Advanced characterization of rare earth element minerals in coal utilization byproducts using multimodal image analysis

International Journal of Coal Geology 195 (2018) 362–372

Contents lists available at ScienceDirect

International Journal of Coal Geology journal homepage: www.elsevier.com/locate/coal

Advanced characterization of rare earth element minerals in coal utilization byproducts using multimodal image analysis

T

⁎

Scott N. Montrossa, , Circe A. Verbab, Han Ling Chanc, Christina Lopanob a

ORISE, Research and Innovation Center, National Energy Technology Laboratory, United States Research and Innovation Center, National Energy Technology Laboratory, United States c Electron Microscopy Facility, Oregon State University, United States b

A R T I C LE I N FO

A B S T R A C T

Keywords: Rare earth elements Coal Coal utilization byproducts Microanalysis Focused ion beam scanning electron microscopy

Multimodal microanalytical characterization techniques are applied to identify and quantify rare earth element (REE) and REE + Y (REY) bearing mineral phases in coal utilization byproducts (CUB) from various coal-fired power plants. The characterization work provides quantitative assessments of REE in coal and CUB as obtained from 2- and 3D imaging, elemental mapping, volumetric estimates, and advanced high-resolution pixel classification. REY-bearing phosphate minerals rhabdophane (Ce,La)(PO4·H2O), monazite (Ce,La,Nd,Th)(PO4,SiO4), xenotime (YPO4,SiO4), and apatite (Ca5(PO4)3(F,Cl,OH), as well as REE-enriched calcium oxide were identified via electron microscopy and microprobe analysis. The minerals generally occurred as 1–20 μm-long crystals in the rock and ash samples. The most notable finding was that REEs are present as monomineralic grains dispersed within the ash, as well as fused to or encapsulated by amorphous aluminosilicate glass particles, also referred to as slag. It is indicative that conventional coal combustion processes sequester REE-bearing mineral phases such as rhabdophane, monazite, and zircon from the coal feed into aluminosilicate glass phases. The advanced microscopy and image analysis techniques applied in this study make it possible to deduce the average density of ash particles and quantify REE phases encapsulated in glass. Consequently, these REE phases may be targeted and recovered via density separation. These findings advance resource recovery techniques and commercial REE separation technologies for coal and combustion byproducts.

1. Introduction Since 1985, the United States has consumed > 800 million tons of coal per year with nearly 90% used for electricity generation (US-EIA Energy Review, 2016). Coal-fired electric utilities produce over 100 million tons of coal utilization byproducts (CUB) annually (ACA, 2014). The U.S. Department of Energy (DOE) defines CUB as fly ash, bottom ash, boiler slag, fluidized-bed combustion ash, or flue gas desulphurization materials produced from the combustion of coal or the cleaning of stack gases (US-DOE Topical Report 24). Industrial and engineering uses for CUB solid residues have reduced fly ash waste. However, nearly 75 million tons of CUB are disposed of in landfills or storage ponds annually (ACA, 2014). Given the reliance on imported REYs, there is a need to identify a viable domestic source of critical elements (e.g., REY, Li, Co, Sc, and Ga) in order to increase domestic production for several sectors of the U.S. economy (e.g. catalysts, electronics, magnets, batteries, and other applications related to national security and energy independence). Rare earth elements (REEs or lanthanides) and REE + yttrium (REY) ⁎

Corresponding author. E-mail address: [email protected] (S.N. Montross).

https://doi.org/10.1016/j.coal.2018.06.018 Received 28 February 2018; Received in revised form 21 June 2018; Accepted 23 June 2018 Available online 25 June 2018 0166-5162/ © 2018 Elsevier B.V. All rights reserved.

concentrations are reported for coal beds and coal fly ash from around the world (Dai et al., 2017a; Dai et al., 2016; Franus et al., 2015; Hower et al., 1999, 2013; Rozelle et al., 2016; Seredin, 1996; Schatzel and Stewart, 2003; Warwick et al., 1997). Seredin and Dai (2012) comprehensively reviewed coal deposits and coal ash as an alternative source of REY and showed that REY is enriched in coal seams as well as the overlying and underlying strata. Enrichment of REY in coal is typically equal to or exceeds estimates of REY concentrations in conventional REY ore deposits (Dai and Finkelman, 2018; Seredin and Dai, 2012). The physical and chemical composition of CUB produced by power plants is based on a combination of geologic history, industrial preparation methods, and specific combustion processes (Brownfield et al., 2005; Taggart et al., 2016). High-rank (e.g., bituminous and anthracite) coal and coal ash tend to have the highest REE content, as the enrichment of REE in coal varies greatly and depends on geological conditions (Ketris and Yudovich, 2009; Rudnick and Gao, 2012). Rare earth element concentrations for > 700 CUB and coal preparation samples, including fly ash, bottom ash, feed coal, reject coal, and underclay, are

International Journal of Coal Geology 195 (2018) 362–372

S.N. Montross et al.

