Environ. Sci. Technol. 2005, 39, 4533-4539
Organic Coprecipitates with Calcite: NMR Spectroscopic Evidence BRIAN L. PHILLIPS,* YOUNG J. LEE, AND RICHARD J. REEDER Center for Environmental Molecular Science, Department of Geosciences, State University of New York, Stony Brook, New York 11794-2100
Dissolved organic ligands are well known to interact strongly with the calcite surface, altering precipitation and dissolution rates, crystal morphology, and possibly the ability of calcite to sequester metal contaminants. We show, using NMR spectroscopic techniques, that some of the citrate molecules present in a solution of precipitating calcite are incorporated structurally into the calcite crystal. Calcite grown by a seeded constant-addition method contains approximately 1 wt % coprecipitated citrate and yields 13C{1H} cross-polarization magic-angle spinning NMR spectra that contain broad peaks for citrate plus a signal from carbonate. Results from 13C{1H} heteronuclear correlation NMR experiments show that citrate is located in close spatial proximity to carbonate groups. In addition, calcite/citrate coprecipitates contain about 0.4 wt % excess water, which is not present as fluid inclusions, and some of which occurs as rigid structural water. These results suggest that water and hydrogen-bonding interactions play a role in the interface between included organic molecules and the calcite host. Additional NMR data obtained for calcite coprecipitates of aspartic and glutamic acids suggest they are also incorporated structurally but at concentrations much lower than for citrate, whereas no evidence was found for phthalate incorporation.
Introduction Calcite is a common and dynamic component at near-surface conditions as well as in sediments. As a consequence, interactions at the calcite/water interface play a crucial role in modulating natural water chemistry and in determining the mobility of some metal contaminants. Dissolution and precipitation of calcite can affect uptake and release of many metal cations (1, 2). Dissolved organic matter (DOM) is also a common component of these systems, especially in surficial aquatic environments, and can affect the ability of calcite to sequester contaminants either by mediating metal partitioning between fluid and solid phases, through ligand/metal complexation, or by altering the nature of the calcite/fluid interface. That organic ligands modify the calcite/water interface is apparent from the generally strong sorption characteristics (3-7) and well-established observations that organic ligands can inhibit calcite precipitation (8-14), direct precipitation of metastable CaCO3 polymorphs (15-17), and affect calcite dissolution rates (18-22). Most of these studies have focused on molecules containing carboxylate func* Corresponding author phone: (631)-632-6853; fax: (631)-6326840; e-mail:
[email protected]. 10.1021/es048733x CCC: $30.25 Published on Web 05/17/2005
2005 American Chemical Society
tionality because it is believed generally that these processes involve binding of the organic molecules specifically to the Ca2+ on calcite surfaces. Much of the previous work in this area has been driven by interest in biomineralization (23) and prevention of carbonate scale formation in industrial processes (24). But, it is clear that natural inorganic processes can be affected as well. From these studies, it is apparent that interaction between organic acids and the calcite surface depends to a large degree on the detailed structure of the molecule, even for those molecules that interact primarily through carboxylate groups. For example, sorption of dicarboxylates is strongest for those forming five-membered chelate rings and citrate sorbs more strongly than tricarballylic acid (4). Amjad (8) found that benzenehexacarboxylic acid was effective at inhibiting calcite growth, whereas 1,3,5 benzenetricarboxylic acid had no effect. Although many aromatic and other ring-structured carboxylic acids effectively inhibit calcite growth, those with a linear backbone appear to have much smaller effects on the calcite growth rate (8, 10, 17). Recent studies using atomic force microscopy (25, 26) have shown that dissolved organic molecules are preferentially sorbed at the calcite surface, through their influence on spreading velocities for different steps. Sorption on different growth steps influences the morphology of the calcite particles during crystallization. These observations suggest that specific molecular configurations and compatibility with the calcite surface structure are important factors. Nevertheless, little is known about the incorporation of dissolved organic molecules, particularly the local environment of organic molecules in the calcite structure, beyond macroscopcoic observations, despite the wide occurrence of dissolved organic matter (DOM) at the mineral-water interface and