Environ. Sci. Technol. 2006, 40, 7054-7059
Cadmium and Lead Ion Capture with Three Dimensionally Ordered Macroporous Hydroxyapatite M A D H A V I S R I N I V A S A N , * ,† CRISTIANO FERRARIS,‡ AND TIM WHITE† School of Materials Science and Engineering, Nanyang Technological University, Blk N4.1-01.-30, Nanyang Avenue, Singapore 639798, and Department Histoire de la Terre, Museum National d’Histoire Naturelle, Unite´ Scientifique de Mine´ralogie CNRS 7160, 61 rue Buffon, F-75005 Paris, France
The capability of three dimensionally ordered macroporous (3DOM) hydroxyapatite, Ca10(PO4)6(OH)2 (HAp), to capture cadmium and lead ions from their respective salt solutions was studied as a function of temperature. Synthesis of 3DOM material was achieved by colloidal crystal templating of polystyrene spheres (1 µm diameter) using calcium nitrate (Ca(NO3)2) and orthophosphoric acid (H3PO4) as precursors. The macroporous product consisted primarily of HAp (>80% depending on the sintering temperature) together with amorphous calcium phosphate. The sorption ability of 3DOM material to Cd/Pb ion was benchmarked against HAp powder prepared via the same route without the template. On the basis of quantitative X-ray diffraction (XRD) and analytical transmission electron microscopy (ATEM) 3DOM HAp demonstrated a higher uptake of cadmium, viz. x ) 0.71 in Ca10-xCdx(PO4)6(OH)2 than nonporous HAp (x ) 0.42). The incorporation of Cd was homogeneous in the 3DOM HAp crystals (as compared to the powder) leading to a decrease in lattice parameters as Cd2+ has a smaller ionic radius compared to Ca2+. A preference for Cd to enter the CaII tunnel site of HAp was consistent with this being the readily exchangeable site. The leadbearing solution acted to collapse the macropores through the rapid crystallization of pyromorphite (Pb10(PO4)6(OH)2) via a dissolution-precipitation mechanism, possibly promoted by the amorphous component, that overwhelmed HAp ion exchange. The rapid crystallochemical incorporation of Cd and fixation of Pb by 3DOM HAp demonstrates the potential of thin-walled porous structures for the treatment of contaminated waters.
Introduction Contamination of surface and groundwaters by heavy metals including cadmium, lead, and mercury can result in wide dispersal of these pollutants and their accumulation in biota, leading to a range of illnesses including osteoporosis (lead) and itaıˆ-itaıˆ disease (cadmium). Removal of metals from water and wastewater commonly involves precipitation followed by a chemically or physically induced transformation to nonlabile forms. This durable matrix to which the metals are transferred should be stable over a range of conditions, especially with respect to pH and temperature. Apatite, with * Corresponding author e-mail:
[email protected]. † Nanyang Technological University. ‡ Unite ´ Scientifique de Mine´ralogie CNRS 7160. 7054
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 22, 2006
the prototypical chemistry Ca10(PO4)6(OH,F,Cl)2, is a promising candidate for heavy metal stabilization (1-9), as its crystal structure is tolerant to many ionic substitutions and complete replacement of Ca2+ by Ba2+, Sr2+, Cd2+, and Pb2+ and P5+ by V5+, Cr5+, and As5+ is possible. Hydroxyapatite (HAp, Ca10(PO4)6(OH)2) has been extensively studied because of its relatively high sorption capacity, low solubility especially in basic and neutral pH, stability under reducing and oxidizing conditions, ready availability, and low cost. Numerous studies of heavy metal exchange in powdered forms of HAp, especially with respect to lead and cadmium have been reported (1-8, 10-16), although ‘ionexchange’ sensu stricto is not often demonstrated, with the experimental data rather implying that dissolution-precipitation is operative. Indeed, while the stabilization of many metals using HAp (1-16) has been confirmed, the precise mechanisms of fixation remain poorly understood. In general, ion-exchange is favored in crystal structures that consist of an adaptable framework, penetrated by channels along which ions migrate, with the framework adjusting to facilitate the passage of species through the tunnels. Members of the apatite structural family, which commonly display P63/m hexagonal symmetry and have the ideal crystallochemical formula [AI4][AII6][(BO4)6]X2, can be described as one-dimensional ion