Environ. Sci. Technol. 2010, 44, 8409–8414
Biogenic Formation and Growth of Uraninite (UO2) SEUNG YEOP LEE,* MIN HOON BAIK, AND JONG WON CHOI Korea Atomic Energy Research Institute, 1045 Daedeok-daero, Yuseong-gu, Daejeon, Korea
Received June 4, 2010. Revised manuscript received October 12, 2010. Accepted October 18, 2010.
Biogenic UO2 (uraninite) nanocrystals may be formed as a product of a microbial reduction process in uranium-enriched environments near the Earth’s surface. We investigated the size, nanometer-scale structure, and aggregation state of UO2 formed by iron-reducing bacterium, Shewanella putrefaciens CN32, from a uranium-rich solution. Characterization of biogenic UO2 precipitates by high-resolution transmission electron microscopy (HRTEM) revealed that the UO2 nanoparticles formed were highly aggregated by organic polymers. Nearly all of the nanocrystals were networked in more or less 100 nm diameter spherical aggregates that displayed some concentric UO2 accumulation with heterogeneity. Interestingly, pure UO2 nanocrystals were piled on one another at several positions via UO2-UO2 interactions, which seem to be intimately related to a specific step in the process of growing large single crystals. In the process, calcium that was easily complexed with aqueous uranium(VI) appeared not to be combined with bioreduced uranium(IV), probably due to its lower binding energy. However, when phosphate was added to the system, calcium was found to be easily associated with uranium(IV), forming a new uranium phase, ningyoite. These results will extend the limited knowledge of microbial uraniferous mineralization and may provide new insights into the fate of aqueous uranium complexes.
Introduction Dissimilatory metal-reducing bacteria (DMRB) can couple the oxidation of organic matter or H2 to the reduction of oxidized radionuclides. Usually, oxidized uranium(VI) is much more soluble than the reduced form, uranium(IV), and typically exists in groundwater as uranyl carbonate complexes (1–3). Oxidized uranium is readily reduced by DMRB under anoxic conditions, resulting in the precipitation of UO2 nanoparticles (4, 5). The rapid rate of oxidized uranium reduction and the low solubility of the reduced form make bioremediation an attractive option for removing uranium from contaminated groundwaters (6–9). Biogenic UO2 is a fascinating and important nanoscale biogeological material. Its long-term structural stability is crucial to the viability of microbial bioremediation strategies (10) that seek to mitigate subsurface uranium contamination. UO2 nanoparticles are potentially highly mobile because of their small size and can redissolve quickly if conditions change (11, 12). Size, shape, structure, degree of crystallinity, and polymer associations all affect * Corresponding author phone: +82 42 868 4735; fax: +82 42 868 8850; e-mail:
[email protected]. 10.1021/es101905m
2010 American Chemical Society
Published on Web 10/27/2010
UO2 solubility, transport in growndwater, and potential for deposition by sedimentation. The fate of uranium in natural systems is of great environmental importance. We report the detailed structure of aggregated nanobiogenic uraninite produced by Shewanella putrefaciens CN32 in the presence of major cations of groundwater. A detailed structural observation of biogenic UO2 with organic polymers has been rarely performed using direct probes (13–16). The work reported here reveals directly observed UO2 nanoparticles and their growth within aggregates to provide insight into the fate of biomineralization products of uranium over longer time scales. Some uranium ore deposits are believed to involve a direct microbial reduction process for uranium(VI) (4, 17), as opposed to an abiotic reduction by reduced species such as sulfide (18), magnetite (19), and green rust (20). We document the aggregation of nanoparticles to form submicrometerscale aggregates by organic polymers, and crystal growth pathways that can lead to morphologies similar to those found in sedimentary environments. Studies of extracellular nanoparticle growth have been rarely reported for radioactive chemical-containing minerals. Here, we present in situ the high-resolution transmission electron microscopy (HRTEM) characterization for the mineralogy and ultrastructure of biogenic UO2 formed by an iron-reducing bacterium and direct evidence that UO2 nanocrytals are grown to larger sizes by crystallographic attachment.