Table 1 Sample descriptions and attributes. ID

Description

251 339 345 357 443 508

Ponded and landfill fly ash collected from an ash disposal site at a power plant in Kentucky, USA. Feed coal from central Appalachian basin coal region. Ash pond sample containing mostly fly ash with a lesser amount of bottom ash. Collected from a pulverized coal power plant in Ohio, USA. Coal seam(s) unidentified. Dry fly ash collected from a pulverized coal power plant in Ohio, USA. Coal seam(s) unidentified. Dry ash pond sample containing mostly bottom ash with lesser amount of fly ash. Collected from a pulverized coal power plant in Ohio, USA. Coal seam(s) unidentified. Pit Flint Clay Roof Rock. As-received sample composed of a fine powder (clay-sized) grain fraction and larger, angular rock fragments (1–20 mm long). Heavy Media Cyclone (HMC) coal preparation plant reject. As-received sample composed of a fine powder fraction and larger, rounded rock fragments (1–10 mm long) relatively homogenous in size.

for REE recovery. Detailed elemental and mineral phase data are combined with multimodal microanalytical imaging and image processing techniques to classify the two- and three-dimensional morphological, elemental, and mineralogical properties of six CUB samples: three fly ash, one bottom ash, one roof rock, and one pulverized coal reject. This characterization work aids in the discovery of targeted, efficient extraction techniques that bring REY resources into the market in the safest and most economically viable way.

reported on the National Energy Technology Laboratory (NETL) Research and Innovation Center database housed on the Energy Data Exchange website (www.edx.netl.gov). Higher than average concentrations of rare earths in CUB, along with the abundance of CUB available for extraction, make coal and associated combustion byproducts a viable target for REE resource recovery. Analysis of the United States Geological Survey Coal Quality Database reveals that domestic coal seams are a significant potential resource for REY as well as non-REY critical elements such as Li, Sc, Co, and Ga (Ekmann, 2012; Lin et al., 2018; Luttrell et al., 2016). Rozelle et al. (2016) expanded the coal-associated resource potential by showing that U.S. coal production byproducts are a source of REE. Rare earth element trends in CUB are predominantly based on the origin of feed coal (Kutchko and Kim, 2006). A study of major coal feedstocks in the United States ranked the REE content of coal from different geographic locations (Taggart et al., 2016). Taggart et al. (2016) characterize U.S. fly ashes of different geologic origin and show an average total REE content in Appalachian coal combusted ashes of ~591 ppm, greater than the Illinois and Powder River Basin coals at 403 and 337 ppm, respectively. The concentration of major elements (Al, Si, and Fe) in ash particles vary among coal sources. The most notable difference lies in the percentages of SiO2, Al2O3, and Fe2O3 in the ash, with the major elemental signatures for three coal regions as follows: Appalachian (high Si and Al, low Ca), Illinois Basin (high Fe, low Ca), and Powder River Basin (low Si and Al, high Ca). The high total REE and ratio of heavy elements (Ho, Er, Tm, Lu, and Yb) to light elements (La, Ce, Nd, and Sm) in low-ash samples indicates REE may preferentially associate with mineral species dispersed within the organic matter of the clean coal samples (Taggart et al., 2016). Ash particle formation in a pulverized fuel boiler is a complex process that is controlled by multiple factors –mineral characteristics of the feed, grind size, combustion conditions, and the direction and velocity of the flue gas (Benson et al., 1993). Ash formation during pulverized coal combustion is described by Seames (2013) and Kutchko and Kim (2006). Studies on the mobility of elements during coal combustion have shown that REE may be homogeneously distributed as micron to nanometer-sized particles in amorphous aluminosilicate glass, also referred to as slag (Hower et al., 2013; Kolker et al., 2017; Mardon and Hower, 2004; Thompson et al., 2018). Rhabdophane, monazite and xenotime (REE-PO4) were observed as individual crystals in CUB fly ash matrix or encapsulated by aluminosilicate glass (Thompson et al., 2018). Mardon and Hower (2004) and Hower et al. (1999) demonstrated the impact that source coal properties and combustion techniques, such as temperature, blend, and ash collection points, have on the quality and composition of the combustion byproducts and associated REE. Hower et al. (2013) demonstrate that vaporization, condensation, and crystallization of carbonatite minerals in Ca-bearing igneous rocks is analogous to the formation of solid inorganic mineral particles (e.g., ash particles) during coal combustion and subsequent collection in the bag house. The objective of this work is to characterize CUB samples that are representative of byproducts produced during the utilization of coal in power generation systems. Pulverized roof rock or reject, produced ash, and disposed materials from coal utilization all provide opportunities