conductors where the [AI4][(BO4)6] framework is constructed of face-sharing AIO6 metaprism columns linked through corner-connection to isolated BO4 tetrahedra (17). The resulting channels contain AII and X ions that are potentially ion conducting and exchangeable (Figure 1). Framework flexibility arises by twisting of the triangular faces of the AIO6 metaprisms through an angle (φ), typically over the range 5-25°(18), that tunes tunnel size in response to its filling and chemistry. As apatite channels lie parallel to [001] and metaprism twisting occurs in (001), it follows that variation in tunnel content will result in large changes in the a lattice constant, while the c cell edge will be relatively invariant. Clearly, the surface area and crystal morphology of HAp will regulate ion sorption, and various morphologies have been synthesized to optimize metal capture, including fibrous and spherical carbonate/magnesium containing apatite agglomerates, which fixed Pb2+ >> Cd2+ > Ni2++ (19). Porous HAp based ceramic filters have been fabricated for the selective removal of various metal ions (Cr3+, Co2+, Ni2+, Al3+, Cu2+, Pb2+, Fe3+) (20), and Furuta et al. (21, 22) reported complete removal of Pb2+ ions from aqueous solutions (200 ppm < [Pb2+] < 400 ppm) using porous HAp monoliths that were synthesized from gypsum waste and possessed a bimodal pore distribution (0.1-0.3 and 10-20 µm). A nanocomposite of zeolite (type-A) covered with acicular HAp has also been studied for the capture of toxic and radioactive ions (23). All these methods have in common the desire to maximize the surface available for sorption and possible ionexchange while enhancing the accessibility of contaminants to those surfaces. Templating using polymer colloidal crystals is a wellestablished method to synthesize macropores of uniform size (100 nm-1 µm diameter) which are three dimensionally ordered (3DOM). This approach is applicable to the preparation of oxides to be used in photonics, optoelectronics and catalysis, and as a bio- and magnetic materials (24-30). However, 3DOM architectures for environmental remediation are yet to be extensively canvassed. In one case, thiol functionalized 3DOM titania and zirconia were prepared as adsorbents for mercury(II) and lead(II) ions with the thiol groups attached to the porous structure to capture these 10.1021/es060972s CCC: $33.50
2006 American Chemical Society Published on Web 10/18/2006
Experimental Methods
FIGURE 1. Polyhedral representations of [AI4][AII6][(BO4)6]X2 apatite emphasizing the AIO6 metaprisms and BO4 tetrahedra. (a) The [001] projection showing the topological relationship between the polyhedra and exchangeable A(II) and X ions. By adjusting the metaprism twist angle O the volume of the tunnels adjust to accommodate different tunnel contents, either in terms of ionic size or occupancy. (b) The [001] projection accentuating the isolated BO4 tetrahedra and their connection to the metaprisms. (c) Clinographic projection of a single one-dimensional channel. metals. Although a high level of thiol functionalization was achieved, less than stoichiometric adsorption took place in these hybrid materials due to limited accessibility to the thiol groups on the interior of the macroporous walls (29). To circumvent this issue, a novel 3DOM HAp was recently reported, in which the walls themselves can directly capture heavy metals by surface sorption/ion exchange (30). The specific surface areas of 3DOM materials are typically lower than their mesoporous counterparts, but the large interconnected cavities enhance accessibility to the active sites and the short diffusion path of the thin pore walls accelerates ion-exchange (24-30). The present investigation was undertaken to determine the ion sorption capacity of 3DOM materials, consisting predominantly of HAp and lesser amounts of amorphous calcium phosphates, when contacted with aqueous solutions of cadmium or lead, and compare the performance with nonporous counterparts synthesized under similar conditions. Quantitative X-ray diffraction and electron microscopic techniques were used to examine modifications in the morphology and crystal structure of the HAp component during Cd/Pb exposure and study the mechanism of ion exchange. The competing role of amorphous phosphates during sorption and fixation is also addressed.
