1 Characterizing the Discoloration of Methylene Blue in Fe0/H2O Systems. 1 C. Noubactep 2 Angewandte Geologie, Universität Göttingen, Goldschmidtstraße 3, D - 37077 Göttingen, Germany. 3 Tel.: +49 551 39 3191, Fax.: +49 551 399379, e-mail: cnoubac@gwdg.de 4 5 Abstract 6 Methylene blue (MB) was used as a model molecule to characterize the aqueous reactivity of 7 metallic iron in Fe 0/H2O systems. Likely discoloration mechanisms under used experimental 8 conditions are: (i) adsorption onto Fe0 and Fe 0 corrosion products (CP), (ii) co -precipitation with 9 in-situ generated iron CP, (iii) reduction to colorless leukomethylene blue (LMB). MB 10 mineralization (oxidation to CO 2) is not expected. The kinetics of MB discoloration by Fe 0, 11 Fe2O3, Fe 3O4, MnO 2, and granular activated carbon were investigated in assay tubes under 12 mechanically non -disturbed conditions. The evolution of MB discoloration was monitored 13 spectrophotometrically. The effect of availability of CP, Fe 0 source, shaking rate, initial pH 14 value, and chemical properties of the solution were studied. The results present evidence 15 supporting co -precipitation of MB with in -situ generated iron CP as main discoloration 16 mechanism. Under high shaking intensities (> 150 min -1), increased CP gener ation yields a 17 brownish solution which disturbed MB determination, showing that a too high shear stress 18 induced the suspension of in -situ generated corrosion products. The present study clearly 19 demonstrates that comparing results from various sources is di fficult even when the results are 20 achieved under seemingly similar conditions. The appeal for an unified experimental procedure 21 for the investigation of processes in Fe0/H2O systems is reiterated. 22 23 Keywords: Adsorption; Co-precipitation; Iron Corrosion, Methylene Blue; Zerovalent Iron. 24 25 2 Introduction 25 Permeable reactive barriers using elemental iron -based alloys (Fe 0-based alloys widely termed as 26 zerovalent iron) as a reactive medium have been proven to be an efficient and affordable 27 technology for removing i norganics and organics species from groundwater [1 -7]. Even living 28 species like viruses have been successfully removed [8]. Despite 15 years of intensive 29 investigations, the removal mechanisms of contaminants in Fe 0 treatment systems are still not 30 well und erstood [9,10]. In fact, the well -established premise that contaminant removal results 31 from the low electrode potential of the redox couple Fe II/Fe0 (E0 = -0.44 V) can not explain why 32 redox-insensitive species are quantitatively removed [11,12]. However, understanding the nature 33 of primary processes yielding to contaminant removal in Fe 0/H2O systems is of fundamental 34 importance for advancing technological applications. The accurate knowledge of these processes 35 will favor the identification of factors domina ting the general reactivity of Fe 0/H2O systems, 36 which is of fundamental importance for the long -term stability of iron reactive barriers. A more 37 rational devising of Fe 0 treatment systems for an effective and economical contaminant removal 38 could be achieved. 39 Fe0 oxidation releases dissolved iron species (Fe II, Fe III) which hydrolyse with increasing pH and 40 precipitate primarily as hydrous oxides (oxide-film) or corrosion products (CP). Oxide-films (CP) 41 of varied composition and thickness develop at all aqueo us Fe 0/H2O interfaces [13,14]. 42 Therefore, an aqueous Fe 0 treatment system (Fe 0/H2O system) is made up of Fe 0, iron oxides 43 (oxide-film), and water (H2O). Contaminant adsorption onto the oxide -film and reduction by Fe 0 44 have mostly been evaluated as separate, independent processes that occur simultaneously or 45 sequentially on metal surfaces. However, contaminants may be primarily quantitatively 46 sequestered by in situ generated hydrous iron oxides (co -precipitation) [11,12]. Initial corrosion 47 products polymerise and precipitate, first as very reactive oxides having short -range crystalline 48 3 order and after aging as crystalline oxides [15 -18]. Subsequent abiotic direct reduction (electrons 49 are transferred from Fe 0) or indirect reduction (electrons from Fe II, H/H 2) of adsorbed or co -50 precipitated contaminants is possible. As a rule co -precipitation occurs whenever the 51 precipitation of a major species (e.g., iron oxide) takes place in the presence of foreign species 52 (e.g., contaminants) and has been documented for organ ics [16,17,19,20], inorganics [21 -23] and 53 living species [8] under various conditions. Generally, adsorption and co -precipitation are 54 considered to be related such that in order for co -precipitation to occur, sorption to the surface of 55 a forming solid occu rs and the adsorbed species is then sequestered in the matrix of the 56 precipitating phase (e.g. iron hydroxide). However, co -precipitation in Fe 0/H2O systems may be 57 primarily regarded as a non -specific removal mechanism [11,17] as to be demonstrated in this 58 study of a process involving the discoloration of methylene blue. 59 Methylene blue (MB) is a well-known redox indicator [24] and is a cationic thiazine dye with the 60 chemical name tetramethylthionine chloride. It has a characteristic deep blue colour in the 61 oxidized state; the reduced form (leukomethylene blue - LMB) is colorless. MB has been widely 62 used in environmental sciences primarily to access the suitability of various materials for 63 wastewater discoloration [25-29]. The mechanism of MB removal by Fe 0-based materials which 64 may be suitable for environmental remediation (cast iron, low alloy steel) has not been yet 65 systematically investigated. Imamura et al. [30] investigated the mechanism of adsorption of 66 methylene blue and its congeners onto stainless st eel particles. MB has also been used for 67 corrosion inhibition of mild steel in acid solutions [31]. 68 The literature on “Fe 0 technology” is characterized by the fact that, since the effectiveness of Fe 0 69 reactive walls to degrade solvents was demonstrated, th e feasibility of applying Fe 0 to treat other 70 compounds (or group of compounds) are performed without previous systematic investigations 71 [9]. For example, while presenting the discoloration of MB by a Fe/Cu bimetallic system, Ma et 72 4 al. [28] referenced sever al works dealing with dyes in general [32 -34]. The authors did not 73 specified whether the referenced works have used MB. Furthermore, their experimental 74 procedure did not include a system with Fe 0 alone to evidence the improvement induced by Cu 0 75 addition. 76 Given the diversity of contaminant removal mechanisms in a Fe 0/H2O system, an approach to 77 elucidate the mechanism of contaminant removal in the system is to characterise the removal 78 process of the contaminant in question by a pure adsorbent (e.g. activated carbon - AC), and 79 model iron corrosion products (Fe2O3, Fe3O4) under the same experimental conditions [35]. Here, 80 comparing the evolution of contaminant removal in the systems with pure adsorption (AC, Fe2O3, 81 Fe3O4) and in the system with Fe 0 will help dis cussing the removal mechanism. Another 82 approach consists in introducing MnO 2 to delay the availability of corrosion products in the 83 system [36]. MnO 2 readily reacts with Fe II from Fe 0 corrosion products: reductive dissolution of 84 MnO2 by Fe II [37]. If the p rocess of contaminant removal is coupled with the precipitation of 85 iron, then contaminant removal will be delayed as long as the added amount of MnO 2 consumes 86 FeII for reductive dissolution as it will be presented later. 