2. Materials and methods 2.1. Materials and elemental analysis Six representative samples of CUB were selected from the National Energy Technology Laboratory (NETL) coal and coal byproduct field samples database housed on the Energy Data Exchange (National Energy Technology Laboratory Energy Data Exchange (NETL-EDX), 2018). A detailed description of each sample is provided in Table 1 and photographs of the samples are shown in Fig. 1. Major, minor, and trace element concentrations were measured for dried samples by Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES) and Inductively Coupled Plasma-Mass Spectrometry (ICP-MS). Detailed methods for the measurement of REE and other trace elements in geologic samples by ICP-OES/MS are discussed in Bank et al. (2016). Ratios of heavy, medium, and light REY are calculated based on the geochemical classification by Seredin and Dai (2012), which divides lanthanides and yittrium (REY) into LREY (La, Ce, Nd, Sm), MREY (Eu, Gd, Tb, Dy, Y), and HREY (Ho, Er, Tm, Lu). 2.2. Microanalysis sample preparation Fifty grams of each as-received sample (see Fig. 1) was oven dried at 55 °C for 8 h and stored in a nitrogen-purged desiccator prior to analysis. Each dried sample was mixed thoroughly, and subsamples were taken for grain size analysis, X-ray diffraction (XRD) analysis, and SEM analysis. Rock and ash samples for XRD were ground to a uniform powder that passed through a < 63 μm (230-mesh) sieve. Samples for electron imaging and microanalysis were mounted as powder grain mounts on 1-cm diameter metal stubs backed with conductive carbon tape, or mounted in epoxy and polished. 2.3. X-ray diffraction Bulk mineralogy of rock and ash samples was determined by X-ray diffraction (XRD) of a randomly-oriented powder mount. Powdered samples were mounted in cavity mounts on an automatic sample changer with a spinner. XRD patterns were collected using a PANalytical X'Pert Pro diffractometer equipped with copper radiation and an X'Celerator parallel plate detector. The samples were scanned at 45 kV, 40 mA over a range of 4.0–70° two-theta in continuous mode with a step size of 0.033° and a count time of 1000 s. Phase IDs, peak alignments, and mineral identifications were made via comparison of the diffraction peaks against the PDF4+ - ICDD database (ICDD, 2016) in PANalytical HighScore Plus. Semi-quantitative analysis of the 363

International Journal of Coal Geology 195 (2018) 362–372

S.N. Montross et al.

Fig. 1. Photograph of coal byproduct samples. 251-fly ash, 339-fly ash and bottom ash, 345-fly ash, 357-fly and bottom ash, 443-flint clay roof rock, and 508pulverized coal reject. Scale bar (bottom right) is 25 mm.

sample 345 using a CAMECA SX Five electron probe outfitted with BSE, energy dispersive (EDS), and wave dispersive (WDS) detectors. EPMA imaging was done at 15 kV, 95 nA, and 25 ms dwell time. Elemental Xray maps for REE and P were acquired by WDS and maps for Al, Si, and Ca were acquired by EDS. The concentrations of REE and P were acquired by EDS and WDS from identical spots on three individual monazite grains in sample 345. The calculated percent difference between EDS and WDS results for individual elements are as follows: P (1.3%), La (5.0%), Ce (2.4%), Nd (5.0%), and Pr (0.45%). These values confirm the robustness and reliability of our EDS standards to quantify REE in the samples. The extent of pore space, organic matter, and mineral phases was investigated by volume reconstructions of serial slices using focused ion beam scanning electron microscopy (FIB-SEM) on characterized samples. The FIB-SEM sites for the flint clay roof rock (443), rejected pulverized coal sample (508), and fly ash sample (345) were chosen based on the presence of REE-bearing minerals identified by previous SEMEDS analysis. These sites were prepared for milling at high current in a FEI Helios 650 dual beam FIB-SEM. A Pt-C-Pt pad (20-50-20 nm layer thickness; total 90 nm pad) was put down on the sample at the milling location to protect the sample during milling and prevent curtaining during image acquisition. Fe2O3 was used as the source material to trench out the region of interest. The fiducials were set and milling was conducted at a 52o angle. Detailed analytical conditions for each site can be found in Table 2. Auto Slice & View (ASV) was used to ion mill and image sequentially. This software allows for automatic acquisition of high-resolution 3D images from a through-the-lens detector (TLD) in secondary and backscatter electron mode at 2 kV; the data were acquired by milling serial slices at 15 nm with pixel and voxel size dependent on the sample.