Synthesis of 3DOM HAp. The synthesis of 3DOM Ca10(PO4)6(OH)2 (MHAp1) has been described earlier (30). Briefly, monodispersed polystyrene latex microsphere suspensions of mean diameter 1.0 µm (Duke Scientific Corporation) having a size distribution e3% were spun for 48 h at 900 rpm and then air-dried to obtain centimeter-scale well-ordered close packed polymer templates (Figure S1a of the Supporting Information). The precursor solution, prepared by mixing stoichiometric amounts of 0.8 M H3PO4 and 1.64 M Ca(NO3)2‚ 4H2O, was forced into the template voids by vacuum infiltration. These infiltered templates were heated at 400 °C for 3 h, followed by 700 °C for 6 h under flowing oxygen. Nonporous powder hydroxyapatites, Ca10(PO4)6(OH)2 (PHAp6), were synthesized using the same conditions sans templating. Sorption Experiments. The macroporous (MHAp1) and powder (PHAp6) apatite were immersed in a cadmium nitrate solution containing 8 × 10-3 mol/L of cadmium at RT (MHAp2, PHAp7), 50 °C (MHAp3, PHAp8), and 80 °C (MHAp4, PHAp9). After 100 h the MHAp and PHAp were recovered, washed thoroughly with distilled water, and dried at 100 °C. Similarly, MHAp1 was contacted with a lead nitrate solution (1.25 × 10-2 mol/L) at RT for 100 h (MHAp5). For comparative purposes, the concentrations of the metal solutions were similar to those used in earlier studies (2, 8, 10, 46). Characterization Methodology. Secondary electron images (SEI) were obtained using a JEOL JSM-5410LV scanning microscope operating at an accelerating voltage of 15 kV. Powder X-ray diffraction (XRD) patterns of the pristine macroporous apatite and after immersion in the Cd/Pb solutions were collected using a Siemens D5005 diffractometer with CuKR radiation, step-scanned over the 2θ range 10-80° at intervals of 0.02° with an acquisition time of 14 s per step leading to a total collection time of 13.6 h. Scanning transmission and analytical electron microscopy (STEM and AEM) were performed at 300 kV with a JEOL JEM 3010 electron microscope equipped with double-tilt holder, LaB6 cathode, LINK ISIS energy dispersive spectroscopy X-ray spectrometer (EDS), and a backscattered electron-imaging device (BEI) coupled with the STEM. AEM was performed in TEM-EDS mode, with a live counting time of 50 s and a nominal beam diameter of 25 nm. Recalculation and normalization of the AEM analyses were performed assuming the thin-film approximation, using experimental calibrations from silicate standards. AEM compositions were expressed based upon 26 (O, OH) atoms per formula unit (apfu) for the Ca10(PO4)6(OH)2 synthetic apatite. The samples examined by TEM-AEM were from the same batches used for X-ray and scanning electron microscope (SEM) analysis and prepared by gently crushing in an agate mortar under acetone with a few drops of the suspension deposited on a holey carbon film supported by a copper grid. The amounts of Cd or Pb present in the HAp crystals and precipitated phases were deduced by analytical electron microscopy (AEM). Quantitative X-ray Diffraction. Rietveld analysis of the X-ray diffraction patterns was carried out using the fundamental parameter procedure as implemented by TOPAS R (Version 2.1) to quantitatively extract phase content for the materials (31-33). The starting structure used in the refinement was that of Holly Springs hydroxyapatite which adopts P63/m symmetry with a ) 9.424 Å and c ) 6.879 Å (34). In each case, a background polynomial, scale factor, cell parameters, zero point correction, and site occupancy factors were refined, while atom positions and isotropic thermal parameters were fixed. Secondary phases were added as required using appropriate crystallographic models for β-Ca3(PO4)2 (β-TCP) (35), CaCO3 (calcite) (36), β-Ca2P2O7 (37), CdCO3 (otavite) (38), Cd2P2O7 (39), CaHPO4 (monetite) (40), PbCO3 (cerrusite) (41), and Pb10(PO4)6(OH)2 (pyromorphite) VOL. 40, NO. 22, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7055
FIGURE 2. Scanning transmission electron (STEM) images of (a) macroporous apatite (MHAp1), after immersion in cadmium solution for 100 h at (b) RT(MHAp2), (c) 50 °C (MHAp3), and (d) 80 °C (MHAp 4). (42). The weight percent (wt %) of each Bragg diffracting phase was calculated from the refined scale factors (43). Lattice constants and amorphous content were determined absolutely by adding respectively 15 wt % Si (44) and 20 wt % Al2O3 (45) NIST standards. Occupancy of the CaI and CaII sites in Cd sorbed HAp were refined independently with the total occupancy factor (Ca + Cd) constrained to unity for each site (i.e., full site occupation).