87 The present study is an attempt to elucidate the physico -chemical mechanism of MB 88 discoloration in Fe 0/H2O systems by compar ing the kinetics and/or the extent of MB 89 discoloration by Fe0 and different materials: granular activated carbon (GAC or AC), iron oxides 90 (Fe2O3, Fe3O4) and manganese dioxide (MnO 2). Non-disturbed (not shaken or shaking at 0 min -1) 91 batch experiments were performed in order to allow formation and transformation of corrosion 92 products at the surface of Fe 0 as it occurs in the nature and in column experiments. The effects of 93 various factors (initial pH value, mixing intensity, particle size, Fe 0 source, Cl -, HCO3 -, EDTA) 94 on the extent of MB discoloration are discussed. The results show that MB quantitative 95 5 discoloration is mostly due to co -precipitation with in -situ generate d corrosion products. 96 Therefore, MB discoloration occurs within the oxide-film on Fe 0. 97 Background of the Experimental Methodology 98 A survey of the electrode potentials of the redox couples relevant for the discussion in this study 99 [FeII(aq)/Fe 0, Fe III(aq)/Fe II (aq), Fe III (s)/Fe II (s), MnO2/Mn 2+, O2/HO -, and MB +/LMB (Eq. 1 to Eq. 6)] 100 suggests that from the available iron species, Fe 0 and Fe II(s) can reduce MB. Equation 2 is that of 101 the adsorbed Fe II known as structural Fe II. The electrode potential of this redo x couple was 102 determined by White and Patterson [38]. The electrode potential of Eq. 3 to 6 shows that Fe III(aq), 103 dissolved O2 and MnO2 may re-oxidize colorless LMB to blue MB +. 104 Reaction E0 (V) Eq. Fe2+ + 2 e - Û Fe0 -0.44 (1) Fe3+(s) + e - Û Fe 2+ (s) -0.36 to -0.65 (2) MB+ + 2 e- + H+ Û LMB 0.01 (3) Fe3+(aq) + e - Û Fe2+(aq) 0.77 (4) O2(aq) + 2 H2O + 4 e - Û 4 OH- 0.81 (5) MnO2 + 4 H + + 2 e - Û Mn2+(aq) + 2 H2O 1.23 (6) Reductive MB discoloration in this study may be the result of either (i) Fe 0 corrosion (oxidation 105 to Fe II(aq)) (Eq. 1) or (ii) oxidation of adsorbed Fe II (FeII(s) to Fe III (s) - Eq. 2). Additionally, MB 106 adsorption onto in situ generated and aged Fe 0 corrosion products and MB entrapment in the 107 structure of forming corrosion products (co -precipitation) are two further discoloration 108 mechanisms. Therefore, it is difficult to resolve the effect of specific redox reactions on MB 109 discoloration from the effects of other processes. To resolve this problem two additives are added 110 to Fe 0: granular acti vated carbon (GAC) and manganese dioxide (MnO 2). GAC is a pure 111 adsorbent for MB [25] whereas reductive dissolution of MnO 2 has been reported to decolorize 112 6 MB [39]. The presentation above shows that MnO 2 should re -oxidise reduced LMB (no 113 discoloration). The refore, MB discoloration in the presence of MnO 2 could only result from 114 adsorption. On the other hand, MnO 2 is known to be reductively dissolved by Fe II [37, 40]. By 115 consuming FeII, MnO 2 accelerates Fe 0 corrosion, producing more adsorption or co -precipitation 116 agents for MB. Increased adsorption is supported by the fact that iron corrosion products are of 117 higher specific surface area (> 40 m2 g-1) than the used Fe 0 (0.29 m2 g-1). The reductive 118 dissolution of MnO 2 (Eq. 7 and 8) produce further new reactive ad sorbents (MnOOH and 119 FeOOH). 120 Fe2+(aq) + MnO 2 + 2 H 2O Þ FeOOH + MnOOH + 2 H+ (7) 2 Fe2+(aq) + MnO 2 + 2 H2O Þ 2 FeOOH + Mn 2+ + 2 H+ (8) Noubactep et al. [36] have shown that MnO2 retards the availability of free corrosion products for 121 contaminant co-precipitation. 122 The used methodology for the investigation of the process of MB discoloration mechanism by 123 Fe0 consists in following the MB discoloration in the presence of MnO 2 (“Fe 0” and “Fe0 + MnO2” 124 systems). Thus, the availability of corrosion products for MB co-precipitation in the bulk solution 125 is delayed by the addition of MnO 2. It should be kept in mine that MB discoloration and not MB 126 removal is discussed in this study. For the discussion of MB removal TOC measurements for 127 instance should have been necessa ry to account for MB reduction to LMB which remains in 128 solution. 129 Materials and Methods 130 Solutions 131 The MB molecule has a minimum diameter of approximately 0.9 nm [25,41]. As positively 132 charged ions, MB should readily adsorb onto negatively charged surface. T hat is at pH > pH pzc; 133 pHpzc being the pH at the point of zero charge [42,43]. The used initial concentration was 20 mg 134 7 L-1 (~0.063 mM) MB and it was prepared by diluting a 1000 mg L-1 stock solution. All chemicals 135 were analytical grade. 136 Solid Materials 137 The main Fe 0 material (ZVI0 – Tab. 1) is a readily available scrapped iron. Its elemental 138 composition was found to be: C: 3.52%; Si: 2.12%; Mn: 0.93%; Cr: 0.66%. The material was 139 fractionated by sieving. The fraction 1.6 - 2.5 mm was used. The sieved Fe 0 was used without 140 any further pre -treatment. Further 13 commercial Fe 0 samples (ZVI1 through ZVI13) were used 141 in the set of experiments aiming at characterizing the impact of Fe 0 source. The main 142 characteristics of these materials are summarized in table 1, whi ch is quite typical for a large 143 range of powdered and granular Fe 0 used in laboratory investigations and field works. 144 The used granular activated carbon (GAC or AC from LS Labor Service GmbH - Griesheim) was 145 crushed and sieved. The particle sized fraction ranging from 0.63 to 1.0 mm was used without 146 further characterization. Granular activated carbon is used as porous adsorbent for MB [25,26]. 147 Powdered commercial Fe 2O3 (Fluka), Fe 3O4 (Fisher Scientific) and MnO 2 (Sigma -Aldrich) were 148 purchased and used witho ut any further characterization. Fe 2O3 and Fe 3O4 were also used as 149 possible MB adsorbents and are proxies for aged iron corrosion products (Tab. 2). 150 Broken manganese nodules (MnO 2) collected from the deep sea with an average particle size of 151 1.5 mm and ele mental composition of Mn: 41.8%; Fe: 2.40%; Si: 2.41%; Ni: 0.74%; Zn: 0.22%; 152 Ca: 1.39%; Cu: 0.36% were used. These manganese nodules originated from the pacific ocean 153 (Guatemala- basin: 06°30 N, 92°54 W and 3670 m deep). The target chemically active 154 component is MnO 2, which occurs naturally mainly as birnessite and todorokite [44]. MnO 2 was 155 mainly used to control the availability of in situ generated oxides from Fe 0 corrosion [36, 45]. 156 Reductive dissolution of MnO 2 has been reported to degrade a number of o rganic pollutants [39, 157 8 46 and ref. therein]. Zhu et al. [39] reported the quantitative discoloration of MB by deep sea 158 manganese nodules (pelagite). 159 Rationale for Choice of Test Conditions 160 Materials selected for study were known to be effective for adsorbi ng MB (GAC), discoloring 161 MB (Fe 0, MnO 2) or delaying the availability of iron corrosion products in Fe 0/H2O systems 162 (MnO2). Fe 2O3 and Fe 3O4 were used to characterize the reactivity of aged corrosion products. 163 Table 2 summarises the function of the individua l materials and gives the material surface 164 coverage in individual reaction vessels. The detailed method for the calculation of the surface 165 coverage (q) is presented by Jia et al. [47]. The minima of reported specific surface area (SSA) 166 values of the adsorb ents were used for the estimation of surface coverage. The Fe 0 SSA was 167 earlier measured by Mbudi et al. [52]. The value 120 Å 2 is considered for the molecular cross-168 sectional area of MB [25]. From Tab. 2 it can be seen that, apart from Fe 0 (q = 31), all ot her 169 materials were present in excess “ stoichiometry” (q £ 0.2). This means that the available surface 170 of Fe 0 can be covered by up to 31 mono -layers of MB, whereas the other materials should be 171 covered only to one fifth with MB ( q = 1 corresponds to a mono -layer coverage). Therefore, 172 depending on the initial pH value and the affinity of MB for the individual materials (pH pzc) and 173 the kinetics of MB diffusion to the reactive sites (material porosity, mixing intensity), the MB 174 discoloration should be quantitat ive. A survey of the pH pzc values given in Tab. 2 suggests that 175 MB adsorption onto all used adsorbents should be favourable because the initial pH was 7.8. At 176 this pH value all surfaces are negatively charged; MB is positively charged. Because the available 177 Fe0 surface can be covered by up to 31 layers of MB, a progressive MB discoloration in presence 178 of Fe 0 is expected. The tests were performed under mechanically non -disturbed conditions; the 179 effect of the shaking intensity was evaluated in separated exper iments. Because diffusion is the 180 9 main mechanism of MB transport under non -disturbed conditions, long reaction times were 181 experienced to identify the main process of aqueous MB discoloration by Fe 0. 