Crystal Face Distributions and Surface Site Densities of Two Synthetic

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES

Article

Crystal Face Distributions and Surface Site Densities of Two Synthetic Goethites: Implications for Adsorption Capacities as a Function of Particle Size Kenneth J. T. Livi, Mario Villalobos, Rowan K. Leary, Maria Varela, Jon Barnard, Milton Villacís-García, Rodolfo Zanella, Anna Goodridge, and Paul A. Midgley Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b01814 • Publication Date (Web): 15 Aug 2017 Downloaded from http://pubs.acs.org on August 17, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Crystal Face Distributions and Surface Site Densities of Two Synthetic Goethites: Implications for Adsorption Capacities as a Function of Particle Size Kenneth J. T. Livi,1* Mario Villalobos,2 Rowan Leary,3 Maria Varela,4,# Jon Barnard,3 Milton Villacís-García,2 Rodolfo Zanella,5 Anna Goodridge,1 and Paul Midgley3 1. Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, MD 21218 USA 2. Environmental Bio-Geochemistry Group, Geochemistry Dept., Instituto de Geología, Universidad Nacional Autónoma de México (UNAM), CU, CDMX, 04510 México 3. Department of Materials Science & Metallurgy, 27 Charles Babbage Road, University of Cambridge, CB3 0FS, UK 4. Departamento de Fisica de Materiales & Instituto Pluridisciplinar, Universidad Complutense de Madrid, 28040 Madrid, Spain 5. Centro de Ciencias Aplicadas y Desarrollo Tecnológico (CCADET), Universidad Nacional Autónoma de México (UNAM), CU, CDMX, 04510 México

#

Previous address: Materials Science & Technology Division, Oak Ridge National Laboratory.

Oak Ridge, TN 37831.

1 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

CORRESPONDING AUTHOR: [email protected] ABSTRACT

Two synthetic goethites of varying crystal size distributions were analyzed by BET, conventional TEM, cryo-TEM, atomic resolution STEM and HRTEM, and electron tomography in order to determine the effects of crystal size, shape, and atomic scale surface roughness on their adsorption capacities. The two samples were determined by BET to have very different site densities based on CrVI adsorption experiments. Model specific surfaces areas generated from TEM observations showed that, based on size and shape, there should be little difference in their adsorption capacities. Electron tomography revealed that both samples crystallized with an asymmetric {101} tablet habit. STEM and HRTEM images showed a significant increase in atomic-scale surface roughness of the larger goethite. This difference in roughness was quantified based on measurements of relative abundances of crystal faces {101} and {201}for the two goethites, and a reactive surface site density was calculated for each goethite. Singlycoordinated sites on face {210} are 2.5 more dense than on face {101}, and the larger goethite showed an average total of 36% {210} as compared to 14% for the smaller goethite. This difference explains the considerably larger adsorption capacitiy of the larger goethite vs. the smaller sample, and points towards the necessity of knowing the atomic scale surface structure in predicting mineral adsorption processes.

KEYWORDS goethite, adsorption sites, specific surface area, (scanning) transmission electron microscopy, electron tomography


Page 2 of 37

Page 3 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

1. INTRODUCTION Goethite (α-FeOOH) is a highly abundant and thermodynamically stable Fe(III) oxyhydroxide in the environment (1, 2). It is usually found as colloidal and nano-sized particles and coatings ubiquitous in aqueous environments such as soils and sediments, and in wind-blown mineral dusts (3, 4) and influences the transport and fate of numerous aqueous species through sorptive interactions. Both its bulk and surface structures have been extensively investigated, and the surfaces of ideal goethite crystals have served as a reference for the development of various surface complexation models in current use. Despite several decades of investigations, one enigma of goethite surface reactivity has not been explained. Laboratory preparations of goethite, using the classic Atkinson et al. (5,6) method of synthesis, can produce both nanometer- and micrometer-sized particles. When these variable-sized particles are used in absorption experiments, the larger particles exhibit higher adsorption capacities than the smaller particles, when normalized by specific surface area (SSA) (as determined by the Brunauer–Emmett–Teller (BET) method using dry preparations (7-13)). This anomalous behavior occurs despite the fact that the goethite particle shapes (reportedly acicular) apparently remain unaltered for the different sizes, essentially thwarting the adequate thermodynamic description of the goethite adsorption behavior. Rubasinghege et al. (14) have suggested, as a qualitative explanation to this, the evidence that the nanoparticulate goethite exhibits a greater propensity for particle aggregation under aqueous suspension as compared to its micrometer analogs, resulting in considerable surface occlusion. While considerable aggregation of nanoparticulate goethite is not questionable, it is not clear from this explanation if the dry samples, from which the BET SSA were derived, would


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 37

not show similar (or larger) aggregation behavior, or if aggregation would indeed change the adsorption capacity. Indirect quantitative explanations and modeling of this phenomenon have been provided by Villalobos and collaborators (9, 11, 12) for adsorption of protons, and various cations and anions. They proposed that, in contrast to the ideal flat and prevalent crystal faces {101} and {001} (Pnma space group) of the smaller particle sizes, the larger, more reactive goethites show roughened surfaces as a result of the presence of large proportions of more reactive faces normally found only at the particle tips ({010} and {210} faces). The higher reactivity of these latter faces derives from their considerably larger density of reactive sites.

Using the

crystallographic site density values of each face, and experimental data of maximum Cr(VI) adsorption of goethites of different particle sizes, they were able to back-calculate the expected fractional contribution of each type of crystal face (for a simple two-crystal face goethite model made of {101}/{010}) to quantitatively describe the adsorption data observed. Fixing these face and site contributions, they successfully modeled the goethite adsorption behavior using surface complexation affinity constants obtained individually for each type of surface site. The goal of the present work is an expansion and quantification of the ideas put forth in (12) — that is, to directly determine the possible affects of several parameters (size, shape, crystal habit, surface structure, specific gravity and aggregation state) of goethite particles, and to quantitatively measure their contributions to reactivities of goethites with widely differing crystal sizes. Through these comparisons, corrected site density values for each goethite under “real” aqueous conditions will be derived with increased accuracy. Examination of the affects of these parameters will be accomplished by means of advanced electron microscopic techniques.


