Christopher C. Fuller

Surface chemistry of ferrihydrite: Part 1. EXAFS studies of the geometry of coprecipitated and adsorbed arsenate

Abstract EXAFS spectra were collected on both the As and Fe K-edges from samples of two-line ferrihydrite with adsorbed (ADS) and coprecipitated (CPT) arsenate prepared over a range of conditions and arsenate surface coverages. Spectra also were collected for arsenate adsorbed on the surfaces of three FeOOH crystalline polymorphs, α (goethite), β (akaganeite), and γ (lepidocrocite), and as a free ion in aqueous: solution. Analyses of the As EXAFS show clear evidence for inner sphere bidentate (bridging) arsenate complexes on the ferrihydrite surface and on the surfaces of the crystalline FeOOH polymorphs. The bridging arsenate is attached to adjacent apices of edge-sharing Fe oxyhydroxyl octahedra. The arsenic-iron distance at the interface ( 3.28 ±0.01 A ) is close to that expected for this geometry on the FeOOH polymorph surfaces, but is slightly shorter on the ferrihydrite surfaces ( 3.25 ± 0.02 A ). Mono-dentate arsenate linkages ( 3.60 ± 0.03 A ) also occur on the ferrihydrite, but are not generally observed on the crystalline FeOOH polymorphs. The proportion of monodentate bonds appears largest for adsorption samples with the smallest As Fe molar ratio. In all cases the arsenate tetrahedral complex is relatively undistorted with As-O bonds of 1.66 ± 0.01 A . Precipitation of arsenate or scorodite-like phases was not observed for any samples, all of which were prepared at a pH value of 8. The Fe EXAFS results confirm that the Fe-Fe correlations in the ferrihydrite are progressively disrupted in the CPT samples as the As Fe ratio is increased. Coherent crystallite size is probably no more than 10 A in diameter and no Fe oxyhydroxyl octahedra corner-sharing linkages (as would be present in FeOOH polymorphs) are observed at the largest As Fe ratios. Comparison of the number and type of Fe-Fe neighbors with the topological constraints imposed by the arsenate saturation limit in the CPT samples (about 0.7 As Fe ) indicates ferrihydrite units consisting mainly of Fe oxyhydroxyl octahedra arranged in short dioctahedral chains with minimal interchain linking by octahedra corners. This is consistent with an enlarged surface area and a larger proportion of sites for bidentate arsenate bonding in CPT samples as compared to the ADS samples, which saturate with arsenate at lower As Fe ratios. The latter samples have larger crystallite sizes and a definite proportion of ferric octahedra sharing corners. The ratio of corner-sharing to edge-sharing Fe oxyhydroxyl octahedra in the ADS samples, and CPT samples with small As loadings, is very similar to what would be present in very small particles of goethite or akaganeite. The difference in the polymeric structure of ADS and CPT samples at higher As Fe ratios is due to strong arsenate bidentate adsorption that poisons the surface of particles of ferrihydrite precipitated in the presence of substantial arsenate, limiting their normal crystallization, and preventing further Fe-O-Fe polymerization. If the arsenate is applied after precipitation much less adsorption occurs since polymerization has already progressed. In both ADS and CPT samples, Fe-O-Fe polymerization increases with age, though at different rates for each type of sample.

Surface chemistry of ferrihydrite: Part 2. Kinetics of arsenate adsorption and coprecipitation

Abstract The kinetics of As(V) adsorption by ferrihydrite was investigated in coprecipitation and postsynthesis adsorption experiments conducted in the pH range 7.5–9.0. In coprecipitation experiments, As(V) was present in solution during the hydrolysis and precipitation of iron. In adsorption experiments, a period of rapid ( Waychunas et al., 1993) showed that neither ferric arsenate nor any other As-bearing surface precipitate or solid solution was formed. The high adsorption densities are possible because the ferrihydrite particles are extremely small, approaching the size of small dioctahedral chains at the highest As(V) adsorption density. The results suggest that the solid solution model proposed by Fox (1989, 1992) for control of arsenate and phosphate concentrations in natural waters may be invalid.

