- Email: [email protected]
Crustacean derived calcium phosphate systems: Application in defluoridation of drinking water in East African rift valley
Accepted Manuscript Title: Crustacean Derived Calcium Phosphate Systems: Application in Defluoridation of Drinking Water in East African Rift Valley Authors: Agatha W. Wagutu, Revocatus Machunda, Yusufu Abeid Chande Jande PII: DOI: Reference:
S0304-3894(17)30949-4 https://doi.org/10.1016/j.jhazmat.2017.12.049 HAZMAT 19081
To appear in:
Journal of Hazardous Materials
Received date: Revised date: Accepted date:
18-7-2017 23-11-2017 18-12-2017
Please cite this article as: Wagutu AW, Machunda R, Jande YAC, Crustacean Derived Calcium Phosphate Systems: Application in Defluoridation of Drinking Water in East African Rift Valley, Journal of Hazardous Materials (2010), https://doi.org/10.1016/j.jhazmat.2017.12.049 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Crustacean Derived Calcium Phosphate Systems: Application in Defluoridation of Drinking Water in East African Rift Valley Agatha W. Wagutu*1, Revocatus Machunda2 [email protected] Tel +254-721 996719
and
Yusufu
Abeid
Chande
Jande1.
Email:
1
SC RI PT
Department of Materials, Energy Science and Engineering, Nelson Mandela African Institution of Science and Technology, P.O.BOX 447, Arusha, Tanzania 2
Department of Water, Environmental Science and Engineering, Nelson Mandela African Institution of Science and Technology, P.O.BOX 447, Arusha, Tanzania
M
A
N
U
Graphical abstract
D
Highlights
TE
1. Biogenic source of calcium precursor is explored.
EP
2. Crustacean derived calcium phosphate systems exhibit superior defluoridation efficiency
CC
3. Fluoride removal mechanism follow pseudo 2nd order kinetics
A
4. Fluoride adsorption capacity from field water = 13.5mg/g at pH of 8.5 and initial fluoride of 70 mg/L.
Abstract
N
U
SC RI PT
Calcium phosphate adsorbents, derived from prawns and crabs shell biomass wastes have been developed using wet chemistry and low temperature treatment. The adsorbents were characterized by X-ray diffractometry and Fourier transform infrared spectroscopy. Batch adsorption test were carried out to investigate their effectiveness in adsorption of fluoride from ground and surface waters. Adsorption capacities were compared with bone char and synthetic hydroxyapatite (CCHA). Results indicate that prawns derived adsorbent (PHA) formed hexagonal structure with phases identifiable with hydroxyapatite while crabs based adsorbent (CHA) formed predominantly monoclinic structure with crystalline phase characteristic of brushite. Vibrational analysis and kinetic studies predicted defluoridation occurred mainly by ion exchange and ion adsorption mechanisms. Defluoridation capacity of the adsorbents was found to be superior compared to bone char and CCHA. CHA was the most effective with efficiencies above 92% and highest capacity of 13.6 mg/g in field water with fluoride concentration of 5-70 mg/L. PHA had highest capacity of 8.5 mg/g which was still better than 2.6 mg/g recorded by CCHA and bone char. Adsorption was best described by pseudo 2nd order kinetics. The findings indicate that crustacean derived calcium phosphate systems have better potential for defluoridation than traditional bone char and synthetic systems.
A
Keywords
M
Adsorption; Crabs; Fluoride; Prawns; Kinetics 1. Introduction
EP
TE
D
Two of the five major fluoride (F-) belts in the world are found in Africa, one in the northern part and another along the East African Rift Valley (EARV) [1, 2]. Though dietary F- is beneficial for prevention of dental caries, when consumed in excess (> 1.5 mg/L) it can cause dental and skeletal fluorosis [3]. Due to climate change and persistent drought, groundwater from high F- water beds continue to be the main source of drinking water in EARV. This threatens over 80 million people with fluorosis [4].
A
CC
In high F- endemic Northern Tanzania, defluoridation has been studied as alternative means of providing safe drinking water. Application of NALGONDA technology using alum and lime proved unsustainable due to large dosage, low efficiency in high F- water and requirement for the chemicals to be imported [5]. Bone char (BCA) was introduced based on its tendency to exhibit positive charge on the surface allowing it to attract F- from solution [6] and the fact that it could be produced locally [7, 8]. Currently, BCA technology is used in Tanzania, Kenya and Ethiopia. However, it has limitations; cost of processing the bone at high temperatures is high, efficiency in high F- water is low - requiring enrichment with calcium and phosphate components [9], bacterial growth and social unacceptability [2]. These limitations have steered proposals to replace BCA with synthetic calcium phosphate system (CAPs) [10].
SC RI PT
It is thus recognized that CAPs, are potential solution to F- problem. Apart from mammalian bones, calcium precursor can be extracted from natural sources such as egg shells, crustacean shells, corals and plants [11]. Crustacean shells are of particular interest because of the abundance from sea food industry as process waste and environmental benefits associated with recycling them. It is estimated that 6-8 million tons of shells are produced annually in the world and most of it is left to rot in the sea or is dumped as landfill [12]. This work reports new approach to utilize crustacean shell. Synthesis and characterization of CAPs derived from prawn and crab shells is discussed. Defluoridation capacities were investigated to evaluate their potential in treatment of water contaminated with F-.
2. Experimental 2.1 Raw materials
N
U
Chemical reagents used were of analytical grade from sigma Aldrich. Samples of white prawns (Fenneropenaeus indicus) and swimming crabs (Portunus pelagicus) shells were collected from Dar es Salaam, Tanzania. Moisture content of clean and dry shells was determined by oven drying at 110 °C for 12 h and ash content using National Renewable Energy Laboratory (NREL) standard procedure LAP-005 [13].
M
A
2.2 Preparation of CAPs
TE
D
Dried shells were ground into 2 mm size, deproteinated using 0.25 M NaOH at 70 °C for 2 h, rinsed to neutral pH and oven dried at 50 °C for 12 h. Protein free shells were treated with 2 M HCl (solid to acid ratio 1:4 for prawns and 1:8 for crab shells) at room temperature (25±2 °C) for 24 h while stirring. The filtrate obtained was titrated with 0.1 M EDTA and calcon indicator (C21H14N2O7S.2H2O) to determine Ca2+ concentration. For comparison, filtrate was also precipitated using 0.5M Na2CO3. All the CAPs were synthesized following Equation 1 and procedure described in [14], without calcination.
EP
10Ca2+(aq) + 6PO43-(aq) +2OH- (aq) = Ca10(PO4)6(OH)2 (s)
( 1)
A
CC
Synthetic hydroxyapatite (CCHA) was prepared by reacting 250 ml of 1.09 M Ca(NO3)2.4H2O and equal volume 0.653 M (NH4)2HPO4, in a 1000 ml beaker under vigorous stirring, pH was adjusted to 10.4 using 28% ammonia solution and stirring continued at 40 °C for 4 h. The mixture was left to age at room temperature for 24 h while stirring. Precipitate formed was filtered, rinsed with distilled water to neutral pH followed by methanol rinse, oven dried at 80 °C for 24 h and ground into powder. Crustacean derived CAPs were synthesized using same general procedure; For CHA, 300 ml CaCl2 liquor (crab shell acid extract) with Ca2+ concentration of 43.69 g/l (1.09 M) was reacted with 300 ml of 0.653 M (NH4)2HPO4. PHA was synthesized using 300 ml CaCl2 liquor from prawns shell with Ca2+ concentration of 27.65 g/l (0.690 M) and 300 ml of 0.416 M (NH4)2HPO4.
2.3 Material characterization
SC RI PT
Chemical composition of CAPs was evaluated using energy dispersive X-ray (EDS) spectroscopy (Shimadzu EDX-800HS). XRD patterns were determined using Bruker D2 PHASER bench-top diffractometer with monochromatized CuKα radiation (λ= 0.15406 nm) over 2θ range of 3 -75 °, scanning speed of 3° per minute and step size of 0.02°. Functional groups analysis was conducted by FT-IR using Tensor 27 spectrometer fitted with a high-throughput screening device (HTS-XT). Tests were conducted in absorbance mode in spectral range 4000- 400 cm-1. 2.4 Batch adsorption tests
M
A
N
U
Simulated water (1000 mg/L) was prepared by dissolving 2.21 g of NaF in 1 L of distilled water. Dilution were made to obtain different concentrations of F-. Field water samples were collected from Ngarenyanyuki ward (S03°10.931’E036°51.677’), Arusha, Tanzania. Physical parameters: Electrocuductivity (EC), dissolved oxygen (DO), oxidation reduction potential (ORP), total dissolved solids (TDS) and temperature were measured using Hanna HI 9829 Multiparameter. Fconcentration was determined using ion selective electrode (ISE) connected to a Mettler Toledo seven compact pH/Ion S220 meter. Fluorinated water was mixed with total ionic strength buffer (TISAB) ratio of 1:1 before F- measurements. TISAB preparation was as follows; 57.0 ml glacial acetic acid, 58.0 g NaCl, 4 g cyclohexanediaminetetraacetic acid (CDTA) and approximately 150 ml of 6 M NaOH to adjust pH to 5.3- 5.5 in 1L. Orion Star A211 meter was used to record pH. Microbes were determined by plate count on agar media incubated at 44.5 °C for 24 h. Levels of phosphate and color were determined using HACH DR/2800 spectrophotometer.
EP
TE
D
Batch adsorption experiments were conducted using 0.25 g of adsorbent/50 ml of solution. BCA was included in field water tests for comparison. Mixtures were shaken in 250 ml plastic bottles on Retch AS 200 shaker with amplitude of 70 for 12 h. Final pH and residue F- was determined and adsorption capacity calculated using Equation 2. Effect of contact time, pH and temperature on F- removal was determined.
(C0 Ce )V m
CC
qe
(2)
A
qe = F- adsorbed at equilibrium (mg/g), C0 = Initial F- concentration (mg/L), Ce = F- concentration at equilibrium (mg/L), V= Volume (L), m = mass of adsorbent (g). 3. Results and Discussion 3.1 Preparation of CAPs Mineral content, Ca2+ concentration in acid liquors and yields of adsorbents are shown in Table 1. Theoretical yields were calculated following Equation 1. Molarity of liquors was calculated from Ca2+ concentration obtained by EDTA complexation. Mineral content by precipitation was
slightly higher than that obtained by EDTA method. This was attributed to presence of other cations in the acid liquor which co-precipitated with Ca2+. This observation is supported by presence of Si, Sr and Mn in EDS data (Table 2). Table 1: Mineral content, Ca2+ concentration in acid liquors and yields of CAPs. Prawns shell 7.0 34.0 34.5 32.6 27.7 300 0.69 17.5 20.9 83.8
Crabs shell 7.0 68.5 68.8 46.8 43.7 300 1.09 32.1 32.8 97.7
U
SC RI PT
Synthetic Ca2+ 250 1.09 27.2 27.7 99.6
Parameter Moisture (%) Ash (%) Mineral content by acid dissolution (%) Mineral content by CO32- precipitation (g/l) Ca2+ by EDTA complexation (g/l) Volume of Ca2+ solution (ml) Molarity of Ca2+ solution (M) Yield of CAPs Experimental (g) Yield of CAPs Theoretical (g) Efficiency (%)
N
3.2 Characterization of CAPs
M
A
Chemical composition of the CAPs by EDS analysis are presented in Table 2. Traces K2O and SrO were higher in PHA than in CHA while SiO2 and MnO were only identified CHA. Table 2: Chemical composition of CAPs (wt%) P2O5 29.2 30.8 35.7
D
CaO 70.8 65.0 60.1
K2 O 2.5 1.8
SiO2 1.5
SrO 1.6 0.6
MnO 0.3
EP
TE
Adsorbent CCHA PHA CHA
A
CC
XRD patterns are shown in Fig. 1. Diffraction phases of CCHA and PHA were identified by orientations 002, 211, 202 and 213 occuring at 2θ values 25.9, 32.0 34.1 and 49.5 respectively. Calculated lattice parametrs a = b = 9.434 and c = 6.920 corresponds to hexagonal phases of hydroxyapatite (HAP), JCPDS 09-0432 (a = b = 9.432, c = 6.881).