Page 5 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

The affects of varying physical parameters of crystals can be estimated by creating model crystals based on electron microscope observations. Specifically, changes in the types of terminating faces (habit) of nano- and microscale crystals could result in differences in total surface normalized reactivity if these faces have differing site densities. Following this, changes in the ratios of terminating faces surface areas on the macroscale (shape, i.e., acicular versus equant) could also alter site densities. An aspect that is related to the ratio of face areas is the presence of atomic scale surface steps (surface structure, i.e., smooth versus rough). Steps on surface faces can act as additional adsorption sites and change the adsorption capacity of a face. The importance of this phenomenon was revealed by (15) on rutile nanoparticle surfaces. Lastly, the amount of occluded faces due to aggregation or oriented attachment during adsorption experiments (in aqueous solution) or surface area measurements (oven-dried BET measurements) can affect the estimation of the maximum loading of adsorbant or the surface area. Comparison of BET and electron microscopy methods of determining SSA. The typical method for determining the SSA of powdered constituents is the BET method. This method requires that the powder be dried before nitrogen gas is allowed to adsorb onto the surface of the crystals. In many nanoparticle samples, the drying process may potentially promote aggregation of crystals that would effectively reduce the total measured surface area. Different methods of drying can lead to different amounts of aggregation (i.e., freeze drying vs air or oven drying). We study here the affects of simple oven drying on the aggregation state of BET samples. In addition, recent investigations of new pathways for mineral growth have revealed that particle mediated growth or oriented attachment (OA) is a possible mechanism for growth of relatively insoluble phases (16, 17) in solution. In OA growth, nanoparticles attach to


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

each other by aligning crystallographic faces in solution to produce a crystallographicallycontinuous larger crystal. This has been shown to occur significantly during goethite growth (18). These observations suggest that face attachment may be active both in solution during adsorption experiments and during the drying process. An alternative to BET measurements of surface area is transmission electron microscopy (TEM), which can be used to determine the average crystal size, shape and habit of goethites and produce a model surface area and site density. In addition, conventional TEM (CTEM) observations can qualitatively assess whether aggregation and OA of the crystals after ovendrying the preparation for BET methods are important. In contrast, Cryo-TEM methods are capable of freezing in the aggregation state of crystals as they are in solution and, therefore, can be compared to the oven-dried state to determine whether there is a change in aggregation state. Augmenting CTEM observations is electron tomography (ET)(19). Electron tomography produces high-resolution 3D reconstructions of nano and microscale particles, which can enable unambiguous identification of the crystal faces that are present and their areas. At the smallest scale, the atomic surface roughness of nanoparticles can be estimated by aberration-corrected scanning transmission electron microscopy (STEM) and high-resolution TEM (AC-HRTEM) imaging if the thin perimeter edge of the crystal is representative of the surface steps. Livi et al. (15) established methods for STEM imaging of surface roughness on nanorutile. The structure and habit of goethite crystals permits the viewing down the [001] direction that separates the Fe and O atoms into columns. This allows for unambiguous interpretation of the orientation of surface steps on the crystal perimeters for both STEM and AC-HRTEM. Combining these observations, we present a more accurate model of the surface state of differentially-sized goethite particles and the reasons for their disparate adsorption site densities.


Page 6 of 37

Page 7 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

2. EXPERIMENTAL 2.1 SAMPLE SYNTHESIS Goethite crystals were synthesized by hydrolyzing Fe(III) solutions with NaOH to pH>12 and aging at 60 oC for one day. Experimental details for obtaining preparations of different specific surface areas (SSAs), within this protocol are described elsewhere (6). In summary, the crucial experimental parameter used to control the final goethite SSA was the speed of base addition to the Fe(III) solution, being inversely proportional to the resulting SSA. Goethites with a range of BET-SSAs from 42 m2/g to 101 m2/g were obtained. The two extreme cases, GOE42 and GOE101, were selected for processing through the different experimental procedures described in the following sections.

2.2 ADSORPTION EXPERIMENTS Concentrations of maximum Cr(VI) adsorbed on the selected goethites were determined in 10-15 replicates according to the procedure described by (11).

Briefly, conditions for

chromate surface oversaturation (4x10-3 mol/L) were set to equilibrate with goethite suspensions (1.8 g/L) at ionic strengths of 0.1 M NaClO4 and constant pH of 4. The suspensions were ultrasonicated for 1 min and shaken for 72 h, after which they were centrifuged and filtered through 0.05 µm membranes.

Cr(VI) remaining in solution was quantified with a Jenna-

Analitycs SPECORD 210 PLUS UV/Vis spectrophotometer at a wavelength of 348 nm. The difference from the total added initially was assumed to be the adsorbed concentration.

2.3 BET MEASUREMENTS


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The specific surface area was calculated by the BET method on a Quantachrome Autosorb 1. Before nitrogen adsorption, 200-250 mg of the oven-dried and (mortar-) dispersed goethites were placed on a Quantrachrome 9 mm cell, and outgassed at 105 °C for 24 h to remove any adsorbed water after storage and transport of the goethite. Nitrogen adsorption isotherms were programmed with a 44 data point collection, of which the first 11 were used for SSA calculations. 2.4 SPECIFIC GRAVITY (SG) MEASUREMENTS Picnometers of 10 mL capacity were used, but all volumes were determined gravimetrically. Before use picnometers were carefully washed with a 50% ethanol solution in water and rinsed several times with water. Each picnometer was filled with water to the mark by overflow, carefully paper-drying the overflow, and replicated five times to register an average water-filled weight (Wp+w). Goethite samples previously dried at 105 ºC for 12 h were accurately weighed to between 0.3 and 0.4 g inside the picnometers (Wg) and were half-filled with water. These were ultrasonicated for 1 h and subsequently were left stationary to let solids settle towards the bottom of the picnometer. Additional water was added to reach the lower part of the narrow neck of the picnometer taking care to stir the least possible the underlying suspension, the lid was placed and any water overflow was paper-dried off. The total weight was computed (Wp+w+g). Assuming a water density of 1 g/cm3, the calculations performed were: Goethite SG = (Wg)/(Wg+Wp+w-Wp+w+g)

