Physicochemical, crystalline and morphological characteristics of bricks used for ground waters purification in Bangui region (Central African Republic)

Applied Clay Science 59–60 (2012) 69–75

Contents lists available at SciVerse ScienceDirect

Applied Clay Science journal homepage: www.elsevier.com/locate/clay

Research paper

Physicochemical, crystalline and morphological characteristics of bricks used for ground waters purification in Bangui region (Central African Republic) S.C. Dehou a, M. Wartel b, P. Recourt c, B. Revel d, J. Mabingui a, A. Montiel e, A. Boughriet f,⁎ a

Chaire Unesco “sur la gestion de l'eau”, Laboratoire Hydrosciences Lavoisier, Université de Bangui, Faculté des Sciences, B.P. 908, République Centrafricaine Université de Lille1 Laboratoire Géosystèmes Equipe Chimie Analytique et Marine, UMR 8217 CNRS, BâtC8, 59655 Villeneuve d'Ascq cedex, France Laboratoire Géosystèmes (UMR 8217, Lille1/CNRS), UFR des Sciences de la Terre-bâtiment SN5, 59655 Villeneuve d'Ascq cedex, France d Service RMN (Résonnance Magnétique Nucléaire), Bld Langevin, Bâtiment C4, Université des Sciences et Technologies de Lille1, 59655 Villeneuve d'Ascq cedex, France e Société Anonyme de Gestion des Eaux de Paris, 9 rue Schoelcher, 75675 Paris Cedex 14, France f Université Lille Nord de France, I.U.T de Béthune Département de Chimie, Rue de l'Université, B.P. 819, 62408 Béthune Cedex, France b c

a r t i c l e

i n f o

Article history: Received 9 March 2011 Received in revised form 3 February 2012 Accepted 8 February 2012 Available online 28 March 2012 Keywords: Brick Kaolinite Metakaolinite Adsorption Iron oxides Water treatment

a b s t r a c t The development and improvement of natural water and wastewater purification technologies utilizing low-cost raw materials like bricks are necessary to make possible easy application in poor countries. The present work concerns the detailed studies on a soil which is commonly used to make bricks by craftsmen in Bangui region (Central African Republic). The chemical and mineralogical composition of this soil before and after thermal transformation, and its crystalline, morphological and surface properties were determined by combining several techniques: X-ray Diffraction (XRD), ThermoGravimetric Analyses (TGA), Differential Thermal (DTA) Analyses, TGA/Mass Spectrometry (MS), 27Al and 29Si NMR spectroscopy, and Environmental Scanning Electron Microscopy (ESEM; an apparatus equipped with an Energy Dispersive X-ray Spectrometer, EDS). The basic brick making led to an interesting mesoporous material that was found to be a good adsorbent for Fe(II) removal from contaminated natural waters. Column experiments further revealed significant improvement of the sorption capacity of this brick when its surface was coated with iron oxy-hydroxide. ESEM/EDS micro-analyses revealed that FeOOH was preferentially deposited onto brick clays (mainly disordered metakaolinite), thus showing the key role played by these minerals in the study water treatment when compared to the sorption performances obtained with raw sand and FeOOH-coated sand. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Numerous water-treatment processes had been used in the past to remove metallic pollutants from ground- and surface-waters. Among them, adsorption methods are nowadays currently considered as economic, efficient and low-cost technologies for water purification in rural and urban areas lacking centralized water supplies. For that purpose, various cheap materials like: metal oxides (Chan et al., 2009; Chang et al., 2010; Litter et al., 2010; Maliyekkal et al., 2009; Rodrigues et al., 2010; Wu and Zhou, 2009); zeolites (Motsi et al., 2009; Qiu and Zheng, 2009; Rios et al., 2008; Wang and Peng, 2010; Wang et al., 2009); iron oxide-coated sand (Boujelben et al., 2009); and clay minerals (Jiang et al., 2010; Novakovic et al., 2008; Oubagaranadin and Muthy, 2010; Oubagaranadin et al., 2010; Ouhadi et al., 2006; Stathi et al., 2007; Vieira et al., 2010), had recently been tested with the aim to demonstrate their capacity of metal removal ⁎ Corresponding author at: Université de Lille1 Laboratoire Géosystèmes, Equipe Chimie Analytique et Marine, FRE CNRS 3298, BâtC8, 59655 Villeneuve d'Ascq cedex, France. Fax: + 33 3 20 43 48 22. E-mail address: [email protected] (A. Boughriet). 0169-1317/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.clay.2012.02.009

from contaminated waters. But, the solution of specific water problems encountered in poor/rural communities of developing countries had further necessitated the elaboration of proven and locally more adapted water-treatment procedures at low costs. In the last few years, the ability of crushed bricks used as a cheap material to remove soluble toxic metal contaminants from wastewaters had been evidenced (Aziz et al., 2008; Djeribi and Hamdaoui, 2008). Particularly, the sorption characteristics of brick were found to be better than those of sand (Arias et al., 2006; Boujelben et al., 2009; Han et al., 2006a,b; Selvaraju and Pushpavannam, 2009) and to increase with the quantity of oxides and/ or hydroxides of iron, aluminum and manganese associated with the material (Boujelben et al., 2008). Furthermore, the adsorption performance of crushed bricks for a wide variety of salts was found to be comparable with that of activated charcoal (Selvaraju and Pushpavannam, 2009; Yadav et al., 2006). Our preliminary works on crushed bricks showed excellent data about their capacity to remove dissolved iron(II) from ground waters which were sampled in Bangui-region wells in Central African Republic. Indeed, these ground waters do necessitate a purification treatment before any domestic uses because they are strongly polluted by dissolved iron at concentration levels reaching up to 10 mg per liter; these waters when exposed to air oxygen, turn rapidly cloudy and in