SC RI PT U N A
M
Fig 1: XRD patterns of CCHA, PHA and CHA with main peak orientations assigned
A
CC
EP
TE
D
XRD pattern of CHA revealed presence of mostly monoclinic structure with main phase orientation 020, 12-1, 14-1 at 2θ values 11.6, 20.9 and 29.3 respectively and lattice parameter, b = 15.32. The phases corresponds to brushite (CaHPO4.2H2O) peaks with lattice parameter b = 15.18, JCPDS 72-0713. HAP peaks (002, 211 and 202) were also identified but to a lesser extent. Presence of brushite in HAP crystals have been identified as factor of aging time, in which after 24 h all brushite and other calcium phosphate should convert to more stable HAP [15]. This does not however, explain the presence of major brushite phases in CHA, which was synthesized using the same reaction conditions of time and temperature as CCHA and PHA. Pattern of corresponds to that of HAP synthesized using egg shells [16], while CHA pattern was simillar to that reported for brushite crystals prepared by single diffusion gel technique [17]. Average crystallite sizes were calculated using Scherrer’s relationship; L=0.94λ/βcosθ. Peak 002 was used to calculate CCHA and PHA sizes, giving 31 nm and 15 nm respectively. Value for CCHA compare with that reported by Rusu (2005), L = 34 nm [18]. CHA gave avarage crstallize size, L = 47.9 nm using 12-1 principle peak. The three sample thus produced nano-crytalline sizes as evidenced in Table 3.
Table: 3 Selected XRD peaks analysis data (CuKα1λ = 0.15406 nm, β = 0.94) 2θ peak maxi (°) CCHA 25.80
D spacing (Å)
hkl
FWHM (°)
Crystallite size (nm)
3.46
002
0.2778
30.65
3.08 2.81 2.56 1.95
210 211 310 222
0.4690 1.2738 0.8878 0.5026
18.28 6.78 9.95 18.19
3.44 3.09 2.79 2.28 1.94
002 210 211 310 222
0.5877 2.4484 1.8960 1.6613 0.9108
14.90 3.50 4.56 5.21 9.93
7.61 4.25 3.05 2.81 2.17
020 12-1 210 202 15-2
0.1591 0.1761 0.2097 0.4377 0.2300
SC RI PT
28.90 32.00 39.88 46.49 PHA 25.89 28.85 32.06 39.56 46.80 CHA 11.62 20.90 29.23 34.12 41.50
52.45 47.93 40.91 19.84 38.59
A
CC
EP
TE
D
M
A
N
U
FT-IR spectra determined using diffuse reflectance ZnSe prisim are given in Fig. 2. Frequencies were assigned according to literature values [16, 19]. Modes related to structural OH were observed in two regions; 620- 635 cm-1 and 3480- 3570 cm-1 while intense P=O stretching bands associated with PO43- were seen at 1160-1204 cm-1, Table 4.
Fig 2: FT-IR spectra of CAPs adsorbents.
Spectra of CHA and PHA after defluoridation of groundwater with initial F- of 29.4 mg/L are compared in Fig. 3. Slight difference was observed in spent PHA (FPHA); split in 630 cm-1 peak forming two peaks, one to longer wavelength (620 cm-1) and another to shorted wavelength 635 cm-1 which is accompanied by decrease in absorption intensity (Fig. 3a). This change has been identified earlier in HAP by [20] but in their case the 630 cm-1 disappeared and new peak formed
at 738 cm-1. Intense peak at 1175 cm-1 also shifted slightly to longer wavelength with reduced intensity. Results indicate that some amount of F- was adsorbed by PHA as evidenced from batch adsorption tests.
A
N
U
SC RI PT
(a)
EP
TE
D
M
(b)
CC
Fig 3: Comparisons FT-IR spectra PHA and CHA before and after F- adsorption.
A
In CHA spectrum, decrease in absorption intensities was observed for most peaks. Peak 624 cm-1 shifted slightly to shorter wavelength and split into two, 627 and 639 cm-1. Broad intense peak at 1202 cm-1 became narrower and shifted to longer wavelength accompanied by increase in intensity (Fig. 3b). Bands at 1095 cm-1 and 3481 - 3552 cm-1 which are associated with HPO4 and structural OH stretching respectively, disappeared completely in FCHA. This indicate the OH and HPO4 groups were effectively replaced by F-. Similar observations has been reported in [21, 22] and is also well supported from batch adsorption tests which show that CHA removed almost 95% of F- from the water sample.
Table 4: IR modes and observed frequencies for CAPs and fluorinated FPHA and FCHA Vibrational frequencies CCHA PHA 588 583 620 630 875 877 963 963 1087 1088 1169 1175 1422 1420 1651 1651 3770 -
FPHA 583 635 873 963 1089 1169 1424 1655 -
3.3 Batch absorption test
CHA 584 624 875 1004 1095 1202 1655 3552
FCHA 586 638 875 961 1081 1169 1424 1655 -
SC RI PT
Assigment IR Mode PO4 bending Structural OH liberation P-O-P stretching PO4 stretching PO4 bending PO4 P=O stretching CO3- group H2O adsorbed Strucutal OH stretch
M
A
N
U
Field parameters of water samples are presented in Table 5. All samples contained F- above permissible limits and ORP below 400-650 mV which is deemed necessary for oxidation of contaminants like bacterial pathogens [23]. Samples F4-F7 failed quality test for EC and TDS. EC of 800-1000 µs/cm and TDS 500 mg/L are acceptable for drinking water. Surface water (F3, F5 and F7) failed quality standards for microbial and color requirements. Color values above 15 units are considered unacceptable but up to 85 units is found in natural drinking water [24]. Table 5: Water quality parameters of field samples ORP (mV)
DO (ppm)
EC (µS/cm)
TDS (ppm)
PO43(mg/)
7.6 7.5 8.3 8.5
92.2 91.6 -6.5 50.6
6.4 5.2 6.3 5.6
483 816 445 2883
242 408 223 1441
24.0 23.0
9.5 8.1
21.3 39.0
7.2 5.72
2062 3500
26.9
8.6
-190.9
3.06
4352
Color (pt Co)
Temp (°C)
F1 F2 F3 F4
Spring Spring River Shallow well Stream Shallow well Swamp
7.0 8.0 215.0 6.0
16.7 19.2 19.8 22.0
EP 3781.0
F(mg/L)
0.27 0.12 0.21 0.13
E.Coli (CFU/100 ml) 00 00 200 00
1032 1749
0.71 0.13
232 00
31.00 33.70
2178
0.65
17
70.10
4.20 6.48 18.70 29.40
A
F7
710.0 6.0
CC
F5 F6
pH
D
Source
TE
ID
Results of F- adsorption are shown in Table 6. Defluoridation capacities (DC) in low F- levels (410 mg/L) were almost same for all adsorbents. In simulated water of 10 mg/L DCs were 1.933, 1.882 and 1.852 mg/g for CHA, PHA and CCHA respectively. These results compare closely with that reported for nano-HAP, 1.845 mg/g [25]. Above 10 mg/L, a distinctive trend was observed; CHA > PHA > CCHA (Fig. 4a). CHA recorded the highest DC of 21 mg/g, this matches monolayer capacity reported for nano-HAP derived from waste phosphogypsum (19.7 mg/g) [26]. In field
water samples, highest DCs were 13.6, 8.5 and 2.6 for CHA, PHA and CCHA respectively (Fig. 4b). No data were found from literature on performance of CAPs in the field water with F- above 10 mg/L. From these data it can be concluded that CHA exhibits superior F- removal in both simulated and field water. PHA and CCHA adsorbent are appropriate for F- levels below 10 mg/L, with PHA performing much better than CCHA.
Sample
pH
F initial (mg/L)
SC RI PT
Table 6: Batch adsorption test for F- in different adsorbents using simulated and field water samples F- residue after 12 h treatment (pH final) CCHA PHA CHA
BCA*
0.607±0.0 (6.6)
-
0.362±0.0 (5.8)
-
0.977±0.0 (5.9)
-
Simulated water 10.27±0.13
1.01±0.2 (8.0)
0.86±0.2 6.8)
S2
3.80
50.56±0.42
21.67±0.8 (8.2)
4.95±0.2 (6.2)
S3
3.87
106.25±3.0
80.0±2.7 (7.7)
56.5±1.5 (6.4)
F1
7.60
4.20±00
0.638±0.3 (8.1)
U
5.90
0.845±0.4 (8.0)
0.275±0.0
(6.8)
1.9 ±0.0 (8.5)
F2
7.53
6.48±0.1
1.21±0.0
(8.2)
0.744±0.2
(8.0)
0.470±0.1
(7.0)
2.6±0.2 (8.5)
F3
8.30
18.70±00
8.66±0.2
(8.0)
2.83±0.0
(7.9)
0.221±0.0
(7.3)
9.0±0.4 (8.8)
F4
8.00
29.40±1.6
14.67±0.8 (8.4)
A
S1
6.02±0.2
(8.4)
0.582±0.2
(8.0)
17.4±0.3 (8.9)
F5
9.50
31.00±1.4
21.70±0.1 (9.3)
9.54±0.4
(9.1)
0.485±0.1
(7.8)
22.0±0.0 (9.4)
F6
8.10
33.70±1.5
18.87±0.7 (8.4)
6.60±0.4
(8.5)
0.569±0.2
(8.2)
21.9±0.2 (8.8)
F7
8.60
70.10±4.2
57.20±1.6
27.57±0.3
(8.6)
2.160±0.4
(8.1)
57.0±0.1 (8.8)
M
N
Field water
D
(8.8)
* BCA was in granular form as obtained from defluoridation center. It thus not included in the analysis CAPs
A
CC
EP
TE
synthesized in the laboratory.
SC RI PT U N
M
A
Fig 4: a) Efficiency of F- removal b) defluoridation capacities for different adsorbents.
EP
TE
D
Effect of adsorbent on water quality was tested for CHA using samples F4 and F5 (Table 7). Results show CHA removed yellow tint in F5 but increased color in F4, both samples remained clear. Ability of CHA to remove color can be explained by presence of MnO in the elemental composition, Table 2. MnO/hydroxides are known to promote degradation of complex organics such as humic acid, polychlorinated biphenyls and phenols [27]. The adsorbent did not have any effect on the microbial count and increased the level of phosphate slightly. No other major changes were observed in other parameters.
CC
Table 7: Water quality parameters before and after treatment with CHA Lab parameters
Color (pt Co)
pH
ORP (mV)
DO (ppm)
EC (µS/cm)
TDS ( ppm)
PO4 3(mg/)
F4
Before
6.00
8.4
149
5.2
2815
1440
After
12.0
7.8
84.1
4.14
2767
Before
710.00
9.6
124
4.4
After
109
7.5
81
4.5
A
ID
F5
F(mg/L)
0.1
E.Coli (CFU/ 100 ml) 00
1385
2.4
00
0.58
1904
959
0.7
232
31.00
1952
977
2.8
232
0.49
29.40
TE
D
M
A
N
U
SC RI PT
3.4 Effect of pH Influence of pH on adsorption of F- was studied using sample, F4. The initial pH was varied from 3-9 at room temperature for 180 min. Results (Fig 5) show efficiency of CCHA and PHA decreased considerably with increase in pH and best performance was achieved at pH = 3 with over 95% removal. This was attributed to protonation of adsorbent’s surface sites in acidic media, thus increasing affinity for F-, a phenomenon observed with most adsorbents [28-30]. Performance of CHA was not significantly affected by pH, narrow range (98-96%) was observed between pH 38.5.