Pb2+ and Zn2+ adsorption by a natural aluminum- and iron-bearing surface coating on an aquifer sand

Pb2+ and Zn2+ adsorption was studied in batch experiments with material collected from a shallow, unconfined aquifer of glacial outwash sand and gravel in Falmouth, Massachusetts, USA. The aquifer solids contain primarily quartz (95% w/w), with minor amounts of alkali feldspars and ferromagnetic minerals. Pb2+ and Zn2+ adsorption experiments with various grain size and mineral fractions of the aquifer solids showed that (1) Zn2+ adsorption was independent of grain size, but Pb2+ was preferentially adsorbed by the <64 μm size fraction and (2) Pb2+ adsorption decreased after removal of the paramagnetic, Fe-bearing mineral fraction, but Zn2+ adsorption was unaffected. Pb2+ and Zn2+ adsorption on mineral separates from the aquifer material compared with metal adsorption on a purified quartz powder indicated that adsorption of both metal ions was dominated by coatings on the quartz fraction of the sediment. Characterization of the coatings by AES, SEM-EDS, and TOF-SIMS demonstrated that the natural quartz grains were extensively coated with Al- and Fe-bearing minerals of variable composition. Thin sections of quartz grains examined by TEM showed that the coatings contained both polycrystalline regions and single mineral crystals. The coating thickness varied from <10 nm up to 30 μm. The coatings were mostly resistant to dissolution by an extraction protocol designed to dissolve noncrystalline phases. The effect on metal adsorption of dissolving surface coatings from the sediment by chemical extraction was also measured. A hydroxylamine-HC] extraction designed to dissolve crystalline Fe oxide phases decreased Pb2+ and Zn2+ adsorption relative to untreated sediment (extracted Fe/Al ∼ 1), but Pb2+ and Zn2+ adsorption were not appreciably changed after sediment was extracted with dithionite-citrate (extracted Fe/Al ∼ 5). Overall, the results suggest that Pb2+ preferred to form complexes with iron hydroxyl sites, while aluminol sites were more important for Zn2+ adsorption. However, a definitive understanding of adsorption reactions in groundwaters will require detailed studies of the extensive coatings formed at mineral-water interfaces by chemical weathering processes.

Historical trends of metals in the sediments of San Francisco Bay, California

Abstract Concentrations of Ag, Al, Cr, Cu, Fe, Hg, Mn, Ni, Pb, V and Zn were determined in six sediment cores from San Francisco Bay (SFB) and one sediment core in Tomales Bay (TB), a reference estuary. SFB cores were collected from between the head of the estuary and its mouth (Grizzly Bay, GB; San Pablo Bay, SP; Central Bay, CB; Richardson Bay, RB, respectively) and ranged in length from 150 to 250 cm. Concentrations of Cr, V and Ni are greater than mean crustal content in SFB and TB sediments, and greater than found in many other coastal sediments. However, erosion of ultramafic rock formations in the watershed appears to be the predominant source. Baseline concentrations of other metals were determined from horizons deposited before sediments were influenced by human activities and by comparing concentrations to those in TB. Baseline concentrations of Cu co-varied with Al in the SFB sediments and ranged from 23.7±1.2 μg/g to 41.4±2.4 μg/g. Baseline concentrations of other metals were less variable: Ag, 0.09±0.02 μg/g; Pb, 5.2±0.7 μg/g; Hg, 0.06±0.01 μg/g; Zn, 78±7 μg/g. The earliest anthropogenic influence on metal concentrations appeared as Hg contamination (0.3–0.4 μg/g) in sediments deposited at SP between 1850 and 1880, apparently associated with debris from hydraulic gold mining. Maximum concentrations of Hg within the cores were 20 times baseline. Greater inventories of Hg at SP and GB than at RB verified the importance of mining in the watershed as a source. Enrichment of Ag, Pb, Cu and Zn first appeared after 1910 in the RB core, later than is observed in Europe or eastern North America. Maximum concentrations of Ag and Pb were 5–10 times baseline and Cu and Zn concentrations were less than three times baseline. Large inventories of Pb to the sediments in the GB and SP cores appeared to be the result of the proximity to a large Pb smelter. Inventories of Pb at RB are similar to those typical of atmospheric inputs, although influence from the Pb smelter is also suspected. Concentrations of Hg and Pb have decreased since the 1970s (to 0.30 μg/g and 25 μg/g, respectively) and were similar among all cores in 1990. Early Ag contamination was perhaps a byproduct of the Pb smelting process, but a modern source of Ag is also indicated, especially at RB and CB.