182 Discoloration studies 183 Unless otherwise indicated, batch ex periments without shaking were conducted. The batches 184 consisted of 5 g L-1 of a reactive material (GAC, Fe0, Fe2O3, Fe 3O4, MnO 2). In some experiments 185 5 g L-1 Fe 0 was mixed with 0 or 5 g L -1 AC and MnO 2 respectively. An equilibration time of 186 about 30 days w as selected to allow a MB discoloration efficiency of about 80% in the reference 187 system (ZVI0 alone). The extent of MB discoloration by AC, Fe 0, MnO 2, aged (Fe 2O3, Fe 3O4) 188 and in situ generated iron oxides was characterized. For this purpose 0.11 g of Fe 0 and 0 or 0.11 g 189 of the additive were allowed to react in sealed sample tubes containing 22.0 mL of a MB solution 190 (20 mg L-1) at laboratory temperature (about 20° C). The tubes (20 mL graded) were filled to the 191 total volume to reduce the head space in the re action vessels. Initial pH was ~7.8. After 192 equilibration, up to 5 mL of the supernatant solutions were carefully retrieved (no filtration) for 193 MB measurements. In order to fit the calibration curve for quantitative measurements, the 194 maximal dilution factor was four (4). 195 Apart from experiments aiming at investigating the impact of mixing intensity and that of the 196 initial pH value, the contact vessels were turned over-head at the beginning of the experiment and 197 allowed to equilibrate in darkness to avoid poss ible photochemical side reactions. At the end of 198 the equilibration time no attempt was made to homogenize the solutions. 199 Analytical methods 200 MB concentrations were determined by a Cary 50 UV -Vis spectrophotometer at a wavelength of 201 664.5 nm using cuvettes w ith 1 cm light path. The pH value was measured by combined glass 202 electrodes (WTW Co., Germany). Electrodes were calibrated with five standards following a 203 multi-point calibration protocol [53] in agreement with the current IUPAC recommendation [54]. 204 10 Each experiment was performed in triplicate and averaged results are presented. 205 Results and Discussion 206 After the determination of the residual MB concentration (C) the corresponding percent MB 207 discoloration was calculated according to the following equation (Eq. 9): 208 P = [1 - (C/C0)] * 100% (9) 209 where C0 is the initial aqueous MB concentration (about 20 mg L -1), while C gives the MB 210 concentration after the experiment. The operational initial concentration (C 0) for each case was 211 acquired from a triplicate contr ol experiment without additive material (so -called blank). This 212 procedure was to account for experimental errors during dilution of the stock solution (1000 mg 213 L-1), MB adsorption onto the walls of the reaction vessels and all other possible side reaction 214 during the experiments. 215 MB discoloration by different agents and discoloration mechanism by Fe 0 216 Figure 1 shows the time dependent MB discoloration curve for all the investigated materials. The 217 reference system is a blank experiment as presented above. It c an be seen that commercial Fe 2O3 218 and MnO 2 did not significantly decolourise MB over the whole duration of the experiments. It is 219 well-known, that poorly crystalline natural MnO 2 are more reactive than land -born and synthetic 220 MnO2 [39,44]. The decreasing order of discoloration efficiency at the end of the experiment was: 221 Fe0 > GAC > Fe 3O4 > MnO 2. However, the evolution of the individual systems was very 222 different. 223 (i) As expected from the surface coverage (q = 31), Fe0 presents a progressive MB discoloration 224 over the duration of the experiment. The discoloration mechanism can be the reduction to LMB 225 by Fe 0 and Fe II(s) species, adsorption onto in situ generated corrosion products and/or MB co -226 precipitation with these new corrosion products. 227 11 (ii) Fe 3O4 (20 g L -1) shows a rapid discoloration kinetic for the first 8 days. The discoloration 228 efficiency then remains constant to approximately 60% through the end of the experiment. This 229 behaviour is typical for non-porous adsorbents. Alternatively available pores may be inaccessible 230 for MB. 231 (iii) MB discoloration through GAC is insignificant at the start of the experiment (10% after 10 232 days) and then increases progressively to 75% at the end of the experiment (day 36). This 233 behaviour is typical for porous adsorbents. 234 (iv) MnO 2 (nat) shows the same behaviour as GAC but the extent of MB discoloration is 235 significantly lower (50% at day 36). Natural MnO 2 acts mostly as adsorbent. MB oxidative 236 discoloration as reported Zhu et al. [39] is not likely to occur under the experime ntal conditions 237 of this work. Note that, on the contrary to Zhu et al. [39], the experiments in this study were 238 performed under mechanically non -disturbed conditions. While investigating the effect of 239 dynamic conditions, Zhu et al. [39] did not include any non -disturbed system. They just 240 compared shaking (145 min -1) versus motor-stirring (550 min -1) and air-bubbling versus nitrogen 241 bubbling (both 32 mL s-1). These mixing conditions are pertinent to wastewater treatment 242 systems but are not reproducible in fi eld-Fe0 treatment walls, mixing could have favour MB 243 mineralisation (oxidation to CO2) which is an irreversible discoloration. 244 To better characterize the MB discoloration from aqueous solution by Fe 0, five further 245 experiments have been performed for 36 day s with 5 g L-1 Fe 0 and 0 or 5 g L -1 of GAC and 246 natural MnO2. 247 Figure 2a summarizes the results of MB discoloration in these five systems and Fig. 2b depicts 248 the evolution of MB discoloration for 5 g L -1 Fe 0 and additive (AC or MnO 2) dosages varying 249 from 0 to 9 g L -1 for an experimental duration of 36 days. Fig. 2a shows a regular evolution for 250 the systems involving AC and Fe 0. The MB discoloration efficiency decreases in the order “Fe 0 + 251 12 AC” > Fe 0 > AC. Considering AC and Fe 0 as pure adsorbents it is expected that the mixture 252 (maximal available binding sites) depicts a larger MB discoloration efficiency than individual 253 materials (Tab. 2). This trend was not observed for systems involving MnO 2. Here, the 254 decreasing order of MB discoloration efficiency was: Fe 0 > “Fe 0 + MnO 2” @ MnO 2. These 255 observations were described by Noubactep et al. [36,45,55] for uranium removal by Fe 0. A 256 “MnO2 test” was proposed for mechanistic investigations in Fe 0/H2O systems. The major feature 257 of the “MnO2 test” is that in reacting with Fe II from Fe 0 oxidation, MnO 2 delays the availability 258 of “free” corrosion products which entrapped contaminants while polymerising and precipitating. 259 “Free” corrosion products are Fe -oxides generated in the vicinity of metallic iron grains. As long 260 as MnO 2 is reductively dissolved, Fe -oxides are generated at its surface or in its vicinity. 