Results and Discussion Cadmium Fixation. As prepared 3DOM apatite (MHAp1) consisted of well-ordered macropores (>5 µm in extent) of uniform void size (0.8 µm) (Figure S1b of the Supporting Information, Figure 2a ). The average wall thickness was ∼25 nm with the apatite crystals 50-60 nm in length and 20 nm wide (see Figure 3b in ref 30) elongated along [001]. The macroporous architecture remained intact (Figure 2b) during cadmium salt solution treatment at room temperature (MHAp2); however, recrystallized components attached to the apatite walls were evident. Immersion in the cadmium solution at 50 °C (MHAp3) and 80 °C (MHAp4) lead to progressively greater recrystallization (Figure 2c,d) and partial collapse of the 3DOM structure. The quantity of cadmium ion sorbed/substituted for calcium, viz. x in Ca10-xCdx(PO4)6(OH)2, in the 3DOM and powder HAp deduced by ATEM and Rietveld analysis are summarized in Table 1. Crystal-to-crystal cadmium content in 3DOM apatite (MHAp2-MHAp4) was uniform at all temperatures. The value of x (averaged over 20-30 apatite grains) determined by ATEM gradually rose from 0.16 apfu at room-temperature reaction (MHAp2) to 0.86 apfu at 80 °C (MHAp4) due to increasingly rapid sorption/ion exchange kinetics. This observation was distinct from the powdered HAp (PHAp6-PHAp9) where ion incorporation in HAp was less homogeneous. Unlike the macroporous walls which contain regular, small apatite crystallites, the powder counterpart was composed of crystals ranging from 10-200 nm, with more extensive cadmium replacement (x = 0.86) seen in smaller crystals (10-20 nm) and negligible quantities (x = 0) found in larger crystals (100-200 nm). This result is broadly similar to the recent observations of Marchat et al. (46) on cadmium substitution in synthetic hydroxyapatite 7056
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 22, 2006
FIGURE 3. Representative diffraction patterns to illustrate the formation of secondary phases. X-ray diffraction profile (experimental (-), calculated (°), and difference) of cadmium ion treated (a) 3DOM at 50 °C (MHAp3) and (b) powder at 80 °C (PHAp9) along with NIST Si standard for accurate unit cell determination. Bragg reflection markers from top to bottom for (a) are CdCO3 (*), monetite (b), β-Ca3(PO4)2 (#), CaCO3 (@), Ca2P2O7 (+), Cd2P2O7 (O), and HAp and for (b) are HAp, CdCO3 (*), Si, and Ca2P2O7 (+). The main lines of the secondary phases are marked on the pattern. powders. Rietveld refinement showed that x increased gradually for 3DOM apatite from 0.29 (MHAP2) to 0.71 (MHAP4), whereas for the powder counterpart x was always 0.35-0.42 whatever the reaction temperature. Thus, there is reasonable concurrence in the results from the two techniques when it is considered that ATEM analysis were obtained from individual exchanged crystals, while XRD is averaging over tens of thousands of crystals. It is concluded that the total cadmium sorbed and captured by the HAp component of the 3DOM material is greater and more homogeneous compared to the powder. This can be attributed to the enhanced mobility of the Cd salt solution through the open porous framework and nearly complete accessibility of apatite crystals for ion sorption. The maximum uptake of Cd (x ∼ 0.86) obtained in this