EP
Fig 5: Effect of initial pH on the adsorption of F- on CAPs
CC
3.5 Adsorption isotherms
A
Adsorptive characteristics of the CAPs was tested at room temperature using two parameter isotherms; Langmuir, Freundlich, Dubinin-Radushkevich (D-R) and Temkin models in linear and non-linear forms as described in [31]. Nonlinear isotherms are shown in Fig. 6 and the corresponding data in Table 8. Linearized data are illustrated in Table 9. The general observation indicate that adsorption in simulated water is very different from field water in terms of adsorption capacity and correlation coefficients.
CCHA was best modeled by linear Langmuir Equation 3, with R2 > 0.98. Maximum adsorption determined experimentally for simulated and field water are 6.4 and 3.0 mg/g and corresponds well to calculated monolayer adsorption 6.2 and 3.2 mg/g respectively. This signifies a monolayer coverage onto a structurally homogeneous surface with finite number of identical sites [32]. This observation can be supported from XRD and EDX data which show only hydroxyapatite phases are present in the material.
SC RI PT
Ce 1 C e qe bK LCe qm
(3)
where Ce is F- concentration at equilibrium (mg/L), qe is adsorption capacity at equilibrium (mg/g), qm is monolayer adsorption capacity (mg/g) and KL is Langimuir constant (L/mg)
RT ln( KT Ce ) bT
(5)
A
qe
(4)
N
qe (qm ) Exp ( K ad 2 ) , ε = RT ln[1+1/Ce]
U
For PHA, simulated water was best modeled by D-R with R2>0.98 and field water by Temkin with R2= 0.93, Equations 4 and 5. The two isotherms describe empirical interaction of adsorbate and adsorbent, in which D-R assumes adsorption mechanism with Gaussian energy distribution on to heterogeneous surface while Temkin indicates a linear decrease in adsorption energy with coverage without the need for homogenous surface [31].This could indicate multimolecular layers coverage in nanosized adsorbent as seen with XRD crystallite size of 15 nm in Table 3.
D
M
where, Kad is adsorption equilibrium constant (mol2/kJ2), ε is Polanyi potential, E is mean free energy of adsorption (kJ/mol), R is gas constant (8.314J/mol/K), T is Temperature at 298K, KT is equilibrium constant (L/g) and b is constant related to heat of adsorption (J/mol).
CC
EP
TE
CHA was barely modeled by linear regression. Nonlinear D-R isotherm fairly represented adsorption in simulated water with R2= 0.82. Field water was equally modeled by all isotherms including Freundlich Equation 6 with R2 = 0.77-0.80. The difficulty in modelling CHA was attributed to multiphase structure which showed presence of mainly brushite and HAP in the crystal lattice. Presence of MnO which has catalytic properties on the surface may also have contributed to the heterogeneity resulting in different adsorption characteristics.
1/ n
qe K F Ce
(6)
A
where KF is Freundlich constant (mg/g) and n relates to adsorption intensity.
Table 8: Nonlinear Langmuir, Fruendlich, Temkin and Dubinin-Radushkevich isotherms constants for adsorption of F- onto CAPs Field water CCHA
PHA
CHA
6.176 0.325 0.900
11.707 0.428 0.910
2913.541 0.006 0.409
3.204 0.470 0.790
10.524 0.124 0.930
32.458 0.336 0.804
2.523 4.743 0.756
4.750 4.380 0.646
20.554 0.669 0.461
1.411 4.583 0.575
1.741 2.070 0.930
7.701 1.331 0.799
7.934 2.526 0.833
5.987 1.196 0.774
4.113 0.205 0.325
7.726 4.422 0.684
1.858 1.259 0.930
5.556 0.481 0.768
6.673 2.027 0.903
11.481 1.236 0.984
18.177 0.635 0.815
2.854 1.381 0.776
6.782 0.279 0.788
15.446 3.216 0.793
SC RI PT
CHA
A
CC
EP
TE
D
M
A
N
Freundlich KF (mg/L) N R2 Temkin KT (L/g) bT R2 Dubinin Radushkevich qm (mg/g) E (KJ/mol) R2
Simulated water CCHA PHA
U
Parameter/ model Langmuir qm (mg/g) KL (L/mg) R2
Fig 6: Non-linear adsorption isotherms of F- in different CAPs. a,b,c are simulated water in CCHA, PHA and CHA respectively while d,e and f are isotherm for field water in CCHA, PHA and CHA respectively.
Table 8 : Linear Langmuir, Fruendlich, Temkin and Dubinin-Radushkevich isotherms constants for adsorption of F- onto CAPs PHA
CHA
CCHA
PHA
CHA
5.490 0.015 0.986
10.360 0.015 0.994
0.001
2.650 0.021 0.985
11.060 0.076 0.883
0.008
2.054 3.662 0.872
3.151 2.648 0.725
0.095
1.000 2.925 0.800
1.201 1.526 0.894
0.445
7.933 0.977 0.833
5.987 2.071 0.774
0.324
7.726 0.560 0.684
1.858 1.968 0.930
5.610 5.151 0.768
5.174 0.361 1.176 0.820
10.270 0.477 1.024 0.991
0.04
2.692 0.269 1.363 0.851
5.302 0.435 1.073 0.858
0.376
SC RI PT
CCHA
U
Parameter/model Langmuir qm (mg/g) KL (L/mg) R2 Freundlich KF (mg/L) n R2 Temkin KT(L/g) bT R2 Dubinin- Radushkevich qm (mg/g) K (mol2/kJ2) E (KJ/mol) R2
M
A
N
Maximum qm calculated from D-R non-linear isotherm were compared with qm of other adsorbents to validate PHA and CHA, Table 10. It can be seen that our results are comparable to best performing HAP based adsorbents reported. High DC obtained in field water at natural pH is unique for our adsorbents, unfortunately no comparative data from literature as best performances are only based on simulated water and adjusted pH.
D
Table 9: Monolayer F- adsorption capacities on various adsorbent materials at 298 ±2 K Adsorbent
C0 max (mg/L)
qm (mg/g)
Reference
50.0
0.1 g/100 ml
19.7
[26]
200.0
5 mg/10 ml
25.8
[33]
10.0
0.25 g/50 ml
1.3
[25]
Nano sized -HAP
95.0
2 g/L
11.0
[34]
Mg/Ce/Mn/O-DE*
100.0
0.6 g/100 ml
13.3
[35]
Heat activated Dolomite
500.0
2 g/L
228.8
[36]
2.7
[30]
HAP-nano wires
CC
EP
Nano-HAP
TE
Phospogypsum HAP
Adsorbent loading
Bone char
A
PHA –prwan shell CAP
Brushite CHA-crab shell CAP
20.0
1 g/480 ml
106.0
11.5
70.0 **
0.25 g/50 ml
6.8
25.0
0.2 g/25 ml
6.6
106.0 70.0 * *
*DE= diatomaceous earth ** Field water sample
18.2 0.25 g/50 ml
15.4
This study
[37] This study
3.6 Effect of contact time and adsorption kinetics
D
M
A
N
U
SC RI PT
Kinetic results indicate rapid uptake within the first 10 min, with > 75% of the possible F- adsorbed. CCHA for instance removed 37% of the possible 42% maximum in the first 10 min. The process proceeded slowly for the rest of reaction time (Fig. 7).
TE
Fig 7: Variation of % F- removal with contact time on CAPs using groundwater sample (F4) with initial Fof 29.4 mg/L and initial pH = 8.5.
A
CC
EP
Experimental data on rate of F- adsorption was analyzed using four kinetic models, Table 11 [38, 39], results are shown in Table 12. Pseudo 2nd order model described the kinetic process more appropriately compared to other models (Fig. 8). Calculated qe values agree well with the experimental values and R2 > 0.99 shows that the model can be applied for the entire adsorption process, a phenomenon that confirms existence of chemisorption [40]. CHA was well modeled by all equations with R2 > 0.9. This implies that several mechanisms could be involved, hence the high DCs observed.
Table 10: Kinetic models equations Linear from 1st
Pseudo 2nd order Intra-particle diffusion Elovich
log (qe qt ) log qe
K1t 2.302
t 1 t 2 qt K s (qe ) qe
qt Ki pt1/ 2 C 1
ln( )
Kip=intra-particle diffusion (mg/(g.min1/2)) C= intercept 1
ln t
α= initial adsorption (mg/(g.min)) β= desorption constant (g/mg)
CC
EP
TE
D
M
A
N
U
qt
Parameters qe =F- adsorbed at equilibrium (mg/g) qt = F- adsorbed at time ‘t’ (mg/g) K1=Rate constant for pseudo 1st order (min-1) Ks=Rate constant for pseudo 2nd order (g/(mg.min)
SC RI PT
Model Pseudo order
A
Fig 8: Pseudo-2nd order kinetic plots for fluoride adsorption by CAPs.
Table 11: Kinetic parameters of CAPs PHA
CHA
1.00 2.22 0.64 0.29
0.84 3.88 -0.0039 0.18
0.98 5.40 1.97 0.94
2.26 2.22 0.180 1.00
3.65 3.88 0.06 0.99
1.91 0.03 0.88
2.65 0.08 0.55
9.64 0.96
3.37 0.55
U
SC RI PT
CCHA
5.62 5.40 0.03 1.00
-32.52 8.22 0.94 2.48 0.91
A
N
Parameter/model Pseudo 1st order qm cal (mg/g) qe exp (mg/g) K1 (L/mg) R2 Pseudo 2nd order qm cal (mg/g) qm exp (mg/g) Ks (g/mg.min) R2 Intra-particle diffusion C Kip (mg/g.min1/2) R2 Elovich Β (mg/(g.min)) R2
M
3.7 Effect of temperature and thermodynamic parameters
D
Influence of temperature on the adsorption was tested at 300 -327 K for 180 min (Fig. 9). Thermodynamic parameters; Gibbs energy (ΔG◦), enthalpy of reaction (ΔH0) and entropy (ΔS0) were calculated using equations 7-8 as described in [41], results shown in Table 13.
qe ) Ce
qe H 0 1 S 0 ) Ce R T R
EP
ln(
TE
G 0 RT ln(
(7)
(8)
A
CC
Positive ΔG0 indicated adsorption process was nonspontaneous for CCHA, PHA and spontaneous for CHA. The best temperature for adsorption for all the CAPs was identified as T= 307 K. Positive ΔH0 suggested the endothermic nature of adsorption, which is evidenced from slight increase in removal between 300-308 K. Positive ΔS0<<<<1 suggested no changes in randomness occurred at the solid/liquid interface during the adsorption process.
SC RI PT U N A
D
M
Fig 9: Variation of F- adsorption with temperature of groundwater sample (F4)
Table 12: Thermodynamic parameters of interaction of F- with CAPS
EP
CC
PHA
A
CHA
Temp (K) 302 307 314 325
TE
Adsorbent CCHA
302 307 314 325 302 307 314 325
ΔG (kJ mol-1) 4.75 4.42 5.19 4.32 2.10 1.46 1.65 1.92 -4.47 -6.81 -6.19 -7.84
ΔH (kJ mol-1)
ΔS (kJ mol-1)
6.61
0.009
1.99
0.0007
29.78
0.12
3.8 Elucidation of F- removal mechanism
SC RI PT
Mechanisms of F- uptake from solutions has been evaluated by several authors [22, 25, 26, 34]. Three main processes are highlighted; Ion exchange, electrostatic adsorption and precipitation. In this study, mechanism was elucidated from kinetic and FT-IR analysis. In the spectra (Fig. 3), shift or disappearance of bands related to OH and HPO4 bands accompanied by changes in intensity after adsorption was assumed to indicate F- substitution and formation of fluorapatite by CHA and PHA as follows; + 2F- (aq) =
Ca10(PO4)6(OH)2 (s) CaHPO4.2H2O (s) + 2F- (aq)
Ca10(PO4)6F2 (s) + 2OH-(aq)
(9)
= Ca10(PO4)6F2 (s) + 4HPO42- (aq) + 6H+(aq) + 20H2O (l) (10)
N
U
Kinetic tests showed equilibrium pH increased slightly in PHA treated water while it decreased in CHA during the rapid uptake. This was seen to confirm release of OH- and H+ into solution from the above reactions. Slow process was predicted to indicate electrostatic adsorption which increased with decrease in pH. It can thus be concluded that though ion exchange plays a significant role, it is not the only force driving the removal of F- from solution by the CAPs.