70

S.C. Dehou et al. / Applied Clay Science 59–60 (2012) 69–75

reddish brown color. This chemical phenomenon reveals the formation/ precipitation of colloidal iron oxides/hydroxides on which bacteria can be trapped, proliferate, and hence, can cause health problems among local population. This explains why Bangui ground waters become organoleptic, undrinkable, and in addition inappropriate for washing and doing the laundry. To enhance the efficiency of water purification processes with brick pellets, this adsorbent must be improved and modified in its chemical nature, crystalline structures, surface properties and preparation processes. Before undertaking this complex task involving the consideration of many physicochemical parameters, we have preliminary focused our attention on the determination of the chemical composition and crystalline/morphological characteristics of the crude brick, as well as on the investigations of the different chemical and crystalline transformations of the starting material (which is a clays-rich soil extracted at ≥0.2 m below ground) at oven temperatures commonly used by local craftsmen living in Bangui region to make bricks. Another objective of the present work has been: (a) to investigate the performances of this brick for Fe(II) removal from synthetic aqueous solutions before and after its coating with iron oxy-hydroxide by carrying out column experiments, and to compare data to those obtained with columns filled with Bangui sand; and (b) to better understand the variable sorption capacity of untreated and treated bricks by means of the BET method in order to reach specific surface area, specific volume and pore size distribution, and using the ESEM/EDS technique on micro-specimens in order to characterize chemical composition, atomic distribution, morphology and surface textures . 2. Experimental All chemicals [Fe(NO3)3, 9H2O, Fe(SO4)2(NH4)2, 6H2O, HCl, NaOH, HNO3] used for the chemical treatments of sorbent materials and for sorption experiments were of analytical quality. All solutions were prepared using Milli-Q water. Bricks were made by African craftsmen and used for construction activity by local people in Bangui region (Central African Republic). Bricks makers extracted starting material directly near their homes at ≥0.2 m below ground. Briefly, extracted soils were mixed with water and the obtained mud was shaped with molds; Resulting airdried (48 h) bricks were placed in efficient stackings with air flows in order to constitute a basic oven – that was built simply on ground – heat treated with fired dry wood for a period of about 3 days at temperatures ranging from 500 to 900 °C, and finally cooled progressively up to ambient temperature for 2/3 days. In order to increase the surface area of the brick material, the latter was manually broken in grains by using a hammer. Brick particles were afterwards sieved with mechanical sieves and the fraction containing particles sizes that varied from 0.7 to 1.0 mm, was kept for our experiments. This fraction was washed with Milli-Q water and afterwards decanted; after settling, water was eliminated and brick grains were dried at 105 °C. The dried solid particles thus obtained were ready for the following chemical treatments: (i) 1-M HCl leaching for 1 day; (ii) a deposition of FeOOH onto brick grains surfaces through the precipitation of a 0.25-M ferric nitrate solution in the presence of a 6-M NaOH solution, which was followed by the addition of a 1-M NaOH solution in order to adjust the pH between 6 and 7; and (iii) finally, the resulting pellets were gently washed several times with Milli-Q water in order to eliminate the excess of FeOOH not attached to brick grains before using them for the purification treatment of iron(II) contaminated ground waters . The alkaline fusion AFNOR procedure (AFNOR, 2004) was used in this work for carrying out the total attack of both soil and (raw and iron oxy-hydroxide-coated) brick samples and for determining the total contents of elements. Briefly, 200 mg of ground solids were put inside a platinum crucible, heated progressively up to 450 °C for 1 h, and maintained at this temperature for 3 h. After that, the

