www.elsevier.com/locate/jcis
Modeling and electrokinetic evidences on the processes of the Al(III)
sorption continuum in SiO
2(s)
suspension
Wen Hui Kuan,
a,∗Shang Lien Lo,
band Ming Kuang Wang
caDepartment of Environmental and Safety Engineering, Ming-Chi Institute of Technology, 84, Gunjuan Rd., Taishan, Taipei Hsien 243, Taiwan bGraduate Institute of Environmental Engineering, National Taiwan University, Taipei 106, Taiwan
cDepartment of Agriculture Chemistry, National Taiwan University, Taipei 106, Taiwan
Received 30 June 2003; accepted 12 December 2003
Abstract
Reactions of Al(III) at the interface between SiO2(s) and aqueous solution were characteristically and quantitatively studied us-ing electrophoretic methods and applyus-ing a surface complexation/precipitation model (SCM/SPM). The surface and bulk properties of Al(III)/SiO2suspensions were determined as functions of pH and initial Al(III) concentration. Simulated modeling results indicate that the SCM, accounting for the adsorption mechanism, predicts sorption data for low surface coverage only reasonably well. Al(III) hy-drolysis and surface hydroxide precipitation must be invoked as the Al(III) concentration and/or pH progressively increase. Accordingly, the three processes in the Al(III) sorption continuum, from adsorption through hydrolysis to surface precipitation, could be identified by the divergence between the SCM/SPM predictions and the experimental data. SiO2(s) suspensions with low Al(III) concentrations (1× 10−4and 1× 10−5M) exhibit electrophoretic behavior similar to that of a pure SiO2(s)system. In Al(III)/SiO2systems with high Al concentrations of 1× 10−3, 5× 10−3and 1× 10−2M, three charge reversals (CR) are observed, separately representing, in order of increasing pH, the point of zero charge (PZC) on the SiO2substrate (CR1), the onset of the surface precipitation of Al hydroxide (CR2), and at a high pH, the PZC of the Al(OH)3coating (CR3). Furthermore, in the 1× 10−3M Al(III)/SiO2(s)system, CR2 is consistent with the modeling results of SCM/SPM and provides evidence that Al(III) forms a surface precipitate on SiO2(s)at pH above 4. SiO2(s)dissolution was slightly inhibited when Al(III) was adsorbed onto the surface of SiO2(s), as compared to the dissolution that occurs in a pure SiO2(s) suspension system. Al hydroxide surface precipitation dramatically reduced the dissolution of SiO2(s)because the Al hydroxide passive film inhibited the corrosion of the SiO2(s)surface by OH−ions.
2004 Elsevier Inc. All rights reserved.
Keywords: Al(III); Interface reactions; SiO2; Electrophoretic method; Surface complexation model; Surface precipitation model
1. Introduction
Silicon is the second most abundant element in the earth’s crust and is present as silicates and silica. Iler [1] divided silica into five different phases, of which anhydrous and hy-drous amorphous silica are the most important from a col-loidal perspective, since it is this phase that forms colcol-loidal and microporous structures. In acidic lakes and streams throughout the world, aluminum is present in concentrations from several µmol/l to several hundred µmol/l [2], that is, at a level toxic to aquatic life [3]. Since the toxicity of Al depends strongly on its speciation (free and complexed Al)
* Corresponding author.
E-mail address: whkuan@ccsun.mit.edu.tw (W.H. Kuan).
and mobility (soluble, colloid, or precipitated) [4,5], its inter-action with silica may significantly alter the bioavailability of this element and its detrimental effects on aquatic or-ganisms [6,7]. In the water-treatment industry the interac-tions between silica and aluminum also play a determining role, because Al salts are common coagulants and silica is a prevalent constituent of solid matter in raw water. Moreover, hydrolyzed aluminum is used to set rosin and dyes in the production of paper and to improve the retention of fillers. The increased use of hydrolyzed aluminum salts has height-ened commercial and research interest in aluminum hydrol-ysis [8,9]. A number of studies on aluminum hydrolhydrol-ysis have emphasized the stoichiometry or kinetics of complex hydrolysis, precipitation, and surface chemical reactions of a pure aluminum system [9,10]. Some researchers [11–14]
0021-9797/$ – see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2003.12.034
have pointed out that soluble aluminum has a high poten-tial to form mixed oxides with other component oxides in natural environments, and the characteristics of such mixed oxides deviate significantly from either average collective properties for the group or a collection of discrete pure solid phases. However, the understanding of when and how the sorption continuum for Al(III), from soluble free ions to a surface or bulk oxide precipitate, proceeds in SiO2(s)
suspen-sions is quite limited. Limited knowledge of these processes has restricted the accurate prediction of the distribution of Al(III) in aquatic systems and the adsorption properties of other trace elements onto the undefined surfaces of these mixed oxides.
When a metal ion is present in an oxide suspension sys-tem, a sorption continuum [15] may occur from mononu-clear adsorption to hydrolysis, to multinumononu-clear adsorption, or to precipitation on the surface of oxide or in bulk solu-tion; during these processes, system pH or sorbate/sorbent ratios increase [16–19]. At a low pH or sorbate/sorbent ra-tio, cations coordinate with surface functional groups of an oxide; the enhanced interaction between adjacent cations results in a sorption continuum with pH and/or the sor-bate/sorbent ratio increasing. Because adsorption, hydroly-sis, and precipitation of Al(III) occur over a very narrow pH range, each process of the sorption continuum is diffi-cult to interpret and distinguish [20,21]. Xu et al. [22] also indicated that such processes can proceed simultaneously, and distinguishing them requires analytical methods with molecular-scale resolution.