2.5 CONVENTIONAL TEM SAMPLE PREPARATION A small portion of the crystals were suspended in double-distilled deionized water and ultrasonicated for 3 minutes. A TEM grid covered with a lacey-carbon support film was dipped


Page 8 of 37

Page 9 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

immediately into the goethite suspension and allowed to dry. Examination of the air-dried samples was performed on a Philips CM 300 FEG TEM at 300 kV at Johns Hopkins University. Since the 300 kV beam is capable of damaging the goethite structure, care was taken not to damage the smallest crystals by performing low-dose procedures.

2.6 CRYO-TEM SAMPLE PREPARATION A small portion of the crystals were suspended in double-distilled deionized water and ultrasonicated for 3 minutes. A droplet of the suspension was placed on a lacey-carbon grid, blotted, and plunge-frozen using an FEI Vitrobot and liquid ethane. The sample was then transferred to a Gatan cryo holder under liquid nitrogen and placed into an FEI T12 TEM operating at 120 kV at Johns Hopkins University. Low doses of electrons were used to image the amorphous H2O encased crystals, which was aided by the use of a high-sensitivity FEI Eagle CCD camera.

2.7 ELECTRON TOMOGRAPHY CONDITIONS Electron tomography (ET) tilt series using high-angle annular dark-field (HAADF) STEM imaging were acquired on an FEI Titan3 80-300 (S)TEM equipped with a CEOS probe aberration-corrector and operated at 300 kV. In HAADF-STEM, image contrast is mainly due to mass-thickness and thus is analogous to X-ray tomography in many regards, but achieves analysis on the nanometer scale. The tilt series spanned between -75° to +70° with a 5° increment for GOE101, and between -70° to +74° with a 2° increment for the larger GOE42 crystal analyzed, and between -72° to +72° with a 2° increment for the smaller GOE42 crystal. The GOE101 and larger GOE42 crystal tilt series were aligned using the methods described in


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 37

(20) (implemented in MATLAB) and (21) (implemented in Fiji/TomoJ), respectively, and the smaller GOE42 crystal by standard cross-correlation (implemented in FEI Inspect3D). A background subtraction was performed for all tilt series images so as to isolate the crystal/aggregate of interest, removing the support film and other crystals in the field of view. This was achieved by masking the crystal of interest throughout the tilt-series (by thresholding followed by manual refinement), use of the roifill command in MATLAB to inpaint the background beneath the crystal and subtraction of the obtained background image from the original. Tomographic reconstruction of the background-subtracted tilt series, also implemented in MATLAB, was carried out by a compressed sensing approach (22) with 3D total variation (TV) minimization, using the primal-dual hybrid gradient method (23) in conjunction with a projection

operator

from

the

ASTRA

toolbox

(24,

25)

(available

at

https://sourceforge.net/projects/astra-toolbox/) and making use of the object oriented framework for inverse problems provided in (26). The 3D visualizations of the ET reconstructions are volume renderings, generated using Avizo Fire (Visualization Sciences Group). Segmentation of the 3D tomographic reconstructions to identify {101} and {210} surface areas was also carried out in Avizo Fire and is described in the supporting information.

2.8 ABERRATION-CORRECTED STEM CONDITIONS STEM imaging was obtained using a NION UltraSTEM 200 microscope at Oak Ridge National Laboratories, equipped with a cold-field emission gun emitter, with a C5/C3 aberration corrector operating at 200 kV. Aberration correction resulted in a spatial resolution better than 1 Å (probe size). Beam convergence was set at an approximate half angle of 32 mrad, and the beam current used was of tens of pA. STEM HAADF images collected information from 120


Page 11 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

mrad and STEM BF images were collected at 5 mrad and less. Due to the possibility of alteration of the atom positions from beam fluence, care was taken to orient the nanocrystals using the center of the crystals, thus minimizing damage to the perimeter. Images were taken along the perimeter so that area scans contained only a small overlap. The 200 kV beam proved to be too reactive for the smallest goethite crystals as they damaged very rapidly. However, the largest crystals withstood the beam long enough to acquire excellent atomic-resolution images without damage to the crystal perimeters. The smaller crystals were imaged in an aberration-corrected TEM in HRTEM imaging mode.

2.9 ABERRATION-CORRECTED HRTEM CONDITIONS AC-HRTEM imaging was performed on an FEI Titan 60-300 TEM at the Center for Electron Microscopy and Analysis, Ohio State University. This was necessary since STEM imaging damaged these crystals very rapidly. Conventional (non aberration-corrected) TEM is not appropriate for determining the perimeter structure since aberrations in the lenses will delocalize lattice fringe information and create ambiguous terminations of the crystals. The smallest goethite crystals were imaged under both 300 kV and 60 kV. Imaging at 300 kV, but with low-dose imaging conditions, produced the best images.

2.10 COMMENT ON SAMPLE SIZE FOR ET, AC-STEM AND AC-HRTEM Electron tomography, AC-STEM and AC-HRTEM are time consuming and expensive methods that are inherently limited to small numbers of particle investigations. The results presented below represent data from only a few crystals for each method and no crystal has undergone examination by more than one of these methods. Therefore, it is possible that that


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 37

results may not be representative of all goethite crystals in the two samples. However, it is unlikely that these crystals are rare aberrant forms and present vastly different shapes and surface roughness from the bulk.

