ARTICLE IN PRESS WAT E R R E S E A R C H
40 (2006) 3167 – 3176
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/watres
Removal of Mn2+ ions from drinking water by using Clinoptilolite and a Clinoptilolite–Fe oxide system Maria K. Doula National Agricultural Research Foundation of Greece, Soil Science Institute of Athens, 1 Sof. Venizelou St., 14123 Lycovrisi, Greece
art i cle info
A B S T R A C T
Clinoptilolite, a natural zeolite, was used for the synthesis of a high surface area
Received 11 February 2006
Clinoptilolite–Iron oxide system, in order to be used for the removal of Mn2+ ions from
Received in revised form
drinking water samples. The new system was obtained by adding natural clinoptilolite in
1 July 2006
an iron nitrate solution under strongly basic conditions. The Clin–Fe system has specific
Accepted 7 July 2006
surface area equal to 151.0 m2/g and is fully iron exchanged (Fe/Al ¼ 1.23). Batch adsorption
Available online 30 August 2006
experiments were carried out to determine the effectiveness of the Clin and the Clin–Fe
system in removal of manganese from drinking water. Adsorption experiments were
conducted by mixing 1.00 g of each of the substrates with certain volume of water samples
contaminated with 10 different Mn concentrations (from 3.64 106 to 1.82 102 M or
from 0.2 to 1000 ppm). For the present experimental conditions, the Mn adsorption capacity
of Clin was 7.69 mg/g, whereas, of Clin–Fe system was 27.12 mg/g. The main factors that
contribute to difference adsorption capacity of the two solids are due to new surface
species and negative charge of Clin–Fe system. In addition, the release of counterbalanced ions (i.e., Ca2+, Mg2+, Na+ and K+) was examined as well as the dissolution of framework Si and Al. It was found that for the most of the samples the Clin–Fe system releases lower concentrations of Ca, Mg and Na and higher concentrations of K than Clin, while the dissolution of Si/Al was limited. Changes in the composition of water samples as well as in their pH and conductivities values were reported and explained. & 2006 Elsevier Ltd. All rights reserved.
Manganese exists in water as a groundwater mineral but may also be present due to underground pollution sources. Manganese may become noticeable in tap water at concentrations higher than 0.05 mg/l of water by imparting a color, odor, or taste to the water. According to Division of Environmental Epidemiology and Occupational Health of Connecticut (2001), health effects from manganese are not a concern until concentrations are approximately ten times higher. The levels of Mn in groundwater from natural leaching processes can vary widely, depending upon the types of rock and minerals Tel./fax: +3 210 5053455.
E-mail address: [email protected]
. 0043-1354/$ - see front matter & 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2006.07.013
present at the water table. Typically, Mn concentrations from natural processes are low but can range up to 1.50 mg/l or higher. Sources of pollution, rich in organic matter (e.g., runoff from landfills, composts, brush or silage piles, or chemicals such as gasoline), can add to the background level by increasing Mn release from soil or bedrock into groundwater. Several treatment technologies, such as chemical precipitation, ultra filtration, adsorption and ion-exchange, reverse osmosis, electrodialysis, have been developed for eliminating manganese and heavy metals in general from solution phase (Erdem et al., 2004). A number of adsorbent materials have
ARTICLE IN PRESS 3168
WAT E R R E S E A R C H
40 (2006) 3167– 3176
C0 C CSi
is the initial Mn concentration (mol/l or ppm) is the solution Mn concentration (mol/l)
CAl X QP QA
is the solution concentration of dissolved Al (mol/l) is the adsorbed Mn concentration (mol/kg) is the maximum processing capacity (l/kg) is the practical specific capacity (eq/kg)
is the solution concentration of dissolved Si (mol/l)
been studied for their capacity to remove heavy metals, including activated carbon, activated alumina, ion-exchange resins, chushed coals, zeolites, clay minerals and other aluminosilicates. Some of these materials, such as ionexchange resins, are totally effective but expensive, and others such as coal and straw are inexpensive but ineffective. Clinoptilolite (natural zeolite), vermiculite, peat moss, slow sand filters and other natural materials have been found to have high heavy metals adsorption capacity (Doula et al., 2002; Inglezakis et al., 2002; Bosco et al., 2005). Especially for manganese removal it was found that, for instance, Namontmorillonite has an adsorption capacity equal to 3.22 mg/ g (Abollino et al., 2003); dolomite equal to 2.21 mg/g; marble equal to 1.20 mg/g; quartz equal to 0.06 mg/g (Kroik et al., 1999); clinoptilolite from Turkey equal to 4.22 mg/g (Erdem et al., 2004); granular activated carbon equal to 2.54 mg/g (Jusoh et al., 2005). Although these adsorption capacity values correspond to different experimental conditions and there is no experimental relation between them, they are representative of the solids’ tendency to retain Mn2+ ions. Zeolites are microporous crystalline hydrated aluminosilicates that can be considered as inorganic polymers built from an infinitely extending three-dimensional network (similar to honeycomb) of tetrahedral TO4 units, where T is Si or Al, which form interconnected tunnels and cages. Each aluminum ion that is present in the zeolite framework yields a net negative charge, which is balanced by an extra framework cation, usually from the group IA or IIA. Water moves freely in and out of zeolite pores but the framework remains rigid. The porous zeolite is host to water molecules and ions of positive charge and the ability to exchange cations is one important property of them (Ruthven, 2001; Doula and Ioannou, 2003). In the past many researchers have tried to synthesize mixed systems of clay minerals or zeolites with iron oxides and proved that these systems are not only capable to adsorb high concentrations of inorganic species (Dimirkou et al., 1996, 2002; Ioannou et al., 1996) but also in the case of Fe–zeolite’s systems, to improve zeolites catalytic characteristics (Morturano et al., 2001). It is worthy to note that these mixed Fe–zeolites systems have physicochemical properties which can be applied to a lot of research fields. In the following, we report the synthesis of a high surface area clinoptilolite–iron solid system suitable for metal adsorption applications. We also compare its adsorption ability with the respective of the untreated clinoptilolite by using drinking water samples contaminated with Mn2+ ions. Although clinoptilolite is an adsorbent with high heavy metals adsorption capacity, authors aimed to modify this substrate in order to synthesize an adsorbent with even higher retention ability than that of the parent material. The formation of a Fe-oxide phase onto Clin increases the number
of active sites of the adsorbent because of the presence of the Fe–OH groups in external and internal surface positions of Clin and thus, it is expected for modified Clin to have significantly better adsorption behavior than its untreated form.