crystalline components in samples 443 and 508 was performed using the Reference Intensity Ratio (RIR) method in X'Pert HighScore. Quantitative analysis of crystalline and non-crystalline components in fly and bottom ash samples (251, 339, 345, and 357) was performed by mixing the samples with ~ 10% NIST Corundum (Al2O3) internal standard. Basic Rietveld full-pattern fitting was performed using the PANalytical HighScore Plus software to quantify mineral percentages and estimate amorphous content (wt%). 2.4. Microscopy and microanalysis Geochemical, mineralogical, and physical properties of CUB rock and ash samples were investigated using microscopy and microanalysis techniques. Images and elemental data from SEM-EDS analysis were combined with multimodal image processing techniques to characterize the microscale properties of both the amorphous and crystalline components in the rock and ash samples. Polished and grain mounted samples were coated with ~10 nm of Pd by evaporation. A field emission-scanning electron microscope (FE-SEM, FEI Inspect F) equipped with an energy-dispersive X-ray spectrometer (EDS, Oxford Instruments) was used to observe the microstructure of rock and ash samples and collect elemental data with accompanying backscatter electron (BSE) micrographs. The microscope was operated at 20 kV at a working distance of 10 mm, a beam current of ~100 nA, and beam aperture 3. All samples were scanned in an x-y criss-cross pattern across the entire area (5.0 cm2) of the mounted sample. Standard-based quantitative EDS was achieved for all analyses using certified rare earth element standards for phosphates (REEP25–15 + FC, Astimex Standards Ltd.) and oxides (Standard block #489, Geller Microanalytical Laboratory). Standard block #489 is certified to ISO 9001 and 17,025 standards. Electron Probe Microanalysis (EPMA) was conducted on fly ash Table 2 Analytical conditions and attributes for FIB-SEM analysis on a FEI Helios 650. Sample

e-beam current (pA)

i-beam voltage (kV)

# Slices

Slice thickness (nm)

X (μm)

Y (μm)

Z (μm)

Pixel size (nm)

Voxel size (μm3)

345 443 508

100 800 400

30 25 30

1500 318 466

15 15 15

38.1 14.9 34.4

37.0 10.5 31.1

22.5 4.78 6.99

14.5 7.29 22.3

3.15 0.80 7.45

364

International Journal of Coal Geology 195 (2018) 362–372

S.N. Montross et al.

2.5. Advanced image processing

Table 4 Summary of REE and REY concentration and ratio in CUB samples.

Backscattered electron images and elemental maps of individual grains were collected on independent, single fields of view and on larger areas using montaging (up to 1024 fields of view at 1024 × 1024 resolution). Single montage images were collected at a resolution of 16.07 pixels/μm at 1024 × 1024 and a magnification of 4,000×. Montaging was completed using Oxford INCA software. This technique allows for identification and mapping of the distribution of REE minerals and other mineral phases in CUB samples over a 1-cm2 area. Elemental data is displayed as an overlay on BSE images, or as cameo images. Cameo images are colour-enhanced images where each colour represents a distinct range of X-ray emission spectra that can be used to highlight distinct phases. Pixel classification and object/phase identification were performed on individual stacks of multiple images, and montaged SEM images using Ilastik 1.1.9 software. The particle morphology and the distribution and occurrence of features in the classified images were quantified using FIJI open source software (Schindelin et al., 2012). REE mineral phases were identified by SEM-BSE and elements were quantified by WDS or EDS. Acquired FIB-SEM images (where applicable) were combined to determine REE by mass. Individual image slices generated by FIB-SEM were realigned and the reconstructed milled volume and features of interest were segmented using ThermoFisher Scientific (formerly FEI) Avizo9 Digital Rock Physics or PerGeos software. Once the volume distribution of REE-phosphates or other minerals was reconstructed, the atomic weight of each element (e.g., Ce, La, and Nd) were used to calculate its mass fraction in the mineral phase or glassy material. The occurrence of REE minerals in the sample were tracked and quantified using image files generated by pixel and object classification. Characterization data and advanced image processing using individual, stacked, and montaged BSE and elemental maps provided both qualitative and quantitative data on the occurrence of REY minerals and other mineral phases of interest in CUB. The methods were successful at (1) determining the bulk mineralogy and relative mineral abundances for each sample using XRD, (2) discovering REY mineral phases in rock and ash samples, (3) imaging in 3D the morphology and chemical composition of ash particles, mineral solids, and rock fragments in the samples, and (4) imaging and quantification of REYbearing minerals encapsulated in fly ash glass particles.