The following definitions are adopted in the discussion of the mechanism of the interface reactions. According to the relative concentrations and contact methods, interactions between hydrous oxides and metals include (1) sorption, re-ferring to the uptake of a dissolved metal by a solid phase, irrespective of the mechanism [16,23,24]; (2) adsorption, referring to a surface complexation reaction between sur-face sites and a metal sorbate [25]; and (3) sursur-face precip-itation, referring to the formation of a precipitate, induced by the sorbent surface under solution conditions that are undersaturated with respect to any known phase [26–28]. Depending on the type of interaction, metals exhibit vari-ous solubilities that can markedly alter their environmental mobility and removal efficiency during wastewater treat-ment [29–31].
This study aims to identify the reactions that occurred during this sorption continuum of Al(III) in SiO2(s)
suspen-sion and to clarify how environmental conditions influence these processes. The effects of a wide range of sorbate-to-sorbent ratios and pH were modeled using SCM/SPM to clarify the onset of each process in the sorption con-tinuum. Electrokinetic measurements were performed to determine the nature of the Al(III) interaction with the SiO2(s) surface [32]. The measurement of silica
dissolu-tion yielded informadissolu-tion on the onset of surface precipita-tion.
2. Materials and methods
2.1. Batch experiments
Batch experiments concerning Al(III) sorption onto SiO2
were conducted in 50-ml polypropylene bottles with caps. The used silica (SiO2) was Cab-O-Sil M5 (Cabot Corp.,
Tus-colca, IL), amorphous and fumed silica with a BET surface area of 200 m2/g (analyzed by ASAP 2000). Before the batch experiments were performed, the SiO2suspension was
aged at 25◦C in an N2atmosphere for 2 h.
Solutions were prepared using reagent-grade chemicals and Millipore-Q water, following standard methods [33]. The initial concentration of Al(III) (Al(NO3)3·9H2O) was
0–1× 10−2 M; the SiO2 concentration was maintained at
1 g/l, and the background electrolyte concentration was ad-justed to 0.1 M by KNO3 electrolyte solution. All
experi-ments were performed at 25◦C, after adjustment to the de-sired pH using KOH and HNO3solutions, with shaking at
200 rpm. Since preliminary kinetic experiments suggested that pseudo-equilibrium was reached within 24 h, 24 h was selected as the reaction time for the equilibrium experi-ments [34].
To elucidate these processes in the sorption continuum, samples were prepared in which SiO2(s)powder was
equi-librated with Al(III) aqueous solutions with concentrations from undersaturated to oversaturated, with respect to the bulk phase precipitation in the absence of a solid sorbent. Two parameters, pH and total Al(III) concentration, were studied in batch experiments to determine sorption pH edges and isotherms, respectively. After the reaction, a subsample of the suspension was taken for electrophoresis measure-ments; the rest of the suspension was centrifuged (Kubota 6800) at 10,000 rpm for 15 min and the supernatant was passed through a 0.2-µm membrane filter to analyze the sol-uble Al and the Si dissolved out of bulk SiO2(s)solid by an
ICP-AES (Perkin–Elmer, Optima 2000DV).
2.2. Electrophoresis measurements
Electrophoretic mobility (EM) was measured at various pH and Al(III) concentrations to detect changes in net to-tal particle surface charge density associated with the sorp-tion of Al(III). The electrophoretic mobility of the particles in SiO2suspensions with various amounts of Al(III) at pH
values between 1.5 and 12.0 was measured using a laser Doppler electrophoretic light-scattering apparatus (Malvern Instrument, Zetasizer 2000). Triplicate measurements were made in crossed-beam mode with a 30-s count time, an applied voltage of 150 V, and a modulator frequency of 1000 Hz.
2.3. Modeling approaches
Three categories of surface reaction were considered— surface complexation, surface precipitation, and bulk
precip-Table 1
Category of surface reactions
Category Applied model Forming species Abbreviation
Surface complexation SCM Adsorption of hexaqua-Al3+ion (Al(H2O)36+) onto SiO2(s) AA
Adsorption of 1st hydrolyzed hexaqua-Al3+ion (Al(OH)(H2O)25+) onto SiO2(s) AH
Surface precipitation SPM Coprecipitation of Al(III) hydroxide phase as a solid solution with the original SiO2(s)
surface phase
PS
Bulk precipitation Precipitation of Al(OH)3(s)in bulk solution PB
Table 2
Fixed parameters for modeling with SCM and SPM
Site concentration of SiO2(s)(mmol/g) 1.67a
Surface area of SiO2(s)(m2/g) 200b
log Kaint1 −1.5 a log Kaint2 −5.5 a log Ksp_Al 10.38c log Ksp_Si −2.74c
aMeng and Letterman [13,14]. b Measured in this study.
cDatabase of MINTEQA2 program.
itation—to describe the sorption continuum. In each reaction category, Al(III) forms several surface species, depending on the system conditions (Table 1).