Materials and methods
The Clinoptilolite used in the present investigation comes from a layer situated in Thrace (North Greece). This material has been used for metal adsorption experiments (Doula et al., 2002; Doula, 2003) and consequently its physicochemical properties are well known. According to its chemical composition it is a Ca-rich clinoptilolite with almost no Fe content (Na0.2K0.6Mg0.7Ca2.0Al6.2Si29.8O72 19.6H2O). Its characterization as clinoptilolite is further supported by the Si/Al ratio, which is equal to 4.8. The estimated cation exchange capacity (C.E.C.) of Clin, with respect to its formula, is 2.35 meq/g. The powder XRD study showed that zeolite, feldspars and total micas+clays are present through the tuff whereas the clinoptilolite used in the present study comes from a layer with very high clinoptilolite content (up to 90%). Its specific surface area is equal to 30.98 m2/g and the average pore ˚ (Doula et al., 2002). Before the synthesis diameter is o20.0 A of the new system as well as the adsorption experiments the zeolite was finely ground and sieved to o0.02 mm.
Clinoptilolite–Fe oxide system
The Clin–Fe system was synthesized by following the method of pure goethite preparation, as described by Schwertmann and Cornell (1991). The change made by the authors concerns the presence of clinoptilolite, which was added in the experimental flasks. The system was prepared by mixing 10.0 g of clinoptilolite, 100 ml of freshly prepared 1 M Fe(NO3)3 solution, and 180 ml of 5 M KOH solution in a 2 l polyethylene flask. The addition of KOH solution was rapid and with stirring. The suspension was diluted to 2 l with twice distilled water and was held in a closed polyethylene flask at 70 1C for 60 h. After the appropriate period the reaction vessel was removed from the oven, and the precipitate was centrifuged, washed (until free of NO 3 ions) and finally dried. According to the method of goethite preparation, during the 60 h period the suspension is converted from a red-brown suspension of ferrihydrite to a compact, yellow-brown precipitate of goethite. However, in the presence of Clin the precipitate maintains its initial red-brown color. The specific surface area of the Clin–Fe system is 151.0 m2/g and the
ARTICLE IN PRESS WAT E R R E S E A R C H
˚ . Spectroscopic methods average pore diameter is p20.0 A (XRD, EPR, FTIR TG/DSC) were used for the characterization of the newly synthesized Clin–Fe oxide system and discussed in detail elsewhere (Doula, 2003).
The water sample used comes from the Athens drinking water network and its physicochemical analysis is shown in Table 1. The concentrations of Na and K were measured by using a Korning 410 flame photometer, whereas Si and Al were measured by using a Varian Liberty 220 ICP emission spectrometer. Ca and Mg were volumetrically determined and from these two concentrations the value of sample hardness (in ppm CaCO3) was estimated.
solution conductivities were measured. Mn concentrations were measured in the liquid phase by using a SpectrAA 300 Varian flame atomic spectrometer. The concentrations of Mn species adsorbed were calculated from the difference in measured and initial concentrations of the species in solution. The solution concentrations of Ca, Mg, Na, K, Si and Al were also measured. All experimental stages took place in triplicate, in a water bath at constant temperature (25 1C), and under a N2 atmosphere by placing the water bath in a glove box. The relative standard deviations (RSD) for concentration and pH measurements were 0.8% and 2.1%, respectively. The maximum processing capacity (QP) expresses the solution volume that can be purified from an ion with initial concentration C, by using a certain mass of sorbent or exchanger (Leinonen, 1999): QA ð1=kgÞ, QP ¼ ½C0
Stock manganese solutions
Ten stock Manganese solutions from 4 to 20 000 ppm were prepared by dissolving Mn(NO3)2 4H2O in twice distilled water. Their pH values ranged from 7.00 to 4.06 whereas their conductivities from 18.6 mS/cm to 53.3 mS/cm.