3. RESULTS 3.1

BET,

MAXIMUM

ADSORPTION

LOADING

AND

DENSITY

MEASUREMENTS Results of BET, maximum adsorption loading of Cr(VI) and density of goethite samples are presented in Table 1. From these data, maximum loadings of 2.89 and 4.39 µmol/m2 of Cr(VI) adsorbed to GOE101 and GOE42 may be calculated, respectively. Based on the findings of (27) that bidentate complexation dominates under Cr(IV) saturated conditions, the calculated site densities would be 3.48 sites/nm2 and 5.28 sites/nm2 for GOE101 and GOE42, respectively. If other modes of Cr(VI) binding were present, such as monodentate and outer-sphere complexes, as more recent work on ferrihydrite points out (28), the latter values would represent the higher limits of site densities expected on goethite. Nevertheless, we can safely state that the large GOE42 sample has 1.5 times the surface reactivity of the smaller GOE101, when normalized by SSA.

3.4 CRYSTAL HABIT AND SIZE The crystal habit of GOE101 and GOE42 were investigated by conventional TEM imaging. From TEM images of crystals prepared by air-dried methods, goethite particles of both samples crystallized as laths that were elongated along the b-axis (Pnma setting)(Fig. 1). The average


Page 13 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

dimensions of crystal lengths and widths as determined from CTEM for both samples are given in Table 2. The tips of the laths were often complex shapes, but in general followed traces parallel to {210} faces or combinations of {210} faces constructing terminations roughly parallel to the {010} trace. The prism faces were difficult to determine in most TEM images due to the flatlaying preferred orientation created by sample preparation ([001] parallel to the electron beam). However, some crystals were imaged down the b-axis for GOE101, which revealed that the prism faces for at least some of the crystals were in the {101} form. In order to determine the actual habit of prisms, ET was needed. The ratio of {101}:{001} areas can have a profound effect on the calculated total surface area and volume, which in turn effects the estimated SSA and site density. GOE42 was comprised of a large range of particle sizes (from 2000 x 100 nm to 100 x 15 nm) and appeared to contain crystals of possibly different habits (Fig 1a). In contrast, GOE101 (Fig 1b) exhibited a smaller range of crystal sizes and shapes (from 250 x 50 nm to 50 x 5 nm) and was more homogeneous than GOE42. As a consequence of the varied crystal population, the average length and width estimate for GOE42 is not as precise as that for GOE101.

3.4 CRYO-TEM CryoTEM imaging was performed in order to estimate the extent of agglomeration or oriented attachment of the crystals during the adsorption experiments. The amount of ultrasonication and stirring during the adsorption experiments were duplicated prior to plunge freezing. Comparisons of air-dried and cryoTEM images of GOE42 and GOE101 are shown in Figs 2a,b, respectively. There are significant reductions in the amount of overlapping particles in the cryoTEM images as


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 37

compared to air-dried. Quantitative estimates of the amount of area occluded by surface-tosurface attachments were not possible without 3-dimensional tomographic reconstructions of the aggregation states, and thus are not calculated. Needless to say, these images indicate that BET of air-dried goethite most likely will underestimate the surface area available for adsorption in solution due to aggregation of goethite laths. The cryoTEM images also clarify that many of the laths in GOE42 form star twin growth habits.

3.5 ELECTRON TOMOGRAPHY For GOE101, only aggregates of crystals were successfully imaged in tilt series imaging experiments. Fig. 3a shows an electron tomography (ET) 3D reconstruction of a typically observed tightly packed aggregate (see also movie in Fig. SI1). Although this prevented clear examination of a single crystal, an unexpected benefit was gained since the organizational state of the aggregation could be imaged. First, the faces found on individual GOE101 crystal prisms were in the {101} form only. However, from orthogonal sections through the 3D reconstruction (Fig. 3b), the cross sectional habit of the crystal prisms was tablet and not diamond as would be expected for the ideal {101} prism. Apparently, one set of the two symmetrically related face pairs grew larger than the other – flattening the crystal into tablets. When examining the aggregation as a whole, it is apparent that the outline of the aggregate forms a 50° angle, which mimics the {101} contact angle. Although these aggregates formed during drying and cannot be considered to be permanently joined, the individual particles have undergone the first step of OA. If they were to physically merge, then they could be considered as having grown through OA.


Page 15 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

For GOE42, one large and one relatively small crystal were chosen for ET analysis, to determine if there were significant differences in morphology. A 3D rendering of the large GOE42 crystal (1250 x 150 nm), which exhibits twinning, is shown in Fig. 4a (see also Fig. SI2). From the reconstruction, it can be concluded that GOE42 prism faces are not smooth but contain numerous large ridges that run parallel to the b-axis. Many of these ridges terminate before reaching the main tips of the crystal, which we call “side tips” here. In Fig. 4b, orthogonal cross sections of the crystal revealed that all the faces present in the prism belonged to the {101} form and no {100} faces were found. The tips of GOE42 crystals are composed of a complex arrangement of {210} faces, as revealed by high-resolution STEM imaging (see below). The presence of these {101} ridges effectively thins the overall thickness of the crystals relative to a crystal with a perfect {101} prism. Thus, the surface area to volume ratio would be greater than an ideal {101} prism and decreases the relative surface area of {210}:{101} faces. However, the presence of side tips would increase the {210}:{101} surface area. The surface expression of the ridges suggests that these crystals could have been formed by aggregation of smaller needleshaped crystals as is proposed by the OA non-classical growth method (18). In the 3D analysis of the smaller crystal (410 x 30 nm), the same ridges are present as are the side tips (SI3). Figure 5 summarizes the findings of the goethite habit found by ET. The surface area of specific faces and the total volume can be estimated by segmenting and quantifying the 3D ET reconstructions. The identified areas for the {101} and {210} faces are differentially colored in the reconstructions in Figs. 3c, 4c and S1c, and the obtained surface areas and volumes are given in Table 3. However, since it is only possible to analyze a small number of crystals by ET, the surface areas calculated by ET cannot be used to estimate the


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 37

representative surface area for the two samples. Instead the TEM measurements are more appropriate to obtain the representative size, surface area and volumes.

