- Email: [email protected]
The dissolution of biotite and chlorite at 25°C in the near-neutral pH region
JOURNAL OF
Contaminant Hydrology ELSEVIER
Journal of Contaminant
Hydrology
21(1996)
201-213
The dissolution of biotite and chlorite at 25°C in the near-neutral pH region Maria Malmstrijm
a, Steven Banwart aT*, Jeanette Lewenhagen Lara Duro b, Jordi Bruno ’
a,
aDepartment of Inorganic Chemistry, The Royal Institute of Technology,Stockholm, Sweden ’Department d’Enginyeria Quimica, Universitat Polit&nica de Catalunya,Barcelona, Spain ’h4BT Technologica Ambiental,Cerdanyola, Spain Received 17 December
1993; accepted
1.5 December
1994 after revision
Abstract We studied
the dissolution
of biotite
and chlorite
in laboratory
systems
with flow-through
and
pH region, under N2(sj atmosphere is highly non-stoichiometric. A slow linear release of iron during a period of weeks indicates a surface-chemical-reaction-controlled mechanism of release for iron. The release of potassium is much faster than that of iron. A parabolic dependence of accumulated release with time suggests a diffusion-controlled mechanism of potassium release. The rates of magnesium, aluminium and silicon release are between those for potassium and iron and approach that of iron with time. There is no significant influence of (bi)carbonate or pH on biotite dissolution rate or stoichiometry in the pH region 7 < pH < 8.5. The release rates of elements from chlorite are close to stoichiometric and similar to the iron release rate from biotite. In closed batch reactors at near-basic pH the composition of test solutions in contact with biotite is apparently controlled by gibbsite (Al), kaolinite (Si) and Fe(III)-(hydrjoxide. We estimated a turn-over time (lo’-lo2 yr) for molecular oxygen and a time scale (10 months) to develop characteristic Fe(U) concentrations for a granitic groundwater. batch
reactors.
The
initial
dissolution
of
biotite
in the
near-neutral
1. Introduction The most critical safety aspect for long-term disposal of high-activity nuclear waste is the design of engineered barriers for waste containment. Safety assessment, however,
* Corresponding author. Present address: Department Bradford, Bradford, BD7 lDP, UK. Elsevier Science B.V. SSDI 0166-3542(95)00047-X
of Civil and Environmental
Engineering,
University
of
202
M. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
must consider eventual failure of these barriers. In this case, the time scale for migration of long-lived radionuclides from a deep aquifer to the biosphere is dependent on hydrologic processes and aqueous speciation controlling solubility and adsorption behaviour. Redox conditions are especially important concerning the migration of uranium, neptunium, plutonium and technetium. All form sparingly soluble (hydr)oxides under reducing conditions but are highly soluble, and thus mobile, in their higher
T
14A
Fig. 1. Idealized crystal structure [modified from Bamhisel and Bertsch (1989)] of: (a) biotite; and (b) chlorite. An octahedral sheet of cations lies between by two tetrahedral sheets of silicon and aluminium. Biotite maintains charge balance by K + ions whereas chlorite has a hydroxy-interlayer.
M. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
203
valence states. During its operating lifetime, a deep repository for high-activity nuclear waste will be open to oxic conditions and surface water inflow. After closing the repository, it is important to anticipate the extent of the oxidizing disturbance, and the time scale for the aquifer to return to reducing conditions. In addition to respiration of organic carbon, uptake of molecular oxygen by reducing minerals can contribute to the development of anoxia in groundwaters (Grenthe et al., 1992). Molecular oxygen can be reduced directly at Fe(I1) sites on the mineral surfaces, or in solution subsequent to release of structural Fe(I1) (White and Yee, 1985). Biotite and chlorite are two important reservoirs of Fe(I1) in granitic bedrocks (Banwart et al., 1994). Here we report laboratory results on the thermodynamics and kinetics of biotite dissolution, and chlorite dissolution kinetics, at neutral and near-basic pH. These wet chemical experiments do not specifically address structural transformation of the mineral phases or redox speciation in solution, but focus on the stoichiometry of element release. Specific objectives are: (1) to determine the rate and stoichiometry of biotite and chlorite dissolution under near-neutral anoxic conditions; (2) to assess the effect of the (bi)carbonate ligand on biotite dissolution kinetics; and (3) to assess major-element speciation, and saturation with respect to possible secondary phases, in solutions reacted with biotite.
Table 1 Mineralogical
data on the specimen Biotite
Chlorite
Elemental ratio:
Si Fe Mg Al K
3.81 1.65 1 1.50 0.95
Fe(U) /Fe(tot)
kO.6 +O.l + 0.04 + 0.1
34.2 f 0.6 1 26.9 + 0.9 6.7kO.l -
ratio (o/o):
Spectrophotometric Mijssbauer
a
74.9 f 4.8 85.5
70.9 f 5.3 52.1
X-ray d(001) spacing (A)
14.40
Specific surface area (m’ g- I):
75-125 125-300
/em wm
Error estimates are a Method modified b Determined from ’ Determined from
1.81 +O.l b given as one standard deviation. from Pirhonen and Pitklnen (1991). Brunauer-Emmett-Teller adsorption BET analysis using helium.
4.38 ’ 3.36 ’
isotherm (BET) analysis using krypton.
204
M. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
Previous laboratory investigations of biotite (Schnitzer and Kodama, 1976; Acker and Bricker, 1992; Trotignon and Turpault, 1992 - up to pH 7) and chlorite (Ross, 1969) dissolution kinetics focused on the acid pH region.
2. Experiments 2.1. Mineral characterization
and sample preparation
Biotite (Newman and Brown, 1987; Fanning et al., 1989) and chlorite (Newman and Brown, 1987; Barnhisel and Bertsch, 1989) are structurally similar (Fig. 1) consisting of a sheet of octahedrally coordinated cations surrounded by two tetrahedrally coordinated sheets of silicon and aluminium. Biotite maintains charge balance with exchangeable K+ ions in the silicate interlayers whereas chlorite has brucite/gibbsite-like hydroxyinterlayers. Naturhistoriska Riksmuseet (Stockholm, Sweden) provided the natural mineral specimens used in this study (Biotite LK 5210, Chlorite 740665). Table 1 lists characteristics of the mineral specimens. Powder X-ray diffractometry confirmed the minerals to be biotite and chlorite with no detectable crystalline impurities. Based on the results in Table 1 and assignment of structural positions according to Newman and Brown (1987) and Bailey (1972, equation 1) the mineral compositions are: Biotite:
K o.so(Mg,.,,Fe:‘.,,Fe~.~~~~.~~)(Si,.,,~”.,,)O,”(OH)*
Chlorite:
(Mg,.,~o.~FeZ’.,Fe~.‘,)(Si,.,~,.,)O,,(OH)s
FEEDVESSEL
u
MEASURING CELL
REACTOR
_SAMPLE BOTTLE Fig. 2. Experimental setup for the thin-film continuous flow-through system. Test solution of known composition is pumped through a thin-film reactor where the mineral powder is held between two membrane filters. The outflow solution passes a closed U-tube where pH is measured continuously. The reacted solution is collected in acid-containing polypropylene bottles.
M. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
205
Mineral powder was prepared by peeling small sheets from the mineral surface, size fractionating with sieves, and washing ultrasonically three times in ethanol to remove ultrafine particles (Schott et al., 1981). 2.2. Experimental
procedure
Experiments were carried out at 25 f 1°C 1-atm pressure, using the ionic medium: [Na+] + [H+] = [HCO;] + [ClO;] + [OH-] = 0.500 M. High-purity chemicals (pro analysi or better) and gases (99.5% purity) were used. Gases were cleaned by bubbling through an oxygen trap (amalgamated zinc/vanadous solution) and a wash bottle containing ionic medium to saturate the gas stream with water vapour at experimental ionic strength and temperature. Reacted solution samples were analyzed for Si, Mg, Al, total Fe and K where applicable using an ARL” (Applied Research Laboratory) model 3520B ICP (inductively coupled plasma emission spectrometer) analyzer. Standards were prepared by dilution of Atomic Spectral Standards (Analytical Standards, Kungsbacka, Sweden) into appropriate matrix solutions. Proton concentrations were determined using a glass electrode-Ag/AgCl reference electrode couple. The electrode couple was normally calibrated according to Gran (1981) before and after each experiment. Membrane filters used in the experiments were 0.2-pm pore size filters of PTFE-type (polytetrafluoroethylene, MFS@ J020A025A or Millex” FG SLFG025BS). 2.2.1. Kinetic experiment The experiments were performed using a continuous flow-through reactor (Bnmo et al., 1991) (Fig. 2), consisting of a thin layer (< 3 mm) of mineral powder of known mass, held between two 25-mm-diameter membrane filters, across which test solutions of known composition were pumped at specified rates. The pH of the reactor outflow was recorded continuously. The following equation calculates the surface area normalized mineral dissolution rate from concentration [C (mol/g solution)] of element j in sample i: Ri,j
=
Fici,j mop
(moles of mineral hh’ m-‘) I
(1)
where F = flow rate (g h- ’); m = mass of mineral (g); A = specific surface area (m* g-l); and P is a stoichiometric factor (moles of element/moles of mineral). The following equation calculates the accumulated release of an element during an experiment: n,( ti) = zR,,/(
ti - ti_ 1)
(moles of mineral m-‘)
(2)
where n denotes accumulated release (moles of mineral m-*1; and t the time (h) elapsed from the start of the experiment. In addition to continuously purging feed solutions with N,,,, the carbon dioxide-free experiments were carried out in a glove box under N2(g) atmosphere. In these experiments the proton concentrations in the feed solutions were calculated from dilution of
206
M. Malmstriim et al./Journal
of Contaminant Hydrology 21 (1996) 201-213
standard acid (HClO,) or base (NaOH) solutions in NaClO,. A fresh mineral sample was used for each experiment. Solution samples were collected in polypropylene, acid-washed bottles, containing a small volume of concentrated HNO,. 2.2.2. Equilibrium
experiments
2.2.2.1. Batch experiments. A known mass of mineral was placed in a Teflon” or polypropylene bottle. The bottle was filled with ionic medium at specified pH, sealed and placed onto a shaker table under NX,) atmosphere. The system was left to equilibrate for 60 days. The final pH was measured, the solid phase separated by membrane filtration and the filtrate acidified (to pH 5 1) by a small volume of nitric acid. 2.2.2.2. Titration experiments. The titration was started at pH = 4 and was brought to pH 10.7 by consecutive base additions. After each addition, equilibrium was assumed when the potential drift in the glass electrode assembly was < 10m4 mV ss’. At equilibrium the solution was sampled by removal of some suspension, separation of the solid phase by membrane filtration and acidification (to pH 5 1) of the filtrate by a small volume of nitric acid. The titration vessel was maintained under Nzcgj atmosphere.
3. Results and discussion 3.1. Kinetic experiment Fig. 3a shows the accumulated release of elements from biotite with time at pH 7 and 1 atm N2(sj. The release of all elements is rapid at the beginning of the experiment and decreases to quasi steady-state rates with time. Such a decrease of dissolution rates with time is typical for dissolution kinetic experiments using fresh mineral samples (Furrer and Stumm, 1986; Caroll-Webb and Walther, 1988; Wieland and Stumm, 1992), and is due to dissolution of ultrafine particles (Holdren and Berner, 1979; Schott et al., 1981) and more reactive surface sites (Wehrli, 1989). The dissolution is initially highly non-stoichiometric with total iron release relatively slow and linear with time. Linear release, obtained at a fixed flow rate, reflects constant output concentrations from the thin-film reactor. This may be due to steady-state kinetics or mineral-solution equilibrium. At the end of this experiment we ran the thin-film reactor at half the rate, yielding approximately double concentrations of iron, aluminium and magnesium in the outflow solution, excluding the possibility of equilibrium conditions or transport-controlled release of these elements. Steady-state kinetics is expected for a surface-chemical-reaction-controlled dissolution mechanism with constant concentration of reactive surface iron sites (Furrer and Stumm, 1986; Blum and Lasaga, 1988; Wieland et al., 1988; Stumm and Wieland, 1990; Stumm and Wollast, 1990). Potassium release is very fast at the beginning of the experiment and apparently becomes diffusion controlled after N 10 days as indicated by a linear plot of accumulated release against Jt (not shown; Schott and Berner, 1985). Aluminium and magnesium release rates are between the extremes
of Contaminant Hydrology 21 (1996) 201-213
M. Malmstriim et al. /Journal
.2E-O3
5E-05
7
E
3;
2
g
207
4E-05 3E-05
“6 2E-05 I. 2 IE-05 c
0
0 0
5
10
20
15
25
30
Time [Days]
-
3E-05
- 2.OE-03
-lbj
3
0
5
10
15
20
25
30
Time [Days] Fig. 3. Accumulated release of minerals calculated (Eq. 2) for each element (A = Si; v = Fe; ??= Al; X = K) at three different experimental conditions. a. Biotite dissolution pH 7 (CO,-free). b. Biotite dissolution pH 8.1, 1% COys,. c. Chlorite dissolution pH 8.2, 1% COq,). The scale for K+ release is shown on the vertical axis at the right. The experiment at pH 7 was run flow rate during the final 3 days of the experiment yielding double concentration of Fe, Mg and outflow solution. This excludes the possibility of equilibrium or transport controlled release of these
0 = Mg;
at half the Al in the elements.
exhibited by Fe and K, and approach the iron release rate with time. Except for the anomalous increase after day 20, Si release follows that of Al and Mg. Table 2 lists the dissolution rates calculated from each element in four kinetic experiments. We want to emphasize that these experiments were performed in a flow-through system where the feed solution flushes the released elements from the reactor, preventing precipitation of secondary reaction products. A scanning electron
.pallodal uaaq amq aleuoq.m(!q) dq uognlossyp p2lauy 30 (1661 ‘.IaqlleM pm snga%!oM) uoyq~qu! pue (~661 ‘JEMU~~ pm %a8 fz661 ‘y2 la 0un.Q uogwalame qloq amle~alg aql UI .uo+?al Hd lw)nau-Jeau aql u! d_IlauIo!q~~ols IO alw uoyI~oss!p uo Hd JO aleuoqlm(!q) 30 lDa33a lue~3!u8!s ou s! alaql leql s~oys (z a[ql?J_,) sluampadxa algo!q aalql aql 30 uosydmo:, v .luawpadxa sno!AaId aql 01 .w[y!s s! sluaura[a 11~ 30 mo!lzeqaq aq$ SMOGS qE ‘%d aqL .(“)zo3 urle I()‘() lapun 1’8 Hd le algo!q 30 uoyqoss!p waw!ladxa aql30 pua aql 1~ ad 30 leql my1 Jai923 [I!ls s! 9 aurg qi!~ shoos +x 30 asea[al aql q8noql -w .aseaIal IuauraIa pa~~orluo3-uo!l3ea~-le3ruray3-a~~~ns 1saMoIs aql ‘dq pal!uuI pm ‘01 zyawoyq~~ols ‘asealal Jeaug 01 qzoldde (0 pue flakI palaIdap E q8nolql uo:sn33Fp 01 anp aseaIaJ 3qoqeled 01 uo!l!sue~l (z) fsal!s aa!lzzal ~1013 aseaIa1 paIIoquoD-uog -3eal-Ie~ruIaq~-a3e3ms lse3 (1) :sa%ls aaJq1 qZnoJq$ sassed uoyqoss!p cgaworq~rols 01 uo!ssa.Bold aqL .qled uoyg!p IaZuo[ d[aA!ssaJ8old e 30 u0ym.103 01 anp ‘(ad) pasl?aIaJ 4uauIaIa jsamop aql ql!M ~!J4amo~q3!ols auIoDaq II!M +x paseaIa1 dIp!del A~pz!lru! aql ‘au!1 qlrm leql swpald [apow s!qL .palyg-aw uO!Sn33fp sauro9aq asea1a.I +‘>I aql qD!qM q8nolql uo@aI palaldap-+x e surlo3 pm s~adepalu! al!lo!q aql u10_13+y saAouIal +H Pue +EN uognlos q$!M a8ueqDxa p!dea Xap!xo xalduro3 30 uognIoss!p luan.x%om! 103 (0661) loouqDg 30 [apour Ienldamo3 aql E?uysn slInsa1 asaql 1aJdIalur aM ‘al!loTq papeamn aql ~11013samala33rp p@oloq&ouI JO samydr3ald ou smoqs L Hd le palXal aIdtues algoyq aql 30 (umoqs IOU) a&u! (~2s) adoD,soguI
_
_
pq
,_01.8’6 uoyalap .+tolaq
L
-01
g_OI.P’9
(Z’O T P’Z)
,_OI.Z'Z b-OI.(Z’ZTP’9)
,_01.9'9
,_01.8'S ,_oI~(S'o+I'z)
~_OI~(Z'ZTSS-)
,_OI.Z'I b_01.(9'0TE'6)
g-01.8'2 g-OI~(P'OTO'1)
'N"! '03
pOI.1’9
i -01 ‘(O’I T P.9)
pOI.5'9 x _OI'(Z'IT6'2)
I(_OI'E'P ,_OI~(Z'IfPW
i_OI,C'I i_oI'(z'oTz'I)
pOI,E’9
i_oI.(o’IT9’s)
L-01 ,O'I ,_OI.(L'OTZ'L)
+01.0'8 b-oI~(z‘Ir8'L)
i-OI.O'I i_OI.(I'OT1'I)
xu
(wu!) (te=g) x~
(tevur) lv'tl (rw3) ?