261 Thereafter, if co-precipitation is the primary mechanism of contaminant removal, no quantitative 262 removal could occur until enough free corrosion products are available to entrap them while 263 ageing [36]. To confirm this statement the experiment presented in Fig. 2b was conducted. 264 From Fig. 2b it can be seen that about 4 g L -1 activated carbon are sufficient to achieve almost 265 100% MB discoloration. For [AC] > 4 g L-1 no additional discoloration was possible. The system 266 with MnO 2 depicts a progressive decrease of MB discoloration with increasing MnO 2 mass 267 loading. The reaction of Fe II species yielding reductive dissolution of MnO 2 is well documented 268 [37,40,56] and yields more adsorbents (e.g., FeOOH, MnOOH – Eq. 7 and 8). However, MB 269 discoloration is only quantitative when the oxidative capacity of available MnO 2 for Fe II is 270 exhausted. Thus, MB is removed from the aqueous solution through co -precipitation with in situ 271 generated iron corrosion products. The characterization of the impact of MnO 2 on contaminant 272 removal by Fe0 occurs ideally under non-disturbed conditions [57]. Note that, if the experiments 273 are performed under (too high) mixing conditions or in columns, increased c ontaminant removal 274 efficiency in the presence of MnO 2 could have been reported. For example, Burghardt and 275 13 Kassahun [58] reported increased uranium and radium removal in “Fe 0 + MnO 2” systems 276 comparatively to the system with Fe 0 alone. The results of Burgha rdt and Kassahun [58] are by 277 no means contradictory to those reported here and elsewhere [40] because the net effect of MnO 2 278 is to promote iron hydroxide formation (or to sustain corrosion) resulting in an increased 279 contaminant removal capacity. Similarly, while Noubactep et al. [36,45,57] reported a delay of U 280 removal by Fe 0 in the presence of pyrite in non -disturbed experiments, Lipczynska-Kochany et 281 al. [59] reported increased carbon tetrachloride degradation in the presence of pyrite. Pyrite is 282 known fo r its pH lowering capacity, and thus increasing iron corrosion. Non -disturbed 283 experiments allow a better characterization of the progression of involved processes. 284 Effect of Fe0 source 285 Experiments were conducted with 14 different Fe 0 materials: ZVI0 through ZVI13. ZVI1, ZVI2, 286 ZVI3 and ZVI12 were powdered materials. The 10 other samples were granulated materials. The 287 results of MB discoloration are summarised in table 1. The experimental duration was 35 days. It 288 is shown that powdered materials are more effi cient in removing MB than granulated materials 289 (Tab. 1). The discoloration efficiency for granulated materials varies from 65% for ZVI7 to 80% 290 for ZVI2 (absolute values). That is 15% reactivity difference while the maximum standard 291 deviation for the tripli cates in individual experiments was 8.5% (for ZVI12). Therefore, the Fe 0 292 source (intrinsic reactivity) is a significant operational parameter for laboratory studies. Similar 293 results were reported by Miehr et al. [60] who reported differences in constants of contaminant 294 reduction up to four orders of magnitude when comparing nine types of Fe 0. Therefore, 295 comparing results obtained with different granulated Fe 0 under comparable experimental 296 conditions may lead to erroneous conclusions. 297 14 Effect of shaking intensity 298 Figure 3 clearly shows that MB discoloration efficiency increases with the shaking intensity. The 299 experimental duration was 24 h (1 day). The reaction vessels were shaken on a rotary shaker. The 300 MB discoloration rate of 5% at 0 min-1 (non-disturbed conditions) increased to 96% at 200 min-1. 301 Between 100 and 150 min -1 the MB discoloration rate was constant to 55%. Parallel experiments 302 in 100 mL Erlenmeyer shows comparative results but at 200 min -1 the solution was no more 303 limpid and depicted a brown colo ration that persisted even after the solutions were allowed to 304 settle for 5 hours. Therefore, a mixing intensity of about 150 min -1 can be seen as the critical 305 intensity below which MB discoloration studies should be performed. Since applied mixing 306 intensities have not been tested in preliminary works, it is likely that some used mixing 307 operations have been too massive and impractical to mimic subsurface conditions [11]. Mixing 308 intensities as higher as 500 min-1 [61,62] have been used to “keep the iron powder suspended”. 309 Generally, Fe0-based materials show greater contaminant removal efficiency under mixed than 310 under non-disturbed conditions . This removal efficiency is usually attributed to direct reduction 311 whenever the thermodynamics are favourable. However , the open literature on mixed batch 312 experiments demonstrates that a minimum mixing intensity (bubbling, shaking or stirring) is 313 required for complete suspension of solid particles in a liquid medium (e.g., an aqueous solution). 314 Below this critical mixing intensity, the total surface area of the investigated particles is not 315 directly accessible for reaction and the rate of mass transfer depends strongly on stirring rate. 316 Kinetic studies aiming at distinguishing between diffusion -controlled and chemistry-controlled 317 processes have to be conducted at mixing intensities above this critical value [56]. Noubactep 318 [11] has demonstrated that experiments in Fe 0/H2O systems aiming at investigating processes 319 pertinent to subsurface situations should be conducted below the critical value (mass transfer 320 dependent). For Fe0, it is obvious, that the value of this critical mixing intensity depends on the 321 15 particle size (nm, mm, mm). Choe et al. [63] reported a critical value of 40 min -1 for nano -scale 322 Fe0 and performed their experiments at a mixing intensity of 60 min -1. According to the 323 presentation above, Choe et al. [63] would have worked with mixing intensities below 40 min-1 to 324 obtain results relevant for groundwater conditions. Furthermore, working at mixing intensities 325 above 40 min -1 accelerates iron corrosion yielding more corrosion products which are equally 326 kept suspended in the reaction medium. In the course of corrosion products formation, 327 contaminants are entrapped in the matrix of iron oxides (co -precipitation). It is well know that 328 even low adsorbable species are readily removed from aqueous solutions when precipitation 329 occurs in their presence [16,17,20,21]. As discussed above, MB discoloration mainly occurs 330 through co-precipitation with newly generated corrosion products (see above: “MB discoloration 331 by different agents and discoloration mechanism by Fe 0”). MB discoloration by aged corrosion 332 products was insignificant (Fe2O3) or very limited (Fe3O4). 333 Effect of the initial pH value 334 The effect of the initial pH on MB discoloration was investigated over the pH range of 1.5 to 335 10.0. The initial pH was adjusted by addition of 1.0 M NaOH or HCl. The experiments were 336 conducted under shaken conditions (100 min-1). The pH of the solutions was monitored at the end 337 of the experiments (24 and 48 h). The results are summarised in Fig. 4. MB discoloration was 338 negligible when the final pH was lower than 4 (P < 10%). Once the finial pH exceeded this 339 critical value, MB quantitative discoloration occurred and the extent was pH -independent (60% 340 after 24 h and 76% after 48 h). This observation is consistent with the two main types of aqueous 341 iron corrosion under oxic conditions [64,65] : (i) hydrogen evolution type (pH < 4) and (ii) oxygen 342 absorption type (pH > 4). The characteristic fe ature of “hydrogen evolution corrosion” is the 343 liberation of hydrogen as hydrogen gas (H 2) at the cathode. Hydrogen evolution corrosion is 344 normally associated with acid electrolytes (e.g., acid mine drainage) and is not relevant for the 345 16 majority of groundw aters, unless the aquifer is strictly anoxic. The “oxygen absorption” type of 346 immersed Fe 0 corrosion is characteristic of neutral waters. At these pH values (pH > 4.0) iron 347 solubility is low [66]. Thus iron oxide precipitates and MB are removed from the aq ueous 348 solution by sequestration (co -precipitation). The results from Fig. 4 validate the concept that all 349 contaminants are primarily adsorbed or/and sequestered by iron corrosion products (co -350 precipitation) [11,12]. In fact MB discoloration was quantitativ e only at final pH > 4, where iron 351 oxides precipitate due to the low solubility of Fe. Within the oxide -film, redox reactions driven 352 by Fe II species have been reported [67]. Therefore, co -precipitated MB can be reduced to LMB 353 but this reaction could not co ntribute to recorded MB discoloration. 