of organic matter in naturally occurring calcite. Naturally occurring humic substances have been shown to effectively inhibit calcite growth in both laboratory (12, 14) and field (9, 27) settings and to inhibit proton-promoted dissolution (28), suggesting that DOM might play an important role in determining the reactivity of calcite in natural waters. Calcite interaction with the citrate molecule appears to present a specifically interesting case. Although citrate sorbs strongly to the calcite surface (4), Reddy and Hoch (10) found that it did not significantly influence calcite growth rates under seeded, constant-composition conditions. Other studies (11, 13) report significant calcite growth inhibition by citrate in unseeded, pH-drift experiments. Crystal growth in the presence of a strong surface sorbate suggests that the sorbate might be included within the crystal. Ueyama et al. (17) precipitated CaCO3 (both calcite and vaterite polymorphs) in the presence of carboxylate-containing polyamides and found that significant concentrations of the organic polymers remained after thorough rinsing. 13C{1H} cross-polarization magic-angle spinning (CP-MAS) NMR spectra of these calcite crystals contain signals from the organic material, suggesting that the organic molecules are bound within the calcite. It is not apparent, however, how the calcite structure accommodates such large impurities. Understanding the mechanisms by which dissolved organic molecules can be incorporated into the bulk structure of calcite during crystallization is essential for assessing the effects of organic material on long-term metal retention in calcite, and for predicting the effects of DOM on calcite reactivity and metal-ligand complexation on metal coprecipitation in calcite. In this study we report coprecipitation of citrate (Cit) in calcite crystals grown under seeded, constant-addition VOL. 39, NO. 12, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4533
TABLE 1. Calcite Samples Used for This Study organic concentration (wt %) sample
organic coprecipitate
synthesis medium
pH
mass (g)
total
overgrowth
H content (10-3 mol/g)
cc3101 cc3107 cc3109 cc3113 cc0387 cc4003
citrate citrate citrate none aspartate (13C, 15N) glutamate (13C, 15N)
H2O H2O, 13CO3 D2O, 13CO3 H2O, 13CO3 H2O H2O
8.3 8.3 8.4 8.25 8.3 8.3
0.75 0.76 0.74 0.77 0.76 0.78
n.a. 0.67a 0.76a n.a. 0.01b 0.003b
n.a. 1.10 1.13 n.a. 0.02 0.005
0.69 0.69 0.21 0.07 n.a. n.a.
a
Measured by ion chromatography.
b
Measured by HPLC; n.a.: not analyzed.
conditions in the presence of citric acid. We use solid-state NMR spectroscopy to demonstrate that the citrate is located within the calcite structure and to deduce some of the structural features of the calcite/citrate coprecipitate. In particular, we find that significant amounts of structural water accompany citrate in the structure. NMR spectra of calcite grown with glutamic (Glu) and aspartic (Asp) acids also show evidence for coprecipitation but at concentrations much lower than for citrate.
Materials and Methods Sample Preparation. A series of samples for this study was synthesized at room temperature using a modified version of the seeded constant-addition method (29, 30). Briefly, 500 mg of analytical reagent calcite (Alfa Aesar) with an average particle size of 5 µm (0.2 m2 g-1 Brunauer-Emmett-Teller (BET) surface area) was added to a reaction vessel containing 700 mL of an initial growth solution that was 0.007 M CaCl2, 0.007 M NaHCO3, and 0.050 M NaCl. This initial growth solution was slightly oversaturated with respect to calcite. The reaction vessel was maintained by addition of CaCl2 (0.100 M CaCl, 0.050 M NaCl) and Na2CO3 (0.100 M Na2CO3, 0.050 M NaCl) solutions from syringes at a constant rate of 150 µL/min. Air was continuously bubbled through the solutions, which were stirred throughout the synthesis. Previous experiments had shown that the system rapidly evolves to a steady-state solution composition with uniform conditions for calcite growth. Replicate samples of the suspensions were periodically collected, filtered (0.22 µm), and analyzed by DCP emission spectrophotometry for total Ca and flow injection analysis (FIA) for total carbonate. The detection limits are less than 5 ppb Ca for DCP and 0.1 mM for the carbonate FIA analyses; uncertainty for each is less than 1%. Citrate/calcite coprecipitates were prepared by addition of citric acid into both the CaCl2 syringe (0.5 mM citrate, pH ) 5.6) and the initial growth solution (0.05 mM citrate). A similar method was applied using 13C-labeled phthalate (13C2, carboxyl only) at 0.02 mM. Solutions for citrate coprecipitation in D2O were prepared with 99.9% D2O (Aldrich) for the growth solutions and for the CaCl2 and Na2CO3 solutions. 13C-enriched calcite samples were synthesized by using Na 2 13CO (95% 13C; Aldrich Isotec) for the carbonate syringe. 