Geometry of sorbed arsenate on ferrihydrite and crystalline FeOOH: Re-evaluation of EXAFS results and topological factors in predicting sorbate geometry, and evidence for monodentate complexes

Manceaus (1995) reinterpretation of some of our EXAFS results (Waychunas et al., 1993) has been analyzed using both old and newly collected data in an attempt to clarify the nature of proposed monodentate and edge-sharing bidentate arsenate complexes on the ferrihydrite surface. It is shown that EXAFS analysis utilizing data with sufficient k-range does indicate the presence of relatively short AsFe bonds, suggestive of an edge-sharing complex as indicated by Manceau (1995). However, a variety of data analysis factors and crystal chemical considerations create doubt in this assignment. Most significantly, X-ray scattering data collected on a sample of ferrihydrite with a large density of sorbed arsenate, which should show a substantial fraction of the edge-sharing complex, does not show any such correlation within fitting uncertainty. We also suggest that it is unnecessary to invoke the presence of edge-sharing bidentate arsenate to explain the surface growth poisoning of ferrihydrite with increasing sorbed arsenate, as Manceau (1995) claims. Further, we show that a model based on the topology of close packed oxygen ions offers a clear explanation why monodentate arsenate should appear on some surfaces and not on others, and why differing AsFe distances might be observed on a single surface with a single type of complex. This model also explains why bidentate sorbed arsenate can occupy positions with consistent “tilt” angles. Without such consistency, the sorbed arsenate would be highly positionally disordered, and difficult to detect accurately via EXAFS methods.

Sedimentary record of anthropogenic and biogenic polycyclic aromatic hydrocarbons in San Francisco Bay, California

Dated sediment cores collected from Richardson and San Pablo Bays in San Francisco Bay were used to reconstruct a history of polycyclic aromatic hydrocarbon (PAH) contamination. The sedimentary record of PAHs in Richardson Bay shows that anthropogenic inputs have increased since the turn of the century, presumably as a result of increasing urbanization and industrialization around the Bay Area. Concentrations range from about 0.04–6.3 μg g−1. The dominant origin of the PAHs contributing to this modern contamination is from combustion processes. Depth profiles in San Pablo Bay indicate higher concentrations of PAHs since the 1950s than during the late 1800s, also presumably resulting from an increase in urbanization and industrialization. Total PAHs in San Pablo Bay range from about 0.04–1.3 μg g−1. The ratios of methylphenanthrenes/phenanthrene and (methylfluoranthenes+methylpyrenes)/fluoranthene were sensitive indicators of anthropogenic influences in the estuary. Variations in the ratio of 1,7-dimethylphenanthrene/2,6-dimethylphenanthrene indicate a gradual replacement of wood by fossil-fuel as the main combustion source of PAHs in San Francisco Bay sediments. The profile of perylene may be an indicator of eroding peat from marshlands.

Wide angle X-ray scattering (WAXS) study of “two-line” ferrihydrite structure: Effect of arsenate sorption and counterion variation and comparison with EXAFS results