3.6 STEM IMAGING GOE42 particles were imaged down the [001] axis, which aligned and separated oxygen from iron atoms in the structure. In this orientation, atomic-resolution STEM HAADF resolved individual Fe atom columns and revealed the atomic-scale morphology of the goethite crystal perimeter (Fig. 6a,b). From these images, the atomic roughness of the perimeter can be directly measured by counting Fe pairs parallel to the {101} prism faces and pairs parallel to {210} steps. Over one thousand Fe-pairs were counted using the method of (15) and the results are presented in Table 3. The STEM results showed that there was a significant amount of {210} steps present on the crystal perimeters. If steps observed at the perimeter are the final expression of steps found on the faces of goethite crystals, then perimeter roughness measurements are faithful estimates of the total surface atomic roughness.

3.7 ABERRATION-CORRECTED HRTEM IMAGING Since STEM imaging quickly damaged the smaller crystals of GOE101, AC-HRTEM imaging was used to reveal the atomic roughness at their perimeters (Fig. 6c). Results are presented in Table 4. The important observations of the smaller goethite are that they have less overall atomic surface roughness.

4. CALCULATION OF MODEL SPECIFIC SURFACE AREAS AND SITE DENSITIES


Page 17 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

In order to estimate the representative physical properties of a population as heterogeneous as the two samples studied here, each individual crystal measured by TEM had a full surface area model calculated based on simple geometric relationships related to the angle between the {101} and {210} faces and the width and length of crystals. The average of all calculated crystal properties are presented in Table 5 and designated as the TEM Model. As expected, the SSA for GOE101 is more than triple that of GOE42. The site densities were then calculated from the TEM model SSA and measured specific gravities (Table 6). In the TEM Model, the prism faces only contain {101}-type sites, for which crystallographic analyses show a site density of singly-coordinated sites of 3.03 sites/nm2 (12), and the tip faces only have {210}-type sites with 7.5 singly-coordinated sites/nm2 densities. Based simply on the TEM measurements, ideal crystal geometry and atomically smooth faces, the calculated site densities for GOE42 and GOE101 are not very different from each other. However, the measured site densities based on the adsorption experiments show that this is not the case (BET Model Table 6). If surface roughness is considered (STEM Model), the site densities change significantly for GOE42 and lesser so for GOE101 due to the increased atomic roughness of GOE42 over GOE101. Comparing STEM Model with BET Model singly-coordinated site densities, the GOE101 STEM Model becomes somewhat higher than the BET Model (3.68 vs. 3.48 sites/nm2), while the GOE42 STEM Model approaches the BET Model (4.68 vs. 5.28 sites/nm2). The inclusion of surface roughness presents a better fit to the BET data for GOE42 than the simple ideal TEM Model. The TEM model SSA for GOE101 was 63% higher than the BET measurement (Table 5). This may be due to oriented attachment agglomeration during oven-drying preparation for


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 37

BET, which would lower the measured BET SSA value. This might be mitigated by performing freeze-drying preparation of powders for BET measurements. The TEM Model SSA for GOE42 is very similar to BET values indicating that agglomeration may not be as significant for the larger crystals as it may be for the smaller crystals. Table 7 includes the calculated number of sites for the singly, doubly and triply coordinated binding models for each sample. From these values, the calculated total site densities for GOE101 are: singly coord. = 3.68, doubly coord. = 0.53, triply coord. = 2.60 sites/nm2; and for GOE42: singly coord. = 4.68, doubly coord. = 1.37, triply coord. = 1.92 sites/nm2. These values together with the estimated SSAs from TEM models are necessary for adequate surface complexation modeling: for proton charging both singly- and triply-coordinated site densities are required, while for ion binding singly-coordinated site densities are crucial and perhaps in some cases doubly-coordinated sites may be required. The labor-intensive measurements performed in the present work will require finding adequate correlations of parameters for a simpler way to predict the SSAs and site densities for goethite preparations of different SSAs, especially for those below ca. 80 m2/g measured by BET. Those found here for GOE101 are representative for ideal goethite preparations with BET-SSAs above ca. 80 m2/g.

5. CONCLUSIONS From the conventional TEM, STEM, and ET observations, a simple explanation of the causes for the differences between GOE101 and GOE42 can be derived. The data show that, even though the size of GOE101 and GOE42 crystals are orders of magnitude different, if smooth crystal faces are assumed, there would be no difference in adsorption capacities. The ET data reveal that the habit of synthetic goethite made by the Atkinson et al. (1967) (6) method


Page 19 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

does not contain significant {001} faces and has a unique asymmetric {101} prism. However, this is in common for both samples and cannot explain adsorption differences. The ET data also demonstrate that the GOE101 crystals aggregate into clusters that mimic the macrocrystalline shape containing 50° cluster faces. This demonstrates the tendency for goethite to grow by oriented attachment. The larger GOE42 crystals appear to have grown by oriented attachment and contain remnant {210} tips on the prism sides. The only significant difference between the two samples is surface roughness, and this has a profound effect on the capacity of synthetic goethite to adsorb CrVI (and most other ions, including protons). Thus, although crystal habit can play some part in altering adsorption capacity, it is the surface roughness that explains the differences of large and small synthetic goethites. This indicates that it is important to consider surface roughness for both natural and synthetic samples when trying to predict adsorption capacity. In the case of goethite, the {101} and {210} steps present very different site densities and varying amounts of steps on the surfaces of major crystal faces will greatly affect the total site density. Future surface complexation modeling efforts must take into account the corrected site densities and specific surface areas of the goethite crystal preparation used. ACKNOWLEDGEMENTS The original idea for this paper came from Dimitri Sverjensky, Johns Hopkins University, to whom we are very grateful for his many discussions and contributions to the evolution of this paper. We thank Dr Martin Benning for valuable interactions regarding the electron tomography reconstruction. M.V. and M.V-G. would like to acknowledge funds provided by UNAM-PAPIIT project IT100912 and CONACyT Ciencia Básica 2010 project number 153723; and also thank the assistance of Katherine Vaca-Escobar in measuring goethite crystal dimensions. M.V.-G. is grateful to CONACyT for the Ph.D. student fellowship received. We acknowledge V. Maturano for technical support in BET measurements. Electron microscopy work at ORNL sponsored by the U.S. Department of Energy (DOE), Basic Energy Sciences (BES), Materials Sciences and Engineering Division. MV acknowledges support from MINECO grant MAT2015-66888-C3-3R. R.K.L. acknowledges a Junior Research Fellowship from Clare College. The research leading 19 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