crucible was cooled at room temperature, and 200 mg of lithium tetraborate (Li2B4O7) and 800 mg of lithium metaborate (LiBO2) were both loaded in this crucible and heated at 1000–1100 °C in order to obtain a total dissolution of the mixture in a time frame of 10 min. After cooling, the resulting residue was dissolved with 200 mL of a 0.5 M nitric acid solution. The concentration of element ions in the resulting solution was determined by Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AES; model: Varian Vista Pro axial view). The physical characteristics (specific surface area, specific volume and pore size distribution) of raw brick and starting material which was used to make bricks, were determined by the nitrogen adsorption isotherm (BET analysis) using Sorptomatic 1990 Carlo Erba at 77°K. Micrographies of brick surfaces were obtained by using an environmental scanning electron microscope model: ESEM, QUANTA 200 FEI. Elemental analyses of solid samples were performed using the ESEM/EDS technique (ESEM, QUANTA-200-FEI, equipped with an Energy Dispersive X-ray Spectrometer EDS X flash 3001, QUANTAX 400, ROENTEC/BRUKER). For ESEM analyses, samples were previously coated with thin carbon film in order to avoid the influence of charge effect during the ESEM operation; EDS measurements were carried out at 20 kV at high vacuum (10 − 6 Torr), and the maximum pulse throughput was 20 kcps. X-ray Diffraction (XRD) analysis was performed on brick and soil grains before and after their chemical/thermal treatments using a Bruker Endeavor D4 diffractometer at a scan speed of 1°2θ in 2 min and a step size of 0.02° using a Ni-filtered Cu-Kα radiation. The ≤2 μm-clays fractions – which were extracted from broken brick particles, as described above – were prepared as smears oriented on glass slides. The semi-quantitative values were deduced from XRD peak heights and surfaces according to Holtzapffel (1985). Thermo-Gravimetric Analysis (TGA) and Differential Thermal Analysis (DTA) of the starting material (which was a soil originated from Bangui region) were performed using a TGA-DTA apparatus: model Setaram SETSYS Evolution. TGA-MS analyses were also undertaken using a SETARAM apparatus model: TGA-92 coupled to a PFEIFFER Mass Spectrometer (MS). TGA and DTA analyses permitted us to detect, follow, and identify the different thermo-chemical transformations/degradations of the study sample in response to a progressive temperature increase ranging from RT to 1200 °C. For that, 27.6 mg of soil powder were loaded in a platinum crucible and heated at a heating rate of 5 °C min − 1 and an air flow of 75 mL min − 1. 27 Al solid-state magic angle spinning NMR spectra were recorded at 208.4 MHz on a Bruker Avance 800 (18.8 T) multinuclear spectrometer equipped with a 3.2 mm probe. The spinning rate of the 3.2-mm zirconia rotor was 24 kHz. Single-pulse sequences were applied with a 1 μs pulse time (π/10 flip angle), the number of scans was 1024, and the recycle delay was fixed to 2 s. The 27Al chemical shifts were referenced to Al(H2O)63 +. 29Si solid-state magic angle spinning NMR spectra were recorded at 79.4 MHz on a Bruker Avance 400 (9.4 T) spectrometer equipped with a 7 mm probe. The spinning rate of the 7-mm zirconia rotor was 5 kHz. Single-pulse sequences were applied with a 5 μs pulse time (π/2), the number of scans ranged from 1920 to 8190, and the recycle delay was fixed to 30 s. The 29Si chemical shifts were referenced to tetramethylsilane (TMS). 2.1. Column tests Different absorbents (uncoated and iron oxy-hydroxide-coated sand and bricks from Bangui region) were packed into a column that was 2 cm in inner diameter and 20 cm in height, and had a medium – porosity sintered – pyrex disk at its bottom in order to prevent any loss of material. Before being used in the experiments, about 5 bed volumes of Milli-Q water were passed through the column: (a) in order to remove any unbound and thin particles/iron oxide(s)/hydroxide(s), (b) in order to

S.C. Dehou et al. / Applied Clay Science 59–60 (2012) 69–75

71

A

Temperature (°C) 0

0.00

3. Results 3.1. Chemical and mineralogical composition of soil and brick from Bangui region

200

400

600

800

1000

1200

1400

Exo.

TGA mass loss (%)

0

1.45

-5

TGA

-5

2.90

DTA

-5

DTA

4.35

Heat flux (µV)

check the absence of soluble iron in the effluent by ICP-AES, and (c) in order to confirm the stability of the FeOOH coating on brick grains. Either synthetic Fe(II) solutions ([Fe2 +]=10 mg/L) or Bangui ground waters samples ([Fe2 +] =2–10 mg/L) were passed through the column at a flow rate of about 5 mL min− 1. Water samples were collected at various time intervals, and the concentration of soluble iron in the effluent was subsequently analyzed using ICP-AES.

-20 5.80

-25

B

Temperature (°C)

0.00

0

200

400

600

800

1000

1200

8.00E-11

1400

7.00E-11

TGA mass loss (%)

1.45

SiO2 > Al2 O3 > Fe2 O3 > K2 O≅TiO2 ≫CaO > MgO≅Na2 O≅MnO:

6.00E-11 5.00E-11

TGA

2.90

4.00E-11 MS(H 2O)

3.00E-11

4.35

2.00E-11 1.00E-11

5.80 MS(CO2)

MS mass loss of CO2/ H2O (mg)

3.1.1. Starting material Soils from the Bangui region in Central African Republic are nowadays widely exploited by the local population for making house bricks. Alkaline fusion was performed on Bangui-soil samples as described in the experimental section (AFNOR, 2004), and the resulting salts were subsequently dissolved in nitric acid Milli-Q water and analyzed by ICP-AES. The results showed the predominance of the elements combined to oxygen in the following order (see Table 1):

The XRD diffractogram of an untreated (raw) soil sample (figure not shown here) revealed the existence of several reflections assigned mainly to: quartz (SiO2), kaolinite [Al2Si2O5(OH)4] and hematite (α-Fe2O3). Also, when a soil sample was decarbonated and the ≤ 2 μm clays-rich fraction was extracted, the XRD analysis of the recovered particles indicated clearly the presence of illite in addition to kaolinite as the predominant clay minerals. Note that, when heating the sample at 490 °C for 2 h, the XRD bands ascribed to kaolinite disappeared because of the dehydroxylation (calcination) of kaolinite to produce metakaolinite (Al2Si2O7), as evidenced below by the combined use the TGA and DTA techniques. Thermo-gravimetric analysis (TGA) and differential thermal analysis (DTA) were performed on a Bangui-soil sample, see Fig. 1. There was initially a weak mass drop below 215 °C corresponding to physisorbed and interlayer water molecules which were loosely bound to clays present in the study soil (Chen et al., 2010; Kakali et al., 2001; Panda et al., 2010; Ptáček et al., 2010a,b; Tãmăşan et al., 2010; Temuujin et al., 1998). There was also another “badly-defined” mass loss step which was observed from 210 °C to 385 °C (see Fig. 1), and which was partly attributed to the removal of strongly-bonded water molecules present in the first coordination sphere of the interlayer ions (Panda et al., 2010). Note that: (a) on the course of this second mass-loss step, the thermal decomposition of organic matter in the soil further took place at temperatures ≥300 °C, and (b) the DTA curve showed the appearance of a weak exothermic peak which overlaid the presumed endothermic one corresponding to H2O strongly embedded inside clay sheets (Chen et al., 2010; Ptáček et al., 2010a,b). Finally, at higher temperatures an endothermic peak