3 Coprecipitates with aspartate and glutamate were synthesized with a technique similar to that for citrate/calcite, by addition of free L-aspartic (Asp) or L-glutamic (Glu) acid into both the CaCl2 syringe (0.5 mM Asp or Glu, pH ) 5.6) and the initial growth solutions (0.05 mM Asp or Glu, initial pH ≈ 7.9). The pH was approximately 8.3 during crystal growth. Samples for NMR analysis were prepared from 99% 13C-enriched L-aspartic (13C4) and L-glutamic (13C5) acids (Aldrich Isotec). Typical synthesis runs lasted approximately 8 h, during which pH was continuously monitored, with total yields of about 0.75 g calcite overgrowth on the 0.5 g of seed crystals (Table 1). The final Ca(II) concentration of the growth solution was measured by DCP spectrophotometry, from which the mass 4534
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 12, 2005
of the calcite overgrowth on the seed crystals was calculated by mass balance. Samples were filtered, washed repeatedly in deionized water (or D2O), and dried at 60 °C. All solid products were finely crystalline and characterized by powder X-ray diffraction. All diffraction lines indexed to calcite, giving no evidence for the presence of impurity phases. Calcite samples synthesized with citrate and phthalate were also subjected to thermogravimetric analysis (TGA), conducted in air at a heating rate of 5 °C min-1 using a Netzsch STA 449C instrument. For comparison, a control sample was also analyzed, which was prepared as described above, but without added organic. Organic Analyses. Portions of the calcite/citrate samples were dissolved with a high purity hydrochloric acid solution, and the concentration of citrate in the solids was determined by ion chromatography (IC) (see Table 1). The concentration of amino acids, L-aspartic acid and L-glutamic acid, in calcite produced by coprecipitation experiments was analyzed using the method described by Ingalls et al. (31). Briefly, ∼100 mg of each calcite sample was dissolved in 12 N HCl to form a final 6 N HCl solution after dissolution. Additional 6 N HCl was added for the hydrolysis, which was conducted at 150 °C for 90 min under N2 atmosphere (32). High-performance liquid chromatography (HPLC) was used to analyze total hydrolyzable amino acid. NMR Spectroscopy. The solid-state NMR spectra were measured on a 400 MHz (9.4 T) Varian Inova spectrometer, using standard 13C{1H} CP-MAS and 1H single-pulse (SP)MAS) techniques. A Varian/Chemagnetics probe assembly configured for 7.5 mm was used for the 13C-detected experiments, with a typical sample mass of approximately 0.5 g. 2D 13C{1H} heteronuclear correlation (HetCor) spectra (33, 34) were obtained for samples synthesized with 13Cenriched carbonate. The hydrogen content of the bulk samples was estimated using quantitative 1H NMR techniques. Details of the NMR experimental methods and acquisition parameters are given as Supporting Information.
Results and Discussion Sample Characterization. Analysis of the digested citrate/ calcite coprecipitates indicate that the synthesized calcite samples contain approximately 0.7 wt % citrate (Table 1), a value too high to be attributed solely to surface sorbates that may have remained after rinsing. Assuming all of the citrate is incorporated into overgrowths produced on the calcite seed crystals yields a citrate concentration of over 1.1 wt % in the overgrowths. TGA results for a calcite/citrate coprecipitate (Figure SI-3, Supporting Information) show significant weight loss between 400 and 600 °C, temperatures well below carbonate decomposition. The difference in weight loss from that for a control sample (about 0.8 wt %) is consistent with decomposition of citrate plus evolution of bound water at the concentrations determined, respectively, by IC and 1H NMR (vide infra). The TGA weight-loss curve for calcite precipitated with phthalate did not differ significantly from that for the control sample, suggesting very little
FIGURE 1. Predominance diagram for Ca(II)-citrate speciation under the experimental conditions for preparation of calcite/citrate coprecipitates in this study. phthalate incorporation during calcite growth. Amino acid analyses of the digested coprecipitates calcite/Asp and calcite/Glu yielded evidence for organic incorporation in the calcite crystals but at concentrations much lower than for citrate (Table 1). During citrate coprecipitation, the solution pH reached a relatively constant value of 8.3-8.4 within about 3 h of the start of the experiment, similar to that observed for calcite precipitation without citrate present. Measurements of both Ca and total carbonate concentrations showed that both reached constant values within 3 h. These observations suggest that our synthesis experiments reach a steady-state condition after 3 h, that the precipitation rate of calcite reached a constant rate after this period, and that the presence of citrate