4 0 (200 6) 316 7 – 317 6
where QA is the practical specific capacity of the sorbent (equal to experimentally determined adsorbed Mn concentration) in eq/kg and C0 the initial Mn concentration in eq/l.
Mn2+ adsorption experiment
The adsorption experiment was divided into two stages. Firstly, several samples of 1.00 g of each solid were equilibrated with 95.0 ml of water sample (referred to as ‘‘equilibrium period’’). This period lasted for 48 h and after this period the sample pH values were measured. Chemical analysis was carried out for some of these equilibrated water samples in order to verify the changes caused by the contact with the adsorbent as well as to verify the exact solution composition before the addition of the Mn2+ solutions. During the second experimental stage (referred as ‘‘Mn adsorption’’) 5.00 ml of Mn stock solutions were added to the above mentioned samples, at concentrations such that the total mixture contained 0, 3.64 106, 9.11 106, 1.82 105, 7.28 105, 1.82 104, 9.11 104, 1.82 103, 3.64 103, 9.11 103, 1.82 102 mol/l Mn (or 0, 0.20, 0.50, 1.00, 4.00, 10.0, 50.0, 100, 200, 500, 1000 ppm Mn). The samples at the second stage were equilibrated for 48 h, centrifuged for 10 min at 15 000 rpm and then the solution pH values as well as the
Results and discussion
Characterization of the Clin–Fe system
The Clin–Fe system has a dark red color and its elemental analysis reveals that it contains 14.1% Fe as amorphous iron species. Its Si/Al ratio is almost equal to the respective ratio of the parent material and its Fe/Al ratio is equal to 1.23, meaning that the new material is fully iron-exchanged as well as contains an additional fraction of Fe ions that do not act as charge balancing species. This high ion-exchange level is explained by the fact that one Fe3+ ion must compensate three spatially separated negative charges of the zeolite matrix. Because only a small amount of Fe3+ ions could be deposited at cationic sites, the additional Fe3+ ions most likely form iron oxo- or hydroxo-cations that undergo complex chemical transformation during subsequent washing. The Feoverexchanged clinoptilolite (Clin–Fe system) can hold outside of its framework one part of its Fe atoms in exchanged sites and the rest as neutral species. Amorphous extra
Table 1 – Chemical analysis results of water samples before and after their contact with Clin and Clin–Fe system
Waterc Clin Clin–Fe
7.80 7.53 8.14
275 293 395
7.73 104 8.40 104 1.48 104
2.70 104 2.78 104 9.09 105
E0 4.00 105 2.86 103
1.80 104 5.00 104 2.51 104
2.97 105 3.36 104 8.16 105
1.05 105 1.10 105 2.41 106
104 112 23.9
The concentrations of Ca, Mg, Na, K, Si, and Al are in mol/l. a Conductivity in mS/cm. b In ppm CaCO3. c Water samples before treatment.
ARTICLE IN PRESS 3170
WAT E R R E S E A R C H
40 (2006) 3167– 3176
framework metallic cations can be deposited on both internal and external surface (Urquieta-Gonza`lez et al., 2002). The new material, due to the presence of the Fe-oxides located in the zeolite channels or on its external sites, is characterized by the presence of additional active sites (Fe–OH) which are influenced by the solution pH and are potential adsorption sites (Cornell and Schwertmann, 1996). Many researchers have tried to identify the structure of Fe species in zeolites in order to analyze the activity of those species. However, due to the highly variable chemistry of iron in an environment of oxygen-containing ligands and the simultaneous presence of various types of Fe species, whose populations depend on the procedure of Fe–zeolite preparation and treatment, such relationships have not yet been unambiguously established. Generally, like the complex chemistry of iron in solutions, there are many possible structures of Fe species in zeolites. It is evident that, in most cases, single di- and trivalent Fe ions, oxo- and hydroxocomplexes, polymeric oxidic species and iron oxide species are present simultaneously (Chen et al., 1998; Ko¨gel et al., 1999).
During the equilibrium period the chemical composition of the water samples was changed due to their contact with Clin and Clin–Fe system, as reported in Table 1. The two substrates changed the composition of the water samples; however, the characteristics of these changes were different. The solution concentrations of Ca2+, Mg2+, Na+, K+ and framework Si/Al were slightly increased when Clin was used as a substrate. Especially for Na+ ions, the increase in their solution concentration is mainly due to the dissolution of zeolites impurities and also due to ion exchange process. The limited release of Ca, Mg and K was assumed to occur through ion-exchange reactions mainly between these surface cations and water molecules as well as with Na+ ions from impurities dissolution. Generally, this process is a dynamic one and occurs when aluminosilicates materials, such as clinoptilolite, are embedded in aquatic solutions. The motion of Si and Al, which are framework cations, from the substrate toward solution is a solid dissolution process and can cause a defect to zeolite framework structure and contamination of water samples (Stumm, 1991; Oelkers and Schott, 1995). The new system not only retains its Ca and Mg but also adsorbs these ions from the solution so decreasing the total hardness of the samples. Na+ release is also limited as well as the dissolutions of Si and Al. The solution concentration of K+ ions is increased due to the dissolution of K+ ions deposited on Clin–Fe sites during the synthesis procedure, as will be better explained in the following. However, despite this process, the new system uses its available active sites to retain Ca and Mg from solution as well as water molecules. Moreover, the simultaneous adsorption of H+ ions from solution resulted in an increase in sample pH values. Because the K+ released concentrations are higher than the retained concentrations of the other cations a final increase in samples conductivity values were measured.
