Anda di halaman 1dari 12

Chemical Geology 151 Ž1998.

247–258

General implications of aluminium speciation-dependent kinetic


dissolution rate law in water–rock modelling
´ ´ Gerard
Frederic ´ ),1
´
, Bertrand Fritz, Alain Clement, Jean-Louis Crovisier
´
Centre de Geochimie de la Surface, CNRS UPR 6251. 1, rue Blessig, 67084 Strasbourg, France

Abstract

Recent experimental and theoretical work has demonstrated that the dissolution rates of many aluminosilicate minerals
are inversely proportional to the activity of aqueous Al 3q. The consequences of these observations on the rates of natural
geochemical processes have been calculated by the KIRMAT hydrochemical code. Comparisons are performed at the steady
state limit of the pure advective transport through a homogeneous semi-infinite isothermal porous media at 25 and 1508C.
K-feldspar, albite, and muscovite dissolution kinetics are studied over a broad range of initial pH Ž2–10. and aluminium
concentration Žfrom 1 = 10y9 to 1 = 10y3 molal. matching most the natural conditions. Regardless of the mineral, the
characteristic distance requires to reach equilibrium Ž l eq . is two and three orders of magnitude larger and lower than
predicted using the standard Transition State Theory ŽTST. law, respectively. The maximum decrease in muscovite and
alkali-feldspar dissolution rates due to aqueous aluminium at 258C is found at near to neutral pH, and at 1508C it is found at
basic pH. The maximum dissolution rate increase at 258C at acid pH, but at 1508C it is found at basic pH. These results
demonstrate that consideration of the effect of the aluminium speciation on aluminosilicate dissolution rates is required to
improve the accuracy in water–rock interaction modelling. q 1998 Published by Elsevier Science B.V. All rights reserved.

Keywords: Reactive transport; Modelling; KIRMAT; Kinetic dissolution; Transition State Theory; Aluminium speciation

1. Introduction rate laws for mineral dissolution and growth Že.g.,


Coudrain-Ribstein, 1988; Steefel and Lichtner, 1994;
For several years, much effort has been placed on Steefel and Lasaga, 1994; White, 1995..
hydrochemical or reactive transport computer mod- Far more silicate dissolution experiments have
elling in porous media. A large number of programs been performed than mineral growth experiments.
simulating coupled chemical and mass transport in Silicate dissolution kinetics are surface controlled to
porous media are currently available. The most ad- hydrothermal temperatures. According to Transition
vanced of these hydrochemical codes employ kinetic State Theory ŽTST. ŽEyring, 1935., the rate limiting
step of the overall reaction is the formation of an
activated complex at high energy surface sites
)
ŽLasaga, 1981; Aagaard and Helgeson, 1982; Helge-
Corresponding author. Fax: q33 3 83 39 40 69; e-mail:
gerard@nancy.inra.fr
son et al., 1984.. It is commonly assumed that the
1
Present address: INRA, Centre de Nancy, Equipe Cycles concentration of activated complex is proportional to
´
Biogeochimiques, F-54280 Champenoux, France. a precursor complex at the mineral surface and both

0009-2541r98r$ - see front matter q 1998 Published by Elsevier Science B.V. All rights reserved.
PII: S 0 0 0 9 - 2 5 4 1 Ž 9 8 . 0 0 0 8 3 - 7
248 ´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258

complexes are in state of local equilibrium ŽWieland affinity dependence. Given alkali-feldspars, and to a
et al., 1988; Guy and Schott, 1989.. Until recently, lesser extent muscovite, are ubiquitous rock-forming
precursor complexes for multi-oxide minerals were minerals, most water–rock reaction paths will be
believed to consist of either hydrogenated, hydroxe- modified by these effects.
nated, or hydrated surface sites. Taking account of The aluminium speciation-dependent dissolution
these assumptions, TST dictates that at far-from- rate equation has been introduced into the KIRMAT
equilibrium, dissolution rates only depend on pH. ´
reactive transport code ŽGerard et al., 1996, 1997;
The kinetic effects of solution pH on silicate dissolu- ´
Gerard, 1997., which also contains the standard ki-
tion has been thoroughly documented ŽHelgeson et netic rate expression. This innovation permits the
al., 1984; Knauss and Wolery, 1986, 1989; Tole et study of effects of the aluminium speciation depen-
al., 1986; Murphy and Helgeson, 1987; Carroll and dent rate law in hydrochemical modelling. The large
Walther, 1990; Acker and Bricker, 1992; Hellmann, number of factors that now influence aluminosilicate
1994.. The abundance of data has driven most hy- dissolution rates suggest its effects are case depen-
drochemical and geochemical code developers to dent. The goal of this study is to estimate the upper
implement the first order, pH-dependent, rate expres- and lower boundaries of the effect of aqueous alu-
sion in their codes to forecast silicate dissolution. minium on dissolution rates. These results allow
Nonetheless, except for the silica phases ŽRimstidt evaluation of the trends and potential implications of
and Barnes, 1980., the first order approach of the this dissolution rate law on water–rock systems.
overall dissolution rate toward equilibrium had not
been experimentally validated until recently. More-
over, other catalyticrinhibitor effects due to dis-
2. Theory
solved inorganic and organic species have been
pointed out for several years ŽWogelius and Walther,
1991; Amrhein and Suarez, 1992; Welch and Ull- 2.1. Kinetic equations
man, 1993; Bennett and Casey, 1994; Stillings et al.,
1996.. However, they provide no mechanistic kinetic Mineral dissolution rate laws are commonly de-
expression or their applicability range is too narrow rived from TST, assuming both that Hq, H 2 O, and
to be effective in water–rock interaction modelling. OHy are the only aqueous species involved in the
Recently, a new mechanistic kinetic rate law con- reaction forming the activated complex, and that one
sistent with TST and surface chemistry was proposed mole of activated complex is formed from each mole
in the literature ŽGautier et al., 1994; Oelkers et al., of mineral. These assumptions lead to the following
1994; Schott and Oelkers, 1995.. This new rate law equation:
takes account of the inhibitor effect of dissolved
aluminium on alkali-feldspars, kaolinite, muscovite, A Ea

and kyanite dissolution. In addition, alkali-feldspars


and kaolinite apparently have a larger dissolution
ž
rd s k d SaHn q 1 y exp y ž RT // A expy RT . Ž 1.