had little effect on the synthesis conditions. Aqueous speciation of the growth solution (Figure 1), calculated with the program PHREEQC using literature values for the thermodynamic data (35), indicates that over 96% of the citrate is present as the 1:1 Ca(II)-citrate complex. In the calculation, citrate and Ca(II) concentrations were fixed at total 0.1 and 5.0 mM, respectively. Although citrate complexation to Ca(II) is strong under the experimental conditions, the saturation state of the growth solution with respect to calcite should not be affected because of the high metalto-ligand ratio. Because of its predominance in our growth solution, however, it is likely that the aqueous Ca(II)-citrate complexation is principally involved in the incorporation of citrate into the bulk of calcite during coprecipitation. 13 C{1H} CP-MAS NMR Spectroscopy. The 13C{1H} CPMAS spectra of calcite/citrate coprecipitate cc3101 (Figure 2a) contain weak, broad signals at chemical shifts near 182.7, 77.9, and 47 ppm consistent with assignment to the carboxyl, central C-OH, and methylene carbons of citrate, respectively. The chemical shift observed for the carboxyl carbons lies in the range for carboxyl groups bound to Ca(II), represented by 3:2 Ca(II) citrate tetrahydrate (Figure 2b; reagent grade, Aldrich), but differs from those reported for uncomplexed citric acid in the monohydrate solid, 174-178 ppm (36). Significant differences between spectra of the calcite/citrate coprecipitate and crystalline Ca(II)-citrate, including lack of resolution between the terminal and central carboxyl groups and the peak position of the central C-OH carbon (77.9 vs 75.6 ppm), indicate that the citrate signal from the calcite does not represent a surface precipitate. Furthermore, the large width of the peaks from the coprecipitated citrate, 6-10 ppm full width at half maximum, suggests that it occurs in a wide range of structural configurations. For coprecipitates of calcite with L-aspartic and L-glutamic acids, a significant 13C{1H} CP-MAS signal could be detected
FIGURE 2. Typical 13C{1H} CP-MAS NMR spectra of samples used in this study. (a) cc3101 calcite/citrate, plotted at 3× vertical scale. (b) Crystalline 3:2 Ca(II)-citrate tetrahydrate. (c) cc3109 calcite(13C)/citrate. (d) cc387 calcite/aspartate (13C4). (e) cc4003 calcite/ glutamate (13C5). Asterisks denote SSB positions. only from samples prepared with 13C-labeled amino acids. This result is consistent with the amino acid concentrations obtained from the HPLC analysis (Table 1) and our estimated detection limit. Spectra of coprecipitates calcite/Asp (sample cc387) and calcite/Glu (cc4003) are shown in parts d and e of Figure 2, respectively. These spectra resemble those for the calcite/citrate coprecipitate in that broad peaks are present for each of the structurally distinct carbon positions and the chemical shift of the carboxyl peak is consistent with coordination to Ca(II). No signal was detected for calcite grown in the presence of phthalate that was 13C labeled at the carboxyl position. The 13C{1H} CP-MAS spectra of all the calcite/organic coprecipitates contain a narrow peak at 168.7 ppm due to carbonate carbons in calcite (37). The carbonate groups observed in these spectra must occur near hydrogen-bearing defects, because the CP-MAS signal arises from transfer of magnetization from protons to 13C nuclei over distances of less than 1 nm. These samples also give strong 1H NMR signals, with several distinct resonances (Figure SI-1, Supporting Information). Quantitative measurement of 1H NMR signal intensities indicates that the proton concentration in the calcite is much lower for a calcite/citrate coprecipitate prepared in D2O and for calcite precipitated in H2O but without citrate (Table 1). Assuming that the estimated 1H content for sample cc3109 (prepared with citrate in D2O) arises solely from citrate yields a maximum citrate concentration of 0.8-1.0 wt %, in good agreement with the IC results (Table 1). However, a sample prepared similarly but in normal water (cc3107) contains over three times more 1H than expected from its citrate concentration. Expressed as water, this excess H concentration corresponds to approximately 0.4 wt %, or about six water molecules per citrate, after subtracting the background level of H-bearing defects VOL. 39, NO. 12, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4535
TABLE 2. 13C{1H} Cross-Polarization Parameters for Calcite Samples with 13C-Enriched Carbonate in the Overgrowths, Obtained from Least-Squares Fits to Eq 1a sample
conditions
I0 (arb. units)b
TC-H (ms)
T1G,H (ms)
cc3107 cc3109 cc3113
citrate, H2O citrate, D2O none, H2O
100(4) 85(4) 6.1(4)
0.91(5) 3.5(2) 0.89(8)
15(1) >50 51(11)
a Values in parentheses are uncertainties in last digits. b Values normalized to that for sample cc3107; see Supporting Information for details.