In order to study the Mn adsorption from the two substrates an extent range of Mn initial concentrations were examined. Manganese concentrations used are exceedingly higher than the water quality standards of 0.05 ppm for drinking water and 0.20 ppm for agricultural water (Green et al., 2003) but also higher than the value of almost 10.0 ppm which is characterized as high Mn concentration. Concentrations higher than 10.0 ppm Mn are more likely to be found in wastewater but the aim of the authors was to test the adsorption abilities and capacities of the two substrates under extreme conditions. After the equilibrium period, Mn was added to the mixtures at 10 different concentrations in aqueous solutions. Mn adsorption isotherms and the percentage of adsorbed Mn concentrations for both substrates are presented in Figs. 1 and 2. Clinoptilolite has a satisfactory adsorption behavior and is capable to adsorb Mn2+ species from the samples. The adsorption isotherm consists of two regions. In the first, the amount of adsorbed Mn increases gradually up to 0.14 mol/kg (7.69 mg/g). The second is a plateau region where Mn adsorption is almost constant. This specific concentration value is considered as the maximum Mn adsorption capacity of clinoptilolite under our experimental conditions. The maximum percentage adsorption of Mn reaches 65% while the lowest is almost 7%. The Clin–Fe system has a significantly higher adsorption capacity than its parent material (0.494 mol/kg or 27.12 mg/g). A two-region adsorption isotherm characterizes also the Clin–Fe system, however, it seems that the adsorption does not reach a plateau although the slope of the isotherm is gradually decreased. From Figs. 1 and 2 it is obvious that when Mn2+ ions concentration is lower than 1.82 103 M (or o100 ppm), the newly synthesized system has the ability to retain almost the entire Mn2+ solution concentration. Table 2 summarizes the calculated QP values of the two substrates for each one of the initial Mn concentrations. QP values are low for Clin and consequently is questionable if such a material, with no pre-treatment, could be used in a potential water purification process. For Clin–Fe system the higher values of QP values were obtained within 0.50 and 100 ppm Mn, while for Clin within 1.00 and 4.00 ppm. Since QP expresses the solution volume that can be purified, one concludes that the Clin–Fe system is capable to purify almost double solution volume than Clin for initial Mn concentration lower than 10.0 ppm, whereas for higher initial Mn concentration the Clin–Fe system can purify a solution volume almost three or four times larger than untreated clinoptilolite. It should also be pointed out that these results were obtained from only one-round sample treatment and significantly better results are expected from an integrated water treatment process. Moreover, because retention values are dependent on the adsorbent dose, higher retention values, for both substrates, could be achieved by using higher sorbents dose. By comparing the Mn adsorption capacity of the Clin–Fe system with the adsorption capacity of other natural materials, mentioned above, we conclude that such a mixed system could be used for the purification of heavy contaminated
ARTICLE IN PRESS WAT E R R E S E A R C H
4 0 (200 6) 316 7 – 317 6
Clinoptilolite Clinoptilolite-Fe system
Fig. 1 – Mn adsorption isotherms for Clinoptilolite and Clinoptilolite–Fe system. X: concentration of adsorbed Mn; C: concentration of Mn in equilibrium solutions.
Clin Clin-Fe system 100 90 80
% Mn Adsorption
70 60 50 40 30 20 10 0 Co,
Fig. 2 – Percentage adsorption of Mn for Clin and Clin–Fe system.