rate dependence on chemical affinity than previously With


assumed. A s yRT ln Ž QrK . Ž 2.
Several convincing arguments lead us to believe
that the new rate law might significantly affect the where rd is the dissolution rate Žmol sy1 ., k d the
chemical evolution of natural water–rock systems. dissolution rate constant Žmol cmy2 sy1 ., S the
Above all, aqueous aluminium effects appear strong reactive surface area of the mineral, a H q the in-
enough to induce an apparent chemical affinity de- hibitorrcatalytic term for pH effects and n is an
pendence on rates at far-from-equilibrium conditions. experimental exponent, A denotes the chemical
Moreover, dissolved aluminium is common in aque- affinity Žkcal moly1 ., R the gas constant Ž1.99 =
ous solutions and has a significant concentration 10y3 kcal moly1 Ky1 ., A the pre-exponential factor
range. Finally, the rate of approach toward equilib- of the Arrhenius rate law, Ea the apparent activation
rium is intrinsically slowed due to its larger chemical energy of the reaction Žkcal moly1 ., which is a
´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258 249

function of solution pH. Finally, Q is the ionic namic constant of the precursor formation reaction.
activity product ŽIAP. of the mineral and K its The rate law derived from the above formation reac-
thermodynamic equilibrium constant. tion is given by Žfor details see Oelkers et al., 1994.:
According to TST, the exponent n is equal to the
number of protons required to form the activated a 3Hmq
complex ŽHelgeson et al., 1984; Murphy and Helge-
son, 1987.. The rate constant k d and the exponent n
usually have different values depending on pH re-
gion: pH 1 delimits the acid from the neutral pH
region and pH 2 the neutral from the basic pH region.
rd s k d S
− <
1q
m
aAl 3q

K8a H3 mq
m
aAl 3q

A Ea
This pH dependence leads to the well-known U-shape
function. Many efforts have been made to relate pH 1
with the pH of the point of zero net proton charge,
ž
= 1 y exp y ž s RT // A expy RT Ž 4.

commonly noted pH pznpc Žsee for example, Blum and where s is equal to the number of silica mole per
Stillings, 1995; Hochella and Banfield, 1995.. Al- mole of aluminosilicate. The aqueous activity term
though some consistent results are obtained on sim- controls the dependence of the dissolution rate on
ple oxides such as SiO 2 Že.g., Brady and Walther, aluminium speciation at far-from-equilibrium condi-
1989, 1990; Bennett, 1991., inconsistent results have tions; at close to equilibrium conditions the effects of
been obtained for more complex oxides such as chemical affinity become important. When the num-
alkali-feldspars and kaolinite Žsee for instance the ber of precursor complexes at the mineral surface is
3m

discussion in Chen and Brantley, 1997.. Moreover, relatively small, K8am H < 1 and Eq. Ž4. can be
q

aAl 3q
as proposed by several authors Že.g., Chou and Wol- rewritten in terms of total dissolved aluminium con-
last, 1985; Garnor et al., 1995., the temperature centration Žnoted wAlx. and pH ŽOelkers et al., 1994.:
dependence of the exponent n Žsee Carroll and ym
Walther, 1990; Hellmann, 1994; compare Chou and rd s k d S w Al x Ž a3H q b 3 a2H qq b 2 aH q
q

Wollast, 1985 to Knauss and Wolery, 1986. might


m A
be due to the presence of some unrecognized in-
hibitor effects thus pointing out the failure of Eq. Ž1.
qb4 q b 1 ay1
Hq . ž 1 y exp yž s RT // . Ž 5.
to explain the dissolution kinetic of the complex Where b 1 is the thermodynamic equilibrium con-
oxides studied. Although this deviation from the stant for AlŽOH.y 3q
4 formation reaction from Al and
standard theory can be understood with the help of q
4H 2 O, b 2 for AlŽOH. 2 formation from Al 3q
and
surface chemistry Žsee Brady and Walther, 1992; 2H 2 O, b 3 for AlŽOH. 2q formation from Al 3q and
Casey and Sposito, 1992., both discrepancies can be H 2 O, and finally b4 for AlŽOH. 38 formation from
explained by taking into account the effect of dis- Al 3q and 3H 2 O.
solved aluminium. The presence of an aluminium-
deficientrsilica-rich rate controlling precursor com-
plex was considered by Gautier et al. Ž1994.; Oelkers 2.2. The code KIRMAT
et al. Ž1994., Oelkers and Schott Ž1995., Schott and
Oelkers Ž1995., and Oelkers Ž1996.. It was proposed The hydrochemical code KIRMAT ŽGerard ´ et al.,
that this precursor complex is formed by the follow- ´
1996, 1997; Gerard, 1997. has been developed by
ing exchange reaction: adding provision for chemical mass transport into the
thermodynamic and kinetic geochemical code
K8 KINDISP ŽMade´ et al., 1994.. KINDISP’s thermo-
3 m Hqq M–O m P8 q m Al 3q Ž 3. dynamic database contains the data of Helgeson et
al. Ž1978. and is used in the KIRMAT hydrochemi-
where M–O represents a potentially reactive surface cal code. The dissolution rate laws ŽEqs. Ž1. and Ž4..
site Žmetal–oxygen., m is a stoichiometric coeffi- are encoded in KIRMAT, in addition to other kinetic
cient, P8 the precursor species, and K8 the thermody- formulations for dissolution and precipitation. The
250 ´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258