represented by control sample cc3113 (no citrate, normal water). CP-MAS Intensities. To determine the relationship between the proton-bearing species associated with carbonate groups and the coprecipitated citrate, we measured the variation in the fraction of carbonate groups near protons for samples prepared with and without citrate (cc3107 and cc3113, respectively), and with substitution of solvent protons by deuterium (cc3107 vs cc3109; Table 2). These samples were prepared with 13C-labeled carbonate in the overgrowths to improve sensitivity and to distinguish the overgrowths from the seed crystals. The maximum CP-MAS intensity for these samples (I0) is related to the fraction of carbonate groups near protons, because the concentration of 13C-labeled carbonate far exceeds that for protons. 13C nuclei for which all nearby hydrogen are 2H should not contribute to the signal. To quantitatively compare signal intensity among the different samples it is necessary to take into account its dependence on the experimental contact time (τ), which depends on relaxation times of the nuclear spin systems according to
[
I(τ) ) I0 1 -
][
TC-H T1F,H
-1
(
1 - exp -
)] (
TC-H τ
exp -
)
T1F,H τ (1)
where TC-H is the cross-relaxation time for 1H f 13C magnetization transfer and T1F,H is the time constant for decay of the 1H magnetization. The intensity data are shown in Figure 3 and the resulting best-fit parameters in Table 2. Figure 2c presents a typical NMR spectrum. Comparison of the results for samples cc3107 (citrate, H2O) and cc3109 (citrate, D2O) shows that precipitating calcite in D2O solvent reduces the fraction of carbonate groups near protons by only about 15% (Table 2), even though the proton concentration in cc3109 is lower by a factor of over 3 (Table 1). This result suggests that the excess water present in the coprecipitates is clustered near the citrate groups. Clustering of the excess water near citrate is supported further by the longer TC-H observed for the sample prepared in D2O. Under constant experimental conditions, TC-H is related to a sum over the carbon-proton distances: TC-H-1 ∝ ∑i[d(C-1Hi)]-6, where d(C-1Hi) is the internuclear distance between the carbonate C and the ith nearby proton (38). Larger values of TC-H can result from longer C-1H distances or a reduction in the number of nearby 1H, both of which should result with substitution of 2H on exchangeable hydrogen positions in citrate/water clusters. However, deuteration of included water should have little affect on the number of carbonate groups near the citrate ion and hence on I0, as we observe. If the excess water were distributed throughout the crystal, then substitution by D2O would significantly reduce the number of carbonate groups near 1H but perhaps not greatly affect cross polarization of carbonate carbons to nonexchangeable hydrogen on the citrate groups. 4536
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 12, 2005
FIGURE 3. Carbonate 13C{1H} CP-MAS intensities for calcite/citrate coprecipitates cc3107 (b) and cc3109 (2; D2O) and for calcite sample cc3113 (9). All samples were prepared with carbonate overgrowths enriched to 99% in 13C. Intensity values are normalized to represent absolute intensity per scan and sample mass, such that I0 ≡ 100 for sample cc3107. Lines represent least-squares fits to eq 1. The CP-MAS results for the control sample (cc3113) indicate that the fraction of carbonate groups near protonbearing species is considerably lower in the absence of coprecipitated organics but remains sufficient to yield a 13C{1H} CP-MAS signal, as was also found by Ueyama et al. (17). Possible H-bearing species include bicarbonate, Ca-OH groups, and bound structural water at crystal defects. It follows that observation of a carbonate peak in 13C{1H} CPMAS spectra is generally not sufficient to conclude that organic molecules are present in calcite. 13C{1H} Heteronuclear Correlation (HetCor) Spectra. More direct information on the nature of the hydrogenbearing species near carbonate was obtained from 2-D 13C{1H} HetCor spectra. For samples cc3107 and cc3109, the 13C dimension of the 2-D contour plots (Figure SI-2, Supporting Information) shows signal only for the carbonate resonance. The 1H slices (cross sections) taken at the carbonate peak position (parts a and b of Figure 4) correspond to 1H NMR spectra of the hydrogens that are in close proximity to the 13C-labeled carbonate groups. The HetCor data for sample cc3107 (citrate/H2O, Figure 4a) show at least three components, including a narrow peak at 7.2 ppm, a somewhat broader peak at 5.3 ppm, and a broad spinning sideband (SSB) pattern that contains most of the intensity (>72%). The system of SSBs corresponds to H environments in which the hydrogens occur as rigid pairs, which is consistent with either bound structural water or methylene groups on the citrate. However, the chemical shift of the SSB pattern, 6.5 ppm (the average position of sideband pairs equidistant from the center band), suggests that it arises principally from rigid structural water. A very similar SSB envelope occurs in the 1H SP-MAS spectrum of sample cc3101 (Figure 4c) but at a much lower intensity compared to the narrower peak at 5.3 ppm. The much lower SSB intensity in the SP-MAS spectrum indicates that much of the excess water in the coprecipitate exhibits some restricted motion and does not cross polarize strongly to the carbonate groups, although the peak at 5.3 ppm is too broad to attribute to fluid inclusions. The HetCor spectrum of sample cc3109 (citrate/D2O, Figure 4b) appears similar to that of cc3107, except for the absence of a peak at 7.2 ppm, which was also noted in the 1H SP-MAS spectra (Figure SI-1, Supporting Information), and a decrease in the intensity of the narrower peak at 5.3
acids Asp and Glu also appear to coprecipitate with calcite but at much lower concentrations than for citrate. Unfortunately the amino acid concentrations are too low for use of NMR techniques to directly establish spatial proximity between carbonate and the Asp or Glu. The results of this study raise a number of important questions regarding incorporation mechanisms for metals and organics in calcite. Clearly, more work is needed to determine the partitioning of the organic molecules between calcite and the solution and how the partitioning changes with solution composition and ligand stereochemistry. For example, solution pH could exert significant influence through both surface charge effects and aqueous phase complexation. Under the synthesis conditions of this study, the citrate exists in solution primarily as the negatively charged Ca(II) monocitrate complex, whereas the surface charge is nearly neutral (39). By assumption that sorption precedes incorporation, coprecipitation might be reduced at pH > 10, as calcite surface charge becomes increasingly negative.