A
4. Conclusion
CC
EP
TE
D
M
CAPs derived from crustacean waste have been developed. The process of calcium isolation is versatile because it also allows recovery of other bioactive components like chitin which has valuable applications. Analytical characterization using XRD and FT-IR show that prawn shell produced HAP with nanocystallite size (~15 nm), while crab shell gave mostly brushite with average crystallite size of 50 nm. The adsorbents are found to have excellent DC in field water with F- concentration 5-70 mg/L. Highest capacity recorded are 13.5 mg/g and 8.5 mg/g for CHA and PHA respectively, which are higher than that of CCHA and bone char (2.6 mg/g). Adsorption was best described by pseudo 2nd order kinetics and fairly defined by non-linear D-R isotherm. Anion exchange was proposed as the main driving force in F- uptake followed by slow adsorption process. Overall results indicate that PHA is suitable for treating water with F- of 10 mg/L and below while CHA is efficient at higher levels. The CAPs are thus applicable for high F- water found in EARV. The crustacean derived adsorbents also reduce color in tinted water, which adds to quality of the treated water. Slight increase in phosphate level and failure to remove microbes were identified as limitations of the adsorbents.
A
Acknowledgement This project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement Number 690378. The authors are also thankful to Mwalimu Nyerere African Union Scholarship Scheme (MNAUSS) for the award of fellowship to Agatha Wagutu.
References
[1] R. Brunt, L. Vasak, J. Griffioen, Fluoride in groundwater: probability of occurrence of excessive concentration on global scale, IGRAC, 2004.
SC RI PT
[2] E. Dahi, Africa’s u-turn in defluoridation policy: from the Nalgonda technique to bone char, in: Fluoride 2016, pp. 401-416. [3] T. Rango, A. Vengosh, M. Jeuland, G.M. Whitford, R. Tekle-Haimanot, Biomarkers of chronic fluoride exposure in groundwater in a highly exposed population, Science of The Total Environment, 596-597 (2017) 1-11. [4] M. Amini, K. Mueller, K.C. Abbaspour, T. Rosenberg, M. Afyuni, K.N. Møller, M. Sarr, C.A. Johnson, Statistical Modeling of Global Geogenic Fluoride Contamination in Groundwaters, Environmental Science & Technology, 42 (2008) 3662-3668.
U
[5] J.K. Fawell, K. Bailey, Fluoride in drinking-water, World Health Organization, 2006.
A
N
[6] L.R. Brunson, D.A. Sabatini, Practical considerations, column studies and natural organic material competition for fluoride removal with bone char and aluminum amended materials in the Main Ethiopian Rift Valley, Science of The Total Environment, 488-489 (2014) 580-587.
M
[7] H. Mjengera, G. Mkongo, Appropriate deflouridation technology for use in flourotic areas in Tanzania, Physics and Chemistry of the Earth, Parts A/B/C, 28 (2003) 1097-1104.
D
[8] E. Dahi, Optimisation of bone char production using the standard defluoridation capacity procedure, International Society for Fluoride Research, Chiang Mai, (2015) 29-36.
EP
TE
[9] J. Albertus, H. Bregnhøj, M. Kongpun, Bone char quality and defluoridation capacity in contact precipitation, in: 3rd International workshop on fluorosis prevention and defluoridation of water, 2000, pp. 61-72.
CC
[10] M. Meta, Improving access to safe drinking water: prospection for low-fluoride sources. Brief report presenting main findings, in, Natural Environment Research Council, 2014. [11] M. Akram, R. Ahmed, I. Shakir, W.A.W. Ibrahim, R. Hussain, Extracting hydroxyapatite and its precursors from natural resources, Journal of materials science, 49 (2014) 1461-1475.
A
[12] N. Yan, X. Chen, Don't waste seafood waste: Turning cast-off shells into nitrogen-rich chemicals would benefit economies and the environment, Nature, 524 (2015) 155-158. [13] A. Sluiter, B. Hames, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton, Determination of ash in biomass: LAP-005 NREL analytical procedure, National Renewable Energy Laboratory, Golden, (2004).
[14] F. Chen, Z.-C. Wang, C.-J. Lin, Preparation and characterization of nano-sized hydroxyapatite particles and hydroxyapatite/chitosan nano-composite for use in biomedical materials, Materials letters, 57 (2002) 858-861.
SC RI PT
[15] H.N. M. Rusu, M. Wilke, B. Tiersch, P. Fratzl and G. Peter, Size-controlled hydroxyapatite nanoparticles as self-organized organic-inorganic composite materials, Biomaterials, 26 (2005) 5414-5426. [16] V. Kashkarov, D. Goloshchapov, A. Rumyantseva, P. Seredin, E. Domashevskaya, I. Spivakova, B. Shumilovich, X-ray diffraction and IR spectroscopy investigation of synthesized and biogenic nanocrystalline hydroxyapatite, Journal of Surface Investigation: X-ray, Synchrotron and Neutron Techniques, 5 (2011) 1162-1167. [17] K. Rajendran, C. Dale Keefe, Growth and characterization of calcium hydrogen phosphate dihydrate crystals from single diffusion gel technique, Crystal Research and Technology, 45 (2010) 939-945.
N
U
[18] V. Rusu, C. Ng, M. Wilke, B. Tiersch, P. Fratzl, M. Peter, Size-controlled hydroxyapatite nanoparticles as self-organized organic?inorganic composite materials, Biomaterials, 26 (2005) 5414-5426.
M
A
[19] R. Panda, M. Hsieh, R. Chung, T. Chin, FTIR, XRD, SEM and solid state NMR investigations of carbonate-containing hydroxyapatite nano-particles synthesized by hydroxide-gel technique, Journal of Physics and Chemistry of Solids, 64 (2003) 193-199.
D
[20] R. Ebrahimi-Kahrizsangi, B. Nasiri-Tabrizi, A. Chami, Characterization of single-crystal fluorapatite nanoparticles synthesized via mechanochemical method, Particuology, 9 (2011) 537544.
EP
TE
[21] H.E. Namin, H. Hashemipour, M. Ranjbar, Effect of aging and calcination on morphology and properties of synthesized nanocrystalline TiO 2, International Journal of Modern Physics B, 22 (2008) 3210-3215. [22] M. Tafu, T. Chohji, Reaction between calcium phosphate and fluoride in phosphogypsum, Journal of the European Ceramic Society, 26 (2006) 767-770.
A
CC
[23] V. Goncharuk, V. Bagrii, L. Mel’nik, R. Chebotareva, S.Y. Bashtan, The use of redox potential in water treatment processes, Journal of Water Chemistry and Technology, 32 (2010) 19. [24] WHO, Guidelines for Drinking-water Quality, in, 2011. [25] C.S. Sundaram, N. Viswanathan, S. Meenakshi, Defluoridation chemistry of synthetic hydroxyapatite at nano scale: equilibrium and kinetic studies, Journal of hazardous materials, 155 (2008) 206-215.
[26] D. Zhang, H. Luo, L. Zheng, K. Wang, H. Li, Y. Wang, H. Feng, Utilization of waste phosphogypsum to prepare hydroxyapatite nanoparticles and its application towards removal of fluoride from aqueous solution, Journal of hazardous materials, 241 (2012) 418-426.
SC RI PT
[27] B.M. Tebo, J.R. Bargar, B.G. Clement, G.J. Dick, K.J. Murray, D. Parker, R. Verity, S.M. Webb, Biogenic manganese oxides: properties and mechanisms of formation, Annu. Rev. Earth Planet. Sci., 32 (2004) 287-328. [28] Y. Tang, X. Guan, J. Wang, N. Gao, M.R. McPhail, C.C. Chusuei, Fluoride adsorption onto granular ferric hydroxide: Effects of ionic strength, pH, surface loading, and major co-existing anions, Journal of hazardous materials, 171 (2009) 774-779. [29] G. Alagumuthu, M. Rajan, Equilibrium and kinetics of adsorption of fluoride onto zirconium impregnated cashew nut shell carbon, Chemical Engineering Journal, 158 (2010) 451-457.
N
U
[30] N.A. Medellin-Castillo, R. Leyva-Ramos, R. Ocampo-Perez, R.F. Garcia de la Cruz, A. Aragon-Pina, J.M. Martinez-Rosales, R.M. Guerrero-Coronado, L. Fuentes-Rubio, Adsorption of fluoride from water solution on bone char, Industrial & Engineering Chemistry Research, 46 (2007) 9205-9212.
A
[31] K. Foo, B. Hameed, Insights into the modeling of adsorption isotherm systems, Chemical Engineering Journal, 156 (2010) 2-10.
M
[32] A. Günay, E. Arslankaya, I. Tosun, Lead removal from aqueous solution by natural and pretreated clinoptilolite: adsorption equilibrium and kinetics, Journal of hazardous materials, 146 (2007) 362-371.
TE
D
[33] J. He, K. Zhang, S. Wu, X. Cai, K. Chen, Y. Li, B. Sun, Y. Jia, F. Meng, Z. Jin, Performance of novel hydroxyapatite nanowires in treatment of fluoride contaminated water, Journal of hazardous materials, 303 (2016) 119-130.
EP
[34] V. Sternitzke, R. Kaegi, J.N. Audinot, E. Lewin, J.G. Hering, C.A. Johnson, Uptake of fluoride from aqueous solution on nano-sized hydroxyapatite: examination of a fluoridated surface layer, Environ Sci Technol, 46 (2012) 802-809.
CC
[35] W.M. Gitari, A.A. Izuagie, J.R. Gumbo, Synthesis, characterization and batch assessment of groundwater fluoride removal capacity of trimetal Mg/Ce/Mn oxide-modified diatomaceous earth, Arabian Journal of Chemistry, (2017).
A
[36] V.C.a.S. Prasad, Rapid removal of fluoride from aqueous media using activated dolomite, Analytical Methods, 7 (2015) 8304-8312. [37] M. Mourabet, H. El Boujaady, A. El Rhilassi, H. Ramdane, M. Bennani-Ziatni, R. El Hamri, A. Taitai, Defluoridation of water using Brushite: Equilibrium, kinetic and thermodynamic studies, Desalination, 278 (2011) 1-9.
[38] E. Demirbas, M. Kobya, E. Senturk, T. Ozkan, Adsorption kinetics for the removal of chromium (VI) from aqueous solutions on the activated carbons prepared from agricultural wastes, Water Sa, 30 (2004) 533-539.
SC RI PT
[39] S. Ayoob, A.K. Gupta, P.B. Bhakat, V.T. Bhat, Investigations on the kinetics and mechanisms of sorptive removal of fluoride from water using alumina cement granules, Chemical Engineering Journal, 140 (2008) 6-14. [40] H. Qiu, L. Lv, B.-c. Pan, Q.-j. Zhang, W.-m. Zhang, Q.-x. Zhang, Critical review in adsorption kinetic models, Journal of Zhejiang University-SCIENCE A, 10 (2009) 716-724.