classic explicit finite difference scheme is used to Ž1986. and Hellmann Ž1994. at 25, 70, and 1008C,
solve mass transport equations. The advective flux respectively. A good agreement is observed for m s
may be computed by either centred or upstream 1r3 ŽFig. 1., which is in accordance with that re-
weighing. The later spatial scheme allows modelling ported by Oelkers et al. Ž1994. at 1508C. The knowl-
of purely advective systems. The reaction and trans- edge of k d at different temperatures allows calcula-
port mass balance equations are fully coupled in a tion of an apparent activation energy Ž Ea ., which is
one-step algorithm. The explicit calculation method
enables us to keep the matrix rank identical to that
solved in KINDISP and, accordingly, to perform
calculations mesh by mesh. KIRMAT provides the
user with mineral volume fractions, porosity, perme-
ability, reactive surface area, and tortuosity over
geologically timeframes in a 1D finite difference
grid. The time evolution of these governing parame-
ters may be computed by either assuming the quasi-
stationary state approximation Že.g., Lichtner, 1988;
Steefel and Lasaga, 1994. or in a transient mode.

3. Computational methods

3.1. The characteristic distance to reach equilibrium

The effect of the aluminium speciation-dependent


rate law ŽEq. Ž4.. on aluminosilicate dissolution rates
is studied by comparing the values of the characteris-
tic distance to reach equilibrium calculated with
KIRMAT with alternatively with Eqs. Ž1. and Ž4..
The characteristic distance for a mineral to reach its
thermodynamic equilibrium, l eq , is inversely propor-
tional to the average mineral reaction rate. The char-
acteristic distance to reach equilibrium was first used
by Knapp Ž1989. to assess the validity of the local
equilibrium approximation ŽLEA. for quartz dissolu-
tion by comparison of computed results with the
scale of interest Že.g., size of a sedimentary basin,
column..

3.2. Kinetic parameters of the aluminium


speciation-dependent rate law

As described in Oelkers et al. Ž1994., it is possi-


ble to reinterpret published rate data initially fitted
with the Eq. Ž1. using Eq. Ž5. and the experimental
values of the total aluminium concentration. This Fig. 1. Graphic reinterpretations of dissolution rate data with Eq.
method has been used on the albite dissolution data Ž5.. Ža. Chou and Wollast Ž1985. at 258C, Žb. Knauss and Wolery
of Chou and Wollast Ž1985., Knauss and Wolery Ž1986. at 708C, Žc. Hellmann Ž1994. at 1008C.
´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258 251

Table 1 Ž s s 3.. The kinetic parameters for albite, K-feldspar


Kinetic parameters used in the dissolution rate equation given by and muscovite used in Eq. Ž4. are listed in Table 1.
Eq. Ž4. for albite, K-feldspar, and muscovite
The dissolution kinetic behaviour of kaolinite is not
Mineral Ea m K8 kd s studied because it is usually a secondary product
Žkcal moly1 . Žmol cmy2 sy1 .
at 258C rather than a primary rock-forming mineral.
Albite 3.6 1r3 0 y12.4 3
K-feldspar 3.6 1r3 0 y13.1 3
muscovite 3.8 1r3 0 y14.3 3
3.3. Kinetic parameters of the pH-dependent rate
law

The kinetic parameters of the pH-dependent rate


independent of pH. The resulting value is Ea s 3.6 law ŽEq. Ž1.. are listed in Table 2. The pH values
kcal moly1 . This value corresponds to the lower delimiting the acid from the neutral region and the
limit for the control by adsorption-surface precursor neutral from the basic regions are noted as pH 1 and
formation of the overall dissolution ŽGuy and Schott, pH 2 , respectively. No apparent activation energy is
1989.. Insufficient data is available to calculate Ea available for muscovite under neutral and basic pH
for K-feldspar dissolution. Based on the structural conditions. The theoretical value proposed by Lasaga
similarity of K-feldspar and albite, we assume they Ž1984. is accepted by default in the present study.
have identical apparent activation energies. This value is widely used in modelling kinetic disso-
The muscovite dissolution data of Knauss and lution processes Žsee for instance, Made,´ 1991; Ben
Wolery Ž1989. at 708C has also been reinterpreted Baccar et al., 1993; Bertrand et al., 1994; Made´ et
Žsee Oelkers et al., 1994.. Interestingly, the exponent al., 1994.. The activation energies associated with
m s 1r3, and the apparent activation energy: Ea s the standard TST rate law are larger than those
3.8 kcal moly1 ŽOelkers, pers. com.. is similar to obtained for the aluminium speciation-dependent ki-
that of albite. The similar behaviour of albite and netic equation. These differences stem from the ef-
muscovite supports the assumption that the reaction fect of dissolved aluminium on the standard dissolu-
order of muscovite dissolution is equal to three tion rate.