354 A certain commonly misconception may be found in the literature concerning the process of 355 contaminant removal in Fe 0/H2O systems due to improper consideration of the two main 356 mechanisms of iron corrosion. Ideally, whe never the initial pH is lower than 4, the pH should be 357 carefully monitored and used to interpret results. From Fig. 4 it can be seen for example, that for 358 an initial pH of 3.0 the final pH was 4.3 and the extent of MB discoloration was slightly lower 359 than that of the experiment with initial pH values ³ 4 (for the given experimental duration). 360 Consequently, the repeatedly reported lag time for contaminant removal [61,68] is the time to 361 exceed pH 4 (or to enable generation of enough corrosion products for con taminant co -362 precipitation/sequestration). It must be emphasised that for contaminants (e.g., Cr IV) which are 363 also reducible by aqueous Fe II the extent of their removal at pH < 4 depends on their relative 364 solubility of their reduced form. Regardless from th e redox reactivity co -precipitation of 365 contaminant and reaction products occurs at pH > 4. Contaminants, intermediates and final 366 products are possibly entrapped in the matrix of corrosion products. 367 17 Effect of solution chemistry 368 The effect of solution parame ters on MB discoloration by Fe 0 was studied using 0.2 mM of 369 Al(NO)3, BaCl 2, CaCl 2, CuCl 2, EDTA, (NH 4)2CO3, and NiCl 2. Further non-disturbed 370 experiments were performed for 35 days with concentrations of CaCl 2, CuCl 2 and NaHCO 3 371 varying from 0 to 4 mM (Figure 5). Figure 5a shows that apart from (NH 4)2CO3 (90%) all other 372 additives lower the extent of MB discoloration by Fe 0 (78%). The lowest discoloration efficiency 373 (15%) was observed in the presence of EDTA and is consistent with the fact that complexing 374 FeII/FeIII delays the iron oxide precipitation [69 -71] and hence retards MB discoloration. For the 375 four systems containing chloride ions (Cl -), NiCl 2 depicts the lowest MB discoloration efficiency 376 (33%) and CaCl 2 the highest (72%). BaCl 2 and CuCl 2 show very com parable discoloration 377 efficiency (about 60%). This observation is partly consistent with reported results from the 378 literature on corrosion stating that: (i) at low concentration CO2 3- is corrosive, (ii) hardness (Ca2+) 379 is corrosive, while Ni 2+ has inhibiti ve properties for iron corrosion. Cu 2+ would have accelerated 380 Fe0 corrosion yielding more corrosion products for MB discoloration than in the reference system 381 (Fe0 alone). Because this was not the case, the experiments reported in Fig. 5b were performed. 382 It can be seen that NaHCO 3 enhances MB discoloration for all tested concentrations. The 383 discoloration efficiency increased from 77% at 0.0 mM NaHCO 3 to 90% at 0.8 mM NaHCO3 and 384 remains constant for higher NaHCO 3 concentrations ( £ 4 mM). In the experiments w ith CaCl 2 385 and CuCl 2 the initial discoloration rate of 77% first decreases to 70 and 64% respectively at an 386 additive concentration of 0.2 mM and subsequently increases to about 74% and remains constant. 387 However, for 4 mM CuCl 2 the discoloration efficiency ( 73% at 2 mM) drops to 30% at 4 mM 388 while the discoloration efficiency in the presence of CaCl 2 remains constant (74%). The 389 behaviour of the system with CuCl 2 was not further investigated but suggests that if Cu 2+ is 390 quantitatively produced in a Cu/Fe bimeta llic system the reactivity of Fe 0 may be inhibited. This 391 18 issue is yet to be considered in the Fe0 technology. Similarly, the comparatively low discoloration 392 efficiency observed in the system with 0.2 mM NiCl 2 (33% against 60% for CaCl 2) should 393 question the concept of using Ni and Cu as additive metals to form nickel bimetallic systems to 394 “improve the reduction capacity of Fe 0” [28]. No such improvement could be observed in this 395 study (Fig. 5). Discussing the validity of the concept of using bimetallics to improve Fe 0 396 reactivity is over the scope of this work (see ref. [72]). 397 Another important issue from the discussion above is the importance of the cation nature in 398 chloride salts on the extent of MB removal. Generally, chloride ions are known to promote iron 399 corrosion, and therefore increase, sustain or restore Fe 0 reactivity. These observations are mostly 400 attributed to pitting iron corrosion or avoiding the formation of oxide-layers on iron [73, 74]. The 401 discussion above demonstrated clearly that the nature of the used salt should be considered in 402 comparing results from independent sources. 403 Conclusions 404 In summary, despite the low adsorptivity exhibited by MB towards Fe 0, Fe2O3 and Fe 3O4, under 405 the experimental conditions, MB was quantitatively discolored as F e0 corrosion proceeded. The 406 extent of MB discoloration was insignificant in experiments in which the availability of in situ 407 generated corrosion products was delayed (MnO 2 addition). Data from the experiments with the 408 systems “Fe0” and “MnO 2” clearly showed that the kinetics of MB adsorption and reduction by 409 MnO2 is slower than MB co -precipitation. Thus, even in systems where direct contaminant 410 reduction (electrons from Fe 0) is likely to occur, co -precipitation will interfere with (or even 411 hamper) mass transport involving Fe0. 412 The concept that methylene blue (MB) discoloration from aqueous solution in presence of 413 metallic iron is caused by MB co -precipitation with Fe 0 corrosion products is consistent with 414 many experimental observations, in particular the effects of the initial pH value and the impact of 415 19 MnO2 on MB discoloration. Generally, a queous contaminant removal in Fe 0/H2O systems can be 416 viewed as a “trickle down” in which a fraction of the targeted contaminant is continuously adsorb 417 onto in situ generat ed high reactive corrosion products [11]. Contaminants are subsequently 418 entrapped into the structure of ageing corrosion products. In this situation, no observable 419 equilibrium is attained. Therefore, the use of adsorption isotherms (e.g., Freundlich, Langm uir) to 420 interpret data from removal experiments in Fe 0/H2O systems is not justified (e.g. ref. [75]). 421 Furthermore, adsorbed or co -precipitated contaminants can be further reduced both by a direct 422 and an indirect mechanism [11.12]. The direct contaminant re duction is only possible when the 423 oxide-film on Fe 0 is electronic conductive or if so -called electron mediators are available [34, 424 76]. Noubactep [11] has clearly shown that the concept of contaminant adsorption and co -425 precipitation as fundamental removal mechanism is more accurate and considers inherent 426 mistakes of the reductive transformation concept. 