A
CC
EP
TE
D
M
A
N
U
[41] I. Anastopoulos, G.Z. Kyzas, Are the thermodynamic parameters correctly estimated in liquid-phase adsorption phenomena?, Journal of Molecular Liquids, 218 (2016) 174-185.
S0304-3894(17)30949-4 https://doi.org/10.1016/j.jhazmat.2017.12.049 HAZMAT 19081
To appear in:
Journal of Hazardous Materials
Received date: Revised date: Accepted date:
18-7-2017 23-11-2017 18-12-2017
Please cite this article as: Wagutu AW, Machunda R, Jande YAC, Crustacean Derived Calcium Phosphate Systems: Application in Defluoridation of Drinking Water in East African Rift Valley, Journal of Hazardous Materials (2010), https://doi.org/10.1016/j.jhazmat.2017.12.049 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Crustacean Derived Calcium Phosphate Systems: Application in Defluoridation of Drinking Water in East African Rift Valley Agatha W. Wagutu*1, Revocatus Machunda2 [email protected] Tel +254-721 996719
and
Yusufu
Abeid
Chande
Jande1.
Email:
1
SC RI PT
Department of Materials, Energy Science and Engineering, Nelson Mandela African Institution of Science and Technology, P.O.BOX 447, Arusha, Tanzania 2
Department of Water, Environmental Science and Engineering, Nelson Mandela African Institution of Science and Technology, P.O.BOX 447, Arusha, Tanzania
M
A
N
U
Graphical abstract
D
Highlights
TE
1. Biogenic source of calcium precursor is explored.
EP
2. Crustacean derived calcium phosphate systems exhibit superior defluoridation efficiency
CC
3. Fluoride removal mechanism follow pseudo 2nd order kinetics
A
4. Fluoride adsorption capacity from field water = 13.5mg/g at pH of 8.5 and initial fluoride of 70 mg/L.
Abstract
N
U
SC RI PT
Calcium phosphate adsorbents, derived from prawns and crabs shell biomass wastes have been developed using wet chemistry and low temperature treatment. The adsorbents were characterized by X-ray diffractometry and Fourier transform infrared spectroscopy. Batch adsorption test were carried out to investigate their effectiveness in adsorption of fluoride from ground and surface waters. Adsorption capacities were compared with bone char and synthetic hydroxyapatite (CCHA). Results indicate that prawns derived adsorbent (PHA) formed hexagonal structure with phases identifiable with hydroxyapatite while crabs based adsorbent (CHA) formed predominantly monoclinic structure with crystalline phase characteristic of brushite. Vibrational analysis and kinetic studies predicted defluoridation occurred mainly by ion exchange and ion adsorption mechanisms. Defluoridation capacity of the adsorbents was found to be superior compared to bone char and CCHA. CHA was the most effective with efficiencies above 92% and highest capacity of 13.6 mg/g in field water with fluoride concentration of 5-70 mg/L. PHA had highest capacity of 8.5 mg/g which was still better than 2.6 mg/g recorded by CCHA and bone char. Adsorption was best described by pseudo 2nd order kinetics. The findings indicate that crustacean derived calcium phosphate systems have better potential for defluoridation than traditional bone char and synthetic systems.
A
Keywords
M
Adsorption; Crabs; Fluoride; Prawns; Kinetics 1. Introduction
EP
TE
D
Two of the five major fluoride (F-) belts in the world are found in Africa, one in the northern part and another along the East African Rift Valley (EARV) [1, 2]. Though dietary F- is beneficial for prevention of dental caries, when consumed in excess (> 1.5 mg/L) it can cause dental and skeletal fluorosis [3]. Due to climate change and persistent drought, groundwater from high F- water beds continue to be the main source of drinking water in EARV. This threatens over 80 million people with fluorosis [4].
A
CC
In high F- endemic Northern Tanzania, defluoridation has been studied as alternative means of providing safe drinking water. Application of NALGONDA technology using alum and lime proved unsustainable due to large dosage, low efficiency in high F- water and requirement for the chemicals to be imported [5]. Bone char (BCA) was introduced based on its tendency to exhibit positive charge on the surface allowing it to attract F- from solution [6] and the fact that it could be produced locally [7, 8]. Currently, BCA technology is used in Tanzania, Kenya and Ethiopia. However, it has limitations; cost of processing the bone at high temperatures is high, efficiency in high F- water is low - requiring enrichment with calcium and phosphate components [9], bacterial growth and social unacceptability [2]. These limitations have steered proposals to replace BCA with synthetic calcium phosphate system (CAPs) [10].
SC RI PT
It is thus recognized that CAPs, are potential solution to F- problem. Apart from mammalian bones, calcium precursor can be extracted from natural sources such as egg shells, crustacean shells, corals and plants [11]. Crustacean shells are of particular interest because of the abundance from sea food industry as process waste and environmental benefits associated with recycling them. It is estimated that 6-8 million tons of shells are produced annually in the world and most of it is left to rot in the sea or is dumped as landfill [12]. This work reports new approach to utilize crustacean shell. Synthesis and characterization of CAPs derived from prawn and crab shells is discussed. Defluoridation capacities were investigated to evaluate their potential in treatment of water contaminated with F-.
2. Experimental 2.1 Raw materials
N
U
Chemical reagents used were of analytical grade from sigma Aldrich. Samples of white prawns (Fenneropenaeus indicus) and swimming crabs (Portunus pelagicus) shells were collected from Dar es Salaam, Tanzania. Moisture content of clean and dry shells was determined by oven drying at 110 °C for 12 h and ash content using National Renewable Energy Laboratory (NREL) standard procedure LAP-005 [13].
M
A
2.2 Preparation of CAPs
TE
D
Dried shells were ground into 2 mm size, deproteinated using 0.25 M NaOH at 70 °C for 2 h, rinsed to neutral pH and oven dried at 50 °C for 12 h. Protein free shells were treated with 2 M HCl (solid to acid ratio 1:4 for prawns and 1:8 for crab shells) at room temperature (25±2 °C) for 24 h while stirring. The filtrate obtained was titrated with 0.1 M EDTA and calcon indicator (C21H14N2O7S.2H2O) to determine Ca2+ concentration. For comparison, filtrate was also precipitated using 0.5M Na2CO3. All the CAPs were synthesized following Equation 1 and procedure described in [14], without calcination.
EP
10Ca2+(aq) + 6PO43-(aq) +2OH- (aq) = Ca10(PO4)6(OH)2 (s)
( 1)
A
CC
Synthetic hydroxyapatite (CCHA) was prepared by reacting 250 ml of 1.09 M Ca(NO3)2.4H2O and equal volume 0.653 M (NH4)2HPO4, in a 1000 ml beaker under vigorous stirring, pH was adjusted to 10.4 using 28% ammonia solution and stirring continued at 40 °C for 4 h. The mixture was left to age at room temperature for 24 h while stirring. Precipitate formed was filtered, rinsed with distilled water to neutral pH followed by methanol rinse, oven dried at 80 °C for 24 h and ground into powder. Crustacean derived CAPs were synthesized using same general procedure; For CHA, 300 ml CaCl2 liquor (crab shell acid extract) with Ca2+ concentration of 43.69 g/l (1.09 M) was reacted with 300 ml of 0.653 M (NH4)2HPO4. PHA was synthesized using 300 ml CaCl2 liquor from prawns shell with Ca2+ concentration of 27.65 g/l (0.690 M) and 300 ml of 0.416 M (NH4)2HPO4.
2.3 Material characterization
SC RI PT
Chemical composition of CAPs was evaluated using energy dispersive X-ray (EDS) spectroscopy (Shimadzu EDX-800HS). XRD patterns were determined using Bruker D2 PHASER bench-top diffractometer with monochromatized CuKα radiation (λ= 0.15406 nm) over 2θ range of 3 -75 °, scanning speed of 3° per minute and step size of 0.02°. Functional groups analysis was conducted by FT-IR using Tensor 27 spectrometer fitted with a high-throughput screening device (HTS-XT). Tests were conducted in absorbance mode in spectral range 4000- 400 cm-1. 2.4 Batch adsorption tests
M
A
N
U
Simulated water (1000 mg/L) was prepared by dissolving 2.21 g of NaF in 1 L of distilled water. Dilution were made to obtain different concentrations of F-. Field water samples were collected from Ngarenyanyuki ward (S03°10.931’E036°51.677’), Arusha, Tanzania. Physical parameters: Electrocuductivity (EC), dissolved oxygen (DO), oxidation reduction potential (ORP), total dissolved solids (TDS) and temperature were measured using Hanna HI 9829 Multiparameter. Fconcentration was determined using ion selective electrode (ISE) connected to a Mettler Toledo seven compact pH/Ion S220 meter. Fluorinated water was mixed with total ionic strength buffer (TISAB) ratio of 1:1 before F- measurements. TISAB preparation was as follows; 57.0 ml glacial acetic acid, 58.0 g NaCl, 4 g cyclohexanediaminetetraacetic acid (CDTA) and approximately 150 ml of 6 M NaOH to adjust pH to 5.3- 5.5 in 1L. Orion Star A211 meter was used to record pH. Microbes were determined by plate count on agar media incubated at 44.5 °C for 24 h. Levels of phosphate and color were determined using HACH DR/2800 spectrophotometer.
EP
TE
D
Batch adsorption experiments were conducted using 0.25 g of adsorbent/50 ml of solution. BCA was included in field water tests for comparison. Mixtures were shaken in 250 ml plastic bottles on Retch AS 200 shaker with amplitude of 70 for 12 h. Final pH and residue F- was determined and adsorption capacity calculated using Equation 2. Effect of contact time, pH and temperature on F- removal was determined.
(C0 Ce )V m
CC
qe
(2)
A
qe = F- adsorbed at equilibrium (mg/g), C0 = Initial F- concentration (mg/L), Ce = F- concentration at equilibrium (mg/L), V= Volume (L), m = mass of adsorbent (g). 3. Results and Discussion 3.1 Preparation of CAPs Mineral content, Ca2+ concentration in acid liquors and yields of adsorbents are shown in Table 1. Theoretical yields were calculated following Equation 1. Molarity of liquors was calculated from Ca2+ concentration obtained by EDTA complexation. Mineral content by precipitation was
slightly higher than that obtained by EDTA method. This was attributed to presence of other cations in the acid liquor which co-precipitated with Ca2+. This observation is supported by presence of Si, Sr and Mn in EDS data (Table 2). Table 1: Mineral content, Ca2+ concentration in acid liquors and yields of CAPs. Prawns shell 7.0 34.0 34.5 32.6 27.7 300 0.69 17.5 20.9 83.8
Crabs shell 7.0 68.5 68.8 46.8 43.7 300 1.09 32.1 32.8 97.7
U
SC RI PT
Synthetic Ca2+ 250 1.09 27.2 27.7 99.6
Parameter Moisture (%) Ash (%) Mineral content by acid dissolution (%) Mineral content by CO32- precipitation (g/l) Ca2+ by EDTA complexation (g/l) Volume of Ca2+ solution (ml) Molarity of Ca2+ solution (M) Yield of CAPs Experimental (g) Yield of CAPs Theoretical (g) Efficiency (%)
N
3.2 Characterization of CAPs
M
A
Chemical composition of the CAPs by EDS analysis are presented in Table 2. Traces K2O and SrO were higher in PHA than in CHA while SiO2 and MnO were only identified CHA. Table 2: Chemical composition of CAPs (wt%) P2O5 29.2 30.8 35.7
D
CaO 70.8 65.0 60.1
K2 O 2.5 1.8
SiO2 1.5
SrO 1.6 0.6
MnO 0.3
EP
TE
Adsorbent CCHA PHA CHA
A
CC
XRD patterns are shown in Fig. 1. Diffraction phases of CCHA and PHA were identified by orientations 002, 211, 202 and 213 occuring at 2θ values 25.9, 32.0 34.1 and 49.5 respectively. Calculated lattice parametrs a = b = 9.434 and c = 6.920 corresponds to hexagonal phases of hydroxyapatite (HAP), JCPDS 09-0432 (a = b = 9.432, c = 6.881).
