Soil Chemical Processes. " Academic Press, San Diego, CA.
Sparks, D. L. (1992) . Soil kinetics. In "Encydopedi:t of Earth Systems Science" (\VJ. A. Nirenberg,
.
ed.), Vol. 4, pp. 219-229. Academic Press, San Diego, CA .
Sparks, D. L.. and Suarez, D. L., eds. ( 1991 ). "Rates of Soil Chemical Processes'. SSSA Spec. Pub l.
No. 27, Soil Sci. Soc. Am., Madison, Wl.
Srumm , W., ed. ( 1990). "Aquatic Chemical Kinetie>. " Wiley, Kew York.