427 It must be concluded that natural Fe 0/H2O systems consist of core Fe 0 and essentially amorphous 428 Fe oxides that remain to be characterized. In this regard, m any investigators have shown the 429 presence of various Fe oxyhydroxides and discussed their role in the process of contaminant 430 removal [77 -82]. Strictly, these oxyhydroxides should be considered as transient states as 431 Fe0/H2O systems are transforming systems . Therefore, a continuously reacting Fe 0/H2O system 432 can not be simply treated being at thermodynamic equilibrium. Thus, characterising the system 433 composition at certain dates is very useful but should be completed by continuously 434 characterizing the system as the contaminants are removed and/or transformed. 435 With this study, the potential of bulk reactions with selected additives for providing mechanistic 436 information [36] on aqueous contaminant removal is confirmed for the first time using an organic 437 compound. This study also demonstrates the significant impact of selected operational 438 experimental parameters (iron type, shaking intensity, solution chemistry) on the process of MB 439 20 co-precipitation in Fe 0/H2O systems. A unified experimental procedure is needed to : (i) avoid 440 further data generation under non relevant experimental conditions , and (ii) facilitate the inter -441 laboratory comparison of data. At the term such efforts will provide a confident background for a 442 non-site-specific iron barrier design [83] . Keep ing in mine the large spectrum of contaminants 443 that can be removed in Fe 0/H2O systems and the diversity of Fe 0 materials that are used by 444 individual research groups, it is obvious, that the development of such an unified experimental 445 procedure should be a concerted effort. 446 Acknowledgments 447 For providing the iron materials (Fe 0) investigated in this study the author would like to express 448 his gratitude to the branch of the MAZ (Metallaufbereitung Zwickau, Co) in Freiberg (Germany), 449 Gotthart Maier Metallpulver GmbH (Rheinfelden, Germany), Connelly GPM Inc. (USA), Dr. R. 450 Köber from the Institute Earth Science of the University of Kiel and Dr. -Ing. V. Biermann from 451 the Federal Institute for Materials Research and Testing (Berlin, Germany). Mechthild Rittmeier 452 and Emin Özden are acknowledged for technical support. The work was granted by the Deutsche 453 Forschungsgemeinschaft (DFG-No 626/2-2). 454 455 References 456 [1] A.D. Henderson, A.H. Demond, Long -term performance of zero -valent iron permeable 457 reactive barriers: a critical review. Environ. Eng. Sci. 24 (2007), 401–423. 458 [2] D.F. Laine, I.F. Cheng, The destruction of organic pollutants under mild reaction conditions: 459 A review. Microchem. J. 85 (2007), 183–193. 460 [3] S.F. O´Hannesin, R.W. Gillham, Long-term performance of an in situ "iron wall" for 461 remediation of VOCs. Ground Water 36 (1998), 164–170. 462 21 [4] M.M. Scherer, S. Richter, R.L. Valentine, P.J.J. Alvarez, Chemistry and microbiology of 463 permeable reactive barriers for in situ groundwater clean up. Rev. Environ. Sci. Technol. 30 464 (2000), 363–411. 465 [5] P.G. Tratnyek, M.M. Scherer, T.J. Johnson, L.J. Matheson, Permeable reactive barriers of 466 iron and other zero -valent metals. In Chemical Degradation Methods for Wastes and 467 Pollutants: Environmental and Industrial Applications, Tarr, M.A., Ed., Marcel Dekker: New 468 York, (2003) 371–421. 469 [6] N.A. VanStone, R.M. Focht, S.A. Mabury, B.S. Lollar, Effect of iron type on kinetics and 470 carbon isotopic enrichment of chlorinated ethylenes during abiotic reduction on Fe(0). 471 Ground Water 42 (2004), 268–276. 472 [7] S.D. Warner, D. Sorel, Ten years of permeable reactive barriers: Lessons learned and future 473 expectations. In: Chlorinated Solvent and DNAPL Remediation: Innovative Strategies for 474 Subsurface Cleanup, Henry, S.M., Warner, S.D., Eds, American Che mical Society: 475 Washington, DC, ACS Symp., Ser. 837 (2003), 36–50. 476 [8] Y. You, J. Han, P.C. Chiu, Y. Jin, Removal and inactivation of waterborne viruses using 477 zerovalent iron. Environ. Sci. Technol. 39 (2005), 9263–9269. 478 [9] B. Jafarpour, P.T. Imhoff, P.C. Chiu, Quantification and modelling of 2,4 -dinitrotoluene 479 reduction with high-purity and cast iron. J. Contam. Hydrol. 76 (2005), 87–107. 480 [10] J.A. Mielczarski, G.M. Atenas, E. Mielczarski, Role of iron surface oxidation layers in 481 decomposition of azo -dye water pollutants in weak acidic solutions. Appl. Catal. B56 (2005), 482 289–303. 483 [11] C. Noubactep, Processes of contaminant removal in “Fe 0–H2O” systems revisited . The 484 importance of co-precipitation. Open Environ. J. 1 (2007), 9–13. 485 22 [12] C. Noubactep, A critic al review on the mechanism of contaminant removal in Fe 0–H2O 486 systems. Environ. Technol. 29 (2008), 909–920. 487 [13] M. Cohen, The formation and properties of passive films on iron. Can. J. Chem. 37 (1959), 488 286–291. 489 [14] K.J. Vetter, General kinetics of passiv e layers on metals. Electrochim. Acta 16 (1971), 490 1923–1937. 491 [15] B. Gu, J. Schmitt, Z. Chen, L. Liang, J.F. McCarthy, Adsorption and desorption of natural 492 organic matter on iron oxide: mechanisms, and models. Environ. Sci. Technol. 28 (1994), 38–493 46. 494 [16] Y. Satoh, K. Kikuchi, S. Kinoshita, H. Sasaki, Potential capacity of coprecipitation of 495 dissolved organic carbon (DOC) with iron(III) precipitates. Limnology 7 (2006), 231–235. 496 [17] U. Schwertmann, F. Wagner, H. Knicker, Ferrihydrite –Humic associations magn etic 497 hyperfine interactions. Soil Sci. Soc. Am. J. 69 (2005), 1009–1015. 498 [18] W. -C. Ying, J.J. Duffy, M.E. Tucker, Removal of humic acid and toxic organic compounds 499 by iron precipitation. Environ. Progr. 7 (1988), 262–269. 500 [19] E. Tipping, The adsorption of aquatic humic substances by iron oxides. Geochim. 501 Cosmochim. Acta 45 (1981), 191–199. 502 [20] E. Tipping, Some aspects of the interactions between particulate oxides and aquatic humic 503 substances. Mar. Chem. 18 (1986), 161–169. 504 [21] R.J. Crawford, I.H. Hardi ng, D.E. Mainwaring, Adsorption and coprecipitation of single 505 heavy metal ions onto the hydrated oxides of iron and chromium. Langmuir 9 (1993), 3050 –506 3056. 507 [22] H. Füredi-Milhofer, Spontaneous precipitation from electrolytic solutions. Pure Appl. Chem. 508 53 (1981), 2041–2055. 509 23 [23] I. Nirdosh, S.V. Muthuswami, M.H.I. Baird Radium in uranium mill tailings - Some 510 observations on retention and removal. Hydrometallurgy 12 (1984), 151–176. 511 [24] B.D. Jones, J.D. Ingle, Evaluation of redox indicators for determining sulfate-reducing and 512 dechlorinating conditions. Water Res. 39 (2005), 4343–4354. 513 [25] A.A. Attia, B.S. Girgis, N.A. Fathy, Removal of methylene blue by carbons derived from 514 peach stones by H3PO4 activation: Batch and column studies. Dyes and Pigments 76 (2008), 515 282–289. 516 [26] J. Avom, J.B. Ketcha, C. Noubactep, P. Germain, Adsorption of methylene blue from an 517 aqueous solution onto activated carbons from palm-tree cobs. Carbon 35 (1997), 365–369. 518 [27] Dutta, K., Mukhopadhyay, S., Bhattacharjee, S., Chaudhuri, B., 2001. Chemical oxidation of 519 methylene blue using a Fenton-like reaction. J. Hazard. Mater. 84, 57 –71. 520 [28] L.M. Ma, Z.G. Ding, T.Y. Gao, R.F. Zhou, W.Y. Xu, J. Liu, Discoloration of methylene 521 blue and wastewater from a plant by a Fe/Cu bimetallic syst em. Chemosphere 55 (2004), 522 1207–1212. 523 [29] S. Pande, S.K. Ghosh, S. Nath, S. Praharaj, S. Jana, S. Panigrahi, S. Basu, T., Pal, Reduction 524 of methylene blue by thiocyanate: Kinetic and thermodynamic aspects. J. Colloid Interf. 525 Sci. 299 (2006), 421–427. 526 [30] K. Imamura, E. Ikeda, T. Nagayasu, T. Sakiyama, K. Nakanishi, Adsorption behavior of 527 methylene blue and its congeners on a stainless steel surface. J. Colloid Interf. Sci. 245 528 (2002), 50–57. 