FIGURE 4. (a-b) 1H slices from 2D 13C{1H} HetCor spectra of calcite/ citrate coprecipitates, taken at the 13C carbonate peak position. (a) Sample cc3107 (13C-enriched carbonate, H2O); (b) cc3109 (13Cenriched carbonate, D2O); (c) 1H SP-MAS NMR spectrum of sample cc3101 for comparison. The dotted line is an aid to illustrate the differences in spinning sideband centroid. All spectra were acquired at a spinning rate of 5.0 kHz. ppm relative to the SSB envelope. More significant is a change in the centroid of the SSB envelope to a chemical shift of 3.2 ppm, which is consistent with assignment to methylene protons on the citrate. This assignment is further supported by the large width of the SSB envelope, despite the absence of a significant concentration of 1H in the water molecules. Because of the large 2H/1H ratio, any remaining protons in the structural water should occur primarily as HDO and should not give significant SSB intensity. These observations indicate that the carbonate CP-MAS signal from sample cc3109 results primarily from direct magnetization transfer from hydrogen in the coprecipitated citrate and firmly establishes close spatial proximity between citrate and the carbonate host. However, cross polarization of carbonate to primarily structural water in cc3107 is consistent with the presence of citrate/water clusters such that carbonate carbons are closer to the water protons, as we inferred from the CPMAS intensities. Magnetization transfer from the citrate protons is evident when the water protons are replaced by deuterons. Structural Implications. Our results indicating coprecipitation of citrate with calcite are consistent qualitatively with previous work that shows citrate sorbs strongly to the calcite surface (4) but does not greatly alter calcite growth rates under conditions similar to those employed here (13). The NMR spectroscopic data indicate that much of the citrate is incorporated structurally into the calcite. In particular, the HetCor cross peaks between the carbonate 13C peak and the citrate methylene 1H resonance show that these components are able to exchange magnetization via dipolar coupling, which requires distances within a few angstroms. The citrate CP-MAS signal from the coprecipitates cannot arise solely from a separate Ca(II)-citrate phase. Furthermore, the citrate cannot be attributed primarily to a surface sorbate; distribution of the citrate content over the measured specific surface area would require a surface loading > 100 nm-2. The amino
Structural and stereochemical factors are also likely to play important roles in determining the extent of coprecipitation. The 13C NMR chemical shifts of the citrate carboxyl groups are consistent with binding to Ca(II), and no evidence was found for protonated carboxyl groups. These results are consistent with citrate incorporation through complexation with Ca(II) ions. Although no direct information on the details of Ca(II)-citrate coordination is available from these NMR data, the broad 13C NMR peaks of the included citrate indicate that it occurs in a broad range of configurations. The ability of a ligand to bind to surface Ca(II) sites, however, appears insufficient for coprecipitation; no evidence for phthalate coprecipitation was found even though it is known to sorb strongly to the calcite surface (4). The size of the citrate ion would appear to preclude its substitution directly for any structural units of calcite, requiring some structural accommodation for coprecipitation. Factors contributing to the ability of calcite to incorporate citrate at minor concentrations could include a favorable fit between citrate and the calcite structure and the conformational flexibility of citrate (40). For example, the citrate C1-C5 (carboxyl) distance, 5.1 Å, is very close to one of the distances between adjacent carbonate groups parallel to the calcite cleavage surface (4.991 Å). The position of the central R-hydroxycarboxyl group allows citrate to coordinate several different metal sites in a wide range of configurations, as demonstrated by the crystal structures of the divalent metal citrates (41-43). This feature could allow citrate complexed to surface Ca(II) sites to bind additional Ca(II) in an advancing layer or step. Hydrogen-bonding interactions could also play a crucial role in accommodating such large defects in the calcite structure, as suggested by the occurrence of significant amounts of water in the calcite/ citrate coprecipitates, including rigid structural water. Possible roles for water include completion of Ca(II) coordination shells disrupted by carbonate vacancies and hydrogen bonding to carboxyl or hydroxyl groups (41-43). For citrate coprecipitation, local charge balance considerations seem to require also the presence of additional charged species to balance isolated trivalent citrate in the calcite structure, such as bicarbonate or hydroxyl ions. Paired citrate defects would be another possibility but seem unlikely. A signal that could be attributed to bicarbonate was observed in both 1H SP-MAS spectra and in the HetCor data, although no direct evidence was found to associate these species directly with the citrate. Lack of 13C NMR signal intensity in the coprecipitate at chemical shifts similar to those of citric acid indicates that protonation of carboxyl groups is unlikely to be a significant mechanism for accommodating isolated VOL. 39, NO. 12, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4537
citrate molecules in the structure under the synthesis conditions of this study (pH ≈ 8.3).
Conclusions The results of this study show that significant concentrations of organic molecules can be incorporated into the calcite structure but that the concentration of the coprecipitate depends sensitively on the detailed structure of the organic ligand. One implication of this study is that strong complexation of metals by organic ligands in solution might not prevent metal contaminants from being incorporated into growing calcite crystals. If the organic molecules can be incorporated into the structure, then it might be possible to incorporate a metal chelate complex in the structure as well. Such behavior could either enhance or diminish coprecipitation of metal cations with calcite.