Table 2
Kinetic parameters used in the dissolution rate equation given by Eq. Ž1. for albite, K-feldspar, and muscovite dissolution
Mineral log k dŽH q. log k dŽH 2 O . log k dŽOH y. pH 1 pH 2 n ŽpH - pH 1 . n ŽpH ) pH 2 .
Žmol cmy2 sy1 . a Žmol cmy2 sy1 . a Žmol cmy2 sy1 . a
Albite y14.0 y15.7 y18.3 5 8 0.36 y0.32
K-feldspar y14.2 y15.8 y18.2 5 8 0.3 y0.27
muscovite y15.3 y18.0 y19.5 5 7 0.54 y0.21

Mineral EaŽH q. EaŽH 2 O . EaŽH q. s


Žkcal moly1 . Žkcal moly1 . Žkcal moly1 .
Albiteb 21.3 16.5 20.3 1
K-feldspar b 12.4 8.6 13.9 1
muscovite 6d 14.3 d 14.3 e 1
a
At 258C: Chou and Wollast Ž1985. for albite; Made´ Ž1991. for K-feldspar and Knauss and Wolery Ž1989. for muscovite.
b
Hellmann Ž1994..
c
Blum and Stillings Ž1995..
d
Nagy Ž1995..
e
Theoretical value for a reaction control by the formation of an activated complex ŽLasaga, 1984..
k dŽH q. , k dŽH 2 O . and k dŽOH y. stand for dissolution rate constants, EaŽH q. , EaŽH 2 O . and EaŽOH . for the apparent activation energies in acid
ŽpH - pH 1 ., neutral ŽpH 1 - pH - pH 2 ., and basic pH conditions, respectively.
252 ´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258

3.4. Modelling strategy may be reached in an acid mine drainage, some ore
recovery techniques, and is caused by organic acids
Aluminosilicate dissolution rates are function of during sedimentary basin evolution. For instance, the
pH and chemical affinity variations along the reac- upper pH range is attained in concrete formation
tion path, whatever the kinetic formulation em- water. Concrete will be widely used in nuclear waste
ployed. For rates that follow Eq. Ž4., however, aque- storage. The spreading of a hyper-alkaline plume
ous aluminium concentration affects not only the may occur and its geochemical impact needs has
affinity term but also the activity term. Aluminosili- already been studied by modelling water–rock inter-
cate dissolution rates are thereby self-inhibited by actions Žsee for example, Eikenberg and Lichtner,
the aluminium release during dissolution. Because 1992; Steefel and Lichtner, 1994.. The present calcu-
secondary products usually contain silica and alu- lations are performed at 25 and 1508C. The most
minium, they can affect the average mineral dissolu- important inhibitor effects are sought in this study.
tion rate by altering the aluminium concentration in Silica and alkali concentrations in the initial solution
addition to chemical affinity and pH along the flow are very low Ž1 = 10y2 0 molal. to have a maximum
path. It follows that the identity of the secondary initial chemical affinity, which results in the largest
product minerals will have a large effect on alkali- aluminium concentration at equilibrium. A more re-
feldspar and muscovite dissolution rates. For in- alistic initial solution composition, a silica concentra-
stance, a secondary product with a high aluminium tion to quartz or amorphous silica saturation for
to silica ratio Že.g., kaolinite. should decrease to a example, would lead to intermediate results by low-
lesser extent alkali-feldspar and muscovite dissolu- ering aluminium concentration at equilibrium.
tion rates than the precipitation of pure silica. The formation of polynuclear dissolved alu-
The above discussion demonstrates that a larger minium species ŽAl 2 ŽOH. 4q Ž . 5q
2 , Al 3 OH 4 , Al 13 when
.
number of factors affects the average dissolution rate the aluminium concentration is high has not been
when controlled by Eq. Ž4.. In particular, both pre- considered in the calculations. These species are
cipitation and dissolution of secondary minerals have absent from the KIRMAT thermodynamic database.
an additional influence on alkali-feldspar and mus- Furthermore, the current knowledge is insufficient in
covite dissolution rates. As a consequence, a multi- thermodynamic parameters ŽPlyasunov and Grenthe,
step approach is followed to evaluate the effect of 1994; Ohman and Sjoberg, ¨ 1996., which prevents
rate expressions on dissolution rates. The first step is their addition to the database in keeping its consis-
to compute dissolution rates of single minerals alter- tency. Moreover, it is uncertain that the formation of
natively with Eqs. Ž1. and Ž4. using KIRMAT. The the polymer Al 13 from Al 3q is not kinetically con-
second step is to add provision for secondary mineral strained ŽHem and Roberson, 1990; Furrer et al.,
precipitation to these systems. The third step is to 1992.. By the same token, the formation of aqueous
add to the system other dissolving minerals with complexes between alkali ŽNa, K. and AlŽOH.4 y
different stoichiometries and reaction rates. has also been ignored in the present study. The state
3.5. Computational conditions parameters are known ŽPokrovskii and Helgeson,
1995; Diakonov et al., 1996. but only a negligible
KIRMAT calculations are done assuming steady amount is formed over the temperature and concen-
state over a 1D homogeneous semi-infinite system. tration ranges considered.
Pure advective mass transport is considered to keep The approximation leading to Eq. Ž5. is assumed
the comparisons between the different l eq indepen- valid for the three aluminosilicates with the tempera-
dent of the dispersion effects; dispersive transport is ture and pH ranges considered in this study. The
proportional to the concentration gradient, which above experimental reinterpretations performed with
may vary with the rate law used. The range of initial Eq. Ž5. have shown that K8 ™ 0 at various tempera-
conditions is chosen to correspond to those found in tures and a large pH range. None of the laboratory
nature. The initial pH and total dissolved aluminium experiments performed on albite and K-feldspar at
concentration range from 2 to 10 and from 1 = 10y0 9 1508C have led to a different finding ŽOelkers et al.,
to 1 = 10y0 3 molal, respectively. A pH equal to 2 1994; Gautier et al., 1994..
´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258 253

4. Results

4.1. Single mineral systems

The results obtained with KIRMAT at steady state


by assuming that only the presence of a single

Fig. 3. Initial Qr K ratio for K-feldspar vs. the initial pH. The
chemical affinity A is related to Qr K through Eq. Ž2.. Open
symbols: T s 258C; filled symbols: T s1508C; the dashed lines
stand for the results with wAlx init.s1=10y0 9 molal, continuous
line for wAlx init.s1=10y0 3 molal.