TGA

0.00E-11

Fig. 1. (A) Thermogravimetric analysis (TGA), differential thermal analysis (DTA), and (B) Mass spectrometric (MS) analysis of a Bangui-soil sample.

appeared at ~450 °C in the DTA plot of the soil sample (Fig. 1), and was ascribed to the kaolinite dehydroxylation, as observed previously (Ptáček et al., 2010a,b; Tãmăşan et al., 2010): hydroxyl groups in clays condensed and dehydrated, leading to the formation of a disordered metakaolinite phase (e.g., Chen et al., 2010; Feng et al., 2009; Panda et al., 2010; Ptáček et al., 2010a,b). Illite was also subject to a slight H2O mass drop commencing at 450 °C up to 600 °C and corresponding to a dehydroxylation process similar to that observed for kaolinite (Carroll et al., 2005). This phenomenon was detected in the DTA curve (an overlapping observed from 400 to 600 °C; see Fig. 1A) as well as in the MS spectrum (a weak trail behind the H2O peak that was ascribed primarily to the loss of kaolinite water; see Fig. 1B). These interpretations were sustained by MS analyses of H2O vapor generated during thermal soil decomposition, showing the appearance of two MS peaks during the course of the 1st step and the 2nd step of the TGA curve (see curve MS(H2O) in Fig. 1B).

Table 1 Chemical compositions of Bangui soil, raw brick and FeOOH-coated brick analyzed by ICP-AES. For each of these materials, “n” represented the number of samples analyzed and the data obtained were used for assessing the absolute error mentioned in wt.% values. Materials

Bangui soil Raw brick 1-M HCl leached brick FeOOH coated brick DL: detection limit

SiO2

Al2O3

Fe2O3

MnO

MgO

CaO

Na2O

K2O

TiO2

Loss of ignition Total(Σ) n

%

%

%

%

%

%

%

%

%

%

%

74.51 ± 1.42 67.33 ± 2.88 69.00 ± 3.54 68.97 ± 2.02

12.84 ± 1.58 19.31 ± 1.66 18.06 ± 0.91 17.39 ± 1.18

2.49 ± 0.22 6.30 ± 1.23 5.61 ± 0.93 6.92 ± 1.39

0.11 ± 0.03 0.04 ± 0.01 0.03 ± 0.02 0.03 ± 0.01

0.28 ± 0.05 0.43 ± 0.02 0.33 ± 0.01 0.21 ± 0.05

0.40 ± 0.13 0.13 ± 0.02 b DL b DL

0.17 ± 0.02 b DL b DL b DL

1.60 ± 0.14 1.01 ± 0.31 1.05 ± 0.35 0.73 ± 0.25

1.47 ± 0.10 1.74 ± 0.40 1.63 ± 0.31 1.49 ± 0.24

6.40 ± 1.07 5.60 ± 0.38 4.60 ± 1.64 4.85 ± 0.91

100.50 101.30 100.30 100.6

4 4 6 5

72

S.C. Dehou et al. / Applied Clay Science 59–60 (2012) 69–75

From the global losses of water molecules measured for kaolinite and illite, the weight percentages of these minerals in the study sample were assessed: for kaolinite (22–23 wt.%) and for illite (≈3.9 wt.%), as well as for metakaolinite (20–21 wt.%). Note that a CO2 peak appearing in the MS spectrum of Fig. 1B was ascribed to the thermal decomposition of carbonates. 3.1.2. Brick By applying the French guideline concerning the alkaline fusion procedure to our samples (AFNOR, 2004), chemical compositions of crushed bricks before and after different chemical treatments were reported in Table 1. As a whole, the results obtained for raw brick showed the predominance of SiO2 (~67 wt.%), Al2O3 (~19 wt.%), and Fe2O3 (~6 wt.%), and to a lesser extent, TiO2 (~ 1.5 wt.%), and K2O (~1.5 wt.%). After leaching brick pellets with a 1 M HCl solution at ambient temperature for 1 day and after the precipitation of ferric hydroxides on the surface of brick grains (according to the procedure described in the experimental part), the samples' analyses revealed a slight increase of Fe2O3 with total amounts reaching up to 7 wt.% (in comparison with ~5–6 wt.% in the initial product). The 29Si NMR spectra of crushed raw brick and its ≤2 μm claysrich fraction (Fig. 2) displayed: (a) both a weak and thin resonance signal at − 107 ppm assigned to silicon atoms in quartz grains (which were not thoroughly eliminated during sieving operations), and (b) a broad 29Si peak at −90 ppm attributed to 29Si atoms in

A

27Al

NMR at 18.8T 4.6

65.8 28.0 4.9

65.8

RT (brick)

28.9 500°C (soil)

5.5

71.7 RT (soil)

80

60

B

40 29Si

20

0

[ppm]

3.2. Column experiments performed on sand and brick from Bangui region

NMR at 9.4T -92.5

-107.4 RT (brick)

-91.6

-107.5 500°C (soil)