water samples. This property along with the fact that the new system is inexpensive, easily regenerated and harmless for human beings as well as for the environment could characterize the Clin–Fe system as a very promising Mn adsorbent. The differences in chemical behavior, in the adsorption of Mn2+ ions, in counterbalanced ions release and in Si/Al dissolution of the two substrates are owed to the different surface species which are located on the two substrates as mentioned above. Owing to the presence of non-crystalline Fe formations located in cationic positions in the zeolite channels, of Fe
binuclear and, in general, iron complexes in extra-framework positions as well as amorphous iron oxides FeOx located at the surface of the zeolite crystal, the Clin–Fe system has higher specific surface area and thus higher adsorption capacity than untreated Clin (Pe´rez-Ramirez et al., 2002). The higher adsorption detected for Clin–Fe system is also due to its high negative surface charge. After the equilibrium period the new system stabilized the samples solution pH at values within the basic pH range (8.14). Thus, the Si–O, Al–O and Fe–O groups predominate, giving the surface a net negative charge which is available to adsorb Mn from solution. On the contrary, Clin stabilized the solution pH at
ARTICLE IN PRESS 3172
WAT E R R E S E A R C H
40 (2006) 3167– 3176
Table 2 – Conductivities, pH, hardness and QP values for all water samples treated with Clin and Clin–Fe system Clinoptilolite
0.20 0.50 1.00 4.00 10.0 50.0 100 200 500 1000
7.39 7.45 7.38 7.51 7.39 7.18 7.09 7.00 6.99 6.85
280 280 281 291 318 477 674 1071 2170 3850
103 104 106 105 112 136 153 167 196 197
38.7 58.6 62.7 65.1 57.7 33.1 29.5 23.7 14.6 5.31
8.23 8.26 8.27 8.27 8.24 7.85 7.48 7.09 6.83 6.72
374 376 380 390 417 576 766 1159 2230 3980
23.4 25.3 23.8 26.2 28.6 73.3 128 193 278 319
69.4 92.0 98.2 98.0 98.7 98.4 92.2 69.3 40.7 26.2
Conductivity in mS/cm; hardness in ppm CaCO3; QP in l/kg.
lower value (7.53) and thus it is expected to carry less negative surface charge. The pH of the aqueous solution is an important controlling parameter in the sorption process and metal removing typically increases with increasing pH values. Except for the effect on the ionization degree of the surface groups of the sorbent the solution pH influences metal speciation. Heavy metal ions (Mz+) may form complexes with inorganic ligands such as OH and the extent of the complex formation varies with pH. The exact specifiation of a metal has a significant impact on the removal efficiency of the sorbents. Depending on the pH and metal concentration, Mn may form complexes with OH (for example Mn(OH)+, 2 Mn(OH)2, Mn(OH) 3 , Mn(OH)4 ) at higher pH values, and as a result Mn-hydroxyl species may participate in the adsorption and precipitation onto the zeolite structure (Hui et al., 2005). However, by using a chemical equilibrium model (CHEAQS) it was proved that the entire quantity of Mn is present in solution as Mn2+ for both substrates (Verweij, 2005). There is only one exception for Clin–Fe system when the Mn initial concentration is 1000 ppm. In this case the percentage of Mn2+ is 94.56%; of Mn(OH)+ is 0.18%; and of Mn2(OH)3+ is 2.72%.
Conductivity and samples pH
Table 2 presents the conductivities and pH values for all water samples after Mn2+ adsorption experiment. According to this data, clinoptilolite resists pH changes and after the adsorption process all the samples have pH values within the neutral pH range. The samples conductivity values are also low. Extremely high increase in the sample’s conductivity was reported only for the three higher Mn concentrations. On the contrary, the Clin–Fe system stabilizes the sample solution pH to a higher values then the untreated Clin. By considering the decrease in solution pH values for both substrates one may conclude that the two solids release H+, during Mn adsorption experiments. However, in order to calculate the real concentration of released/adsorbed H+ we have to consider into our calculations the concentration of H+ in equilibrated solutions (before and after adsorption) as well as the concentration of H+ added along with Mn stock
solutions (the increase in Mn concentration results in a decrease in solution pH values). After these calculations, it was confirmed that the adsorption of H+ is the predominant process for both solids, although H+ release was detected for Clin but only for the two lower Mn initial concentrations. Thus, the continuous decrease in solution pH values is owed to the continuous increase in H+aq by external addition. Trying to neutralize the surrounding environments the two solids adsorb H+ ions from solutions and despite the final decrease in solutions pH values, the calculations prove that the predominant process is the adsorption of H+. Generally, zeolites and zeolite systems tend to neutralize the solutions, acting either as proton acceptors or as proton donors, exhibiting thus an amphoteric character (Filippidis et al., 1996). During Mn adsorption process the H+ ions participate in many reactions in the solution and the solid phase. Hydrogen ions were added to the water samples along with Mn2+ solutions and their concentrations were increased as higher Mn2+ concentrations were used. These H+ species together with those released from the surface as a result of surface reactivity participate in a lot of solution and surface processes. They are not only capable to provoke surface protonation (Eqs. (2) and (3)) but also to affect all reactions in which H+ or OH are involved: S2OH þ Hþ aq
S2OH2 þ ;
S2O þ Hþ aq
where S corresponds to the surface central metal (i.e., Si, Al). Especially for Clin–Fe system, it was calculated that it adsorbs larger amounts of H+aq than Clin for the six lower Mn concentrations. For the rest of Mn concentrations the adsorption of H+ is almost equal for the two sorbents. This behavior is expected because for low Mn concentrations (and low H+aq concentrations as well as limited movement of the counterbalanced ions) there is more negative charge available on Clin–Fe system capable to adsorb positive ions from the solution. The substrate saturation is also obvious by considering the adsorption data of Figs. 1 and 2, where the decrease in the isotherm slope as well as the decrease in the
ARTICLE IN PRESS WAT E R R E S E A R C H
percentage adsorption of Mn start from almost the same value of Mn initial concentration.