to these results has received funding from the European Research Council under the European Union’s Seventh Framework Programme (FP7/2007-2013)/ERC grant agreement 2915223DIMAGE, as well as from the European Union Seventh Framework Programme under Grant Agreement 312483-ESTEEM2 (Integrated Infrastructure Initiative – I3).

Supporting Information Available: SI1. Movie of ET reconstruction of GOE101 aggregate. SI2. Movie of ET reconstruction of GOE42 large crystal. SI3. Movie of ET reconstruction of GOE42 small crystal.

This material is available free of charge via the Internet at http://pubs.acs.org.


Page 20 of 37

Page 21 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Table 1. Experimental goethite characterization parameters. Specific Surface Cr(VI) adsorption Mass densityc Goethite sample Areaa (m2/g) (g/cm3) maximab (µmol/g) GOE101 GOE42

101 ±3

292 ±3.5

3.55 ±0.05

42 ±2

188 ±12

3.30 ±0.07

a

From nitrogen adsorption BET measurements From 10-15 replicates at pH 4 c From 5 replicates b

Table 2. TEM measurements of Crystal dimensions and calculated surfaces areas and volume. Sample Length (nm) St Dev (1s) Error of Mean Min Max

GOE101 107 42 1.6 22 255

GOE42 681 416 24 123 2097

Width (nm) St Dev (1s) Error of Mean Min Max Number of particles measured

21 7 0.3 5 48 645

99 24 1 4 457 304

L:W

5.3

8.5

Table 3. Surface Area and Volume Calculations from ET Sample

Length (nm)

Width (nm)

Face

ET Surface Area (nm2)

{101} 1.44 x103 GOE101* NA NA {210} 1.69 x102 {101} 5.81 x106 GOE42 1250 150 large crystal {210} 1.87 x104 {101} 2.84 x104 GOE42 410 30 small crystal {210} 2.09 x102 *Only the tip of a crystal analyzed (see Fig. 2c).

ET Total Surface Area (nm2)

ET Volume (nm3)

1.60 x103

4.39 x103

5.99 x105

8.52 x106

2.87 x104

9.93 x104


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 37

Table 4. Surface Step proportions on GOE101 (AC-HRTEM) and GOE42 (STEM) Number Singly- Number Steps Coordinated Sites (STEM) (x104) (TEM)(x104)

Sample

Step Type

Percent of Prism

Percent of Tip

GOE101

{101}

90 ± 2

20 ± 1

1.65 ± 0.04

1.85 ± 0.05

{210}

10 ± 0.5

80 ± 4

0.25 ± 0.01

2.18 ± 0.01

{001}

0

0

{101}

67 ± 1

20 ± 1

60.1 ± 4.4

89.6 ± 4.4

{210}

33 ± 1

79 ± 1

14.0 ± 1.1

12.4 ± 1.1

{001}

0.3 ± 0.1

1 ± 0.5

GOE42

Table 5. Calculated Surface areas and Volumes of single crystals based on TEM measurements. Sample

Faces Present

GOE101

{101} {210}

GOE42

{101} {210}

TEM TEM Total TEM Vol Surface (x104 nm3) Face Area Area (x104 (x104 2 nm ) nm2) 0.53 ± 0.559 ± 1.38 ± 0.52 0.01 0.014 0.029 ± 0.001 19.8 ± 1.5 21.7 ± 1.6 417 ± 51

TEM Model SSA (m2/g)

0.187 ± 0.14


165 ± 2.4

50.4 ± 1.8

Page 23 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Table 6. Calculated singly-coordinated surface sites based on TEM and STEM surface roughness measurements.

Sample

Faces

TEM STEM Model Model 4 Sites (x10 ) Sites (x104)

TEM Model site density (#/nm2)

STEM Model site density (#/nm2)

BET Model site density (#/nm2)

GOE101

Prism

1.61 ± 0.40

1.85 ± 0.46

3.31 ± 0.01

3.68 ± 0.10

3.48 ± 0.14

Tip

0.25 ± 0.07

0.22 ± 0.06

Prism

60.1 ± 4.4

89.6 ± 6.6

3.36 ± 0.01

4.68 ± 0.10

5.28 ± 0.58

Tip

14.0 ± 1.1

12.4 ± 0.95

GOE42


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 37

Table 7. Summary of calculated sites per representative crystal Sample GOE101

Step Type

Percent of Face

Prism

{210}

10 ± 0.5

Prism

{101}

90 ± 2

Tip

{210}

79 ± 4

Tip

{101}

20 ± 1

Tip

{010}

1±1

Total

{210}

14 ± 2

Total

{101}

86 ± 5

Total

{010}

0.06 ± 0.1

Sample GOE42 Prism

{210}

33 ± 1

Prism

{101}

67 ± 1

Prism

{010}

0.3 ± 0.1

Tip

{210}

79 ± 1

Tip

{101}

20 ± 1

Tip

{010}

1±1

Total

{210}

36 ± 2

Total

{101}

63 ± 2

Total

{010}

0.4 ± 0.2

Area of Steps (x104 nm2) 0.053 ± 0.003 0.477 ± 0.02 0.026 ± 0.001 0.007 ± 0.001 0.0003 ± 0.0001 0.079 ± 0.004 0.484 ± 0.02 0.0003 ± 0.0001