(wu!)
"wbl
(vv3) Bwx
(p!p!) !slf (lF3) lS2f
'NU! '03 %Z0'0f00'1 avwq
%z0'0T00'1 aiypp 2'8
'N %OOI appq
I'8
(z ,ba) luaruala qxa
'N%OOI avwq
S'8
svo le_raugq Hd
L
~013 pa,epm[e3 (, _q z_tu [eJau!m salour) saw uoqn[oss!a z at9e.L
A4. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
209
Fig. 3c shows the dissolution of chlorite at pH 8.2 under 0.01 atm CO,,,,. The final accumulated release of chlorite is much lower than for biotite (Fig. 3a and b) due mainly to a lower initial dissolution rate. Incorporation of hydrated ions into the biotite interlayer expands the mica structure and exposes additional mineral surface to solution attack. This is not expected for the chlorite hydroxy-interlayers. The initial release rates of the different elements are nearly stoichiometric. After 16-17 days the outflow concentrations of iron and aluminium are below detection limits. However, the final dissolution rate (Table 2), calculated from silicon and magnesium release, is similar to that of biotite. 3.2. Equilibrium
experiments
Fig. 4a-e shows the equilibrium concentration of elements released from biotite as a function of pH in the neutral to near-alkaline region. The plotted lines represent the solubility of possible secondary phases calculated by HARPHREEQ (Brown et al., 1991) using the HATCHES database (Cross and Ewart, 1989). The equilibrium data cannot be explained by congruent dissolution of biotite. The solution composition is rather controlled by: (1) exchange of interlayer ions; (2) incongruent dissolution and transformation of the silicate sheet (Fanning et al., 1989 and references therein); and (3) the subsequent precipitation of secondary phases. Silicon (Fig. 4a) and aluminium (Fig. 4b) concentrations appear to be regulated by simultaneous precipitation of kaolinite and gibbsite. A recent work of Nagy and Lasaga (1993) indicates the likelihood of such a process at low temperature. The formation of kaolinite during biotite weathering in natural system has been reviewed by Fanning et al. (1989), and recently reported by Banfield and Eggleton (1988) and Fordham (1990). Fig. 4c shows the total concentration of Fe. In this study we did not address redox speciation. Not knowing the redox speciation of iron in solution, we can only discuss the two limiting cases where the total iron is dominated by one of the redox states: (1) If total iron reflects Fe(III), i.e. Fe(I1) released from the biotite has been oxidized by trace amounts of oxygen, the solubility is apparently controlled by amorphous Fe(III)-hydroxide. (2) If, on the other hand, total iron is dominated by Fe(II), any Fe(II1) released from the biotite has precipitated as a less soluble Fe(III)-(hydr)oxide. The total iron concentration is too low to be controlled by a Fe(B)-(hydr)oxide or simple Fe(II)-silicate (fayalite). The concentration of magnesium (Fig. 4d) in solution is likewise far too low to be explained by brucite (Mg(OH),) or magnesium silicate (chrysotile, sepiolite, forsterite, enstatite) equilibrium. The solution concentration of these elements may instead be controlled by equilibrium with a complex alteration product of biotite. Secondary sheet silicates such as vermiculite, smectite and mixed-layer clays have been observed as weathering products of biotite in natural systems (Banfield and Eggleton, 1988; Fanning et al., 1989; Fordham, 1990). Alternatively, the concentration of Fe and Mg could be kinetically controlled. When biotite dissolves in a batch or titration experiment, the solution rapidly becomes oversaturated with respect to some secondary mineral phases. These phases may precipitate on the biotite surface, yielding a biotite core covered by precipitates and
M. Malmstr&
et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
(b) -2 ,,
6
7
8
9
10
7
6
8
PH
9
10
PH (d) -2
8
9
10
-107
’
6
I 7
8
4
’ 9
10
PH
PH
-107 6
, 7
8
9
10
PH Fig. 4. Equilibrium concentrations of elements released to solution during biotite dissolution ( ??= batch data; ? ?= titration data). Model lines, calculated by the HARPHREEQ code (Brown et al., 1991) using the HATCHES database (Cross and Ewart, 1989), represent solubility of some possible secondary phases: (a) silicon (- - - = kaolinite; -= quartz); (b) aluminium (- ~ - = kaolinite; - - = gibbsite); (c) iron (- - = Fe(Q%,,, fresh); -= Fe(Q%,,, agedj); (d) magnesium; and (e) potassium.
alteration products. Elements that are not in equilibrium with secondary phases may be prevented from further accumulation in solution due to dissolution inhibition by weathering products blocking the reactive biotite surface. Fig. 4e shows the potassium concentrations in solution which are approximately constant (lop3 M) due presumably to the ion-exchange reaction. The results from the batch and the titration experiments are in fair agreement. The silicon data show the biggest discrepancy with lower concentrations obtained from the batch experiment. This may be due to extensive alteration of the solid phase at pH 4, prior to equilibration at higher pH during the continuous titration.
M. MalmstrGm et al. /Journal
4. Implications
for groundwater
of Contaminant Hydrology 21 (1996) 201-213
211
redox chemistry
An operating deep repository for high-activity nuclear waste will be open to oxic conditions. After closing the repository, what are the time scales to develop reducing conditions? We consider two cases: (1) depletion of dissolved molecular oxygen (i.e. reaching anoxic conditions); and (2) production of reducing capacity in solution, in the form of dissolved Fe(U), after oxygen depletion. The following equation calculates the time scale to deplete dissolved molecular oxygen: = [021dissolved thj
7anoxic
(3)
R O,(uptake)
is the rate of oxygen consumption (mol L-’ hP ‘1. ROz(uptake) Assuming that the release rate of iron from biotite or chlorite is limiting the following reaction:
where
4Fe(II)
+ 0, + 4Hf+
4Fe(III)
the rate of oxygen uptake is calculated R O,(uptake)
AWfmRdiss.PFe(ll) =
the rate of
+ 2H,O by: (mol L-’ hh’)
4
where A, = wetted surface area Cm2 L-r); f,,,= mole fraction of mineral m (moles mineral/moles total); and PFecl,) = mole fraction of Fe(II1 in mineral m (moles Fe(II)/moles mineral). In this case the turnover time of molecular oxygen is estimated on the order of 1O1-lo2 yr (Table 3). Alternatively, oxygen may react directly with Fe(U) sites in the rock. White and Yee (19851 reported oxygen uptake by iron silicates.
Table 3 Characteristic Mineral
Biotite Chlorite Hornblende Augite
time scales for oxygen depletion
and development
of reducing
capacity
in solution
Oxygen depletion based on uptake by minerals bX‘, d
Build-up of Fe(U) in solution asc, e
(yd
(yd
(days)
52 285 n.a. n.a.
240 n.a. 59 84
189 2,083 n.a. n.a.
Oxygen depletion based on iron release rate a, b, ’
n.a. = data not available. a From this study. b We used A w = 10 m* L-’ determined for the wetted surface area of a fracture zone during the Stripa (Sweden) field investigations (Andersson et al., 1989). We used f, = 0.1 according to fb,o,ite = 0.1 in the host rock of a fracture zone at the &pii Hard Rock Laboratory, Sweden (Banwart et al., 1994). ’ We assumed an initially oxic (8 mg L- ’ 0,) solution. d Estimated using data from White and Yee (1985, fig. 12). e We used a characteristic concentration of Fe(B), [Fe(B)] = lo-’ mot L- ‘, observed in an undisturbed fracture zone at 70-m depth at the Aspi5 Hard Rock Laboratory (Banwart et al., 1994).