Movement of counterbalanced ions
Figs. 3 and 4 present the amounts of Ca2+, Mg2+, Na+ and K+ (in eq/l) released during Mn adsorption as a function of initial Mn concentrations. They also present the overall release of positively charged ions from the two sorbents as well as the overall adsorption of positively charged ions (i.e., H+ and Mn2+) by the two solids. During adsorption stage, release of counterbalanced ions takes place partly as a consequence of Mn2+ adsorption. Retention of Mn2+ is undoubtedly the process which determines and forces directly and indirectly, the release of counterbalanced ions. Metals form either outer- or innersphere complexes with surface sites. The outer-sphere complexation involves the ion-exchange reactions between metal ions and surface counterbalanced ions: ð S2O Þ2 . . . Cnþ 3n þ Mn2þ !ð S2O Þ2 . . . Mn2þ þ ð3 nÞCnþ ;
where C is the counterbalanced ion with charge n+ (n ¼ +1 or +2). During inner-sphere complexation hydrogen ions are released as products and the process causes a total decrease in solution pH (Stumm, 1991): S2OH þ Mn2þ !
S2O2Mnþ þ Hþ ;
4 0 (200 6) 316 7 – 317 6
2S2OH þ Mn2þ ! ð S2OÞ2 2Mn þ 2Hþ :
By comparing the two figures it is obvious that the behaviors of Na+ and K+ ions are significantly different for the two solids. Clin releases higher concentrations of Na ions and lower concentrations of K+ ions than Clin–Fe system. However, both released concentrations are constant during adsorption and equal to the concentrations released during equilibrium stage. Such an observation was also reported by the authors for Cu adsorption by clinoptilolite (Doula et al., 2002). It was suggested then, that almost the entire concentrations of Na and K were released during equilibrium stage and these ions did not participate in ion-exchange reactions. Another reason for this constant release is that Clin contains impurities which are dissolved and enriched the solution with Na+ ions. Owing to this dissolution the relation between released and retained equivalents of positive charges is not stoichiometric. For the Clin–Fe system the release of K+ is significantly higher than the release of the other counterbalanced ions except for the two higher Mn initial concentrations. The release of K+ ions remain constant with a small increase for the highest Mn concentration. The high release of K+ is owed to the synthesis procedure of the Clin–Fe system: the new system was in contact for 60 h with a 5 M KOH solution and thus, high concentrations of this ion were deposited on system sites which were dissolved during equilibrium and adsorption stage. The observation that the concentrations of K+ and Na+ ions do not vary as a function of Mn adsorption is evidence that
Mg Ca K
Total Adsorption Total Desorption
0.20 0.50 1.00 4.00 10.0
Initial Mn concentration, ppm Fig. 3 – Concentrations of released Na+, Mg2+, Ca2+, K+ ions, total adsorbed cations (H+ and Mn2+) and total desorbed cations during Mn adsorption by Clinoptilolite.
ARTICLE IN PRESS 3174
WAT E R R E S E A R C H
40 (2006) 3167– 3176
Clinoptilolite-Fe system Na Mg Ca
K Total Adsorption Total Desorption
Initial Mn concentration, ppm Fig. 4 – Concentrations of released Na+, Mg2+, Ca2+, K+ ions, total adsorbed cations (H+ and Mn2+) and total desorbed cations during Mn adsorption by Clin–Fe system.
they neither influence nor are affected by the adsorption of Mn. For low Mn initial concentrations the release of both Ca2+ and Mg2+ are higher from the surface of Clin than from the surface of Clin–Fe system. The release becomes higher for the Clin–Fe system only for the three higher Mn initial concentrations examined. Thus, the significant result obtained concerns the capability of the Clin–Fe system to adsorb high Mn concentrations from solutions and simultaneously to maintain the Ca and Mg solution concentrations low. As a consequence, the hardness of the water samples treated with the Clin–Fe system is significantly lower than the respective ones for Clin (Table 2). By increasing initial Mn concentrations the two solids are forced to adsorb even higher metal concentrations. Clin does not respond to this demand whereas the newly synthesized system succeeds to adsorb even larger Mn concentrations by simultaneous increase in Ca and Mg release. Ion-exchange process is more likely to predominate for Clin–Fe system when initial concentration of Mn is lower than 3.64 103 M (or 200 ppm). Within this concentration range (0–200 ppm) the release of Ca2+ and Mg2+ follows the adsorption of Mn quantitatively and qualitatively. When Mn concentration becomes higher than 200 ppm the amounts of released Ca and Mg are lower than the adsorbed amounts of Mn. By considering that K+ and Na+ ions do not significantly affect the adsorption one concludes that except ion-exchange between Mn2+ ions and Ca2+/Mg2+ ions there is also another type of Mn retention on the system’s surface. Thus, it is
possible that Mn ions form inner-sphere complexes with surface sites as described by Eqs. (5) and (6). Nevertheless, the fact that the inner-sphere complexation is more obvious for C04200 ppm does not exclude the formation of such complexes for lower Mn concentrations. The release of Mg2+ ions is almost constant for Clin, although there is a slight increase in Mg solution concentration as Mn retention becomes higher. On the contrary, the release of Ca2+ is noticeable high for Clin and it seems that the increase in its presence in equilibrated solutions is affected by Mn adsorption. However, the total amount of released Ca and Mg is higher than the adsorbed amounts of Mn meaning that part of the Ca, Mg release is owed to ion-exchange between them and Mn from solution. The dissolution of zeolite framework can also affect and control the presence of Ca and Mg (but also of other cations) in solutions. Under specific experimental conditions Si and Al from the framework move toward solution and this process is characterized as dissolution and depends mainly on the solution pH, on the extent of surface protonation and on the nature of solution ions (Stumm, 1991). Generally, the dissolution of framework Si and Al causes the local distraction of the framework and the release of more counterbalanced cations. Fig. 5 presents the concentrations of Si and Al found in solutions after Mn adsorption. The dissolution of Si is higher for Clin than for Clin–Fe system and this is a possible explanation for the increased presence of Ca and Mg ions in solutions. Although the Si dissolution seems to be constant and unaffected by Mn adsorption, the equivalents of Si
ARTICLE IN PRESS WAT E R R E S E A R C H
4 0 (200 6) 316 7 – 317 6
3.00 5.00 Clinoptilolite Clinoptilolite-Fe 4.00
CAl x10-5, mol/l
Csi x10-4, mol/l
Fig. 5 – Dissolution of framework Si and Al for Clin and Clin–Fe system as a function of Mn2+ initial concentrations.