2.3.1. Surface complexation model (SCM)
The geochemical speciation code MINTEQA2, Ver-sion 3.11 [35], was used to simulate the equilibrium partition of Al(III) in the SiO2(s)/aqueous solution. The SiO2(s)
sur-face parameters were adopted from the work of Meng and Letterman [13,14] and are presented in Table 2. Both the diffuse layer model (DLM) and the database included in the MINTEQA2 were used for surface complexation modeling. The DLM was a type of simple two-layer model, including one planar surface layer and one diffuse layer of counte-rions [36]. The exponential coulombic term derived from Gouy and Chapman’s electrical double layer theory (EDL) was introduced to correct the surface-reaction mass-law con-stants for surface charge effects [32]. Table 3 summarizes the surface reactions considered in the SCM (Eqs. (1)–(4)).
2.3.2. Surface precipitation model (SPM)
Farley et al. [37] extended the DLM to include surface precipitation by taking into account the formation of a new surface phase, which is described as an ideal solid solu-tion [38,39] of Me(OH)2(s)and Fe(OH)3(s). This approach
yields a continuum between the adsorption of a solute on a surface and the precipitation of this solute as a new bulk solid phase. In this study, the model was modified and applied to the Al(III)/SiO2(s)system.
The success of the surface precipitation model follows from the fact that it provides a mathematical approach to increasing the activity of the coprecipitating solid metal hy-droxide phase continuously from near zero to unit activity as the amount of surface precipitate that forms increases. This continuous increase is accomplished by assuming that an ideal solid solution is generated by
(5) {Al(OH)3(s)} = [Al(OH)3(s)] Ts and (6) {SiO2(s)} =[SiO2(s)] Ts ,
where { } and [ ], respectively, represent the activity and molal concentration of the corresponding species, and Ts represents the total mass of solid material in solid solution and is given by
(7) Ts= [Al(OH)3(s)] + [SiO2(s)].
Additionally, an ideal solid solution is assumed, so
(8)
{Al(OH)3(s)} + {SiO2(s)} = 1.
This representation allows the solubility of the aluminum hydroxide phase to decline at lower surface coverages. For
Table 3
Surface reactions considered in SCM
Reaction Equilibrium constant expression
Surface ionization
SOH+ H+ SOH+2 Kaint (1)
1 = exp(F ϕ/RT )[SOH + 2]/[SOH][H+] SOH SO−+ H+ Kaint (2) 2 = exp(−F ϕ/RT )[SO −][H+]/[SOH] Surface complexation
SOH+ Al3+ SOAl2++ H+ KAAint = exp(2F ϕ/RT )[SOAl2+][H+]/[SOH][Al3+] (3) SOH+ Al3++ H2O SOAlOH++ 2H+ KAHint = exp(F ϕ/RT )[SOAlOH+][H+]2/[SOH][Al3+] (4) F : Faraday constant; φ: the potential at the SiO2(s)surface; R: gas constant; T : absolute temperature.
example, the aluminum hydroxide phase forms when the Al(III) ion activity product (IAP) is exceeded, according to
(9) IAP= {Al3+}{OH−}3= {Al(OH)3(s)}Ksp_Al,
where Ksp_Al is the thermodynamic solubility product
con-stant of Al(OH)3(s). According to Eq. (5), as{Al(OH)3(s)} decreases, the value of the IAP required for the onset of precipitation decreases. Hence, surface precipitation can oc-cur at Al(III) ion concentrations and pH values below those required for precipitation of the pure solid phase from aque-ous solution. A solid-phase activity continuaque-ously increasing from almost zero to unity has been found to be necessary to enable the solid solution model to describe Al(III) ion sorp-tion as a funcsorp-tion of the changing surface coverage.
In this model, Al(III) at the SiO2(s)/aqueous solution
in-terface is treated as a surface species, while Al(III) is not in direct contact with the solution phase; that is, it is buried un-der the SiO2(s)/aqueous solution interface and is treated as a
solid species (denoted by subscript (s)) that forms a solid so-lution. Surface precipitation reactions involved in the uptake of Al(III) by SiO2(s)can then be written as follows:
Adsorption of Al(III) onto SiO2(s) =
=Si–OH0+ Al3++ 3H 2O
(10)
SiO2(s)+ ≡Al–OH+2 + 2H+, KPSint.
Precipitation of Al(III) ≡Al–OH+2 + Al 3++ 3H 2O (11) Al(OH)3(s)+ ≡Al–OH+2 + 3H+, 1/Ksp_Al.
Adsorption of Si(OH)4on Al(OH)3(s) ≡Al–OH+2 + Si(OH)4(aq)
Al(OH)3(s)+ ==Si–OH0+ H++ 2H2O, (12) 1/KPSintKsp_AlKsp_Si. Precipitation of Si(OH)4 = =Si–OH0+ Si(OH) 4(aq) (13) SiO2(s)+ ==Si–OH0+ 2H2O, 1/Ksp_Si.
The surface symbols == and ≡ are used to denote the bonds of the cations with the surface of the solid, and have dif-ferent meanings for Si(IV) and Al(III): ==Si–OH0 repre-sents [SiO2]n because SiO2 is a tetrahedron, with four O
atoms bonded to each Si atom, and ≡Al–OH0 represents
[Al(OH)3]n because Al(OH)3 is an octahedron, with three
O atoms bonded to each Al atom. Thus reactions (10) and (12) are balanced with respect to H and O [37,40]. Although the surface species indicated in reactions (11) and (13) di-vide out of the corresponding mass law expressions, they are included to emphasize the link between adsorption and surface precipitation. The term KPSint represents the intrinsic equilibrium constant for Al(III) precipitation, and Ksp_Siis
the thermodynamic solubility product constant of SiO2(s).