529 [31] E.E. Oguzie, Corrosion inhibition of mild steel in hydrochl oric acid solution by methylene 530 blue dye. Mater. Lett. 59 (2005), 1076–1079. 531 [32] J. Cao, L. Wei, Q. Huang, L. Wang, S. Han, Reducing degradation of azo dye by zero-valent 532 iron in aqueous solution. Chemosphere 38 (1999), 565–571. 533 24 [33] S. Nam, P.G. Tratnyek , Reduction of azo dyes with zero -valent iron. Wat. Res. 34 (2000), 534 1837–1845. 535 [34] E.J. Weber, Iron-mediated reductive transformations: Investigation of reaction mechanism. 536 Environ. Sci. Technol. 30 (1996), 716–719. 537 [35] F. dos Santos Coelho, J.D. Ardisso n, F.C.C. Moura, R.M. Lago, E. Murad, J.D. Fabris, 538 Potential application of highly reactive Fe(0)/Fe 3O4 composites for the reduction of Cr(VI) 539 environmental contaminants. Chemosphere 71 (2008), 90–96. 540 [36] C. Noubactep, G. Meinrath, J.B. Merkel, Investigat ing the mechanism of uranium removal 541 by zerovalent iron materials. Environ. Chem. 2 (2005), 235–242. 542 [37] D. Postma, C.A.J. Appelo, Reduction of Mn-oxides by ferrous iron in a flow system: column 543 experiment and reactive transport modelling. Geochim. Cosmoc him. Acta 64 (2000), 1237 –544 1247. 545 [38] A.F. White, M.L. Paterson, Reduction of aqueous transition metal species on the surface of 546 Fe(II)-containing oxides. Geochim. Cosmochim. Acta 60 (1996), 3799–3814. 547 [39] M.-X. Zhu, Z. Wang, L. -Y. Zhou, Oxidative decolori zation of methylene blue using 548 pelagite. J. Hazard. Mater. 150 (2008), 37–45. 549 [40] D.F.A. Koch, Kinetics of the reaction between manganese dioxide and ferrous ion. Aust. J. 550 Chem 10 (1957), 150–159. 551 [41] H. Valdes, M. Sanchez-Polo, J. Rivera -Utrilla, C.a. Zaror, Effect of ozone treatment on 552 surface properties of activated carbon. Langmuir 18 (2002), 2111–2116. 553 [42] V. Ender, Zur Struktur der Phasengrenze Metalloxid/Elektrolyt -Potentialbildung und 554 Ladungsbilanz. Acta Hydrochim. Hydrobiol. 19 (1991), 199–208. 555 [43] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Environmental applications of 556 semiconductor photocatalysis. Chem. Rev. 95 (1995), 69–96. 557 25 [44] J.E. Post, Manganese oxide minerals: Crystal structures and economic and environmental 558 significance. Proc. Natl. Acad. Sci. USA 96 (1999), 3447–3454. 559 [45] C. Noubactep, A. Schöner, G. Meinrath, Mechanism of uranium (VI) fixation by elemental 560 iron. J. Hazard. Mater. 132 (2006), 202–212. 561 [46] H. Zhang, C. -H. Huang, Oxidative transformation of triclosan and chl orophene by 562 manganese oxides. Environ. Sci. Technol. 37 (2003), 2421–2430. 563 [47] Y. Jia, P. Aagaard, G.D. Breedveld, Sorption of triazoles to soil and iron minerals. 564 Chemosphere 67 (2007), 250–258. 565 [48] T. Liu, D.C.W. Tsang, I.M.C. Lo, Chromium(VI) reductio n kinetics by zero -valent iron in 566 moderately hard water with humic acid: iron dissolution and humic acid adsorption. Environ. 567 Sci. Technol. 42 (2008), 2092–2098. 568 [49] K. Hanna, Adsorption of aromatic carboxylate compounds on the surface of synthesized iron 569 oxide-coated sands. Appl. Geochem. 22 (2007), 2045–2053. 570 [50] S. Yean, L. Cong., C.T. Yavuz, J.T. Mayo, W.W. Yu, A.T. Kan, V.L. Colvin, M.B. Tomson, 571 Effect of magnetite particle size on adsorption and desorption of arsenite and arsenate. J. 572 Mater. Res. 20 (2005), 3255–3264. 573 [51] M.I. Bautista -Toledo, J.D. Méndez -Díaz, M. Sánchez -Polo, J. Rivera -Utrilla, M.A. Ferro -574 García, Adsorption of sodium dodecylbenzenesulfonate on activated carbons: Effects of 575 solution chemistry and presence of bacteria. J. Colloid Interf. Sci. 317 (2008), 11–17. 576 [52] C. Mbudi, P. Behra, B. Merkel, The Effect of Background Electrolyte Chemistry on 577 Uranium Fixation on Scrap Metallic Iron in the Presence of Arsenic. Paper presented at the 578 Inter. Conf. Water Pollut. Natural Porous Media ( WAPO2), Barcelona (Spain) April 11 – 13 579 (2007), 8 pages. 580 26 [53] G. Meinrath, P. Spitzer, Uncertainties in determination of pH. Mikrochem. Acta 135 (2000), 581 155–168. 582 [54] R.P. Buck, S. Rondinini, A.K. Covington, F.G.K. Baucke, C.M.A. Brett, M.F. Camoes, 583 M.J.T. Milton, T. Mussini, R. Naumann, K.W. Pratt, P. Spitzer, G.S. Wilson, Measurement of 584 pH. Definition, standards, and procedures (IUPAC Recommendations 2002), Pure Appl. 585 Chem. 74 (2002), 2169–2200. 586 [55] C. Noubactep, G. Meinrath, P. Dietrich, B. Merkel, Miti gating uranium in ground water: 587 prospects and limitations. Environ. Sci. Technol. 37 (2003), 4304–4308. 588 [56] T. Tekin, M. Bayramoglu, Kinetics of the reduction of MnO 2 with Fe 2+ ions in acidic 589 solutions. Hydrometallurgy 32 (1993), 9–20. 590 [57] C. Noubactep, Investigations for the passive in -situ Immobilization of Uranium (VI) from 591 Water (in German). Dissertation, TU Bergakademie Freiberg, Wiss. Mitt. Institut für 592 Geologie der TU Bergakademie Freiberg, Band 21 (2003), 140 pp, ISSN1433-1284. 593 [58] D. Burghardt, A. Kassahun, Development of a reactive zone technology for simultaneous in 594 situ immobilisation of radium and uranium. Environ. Geol. 49 (2005), 314 –320. 595 [59] E. Lipczynska -Kochany, S. Harms, R. Milburn, G. Sprah, N. Nadarajah, Degradation of 596 carbon tetrachloride in the presence of iron and sulphur containing compounds. Chemosphere 597 29 (1994), 1477–1489. 598 [60] R. Miehr P.G. Tratnyek, Z.J. Bandstra, M.M. Scherer, J.M. Alowitz, J.E. Bylaska, Diversity 599 of contaminant reduction reactions by zerovalent iron: Role o f the reductate. Environ. Sci. 600 Technol. 38 (2004), 139–147. 601 [61] Z. Hao, X. Xu, D. Wang, Reductive denitrification of nitrate by scrap iron filings. J. 602 Zhejiang Univ. Sci. 6B (2005), 182 –187. 603 27 [62] W.S. Pereira, R.S. Freire, Azo dye degradation by recycled waste zero-valent iron powder. J. 604 Braz. Chem. Soc. 17 (2006), 832–838. 605 [63] S. Choe, Y.Y. Chang, K.Y. Hwang, J. Khim, Kinetics of reductive denitrification by 606 nanoscale zero-valent iron, Chemosphere 41 (2000), 1307–1311. 607 [64] G.W. Whitman, R.P. Russel, V.J. Altieri, Effect of hydrogen -ion concentration on the 608 submerged corrosion of steel. Indust. Eng. Chem. 16 (1924), 665 –670. 609 [65] E.R. Wilson, The Mechanism of the corrosion of iron and steel in natural waters and the 610 calculation of specific rates of corrosion. Indust. Eng. Chem. 15 (1923), 127–133. 611 [66] D. Rickard, The solubility of FeS. Geochim. Cosmochim. Acta 70 (2006), 5779–5789. 612 [67] M. Stratmann, J. Müller, The mechanism of the oxygen reduction on rust -covered metal 613 substrates. Corros. Sci. 36 (1994), 327–359. 614 [68] C.G. Schreier, M. Reinhard, Transformation of chlorinated organic compounds by iron and 615 manganese powders in buffered water and in landfill leachate. Chemosphere 29 (1994), 616 1743–1753. 617 [69] L.J. Matheson, P.G. Tratnyek, Reductive dehalogenati on of chlorinated methanes by iron 618 metal. Environ. Sci. Technol. 28 (1994), 2045–2053. 619 [70] C. Noubactep, G. Meinrath, P. Dietrich, M. Sauter, B. Merkel, Testing the suitability of 620 zerovalent iron materials for reactive walls. Environ. Chem. 2 (2005), 71–76. 621 [71] E.M. Pierce, D.M. Wellman, A.M. Lodge, E.A. Rodriguez, Experimental determination of 622 the dissolution kinetics of zero -valent iron in the presence of organic complexants. Environ. 623 Chem. 4 (2007), 260–270. 624 [72] C. Noubactep, On the operating mode of bimetallic systems for environmental remediation. 625 J. Hazard. Mater. (2008), In Press, Available online 13 August 2008. 626 28 [73] R. Hernandez, M. Zappi, C. -H. Kuo, Chloride effect on TNT degradation by zerovalent iron 627 or zinc during water treatment. Environ. Sci. Technol. 38 (2004), 5157–5163. 