SC RI PT U N A
M
Fig 1: XRD patterns of CCHA, PHA and CHA with main peak orientations assigned
A
CC
EP
TE
D
XRD pattern of CHA revealed presence of mostly monoclinic structure with main phase orientation 020, 12-1, 14-1 at 2θ values 11.6, 20.9 and 29.3 respectively and lattice parameter, b = 15.32. The phases corresponds to brushite (CaHPO4.2H2O) peaks with lattice parameter b = 15.18, JCPDS 72-0713. HAP peaks (002, 211 and 202) were also identified but to a lesser extent. Presence of brushite in HAP crystals have been identified as factor of aging time, in which after 24 h all brushite and other calcium phosphate should convert to more stable HAP [15]. This does not however, explain the presence of major brushite phases in CHA, which was synthesized using the same reaction conditions of time and temperature as CCHA and PHA. Pattern of corresponds to that of HAP synthesized using egg shells [16], while CHA pattern was simillar to that reported for brushite crystals prepared by single diffusion gel technique [17]. Average crystallite sizes were calculated using Scherrer’s relationship; L=0.94λ/βcosθ. Peak 002 was used to calculate CCHA and PHA sizes, giving 31 nm and 15 nm respectively. Value for CCHA compare with that reported by Rusu (2005), L = 34 nm [18]. CHA gave avarage crstallize size, L = 47.9 nm using 12-1 principle peak. The three sample thus produced nano-crytalline sizes as evidenced in Table 3.
Table: 3 Selected XRD peaks analysis data (CuKα1λ = 0.15406 nm, β = 0.94) 2θ peak maxi (°) CCHA 25.80
D spacing (Å)
hkl
FWHM (°)
Crystallite size (nm)
3.46
002
0.2778
30.65
3.08 2.81 2.56 1.95
210 211 310 222
0.4690 1.2738 0.8878 0.5026
18.28 6.78 9.95 18.19
3.44 3.09 2.79 2.28 1.94
002 210 211 310 222
0.5877 2.4484 1.8960 1.6613 0.9108
14.90 3.50 4.56 5.21 9.93
7.61 4.25 3.05 2.81 2.17
020 12-1 210 202 15-2
0.1591 0.1761 0.2097 0.4377 0.2300
SC RI PT
28.90 32.00 39.88 46.49 PHA 25.89 28.85 32.06 39.56 46.80 CHA 11.62 20.90 29.23 34.12 41.50
52.45 47.93 40.91 19.84 38.59
A
CC
EP
TE
D
M
A
N
U
FT-IR spectra determined using diffuse reflectance ZnSe prisim are given in Fig. 2. Frequencies were assigned according to literature values [16, 19]. Modes related to structural OH were observed in two regions; 620- 635 cm-1 and 3480- 3570 cm-1 while intense P=O stretching bands associated with PO43- were seen at 1160-1204 cm-1, Table 4.
Fig 2: FT-IR spectra of CAPs adsorbents.
Spectra of CHA and PHA after defluoridation of groundwater with initial F- of 29.4 mg/L are compared in Fig. 3. Slight difference was observed in spent PHA (FPHA); split in 630 cm-1 peak forming two peaks, one to longer wavelength (620 cm-1) and another to shorted wavelength 635 cm-1 which is accompanied by decrease in absorption intensity (Fig. 3a). This change has been identified earlier in HAP by [20] but in their case the 630 cm-1 disappeared and new peak formed
at 738 cm-1. Intense peak at 1175 cm-1 also shifted slightly to longer wavelength with reduced intensity. Results indicate that some amount of F- was adsorbed by PHA as evidenced from batch adsorption tests.
A
N
U
SC RI PT
(a)
EP
TE
D
M
(b)
CC
Fig 3: Comparisons FT-IR spectra PHA and CHA before and after F- adsorption.
A
In CHA spectrum, decrease in absorption intensities was observed for most peaks. Peak 624 cm-1 shifted slightly to shorter wavelength and split into two, 627 and 639 cm-1. Broad intense peak at 1202 cm-1 became narrower and shifted to longer wavelength accompanied by increase in intensity (Fig. 3b). Bands at 1095 cm-1 and 3481 - 3552 cm-1 which are associated with HPO4 and structural OH stretching respectively, disappeared completely in FCHA. This indicate the OH and HPO4 groups were effectively replaced by F-. Similar observations has been reported in [21, 22] and is also well supported from batch adsorption tests which show that CHA removed almost 95% of F- from the water sample.
Table 4: IR modes and observed frequencies for CAPs and fluorinated FPHA and FCHA Vibrational frequencies CCHA PHA 588 583 620 630 875 877 963 963 1087 1088 1169 1175 1422 1420 1651 1651 3770 -
FPHA 583 635 873 963 1089 1169 1424 1655 -
3.3 Batch absorption test
CHA 584 624 875 1004 1095 1202 1655 3552
FCHA 586 638 875 961 1081 1169 1424 1655 -
SC RI PT
Assigment IR Mode PO4 bending Structural OH liberation P-O-P stretching PO4 stretching PO4 bending PO4 P=O stretching CO3- group H2O adsorbed Strucutal OH stretch
M
A
N
U
Field parameters of water samples are presented in Table 5. All samples contained F- above permissible limits and ORP below 400-650 mV which is deemed necessary for oxidation of contaminants like bacterial pathogens [23]. Samples F4-F7 failed quality test for EC and TDS. EC of 800-1000 µs/cm and TDS 500 mg/L are acceptable for drinking water. Surface water (F3, F5 and F7) failed quality standards for microbial and color requirements. Color values above 15 units are considered unacceptable but up to 85 units is found in natural drinking water [24]. Table 5: Water quality parameters of field samples ORP (mV)
DO (ppm)
EC (µS/cm)
TDS (ppm)
PO43(mg/)
7.6 7.5 8.3 8.5
92.2 91.6 -6.5 50.6
6.4 5.2 6.3 5.6
483 816 445 2883
242 408 223 1441
24.0 23.0
9.5 8.1
21.3 39.0
7.2 5.72
2062 3500
26.9
8.6
-190.9
3.06
4352
Color (pt Co)
Temp (°C)
F1 F2 F3 F4
Spring Spring River Shallow well Stream Shallow well Swamp
7.0 8.0 215.0 6.0
16.7 19.2 19.8 22.0
EP 3781.0
F(mg/L)
0.27 0.12 0.21 0.13
E.Coli (CFU/100 ml) 00 00 200 00
1032 1749
0.71 0.13
232 00
31.00 33.70
2178
0.65
17
70.10
4.20 6.48 18.70 29.40
A
F7
710.0 6.0
CC
F5 F6
pH
D
Source
TE
ID
Results of F- adsorption are shown in Table 6. Defluoridation capacities (DC) in low F- levels (410 mg/L) were almost same for all adsorbents. In simulated water of 10 mg/L DCs were 1.933, 1.882 and 1.852 mg/g for CHA, PHA and CCHA respectively. These results compare closely with that reported for nano-HAP, 1.845 mg/g [25]. Above 10 mg/L, a distinctive trend was observed; CHA > PHA > CCHA (Fig. 4a). CHA recorded the highest DC of 21 mg/g, this matches monolayer capacity reported for nano-HAP derived from waste phosphogypsum (19.7 mg/g) [26]. In field
water samples, highest DCs were 13.6, 8.5 and 2.6 for CHA, PHA and CCHA respectively (Fig. 4b). No data were found from literature on performance of CAPs in the field water with F- above 10 mg/L. From these data it can be concluded that CHA exhibits superior F- removal in both simulated and field water. PHA and CCHA adsorbent are appropriate for F- levels below 10 mg/L, with PHA performing much better than CCHA.
Sample
pH
F initial (mg/L)
SC RI PT
Table 6: Batch adsorption test for F- in different adsorbents using simulated and field water samples F- residue after 12 h treatment (pH final) CCHA PHA CHA
BCA*
0.607±0.0 (6.6)
-
0.362±0.0 (5.8)
-
0.977±0.0 (5.9)
-
Simulated water 10.27±0.13
1.01±0.2 (8.0)
0.86±0.2 6.8)
S2
3.80
50.56±0.42
21.67±0.8 (8.2)
4.95±0.2 (6.2)
S3
3.87
106.25±3.0
80.0±2.7 (7.7)
56.5±1.5 (6.4)
F1
7.60
4.20±00
0.638±0.3 (8.1)
U
5.90
0.845±0.4 (8.0)
0.275±0.0
(6.8)
1.9 ±0.0 (8.5)
F2
7.53
6.48±0.1
1.21±0.0
(8.2)
0.744±0.2
(8.0)
0.470±0.1
(7.0)
2.6±0.2 (8.5)
F3
8.30
18.70±00
8.66±0.2
(8.0)
2.83±0.0
(7.9)
0.221±0.0
(7.3)
9.0±0.4 (8.8)
F4
8.00
29.40±1.6
14.67±0.8 (8.4)
A
S1
6.02±0.2
(8.4)
0.582±0.2
(8.0)
17.4±0.3 (8.9)
F5
9.50
31.00±1.4
21.70±0.1 (9.3)
9.54±0.4
(9.1)
0.485±0.1
(7.8)
22.0±0.0 (9.4)
F6
8.10
33.70±1.5
18.87±0.7 (8.4)
6.60±0.4
(8.5)
0.569±0.2
(8.2)
21.9±0.2 (8.8)
F7
8.60
70.10±4.2
57.20±1.6
27.57±0.3
(8.6)
2.160±0.4
(8.1)
57.0±0.1 (8.8)
M
N
Field water
D
(8.8)
* BCA was in granular form as obtained from defluoridation center. It thus not included in the analysis CAPs
A
CC
EP
TE
synthesized in the laboratory.
SC RI PT U N
M
A
Fig 4: a) Efficiency of F- removal b) defluoridation capacities for different adsorbents.
EP
TE
D
Effect of adsorbent on water quality was tested for CHA using samples F4 and F5 (Table 7). Results show CHA removed yellow tint in F5 but increased color in F4, both samples remained clear. Ability of CHA to remove color can be explained by presence of MnO in the elemental composition, Table 2. MnO/hydroxides are known to promote degradation of complex organics such as humic acid, polychlorinated biphenyls and phenols [27]. The adsorbent did not have any effect on the microbial count and increased the level of phosphate slightly. No other major changes were observed in other parameters.
CC
Table 7: Water quality parameters before and after treatment with CHA Lab parameters
Color (pt Co)
pH
ORP (mV)
DO (ppm)
EC (µS/cm)
TDS ( ppm)
PO4 3(mg/)
F4
Before
6.00
8.4
149
5.2
2815
1440
After
12.0
7.8
84.1
4.14
2767
Before
710.00
9.6
124
4.4
After
109
7.5
81
4.5
A
ID
F5
F(mg/L)
0.1
E.Coli (CFU/ 100 ml) 00
1385
2.4
00
0.58
1904
959
0.7
232
31.00
1952
977
2.8
232
0.49
29.40
TE
D
M
A
N
U
SC RI PT
3.4 Effect of pH Influence of pH on adsorption of F- was studied using sample, F4. The initial pH was varied from 3-9 at room temperature for 180 min. Results (Fig 5) show efficiency of CCHA and PHA decreased considerably with increase in pH and best performance was achieved at pH = 3 with over 95% removal. This was attributed to protonation of adsorbent’s surface sites in acidic media, thus increasing affinity for F-, a phenomenon observed with most adsorbents [28-30]. Performance of CHA was not significantly affected by pH, narrow range (98-96%) was observed between pH 38.5.