Importantly, only three of the four equations (10)–(13) are mathematically independent expressions. Table 2 presents the parameters used to model surface precipitation.
Fig. 1. Percentage of Al sorption vs pH and DLM simulation results of 1× 10−4M Al(III)/SiO2(s) system. Three predictions according to
ad-sorption of Al3+ion (AA), adsorption of AlOH2+ion (AH), and surface precipitation (PS) reactions are given for comparison. Fitting parameters are as follows: log KAAint = 0.5, log KAHint = −3, and log KPSint= −4.8.
3. Results and discussion
3.1. Al(III) sorption and simulation of SCM/SPM
The effect of the change in the concentration of the background electrolyte (KNO3) on the Al(III) sorption onto
SiO2(s)was examined before the pH-edge and isotherm
ex-periments to clarify the strength of the bonding between Al(III) and the SiO2(s) surface. The experimental results
revealed that ionic strength did not affect the sorption of Al(III) onto SiO2(s)(figure not shown). This finding suggests
that Al(III) tends to form an inner-sphere complex, which bonds strongly to the o-plane [41]. From previous studies, the structure of strongly sorbed surface species is generally described as monodentate or bidentate inner-sphere com-plexes at low and moderate surface coverage, and as mult-inuclear species and surface precipitation at high surface coverage [17–19,42].
Fig. 1 shows the model predictions of 1× 10−4 M Al(III) reactions in SiO2(s) suspensions at various pH for
the AA, AH, and PS configurations (Table 1) considered in SCM/SPM. The experimental results display that Al(III) sorption onto SiO2(s)increases with the pH value. The AH
and PS configurations excellently fit the experimental data and both appear as one line, while the AA type does not sat-isfactorily describe Al(III) behavior in this region of the en-velope curve. The results of the PB simulation were ignored because homogeneous Al hydroxides cannot precipitate in such a low initial concentration of Al(III), according to the solubility product constant of Al(OH)3(s). As the modeling results suggest, the adsorption mechanism dominates the up-take of 1× 10−4 M Al(III). Crawford et al. [43] stated that the dielectric constant of the solid substrate plays a major role in determining the pattern of the complex species
dur-Fig. 2. Percentage of Al sorption vs pH and DLM simulation results of 1× 10−3M Al(III)/SiO2(s)system. Four predictions according to
adsorp-tion of Al3+ion (AA), adsorption of AlOH2+ion (AH), surface precipita-tion (PS), and adsorpprecipita-tion of AlOH2+ion incorporated with bulk precipita-tion (AH+ PB) reactions are given for comparison. Fitting parameters are as follows: log KAAint = 0.5, log KAHint= −4, log KPSint= −5.8.
ing metal ion adsorption. The dominant sorbing species is the free metal ions when a solid substrate with a dielectric constant () approaching that of water (= 78.3) is used. However, the hydrolysis product of metal ions is the dom-inant adsorbing species when a solid substrate with a low dielectric constant is used. Accordingly, AlOH2+is the dom-inant adsorbing species in this system since the dielectric constant of SiO2(= 6) is much less than that of water.
Fig. 2 presents the model-predicted and experimental data for 1× 10−3 M Al(III) reactions between SiO2(s)and
solu-tion as a funcsolu-tion of pH. As illustrated, attempts to describe the pH-edge data using the adsorption of Al3+ions (AA type in SCM) were unsuccessful above pH 3. This results could be attributed to the fact that the hydrolysis of free Al3+ions occurs above pH 3, and the hydrolyzed species (AlOH2+, Al2(OH)42+, and Al3(OH)54+) begin to dominate [44].
Sim-ulation that involves the AH type successfully describes the interfacial behavior of Al(III) below pH 4, but significantly underestimates the experimental data above pH 4, where high surface loading occurs. Other combinations of reactions were considered to improve the fitness of predictions and ex-plore the events that occurred at high pH. The predictions based on AH configuration incorporated with the reaction of homogeneous precipitation of Al(OH)3(amporphos) (in the
database of the MINTEQA2 Program, v. 3.1) in bulk solu-tion (PB) were denoted as crosses in Fig. 2. The AH+ PB modeling curve has a discontinuous point at pH 5 and then rises gradually to meet the experimental data as pH in-creases. Consequently, the increase in the removal of Al(III) above pH 5 can be attributed primarily to Al hydroxide de-position in bulk solution. Two possible reasons account for the underestimation of SCM from pH 4 to 5: (a) saturation of SiO2(s)surface sites with increasing Al(III) sorption, and
Fig. 3. Modeling Al(III) sorption isotherm using diffusion-layer model. Four predictions according to adsorption of Al3+ ion (AA), adsorp-tion of AlOH2+ ion (AH), surface precipitation (PS), and adsorption of AlOH2+ion incorporated with bulk precipitation (AH+ PB) reactions are given for comparison. Fitting parameters are as follows: log KAAint = 0.5, log KintAH= −4, log KPSint= −5.8.