dissolved are almost equal to 1.60 104, meaning that the release of Ca2+ and Mg2+ could be influenced by framework dissolution. Although the retention of Mn by Clin seems to be controlled mainly by ion-exchange (Fig. 3), it is also possible that Mn ions form inner-sphere complexes with surface sites. A method to confirm the formation of inner-sphere complexation is to carry out desorption experiments because during this process the most available ions (outer-sphere complexed) enter in the solution, while the more stable inner-sphere complexes require more drastic conditions in order to break down their covalent bonds with the surface and to enter the aquatic phase.
The Clin–Fe system used for the adsorption of Mn from
drinking water samples was synthesized by mixing Clinoptilolite with aquatic solution of Fe(NO3)3 under strongly basic conditions (5 M KOH). The new system contains 14.1% Fe as amorphous iron species and its Fe/ Al ratio is equal to 1.23. It has specific surface area equal to 151.0 m2/g, which is significantly higher than the SSA of untreated clinoptilolite (30.98 m2/g). The absolute characterization of the Clin–Fe system is complex, due to many possible states of Fe3+ ions in an environment of oxygen-containing ligands. Generally in such systems, single di- and trivalent Fe ions, oxo- and hydroxo-complexes, polymeric oxidic species and iron oxide species are present simultaneously. For one-round purification experiment and for the specific solid/sample ratio used (1/100) the results indicate that Clin has a satisfactory adsorption behavior. The maximum percentage of Mn adsorption reaches 65%, while the
lowest is almost 7%. For the experimental conditions used, the Mn adsorption capacity of Clin is 7.69 mg/g. The Clin–Fe system has a noticeable higher Mn adsorption capacity (27.12 mg/g) than its parent material. For Mn concentrations lower than 100 ppm the Clin–Fe system is capable to remove almost the entire Mn2+ solution quantity. The high Mn adsorption capacity of Clin–Fe system owed to Fe-clusters located on its surface, to high surface negative charge as well as, to its high specific surface area. After treatment with the Clin–Fe system the water samples have significantly low hardness and it seems that, except the adsorption of Mn2+ ions, the new system acts simultaneously as a water softening material. Despite the high adsorption of Mn2+ ions, the Clin–Fe system maintains the chemical parameters of the water samples to very satisfactory values. One should also point out that these results were obtained from only one-round sample treatment and one expects significantly better results from an integrated water treatment process.
R E F E R E N C E S
Abollino, O., Aceto, M., Malandrino, M., Sarzanini, C., Mentasti, E., 2003. Adsorption of heavy metals on Na-montmorillonite. Effect of pH and organic substances. Water Res. 37, 1619–1627. Bosco, L.M., Jimenez, R.S., Carvalho, W.A., 2005. Removal of toxic metals from wastewater by Brazilian natural scolecite. J. Colloid Interf. Sci. 281, 424–431. Chen, H.Y., Voskoboinikov, T., Sachtler, W.M.H., 1998. Reduction of NOx over Fe/ZSM-5 catalysts: adsorption complexes and their reactivity toward hydrocarbons. J. Catal. 180, 171–183. Cornell, R.M., Schwertmann, U., 1996. The Iron Oxides, Structure, Properties, Reactions, Occurrence and Uses. VCH Verlagsgesellschaft mbH: D-69451 Weinheim, Germany, pp. 87–102.