EP
Fig 5: Effect of initial pH on the adsorption of F- on CAPs
CC
3.5 Adsorption isotherms
A
Adsorptive characteristics of the CAPs was tested at room temperature using two parameter isotherms; Langmuir, Freundlich, Dubinin-Radushkevich (D-R) and Temkin models in linear and non-linear forms as described in [31]. Nonlinear isotherms are shown in Fig. 6 and the corresponding data in Table 8. Linearized data are illustrated in Table 9. The general observation indicate that adsorption in simulated water is very different from field water in terms of adsorption capacity and correlation coefficients.
CCHA was best modeled by linear Langmuir Equation 3, with R2 > 0.98. Maximum adsorption determined experimentally for simulated and field water are 6.4 and 3.0 mg/g and corresponds well to calculated monolayer adsorption 6.2 and 3.2 mg/g respectively. This signifies a monolayer coverage onto a structurally homogeneous surface with finite number of identical sites [32]. This observation can be supported from XRD and EDX data which show only hydroxyapatite phases are present in the material.
SC RI PT
Ce 1 C e qe bK LCe qm
(3)
where Ce is F- concentration at equilibrium (mg/L), qe is adsorption capacity at equilibrium (mg/g), qm is monolayer adsorption capacity (mg/g) and KL is Langimuir constant (L/mg)
RT ln( KT Ce ) bT
(5)
A
qe
(4)
N
qe (qm ) Exp ( K ad 2 ) , ε = RT ln[1+1/Ce]
U
For PHA, simulated water was best modeled by D-R with R2>0.98 and field water by Temkin with R2= 0.93, Equations 4 and 5. The two isotherms describe empirical interaction of adsorbate and adsorbent, in which D-R assumes adsorption mechanism with Gaussian energy distribution on to heterogeneous surface while Temkin indicates a linear decrease in adsorption energy with coverage without the need for homogenous surface [31].This could indicate multimolecular layers coverage in nanosized adsorbent as seen with XRD crystallite size of 15 nm in Table 3.
D
M
where, Kad is adsorption equilibrium constant (mol2/kJ2), ε is Polanyi potential, E is mean free energy of adsorption (kJ/mol), R is gas constant (8.314J/mol/K), T is Temperature at 298K, KT is equilibrium constant (L/g) and b is constant related to heat of adsorption (J/mol).
CC
EP
TE
CHA was barely modeled by linear regression. Nonlinear D-R isotherm fairly represented adsorption in simulated water with R2= 0.82. Field water was equally modeled by all isotherms including Freundlich Equation 6 with R2 = 0.77-0.80. The difficulty in modelling CHA was attributed to multiphase structure which showed presence of mainly brushite and HAP in the crystal lattice. Presence of MnO which has catalytic properties on the surface may also have contributed to the heterogeneity resulting in different adsorption characteristics.
1/ n
qe K F Ce
(6)
A
where KF is Freundlich constant (mg/g) and n relates to adsorption intensity.
Table 8: Nonlinear Langmuir, Fruendlich, Temkin and Dubinin-Radushkevich isotherms constants for adsorption of F- onto CAPs Field water CCHA
PHA
CHA
6.176 0.325 0.900
11.707 0.428 0.910
2913.541 0.006 0.409
3.204 0.470 0.790
10.524 0.124 0.930
32.458 0.336 0.804
2.523 4.743 0.756
4.750 4.380 0.646
20.554 0.669 0.461
1.411 4.583 0.575
1.741 2.070 0.930
7.701 1.331 0.799
7.934 2.526 0.833
5.987 1.196 0.774
4.113 0.205 0.325
7.726 4.422 0.684
1.858 1.259 0.930
5.556 0.481 0.768
6.673 2.027 0.903
11.481 1.236 0.984
18.177 0.635 0.815
2.854 1.381 0.776
6.782 0.279 0.788
15.446 3.216 0.793
SC RI PT
CHA
A
CC
EP
TE
D
M
A
N
Freundlich KF (mg/L) N R2 Temkin KT (L/g) bT R2 Dubinin Radushkevich qm (mg/g) E (KJ/mol) R2
Simulated water CCHA PHA
U
Parameter/ model Langmuir qm (mg/g) KL (L/mg) R2
Fig 6: Non-linear adsorption isotherms of F- in different CAPs. a,b,c are simulated water in CCHA, PHA and CHA respectively while d,e and f are isotherm for field water in CCHA, PHA and CHA respectively.
Table 8 : Linear Langmuir, Fruendlich, Temkin and Dubinin-Radushkevich isotherms constants for adsorption of F- onto CAPs PHA
CHA
CCHA
PHA
CHA
5.490 0.015 0.986
10.360 0.015 0.994
0.001
2.650 0.021 0.985
11.060 0.076 0.883
0.008
2.054 3.662 0.872
3.151 2.648 0.725
0.095
1.000 2.925 0.800
1.201 1.526 0.894
0.445
7.933 0.977 0.833
5.987 2.071 0.774
0.324
7.726 0.560 0.684
1.858 1.968 0.930
5.610 5.151 0.768
5.174 0.361 1.176 0.820
10.270 0.477 1.024 0.991
0.04
2.692 0.269 1.363 0.851
5.302 0.435 1.073 0.858
0.376
SC RI PT
CCHA
U
Parameter/model Langmuir qm (mg/g) KL (L/mg) R2 Freundlich KF (mg/L) n R2 Temkin KT(L/g) bT R2 Dubinin- Radushkevich qm (mg/g) K (mol2/kJ2) E (KJ/mol) R2
M
A
N
Maximum qm calculated from D-R non-linear isotherm were compared with qm of other adsorbents to validate PHA and CHA, Table 10. It can be seen that our results are comparable to best performing HAP based adsorbents reported. High DC obtained in field water at natural pH is unique for our adsorbents, unfortunately no comparative data from literature as best performances are only based on simulated water and adjusted pH.
D
Table 9: Monolayer F- adsorption capacities on various adsorbent materials at 298 ±2 K Adsorbent
C0 max (mg/L)
qm (mg/g)
Reference
50.0
0.1 g/100 ml
19.7
[26]
200.0
5 mg/10 ml
25.8
[33]
10.0
0.25 g/50 ml
1.3
[25]
Nano sized -HAP
95.0
2 g/L
11.0
[34]
Mg/Ce/Mn/O-DE*
100.0
0.6 g/100 ml
13.3
[35]
Heat activated Dolomite
500.0
2 g/L
228.8
[36]
2.7
[30]
HAP-nano wires
CC
EP
Nano-HAP
TE
Phospogypsum HAP
Adsorbent loading
Bone char
A
PHA –prwan shell CAP
Brushite CHA-crab shell CAP
20.0
1 g/480 ml
106.0
11.5
70.0 **
0.25 g/50 ml
6.8
25.0
0.2 g/25 ml
6.6
106.0 70.0 * *
*DE= diatomaceous earth ** Field water sample
18.2 0.25 g/50 ml
15.4
This study
[37] This study
3.6 Effect of contact time and adsorption kinetics
D
M
A
N
U
SC RI PT
Kinetic results indicate rapid uptake within the first 10 min, with > 75% of the possible F- adsorbed. CCHA for instance removed 37% of the possible 42% maximum in the first 10 min. The process proceeded slowly for the rest of reaction time (Fig. 7).
TE
Fig 7: Variation of % F- removal with contact time on CAPs using groundwater sample (F4) with initial Fof 29.4 mg/L and initial pH = 8.5.
A
CC
EP
Experimental data on rate of F- adsorption was analyzed using four kinetic models, Table 11 [38, 39], results are shown in Table 12. Pseudo 2nd order model described the kinetic process more appropriately compared to other models (Fig. 8). Calculated qe values agree well with the experimental values and R2 > 0.99 shows that the model can be applied for the entire adsorption process, a phenomenon that confirms existence of chemisorption [40]. CHA was well modeled by all equations with R2 > 0.9. This implies that several mechanisms could be involved, hence the high DCs observed.
Table 10: Kinetic models equations Linear from 1st
Pseudo 2nd order Intra-particle diffusion Elovich
log (qe qt ) log qe
K1t 2.302
t 1 t 2 qt K s (qe ) qe
qt Ki pt1/ 2 C 1
ln( )
Kip=intra-particle diffusion (mg/(g.min1/2)) C= intercept 1
ln t
α= initial adsorption (mg/(g.min)) β= desorption constant (g/mg)
CC
EP
TE
D
M
A
N
U
qt
Parameters qe =F- adsorbed at equilibrium (mg/g) qt = F- adsorbed at time ‘t’ (mg/g) K1=Rate constant for pseudo 1st order (min-1) Ks=Rate constant for pseudo 2nd order (g/(mg.min)
SC RI PT
Model Pseudo order
A
Fig 8: Pseudo-2nd order kinetic plots for fluoride adsorption by CAPs.
Table 11: Kinetic parameters of CAPs PHA
CHA
1.00 2.22 0.64 0.29
0.84 3.88 -0.0039 0.18
0.98 5.40 1.97 0.94
2.26 2.22 0.180 1.00
3.65 3.88 0.06 0.99
1.91 0.03 0.88
2.65 0.08 0.55
9.64 0.96
3.37 0.55
U
SC RI PT
CCHA
5.62 5.40 0.03 1.00
-32.52 8.22 0.94 2.48 0.91
A
N
Parameter/model Pseudo 1st order qm cal (mg/g) qe exp (mg/g) K1 (L/mg) R2 Pseudo 2nd order qm cal (mg/g) qm exp (mg/g) Ks (g/mg.min) R2 Intra-particle diffusion C Kip (mg/g.min1/2) R2 Elovich Β (mg/(g.min)) R2
M
3.7 Effect of temperature and thermodynamic parameters
D
Influence of temperature on the adsorption was tested at 300 -327 K for 180 min (Fig. 9). Thermodynamic parameters; Gibbs energy (ΔG◦), enthalpy of reaction (ΔH0) and entropy (ΔS0) were calculated using equations 7-8 as described in [41], results shown in Table 13.
qe ) Ce
qe H 0 1 S 0 ) Ce R T R
EP
ln(
TE
G 0 RT ln(
(7)
(8)
A
CC
Positive ΔG0 indicated adsorption process was nonspontaneous for CCHA, PHA and spontaneous for CHA. The best temperature for adsorption for all the CAPs was identified as T= 307 K. Positive ΔH0 suggested the endothermic nature of adsorption, which is evidenced from slight increase in removal between 300-308 K. Positive ΔS0<<<<1 suggested no changes in randomness occurred at the solid/liquid interface during the adsorption process.
SC RI PT U N A
D
M
Fig 9: Variation of F- adsorption with temperature of groundwater sample (F4)
Table 12: Thermodynamic parameters of interaction of F- with CAPS
EP
CC
PHA
A
CHA
Temp (K) 302 307 314 325
TE
Adsorbent CCHA
302 307 314 325 302 307 314 325
ΔG (kJ mol-1) 4.75 4.42 5.19 4.32 2.10 1.46 1.65 1.92 -4.47 -6.81 -6.19 -7.84
ΔH (kJ mol-1)
ΔS (kJ mol-1)
6.61
0.009
1.99
0.0007
29.78
0.12
3.8 Elucidation of F- removal mechanism
SC RI PT
Mechanisms of F- uptake from solutions has been evaluated by several authors [22, 25, 26, 34]. Three main processes are highlighted; Ion exchange, electrostatic adsorption and precipitation. In this study, mechanism was elucidated from kinetic and FT-IR analysis. In the spectra (Fig. 3), shift or disappearance of bands related to OH and HPO4 bands accompanied by changes in intensity after adsorption was assumed to indicate F- substitution and formation of fluorapatite by CHA and PHA as follows; + 2F- (aq) =
Ca10(PO4)6(OH)2 (s) CaHPO4.2H2O (s) + 2F- (aq)
Ca10(PO4)6F2 (s) + 2OH-(aq)
(9)
= Ca10(PO4)6F2 (s) + 4HPO42- (aq) + 6H+(aq) + 20H2O (l) (10)
N
U
Kinetic tests showed equilibrium pH increased slightly in PHA treated water while it decreased in CHA during the rapid uptake. This was seen to confirm release of OH- and H+ into solution from the above reactions. Slow process was predicted to indicate electrostatic adsorption which increased with decrease in pH. It can thus be concluded that though ion exchange plays a significant role, it is not the only force driving the removal of F- from solution by the CAPs.