(b) failure to account for Al(III) polymerization and sur-face precipitation [31]. When the number of sursur-face sites of SiO2is taken to be 8.35 µmol/m2(calculated by data in
Table 2), the 1× 10−3 M Al(III) sorbed by SiO2
suspen-sion in the range pH 4–5 occupies just 14% to 48% of the sites. This phenomenon implies that the postulate (a), insuf-ficiency of surface sites for Al(III) removal, does not apply to this system because of the surface geometry [24]. With reference to postulate (b), Fig. 2 indicates that the surface precipitation model with the solid solution assumption rea-sonably fits the data over the whole range of experimental pH. Therefore, surface precipitation occurs at pH 4, which is about one pH unit lower than for homogeneous Al hydroxide precipitation. Over the pH range from 4 to 5, surface precip-itation is the dominant reaction in this system. A number of researchers have demonstrated that the SCMs are satis-factory only for low surface loading, while SPM can provide reasonable fits to the data over a wide range of coverage con-ditions, from low to high [18,19,30,31,37,40,45]. Increasing the initial concentration of Al(III) by one order of magni-tude (1× 10−4–1× 10−3 M) results in a great variety of governing surface reactions; therefore, both the pH and the sorbate/sorbent ratio importantly affect sorbate reactions at the solid/solution interface.
The effect of sorbate-to-sorbent ratio on surface reac-tion is depicted using sorpreac-tion isotherms at pH 4.80± 0.03 (Fig. 3), where total Al(III) concentration (log[Al]T) is plot-ted against sorption density (log ΓAl, where ΓAl is mole
Al(III) uptake per gram of SiO2 solid). This isotherm
ex-perimental data plotted in double logarithmic format follow a linear trend with unit slope over the whole range of to-tal Al(III) concentrations. Karthikeyan et al. [30,31] tested
the sorption of copper onto hydrous oxides of iron and alu-minum and found out that the copper isotherms for both ox-ides have an initial region with unit slope at low total Cu and a Freundlich region with a slope of less than 1:1 at moderate Cu concentrations. A saturation region also exists, in which Cu removal increases sharply with total Cu, since all the sor-bent reactive sites are fully occupied and Cu is removed by precipitation reactions. The linear and Freundlich regions in the isotherm can account for the two types of reactive sites high and low binding strength, for Cu uptake by iron or aluminum oxide surfaces. However, the isotherm data for Al(III) in this study all follow a linear trend with unit slope over the entire range of total Al(III) concentrations, even in the zone where homogeneous Al hydroxide precipitate may be formed. This result implies that there is only one reac-tive site with the same binding strength for Al(III) uptake by SiO2(s)surface, and at high Al(III) concentrations, some
reactions lead to a transition between adsorption and bulk precipitation. Attempts to describe these isotherm data using SCM, illustrating adsorption of various Al species (AA and AH), SPM, assuming a solid solution formed on the surface (PS), and homogeneous precipitate of Al hydroxide in bulk solution (PB) are shown in Fig. 3. As illustrated in Fig. 3, all models yield suitable fits at low surface coverage, but only SPM reasonably fits the data at high surface coverage. Therefore, the onset and domains of adsorption, hydrolysis, surface precipitation, and bulk precipitation are delineated from the divergence between the model predictions and the isotherm data. Assuming adsorption of free Al3+ions (AA) and AlOH2+(AH) resulted in the model’s fitting satisfacto-rily up to a surface coverage of 10−3.25mol Al(III)/g SiO2(s).
This value of coverage also indicates the onset of surface precipitation. To clarify the surface and bulk precipitation, a homogeneous precipitation reaction of Al(OH)3(s) incor-porated with the adsorption of AlOH2+ (AH+ PB) is also considered. In Fig. 3, the predictions using AH+ PB include a discontinuous point at surface coverage of 10−2.75 mol Al(III)/g SiO2(s)and underestimate the range of surface
cov-erage between 10−3.25 and 10−2.25 mol Al(III)/g SiO2(s),
which can be regarded as the onset of the bulk precipitation of Al(OH)3(s)and the surface precipitation domain,
respec-tively.
3.2. Evidences from electrophoretic mobility (EM) measurements
In the authors’ earlier study, the EM of pure SiO2
sus-pensions over a range of pH values and in three different salt solutions (0, 0.1, 0.5 M) lie on a curve and the isoelec-tric point (IEP) is 2.00 for an SiO2particle [34]. Therefore,
0.1 M KNO3 was adopted as the background electrolyte
in all experiments. Fig. 4 shows the EM data of SiO2
sus-pension with various Al(III) concentrations as a function of pH. The electrophoretic behavior of SiO2 with 1× 10−5
and 1× 10−4 M of Al(III) is similar to that of pure SiO2
suspension. Meng and Letterman [13] also measured the
Fig. 4. The EM of Al(III)/SiO2suspensions as a function of pH in 0.1 M
KNO3electrolyte in the presence or absence of various concentrations of
Al(III).
zeta potential of Al(III)/SiO2(s) suspensions, but the tested
Al(III) concentration was too high to observe electrokinetic behavior similar to that in pure SiO2suspension. James and
Healy [20] examined the electrophoretic behavior of La(III) in a SiO2(s)suspension system and obtained similar results,
with no significant shift in electrophoretic mobility, for low specific adsorption amounts of La(III).