aluminosilicate reacting in an aqueous solution are


illustrated in Fig. 2. The characteristic distance to
reach equilibrium, l eq , calculated with Eq. Ž4. di-
vided by that calculated with Eq. Ž1., referred as the
l eq ratio, is plotted as a function of the initial pH at
25 and 1508C and for a total initial dissolved alu-
minium concentration ŽwAlx init. . of 1 = 10y0 3 and
1 = 10y0 9 molal. Among the three aluminosilicates
studied, K-feldspar shows the maximum l eq ratio at
258C. This ratio is close to 100 at pH init.s 6 and wAlx
y0 3
init.s 1 = 10 molal. The minimum l eq ratio is
close to 10 for albite and muscovite at acid pH
ŽpH s 2. and T s 1508C. The later result is indepen-
dent of total initial dissolved aluminium concentra-
tion. Regardless of the temperature and mineral, the
l eq ratio is insensitive to initial aluminium concentra-
tion at pH init.s 2. The same result is obtained at
basic pH and 1508C. The dissolved aluminium con-
centration at equilibrium plays a key role in these
results. The chemical affinity change required to
reach equilibrium Ži.e., the mineral solubility. con-
trols the final wAlx value and depends on pH ŽFig. 3..
At 1508C, the maximum initial chemical affinity is
attained at acid and basic pH conditions, while at
258C this maximum initial affinity exists only at acid
pH. This relation between log ratio mineral solubility
and initial pH outlines the weak pH variation during
Fig. 2. The l eq ratio vs. initial pH. The l eq ratio is defined as the the reaction path.
characteristic distance to reach equilibrium calculated with Eq. Ž4. The solubility is not the only factor affecting the
divided by that calculated with Eq. Ž1.. Ža. K-feldspar, Žb. albite,
Žc. muscovite. Open symbols: T s 258C; filled symbols: T s l eq ratio ŽFig. 2.. The initial dissolution rates strongly
1508C; the dashed lines stand for the results with wAlx init.s1= affect the l eq ratio. As shown in Fig. 4, the ratio of
10y0 9 molal, continuous lines for wAlx init.s1=10y0 3 molal. the initial dissolution rates computed using Eq. Ž1.
254 ´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258

Fig. 4. Ratio of initial K-feldspar dissolution rates vs. the initial Fig. 5. Effect of anorthite Ždiamonds. and quartz Žtriangles.
pH. rd init.Ž1. demotes the initial dissolution rate calculated with dissolution on the l eq ratio calculated for K-feldspar. Also shown
Eq. Ž1.; rd init.Ž4. stands for the initial dissolution rate calculated in this figure are computed results obtained for anorthite or quartz
with Eq. Ž4.. Open symbols: T s 258C; filled symbols: T s1508C; free systems Žopen squares.. The dashed curves correspond to the
the dashed lines curves correspond to the results with wAlx init.s1 results with wAlx init.s1=10y0 9 molal, solid curves lines for
=10y0 9 molal, continuous curves for wAlx init.s1=10y0 3 molal. wAlx init.s1=10y0 3 molal.

vs. Eq. Ž4. is very close to the ratio of the character-


istic distance needed to reach thermodynamic equi- mineral dissolution ŽFig. 5.. This is due to the simul-
librium ŽFig. 2.. In particular, the position of the taneous silica release, which speeds up the approach
maximum l eq ratio for K-feldspar ŽFig. 2. is clearly to equilibrium and thus diminishes the effect of
due to the difference between the initial dissolution aqueous aluminium on the dissolution rate. Accord-
rates ŽFig. 4.. The l eq ratio decreases from 25 to ingly, quartz dissolution produces a more important
1508C despite the higher temperature dependence of decrease of l eq ratio ŽFig. 5.. The faster additional
the standard TST proton-promoted rate law Žsee aluminium and silica are released in to solution, the
Tables 1 and 2.. This is due to the temperature faster the approach to equilibrium, where the effect
dependence of aluminium speciation, which signifi- of the affinity term supplants the inhibitorrcatalytic
cantly alters the initial dissolution rate ratio com- effects either due to aluminium or pH. The l eq ratio
puted using Eq. Ž4.. tends toward the unity when the local equilibrium
limit is approached ŽFig. 6..
4.2. Influence of other dissolÕing minerals

Comparative calculations have been performed to


evaluate the effects of adding selected dissolving
minerals to alkali-feldspars and muscovite. The
choice of mineral added to the system is based on
their occurrence in natural systems. The aluminium
to silica ratio of the rock-forming minerals associ-
ated with alkali-feldspars and muscovite usually
ranges from the unity to zero. Therefore anorthite
and quartz are considered in the comparative calcula-
tions, stoichiometric anorthite dissolution is as-
sumed. Intermediate results are expected for the
addition of minerals possessing intermediate alu-
Fig. 6. The effect of anorthite dissolution rate on K-feldspar l eq
minium to silica ratios. ratio at T s 258C and pH init.s6. The dashed curve stand for the
Aluminium release from anorthite decreases the results with wAlx init.s1=10y0 9 molal, solid curve for wAlx init.s1
l eq ratio compared to that obtained above for single =10y0 3 molal.
´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258 255

Fig. 7. The l eq ratios at 25 and 1508C for K-feldspar, albite, and muscovite dissolution for various initial conditions as a function of initial
pH. The l eq ratio is defined as the distance required to attain steady state calculated using the aluminium speciation rate law ŽEq. Ž4..
divided by that calculated using the standard pH-dependent law ŽEq. Ž1.. with KIRMAT. The continuous curves denote the single mineral
results. The dashed and dot-dashed curves correspond to l eq ratios obtained for comparisons performed for an initial aluminium
concentration of 1 = 10y0 9 and 1 = 10y0 3 molal, respectively. These latter results were corrected so that the chemical affinity term of both
Eqs. Ž1. and Ž4. were identical Žsee text..
256 ´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258