Experimental Section S. putrefaciens strain CN32 (ATCC BAA-1097) was obtained from the American Type Culture Collection (ATCC), U.S. The S. putrefaciens CN32 was routinely cultured aerobically in a 30 g/L tryptic soy broth (TSB) (Difco Laboratories, Detroit, MI), and stock cultures were maintained by freezing them in 40% glycerol at -80 °C. The aerobically cultured S. putrefaciens CN32 cells were harvested at mid to late log phase by centrifugating them from 30 g/L TSB cultures. The cells were centrifuged at 4,000 rpm for 15 min. The supernatant was discarded and the cell pellets were suspended in a 30 mmol/L NaHCO3 (pH 7) buffer solution and purged with N2 gas. This process was repeated four times and washed cells (>4 × 108 cells mL-1) were used as inoculum. The NaHCO3 buffer solution (30 mM) was extensively flushed with N2 to remove dissolved O2. 100 mL of the buffer solution with lactate (10 mM) as an electron donor were dispensed into 120 mL serum bottles under N2 condition. The headspace of the serum bottles was pressurized with ultrapure nitrogen, then capped with butyl rubber septa and crimped with an aluminum seal. The bottle and solution were sterilized by autoclaving at 121 °C and 15 lb/sq for 20 min. To use background electrolytes similar to groundwaters, several filtered (0.2 µm, Advantec cellulose acetate) stock solutions of major cations were aseptically added by syringe and needle to the serum bottles as soluble forms as follows (mM): calcium chloride, 1.0; potassium chloride, 1.0; magnesium chloride, 1.0. In addition, P was separately injected into some of those bottles as a form of sodium hydrogen phosphate (0.3 mM). Uranium(VI) stock solutions were prepared by dissolving a known amount of UO2(NO3)2 · 6H2O (Aldrich) in a previously acidified HClO4 solution to prevent cation hydrolysis. The stock solution concentrations were about 1 × 10-3 M, and uranium(VI) (5 × 10-5 M) was aseptically added using a VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8409
purged needle and syringe. Finally, washed S. putrefaciens CN32 cells were injected into the serum bottles to create a final concentration of 8 mg/L cell protein. The final pH of the solution of the serum bottle was ∼8.0. The solution pH was measured using a Ross combination pH electrode and an Orion 920A pH/ISE/mV/EC meter. The inoculated serum bottles were then put into a rotary-shaker (120 rpm at 25 °C) in the dark. Periodically, 2 mL samples were aseptically removed by syringe and needle through 0.2 µm cellulose acetate filters and were analyzed for soluble concentrations of major ions, including uranium, using an inductively coupled plasma mass spectrometry (ICP/MS). For X-ray diffraction (XRD) analysis, the solid phase separated by the centrifugation (10 000 rpm for 10 min) was dried under N2 condition and stored in a glovebox. The XRD analysis was carried out using a Bruker AXS instrument equipped with small-angle X-ray scattering (SAXS) and general area detector diffraction (GADD) systems, and the peaks were measured using an area detector. Samples (several mg) were held in 0.5-mm-diameter quartz capillary tubes (Charles Supper Co.). We used Cu-KR radiation at 40 kV/40 mA and a counting time of 300 s. For the electron microscopic observation, the precipitate residue of the centrifuged suspension was washed twice with N2-purged water. At first, the samples were observed using field emission scanning electron microscopy (FESEM; S-4700, Hitach) to acquire morphological and structural information. For SEM observation, samples were dried in an anaerobic chamber for several hours and then uniformly sprayed on a carbon tape pasted on the specimen holder. In order to avoid charging during observation, the samples were coated with a thin OsO4 (∼10 nm) layer. Chemical analysis was also carried out using an energy dispersive X-ray spectrometer (EDS; EMAX, Horiba). For a more detailed investigation, HRTEM analysis was performed with samples collected from the experiment to observe bacterial surfaces and newly formed nanomaterials. In order to see the bacterial surface in detail without disturbance, unstained samples were prepared by drying diluted aliquots of suspension on holey carbon films that were coated with a very thin layer of amorphous carbon (∼5-10 nm thickness) on copper and Formvar-mesh HRTEM grids. The prepared samples were immediately observed using HRTEM (JEOL JEM 2100F) at 200 kV. EDS analysis was also used to detect newly formed nanoparticles and analyze the distribution of major elements.