212
M. Malmstrijm et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
Using their steady-state rates of oxygen consumption to estimate a turnover time of molecular oxygen yields a slightly higher value than what is obtained by using our iron release rates (Table 3). Their experiments were performed under oxic conditions, which may better represent field processes at the time of repository operation. Analogous to Eqs. 3 and 4, we estimated a time scale of 10 months to reach a Fe(I1) concentration of lo-’ mol L- ’ (characteristic for undisturbed conditions in a fracture zone at Asp& Sweden; Banwart et al., 1994) in a granitic groundwater under anoxic conditions (Table 3). We do not imply by these calculations that these processes will necessarily be controlling oxygen consumption in a granitic aquifer, or that data obtained on a laboratory time scale can accurately be extrapolated to time scales of centuries, especially for natural systems where reactive wetted surface area, mineralogy, and kinetic pathways are uncertain. These turnover times, however, can be compared with time scales of hydrologic transport, diffusion in the rock matrix, and microbiological processes, in order to better formulate conceptual and mathematical models for transport and reaction in deep aquifers.
Acknowledgements This work was financed by The Swedish Company (SKB), Stockholm.
Nuclear
Fuel and Waste
Management
References Acker, J. and Bricker, P., 1992. The influence of pH on biotite dissolution and alteration kinetics at low temperature. Geochim. Cosmochim. Acta, 56: 3073-3092. Andersson, P., Gustafsson, E. and Olsson, O., 1989. Investigations of flow distribution in a fracture zone at the Stripa mine, using the radar method, results and interpretation. SKI3 (Swed. Nucl. Fuel Waste Manage. Co.), Stockholm, Tech. Rep. 89-33. Bailey, S.W., 1972. Determination of chlorite compositions by X-ray spacings and intensities. Clays Clay Miner., 22: 381-388. Banfield, J. and Eggleton, R., 1988. Transmission electron microscope study of biotite weathering. Clays Clay Miner., 36(l): 47-60. Banwart, S., Gustafsson, E., Laaksoharju, M., Nilsson, A.-C., Tullborg, E.-L. and Wallin, B., 1994. Large scale intrusion of shallow water into a granite aquifer. Water Resour. Res., 30(6): 1747-1763. Barnhisel, R. and Bertsch, P.M., 1989. Chlorites and hydroxy-interlayered vermiculite and smectite. In: J. Dixon and S. Weed (Editors), Minerals in Soil Environments. Soil Sci. Sot. Am., Madison, WI, pp. 729-787. Berg, A. and Banwart, S., 1994. Anorthite surface speciation and weathering reactivity in bicarbonate solutions at 25°C. In: J. Paul and C.-M. Pradier (Editors), Carbon Dioxide Chemistry: Environmental Issues. The Royal Society of Chemistry, Cambridge, pp. 305-316. Blum, A. and Lasaga, A., 1988. Role of surface speciation in the low-temperature dissolution of minerals. Nature (London), 331: 431-433. Brown, P.O., Howards, A., Sharland, S.N. and Tweed, C.J., 1991. HARPHREEQ - Geochemicdl speciation program based on PHREEQE user guide. Theor. Stud. Dep., Radwaste Disposal Div., Hornwall Lab., Didcot, Rep. B424.2.
M. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
213
Bruno, J., Casas, I. and Puigdomsnech, I., 1991. The kinetics of dissolution of UO, under reducing conditions and the influence of an oxidized surface layer (UO 2+ .): Application of a continuous flow through reactor. Geochim. Cosmochim. Acta, 55: 647-658. Bruno, J., Stumm, W., Wersin, P. and Brandberg, F., 1992. On the influence in mineral dissolution, I. The thermodynamics and kinetics of hematite dissolution in bicarbonate solution at T = 25°C. Geochim. Cosmochim. Acta, 56: 1139-1147. Caroll-Webb, S. and Walther, J.V., 1988. A surface complex reaction model for the pH-dependence of corundum and kaolinite dissolution rates. Geochim. Cosmochim. Acta, 52: 2609-2623. Cross, J.E. and Ewart, F.T., 1989. HATCHES - A thermodynamic database management system. Proc. Conf. on Chemistry and Migration Behaviour of Actinides and Fission Products in the Geosphere, 1989. Fanning, D., Keramidas, V. and El-Desoky, M., 1989. Micas. In: J. Dixon and S. Weed (Editors), Minerals in Soil Environments. Soil Sci. Sot. Am., Madison, WI, pp. 551-634. Fordham, A., 1990. Weathering of biotite into dioctahedral clay minerals. Clay Miner., 25: 51-63. Furrer, G. and Stumm, W., 1986. The coordination chemistry of weathering, I. Dissolution kinetics of rAlu,O, and BeO. Geochim. Cosmochim. Acta, 50: 1847-1860. Gran, G., 1981. Calculation of equivalence volumes in potentiometric titrations. Ph.D. Thesis, Department of Analytical Chemistry, The Royal Institute of Technology, Stockholm. Grenthe, I., Stumm, W., Laaksoharju, M., Nilsson, A.-C. and Wikberg, P., 1992. Redox potentials and redox reactions in deep groundwater systems. Chem. Geol., 98: 131-150.. Holdren, G. and Berner, R., 1979. Mechanism of feldspar weathering, I. Experimental studies. Geochim. Cosmochim. Acta, 43: 1161-1171. Nagy, K. and Lasaga, A., 1993. Simultaneous precipitation kinetics of kaolinite and gibbsite at 80°C and pH 3. Geochim. Cosmochim. Ada, 57: 4329-4335. Newman, A.C.D. and Brown, G., 1987. The chemical constitution of clay. In: A.C.D Newman (Editor), Chemistry of Clays and Clay Minerals. Longman, Essex, pp. 1-128. Pirhonen, V. and PitkInen, P., 1991. Redox capacity of crystalline rocks - Laboratory studies under 100 bar oxygen pressure. SKI3 (Swed. Nucl. Fuel Waste Manage. Co.), Stockholm, Tech. Rep. 91-55. Ross, G.J., 1969. Acid dissolution of chlorites: Release of magnesium, iron and aluminium and mode of acid attack. Clays Clay Miner., 17: 347-354. Schnitzer, M. and Kodama, H., 1976. The dissolution of micas by fulvic acid. Geoderma, 15: 381-391. Schnoor, J.L., 1990. Kinetics of chemical weathering: A comparison of laboratory and field weathering rates. In: W. Stumm (Editor), Aquatic Chemical Kinetics. Wiley, New York, NY, pp. 475-504. Schott, J. and Bemer, R., 1985. Dissolution mechanisms of pyroxenes and olivines during weathering. In: J.I. Drever (Editor), The Chemistry of Weathering. Reidel, Dordrecht, pp. 35-53. Schott, J., Bemer, R. and Sjiiberg, L., 1981. Mechanism of pyroxene and amphibole weathering, I. Experimental studies of iron-free minerals. Geochim. Cosmochim. Acta, 45: 2123-2135. Stumm, W. and Wieland, E., 1990. Dissolution of oxide and silicate minerals: Rates depend on surface speciation. In: W. Stumm (Editor), Aquatic Chemical Kinetics. Wiley, New York, NY, pp. 367-400. Stumm, W. and Wollast, R., 1990. Coordination chemistry of weathering: Kinetics of the surface controlled dissolution of oxide minerals. Rev. Geophys., 28(l): 53-69. Trotignon, L. and Turpault, M.-P., 1992. The dissolution kinetics of biotite in dilute HNO, at 24°C. In: Y.K. Kharaka and A.S. Maest (Editors), Proceedings of the 7th International Symposium on Water-Rock Interaction-WRI-7, Park City, Utah (USA), 13-18 July 1993. Balkema, Rotterdam, pp. 847-850. Wehrli, B., 1989. Monte Carlo simulations of surface morphologies during mineral dissolution. J. Colloid Interface Sci., 132(l): 230-242. White, A.F. and Yee, A., 1985. Aqueous oxidation-reduction kinetics associated with coupled electron-cation transfer from iron-containing silicates at 25°C. Geochim. Cosmochim. Acta, 49: 1263-1275. Wieland, E. and Stumm, W., 1992. Dissolution kinetics of kaolinite in acidic aqueous solutions at 25°C. Geochim. Cosmochim. Acta, 56: 3339-3355. Wieland, E., Wehrli, B. and Stumm, W., 1988. The coordination chemistry of weathering, III. A generalization on the dissolution rates of minerals. Geochim. Cosmochim. Acta, 52: 1969-1981. Wogelius, R. and Walther, J., 1991. Olivine dissolution at 25°C: Effects of pH, CO, and organic acids. Geochim. Cosmochim. Acta, 55: 943-954.