-92.4

-86.1 -107.4

-70

-80

-90

-100

Q3 “sheet-like” layers (Zibouche et al., 2009). Note that this Q3 resonance was affected: (a) by the presence of illite in our samples (Carroll et al., 2005; Zibouche et al., 2009), and (b) by angular changes in Si\O\Si(Al) bonds during kaolinite dehydroxylation that resulted in the formation of metakaolinite (Rios et al., 2009). The 27Al NMR spectra of crushed brick and its ≤2 μm clays-rich fraction (see Fig. 2) displayed three resonances with maxima at 4.7 ppm (assigned to 6-coordinated Al), around 30 ppm (ascribed to 5coordinated Al) and around 60 ppm (attributed to 4-coordinated Al). On the other hand, the 27Al spectrum of the ≤2 μm fraction extracted from a Bangui soil sample showed an intense peak at 6 ppm ascribed to 6-coordinated 27Al atoms in the kaolinite structure (Rios et al., 2009; Zibouche et al., 2009) and a weaker peak at 71 ppm assigned partly to 4-coordinated 27Al atoms in the illite structure (Carroll et al., 2005); also, a thermal treatment of this fraction at 500 °C led to the appearance of metakaolinite 27Al peaks (at 60 ppm, 30 ppm and 4.7 ppm) in addition to the 27Al resonance at ~70 ppm that was attributed, at least partially, to a 27Al peak in illite, because this latter mineral was still thermally stable at the temperature used, as suggested by Carroll et al. (2005). Metakaolinite was found to be a structurally complex amorphous material (Bellotto et al., 1995; Dong et al., 2009; Gualtieri and Bellotto, 1998) with alumina polyhedron sheets (with stacking of its hexagonal layers). This particular structure conferred to metakaolinite specific area/porosity/adsorption characteristics which were sometimes compared with those observed for γ-Al2O3 (Dong et al., 2009; Xu et al., 2009). Fig. 3 shows a typical ESEM micrograph of 1 M HCl-leached brick. EDS analysis on microspecimen surfaces allowed differentiating quartz grains (with high silicon EDS peaks and the absence of aluminum EDS peaks), clay agglomerates (with high Al and Si EDS peaks), and hematite (with high iron EDS peaks), and even minerals at low contents: mica and feldspar, as shown in Fig. 3. Two conclusions can be derived from the above analysis. Firstly the absence of kaolinite in Bangui brick indicates that minimum temperatures in ovens used by local craftsmen were >500 °C. Secondly, illite was always detected in this material, suggesting that during the brick-making process maximum oven temperatures were always b900 °C because an irreversible structural breakdown of illite would have taken place if bricks were treated above this temperature (Carroll et al., 2005). This temperature limit (b900 °C) also explained the absence of mullite in the brick, for the transformation of metakaolinite into mullite necessitated higher oven temperatures ≥1050 °C (Castelein et al., 2001; Chakraborty, 2003; and Chen et al., 2004).

-110

RT (soil)

-120 [ppm]

Fig. 2. Typical 27Al and 29Si NMR spectra of Bangui soil (before and after its thermal treatment at 500 °C for 2 h) and brick made by African craftsmen living in Bangui region (Central African Republic).

Fig. 4 shows the adsorption of iron(II) onto different adsorbents in a column (as described in the experimental part). The influent with a Fe(II) concentration of 10 mg/L was continuously passed through the column at a flow rate of ~ 5 mL min − 1 and at a pH range 5.5–5.9. As can be seen from curves (a) and (b) of Fig. 4, both uncoated and FeOOH-coated Bangui sand removed barely 40% of Fe 2 + ions from the synthetic contaminated water at the beginning of the experiments, and the quantity of soluble iron in the effluent decreased rapidly and reached Fe(II) concentration values close to those initially measured in the influent. As for raw Bangui brick, the iron(II) removal process by this adsorbent was found to be much better (attaining up to 80% removal of soluble iron; see Fig. 4c) than that observed with raw sand as well as with modified sand. This higher sorption capacity of raw brick was attributed to the presence of hematite (α-Fe2O3) in this material, as detected by ESEM/EDS (see Fig. 3). Indeed, as mentioned in the literature (Catalana et al., 2008; Chen and Li, 2010; Rovira et al., 2008; Shuibo et al., 2009; Yin and Ellis, 2009), hematite is known to have a large specific surface area and further high adsorption properties for metal ions in the pH range of most of natural

S.C. Dehou et al. / Applied Clay Science 59–60 (2012) 69–75

73

waters. However, as can be seen in Fig. 4(d) the sorption capacity of Bangui brick was significantly improved when this material was previously coated with iron oxy-hydroxide; thus, 100% removal of iron(II) was obtained at the start of the experiment, and remained at this level up to a 800 mL-volume of influent passed through the column. Under these conditions, Fe 2 + ions were adsorbed on grains surfaces at levels attaining up to 2 mg per gram of brick. Accordingly, our column experiments demonstrated unambiguously that in modified Bangui brick iron oxy-hydroxide is the key adsorbent involved in the mechanism of Fe(II) adsorption, whereas hematite (α-Fe2O3) plays a marginal role in this process.