3.1.1. Geochemical properties Results from ICP-OES/MS analysis were used to calculate the bulk chemical composition, as oxides, for each sample (Table 3, Supplemental Material Table S1). Both the roof rock (443) and reject coal (508) have a similar chemical composition– rich in Al and Si phases

SiO2 Al2O3 Fe2O3 CaO MgO P2O5 K2O Na2O TiO2 ZrO2

357

443

508

71.8 26.9 4.4 0.9 1.0 0.1 2.8 0.4 1.6 0.2

60.9 19.5 15.6 2.5 0.7 0.1 2.1 0.5 1.0 0.1

62.0 19.8 18.0 4.0 0.7 0.2 2.3 0.7 0.9 0.2

61.5 17.9 23.5 2.4 0.6 0.1 1.1 1.0 0.8 0.1

66.0 21.1 5.7 0.5 1.7 0.1 2.6 0.2 1.1 0.1

70.7 21.5 5.2 0.1 2.0 0.1 3.4 0.6 1.0 0.1

345

357

443

508

∑REE (mg/kg) ∑REY (mg/kg) ∑LREE (mg/kg) ∑MREE (mg/kg) ∑HREE (mg/kg) LREE/HREE REE/Th Coutl

743 996 629 324 43 15 17 1.4

363 501 308 172 21 15 16 1.5

445 540 377 139 25 15 20 1.2

494 650 422 203 26 16 19 1.4

285 344 250 81 12 21 20 1.0

325 389 287 89 14 21 18 1.0

*Major (> 50%), intermediate (25-50%), minor (5-25%), and trace (< 5%).

Table 3 Chemical composition of CUB samples reported as weight percent oxide. 345

339

3.1.2. Mineralogical properties The roof rock (443) and HMC reject (508) samples are characterized as aluminosilicate mudrock. Sample 443 is a flint clay that consists of dark gray-black and waxy rock fragments that exhibit conchoidal parting, break into shard-like particles, and are typical of kaolinite (Curtis and Spears, 1970). Diffraction patterns for rock and ash samples are shown in Fig. S2 in the supplemental information. The reject coal (508) and flint clay (443) samples have similar mineralogical compositions and contain the crystalline components quartz, kaolinite, chlorite, muscovite/illite, rutile, and calcite (Table 5). XRD patterns confirmed the presence of various clays in roof and reject rock samples.

3.1. Bulk properties of coal utilization byproduct samples

339

251

with ~10% of the bulk composition as other typical rock-forming elements. On the other hand, the four Class F ashes contain > 70% SiO2 + Al2O3 + Fe2O3, with silicon as the most abundant in the samples (see Table 3). Following the conventional ash classification scheme by Vassilev and Vassileva (2005) samples 339, 345, and 357 can be classified as sialoferric (SiO2 + Al2O3 = 80–90%, Fe2O3 > 10%), whereas fly ash sample 251 is sialic (SiO2 + Al2O3 > 90%). All ash samples belong to a high-temperature-ash group based on the oxide chemistry (Vassilev and Vassileva, 2005). Calcium (as CaO) in ash samples ranged from 0.9 to 4.0 (wt%). The CUB samples analyzed in this study all contain > 300 ppm of REY on a whole-rock basis (Table 4 and Supplemental Material; Table S1). Fly ash sample 251 has the highest concentration of REE/REY, and MREE and HREE concentrations are nearly 2× greater than the other samples analyzed. Fly ash samples (251, 339, 345) have an LREE/HREE ratio of ~15, which is the lowest of the samples analyzed (Table 4). Fly ash samples 251, 339, and 345 have the highest total REE and REY concentrations of the samples analyzed. Fly and bottom ash samples also show enrichment of REY (as LREE, MREE, and HREE) over the noncombusted material –roof rock and coal reject– analyzed in this study. A plot of REE element concentrations normalized to upper continental crust (UCC) values is shown in the supplemental material (Fig. S1). REE values normalized to UCC are slightly enriched in light and medium REY (Sm-Dy, Y) with respect to average crustal values (Rudnick and Gao, 2003). The outlook coefficient Coutl for all samples varies from 1.0–1.5 (Table 4) and is an indication that the fly ash samples analyzed are a promising source material for REY recovery (Dai et al., 2017a).

3. Results

251

Sample ID

Table 5 Semi-quantitative XRD analysis results for flint clay roof rock (443) and heavy media cyclone reject (508).

Quartz Kaolinite Chlorite Muscovite/Illite Calcite Rutile

365

443

508

Minor Minor Minor Intermediate Trace Trace

Intermediate Minor Minor Intermediate Trace Trace

International Journal of Coal Geology 195 (2018) 362–372

S.N. Montross et al.