4.3. Influence of the secondary products Muscovite shows the smallest rate decrease Ži.e.,
the larger l eq ratio. over the temperature range con-
The precipitation of secondary products produces sidered ŽFig. 7.. The effect of the aluminium specia-
the reverse effect on the l eq ratio from that of the tion-dependent rate law is to decrease the muscovite
dissolution of other mineral phases. The largest ef- dissolution rate by one order of magnitude for an
fect is found when an aluminium-rich secondary initial aluminium concentration of Ž1 = 10y0 3 molal..
phase rapidly precipitates. When solubility of this The maximum effect is obtained at neutral and acid
secondary phase is within 1 = 10y0 9 and 1 = 10y0 3 initial pH for 25 and 1508C, respectively. The alu-
molal, the minimum and maximum total initial alu- minium speciation-dependent rate law lowers alkali-
minium concentrations used in this study, the ratio of feldspar dissolution rate by as much as two orders of
the distance to reach equilibrium is equal to that of magnitude, for an initial aluminium concentration of
the initial dissolution rates ŽFig. 4.. Note that these 1 = 10y3 molal. This maximum is attained at 258C
results should be corrected to account for the intrin- for K-feldspar Žneutral pH., and at 1508C for albite
sic higher dependence with respect to the chemical Žacid pH.. The temperature appears to have a larger
affinity of the dissolution rate calculated with Eq. Ž4. effect on l eq for the alkali-feldspars than for mus-
Žsee below..
covite.
Only pure silica precipitation increases the l eq The maximum increase in albite dissolution rate
ratio compared to the single mineral results due to a due to the aluminium speciation-dependent rate law
higher aluminium concentration at equilibrium. is three orders of magnitude at 258C and at acid pH.
However, this possibility is considered unlikely be- For the other minerals tested, the maximum increase
cause the excessively high aluminium concentration is between two and three orders of magnitude, also at
resulting from these calculations are inconsistent with low temperatures and at acid conditions. At 1508C, a
the ubiquity of aluminium-bearing secondary prod- two orders of magnitude rate increase is observed at
ucts in natural systems. basic conditions for all three minerals. According to
4.4. The maximum increase and decrease in dissolu- Eq. Ž4. these rates would increase by another one
tion rates due to the aluminium speciation-dependent order of magnitude lower if the initial aluminium
rate law concentration was lowered by a factor of 1000 to
wAlx init.s 1 = 10y1 2 molal.
The maximum l eq ratio increase, which corre-
sponds to the maximum dissolution rate decrease due
to the effects of aqueous aluminium, can be graphi- 5. Conclusions
cally determined using Fig. 7. This figure illustrates
the l eq ratio for the case of single mineral dissolution The effects of the aluminium speciation-depen-
Žsolid curve. and for the case of mineral dissolution dent rate law ŽEq. Ž4.. have been determined for
in an initial solution containing 1 = 10y3 molal Al K-feldspar, albite and muscovite, at 25 and 1508C
Ždashed curve.. The maximum l eq ratio is given by and over a wide range of initial pH and dissolved
the higher of these two curves at any given initial aluminium concentration. These calculations are pos-
pH. The minimum increase of the l eq ratio cannot be sible by using the KIRMAT hydrochemical code and
larger than unity, as it is set by the local equilibrium a multi-step approach. The first step consists in
limit. On the other hand, the lowest l eq ratio, which modelling single mineral systems. Some additional
corresponds to the largest dissolution rate increase, dissolving and precipitating secondary silicates with
occurs for dissolution in a solution with constant end-member reaction rates and silica–aluminium
aluminium concentration equal to 1 = 10y0 9 molal. stoichiometry were considered in additional calcula-
The effect of the higher chemical affinity depen- tions. This modelling study indicates the need to
dence of Eq. Ž4. is corrected by multiplying the model the reactive transport phenomena with com-
former curve by the l eq ratio calculated for s s 1 pletely validated rate equations; up to three orders of
and s s 3, without considering aluminium specia- magnitude differences in muscovite and alkali-
tion effects. feldspar dissolution rates are found when using an
´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258 257