Results During microbial respiration, the initial aqueous uranium concentration (5 × 10-5 M) continued to slowly decrease to a very low level for about 2 weeks (Figure 1). The continuous decline of aqueous uranium was derived from the bioreduction of uranium(VI) to uranium(IV), which is less soluble and generally tends to precipitate as UO2. When phosphate (PO4) was added in the medium, the decrease in uranium(VI) was initially faster, but finally similar to that in the absence of phosphate. In control (no bacteria) samples, uranium(VI) slightly decreased with time, probably by uranium(VI) coprecipitation with CaCO3, which were formed in a small quantity from the medium. This indicates that the bioreduction and coprecipitation of uranium(VI) can simultaneously concur during microbial respiration, implying a possible competitive uranium(VI) sequestration. Such a competitive reaction for the uranium(VI) removal may frequently occur in an environment where calcium is dissolved in a large quantity with carbonate. The decrease in uranium caused by the components of the medium was large in phosphate-containing solutions, but relatively small (∼10%) in all uranium(VI)-containing solutions. During 8410
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010
FIGURE 1. Uranium(VI) reduction by S. putrefaciens strain CN32 in 30 mmol/L, pH 8.1 HCO3 buffer. Phosphate loading caused a little decrease of aqueous uranium(VI) in both control and CN32 samples. Open square and diamond symbols represent aquatic concentrations of phosphate for CN32+PO4 and control, respectively. uranium(VI) bioreduction, the dissolved phosphate continued to decrease along with the aqueous uranium(VI), and it finally remained at a concentration of ∼0.1 mM. The UO2 formed by the iron-reducing bacterium, S. putrefaciens CN32, was characterized by HRTEM (Figure 2A). Microbial UO2 nanoparticles were highly aggregated by organic polymers. The EDS spectrum for spherical-shape UO2 aggregates shows that they were mainly composed of carbon, uranium, and oxygen. A copper component was derived from the grid used in the HRTEM. The strong carbon peak in the EDS spectrum indicates that organic polymers were associated with uranium precipitates. The detected uranium appeared as aggregate balls measuring 50-100 nm in diameter. Observed aggregates of >100 nm in diameter may have been formed as smaller ones were closely linked by organic ligands. The interlinking of aggregates usually occurred in cases of short distance. Through such a process, the enlarged-aggregate was likely to organize organic-based UO2 networks, serving as stable bases for catching mobile small-sized UO2. Some adsorbed uranium particles were present as locally concentric over the aggregates. The much brighter parts of the aggregates indicate some specific uranium-rich places, indicating that pure uranium particles were not accumulated homogeneously but locally. In a large view of an aggregate investigated by HRTEM (Figure 2B), we can see many uranium nanoparticles gathered on organic substrates. The uranium nanocrystals observed were generally less than 3 nm in diameter and existed as separate entities or as an aggregate. The selected area electron diffraction (SAED) pattern displays diffuse powder rings. The ring electron diffraction pattern obtained from the nanoparticles was indexed for the UO2 cubic unit cell. The (hkl) indices and ring-pattern structure indicate that the remaining nanoparticles consisted of crystallographically oriented, finely crystalline UO2. Uranium nanoparticles were generally aggregated in random crystallographic orientations. But in some cases, nanocrystals appeared to have horizontally attached with orientation sharing their edges (b1 and b2 in Figure 2). The areas b1 and b2 show a group of several attached particles with clear grain edges, and spearheads mark interfaces between the primary particles. Based on the
FIGURE 3. HRTEM images of S. putrefaciens CN32 showing (A) uranium phosphate particles precipitated on the cell surfaces, and (B) an enlarged view of an aggregated particle in (A). Two insets in (A) show SAED pattern (ranges of d-spacings) (top) and EDS spectrum (bottom). Spearheads in (B) mark boundaries of three attached nanoparticles with the orientation.
FIGURE 2. (A) A HRTEM image of aggregated UO2 nanocrystals. White arrows indicate some organic bridges between aggregates. Several highly uranium-rich sites on aggregates are shown by yellow spearheads. (B) An enlarged view of a small square area in (A). More highly dense nanoparticles are indicated by yellow spearheads. An inset in (B) shows (hkl) indices ring pattern of uraninite. (b1) A primary crystallite attached to uraninite {111} lattice fringes with ∼0.32 nm-spaced set. (b2) Three crystallographic attached UO2 particles. Spearheads mark some interfaces between primary crystallites. morphology and microstructural features of the attached nanocrystals, we think that crystallographically oriented attachment is a potential crystal growth mechanism for biogenic uraninite. Similar to the previous findings of concentric uranium on aggregates, our investigation reveals that UO2 overgrowth seems to be facilitated at several specific positions at the nanometer-scale (yellow spearheads). The nanometer-scale gathering of pure UO2 may be intimately related to the stepwise attachment of UO2 nanocrystals. UO2 nanocrystals were likely piled up one on another by the UO2-UO2 interaction.