628 [74] J.S. Kim, P.J. Shea, J.E. Yang, J. -E. Kim, Halide salts accelerate degradation of high 629 explosives by zerovalent iron. Environ. Pollut. 147 (2007), 634–641. 630 [75] D.R. Burris, T.J. Campbell, V.S. Manoranjan, Sorption of trichloroethylene and 631 tetrachloroethylene in a batch reactive metallic iron -water system. Environ. Sci. Technol. 29 632 (1995), 2850–2855. 633 [76] P.G. Tratnyek, M.M. Scherer, B. Deng, S. Hu, Effects of natural organic matter, 634 anthropogenic surfactants, and mode l quinones on the reduction of contaminants by zero -635 valent iron. Wat. Res. 35 (2001), 4435–4443. 636 [77] P.D. Mackenzie, D.P. Horney, T.M. Sivavec, Mineral precipitation and porosity losses in 637 granular iron columns. J. Hazard. Mater. 68 (1999), 1–17. 638 [78] J.A. Mielczarski, G.M. Atenas, E. Mielczarski, Role of iron surface oxidation layers in 639 decomposition of azo -dye water pollutants in weak acidic solutions. Applied Catalysis B: 640 Environ. 56 (2005), 289–303. 641 [79] D.H. Phillips, B. Gu, D.B. Watson, Y. Roh, L. Li ang, S.Y. Lee, Performance evaluation of a 642 zerovalent iron reactive barrier: Mineralogical characteristics. Environ. Sci. Technol. 34 643 (2000), 4169–4176. 644 [80] K. Ritter, M.S. Odziemkowski, R.W. Gillham, An in situ study of the role of surface films 645 on granular iron in the permeable iron wall technology. J. Contam. Hydrol. 55 (2002), 87 –646 111. 647 [81] M.M. Scherer, B.A. Balko, P.G. Tratnyek, The role of oxides in reduction reactions at the 648 metal-water interface. In Kinetics and Mechanism of Reactions at the Minera l/Water 649 Interface, Sparks, D.; Grundl, T., Eds; (1999) pp. 301–322. 650 29 [82] M.M. Scherer, K. Johnson, J.C Westall, P.G. Tratnyek, Mass transport effects on the 651 kinetics of nitrobenzene reduction by iron metal. Environ. Sci. Technol. 35 (2001), 2804–652 2811. 653 [83] K.L. McGeough, R.M. Kalin, P. Myles, Carbon disulfide removal by zero valent iron. 654 Environ. Sci. Technol. 41 (2007), 4607–4612. 655 656 657 30 Table 1: Main characteristics, iron content and percent methylene blue (MB) discoloration (P) of 657 tested Fe0 materials. MB removal were conducted in triplicates for 36 days under non-658 disturbed conditions. The material code (“code”) are from the author, the given form is 659 as supplied; d (mm) is the diameter of the supplied material and the Fe content is given 660 in % mass. 661 Supplier(a) Supplier denotation code form d (mm) Fe (%) P (%) MAZ, mbH Sorte 69(b) ZVI0 fillings - 93(c) 75 ± 2 G. Maier GmbH FG 0000/0080 ZVI1 powder £ 80 92(d) 88 ± 2 G. Maier GmbH FG 0000/0200 ZVI2 powder £ 200 92(d) 89 ± 1 G. Maier GmbH FG 0000/0500 ZVI3 powder £ 500 92(d) 88 ± 1 G. Maier GmbH FG 0300/2000 ZVI4 fillings 200-2000 92(d) 81 ± 4 G. Maier GmbH FG 1000/3000 ZVI5 fillings 1000-3000 92(d) 77 ± 4 G. Maier GmbH FG 0350/1200 ZVI6 fillings 100-2000 92(d) 88 ± 1 Würth Hartgussstrahlmittel ZVI7 spherical 1200 n.d.(e) 66 ± 1 Hermens Hartgussgranulat ZVI8 flat 1500 n.d. 67 ± 2 G. Maier GmbH Graugussgranulat ZVI9 chips n.d. 71 ± 7 ISPAT GmbH Schwammeisen ZVI10 spherical 9000 n.d. 72 ± 6 ConnellyGPM CC-1004 ZVI11 fillings >96 76 ± 4 ConnellyGPM CC-1190 ZVI12 fillings >96 75 ± 9 ConnellyGPM CC-1200 ZVI13 powder >96 84 ± 1 31 (a) List of suppliers: MAZ (Metallaufbereitung Zwickau, Co) in Freiberg (Germany); Gotthart Maier 662 Metallpulver GmbH (Rheinfelden, Germany), ISPAT GmbH, Hamburg (Germany), Connelly GPM Inc. (USA), 663 (b)Scrapped iron material; (c) Mbudi et al. [52]; (d) average values from material supplier, (e)not determined. 664 665 32 665 Table 2: Characteristics, surface coverage and function of the individual reactive materials of this 666 study. Apart from Fe0 the given value of specific surface area (SSA) for are the minima 667 of reported data. Apart from Fe 2O3 the pH at the point of zero charge (pH pzc) is lower 668 than the initial pH value. Therefore, MB adsorption onto the negatively charged 669 surfaces is favorable. The surface coverage is estimated using the method presented by 670 Jia et al. [47]. The total surface that can be covered by the amount of MB present in 22 671 mL of a 0.063 mM is SMB = 0.997 m 2. 672 673 System pHpzc SSA Savailable Coverage Function (m2 g-1) (m2) (1) Fe0 7.6a 0.29 0.032 31.3 MB reductant? Fe0 + MnO2 - - 4.432 0.2 - MnO2 2.0 - 6.0 b 40 4.4 0.2 delays CP availability Fe2O3 7.5 - 8.8 c 60 6.6 0.2 mimics aged CP Fe3O4 6.8 d 40 4.4 0.2 mimics aged CP GAC 7.0 - 8.0e 200 22 0.1 MB adsorbent Fe0 + GAC - (-) 22.032 0.1 - aref. [48], bref [39], cref. [49], dref. [50], e ref. [51]. 674 675 676 677 33 Figure 1 677 678 0 8 16 24 32 40 48 0 20 40 60 80 100 reference Fe 2 O 3 Fe3O4 GAC MnO 2 (nat) MnO 2 (com) Fe0 M B r em ov al / [% ] elapsed time / [days] 679 680 34 Figure 2 680 Fe0 + AC Fe0 AC Fe0 + MnO2 MnO2 0 20 40 60 80 100 (a) [MB]0 = 19.5 mg L -1 t = 36 days [Fe0] = [additive] = 5 g L-1 V = 22 mL M B r em ov al / [% ] added material 681 682 0 2 4 6 8 10 20 40 60 80 100 (b) Exp. conditions: [MB] 0 = 20 mg L -1 t = 35 days [Fe 0] = 5 g L -1 Fe0 + MnO 2 Fe0 + GAC M B r em ov al / [% ] additive loading / [g/L] 683 684 685 35 Figure 3 685 686 0 50 100 150 200 0 20 40 60 80 100 [Fe0] = 10 g L-1 [MB] 0 = 20 mg L-1 t = 24 h essay tube (22 mL) erlenmeyer (100 mL) M B r em ov al / [% ] shaking intensity / [min-1] 687 688 689 690 36 Figure 4 690 691 2 4 6 8 10 0 15 30 45 60 75 90 8.5 9.6 9.9 4.3 10.3 4.3 2.6 [Fe0] = 10 g/L [MB] 0 = 20 mg/L shaken at 100 min-1 24 h 48 h M B r em ov al / [% ] initial pH 692 693 37 Figure 5 693 EDTA NiCl2 Al(NO3)3 BaCl2 CuCl2 ZnSO4 CaCl2 Ref. (NH4)2CO3 -- 20 40 60 80 100 (a)[additive]0 = 0.2 mM [Fe0] 0 = 5 g L-1 [MB] 0 = 20 mg L-1 M B r em ov al / [% ] additive 694 695 0 1 2 3 4 30 45 60 75 90 (b) [Fe0] 0 = 5 g L -1 [MB] 0 = 20 mg L-1 CuCl 2 CaCl2 NaHCO 3 M B r em ov al / [% ] additive / [mM] 696 697 38 Figure Captions 697 Figure 1: 698 Methylene blue removal (%) as a function of equilibration time for the six tested reactive 699 materials. The reference system is a blank experiment without additives. Two sets of experiments 700 with MnO 2 were conducted (see the text). The experiments were conducted in triplicate. Error 701 bars give standard deviations. The lines are not fitting functions, they simply connect points to 702 facilitate visualization. 703 Figure 2: 704 Methylene blue (MB) discoloration by metallic iron (Fe 0), granular activated carbon (AC), 705 manganese nodule (MnO 2), and the mixtures “Fe 0 + AC” and “Fe 0 + MnO 2”. (a) extent of MB 706 discoloration after 36 days, and (b) dependence of the MB di scoloration on the additive loading 707 for 35 days. The experiments were conducted in triplicate. Error bars give standard deviations. 708 The lines are given to facilitate visualization. 709 Figure 3: 710 Effect of the mixing intensity (min -1) on discoloration of MB at initial pH 7.8. The system is 711 mixed on a rotary shaker. The experiments were conducted in triplicate. Error bars give standard 712 deviations. The lines simply connect points to facilitate visualization. 713 Figure 4: 714 Effect of initial pH on discoloration of MB by Fe0 for 24 and 48 h respectively. The experiments 715 were conducted in triplicate. The reported numbers on the plots are the corresponding final pH 716 values. Error bars give standard deviations. The lines simply connect points to facilitate 717 visualization. 718 Figure 5: 719 39 Effect of solution chemistry on MB discoloration by metallic iron (Fe 0): (a) extent of MB 720 discoloration after 36 days for all tested additives, and (b) dependence of the MB discoloration on 721 selected additive concentrations for 35 days. Ref. in figure “a” refers to the experiment in tap 722 water (“no additive”). The experiments were conducted in triplicate. Error bars give standard 723 deviations. The lines simply connect points to facilitate visualization. 724