Wide angle X-ray scattering (WAXS) measurements have been made on a suite of “two-line” ferrihydrite (FHY2) samples containing varying amounts of coprecipitated arsenate. Samples prepared at pH 8 with counter ions chloride, nitrate, and a mixture of both also were examined. The raw WAXS scattering functions show that “two-line” ferrihydrite actually has a large number of non-Bragg (i.e., diffuse scattering) maxima up to our observation limit of 16 A−1. The type of counter ion used during synthesis produces no significant change in this function. In unarsenated samples, Radial Distribution Functions (RDFs) produced from the scattering functions show a well-defined Fe-O peak at 2.02 A in excellent agreement with the mean distance of 2.01 A from extended X-ray absorption fine structure (EXAFS) analysis. The area under the Fe-O peak is consistent with only octahedral oxygen coordination about iron, and an iron coordination about oxygen of 2.2, in agreement with the EXAFS results, the sample composition, and XANES measurements. The second peak observed in the RDFs is clearly divided into two populations of correlations, at 3.07 and 3.52 A, respectively. These distances are close to the EXAFS-derived Fe-Fe subshell distances of 3.02–3.05 and 3.43–3.46 A, respectively, though this is misleading as the RDF peaks also include contributions from O-Fe and O-O correlations. Simulated RDFs of the FeOOH polymorphs indicate how the observed RDF structure relates to the EXAFS pair-correlation function, and allow comparisons with an ordered ferrihydrite structure. The effect of increasing arsenate content is dramatic, as the RDF peaks are progressively smeared out, indicating a wider range of interatomic distances even at moderate surface coverages, and a loss of longer range correlations. At an As/Fe ratio of 0.68, the surface saturation level of arsenate, the RDF shows little order beyond what would be expected from small pieces of dioctahedral Fe oxyhydroxyl chains or small “sheet” units. Analysis of the first RDF peak yields components due to As-O and Fe-O correlations. As the As-O component at 1.67 A increases in size, the Fe-O component decreases, reflecting a decrease in Fe coordination about the average oxygen. This reduction is consistent with a decrease in mean crystallite size as suggested by EXAFS studies. Analysis of the second RDF peak components shows the progressive decrease in Fe-Fe correlations, and the enhancement of As-Fe correlations, as arsenate level increases. Comparison of the experimental RDF from coprecipitated arsenate-saturated FHY2 with simulated RDFs of model iron oxyhydroxyl structures further constrains possible sizes and geometry for the precipitates, and is consistent with sorbed complexes of the bidentate binuclear (apical oxygen sharing) type.

Sediment chronology in San Francisco Bay, California, defined by 210Pb, 234Th, 137Cs, and 239,240Pu

Abstract Sediment chronologies based on radioisotope depth profiles were developed at two sites in the San Francisco Bay estuary to provide a framework for interpreting historical trends in organic compound and metal contaminant inputs. At Richardson Bay near the estuary mouth, sediments are highly mixed by biological and/or physical processes. Excess 234 Th penetration ranged from 2 to more than 10 cm at eight coring sites, yielding surface sediment mixing coefficients ranging from 12 to 170 cm 2 /year. At the site chosen for contaminant analyses, excess 210 Pb activity was essentially constant over the upper 25 cm of the core with an exponential decrease below to the supported activity between 70 and 90 cm. Both 137 Cs and 239,240 Pu penetrated to 57-cm depth and have broad subsurface maxima between 33 and 41 cm. The best fit of the excess 210 Pb profile to a steady state sediment accumulation and mixing model yielded an accumulation rate of 0.825 g/cm 2 /year (0.89 cm/year at sediment surface), surface mixing coefficient of 71 cm 2 /year, and 33-cm mixed zone with a half-Gaussian depth dependence parameter of 9 cm. Simulations of 137 Cs and 239,240 Pu profiles using these parameters successfully predicted the maximum depth of penetration and the depth of maximum 137 Cs and 239,240 Pu activity. Profiles of successive 1-year hypothetical contaminant pulses were generated using this parameter set to determine the age distribution of sediments at any depth horizon. Because of mixing, sediment particles with a wide range of deposition dates occur at each depth. A sediment chronology was derived from this age distribution to assign the minimum age of deposition and a date of maximum deposition to a depth horizon. The minimum age of sediments in a given horizon is used to estimate the date of first appearance of a contaminant from its maximum depth of penetration. The date of maximum deposition is used to estimate the peak year of input for a contaminant from the depth interval with the highest concentration of that contaminant. Because of the extensive mixing, sediment-bound constituents are rapidly diluted with older material after deposition. In addition, contaminants persist in the mixed zone for many years after deposition. More than 75 years are required to bury 90% of a deposited contaminant below the mixed zone. Reconstructing contaminant inputs is limited to changes occurring on a 20-year time scale. In contrast, mixing is much lower relative to accumulation at a site in San Pablo Bay. Instead, periods of rapid deposition and/or erosion occurred as indicated by frequent sand-silt laminae in the X-radiograph. 137 Cs , 239,240 Pu , and excess 210 Pb activity all penetrated to about 120 cm. The distinct maxima in the fallout radionuclides at 105–110 cm yielded overall linear sedimentation rates of 3.9 to 4.1 cm/year, which are comparable to a rate of 4.5±1.5 cm/year derived from the excess 210 Pb profile.