rich spheres were most abundant in fly ash sample 339 (medium) and in bottom ash sample 357 (minor). All samples contained calcium-bearing particles (trace to minor); however, in 345 and 357 calcium (hydr) oxides make up between 15 and 20% (minor) of the total particles. Bottom ash sample 357 and fly ash sample 345 contained minor concentrations of agglomerated particles. The most abundant and largest agglomerated particles are observed in bottom ash sample 357. Agglomerated particles, masses of smaller particles joined together by glass, slag, or mineral flocculation, are most abundant and had the largest diameter (up to 300 μm) in the bottom ash sample (357). Agglomerated particles are composed of a mixed element “slag” matrix that contains other particles of varying chemical and physical composition. An image of a large agglomerated particle in 357 and the accompanying elemental map are shown in Fig. 3. The matrix slag is composed of Al/Si/Ca and contains iron oxide and calcium oxide particles. Fly ash sample 251 contained the highest total REE concentration of all the samples (Table 4 and Supplemental Material Tables S1 and S2). However, no REE minerals were detected in the matrix or encapsulated by Al/Si/Ca slag. The majority of the spherical particles were coated in thick (> 5 μm) or fully encapsulated Al/Si/Ca slag (see Fig. 3). In some cases, calcium-rich slag encapsulates smaller aluminosilicate glass particles or becomes fused to the surface of larger glass particles. In addition, calcium sulfate crystals, identified as gypsum by XRD, were present as mineral crystals and on the surface of larger agglomerated slag particles (Fig. 4a). Aluminosilicate glass and Fe/Ti oxide crusts were adsorbed to the surface of the gypsum crystals. In addition, mixed iron‑sulfur oxides were associated with unburnt carbon phases (Fig. 4b). Size distribution of aluminosilicate glass spheres was determined using SEM image analysis and by point count methods. In general, fly ash particles have a smaller average sphere diameter than bottom ash. Bottom ash displays a bi-modal grain size distribution with maximum abundance in two size fractions, 2.0–2.9 μm and 15.0 to 24.0 μm. Average diameter for fly ash spheres is 4.3, 2.4, and 4.1 μm for samples 251, 339, and 345, respectively. The results of physical characterization - grain size, sphere size distribution, and particle types- for ash particles are shown in the supplemental material.

Table 6 Quantitative XRD analysis results for fly ash and bottom ash samples.

Quartz Mullite Hematite Magnetite Diopside Forsterite Gypsum Amorphous

251

339

345

357

Minor (11) Minor (16) Trace (< 1) Trace (< 1) ND ND ND Major (62)

Minor (8) Minor (9) Trace (3) Trace (3) ND Trace (< 1) ND Major (66)

Minor (6) Minor (7) Trace (3) Minor (8) ND Trace (< 1) ND Major (66)

Trace (2) Minor (6) Trace (4) Trace (3) Trace (2) ND Trace (< 1) Major (71)

Non-crystalline (amorphous) particles are the predominant components in all ash samples. Ash samples (251, 339, 345, 357) contain the crystalline components quartz, mullite, hematite, and magnetite in minor to trace amounts (< 25% wt%) (Table 6). Hematite, magnetite, mullite, forsterite, diopside, and gypsum are exclusive to ash samples in this study and were not detected in roof rock or reject coal samples. *Major (> 50%), intermediate (25-50%), minor (5-25%), and trace (< 5%). Values below detection limit are reported as “ND”. The measured weight percent for each phase is shown in parenthesis.

3.2. Advanced characterization of coal utilization byproducts 3.2.1. Scanning Electron microscopy and microanalysis Petrographic analysis by SEM of rock grains from samples 443 and 508 showed abundant framboidal pyrite clusters in pore spaces and between layers of the clay matrix. Chlorite grains were composed of the endmembers clinochlore (Mg, Al) and chamosite (Fe, Al). SEM analysis provides morphological and elemental data to better identify clay minerals present—specifically kaolinite (Si, Al-rich member) and chlorite (clinochlore/chamosite) — in flint clay and HMC reject samples. Iron in the reject and roof rock is primarily associated with pyrite, chlorite, or siderite, but may also be present as amorphous or poor crystalline oxides/hydroxides. Individual minerals identified by SEM-EDS analysis are quartz, muscovite, clinochlore, monazite, xenotime, pyrite, siderite, barite, titanium oxide, ilmenite, sphalerite, and zircon (see examples in Fig. 2). Ilmenite was present in the sample (< 1%) and co-existed with nickel cobalt sulfide and sphalerite ((Zn,Fe)S). Rock samples contained trace carbon in the form of coal streaks, and no calcium-rich mineral phases were identified in the coal reject or overburden samples. Fly ash and bottom ash samples contained amorphous glass particles, mixed element slag, solid non-spherical particles, cenospheres (hollow and spherical Al/Si particles), and solid spherical particles (Fig. 2). Organic particles were present in ash samples and consisted of unburnt carbon particles and char plerospheres (carbon cage). Particles had surface coatings/crusts common for multiple particles joined together by mineral flocculation. XRD analysis showed that non-crystalline (amorphous) particles comprised a majority (> 62%) of the material in ash samples (see Table 7). SEM-EDS analysis further corroborates the chemical make-up of the amorphous glass, which is primarily composed of Al, Si, and O, and thus generally referred to as aluminosilicate (Al/Si) glass. In addition, cenospheres –hollow Al/Si glass spheres– and solid Al/Si glass spheres/pellets were present in all samples (see Fig. 2 c, e). Mixed element slag was found in all samples and generally contained some proportion of Al, Si, Ca, K, and Fe. Furthermore, Ca was distributed throughout the sample in a variety of different particles such as porous, irregularly shaped grains or solid mineral slag pellets. Porous CaO-rich grains were filled with smaller cenospheres and solid particles, and appeared to have a fibrous micro-texture. Particles in the byproduct samples are classified by morphology and chemical composition. Fly ash samples 251 and 339 contained minor amounts of mixed Fe-oxide and aluminosilicate spheres. Iron‑titanium-