aluminium speciation-dependent rate law vs. the Brady, P.V., Walther, J.C., 1990. Kinetics of quartz dissolution at
standard proton-promoted rate equation. In addition, low temperatures. Chem. Geol. 82, 253–264.
Brady, P.V., Walther, J.C., 1992. Surface chemistry and silicate
secondary product formation produces sensitive dissolution at elevated temperatures. Am. J. Sci. 292, 639–658.
feedbacks on alkali-feldspar and muscovite dissolu- Carroll, S.A., Walther, J.V., 1990. Kaolinite dissolution at 25, 60
tion rates by controlling the dissolved aluminium and and 808C. Am. J. Sci. 290, 797–810.
silica concentrations. Therefore, the effects of aque- Casey, W.H., Sposito, G., 1992. On the temperature dependence
ous aluminium concentration on rates may have a of mineral dissolution rates. Geochim. Cosmochim. Acta 56,
3825–3830.
crucial effect on basin modelling performed to evalu- Chen, Y., Brantley, S.L., 1997. Temperature and pH-dependence
ate the permeability variations and fluid migration. of albite dissolution rate at acid pH. Chem. Geol. 135, 275–
290.
Chou, L., Wollast, R., 1985. Steady-state kinetics and dissolution
mechanisms of albite. Am. J. Sci. 285, 963–993.
Acknowledgements
´´
Coudrain-Ribstein, A., 1988. Transport d’elements ´
et reactions
´
geochimiques `
dans les aquiferes. ` Univ. Louis Pasteur,
These,
Support was from the Centre National de la Strasbourg, France.
´ ´ de Secours
Recherche Scientifique and the « Societe Diakonov, I., Pokrovskii, G., Schott, J., Castet, S., Gout, R., 1996.
des Amis des Sciences ». The authors are indebted to An experimental and computational study of sodium–
aluminum complexing in crustal fluids. Geochim. Cosmochim.
E.H. Oelkers for cheerful encouragement, fruitful Acta 60, 197–211.
discussions, and thoughtful correction of the English Eikenberg, J., Lichtner, P.C., 1992. Propagation of hyperalkaline
language. cement pore waters into the geologic barrier surrounding a
radioactive waste repository. In: Kharaka, Y.K., Maest, A.S.
ŽEds.., Water–Rock Interactions Proc., Balkema, Rotterdam,
pp. 377–380.
References Eyring, H., 1935. The active complex in chemical reactions. J.
Chem. Phys. 3, 107–115.
Aagaard, P., Helgeson, H.C., 1982. Thermodynamic and kinetic ¨
Furrer, G., Trusch, B., Muller, C., 1992. The formation of polynu-
contraints on reaction rates among minerals and aqueous clear Al13 under simulated natural conditions. Geochim. Cos-
solution: I. theorical considerations. Am. J. Sci. 282, 237–285. mochim. Acta 56, 3831–3838.
Acker, J.G., Bricker, O.P., 1992. The influence of pH on biotite ´ J.L., Lasaga, A.C., 1995. The effect of pH
Garnor, J., Mogollon,
dissolution and alteration kinetics at low temperature. on kaolinite dissolution rates and on activation energy.
Geochim. Cosmochim. Acta 56, 3073–3092. Geochim. Cosmochim. Acta 59, 1037–1052.
Amrhein, C., Suarez, D.L., 1992. Some factors affecting the Gautier, J.-M., Oelkers, E.H., Schott, J., 1994. Experimental study
dissolution kinetics of anorthite at 258C. Geochim. Cos- of K-feldspar dissolution rates as a function of chemical
mochim. Acta 56, 1815–1828. affinity at 1508C and pH 8. Geochim. Cosmochim. Acta 58,
Ben Baccar, M., Fritz, B., Made, ´ B., 1993. Diagenetic albitization 4549–4560.
of K-feldspar and plagioclase in sandstone reservoirs: thermo- ´
Gerard, ´
F., 1997. Modelisation ´
geochimique thermodynamique et
dynamic and kinetic modeling. J. Sedim. Petrol. 63, 1100– ´
cinetique ´
avec prise en compte des phenomenes ` de transport
1109. de masse en milieu poreux sature. ´ These,
` Univ. Louis Pasteur,
Bennett, P.C., 1991. Quartz dissolution in organic-rich aqueous Strasbourg, France.
systems. Geochim. Cosmochim. Acta. 55, 1781–1797. ´
Gerard, ´
F., Clement, A., Fritz, B., Crovisier, J.-L., 1996. Introduc-
Bennett, P.C., Casey, W., 1994. Chemistry and mechanisms of tion of mass transport phenomena into the thermodynamic and
low-temperature dissolution of silicates by organic acids. In: kinetic geochemical code KINDIS: the code KIRMAT. C.R.
Pittman, E.D., Lewan, M.D. ŽEds.., Organic Acids in Geologi- Acad. Sci., Paris, Vol. 322 ser. IIa, pp. 377-384.
cal Processes. Springer, Berlin, pp. 162–200. ´
Gerard, ´
F., Clement, A., Fritz, B., 1997. Numerical validation of a
Bertrand, C., Fritz, B., Sureau, J.-F., 1994. Hydrothermal experi- Eulerian hydrochemical code using a 1D multisolute mass
ments and thermo-kinetic modelling of water–sandstone inter- transport system involving heterogeneous kinetically-con-
actions. Chem. Geol. 116, 189–202. trolled reactions. J. Contam. Hydrol. 30, 201–216.
Blum, A.E., Stillings, L.L., 1995. Feldspar dissolution kinetics. In: Guy, C., Schott, J., 1989. Multi-site surface reaction versus trans-
White, A.F., Brantley, S.L. ŽEds.., Chemical Weathering Rates port control during the hydrolysis of a complex oxide. Chem.
of Silicate Minerals, Mineral. Soc. Am., Rev. Mineral., Vol. Geol. 78, 181–204.
39, pp. 291–351. Helgeson, H.C., Delany, J.M., Nesbitt, H.W., Bird, D.K., 1978.
Brady, P.V., Walther, J.C., 1989. Controls of silicate dissolution Summary and critique of the thermodynamic properties of
rates in neutral and basic pH solutions at 258C. Geochim. rock-forming minerals. Am. J. Sci. 278A, 1–229.
Cosmochim. Acta 53, 2823–2830. Helgeson, H.C., Murphy, W.M., Aagaard, P., 1984. Thermody-
258 ´
F. Gerard et al.r Chemical Geology 151 (1998) 247–258