3.3. Physicochemical properties of treated brick pellets In order to gain more information about the chemical, textural and morphological characteristics of iron oxy-hydroxide-coated brick pellets at their surfaces, ESEM/EDS micro-analyses were performed and the detailed examination of various micro-specimens hazardously chosen was undertaken. Particularly, EDS investigations revealed an interesting feature about iron distribution on brick surfaces: Fe atoms were spread preferentially over the surfaces of clay aggregates rather than over those of quartz grains. To support this, a detailed micro-observation within a cross-section of a FeOOH-coated brick sample was performed by the ESEM/EDS technique; Fig. 5 represents the line scannings for Al, Fe and Si showing principally that the intensity of iron increases significantly on the surface of Al-rich clay aggregates but decreases quickly on sand grains. It was clearly observed that the highest intensities of iron were detected on Al and Si-rich specimens, and more particularly on surface areas where Al/Si ratios were found to be the most elevated ones and further close to 1. These ratios corresponded well to the atomic stoichiometry of aluminum and silicon in the clay minerals present predominantly in the studied brick, i.e.: metakaolinite and illite (see Fig. 5). In contrast, Fe contents were found to be lowest in poorer Al grains/aggregates and/or quartz-rich particles. The pore size distribution for 1 M leached-brick pellets showed that the average diameter of pores was 144 Å (±6 Å), and in addition two types of pores predominated in this material with diameters equal to 40 Å and 270 Å. As for FeOOH-coated bricks, their pore size distribution revealed two features: (i) the average diameter of pores (115 ± 4 Å) was found to be lower than that measured on the uncoated brick (144 Å ± 6 Å); and (ii) the number of pores with a diameter of 40 Å increased significantly, whereas those with a diameter of 270 Å decreased slightly. Consequently, these findings seemed to show that the Fe(II)-adsorption mechanism implicating iron oxy-hydroxide should take place preferentially onto pores with the largest sizes, especially, onto 270 Å-size pores. Overall, our investigations confirmed the importance of porous characteristics of brick on the FeOOH-coating procedure, that could be explained only by the presence of clays in this material. Indeed, clays are known to have better adsorption capacities than those observed with sands, as pointed out recently (Boujelben et al., 2008; Selvaraju and Pushpavannam, 2009). As a result, we did believe that these surface area and pore-size distribution properties were intimately related mainly to the presence of disordered metakaolinite, Al2SiO7 (Li et al., 2007; Bellotto et al., 1995; Dong et al., 2009; Gualtieri and Bellotto, 1998; Gualtieri et al., 1995). Furthermore, it was found that the roughness feature of iron oxide/hydroxide-coated material contributed to a slight increase of the specific surface area (~2–3 m2/g) for brick pellets with 0.7–1.0 mm sizes, and a strong increase (up to ~120 m2/g) for brick powder with an average size of ≅50 μm, as already mentioned for other iron-coated materials (e.g., Boujelben et al., 2008).

Fig. 3. A typical ESEM micrograph of brick which was produced in Bangui region (Central African Republic), and EDS analysis of detected micro-specimens.

Pourcentage of adsorbed iron (%)

74

S.C. Dehou et al. / Applied Clay Science 59–60 (2012) 69–75

120 100

(d)

80 60

(c)

40

(b)

20

(a)

0 0

200

400

600

800

1000

1200

1400

1600

Volume of FeII solution passing through the column (mL) Fig. 4. Evolution of the percentage of iron(II) adsorbed onto different adsorbents as a function of the volume of a 10 mg L−1 FeII solution passed through a column (see the Experimental part) at a flow rate of ~5 mL min−1: (a) Bangui sand before and (b) after its FeOOH coating; Bangui brick before (c) and (d) after its FeOOH coating.

each one less than 1%). The thermal decomposition of this soil at 400– 600 °C was mainly characterized by the dehydroxylation of kaolinite, leading to metakaolinite with an amorphous structure. On the other hand, the present work investigated the convenience of Bangui brick as an adsorbent for the removal of iron(II) from aqueous solutions by carrying out column experiments, and results were compared to those obtained with Bangui sand. It was shown unambiguously that the sorption capacity of this brick was largely better than that of raw sand and even of FeOOH-coated sand, in part because of the large specific surface area and high adsorption properties of hematite present in the brick for metal ions in the pH range of most of natural waters. Column tests further revealed that when brick pellets were coated with iron oxy-hydroxide, their sorption performances improved more significantly, achieving up to 100% removal of soluble iron. ESEM/EDS studies permitted us to show that FeOOH was preferentially deposited onto brick-clays surfaces and the roughness feature of these coatings should further contribute to increase specific surface area. Acknowledgments

4. Conclusion The chemical and mineralogical composition of a clays-rich soil – which is commonly used by local craftsmen in Bangui region (Central African Republic) to fabricate bricks – was: quartz (58 ± 3%), kaolinite (24 ± 4%), illite (3 ± 1%), hematite (3 ± 1%), and to a less extent mica and feldspar (the weight percentage of mica and felspar represented

This work is partly funded by the “Agence de l'Eau Artois-Picardie” and the “Region Nord Pas-de-Calais”. This study is part of the firstauthor (St C. Dehou) Ph.D. thesis, and results from the cooperation between the University of Lille1 (France) and the University of Bangui (Central African Republic). This collaboration and the Grant-in Aid to Mr. St C. Dehou for his scientific research are financially supported by the Embassy of France to Bangui. The authors are grateful to Mrs. R.N. Vannier (ENSC, Lille1, France) for TGA and DTA analyses. References

Fig. 5. ESEM/EDS micro-observation within a cross-section of a FeOOH-coated brick sample, showing the line scannings for Al, Fe and Si.