3.3. Rare earth elements in coal utilization byproducts Rare earth-bearing phosphate minerals rhabdophane, monazite, xenotime, and apatite were identified in rock and ash samples via SEM and electron microprobe analysis (EPMA). Rhabdophane (hydrous monazite-group mineral) and monazite have a similar chemical composition (Dai et al., 2017b). However, rhabdophane contains small irregular channels that contain water (Mooney, 1950). Six REY-phosphate grains in sample 345 were analyzed by EPMA. Minerals with appearance and chemical composition similar to monazite but with an EPMA measured total oxide wt% of less than < 96.5%, due to the presence of water, were classified as rhabdophane (Nagy et al., 2002). HMC reject (508), flint clay (443), and fly ash (345) contain both monazite and rhabdophane (Table 7, Figs. 5, 6). REE-bearing phosphate minerals generally occurred as crystals in the rock pore space and ash matrix. Other mineral phases in the sample containing REE are florencite, pyrite, calcium oxide, and zircon. A summary of the rare earthbearing mineral phases in CUB samples is shown in Table 7. 3.3.1. Rare earth elements in reject coal and overburden materials Rare earth element-bearing minerals in samples 508 and 443 ranged in size from 1 to 20 μm long, but were typically observed as smaller 1–5 μm florets of rhabdophane and xenotime. REY mineral phases are present in pore space within the rock matrix adjacent to clinochlore, muscovite, and kaolinite mineral grains (Fig. 5). Monazite grains were also present as individual mineral crystals in the fine portion of the sample. These grains were likely dislodged from larger rock grains 366

International Journal of Coal Geology 195 (2018) 362–372

S.N. Montross et al.

Fig. 2. SEM-BSE images of representative particles in roof rock, fly ash, and bottom ash samples. (a) Titanium oxide and quartz in clay matrix 443, (b) framboidal pyrite in coal in matrix clay, (c) fly ash particles from sample 345 in cross-section, (d) agglomerated slag particle from sample 345 in crosssection, (e) amorphous aluminosilicate glass particle containing cenospheres, (f) spherical and irregular shaped amorphous aluminosilicate glass particles (dull gray) with Fe, Ti oxide spheres and crust (bright white), (g) fibrous calcium oxide with Al, Si spheres and large iron oxide pellet in cross-section, and (h) aluminosilicate glass sphere with iron and titanium oxide (bright white).

detected in Ca-rich slag. Monazite was observed as isolated crystals in the ash matrix in multiple fields of view in sample 345 (see Fig. 7). Due to the heterogeneous nature of fly ash, data acquisition and image processing using pixel/object classification and segmentation allowed for the identification of REE mineral phases (e.g., monazite) in the fly ash matrix (Fig. 7). Phase identification by image analysis was used to determine the distribution of REE-hosted phases in fly ash sample 345 (Fig. 7). Automated SEM imaging and elemental map acquisition combined with pixel classification provides quantitative results on the occurrence of rare earth element-bearing mineral phases in the sample. This technique resulted in the identification of 30 grains composed of REE-PO4 and a montaged area of 0.82 mm2,and provides a visualization of how REE-hosted phases are distributed throughout the fly ash sample.

Table 7 Rare earth element-bearing mineral phases in coal byproducts (as determined by SEM and EPMA analysis). Mineral (REE present) Monazite (Ce, La, Nd) Rhabdophane (Ce, La, Nd) Xenotime (Y, Dy, Er, Yb) Florencite (Ce, La, Nd) Apatite (La, Ce, Nd) Sulfides (La, Ni, Co) Calcium oxide (Ce, Y, Yb) Zircon (no REE by EDS) Zircon (Y, Yb by EDS, Ce, Er, Dy by CL)

251

339

345

357

X

X X X

X

X X

X

X X X

X X

443

508

X X X X X X

X X X

X X

X

during prep plant pulverization. Pore spaces filled with pyrite did not contain monazite or xenotime. Trace minerals present in reject coal (508) and overburden samples (443) were zircon, barite, sphalerite, and galena. Framboidal pyrite grains and associated pore-filling clay contained approximately 1% La and Hf, and in some cases Y, Yb, and Er were detected in zircon. No REE was detected in sphalerite, galena, or barite, or in the clay grains adjacent to these minerals. However, SEMbased cathodoluminescence analysis of mineral particles in fly ash sample 345 identified Ce, Dy, and Er associated with zircon.