When phosphate was loaded in the system, a new solid phase was formed by the same bacterium, S. putrefaciens CN32. Figure 3 shows the bacterium encrusted with nanometer-size crystallites, which were identified as ningyoite, a uranium calcium phosphate mineral, by the SAED pattern. The observed SAED ring-pattern for the phase shows two diffuse bands of d-spacings (0.28-0.34 nm and 0.17-0.22 nm) and an intense diffracted ring of ∼0.30 nm, which is a characteristic strong peak of ningyoite. Well-crystallized large ningyoite has a considerable number of diffracted peaks (i.e., ∼40 peaks) and their intervals are too close to be recognized as individual diffracted peaks. Thus, the diffraction peaks usually tend to appear as two or three diffuse bands for nanometer-size biogenic crystals having considerable organic impurities (21). According to the EDS spectrum, the new phase contains phosphate as its main constituent along with uranium and calcium. An enlarged view of the HRTEM image shows nanometer-size ningyoite particles attached to each other with orientation (Figure 3B). Among the particles, we could find minor quantities of UO2 nanoparticles coexisting with ningyoite. Using SEM, biogenic ningyoite was characterized as having a spheroid form (diameter, ∼50 nm) and being composed of uranium, calcium, and phosphate (Figure 4). This spheroidal morphology accords well with that of the microbially synthesized ningyoite (CaU(PO4)2 · H2O) by VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8411
FIGURE 4. SEM photomicrographs showing spheroidal precipitates of uranium phosphate phase. (A) The upper part of the precipitates was identified as ningyoite (arrows) and the lower as aragonite by XRD analysis (inset in (A)). (B) magnified ningyoite particles that are composed of uranium, phosphate, and calcium.
Khijniak et al. (21). XRD patterns in Figure 4A reveal that minor amounts of biogenic ningyoite was coprecipitated with calcium carbonate, aragonite (CaCO3). The XRD peaks (right three arrows in the inset of Figure 4A) at around 30° 2θ (i.e., d-spacing ≈ 0.30 nm) are strongly diffracted ones (i.e., relative intensity (I/I0) ≈ 1) for biogenic ningyoite. Such d-spacing of ∼0.30 nm agreed fairly well with the strongest SAED ringpattern of TEM, as shown in Figure 3A. Fortunately, no minerals other than ningyoite have such d-spacing as their main indicator (I/I0 ) 1) of uranium phosphate phases.
Discussion Biogenic UO2 Aggregation and Growth. Some researchers reported that c-type cytochromes of DMRB are essential for the reduction of U(VI) and the formation of extracelluar UO2 nanoparticles (22, 23). In particular, the outer membrane (OM) decaheme cytochrome MtrC (metal reduction) has been known to directly transfer electrons to U(VI) (22). Through this process, uranium(VI) is bioreduced to uranium(IV), and the next step involves the precipitation of the biogenic uranium phase. Microbial organic molecules may induce a rapid capture of the uranium phase. Microbial polymers are known to scavenge nanoparticles (24, 25). Microbially-derived extracellular polymeric substance (EPS) could limit the dispersal of the nanoparticulate uranium-bearing phase that may otherwise be transported away from its source by subsurface fluid flow. Marshall et al. (22) observed that the UO2-EPS matrix exhibited glycocalyx-like properties and contained multiple elements of OM, polysaccharide, and 8412
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010
heme-containing proteins. They demonstrated that there was a close association of the extracellular UO2 nanoparticles with MtrC and OmcA (outer membrane cytochrome). Aggregation by the organic substances can generally restrict nanoparticle transport by inducing settling (15.26, 27), and it can drive nanoparticles to grow, leading to decreased solubility (28, 29). The SAED pattern and crystallographic analysis of the two-dimensional HRTEM lattice-fringe image confirmed the structure and orientation of adsorbed uraninite nanoparticles (Figure 2). The HRTEM image manifested an intimate binding between organic molecules and uraninite nanoparticles. The primary