Surface complexation and precipitate geometry for aqueous Zn(II) sorption on ferrihydrite I: X-ray absorption extended fine structure spectroscopy analysis

Abstract “Two-line” ferrihydrite samples precipitated and then exposed to a range of aqueous Zn solutions (10 −5 to 10 −3 M), and also coprecipitated in similar Zn solutions (pH 6.5), have been examined by Zn and Fe K-edge X-ray absorption spectroscopy. Typical Zn complexes on the surface have Zn-O distances of 1.97(.02) A and coordination numbers of about 4.0(0.5), consistent with tetrahedral oxygen coordination. This contrasts with Zn-O distances of 2.11(.02) A and coordination numbers of 6 to 7 in the aqueous Zn solutions used in sample preparation. X-ray absorption extended fine structure spectroscopy (EXAFS) fits to the second shell of cation neighbors indicate as many as 4 Zn-Fe neighbors at 3.44(.04) A in coprecipitated samples, and about two Zn-Fe neighbors at the same distance in adsorption samples. In both sets of samples, the fitted coordination number of second shell cations decreases as sorption density increases, indicating changes in the number and type of available complexing sites or the onset of competitive precipitation processes. Comparison of our results with the possible geometries for surface complexes and precipitates suggests that the Zn sorption complexes are inner sphere and at lowest adsorption densities are bidentate, sharing apical oxygens with adjacent edge-sharing Fe(O,OH) 6 octahedra. Coprecipitation samples have complexes with similar geometry, but these are polydentate, sharing apices with more than two adjacent edge-sharing Fe(O,OH) 6 polyhedra. The results are inconsistent with Zn entering the ferrihydrite structure (i.e., solid solution formation) or formation of other Zn-Fe precipitates. The fitted Zn-Fe coordination numbers drop with increasing Zn density with a minimum of about 0.8(.2) at Zn/(Zn + Fe) of 0.08 or more. This change appears to be attributable to the onset of precipitation of zinc hydroxide polymers with mainly tetrahedral Zn coordination. At the highest loadings studied, the nature of the complexes changes further, and a second type of precipitate forms. This has a structure based on a brucite layer topology, with mainly octahedral Zn coordination. Amorphous zinc hydroxide samples prepared for comparison had a closely similar local structure. Analysis of the Fe K-edge EXAFS is consistent with surface complexation reactions and surface precipitation at high Zn loadings with little or no Fe-Zn solid solution formation. The formation of Zn-containing precipitates at solution conditions two or more orders of magnitude below their solubility limit is compared with other sorption and spectroscopic studies that describe similar behavior.

Processes and kinetics of Cd2+ sorption by a calcareous aquifer sand

The rate of Cd2+ sorption by a calcareous aquifer sand was characterized by two reaction steps, with the first step reaching completion in 24 hours. The second step proceeded at a slow and nearly constant rate for at least seven days. The first step includes a fast adsorption reaction which is followed by diffusive transport into either a disordered surface film of hydrated calcium carbonate or into pore spaces. After 24 hours the rate of Cd2+ sorption was constant and controlled by the rate of surface coprecipitation, as a solid solution of CdCO3 in CaCO3 formed in recrystallizing material. Desorption of Cd2+ from the sand was slow. Clean grains of primary minerals, e.g. quartz and aluminosilicates. sorbed much less Cd2+ than grains which had surface patches of secondary minerals, e.g. carbonates, iron and manganese oxides. Calcite grains sorbed the greatest amount of Cd2+ on a weight-normalized basis despite the greater abundance of quartz. A method is illustrated for determining empirical binding constants for trace metals at in situ pH values without introducing the experimental problem of supersaturation. The binding constants are useful for solute transport models which include a computation of aqueous speciation.