Contaminant Hydrology ELSEVIER
Journal of Contaminant
Hydrology
21(1996)
201-213
The dissolution of biotite and chlorite at 25°C in the near-neutral pH region Maria Malmstrijm
a, Steven Banwart aT*, Jeanette Lewenhagen Lara Duro b, Jordi Bruno ’
a,
aDepartment of Inorganic Chemistry, The Royal Institute of Technology,Stockholm, Sweden ’Department d’Enginyeria Quimica, Universitat Polit&nica de Catalunya,Barcelona, Spain ’h4BT Technologica Ambiental,Cerdanyola, Spain Received 17 December
1993; accepted
1.5 December
1994 after revision
Abstract We studied
the dissolution
of biotite
and chlorite
in laboratory
systems
with flow-through
and
pH region, under N2(sj atmosphere is highly non-stoichiometric. A slow linear release of iron during a period of weeks indicates a surface-chemical-reaction-controlled mechanism of release for iron. The release of potassium is much faster than that of iron. A parabolic dependence of accumulated release with time suggests a diffusion-controlled mechanism of potassium release. The rates of magnesium, aluminium and silicon release are between those for potassium and iron and approach that of iron with time. There is no significant influence of (bi)carbonate or pH on biotite dissolution rate or stoichiometry in the pH region 7 < pH < 8.5. The release rates of elements from chlorite are close to stoichiometric and similar to the iron release rate from biotite. In closed batch reactors at near-basic pH the composition of test solutions in contact with biotite is apparently controlled by gibbsite (Al), kaolinite (Si) and Fe(III)-(hydrjoxide. We estimated a turn-over time (lo’-lo2 yr) for molecular oxygen and a time scale (10 months) to develop characteristic Fe(U) concentrations for a granitic groundwater. batch
reactors.
The
initial
dissolution
of
biotite
in the
near-neutral
1. Introduction The most critical safety aspect for long-term disposal of high-activity nuclear waste is the design of engineered barriers for waste containment. Safety assessment, however,
* Corresponding author. Present address: Department Bradford, Bradford, BD7 lDP, UK. Elsevier Science B.V. SSDI 0166-3542(95)00047-X
of Civil and Environmental
Engineering,
University
of
202
M. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
must consider eventual failure of these barriers. In this case, the time scale for migration of long-lived radionuclides from a deep aquifer to the biosphere is dependent on hydrologic processes and aqueous speciation controlling solubility and adsorption behaviour. Redox conditions are especially important concerning the migration of uranium, neptunium, plutonium and technetium. All form sparingly soluble (hydr)oxides under reducing conditions but are highly soluble, and thus mobile, in their higher
T
14A
Fig. 1. Idealized crystal structure [modified from Bamhisel and Bertsch (1989)] of: (a) biotite; and (b) chlorite. An octahedral sheet of cations lies between by two tetrahedral sheets of silicon and aluminium. Biotite maintains charge balance by K + ions whereas chlorite has a hydroxy-interlayer.
M. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
203
valence states. During its operating lifetime, a deep repository for high-activity nuclear waste will be open to oxic conditions and surface water inflow. After closing the repository, it is important to anticipate the extent of the oxidizing disturbance, and the time scale for the aquifer to return to reducing conditions. In addition to respiration of organic carbon, uptake of molecular oxygen by reducing minerals can contribute to the development of anoxia in groundwaters (Grenthe et al., 1992). Molecular oxygen can be reduced directly at Fe(I1) sites on the mineral surfaces, or in solution subsequent to release of structural Fe(I1) (White and Yee, 1985). Biotite and chlorite are two important reservoirs of Fe(I1) in granitic bedrocks (Banwart et al., 1994). Here we report laboratory results on the thermodynamics and kinetics of biotite dissolution, and chlorite dissolution kinetics, at neutral and near-basic pH. These wet chemical experiments do not specifically address structural transformation of the mineral phases or redox speciation in solution, but focus on the stoichiometry of element release. Specific objectives are: (1) to determine the rate and stoichiometry of biotite and chlorite dissolution under near-neutral anoxic conditions; (2) to assess the effect of the (bi)carbonate ligand on biotite dissolution kinetics; and (3) to assess major-element speciation, and saturation with respect to possible secondary phases, in solutions reacted with biotite.
Table 1 Mineralogical
data on the specimen Biotite
Chlorite
Elemental ratio:
Si Fe Mg Al K
3.81 1.65 1 1.50 0.95
Fe(U) /Fe(tot)
kO.6 +O.l + 0.04 + 0.1
34.2 f 0.6 1 26.9 + 0.9 6.7kO.l -
ratio (o/o):
Spectrophotometric Mijssbauer
a
74.9 f 4.8 85.5
70.9 f 5.3 52.1
X-ray d(001) spacing (A)
14.40
Specific surface area (m’ g- I):
75-125 125-300
/em wm
Error estimates are a Method modified b Determined from ’ Determined from
1.81 +O.l b given as one standard deviation. from Pirhonen and Pitklnen (1991). Brunauer-Emmett-Teller adsorption BET analysis using helium.
4.38 ’ 3.36 ’
isotherm (BET) analysis using krypton.
204
M. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
Previous laboratory investigations of biotite (Schnitzer and Kodama, 1976; Acker and Bricker, 1992; Trotignon and Turpault, 1992 - up to pH 7) and chlorite (Ross, 1969) dissolution kinetics focused on the acid pH region.
2. Experiments 2.1. Mineral characterization
and sample preparation
Biotite (Newman and Brown, 1987; Fanning et al., 1989) and chlorite (Newman and Brown, 1987; Barnhisel and Bertsch, 1989) are structurally similar (Fig. 1) consisting of a sheet of octahedrally coordinated cations surrounded by two tetrahedrally coordinated sheets of silicon and aluminium. Biotite maintains charge balance with exchangeable K+ ions in the silicate interlayers whereas chlorite has brucite/gibbsite-like hydroxyinterlayers. Naturhistoriska Riksmuseet (Stockholm, Sweden) provided the natural mineral specimens used in this study (Biotite LK 5210, Chlorite 740665). Table 1 lists characteristics of the mineral specimens. Powder X-ray diffractometry confirmed the minerals to be biotite and chlorite with no detectable crystalline impurities. Based on the results in Table 1 and assignment of structural positions according to Newman and Brown (1987) and Bailey (1972, equation 1) the mineral compositions are: Biotite:
K o.so(Mg,.,,Fe:‘.,,Fe~.~~~~.~~)(Si,.,,~”.,,)O,”(OH)*
Chlorite:
(Mg,.,~o.~FeZ’.,Fe~.‘,)(Si,.,~,.,)O,,(OH)s
FEEDVESSEL
u
MEASURING CELL
REACTOR
_SAMPLE BOTTLE Fig. 2. Experimental setup for the thin-film continuous flow-through system. Test solution of known composition is pumped through a thin-film reactor where the mineral powder is held between two membrane filters. The outflow solution passes a closed U-tube where pH is measured continuously. The reacted solution is collected in acid-containing polypropylene bottles.
M. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
205
Mineral powder was prepared by peeling small sheets from the mineral surface, size fractionating with sieves, and washing ultrasonically three times in ethanol to remove ultrafine particles (Schott et al., 1981). 2.2. Experimental
procedure
Experiments were carried out at 25 f 1°C 1-atm pressure, using the ionic medium: [Na+] + [H+] = [HCO;] + [ClO;] + [OH-] = 0.500 M. High-purity chemicals (pro analysi or better) and gases (99.5% purity) were used. Gases were cleaned by bubbling through an oxygen trap (amalgamated zinc/vanadous solution) and a wash bottle containing ionic medium to saturate the gas stream with water vapour at experimental ionic strength and temperature. Reacted solution samples were analyzed for Si, Mg, Al, total Fe and K where applicable using an ARL” (Applied Research Laboratory) model 3520B ICP (inductively coupled plasma emission spectrometer) analyzer. Standards were prepared by dilution of Atomic Spectral Standards (Analytical Standards, Kungsbacka, Sweden) into appropriate matrix solutions. Proton concentrations were determined using a glass electrode-Ag/AgCl reference electrode couple. The electrode couple was normally calibrated according to Gran (1981) before and after each experiment. Membrane filters used in the experiments were 0.2-pm pore size filters of PTFE-type (polytetrafluoroethylene, MFS@ J020A025A or Millex” FG SLFG025BS). 2.2.1. Kinetic experiment The experiments were performed using a continuous flow-through reactor (Bnmo et al., 1991) (Fig. 2), consisting of a thin layer (< 3 mm) of mineral powder of known mass, held between two 25-mm-diameter membrane filters, across which test solutions of known composition were pumped at specified rates. The pH of the reactor outflow was recorded continuously. The following equation calculates the surface area normalized mineral dissolution rate from concentration [C (mol/g solution)] of element j in sample i: Ri,j
=
Fici,j mop
(moles of mineral hh’ m-‘) I
(1)
where F = flow rate (g h- ’); m = mass of mineral (g); A = specific surface area (m* g-l); and P is a stoichiometric factor (moles of element/moles of mineral). The following equation calculates the accumulated release of an element during an experiment: n,( ti) = zR,,/(
ti - ti_ 1)
(moles of mineral m-‘)
(2)
where n denotes accumulated release (moles of mineral m-*1; and t the time (h) elapsed from the start of the experiment. In addition to continuously purging feed solutions with N,,,, the carbon dioxide-free experiments were carried out in a glove box under N2(g) atmosphere. In these experiments the proton concentrations in the feed solutions were calculated from dilution of
206
M. Malmstriim et al./Journal
of Contaminant Hydrology 21 (1996) 201-213
standard acid (HClO,) or base (NaOH) solutions in NaClO,. A fresh mineral sample was used for each experiment. Solution samples were collected in polypropylene, acid-washed bottles, containing a small volume of concentrated HNO,. 2.2.2. Equilibrium
experiments
2.2.2.1. Batch experiments. A known mass of mineral was placed in a Teflon” or polypropylene bottle. The bottle was filled with ionic medium at specified pH, sealed and placed onto a shaker table under NX,) atmosphere. The system was left to equilibrate for 60 days. The final pH was measured, the solid phase separated by membrane filtration and the filtrate acidified (to pH 5 1) by a small volume of nitric acid. 2.2.2.2. Titration experiments. The titration was started at pH = 4 and was brought to pH 10.7 by consecutive base additions. After each addition, equilibrium was assumed when the potential drift in the glass electrode assembly was < 10m4 mV ss’. At equilibrium the solution was sampled by removal of some suspension, separation of the solid phase by membrane filtration and acidification (to pH 5 1) of the filtrate by a small volume of nitric acid. The titration vessel was maintained under Nzcgj atmosphere.
3. Results and discussion 3.1. Kinetic experiment Fig. 3a shows the accumulated release of elements from biotite with time at pH 7 and 1 atm N2(sj. The release of all elements is rapid at the beginning of the experiment and decreases to quasi steady-state rates with time. Such a decrease of dissolution rates with time is typical for dissolution kinetic experiments using fresh mineral samples (Furrer and Stumm, 1986; Caroll-Webb and Walther, 1988; Wieland and Stumm, 1992), and is due to dissolution of ultrafine particles (Holdren and Berner, 1979; Schott et al., 1981) and more reactive surface sites (Wehrli, 1989). The dissolution is initially highly non-stoichiometric with total iron release relatively slow and linear with time. Linear release, obtained at a fixed flow rate, reflects constant output concentrations from the thin-film reactor. This may be due to steady-state kinetics or mineral-solution equilibrium. At the end of this experiment we ran the thin-film reactor at half the rate, yielding approximately double concentrations of iron, aluminium and magnesium in the outflow solution, excluding the possibility of equilibrium conditions or transport-controlled release of these elements. Steady-state kinetics is expected for a surface-chemical-reaction-controlled dissolution mechanism with constant concentration of reactive surface iron sites (Furrer and Stumm, 1986; Blum and Lasaga, 1988; Wieland et al., 1988; Stumm and Wieland, 1990; Stumm and Wollast, 1990). Potassium release is very fast at the beginning of the experiment and apparently becomes diffusion controlled after N 10 days as indicated by a linear plot of accumulated release against Jt (not shown; Schott and Berner, 1985). Aluminium and magnesium release rates are between the extremes
of Contaminant Hydrology 21 (1996) 201-213
M. Malmstriim et al. /Journal
.2E-O3
5E-05
7
E
3;
2
g
207
4E-05 3E-05
“6 2E-05 I. 2 IE-05 c
0
0 0
5
10
20
15
25
30
Time [Days]
-
3E-05
- 2.OE-03
-lbj
3
0
5
10
15
20
25
30
Time [Days] Fig. 3. Accumulated release of minerals calculated (Eq. 2) for each element (A = Si; v = Fe; ??= Al; X = K) at three different experimental conditions. a. Biotite dissolution pH 7 (CO,-free). b. Biotite dissolution pH 8.1, 1% COys,. c. Chlorite dissolution pH 8.2, 1% COq,). The scale for K+ release is shown on the vertical axis at the right. The experiment at pH 7 was run flow rate during the final 3 days of the experiment yielding double concentration of Fe, Mg and outflow solution. This excludes the possibility of equilibrium or transport controlled release of these
0 = Mg;
at half the Al in the elements.
exhibited by Fe and K, and approach the iron release rate with time. Except for the anomalous increase after day 20, Si release follows that of Al and Mg. Table 2 lists the dissolution rates calculated from each element in four kinetic experiments. We want to emphasize that these experiments were performed in a flow-through system where the feed solution flushes the released elements from the reactor, preventing precipitation of secondary reaction products. A scanning electron
.pallodal uaaq amq aleuoq.m(!q) dq uognlossyp p2lauy 30 (1661 ‘.IaqlleM pm snga%!oM) uoyq~qu! pue (~661 ‘JEMU~~ pm %a8 fz661 ‘y2 la 0un.Q uogwalame qloq amle~alg aql UI .uo+?al Hd lw)nau-Jeau aql u! d_IlauIo!q~~ols IO alw uoyI~oss!p uo Hd JO aleuoqlm(!q) 30 lDa33a lue~3!u8!s ou s! alaql leql s~oys (z a[ql?J_,) sluampadxa algo!q aalql aql 30 uosydmo:, v .luawpadxa sno!AaId aql 01 .w[y!s s! sluaura[a 11~ 30 mo!lzeqaq aq$ SMOGS qE ‘%d aqL .(“)zo3 urle I()‘() lapun 1’8 Hd le algo!q 30 uoyqoss!p waw!ladxa aql30 pua aql 1~ ad 30 leql my1 Jai923 [I!ls s! 9 aurg qi!~ shoos +x 30 asea[al aql q8noql -w .aseaIal IuauraIa pa~~orluo3-uo!l3ea~-le3ruray3-a~~~ns 1saMoIs aql ‘dq pal!uuI pm ‘01 zyawoyq~~ols ‘asealal Jeaug 01 qzoldde (0 pue flakI palaIdap E q8nolql uo:sn33Fp 01 anp aseaIaJ 3qoqeled 01 uo!l!sue~l (z) fsal!s aa!lzzal ~1013 aseaIa1 paIIoquoD-uog -3eal-Ie~ruIaq~-a3e3ms lse3 (1) :sa%ls aaJq1 qZnoJq$ sassed uoyqoss!p cgaworq~rols 01 uo!ssa.Bold aqL .qled uoyg!p IaZuo[ d[aA!ssaJ8old e 30 u0ym.103 01 anp ‘(ad) pasl?aIaJ 4uauIaIa jsamop aql ql!M ~!J4amo~q3!ols auIoDaq II!M +x paseaIa1 dIp!del A~pz!lru! aql ‘au!1 qlrm leql swpald [apow s!qL .palyg-aw uO!Sn33fp sauro9aq asea1a.I +‘>I aql qD!qM q8nolql uo@aI palaldap-+x e surlo3 pm s~adepalu! al!lo!q aql u10_13+y saAouIal +H Pue +EN uognlos q$!M a8ueqDxa p!dea Xap!xo xalduro3 30 uognIoss!p luan.x%om! 103 (0661) loouqDg 30 [apour Ienldamo3 aql E?uysn slInsa1 asaql 1aJdIalur aM ‘al!loTq papeamn aql ~11013samala33rp p@oloq&ouI JO samydr3ald ou smoqs L Hd le palXal aIdtues algoyq aql 30 (umoqs IOU) a&u! (~2s) adoD,soguI
_
_
pq
,_01.8’6 uoyalap .+tolaq
L
-01
g_OI.P’9
(Z’O T P’Z)
,_OI.Z'Z b-OI.(Z’ZTP’9)
,_01.9'9
,_01.8'S ,_oI~(S'o+I'z)
~_OI~(Z'ZTSS-)
,_OI.Z'I b_01.(9'0TE'6)
g-01.8'2 g-OI~(P'OTO'1)
'N"! '03
pOI.1’9
i -01 ‘(O’I T P.9)
pOI.5'9 x _OI'(Z'IT6'2)
I(_OI'E'P ,_OI~(Z'IfPW
i_OI,C'I i_oI'(z'oTz'I)
pOI,E’9
i_oI.(o’IT9’s)
L-01 ,O'I ,_OI.(L'OTZ'L)
+01.0'8 b-oI~(z‘Ir8'L)
i-OI.O'I i_OI.(I'OT1'I)
xu
(wu!) (te=g) x~
(tevur) lv'tl (rw3) ?