Three charge reversal (CR) points (strictly, the pH values at which the electrokinetic potential reverses), CR1 (charge reversal from+ to −), CR2 (charge reversal from − to +), and CR3 (charge reversal from+ to −), correspond to in-creasing pH for the curves of SiO2with 1× 10−3, 5× 10−3,
and 1× 10−2 M of Al(III). The top right corner of Fig. 4 schematically depicts the general features of these electroki-netic curves, including charge reversals. Several recent spec-troscopic studies [17,18,46,47] demonstrated that the value of 10% surface coverage is a criterion for the structure of surface complexes. The inner-sphere monomeric reactions can describe the sorption phenomenon in the low- and mod-erate coverage range (up to 10%). As coverage increases beyond 10% (high-coverage region), spectroscopic evidence is consistent with the formation of multinuclear species and surface precipitation. This study utilizes two coverage clas-sification schemes, based on the percentage of occupied surface active sites and the area percentage covered by hy-drated Al3+ions with an assumed hydrated radius (H.R.) of 3 Å, to elucidate the surface electrophoretic behavior as a function of surface coverage. Table 4 presents the surface coverage values calculated at the 50% Al(III) uptake point on the pH edge for the specified conditions. Each pH edge yielded a range of surface coverages, extending from 0% to twice the value shown in Table 4. For the 1× 10−5 and 1× 10−4M Al(III)/SiO2(s)systems with electrophoretic
be-havior similar to that of pure SiO2(s) suspension, both the
coverage of the area and the site density are lower than 10%, within the range in which inner-sphere monomeric reactions occur [18]. Both the area and the site
concentra-Table 4
Surface coverage for Al(III) sorption pH edges in SiO2(s)suspension system
Total Al(III) conc. Coverage of 50% Coverage for area, Coverage for site concentration (mol/l) uptake (µmol/m2) H.R. of Al3+= 3 Å (%) 1.67 mmol/g SiO2(s)(%)
1× 10−5 0.025 0.42 0.30
1× 10−4 0.25 4.24 2.99
1× 10−3 2.5 42.41 29.94
5× 10−3 12.5 212.06 149.70
1× 10−2 25 424.12 299.40
tion coverage of the 1× 10−3, 5× 10−3, and 1× 10−2 M Al(III)/SiO2(S)system are beyond 10%, the electrophoretic
behavior in these systems clearly differ from that of the 1× 10−5and 1× 10−4M Al(III)/SiO2(s)systems. Thus, the
electrophoretic behavior reflects the mechanism of Al(III) sorption onto the surface of SiO2(s).
These experiments do not reveal any significant shift in CR1 with the concentration of Al(III). The significance of CR1 is clear in that H+and OH−ions are potential determi-nants for the original oxide; thus, CR1 is the point of zero charge (PZC) of the SiO2suspension.
Abundant evidence now supports the claim that at high pH charge reversal, CR3 reflects the metal hydroxide coat-ing on the colloid substrate [20]. If sufficient metal ions are adsorbed at a pH suitable to form a complete coating of ad-sorbed metal hydroxide, then CR3 is the PZC of the metal hydroxide. Incomplete coating, due to a low concentration of metal or to a high concentration of colloidal substrate parti-cles, reflects both coated and uncoated surfaces being mixed together. Thus, CR3 will occur at pH values at or below the pHpzcof the Al(OH)3, depending on the coverage achieved.
According to Fig. 4, CR3 is made to approach the pHpzc
of amorphous Al(OH)3(9.0) [48] by increasing the Al(III)
concentration. The CR3 of the colloid in the 1× 10−2 M of Al(III)/SiO2suspension is the same as the pHpzcof pure
Al(OH)3, suggesting that Al(OH)3completely covered the
SiO2particles.
In Fig. 4, the CR2 of the curves for SiO2with 1× 10−3,
5× 10−3, and 1× 10−2 M of Al(III) ranges from pH 2 to 4 and increases as the concentration of Al(III) decreases. James and Healy [20] suggested that CR2 might indicate surface precipitation induced at a pH below that of bulk precipitation. They also used the relationship between stan-dard free energy and the dielectric constants of the interfacial medium to derive the conclusion that the electric field of the surface-induced precipitation at the interface is established before bulk precipitation occurs, even though the solution is unsaturated with respect to this precipitated solid. Addi-tionally, CR2 shifts to a lower pH as Al(III) concentration increases. Pugh and Bergström [49] also successfully inter-preted the uptake of Mg(II) on α-ultrafine silicon carbide and α-alumina by applying James and Healy’s model [20], and found that strong adsorption of Mg(II) occurred well below the bulk precipitation threshold. With respect to the 1× 10−3 M Al(III)/SiO2(s) system, the CR2 in Fig. 4 is
coincident with the point of the onset of surface
precip-Fig. 5. Si dissolution of SiO2(s)suspension with or without Al(III) over a
pH range 2–9.
itation (∼pH 4) in Fig. 2, which was determined by the divergence between the SCM/SPM predictions and the ex-perimental data. Accordingly, these electrokinetic findings confirm that the SPM, based on the solid solution assump-tion, can successfully describe the transition from adsorption to bulk precipitation and indicate the onset of surface precip-itation. Furthermore, the occurrence of CR2 near the IEP of SiO2(s), almost obscuring it, could be regarded as
represent-ing the strong hydrolysis of Al(III) [20]. CR2 approaches CR1 in a strongly hydrolytic cation system but approaches CR3 in a weakly hydrolytic cation system.
3.3. Effect of Al(III) on SiO2(s)dissolution
Fig. 5 shows the solubility characteristics of 1 g SiO2(s)/l
suspension with or without Al(III) over the pH range from 2 to 9. The results imply that the release of silicate (rep-resented as Si mg/l) from SiO2(s)solid increases with pH,
perhaps because the ionization of monosilicic acid enhances the dissolution of H4SiO4from SiO2(s), as illustrated below:
(14) SiO2(s)+ 2H2O↔ Si(OH)4(aq),
(15) Si(OH)4(aq)+ OH−↔ Si(OH)3O−(aq)+ H2O.