Acknowledgments This work was supported by the NSF-funded Center for Environmental Molecular Science (CHE-0221934) and NSF Grant EAR-0310200 (B.L.P.). We thank two anonymous reviewers for helpful and insightful remarks that led to significant improvements in the mauscript.
Supporting Information Available Details of the NMR spectroscopic measurements, including 1 H MAS NMR spectra (Figure SI-1) and 2-D contour plots of the 13C{1H} HetCor spectra (Figure SI-2), and results of the thermogravimetric analysis as weight-loss curves (Figure SI3). This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Davis, J. A.; Fuller, C. C.; Cook, A. D. A model for trace metal sorption processes at the calcite surface: adsorption of Cd2+ and subsequent solid-solution formation. Geochim. Cosmochim. Acta 1987, 51, 1477-1490. (2) Curti, E. Coprecipitation of radionuclides with calcite: estimation of partition coefficients based on a review of laboratory investigations and geochemical data. Appl. Geochem. 1999, 14, 433-445. (3) Muller, P. J.; Suess, E. Interaction of organic compounds with calcium carbonate - III. Amino acid composition of sorbed layers. Geochim. Cosmochim. Acta 1977, 41, 941-949. (4) Geffroy, C.; Foissy, A.; Persello, J.; Cabane, B. Surface complexation of calcite by carboxylates in water. J. Colloid Interface Sci. 1999, 211, 45-53. (5) Cicerone, D. S.; Regazzoni, A. E.; Blesa, M. A. Electrokinetic properties of the calcite/water interface in the presence of magnesium and organic matter. J. Colloid Interface Sci. 1992, 154, 423-433. (6) Vdovic, N.; Kralj, D. Electrokinetic properties of spontaneously precipitated calcium carbonate polymorphs: the influence of organic substances. Colloid. Surf. A 2000, 161, 499-505. (7) Backfolk, K.; Lagerge, S.; Rosenholm, J. B.; Eklund, D. Aspects on the interaction between sodium carboxymethylcellulose and calcium carbonate and the relationship to specific site adsorption. J. Colloid. Interface Sci. 2002, 248, 5-12. (8) Amjad, Z. Kinetic study of the seeded growth of calcium carbonate in the presence of benzenepolycarboxylic acids. Langmuir 1987, 3, 224-228. (9) Hoch, A. R.; Reddy, M. M.; Aiken, G. R. Calcite crystal growth inhibition by humic substances with emphasis on hydrophobic acids from the Florida Everglades. Geochim. Cosmochim. Acta 2000, 64, 61-72. (10) Reddy, M. M.; Hoch, A. R. Calcite crystal growth rate inhibition by polycarboxylic acids. J. Colloid Interface Sci. 2001, 235, 365370. (11) Meldrum, F. C.; Hyde, S. T. Morphological influence of magnesium and organic additives on the precipitation of calcite. J. Cryst. Growth 2001, 231, 544-558. (12) Inskeep, W. P.; Bloom, P. R. Kinetics of calcite precipitation in the presence of water-soluble organic ligands. Soil Sci. Soc. Am. J. 1986, 50, 1167-1172. 4538
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 12, 2005
(13) Wada, N.; Kanamura, K.; Umegaki, T. Effects of carboxylic acids on the crystallization of calcium carbonate. J. Colloid Interface Sci. 2001, 233, 65-72. (14) Zuddas, P.; Pachana, K.; Faivre, D. The influence of dissolved humic acids on the kinetics of calcite precipitation from seawater solutions. Chem. Geol. 2003, 201, 91-101. (15) Manoli, F.; Dalas, E. Calcium carbonate crystallization in the presence of glutamic acid. J. Cryst. Growth 2001, 222, 293-297. (16) Manoli, F.; Kanakis, J.; Malkaj, P.; Dalas, E. The effect of amino acids on the crystal growth of calcium carbonate. J. Cryst. Growth 2002, 236, 363-370. (17) Ueyama, N.; Hosoi, T.; Yamada, Y.; Doi, M.; Okamura, T.; Nakamura, A. Calcium complexes of carboxylate-containing polyamide with sterically disposed NH‚‚‚O hydrogen bond: Detection of the polyamide in calcium carbonate by 13C crosspolarization/magic angle spinning spectra. Macromolecules 1998, 31, 7119-7126. (18) Barwise, A. J.; Compton, R. G.; Unwin, P. R. The effect of carboxylic acids on the dissolution of calcite in aqueous solution - Part 2. d-, I- and meso-tartaric acids. J. Chem. Soc., Faraday Trans. 