# Singly Coord Sites (x104) 0.398 ± 0.04 1.45 ± 0.10 0.196 ± 0.020 1.45 ± 0.15 0.003 ± 0.001 0.593 ± 0.11 1.47 ± 0.16 0.003 ± 0.001

# Doubly Coord Sites (x104) 0.199 ± 0.02 0

4.63 ± 0.23 9.42 ± 0.47 0.048 ± 0.002 0.851 ± 0.04 0.215 ± 0.011 0.018 ± 0.001 5.49 ± 0.27 9.63 ± 0.48 0.059 ± 0.003

34.8 ± 6.3 28.5 ± 4.7 0.436 ± 0.21 6.38 ± 1.0 0.652 ± 0.13 0.098 ± 0.11 41.1 ± 8.4 29.2 ± 5.3 0.532 ± 0.34

17.4 ± 3.1 0

0.098 ± 0.001 0 0.002 ± 0.001 0.297 ± 0.06 0 0.002 ± 0.001

0.218 ± 0.11 3.19 ± 0.52 0 0.049 ± 0.056 20.6 ± 4.2 0 0.267 ± 0.17


# Triply Coord Sites (x104) 0 1.45 ± 0.10 0 0.020 ± 0.002 0 0 1.47 ± 0.16 0

0 28.5 ± 4.7 0 0 0.652 ± 0.13 0 0 29.2 ± 5.3 0

Page 25 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

REFERENCES (1) Bigham, J. M.; Fitzpatrick, R. W.; Schulze, D. G. Iron oxides. In Soil mineralogy with environmental applications. SSSA Book Series no. 7 (eds. J.B. Dixon and D.G. Schulze). Soil Science Society of America, Madison, Wisconsin 2002, pp. 323-366. (2) Langmuir, D. Aqueous Environmental Geochemistry. Prentice Hall, USA, 1997. (3) Waychunas, G. A.; Kim, C. S.; Banfield, J. F. Nanoparticulate iron oxide minerals in soils and sediments: unique properties and contaminant scavenging mechanisms. Journal of Nanoparticle Research 2005, 7, 409-433. (4) Raiswell, R. Iron Transport from the Continents to the Open Ocean: The Aging– Rejuvenation Cycle. Elements 2011, 7, 101-106. (5) Schwertmann, U.; Cornell, R.M. Iron oxides in the laboratory – Preparation and characterization. WILEY-VCH Verlag GmbH, D-69469 Weinheim Germany 2000, p. 204. (6) Atkinson, F..J.; Posner, A. M.; Quirk, J. P. Adsorption of Potential-Determining Ions at Ferric Oxide-Aqueous Electrolyte Interface. J. Phys. Chem. 1967, 71, 550. (7) Hiemstra, T.; Van Riemsdijk, W. H. A surface structural approach to ion adsorption: the charge distribution (CD) model. J. Colloid Interface Sci. 1996, 179, 488-508. (8) Hiemstra, T.; Van Riemsdijk, W. H. Fluoride adsorption on goethite in relation to different types of surface sites. J. Colloid Interface Sci. 2000, 225, 94-104. (9) Villalobos, M.; Trotz, M.; & Leckie, J. O. Variability in goethite surface site density: Evidence from proton and carbonate sorption. J. Colloid Interface Sci. 2003, 268, 273-287.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 37

(10) Cwiertny, D. M.; Hunter, G. J.; Pettibone, J. M.; Scherer, M. M.; Grassian, V. H. Surface chemistry and dissolution of α-FeOOH nanorods and microrods: Environmental implications of size-dependent interactions with oxalate. J. Phys. Chem. C. 2009, 113, 2175-2186. (11) Villalobos, M.; Cheney, M. A.; Alcaraz-Cienfuegos, J., Goethite surface reactivity: II. A microscopic site-density model that describes its surface area/normalized variability. J. Colloid Interface Sci. 2009, 336, 412-422. (12) Salazar-Camacho, C.; Villalobos, M. Goethite surface reactivity: III. Unifying arsenate adsorption behavior through a variable crystal face – Site density model. Geochimica et Cosmochimica Acta 2010, 74, 2257-2280. (13) Wijenayaka, L. A.; Rubasinghege, G.; Baltrusaitis, J.; Grassian, V. H. (2012) Surface Chemistry of α-FeOOH Nanorods and Microrods with Gas-Phase Nitric Acid and Water Vapor: Insights into the Role of Particle Size, Surface Structure, and Surface Hydroxyl Groups in the Adsorption and Reactivity of α-FeOOH with Atmospheric Gases. J. Phys. Chem. C 2012, 116, 12566−12577. (14) Rubasinghege, G.; Kyei, P. K.; Scherer, M. M.; Grassian, V. H. Proton-promoted dissolution of α-FeOOH nanorods and microrods: Size dependence, anion effects (carbonate and phosphate), aggregation and surface adsorption. J. Coll. Interface Sci. 2012, 385, 15–23. (15) Livi, K. J. T.; Schaffer, B.; Azzolini, D.; Sverjensky, D.; Hazen, R.; Brydson, R. Atomicscale surface roughness of rutile and implications for organic molecule adsorption. Langmuir 2013, 29, 6876-6883.