namic and kinetic constraints on reaction rates among minerals port in Porous Media, Mineral. Soc. Am., Rev. Mineral., Vol.
and aqueous solutions: II. Rate constants, effective surface 34, pp. 131–191.
area and hydrolysis of feldspars. Geochim. Cosmochim. Acta Oelkers, E.H., Schott, J., 1995. The dependence of silicate disso-
48, 2405–2432. lution rates on their structure and composition. In: Kharaka,
Hellmann, R., 1994. The albite–water system: Part I. The kinetics Y.K., Chudaev ŽEds.., Water–Rock Interactions, Balkema,
of dissolution as a function of pH at 100, 200, and 3008C. Rotterdam, pp. 153–156.
Geochim. Cosmochim. Acta 58, 595–611. Oelkers, E.H., Schott, J., Devidal, J.-L., 1994. The effect of
Hem, J.D., Roberson, C.E., 1990. Aluminium hydrolysis reactions aluminium, pH, and chemical affinity on the rates of alumi-
and products in mildly acidic aqueous systems. In: Melchior, nosilicates dissolution reactions. Geochim. Cosmochim. Acta
D.C., Bassett, R.L. ŽEds.., Chemical Modeling of Aqueous 58, 2011–2024.
Systems II. ACS Symp. Ser. 416, pp. 429–446. ¨
Ohman, L.O., Sjoberg, S., 1996. The experimental determination
Hochella, M.F., Banfield, J.F., 1995. Chemical weathering of of thermodynamic properties for aqueous aluminium com-
silicates in nature: a microscopic perspectives with theoretical plexes. Coord. Chem. Rev. 149, 33–57.
considerations. In: White, A.F., Brantley, S.L. ŽEds.., Chemi- Plyasunov, A.V., Grenthe, I., 1994. The temperature dependence
cal Weathering Rates of Silicate Minerals. Mineral. Soc. Am., of stability constants for the formation of polynuclear cationic
Rev. Mineral., Vol. 39, pp. 352–406. complexes. Geochim. Cosmochim. Acta 58, 3561–3582.
Knapp, R.B., 1989. Spatial and temporal scales of local equilib- Pokrovskii, V.A., Helgeson, H.C., 1995. Thermodynamic proper-
rium in dynamic fluid–rock systems. Geochim. Cosmochim. ties of aqueous species and the solubilities of mineral at high
Acta 53, 1955–1964. pressures and temperatures: the system Al 2 O 3 –H 2 O–NaCl.
Knauss, K.G., Wolery, T.J., 1986. Dependence of albite dissolu- Am. J. Sci. 295, 1255–1342.
tion kinetics on pH and time at 25 and 708C. Geochim. Rimstidt, J.D., Barnes, H.L., 1980. The kinetics of silica–water
Cosmochim. Acta 50, 2481–2497. reactions. Geochim. Cosmochim. Acta 44, 1683–1699.
Knauss, K.G., Wolery, T.J., 1989. muscovite dissolution kinetics Schott, J., Oelkers, E.H., 1995. Dissolution and crystallisation
as a function of pH and time at 708C. Geochim. Cosmochim. rates of silicate minerals as a function of chemical affinity.
Acta 53, 1493–1501. Pure Appl. Chem. 67, 903–910.
Lasaga, A.C., 1981. Transition State Theory. In: Lasaga, A.C., Steefel, C.I., Lasaga, A.C., 1994. A coupled model for transport
Kirkpatrick, R.J. ŽEds.., Kinetics of Geochemical Process. of multiple chemical species and kinetic precipita-
Mineral. Soc. Am., Rev. Mineral., Vol. 8, pp. 169–195. tionrdissolution reactions with applications to reactive flow in
Lasaga, A.C., 1984. Chemical kinetics of water rock interaction. J. single phase hydrothermal systems. Am. J. Sci. 294, 593–620.
Geophys. Res. 86, 4009–4025. Steefel, C.I., Lichtner, P.C., 1994. Diffusion and reaction in rock
Lichtner, P.C., 1988. The quasi-stationary state approximation to matrix bordering a hyperalkaline fluid-filled fracture. Geochim.
coupled mass transport and fluid–rock interaction in a porous Cosmochim. Acta 58, 3595–3612.
media. Geochim. Cosmochim. Acta 52, 143–165. Stillings, L.L., Drever, J.I., Brantley, S.L., Sun, Y., Oxburgh, R.,
´ B., 1991. Modelisation
Made, ´ ´
thermodynamique et cinetique des 1996. Rates of feldspar dissolution at pH 3–7 with 0–8 mM
´
reactions ´
geochimiques `
dans les interactions eau-roche. These, oxalic acid. Chem. Geol. 132, 79–90.
Univ. Louis Pasteur, Strasbourg, France. Tole, M.P., Lasaga, A.C., Pantano, C., White, W.B., 1986. The
´ B., Clement,
Made, ´ A., Fritz, B., 1994. Modeling mineralrsolu- kinetics of dissolution of nepheline ŽNaAlSiO4 .. Geochim.
tion interactions: the thermodynamic and kinetic code Cosmochim. Acta 50, 379–392.
KINDISP. Comput. Geosci. 20, 1347–1363. Welch, S.A., Ullman, W.J., 1993. The effect of organic acids on
Murphy, W.M., Helgeson, H.C., 1987. Thermodynamic and ki- plagioclase dissolution rates and stoichiometry. Geochim. Cos-
netic constraints on reaction rate among minerals and aqueous mochim. Acta 57, 2725–2736.
solution: III. Actived complexes and the pH-dependence of the White, S.P., 1995. Multiphase nonisothermal transport of systems
rates of feldspar, pyroxene, wollastonite and olivine hydroly- of reacting chemicals. Water Resour. Res. 31, 1761–1772.
sis. Geochim. Cosmochim. Acta. 51, 3137–3153. Wieland, E., Werhli, B., Stumm, W., 1988. The coordination
Nagy, K.L., 1995. Dissolution and precipitation kinetics of sheet chemistry of weathering: III. A potential generalization of
silicates. In: White, A.F., Brantley, S.L. ŽEds.., Chemical dissolution rates of minerals. Geochim. Cosmochim. Acta 52,
Weathering Rates of Silicate Minerals, Mineral. Soc. Am., 1969–1981.
Rev. Mineral., Vol. 39, pp. 173–233. Wogelius, R.A., Walther, J.V., 1991. Olivine dissolution at 258C:
Oelkers, E.H., 1996. Physical and chemical properties of rocks effects of the pH, CO 2 , and organic acids. Geochim. Cos-
and fluids for chemical mass transport calculations. In: Licht- mochim. Acta 55, 943–954.
ner, P.C., Steefel, C.I., Oelkers, E.H. ŽEds.., Reactive Trans-

Anda mungkin juga menyukai