1990, 86, 137-144. (19) Compton, R. G.; Brown, C. A. The inhibition of calcite dissolution/precipitation: 1,2-dicarboxylic acids. J. Colloid Interface Sci. 1995, 170, 586-590. (20) Fredd, C. N.; Fogler, H. S. The influence of chelating agents on the kinetics of calcite dissolution. J. Colloid Interface Sci. 1998, 204, 187-197. (21) Wilkins, S. J.; Compton, R. G.; Viles, H. A. The effect of surface pretreatment with polymaleic acid, phosphoric acid, or oxalic acid on the dissolution kinetics of calcium carbonate in aqueous acid. J. Colloid Interface Sci. 2001, 242, 378-385. (22) Burns, K.; Wu, Y. T.; Grant, C. S. Mechanisms of calcite dissolution using environmentally benign polyaspartic acid: A rotating disk study. Langmuir 2003, 19, 5669-5679. (23) Meldrum, F. C. Calcium carbonate in biomineralisation and biomimetic chemistry. Int. Mater. Rev. 2003, 48, 187-224. (24) Amjad, Z. Mineral Scale Formation and Inhibition; Plenum Press: New York, 1995. (25) Teng, H. H.; Dove, P. M. Surface site-specific interactions of aspartate with calcite during dissolution: Implications for biomineralization. Am. Miner. 1997, 82, 878-887. (26) Orme, C. A.; Noy, A.; Wierzbicki, A.; McBride, M. T.; Grantham, M.; Teng, H. H.; Dove, P. M.; DeYoreo, J. J. Formation of chiral morphologies through selective binding of amino acids to calcite surface steps. Nature 2001, 411, 775-779. (27) Reynolds, R. C. Polyphenol inhibition of calcite precipitation in Lake Powell. Limnol. Oceanogr. 1978, 23, 585-597. (28) Compton, R. G.; Sanders, G. H. W. The dissolution of calcite in aqueous acid: the influence of humic species. J. Colloid Interface Sci. 1993, 158, 439-445. (29) Zhong, S. J.; Mucci, A. calcite precipitation in seawater using a constant addition technique: a new overall reaction kinetic expression. Geochim. Cosmochim. Acta 1993, 57, 1409-1417. (30) Tesoriero, A. J.; Pankow, J. F. Solid solution partitioning of Sr2+, Ba2+, and Cd2+ to calcite. Geochim. Cosmochim. Acta 1996, 60, 1053-1063. (31) Ingalls, A. E.; Lee, C.; Druffel, E. R. M. Preservation of organic matter in mound-forming coral skeletons. Geochim. Cosmochim. Acta 2003, 67, 2827-2841. (32) Cowie, G. L.; Hedges, J. I. Improved amino acid quantification in environmental samples: charge-matched recovery standards and reduced analysis time. Mar. Chem. 1992, 37, 223-238. (33) Zumbulyadis, N. Hydrogen-silicon nuclear spin correlations in R-Si:H: a two-dimensional NMR study. Phys. Rev. B 1986, 33, 6495-6496. (34) Schmidt-Rohr, K.; Clauss, J.; Spiess, H. W. Correlation of structure, mobility, and morphological information in heterogeneous polymer materials by two-dimensional wide-lineseparation NMR spectroscopy. Macromolecules 1992, 25, 32733277. (35) Martell, A. E.; Smith, R. M. Critical Stability Constants: Second Supplement; Plenum Press: New York, 1989. (36) Fischer, J. W.; Merwin, L. H.; Nissan, R. A. NMR investigation of the thermolysis of citric acid. Appl. Spectrosc. 1995, 49, 120126. (37) Papenguth, H. W.; Kirkpatrick, R. J.; Montez, B.; Sandberg, P. A. 13C MAS NMR-spectroscopy of inorganic and biogenic carbonates. Am. Miner. 1989, 74, 1152-1158.
(38) Mehring, M. Principles of high-resolution NMR in solids, 2nd ed.; Springer-Verlag: Berlin, 1983. (39) van Cappellen, P.; Charlet, L.; Stumm, W.; Wersin, P. A surface complexation model of the carbonate mineral-aqueous solution interface. Geochim. Cosmochim. Acta 1993, 57, 3505-3518. (40) Glusker, J. P. Structural aspects of citrate biochemistry. Curr. Top. Cell Regul. 1992, 33, 169-184. (41) Johnson, C. K. X-ray crystal analysis of substrates of aconitase. V. magnesium citrate decahydrate [Mg(H2O)6][MgC6H5O7(H2O)]2‚2H2O. Acta Crystallogr. 1965, 18, 1004-&.
(42) Sheldrick, B. Calcium hydrogen citrate trihydrate. Acta Crystallogr. B 1974, 30, 2056-2057. (43) Zacharias, D. E.; Glusker, J. P. Structure of strontium citrate pentahydrate. Acta Crystallogr. C 1993, 49, 1732-1735.
Received for review August 13, 2004. Revised manuscript received April 4, 2005. Accepted April 4, 2005. ES048733X
VOL. 39, NO. 12, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4539