ARTICLE IN PRESS 3176
WAT E R R E S E A R C H
40 (2006) 3167– 3176
Dimirkou, A., Ioannou, A., Kalliannou, Ch., 1996. Synthesisidentification of hematite and kaolinite–hematite (K–H) system. Commun. Soil Sci. Plant Anal. 27, 1091–1106. Dimirkou, A., Ioannou, A., Doula, M., 2002. Preparation, characterization and sorption properties for phosphates of hematite, bentonite and bentonite–hematite system. Adv. Colloid Interf. Sci. 97, 37–61. Division of Environmental Epidemiology and Occupational Health, 2001. Fact Sheet-‘‘What you need to know about Manganese in Drinking water’’. State of Connecticut, Department of Public Health, Hartford /www.state.ct.us/dphS. Doula, M., 2003. Heavy metals adsorption onto inorganic colloids (zeolites). Ph.D. Thesis, National Kapodistrian University of Athens, Athens, Greece. Doula, M., Ioannou, A., 2003. The effect of electrolyte anion on Cu adsorption-desorption by Clinoptilolite. Microp. Mesop. Mater. 58, 115–130. Doula, M., Ioannou, A., Dimirkou, A., 2002. Copper adsorption and Si, Al, Ca, Mg and Na release from Clinoptilolite. J. Colloid Interf. Sci. 245, 237–250. Erdem, E., Karapinar, N., Donat, R., 2004. The removal of heavy metal cations by natural zeolites. J. Colloid Interf. Sci. 280, 309–314. Filippidis, A., Godelitsas, A., Charistos, D., Misaelides, P., KassoliFournaraki, A., 1996. The chemical behavior of natural zeolites in aqueous environments: interactions between low-silica zeolites and 1 M NaCl solutions of different initial pH-values. Appl. Clay Sci. 11, 199–209. Green, C.H., Heil, D.M., Cardon, G.E., Butters, G.L., Kelly, E.F., 2003. Solubilization of manganese and trace metals in soils affected by acid mine runoff. J. Environ. Qual. 32, 1323–1334. Hui, K.S., Chao, C.Y.H., Kot, S.C., 2005. Removal of mixed heavy metal ions in wastewater by zeolite 4A and residual products from recycled coal fly ash. J. Hazardous Mat. B 127, 89–101. Inglezakis, V.J., Loizidou, M.D., Grigoropoulou, H.P., 2002. Equilibrium and kinetic ion exchange studies of Pb2+, Cr3+, Fe3+ and Cu2+ on natural clinoptilolite. Water Res. 36 (11), 2784–2792. Ioannou, A., Dimirkou, A., Theodoropoulou, E., 1996. Phosphate sorption by hematite and kaolinite–hematite (K–H) system as
described by isotherms. Commun. Soil Sci. Plant Anal. 27, 1925–1947. Jusoh, A., Cheng, W.H., Low, W.M., Aini, A.N., Noor, M.J.M.M., 2005. Study on the removal of iron and manganese in groundwater by granular activated carbon. Desalination 182, 347–353. Ko¨gel, M., Mo¨nning, R., Schwieger, W., Turek, T., 1999. Simultaneous catalytic removal of NO and N2O using Fe–MFI. J. Catal. 182, 470–478. Kroik, A.A., Shramko, O.N., Belous, N.V., 1999. Sewage cleaning with applying of natural sorbents. Chem. Technol. Water 21 (3), 310. Leinonen, H., 1999. Removal of harmful metals from metal plating waste waters using selective ion exchangers. In: Report Series in Radiochemistry, vol. 13, University of Helsinki, ISBN 951-458759-6 (PDF Version). Morturano, P., Drozdova´, L., Pirngruber, D., Kogelbauer, A., Prins, R., 2001. The mechanism of formation of the Fe species in Fe/ ZSM-5 prepared by CVD. Phys. Chem. Chem. Phys. 3, 5585–5595. Oelkers, E.H., Schott, I., 1995. Experimental study of anorthite dissolution and the relative mechanism of feldspar hydrolysis. Geochim. Cosmochim. Acta 59, 5039–5053. Pe´rez-Ramirez, J., Mul, G., Kapteijn, F., Moulijn, J.A., Overweg, A.R., Dome´nech, A., Ribera, A., Arends, I.W.C.E., 2002. Physicochemical characterization of isomorphously substituted FeZSM-5 during activation. J. Catal. 207, 113–126. Ruthven, M.D., 2001. Verified Syntheses of Zeolitic Materials. Elsevier, New York. Schwertmann, U., Cornell, R.M., 1991. Iron Oxides in the Laboratory, Preparation and Characterization. VCH Verlagsgesellschaft mbH, Germany, pp. 61–68. Stumm, W., 1991. Chemistry of the Solid–Water Interface. Wiley, New York, p. 187. Urquieta-Gonza`lez, E.A., Martins, L., Peguin, R.P.S., Batista, M.S., 2002. Identification of extra-framework species on Fe/ZSM-5 and Cu/ZSM-5 catalysts typical microporous molecular sieves with zeolitic structure. Mater. Res. 5 (3), 321–327. Verweij, W., 2005. ‘‘Chemical Equilibria in Aquatic SystemsCHEAQS’’-PC calculating programm /http://home.tiscali.nl/ cheaqs/index.htmlS.