(wu!)
"wbl
(vv3) Bwx
(p!p!) !slf (lF3) lS2f
'NU! '03 %Z0'0f00'1 avwq
%z0'0T00'1 aiypp 2'8
'N %OOI appq
I'8
(z ,ba) luaruala qxa
'N%OOI avwq
S'8
svo le_raugq Hd
L
~013 pa,epm[e3 (, _q z_tu [eJau!m salour) saw uoqn[oss!a z at9e.L
A4. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
209
Fig. 3c shows the dissolution of chlorite at pH 8.2 under 0.01 atm CO,,,,. The final accumulated release of chlorite is much lower than for biotite (Fig. 3a and b) due mainly to a lower initial dissolution rate. Incorporation of hydrated ions into the biotite interlayer expands the mica structure and exposes additional mineral surface to solution attack. This is not expected for the chlorite hydroxy-interlayers. The initial release rates of the different elements are nearly stoichiometric. After 16-17 days the outflow concentrations of iron and aluminium are below detection limits. However, the final dissolution rate (Table 2), calculated from silicon and magnesium release, is similar to that of biotite. 3.2. Equilibrium
experiments
Fig. 4a-e shows the equilibrium concentration of elements released from biotite as a function of pH in the neutral to near-alkaline region. The plotted lines represent the solubility of possible secondary phases calculated by HARPHREEQ (Brown et al., 1991) using the HATCHES database (Cross and Ewart, 1989). The equilibrium data cannot be explained by congruent dissolution of biotite. The solution composition is rather controlled by: (1) exchange of interlayer ions; (2) incongruent dissolution and transformation of the silicate sheet (Fanning et al., 1989 and references therein); and (3) the subsequent precipitation of secondary phases. Silicon (Fig. 4a) and aluminium (Fig. 4b) concentrations appear to be regulated by simultaneous precipitation of kaolinite and gibbsite. A recent work of Nagy and Lasaga (1993) indicates the likelihood of such a process at low temperature. The formation of kaolinite during biotite weathering in natural system has been reviewed by Fanning et al. (1989), and recently reported by Banfield and Eggleton (1988) and Fordham (1990). Fig. 4c shows the total concentration of Fe. In this study we did not address redox speciation. Not knowing the redox speciation of iron in solution, we can only discuss the two limiting cases where the total iron is dominated by one of the redox states: (1) If total iron reflects Fe(III), i.e. Fe(I1) released from the biotite has been oxidized by trace amounts of oxygen, the solubility is apparently controlled by amorphous Fe(III)-hydroxide. (2) If, on the other hand, total iron is dominated by Fe(II), any Fe(II1) released from the biotite has precipitated as a less soluble Fe(III)-(hydr)oxide. The total iron concentration is too low to be controlled by a Fe(B)-(hydr)oxide or simple Fe(II)-silicate (fayalite). The concentration of magnesium (Fig. 4d) in solution is likewise far too low to be explained by brucite (Mg(OH),) or magnesium silicate (chrysotile, sepiolite, forsterite, enstatite) equilibrium. The solution concentration of these elements may instead be controlled by equilibrium with a complex alteration product of biotite. Secondary sheet silicates such as vermiculite, smectite and mixed-layer clays have been observed as weathering products of biotite in natural systems (Banfield and Eggleton, 1988; Fanning et al., 1989; Fordham, 1990). Alternatively, the concentration of Fe and Mg could be kinetically controlled. When biotite dissolves in a batch or titration experiment, the solution rapidly becomes oversaturated with respect to some secondary mineral phases. These phases may precipitate on the biotite surface, yielding a biotite core covered by precipitates and
M. Malmstr&
et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
(b) -2 ,,
6
7
8
9
10
7
6
8
PH
9
10
PH (d) -2
8
9
10
-107
’
6
I 7
8
4
’ 9
10
PH
PH
-107 6
, 7
8
9
10
PH Fig. 4. Equilibrium concentrations of elements released to solution during biotite dissolution ( ??= batch data; ? ?= titration data). Model lines, calculated by the HARPHREEQ code (Brown et al., 1991) using the HATCHES database (Cross and Ewart, 1989), represent solubility of some possible secondary phases: (a) silicon (- - - = kaolinite; -= quartz); (b) aluminium (- ~ - = kaolinite; - - = gibbsite); (c) iron (- - = Fe(Q%,,, fresh); -= Fe(Q%,,, agedj); (d) magnesium; and (e) potassium.
alteration products. Elements that are not in equilibrium with secondary phases may be prevented from further accumulation in solution due to dissolution inhibition by weathering products blocking the reactive biotite surface. Fig. 4e shows the potassium concentrations in solution which are approximately constant (lop3 M) due presumably to the ion-exchange reaction. The results from the batch and the titration experiments are in fair agreement. The silicon data show the biggest discrepancy with lower concentrations obtained from the batch experiment. This may be due to extensive alteration of the solid phase at pH 4, prior to equilibration at higher pH during the continuous titration.
M. MalmstrGm et al. /Journal
4. Implications
for groundwater
of Contaminant Hydrology 21 (1996) 201-213
211
redox chemistry
An operating deep repository for high-activity nuclear waste will be open to oxic conditions. After closing the repository, what are the time scales to develop reducing conditions? We consider two cases: (1) depletion of dissolved molecular oxygen (i.e. reaching anoxic conditions); and (2) production of reducing capacity in solution, in the form of dissolved Fe(U), after oxygen depletion. The following equation calculates the time scale to deplete dissolved molecular oxygen: = [021dissolved thj
7anoxic
(3)
R O,(uptake)
is the rate of oxygen consumption (mol L-’ hP ‘1. ROz(uptake) Assuming that the release rate of iron from biotite or chlorite is limiting the following reaction:
where
4Fe(II)
+ 0, + 4Hf+
4Fe(III)
the rate of oxygen uptake is calculated R O,(uptake)
AWfmRdiss.PFe(ll) =
the rate of
+ 2H,O by: (mol L-’ hh’)
4
where A, = wetted surface area Cm2 L-r); f,,,= mole fraction of mineral m (moles mineral/moles total); and PFecl,) = mole fraction of Fe(II1 in mineral m (moles Fe(II)/moles mineral). In this case the turnover time of molecular oxygen is estimated on the order of 1O1-lo2 yr (Table 3). Alternatively, oxygen may react directly with Fe(U) sites in the rock. White and Yee (19851 reported oxygen uptake by iron silicates.
Table 3 Characteristic Mineral
Biotite Chlorite Hornblende Augite
time scales for oxygen depletion
and development
of reducing
capacity
in solution
Oxygen depletion based on uptake by minerals bX‘, d
Build-up of Fe(U) in solution asc, e
(yd
(yd
(days)
52 285 n.a. n.a.
240 n.a. 59 84
189 2,083 n.a. n.a.
Oxygen depletion based on iron release rate a, b, ’
n.a. = data not available. a From this study. b We used A w = 10 m* L-’ determined for the wetted surface area of a fracture zone during the Stripa (Sweden) field investigations (Andersson et al., 1989). We used f, = 0.1 according to fb,o,ite = 0.1 in the host rock of a fracture zone at the &pii Hard Rock Laboratory, Sweden (Banwart et al., 1994). ’ We assumed an initially oxic (8 mg L- ’ 0,) solution. d Estimated using data from White and Yee (1985, fig. 12). e We used a characteristic concentration of Fe(B), [Fe(B)] = lo-’ mot L- ‘, observed in an undisturbed fracture zone at 70-m depth at the Aspi5 Hard Rock Laboratory (Banwart et al., 1994).