Page 27 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(16) Penn, R.; Banfield, J. Oriented attachment and growth, twinning, polytypism, and formation of metastable phases: Insights from nanocrystalline TiO2. American Mineralogist 2015, 83, 1077-1082. (17) De Yoreo, J.J.; Gilbert, P.U.P.A; Sommerdijk, N.A.J.M.; Penn, R.L.; Whitelam,S.; Joester, D.; Zhang, H.; Rimer, J.D.; Navrotsky, A.; Banfield, J.; Wallace, A.F.; Michel, F.M.; Meldrum, F.C.; Cölfen, H.; Dove, P.M. Crystallization by particle attachment in synthetic, biogenic, and geologic environments. Science 2015, 349, DOI: 10.1126/science.aaa6760. (18) Yuwono, V.M.; Burrows, N.D.; Soltis, J.A.; Penn, R.L. Oriented aggregation: Formation and transformation of mesocrystal intermediates revealed. J. Am. Chem. Soc. 2010, 132, 2163– 2165. (19) Weyland, M.; Midgley, P.A. Electron tomography. In RSC Nanoscience and Nanotechnology 2015, 37, 211-299. (20) Sanders, T.; Prange, M.; Akatay, C.; Binev, P. Advanced Structural and Chemical Imaging 2015, 1-4. (21) Sorzano, C. O. S.; Messaoudi, C.; Eibauer, M.; Bilbao-Castro, J. R.; Hegerl, R.; Nickell, S.; Marco, S.; Carazo, J. M. Marker-Free Image Registration of Electron Tomography TiltSeries. BMC Bioinf. 2009, 10, 124. (22) Leary, R.; Saghi, Z.; Midgley, P. A.; Holland, D. J. Compressed Sensing Electron Tomography. Ultramicroscopy 2013, 131, 70−91. (23) Chambolle, A., & Pock, T. A first-order primal-dual algorithm for convex problems with applications to imaging. Journal of Mathematical Imaging and Vision. 2011, 40, 120-145. 27 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 37

(24) van Aarle, W.; Palenstijn, W. J.; De Beenhouwer, J.; Altantzis,T.; Bals, S.; Batenburg, K. J.; Sijbers, J. The ASTRA Toolbox: A Platform for Advanced Algorithm Development in Electron Tomography. Ultramicroscopy 2015, 157, 35−47. (25) Palenstijn, W. J.; Batenburg, K. J.; Sijbers, J. Performance Improvements for Iterative Electron Tomography Reconstruction Using Graphics Processing Units (GPUs). J. Struct. Biol. 2011, 176, 250−253. (26) Benning, M. An Object-Oriented Matlab-Framework for Inverse Problems (OOMFIP) Version 0.5 [dataset]. https://doi.org/10.17863/CAM.281. (27) Fendorf, S; Eick, M. J.; Grossl, P, Sparks, D. L. Arsenate and Chromate Retention Mechanisms on Goethite. 1. Surface Structure. Environ. Sci. Technol. 1997, 31, 315-320. (28) Johnston, C. P. & Chrysochoou, M. Investigation of Chromate Coordination on Ferrihydrite by in Situ ATR-FTIR Spectroscopy and Theoretical Frequency Calculations. Environ. Sci. Technol. 2012, 46, 5851-5858.


Page 29 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

FIGURE CAPTIONS

Figure 1. Conventional TEM images of air-dried GOE42 (A) and GOE101 (B). Note the change in scale and the amount of clustering of particles in both samples. Inset in (B) is at the same scale as (A).

Figure 2. Cryo-TEM images of GOE42 (A) and GOE101 (B). This represents the form of attachments that exist in solution during adsorption experiments. GOE42 exhibits start-shaped twinning while GOE101 exhibits a significant degree of tip-to-tip attachment.

Figure 3. Electron Tomography 3D reconstructions of an air-dried aggregate of GOE101 (A,B). (B) is a cross-section view of the aggregate showing the degree of alignment of primary crystals and the overall adherence of the aggregate outline to the 50° {101} contact angle. (C) illustrates the detail of the surface structure of a tip of a single crystal. The {101} faces are colored green while the {210} tip is blue.

Figure 4. Electron Tomography 3D reconstruction of a twinned GOE42 particle (A,B). (B) shows cross-section views through various parts of the crystal. Note that the face of the crystal is composed of contact angles associated with the {101} form. (C) shows the segmented reconstruction where the {101} faces are blue and {210} are red. Note the occurrence of side tips.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 37

Figure 5. Summary of ET findings and habit description. Left-side-top: diagram of the expected habit of a {101} prism with equant faces and {210} terminations. Right-side-top: diagram of the actual habit of a {101} prism found in this study, exhibiting a tablet form and {210} terminations. Middle: schematic of the arrangement of primary crystals upon drying for GOE101 and during the early stage of crystal growth of GOE42. Bottom: ET cross-section views of GEO101 (left) and GOE42 (right) with crystal outlines.

Figure 6. (A) AC-STEM HAADF of GOE42 of the {101} prism perimeter. White dots represent the Fe-atom column positions looking down the [001] direction. At the very edge of the crystal, the outermost Fe atom is located by a red circle. {101} terraces (locations where two or more Fe atoms form a line parallel to the (101) plane) are delineated by blue lines. {210} terraces (locations where two or more Fe atoms form a line parallel to the (210) plane) are delineated by pink and green lines. (B) AC-STEM HAADF of GOE42 of the {210} tip. (Note that this image is rotated 90° with respect to (A).) The color scheme is the same as (A). (C) AC-HRTEM of GOE101 prism perimeter. Here the {101} terraces are colored white, {210} are green and yellow, while {010} terraces are red.


Page 31 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

A

Figure 1.

ACS Paragon Plus Environment

B

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

A

Page 32 of 37

B

1 µm

200 nm

Figure 2.


Page 33 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

A

B

C

50°

10 nm

10 nm

5 nm

Figure 3.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

A

50°

B

50 nm

Page 34 of 37

50 nm

200 nm Figure 4.


C

Page 35 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Actual Habit

Ideal Habit

{101}

{101}

{101}

{101}

{210}

{101}

{210}

{101}

{101} {101}

GOE101 upon drying GOE42 Early growth stage

10 nm

50 nm

Figure 5.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 37

A b-axis

2 nm

{210} Terrace

{101} Terrace

B 2 nm

b-axis

C

{010} Terrace

{210} Terrace

{101} Terrace

5 nm

Figure 6. ACS Paragon Plus Environment

Page 37 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Goethite α-FeOOH

Adsorption Experiments Cr(VI)

Atomic-Resolution STEM

Electron Tomography


Crystal Face Distributions and Surface Site Densities of Two Synthetic

Recommend Documents