Fig. 5 also indicates that the solubility of SiO2(s)decreases
as the Al concentration increases. In a system with low Al(III) concentration, such as 1× 10−4 and 1× 10−5 M,
the concentration of the dissolved Si is slightly lower than that in a pure SiO2(s)system. Stumm and Wollast [50]
re-ported that a surface-coordinated metal ion, such as Cu(II) or Al(III), can block an oxide surface group, thus suppressing dissolution. Aqueous Al(III) forms an aluminosilicate com-plex on the reactive silica surface site [1, Eq. (16)]; and this complex more strongly suppresses dissolution of the reactive silica surface in an alkaline medium than that in an acidic medium, as compared the dissolution in pure SiO2system.
(16) –Si| |(OH) reactive silica surface site + Al(OH)+2 ↔ – | Si | OAl(OH) aluminosilicate complex + H+.
In the system with a high Al concentration of 1×10−3M, the concentration of dissolved Si is similar to that in the other systems below pH 4, but much lower than that in the pure SiO2(s)system above pH 4. The concentration of the Si in
this system at pH 8.6 approaches 0.83 mg/l, which is two or-ders of magnitude less than that of the pure SiO2(s)system.
Furthermore, the trend of the concentration of Si is quite dif-ferent from that in other systems; the concentration increases with pH for pH < 4 but slightly drops as pH increases for pH > 4. These results imply that, while the surface alumi-nosilicate complex inhibits the dissolution of SiO2(s), other
mechanisms may be responsible for the dramatic restraint of the dissolution. Stein [51] also observed the dissolution reduction of 3CaO·Al2O3solid by surface precipitation of
Al(OH)3, but this reactivity reduction had been prevented
by the presence of SiO2. This different trend of dissolution
in 1× 10−3M Al(III)/SiO2(s), can be explained by the
for-mation of an Al hydroxide surface precipitate, which leads to a passive film on the surface of SiO2(s)and inhibits the
corrosion of the SiO2(s)surface by OH−ions.
4. Conclusions
Al(III) sorbs onto SiO2(s) in different modes at
vari-ous pH and Al(III) concentrations, as an Al(H2O)36+
com-plex at low pH and Al(III) concentrations, hydrolyzed Al(OH)(H2O)26+ complex at medium surface loading, and
surface-induced precipitate at high sorption densities, con-firming earlier suggestions that sorption mode changes with surface loading [17–19,27]. The modeling results of the SCM/SPM prediction showed that the solution from which the surface precipitate formed was not saturated with re-spect to Al(OH)3(s). The SCM is an adequate model only
when Al(III) sorption occurs as complexes binding at surface coordination sites of SiO2(s), while inconsistency between
predicted behavior and data exists when solution conditions (alkaline pH, high sorbate/sorbent ratio) lead to the forma-tion of surface precipitate. The SPM is capable of describing sorption over a wide range of pH and sorbate/sorbent ra-tios because this model allows species not only to sorb at the surface sites but also to form solid solutions on the
SiO2(s) surface. Electrokinetic measurements suggest that
no significant shift in electrophoretic mobility occurs dur-ing the specific adsorption of Al(III) onto SiO2(s). However,
under these solution conditions (1× 10−3, 5× 10−3, and 1× 10−2 M Al(III)), which lead to surface precipitation, three charge-reversal points exist and separately represent the IEP of SiO2, surface precipitation, and the Al
hydrox-ide coating onto the SiO2 surface. In the 1× 10−3 M
Al(III)/SiO2(s)system, CR2 (near pH 4) is coincident with
the onset pH of surface precipitation, as determined by the SPM. The dramatically low concentration of dissolved Si above pH 4 in the 1× 10−3 M Al(III)/SiO2(s) system is
evidence of the formation of an Al hydroxide surface pre-cipitate, which leads to the formation of a passive film on the surface of SiO2(s), inhibiting the corrosion of the SiO2(s)
surface by OH−ions.
Acknowledgments
We thank the National Science Council of Taiwan, Re-public of China, NSC 90-2218-E-131-013 and 91-2211-E-131-002 for financial support.
References
[1] R.K. Iler, The Chemistry of Silica, Wiley, New York, 1979. [2] C.T. Driscoll, K.M. Postek, in: G. Sposito (Ed.), The Environmental
Chemistry of Aluminum, CRC Press, Boca Raton, FL, 1996, p. 363. [3] D.W. Sparling, T.P. Lowe, P.G.C. Campbell, in: R.A. Yokel, M.S.
Golub (Eds.), Research Issues in Aluminum Toxicity, Taylor & Fran-cis, Washington, DC, 1997, p. 47.
[4] G.F.V. Landeghem, M.E.D. Broe, P.C. D’hases, Clin. Biochem. 31 (1998) 385.
[5] P. Nayak, Environ. Res. Sect. A 89 (2002) 101.
[6] S. Bi, S. An, W. Tang, R. Xue, L. Wen, F. Liu, J. Inorg. Biochem. 87 (2001) 97.
[7] T.W. Swaddle, Coord. Chem. Rev. 219–221 (2001) 665.
[8] B.A. Dempsey, H. Sheu, T.M.T. Ahmed, J. Mentik, J. Am. Water Works Assoc. 77 (1985) 74.
[9] R.D. Letterman, S.R. Asolekar, Water Res. 24 (1990) 931.