AFNOR, 2004. Norme française NF ISO 14869-2, Nov. 2002. Evaluation de la qualité des sols. : Méthodes d'analyse chimique, volume 1. AFNOR, Saint-Denis La Plaine (France), pp. 411–418. Arias, M., Da Silva-Carballal, J., Garcia-Rio, L., Mejuto, J., Nύňez, A., 2006. Retention of phosphorus by iron and aluminum-oxides-coated quartz particles. Journal of Colloid and Interface Science 295, 65–70. Aziz, H.A., Adlan, M.N., Ariffin, K.S., 2008. Heavy metals (Cd, Pb, Zn, Ni, Cu and Cr(III)) removal from water in Malaysia: Post treatment by high quality limestone. Bioresource Technology 99, 1578–1583. Bellotto, M., Gualtieri, A., Artioli, G., Clark, S.M., 1995. Kinetic study of the kaolinite-mullite reaction sequence: 1. Kaolinite dehydroxylation. Physics and Chemistry of Minerals 22, 207–214. Boujelben, N., Bouzid, J., Elouear, Z., Feki, M., Jamoussi, F., Montiel, A., 2008. Phosphorus remocal from aqueous solution using iron coated natural and engineered sorbents. Journal of Hazardous Materials 151, 103–110. Boujelben, N., Bouzid, J., Elouear, Z., 2009. Adsorption of nickel and copper onto natural iron oxide-coated sand from aqueous solutions: study in single and binary systems. Journal of Hazardous Materials 163, 376–382. Carroll, D.L., Kemp, T.F., Bastow, T.J., Smith, M.E., 2005. Solid-state NMR characterization of the thermal transformation of a Hungarian white illite. Solid State Nuclear Magnetic Resonance 28, 31–43. Castelein, O., Soulestin, B., Bonnet, J.P., Blanchart, P., 2001. The influence of heating rate on the thermal behaviour and mullite formation from a kaolin raw material. Ceramics International 27, 517–522. Catalana, J.G., Park, C., Fenter, P., Zhang, Z., 2008. Simultaneous inner- and outer-sphere arsenate adsorption on corundum and hematite. Geochimica et Cosmochimica Acta 72, 1986–2004. Chakraborty, A.K., 2003. DTA study of the preheated kaolinite in the mullite formation region. Thermochimica Acta 398, 203–209. Chan, Y.Y., Lee, S.M., Yang, J.-K., 2009. Removal of As(III) and As(V) by natural and synthetic metal oxides. Colloids and Surfaces A: Physicochemical and Engineering Aspects 346, 202–207. Chang, F., Qu, J., Liu, R., Zhao, X., Lei, P., 2010. Practical performance and its efficiency of arsenic removal from groundwater using Fe–Mn binary oxide. Journal of Environmental Sciences 22 (1), 1–6. Chen, Y.-H., Li, F.-An, 2010. Kinetic study on removal of copper(II) using goethite and hematite nano-photocatalysts. Journal of Colloid and Interface Science 347, 277–281. Chen, Y.F., Wang, M.-C.H., Hon, M.-èH., 2004. Phase transformation and growth of mullite in kaolin ceramics. Journal of the European Ceramic Society 24, 2389–2397. Chen, H., Liu, Q., Yang, J., Frost, R.L., 2010. Thermogravimetric analysis of selected coalbearing strata kaolinite. Thermochimica Acta 507–508, 84–90. Djeribi, R., Hamdaoui, O., 2008. Sorption of copper(II) from aqueous solutions by cedar sawdust and crushed brick. Desalination 225, 95–112.

S.C. Dehou et al. / Applied Clay Science 59–60 (2012) 69–75 Dong, L., Zinin, P.V., Cowen, J.P., Ming, L.C., 2009. Iron coated pottery granules for arsenic removal from drinking water. Journal of Hazardous Materials 168, 626–632. Feng, H., Li, C., Shan, H., 2009. In-situ synthesis and catalytic activity of ZSM-5 zeolite. Applied Clay Science 42, 439–445. Gualtieri, A., Bellotto, M., 1998. Modelling the structure of the metastable phases in the reaction sequence kaolinite-mullite by X-ray scattering experiments. Physics and Chemistry of Minerals 25, 442–452. Gualtieri, A., Bellotto, M., Artioli, G., Clark, S.M., 1995. Kinetic study of the kaolinitemullite reaction sequence: 2. Mullite formation. Phys. Chem. Miner. 22, 215–222. Han, R., Lu, Z., Zou, W., Daotong, W., Shi, J., Jiujun, Y., 2006a. Removal of copper(II) and lead(II) from aqueous solution by manganese oxide coated sand II. Equilibrium study and competitive adsorption. Journal of Hazardous Materials B137, 480–488. Han, R., Zou, W., Zhang, Z., Shi, J., Yang, J., 2006b. Removal of copper (II) and lead(II) from aqueous solution by manganese oxide coated sand I. Characterization and kinetic study. Journal of Hazardous Materials B137, 384–395. Holtzapffel, T., 1985. Les minéraux argileux : préparation, analyse diffractométrique et détermination. Société Géologique du Nord. publication n°12, 136 pp. Jiang, M.-Q., Jin, X.-Y., Lu, X.-Q., Chen, Z.-L., 2010. Adsorption of Pb(II), Cd(II), Ni(II) and Cu(II) onto natural kaolinite clay. Desalination 252, 33–39. Kakali, G., Perraki, T., Tsivilis, S., Radogiannis, E., 2001. Thermal treatment of kaolin: the effect of mineralogy on the pozzolanic activity. Applied Clay Science 20, 73–80. Li, Z., Beachner, R., McManama, Z., Hanlic, H., 2007. Sorption of arsenic by surfactantmodified zeolite and kaolinite. Microporous Mesoporous Mater. 105, 291–297. Litter, M.I., Morgada, M.E., Bundschuh, J., 2010. Possible treatments of arsenic removal in Latin American waters for human consumption. Environmental Pollution 158, 1105–1118. Maliyekkal, S.M., Philip, L., Pradeep, T., 2009. As(III) removal from drinking water using manganese oxide-coated-alumina: performance evaluation and mechanistic details of surface binding. Chemical Engineering Journal 153, 101–107. Motsi, T., Rowson, N.A., Simmons, M.J.H., 2009. Adsorption of heavy metals from acid mine drainage by natural zeolite. International Journal of Mineral Processing 92, 42–48. Novakovic, T., Rozic, L., Petrovic, S., Rosic, A., 2008. Synthesis and characterization of acid-activated Serbian smectite clays obtained by statistically designed experiments. Chemical Engineering Journal 137 (2), 436–442. Oubagaranadin, J.U.K., Muthy, Z.V.P., 2010. Isotherm modeling and batch adsorber design for the adsorption of Cu(II) on a clay containing montmorillonite. Applied Clay Science 50 (3), 409–413. Oubagaranadin, J.U.K., Muthy, Z.V.P., Mallapur, V.P., 2010. Removal of Cu(II) and Zn(II) from industrial and wastewater by acid-actived montmorillonite–illite type of clay. Comptes Rendus Chimie 13 (11), 1359–1363. Ouhadi, V.R., Yong, R.N., Sedigi, M., 2006. Desorption response and degradation of buffering capability of bentonite, subjected to heavy metal contaminants. Engineering Geology 85 (1–2), 102–110. Panda, A.K., Mishra, B.G., Mishra, D.K., Singh, R.K., 2010. Effect of sulphuric acid treatment on the physicochemical characteristic of kaolin clay. Colloids and Surfaces A: Physicochemical and Engineering Aspects 363, 98–104. Ptáček, P., Kubátová, D., Havlica, J., Brandštetr, J., Šoukal, F., Opravil, T., 2010a. Isothermal kinetic analysis of the thermal decomposition of kaolinite: the thermogravimetric study. Thermochimica Acta 501, 24–29.