212
M. Malmstrijm et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
Using their steady-state rates of oxygen consumption to estimate a turnover time of molecular oxygen yields a slightly higher value than what is obtained by using our iron release rates (Table 3). Their experiments were performed under oxic conditions, which may better represent field processes at the time of repository operation. Analogous to Eqs. 3 and 4, we estimated a time scale of 10 months to reach a Fe(I1) concentration of lo-’ mol L- ’ (characteristic for undisturbed conditions in a fracture zone at Asp& Sweden; Banwart et al., 1994) in a granitic groundwater under anoxic conditions (Table 3). We do not imply by these calculations that these processes will necessarily be controlling oxygen consumption in a granitic aquifer, or that data obtained on a laboratory time scale can accurately be extrapolated to time scales of centuries, especially for natural systems where reactive wetted surface area, mineralogy, and kinetic pathways are uncertain. These turnover times, however, can be compared with time scales of hydrologic transport, diffusion in the rock matrix, and microbiological processes, in order to better formulate conceptual and mathematical models for transport and reaction in deep aquifers.
Acknowledgements This work was financed by The Swedish Company (SKB), Stockholm.
Nuclear
Fuel and Waste
Management
References Acker, J. and Bricker, P., 1992. The influence of pH on biotite dissolution and alteration kinetics at low temperature. Geochim. Cosmochim. Acta, 56: 3073-3092. Andersson, P., Gustafsson, E. and Olsson, O., 1989. Investigations of flow distribution in a fracture zone at the Stripa mine, using the radar method, results and interpretation. SKI3 (Swed. Nucl. Fuel Waste Manage. Co.), Stockholm, Tech. Rep. 89-33. Bailey, S.W., 1972. Determination of chlorite compositions by X-ray spacings and intensities. Clays Clay Miner., 22: 381-388. Banfield, J. and Eggleton, R., 1988. Transmission electron microscope study of biotite weathering. Clays Clay Miner., 36(l): 47-60. Banwart, S., Gustafsson, E., Laaksoharju, M., Nilsson, A.-C., Tullborg, E.-L. and Wallin, B., 1994. Large scale intrusion of shallow water into a granite aquifer. Water Resour. Res., 30(6): 1747-1763. Barnhisel, R. and Bertsch, P.M., 1989. Chlorites and hydroxy-interlayered vermiculite and smectite. In: J. Dixon and S. Weed (Editors), Minerals in Soil Environments. Soil Sci. Sot. Am., Madison, WI, pp. 729-787. Berg, A. and Banwart, S., 1994. Anorthite surface speciation and weathering reactivity in bicarbonate solutions at 25°C. In: J. Paul and C.-M. Pradier (Editors), Carbon Dioxide Chemistry: Environmental Issues. The Royal Society of Chemistry, Cambridge, pp. 305-316. Blum, A. and Lasaga, A., 1988. Role of surface speciation in the low-temperature dissolution of minerals. Nature (London), 331: 431-433. Brown, P.O., Howards, A., Sharland, S.N. and Tweed, C.J., 1991. HARPHREEQ - Geochemicdl speciation program based on PHREEQE user guide. Theor. Stud. Dep., Radwaste Disposal Div., Hornwall Lab., Didcot, Rep. B424.2.
M. Malmstriim et al. /Journal
of Contaminant Hydrology 21 (1996) 201-213
213
Bruno, J., Casas, I. and Puigdomsnech, I., 1991. The kinetics of dissolution of UO, under reducing conditions and the influence of an oxidized surface layer (UO 2+ .): Application of a continuous flow through reactor. Geochim. Cosmochim. Acta, 55: 647-658. Bruno, J., Stumm, W., Wersin, P. and Brandberg, F., 1992. On the influence in mineral dissolution, I. The thermodynamics and kinetics of hematite dissolution in bicarbonate solution at T = 25°C. Geochim. Cosmochim. Acta, 56: 1139-1147. Caroll-Webb, S. and Walther, J.V., 1988. A surface complex reaction model for the pH-dependence of corundum and kaolinite dissolution rates. Geochim. Cosmochim. Acta, 52: 2609-2623. Cross, J.E. and Ewart, F.T., 1989. HATCHES - A thermodynamic database management system. Proc. Conf. on Chemistry and Migration Behaviour of Actinides and Fission Products in the Geosphere, 1989. Fanning, D., Keramidas, V. and El-Desoky, M., 1989. Micas. In: J. Dixon and S. Weed (Editors), Minerals in Soil Environments. Soil Sci. Sot. Am., Madison, WI, pp. 551-634. Fordham, A., 1990. Weathering of biotite into dioctahedral clay minerals. Clay Miner., 25: 51-63. Furrer, G. and Stumm, W., 1986. The coordination chemistry of weathering, I. Dissolution kinetics of rAlu,O, and BeO. Geochim. Cosmochim. Acta, 50: 1847-1860. Gran, G., 1981. Calculation of equivalence volumes in potentiometric titrations. Ph.D. Thesis, Department of Analytical Chemistry, The Royal Institute of Technology, Stockholm. Grenthe, I., Stumm, W., Laaksoharju, M., Nilsson, A.-C. and Wikberg, P., 1992. Redox potentials and redox reactions in deep groundwater systems. Chem. Geol., 98: 131-150.. Holdren, G. and Berner, R., 1979. Mechanism of feldspar weathering, I. Experimental studies. Geochim. Cosmochim. Acta, 43: 1161-1171. Nagy, K. and Lasaga, A., 1993. Simultaneous precipitation kinetics of kaolinite and gibbsite at 80°C and pH 3. Geochim. Cosmochim. Ada, 57: 4329-4335. Newman, A.C.D. and Brown, G., 1987. The chemical constitution of clay. In: A.C.D Newman (Editor), Chemistry of Clays and Clay Minerals. Longman, Essex, pp. 1-128. Pirhonen, V. and PitkInen, P., 1991. Redox capacity of crystalline rocks - Laboratory studies under 100 bar oxygen pressure. SKI3 (Swed. Nucl. Fuel Waste Manage. Co.), Stockholm, Tech. Rep. 91-55. Ross, G.J., 1969. Acid dissolution of chlorites: Release of magnesium, iron and aluminium and mode of acid attack. Clays Clay Miner., 17: 347-354. Schnitzer, M. and Kodama, H., 1976. The dissolution of micas by fulvic acid. Geoderma, 15: 381-391. Schnoor, J.L., 1990. Kinetics of chemical weathering: A comparison of laboratory and field weathering rates. In: W. Stumm (Editor), Aquatic Chemical Kinetics. Wiley, New York, NY, pp. 475-504. Schott, J. and Bemer, R., 1985. Dissolution mechanisms of pyroxenes and olivines during weathering. In: J.I. Drever (Editor), The Chemistry of Weathering. Reidel, Dordrecht, pp. 35-53. Schott, J., Bemer, R. and Sjiiberg, L., 1981. Mechanism of pyroxene and amphibole weathering, I. Experimental studies of iron-free minerals. Geochim. Cosmochim. Acta, 45: 2123-2135. Stumm, W. and Wieland, E., 1990. Dissolution of oxide and silicate minerals: Rates depend on surface speciation. In: W. Stumm (Editor), Aquatic Chemical Kinetics. Wiley, New York, NY, pp. 367-400. Stumm, W. and Wollast, R., 1990. Coordination chemistry of weathering: Kinetics of the surface controlled dissolution of oxide minerals. Rev. Geophys., 28(l): 53-69. Trotignon, L. and Turpault, M.-P., 1992. The dissolution kinetics of biotite in dilute HNO, at 24°C. In: Y.K. Kharaka and A.S. Maest (Editors), Proceedings of the 7th International Symposium on Water-Rock Interaction-WRI-7, Park City, Utah (USA), 13-18 July 1993. Balkema, Rotterdam, pp. 847-850. Wehrli, B., 1989. Monte Carlo simulations of surface morphologies during mineral dissolution. J. Colloid Interface Sci., 132(l): 230-242. White, A.F. and Yee, A., 1985. Aqueous oxidation-reduction kinetics associated with coupled electron-cation transfer from iron-containing silicates at 25°C. Geochim. Cosmochim. Acta, 49: 1263-1275. Wieland, E. and Stumm, W., 1992. Dissolution kinetics of kaolinite in acidic aqueous solutions at 25°C. Geochim. Cosmochim. Acta, 56: 3339-3355. Wieland, E., Wehrli, B. and Stumm, W., 1988. The coordination chemistry of weathering, III. A generalization on the dissolution rates of minerals. Geochim. Cosmochim. Acta, 52: 1969-1981. Wogelius, R. and Walther, J., 1991. Olivine dissolution at 25°C: Effects of pH, CO, and organic acids. Geochim. Cosmochim. Acta, 55: 943-954.