[10] G. Sposito, The Environmental Chemistry of Aluminum, second ed., Lewis Publishers, Boca Raton, FL, 1996.
[11] P.R. Anderson, M.M. Benjamin, Environ. Sci. Technol. 24 (1990) 692. [12] P.R. Anderson, M.M. Benjamin, Environ. Sci. Technol. 24 (1990)
1586.
[13] X. Meng, R.D. Letterman, Environ. Sci. Technol. 27 (1993) 970. [14] X. Meng, R.D. Letterman, Environ. Sci. Technol. 27 (1993) 1924. [15] D.A. Kulik, Geochim. Cosmochim. Acta 64 (2000) 3161. [16] J.D. Reid, B. NcDuffie, Water Air Soil Pollut. 15 (1981) 375. [17] C.J. Chisholm-Brause, P.A. O’Day, G.E. Brown Jr., G.A. Parks,
Na-ture 348 (1990) 528.
[18] L.E. Katz, K.F. Hayes, J. Colloid Interface Sci. 170 (1995) 477. [19] L.E. Katz, K.F. Hayes, J. Colloid Interface Sci. 170 (1995) 491. [20] R.O. James, T.W. Healy, J. Colloid Interface Sci. 40 (1972) 42. [21] M. Rasmusson, S. Wall, Colloids Surf. A Physicochem. Eng.
As-pects 122 (1997) 169.
[22] Y. Xu, F.W. Schwartz, S.J. Traina, Environ. Sci. Technol. 28 (1994) 1472.
[24] W. Stumm, Chemistry of the Solid–Water Interface Processes at Mineral–Water and Particle–Water Interfaces in Natural Systems, Wiley–Interscience, New York, 1992.
[25] J.A. Davis, J.O. Leckie, J. Colloid Interface Sci. 67 (1978) 90. [26] W. Stumm, E. Wieland, in: W. Stumm (Ed.), Aquatic Chemical
Kinet-ics, Wiley–Interscience, New York, 1990, p. 367.
[27] N.T. Towle, J.R. Bargar, G.E. Brown Jr., G.A. Parks, J. Colloid Inter-face Sci. 187 (1997) 62.
[28] J. Laven, H.N. Stein, J. Colloid Interface Sci. 238 (2001) 8.
[29] K.G. Karthikeyan, H.A. Elliott, F.S. Cannon, Environ. Sci. Technol. 31 (1997) 2721.
[30] K.G. Karthikeyan, H.A. Elliott, J. Chorover, J. Colloid Interface Sci. 209 (1999) 72.
[31] K.G. Karthikeyan, H.A. Elliott, J. Chorover, J. Colloid Interface Sci. 220 (1999) 88.
[32] D.E. Yates, S. Levine, T.W. Healy, J. Chem. Soc. Faraday Trans. 70 (1974) 1807.
[33] L.S. Clesceri, A.E. Greenberg, A.D. Eaton, Standard Methods for the Examination of Water and Wastewater, APHA, AWWA, WEF, Wash-ington, DC, 1998.
[34] W.H. Kuan, S.L. Lo, M.K. Wang, Water Sci. Technol. 42 (2000) 441. [35] J.D. Allison, D.S. Brown, K.J. Novo-Gradac, MINTEQA2/
PRODEFA2, A Geochemical Assessment Model for Environmen-tal Systems: Version 3.11 Database and Version 3.0 User’s Manual, Environmental Research Laboratory, U.S. EPA, Athens, GA, 1991.
[36] D.A. Dzombak, F.M.M. Morel, Surface Complexation Modeling: Hydrous Ferric Oxide, Wiley, New York, 1990.
[37] K.J. Farley, D.A. Dzombak, F.M.M. Morel, J. Colloid Interface Sci. 106 (1985) 226.
[38] D.C. Thorstenson, L.N. Plummer, Am. J. Sci. 277 (1977) 1203. [39] P.D. Glynn, E.J. Reardon, Am. J. Sci. 290 (1990) 164.
[40] J.A. Meima, R.N.J. Comans, Environ. Sci. Technol. 32 (1998) 688. [41] K.F. Hayes, C. Papelis, J.O. Leckie, J. Colloid Interface Sci. 125
(1988) 717.
[42] M. Berka, I. Bányai, J. Colloid Interface Sci. 233 (2001) 131. [43] R.J. Crawford, I.H. Harding, D.E. Mainwaring, Langmuir 9 (1993)
3057.
[44] C.F. Baes, R.E. Mesmer, The Hydrolysis of Cations, Wiley, New York, 1976.
[45] D.A. Dzombak, F.M.M. Morel, J. Colloid Interface Sci. 112 (1986) 588.
[46] C.J. Chisholm-Brause, K.F. Hayes, A.L. Roe, G.E. Brown Jr., G.A. Parks, J.O. Leckie, Geochim. Cosmochim. Acta 54 (1990) 1987. [47] A.L. Roe, K.F. Hayes, C.J. Chisholm-Brause, G.E. Brown, G.A. Parks,
J.O. Leckie, Langmuir 7 (1991) 367. [48] G.A. Parks, Chem. Rev. 65 (1965) 177.
[49] R.J. Pugh, L. Bergström, J. Colloid Interface Sci. 124 (1988) 570. [50] W. Stumm, R. Wollast, Rev. Geophys. 28 (1990) 53.
[51] H.N. Stein, in: G.H. Stewart (Ed.), Science of Ceramics, Academic Press, London, 1967, p. 109.