study is comparable to earlier reports on Cd-sorption/ion exchange in powder apatites (2, 8, 10, 47). However, previous investigations used sodium-doped HAp rather than Ca-pure HAp, and direct comparison may not be warranted. Role of Amorphous Calcium Phosphates. Representative partial Rietveld profiles of cadmium treated 3DOM (MHAp3 (20°-60° 2θ)) and powder material (PHAp9 (20°-60° 2θ), with NIST Si Standard) are shown in Figures 3. Full XRD patterns (10-140° 2θ) of the pristine powder HAP (PHAp6 with NIST Al2O3 and Si) and Cd-treated powder (PHAp9 with NIST Si) are given in Figure S2 of the Supporting Information. The crystalline component of the as-prepared macroporous material was near single-phase HAp (98%) (30). Treatment with cadmium nitrate solution lead to the formation of secondary phases including Ca2P2O7, CaCO3, CaHPO4, β-Ca3(PO4)2, CdCO3, and Cd2P2O7. The strongest reflections corresponding to these secondary phases are labeled in Figure 3, and the weight percent is given in Table S1 of the Supporting Information. The fitting of each phase of MHAp2 (25°-35° 2θ) is shown in greater detail in Figure S3 of the Supporting Information. Their appearance is consistent with the STEM and EDS results where recrystallized components were seen
TABLE 1. Selected Crystallographic Parameters before and after Cd/Pb Solution Treatment of Macroporous (M-HAp) and Nonporous (P-HAp) Apatites. x in Ca10-xCdx(PO4)6(OH)2
lattice parameters sample
a (Å)
c (Å)
c/a
cell volume (Å3)
Ca10(PO4)6(OH)2 (34) MHAp1 MHAp2 MHAp3 MHAp4 MHAp5 PHAp6 PHAp7 PHAp8 PHAp9 Cd10(PO4)6(OH)2b
9.424(4)a 9.4206(2) 9.3815(5) 9.3815(7) 9.3812(6) 9.4283(7) 9.4201(2) 9.4278(2) 9.4201(2) 9.4198(1) 9.3350
6.879(4) 6.8890(1) 6.8941(4) 6.8917(6) 6.8904(6) 6.8881(8) 6.8825(2) 6.8876(2) 6.8824(2) 6.8819(1) 6.6640
0.730 0.731 0.735 0.735 0.735 0.731 0.731 0.731 0.731 0.731 0.714
529.1 529.5(2) 525.5(1) 525.3(1) 525.2(1) 530.3(1) 528.9(2) 530.2(1) 528.9(1) 528.8(1) 502.9
a
Errors are esds.
b
Cd2+ occupancies in Ca site
XRD
Ca(I) 4h site
Ca(II) 6f site
0.00 0.16 0.22 0.86
0.00 0.29 0.41 0.71
0.010 (4) 0.027 (6) 0.065 (7)
0.042 (4) 0.051 (5) 0.075 (5)
0.00 0.00-0.86 0.00-0.31 0.00-0.82
0.00 0.35 0.42 0.38
0.025 (3) 0.023 (4) 0.018 (2)
0.042 (2) 0.054 (4) 0.051 (2)
ATEM
crystallite size (nm)
Rbc (%)
60 63 61 57 57 106 113 111 114
9.4 4.7 8.4 8.0 4.2 4.5 3.6 4.3 3.6
Data for Cd-HAp provided for comparative purposes (55). c Rb - Bragg-R factor (31).
on the 3DOM walls. In the case of powder apatite, the weight percent of HAp was essentially unchanged after Cd-sorption with lesser amounts of CdCO3 and Ca2P2O7 forming. The exact mechanism of formation of these compounds is unclear; however, these are likely to be nonequilibrium phases which form from the amorphous phosphates present in 3DOM materials. The fact that greater quantities of these phases are seen in 3DOM material, as compared to powder, is consistent with amorphous content being decisive in the formation of these phases. A pure ion-exchange between Cd and Ca in HAp would not lead to secondary phase formation. However as noted previously, substantial amorphicity (∼14 wt %) is an inherent property of the 3DOM material as a consequence of the synthesis methodology (30) and is far higher than in the powder counterpart (