75

Ptáček, P., Kubátová, D., Havlica, J., Brandštetr, J., Šoukal, F., Opravil, T., 2010b. The nonisothermal kinetic analysis of the thermal decomposition of kaolinite by thermogravimetric analysis (2010). Powder Technology 204, 222–227. Qiu, W., Zheng, Y., 2009. Removal of lead, copper, nickel, cobalt and zinc from waters by a cancrinite-type zeolite synthesized from fly ash. Chemical Engineering Journal 145, 483–488. Rios, C.A., Williams, C.D., Roberts, C.L., 2008. Removal of heavy metals from acid mine drainage (AMD) using coal fly ash, natural clinker and synthetic zeolites. Journal of Hazardous Materials 156, 23–35. Rios, C.A., Williams, C.D., Fullen, M.A., 2009. Hydrothermal synthesis of hydrogarnet and tobermorite at 175 °C from kaolinite and metakaolinite in the CaO–Al2O3– SiO2–H2O system: a comparative study. Applied Clay Science 43, 228–237. Rodrigues, L.A., Maschio, L.J., Silva, R.E., Silva, M.L.C.P., 2010. Adsorption of Cr(VI) from aqueous solution by hydrous zirconium oxide. Journal of Hazardous Materials 173, 630–636. Rovira, M., Giménez, J., Martinez, M., Martinez-Llado, X., de Pablo, J., Marti, V., Duro, L., 2008. Soption of selenium(IV) and selenium(VI) onto natural iron oxides: goethite and hematite. Journal of Hazardous Materials 150, 279–284. Selvaraju, N., Pushpavannam, S., 2009. Adsorption characteristics on sand and bricks beds. Chemical Engineering Journal 147, 130–138. Shuibo, X., Chun, Z., Xinghuo, Z., Jing, Y., Xiaojian, Z., Jingsong, W., 2009. Removal of uranium (VI) from aqueous solution by adsorption of hematite. Journal of Environmental Radioactivity 100, 162–166. Stathi, P., Litina, K., Gournis, D., Giannopoulos, T.S., Deligiannakis, Y., 2007. Physicochemical study of novel organoclays as heavy metal ion adsorbents for environmental remediation. Journal of Colloid and Interface Science 316 (2), 298–309. Tãmăşan, M., Vulpoi, A., Vanea, E., Simon, V., 2010. Textural properties of the medical Algo clay as influenced by calcination. Applied Clay Science 50 (3), 418–422. Temuujin, J., Okada, K., Mackenzie, K.J.D., Jadambaa, T.S., 1998. The effect of water vapour atmospheres on the thermal transformation of kaolinite investigated by XDR, FTIR and solid state MAS NMR. Journal of the European Ceramic Society 19, 105–112. Vieira, M.G.A., Almeida Neto, A.F., Gilmeres, M.L., da Silva, M.G.C., 2010. Sorption kinetics and equilibrium for the removal of nickel ions from aqueous phase on calcined Bofe bentonite clay. Journal of Hazardous Materials 177, 362–371. Wang, S., Peng, Y., 2010. Natural zeolites as effective adsorbents in water and wastewater treatment. Chemical Engineering Journal 156, 11–24. Wang, C., Li, J., Sun, Xia, Wang, L., Sun, X., 2009. Evaluation of zeolites synthesized from fly ash as potential adsorbents for wastewater containing heavy metals. Journal of Environmental Sciences 21, 127–136. Wu, P., Zhou, Y.S., 2009. Simultaneous removal of coexistent heavy metals from simulated urban stormwater using four sorbents: a porous iron sorbent and its mixtures with zeolite and cristal gravel. Journal of Hazardous Materials 168, 674–680. Xu, T., Han, W., Chen, J.Z., Li, Z., Zou, T., 2009. Preparation of potassium acetate intercalated metakaolinite. Applied Clay Science 45, 201–205. Yadav, A.K., Kaushik, C.P., Haritash, A.K., Kansal, A., Rani, N., 2006. Defluoridation of groundwater using brick powder as an adsorbent. Journal of Hazardous Materials B128, 289–293. Yin, S., Ellis, D.E., 2009. DFT studies of Cr(VI) complex adsorption on hydroxylated hematite (1–102) surfaces. Surface Science 603, 736–746. Zibouche, F., Kerdjoudj, H., d'Espinose de Lacaillerie, J.B., Van Damme, H., 2009. Geopolymers from Algerian metakaolin. Influence of secondary minerals. Applied Clay Science 43, 453–458.