A
4. Conclusion
CC
EP
TE
D
M
CAPs derived from crustacean waste have been developed. The process of calcium isolation is versatile because it also allows recovery of other bioactive components like chitin which has valuable applications. Analytical characterization using XRD and FT-IR show that prawn shell produced HAP with nanocystallite size (~15 nm), while crab shell gave mostly brushite with average crystallite size of 50 nm. The adsorbents are found to have excellent DC in field water with F- concentration 5-70 mg/L. Highest capacity recorded are 13.5 mg/g and 8.5 mg/g for CHA and PHA respectively, which are higher than that of CCHA and bone char (2.6 mg/g). Adsorption was best described by pseudo 2nd order kinetics and fairly defined by non-linear D-R isotherm. Anion exchange was proposed as the main driving force in F- uptake followed by slow adsorption process. Overall results indicate that PHA is suitable for treating water with F- of 10 mg/L and below while CHA is efficient at higher levels. The CAPs are thus applicable for high F- water found in EARV. The crustacean derived adsorbents also reduce color in tinted water, which adds to quality of the treated water. Slight increase in phosphate level and failure to remove microbes were identified as limitations of the adsorbents.
A
Acknowledgement This project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement Number 690378. The authors are also thankful to Mwalimu Nyerere African Union Scholarship Scheme (MNAUSS) for the award of fellowship to Agatha Wagutu.
References
[1] R. Brunt, L. Vasak, J. Griffioen, Fluoride in groundwater: probability of occurrence of excessive concentration on global scale, IGRAC, 2004.
SC RI PT
[2] E. Dahi, Africa’s u-turn in defluoridation policy: from the Nalgonda technique to bone char, in: Fluoride 2016, pp. 401-416. [3] T. Rango, A. Vengosh, M. Jeuland, G.M. Whitford, R. Tekle-Haimanot, Biomarkers of chronic fluoride exposure in groundwater in a highly exposed population, Science of The Total Environment, 596-597 (2017) 1-11. [4] M. Amini, K. Mueller, K.C. Abbaspour, T. Rosenberg, M. Afyuni, K.N. Møller, M. Sarr, C.A. Johnson, Statistical Modeling of Global Geogenic Fluoride Contamination in Groundwaters, Environmental Science & Technology, 42 (2008) 3662-3668.
U
[5] J.K. Fawell, K. Bailey, Fluoride in drinking-water, World Health Organization, 2006.
A
N
[6] L.R. Brunson, D.A. Sabatini, Practical considerations, column studies and natural organic material competition for fluoride removal with bone char and aluminum amended materials in the Main Ethiopian Rift Valley, Science of The Total Environment, 488-489 (2014) 580-587.
M
[7] H. Mjengera, G. Mkongo, Appropriate deflouridation technology for use in flourotic areas in Tanzania, Physics and Chemistry of the Earth, Parts A/B/C, 28 (2003) 1097-1104.
D
[8] E. Dahi, Optimisation of bone char production using the standard defluoridation capacity procedure, International Society for Fluoride Research, Chiang Mai, (2015) 29-36.
EP
TE
[9] J. Albertus, H. Bregnhøj, M. Kongpun, Bone char quality and defluoridation capacity in contact precipitation, in: 3rd International workshop on fluorosis prevention and defluoridation of water, 2000, pp. 61-72.
CC
[10] M. Meta, Improving access to safe drinking water: prospection for low-fluoride sources. Brief report presenting main findings, in, Natural Environment Research Council, 2014. [11] M. Akram, R. Ahmed, I. Shakir, W.A.W. Ibrahim, R. Hussain, Extracting hydroxyapatite and its precursors from natural resources, Journal of materials science, 49 (2014) 1461-1475.
A
[12] N. Yan, X. Chen, Don't waste seafood waste: Turning cast-off shells into nitrogen-rich chemicals would benefit economies and the environment, Nature, 524 (2015) 155-158. [13] A. Sluiter, B. Hames, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton, Determination of ash in biomass: LAP-005 NREL analytical procedure, National Renewable Energy Laboratory, Golden, (2004).
[14] F. Chen, Z.-C. Wang, C.-J. Lin, Preparation and characterization of nano-sized hydroxyapatite particles and hydroxyapatite/chitosan nano-composite for use in biomedical materials, Materials letters, 57 (2002) 858-861.
SC RI PT
[15] H.N. M. Rusu, M. Wilke, B. Tiersch, P. Fratzl and G. Peter, Size-controlled hydroxyapatite nanoparticles as self-organized organic-inorganic composite materials, Biomaterials, 26 (2005) 5414-5426. [16] V. Kashkarov, D. Goloshchapov, A. Rumyantseva, P. Seredin, E. Domashevskaya, I. Spivakova, B. Shumilovich, X-ray diffraction and IR spectroscopy investigation of synthesized and biogenic nanocrystalline hydroxyapatite, Journal of Surface Investigation: X-ray, Synchrotron and Neutron Techniques, 5 (2011) 1162-1167. [17] K. Rajendran, C. Dale Keefe, Growth and characterization of calcium hydrogen phosphate dihydrate crystals from single diffusion gel technique, Crystal Research and Technology, 45 (2010) 939-945.
N
U
[18] V. Rusu, C. Ng, M. Wilke, B. Tiersch, P. Fratzl, M. Peter, Size-controlled hydroxyapatite nanoparticles as self-organized organic?inorganic composite materials, Biomaterials, 26 (2005) 5414-5426.
M
A
[19] R. Panda, M. Hsieh, R. Chung, T. Chin, FTIR, XRD, SEM and solid state NMR investigations of carbonate-containing hydroxyapatite nano-particles synthesized by hydroxide-gel technique, Journal of Physics and Chemistry of Solids, 64 (2003) 193-199.
D
[20] R. Ebrahimi-Kahrizsangi, B. Nasiri-Tabrizi, A. Chami, Characterization of single-crystal fluorapatite nanoparticles synthesized via mechanochemical method, Particuology, 9 (2011) 537544.
EP
TE
[21] H.E. Namin, H. Hashemipour, M. Ranjbar, Effect of aging and calcination on morphology and properties of synthesized nanocrystalline TiO 2, International Journal of Modern Physics B, 22 (2008) 3210-3215. [22] M. Tafu, T. Chohji, Reaction between calcium phosphate and fluoride in phosphogypsum, Journal of the European Ceramic Society, 26 (2006) 767-770.
A
CC
[23] V. Goncharuk, V. Bagrii, L. Mel’nik, R. Chebotareva, S.Y. Bashtan, The use of redox potential in water treatment processes, Journal of Water Chemistry and Technology, 32 (2010) 19. [24] WHO, Guidelines for Drinking-water Quality, in, 2011. [25] C.S. Sundaram, N. Viswanathan, S. Meenakshi, Defluoridation chemistry of synthetic hydroxyapatite at nano scale: equilibrium and kinetic studies, Journal of hazardous materials, 155 (2008) 206-215.
[26] D. Zhang, H. Luo, L. Zheng, K. Wang, H. Li, Y. Wang, H. Feng, Utilization of waste phosphogypsum to prepare hydroxyapatite nanoparticles and its application towards removal of fluoride from aqueous solution, Journal of hazardous materials, 241 (2012) 418-426.
SC RI PT
[27] B.M. Tebo, J.R. Bargar, B.G. Clement, G.J. Dick, K.J. Murray, D. Parker, R. Verity, S.M. Webb, Biogenic manganese oxides: properties and mechanisms of formation, Annu. Rev. Earth Planet. Sci., 32 (2004) 287-328. [28] Y. Tang, X. Guan, J. Wang, N. Gao, M.R. McPhail, C.C. Chusuei, Fluoride adsorption onto granular ferric hydroxide: Effects of ionic strength, pH, surface loading, and major co-existing anions, Journal of hazardous materials, 171 (2009) 774-779. [29] G. Alagumuthu, M. Rajan, Equilibrium and kinetics of adsorption of fluoride onto zirconium impregnated cashew nut shell carbon, Chemical Engineering Journal, 158 (2010) 451-457.
N
U
[30] N.A. Medellin-Castillo, R. Leyva-Ramos, R. Ocampo-Perez, R.F. Garcia de la Cruz, A. Aragon-Pina, J.M. Martinez-Rosales, R.M. Guerrero-Coronado, L. Fuentes-Rubio, Adsorption of fluoride from water solution on bone char, Industrial & Engineering Chemistry Research, 46 (2007) 9205-9212.
A
[31] K. Foo, B. Hameed, Insights into the modeling of adsorption isotherm systems, Chemical Engineering Journal, 156 (2010) 2-10.
M
[32] A. Günay, E. Arslankaya, I. Tosun, Lead removal from aqueous solution by natural and pretreated clinoptilolite: adsorption equilibrium and kinetics, Journal of hazardous materials, 146 (2007) 362-371.
TE
D
[33] J. He, K. Zhang, S. Wu, X. Cai, K. Chen, Y. Li, B. Sun, Y. Jia, F. Meng, Z. Jin, Performance of novel hydroxyapatite nanowires in treatment of fluoride contaminated water, Journal of hazardous materials, 303 (2016) 119-130.
EP
[34] V. Sternitzke, R. Kaegi, J.N. Audinot, E. Lewin, J.G. Hering, C.A. Johnson, Uptake of fluoride from aqueous solution on nano-sized hydroxyapatite: examination of a fluoridated surface layer, Environ Sci Technol, 46 (2012) 802-809.
CC
[35] W.M. Gitari, A.A. Izuagie, J.R. Gumbo, Synthesis, characterization and batch assessment of groundwater fluoride removal capacity of trimetal Mg/Ce/Mn oxide-modified diatomaceous earth, Arabian Journal of Chemistry, (2017).
A
[36] V.C.a.S. Prasad, Rapid removal of fluoride from aqueous media using activated dolomite, Analytical Methods, 7 (2015) 8304-8312. [37] M. Mourabet, H. El Boujaady, A. El Rhilassi, H. Ramdane, M. Bennani-Ziatni, R. El Hamri, A. Taitai, Defluoridation of water using Brushite: Equilibrium, kinetic and thermodynamic studies, Desalination, 278 (2011) 1-9.
[38] E. Demirbas, M. Kobya, E. Senturk, T. Ozkan, Adsorption kinetics for the removal of chromium (VI) from aqueous solutions on the activated carbons prepared from agricultural wastes, Water Sa, 30 (2004) 533-539.
SC RI PT
[39] S. Ayoob, A.K. Gupta, P.B. Bhakat, V.T. Bhat, Investigations on the kinetics and mechanisms of sorptive removal of fluoride from water using alumina cement granules, Chemical Engineering Journal, 140 (2008) 6-14. [40] H. Qiu, L. Lv, B.-c. Pan, Q.-j. Zhang, W.-m. Zhang, Q.-x. Zhang, Critical review in adsorption kinetic models, Journal of Zhejiang University-SCIENCE A, 10 (2009) 716-724.
A
CC
EP
TE
D
M
A
N
U
[41] I. Anastopoulos, G.Z. Kyzas, Are the thermodynamic parameters correctly estimated in liquid-phase adsorption phenomena?, Journal of Molecular Liquids, 218 (2016) 174-185.