3.3.3. 3D characterization of REE by image analysis The application of volume segmentation from FIB-SEM stacks to study the microstructure of the mineral matrix and REE phases was explored for three samples: Flint Clay roof rock (sample 443), coal reject (sample 508), and fly ash (sample 345). Image analysis of FIB-SEM volumes is an effective technique to calculate the mineral volume fraction and grain density, and incorporate standardized SEM-EDS elemental concentrations (wt%) for minerals occurring in the micron to sub-micron size range. The numbers reported are based on a microscale and estimates assuming equal distribution in upscaling to predict economical value. This technique was used to image the volumetric phases in Flint Clay Roof Rock (sample 443) and the estimate the masses of specific

3.3.2. Rare earth elements in fly ash and bottom ash Rare earth phosphate minerals (REY-PO4) rhabdophane, monazite, and xenotime were observed as individual crystals in the fly ash matrix or encapsulated by aluminosilicate glass or slag (Fig. 6). REEs were also 367

International Journal of Coal Geology 195 (2018) 362–372

S.N. Montross et al.

Fig. 3. SEM-BSE image and a EDS cameo image of an agglomerated particle from bottom ash (357) composed of Al,Si,Ca slag (brown-red) containing Fe-oxide (blue) and Ca-rich (green) particles. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

glass (1:10.5) were determined by image analysis on the segmented images. To calculate the mass of glass and zircon we used the volume of each phase determined by image analysis and the density of glass and zircon, 2.65 × 103 kg/m3 and 4.65 × 103 kg/m3, respectively (values from Engineering Toolbox, 2009). The masses of the glass and zircon components were calculated using the volume from 3D image analysis and density of the mineral phase of interest. The mass of zircon mineral (as ZrSiO4) in this particular particle of glass is 0.2 ng, and total elemental Zr is 0.14 ng. Using the calculated mass of zircon and glass within the particle it is possible to back calculate the density of the entire glass particle and encapsulated mineral phase (e.g., glass + encapsulated zircon). The density of the zircon-glass particle shown in Fig. 9c is 2.71 × 103 kg/m3 compared to 2.66 × 103 kg/m3, the density of a glass particle with the same dimensions but without encapsulated zircon.

REEs in the mineral phase. Fig. 8 shows the total segmented volume is occupied by 8.8% monazite, 3% organic matter, 0.4% pore space, and 88% matrix and inorganic minerals (e.g., clay, quartz, hematite, and pyrite). These volume percentages were used to calculate the mass fraction of REE based on the chemical composition provided by standardized SEM-EDS prior to milling. The 3D volume of the monazite is 66.4 μm3 out of a total sample volume of 753 μm3. Assuming an average monazite density of 5.15 × 103 kg/m3 (Engineering Toolbox Basics, 2009), the sub-sample has a calculated mass of 0.091 ng Ce, 0.0385 ng La, and 0.0315 ng Nd. This microscale-calculated mass was scaled up to a mineable mass per volume of 214 kg/m3 of Ce, La, and Nd, assuming homogenous distribution based on 2D SEM object classification. The same technique and data reduction was applied to the prep plant-HMC reject (sample 508); the total volume was measured to be 2126.6 μm3,of which 5% is monazite, 0.06% pore space, and 94% matrix and inorganic minerals (e.g., clay, quartz, hematite, pyrite). The monazite had a segmented volume of 187.09 μm3 and a calculated mass of 0.26 ng Ce, 0.11 ng La, and 0.09 ng Nd. The extrapolated total mass of Ce, La, and Nd is upscaled and, assuming homogeneity, the bulk HMC coal reject is calculated to be 216 kg/m3. FIB-SEM tomography of ash particles (sample #345) was used to quantify the number, volume, and weight % of elements encapsulated in a glass particle (Fig. 9 a, b). The morphology and composition of the glass particles and encapsulated zircon grains is shown in Fig. 9c. Segmentation of the encapsulated phase (Fig. 9 b-d) revealed 79 individual grains of zircon -with REE - encapsulated in one glass sphere that is approximately 25 μm in diameter. Fig. 9d shows the distribution of individual zircon crystals separated from the glass. The volume of the glass sphere (3.04 × 103 μm3) and volume fraction ratio of zircon to

4. Discussion The characterization of rock and pulverized coal combustion ash samples suggests much of the REE content in the samples may be too dispersed to quantify via XRD and SEM analysis of a powdered sample. Bulk elemental analysis of the samples by inductively coupled plasma mass spectrometry (ICP-MS) indicates that they have > 300 ppm REE + Y. The lack of monomineralic REE mineral grains observed by SEM, and evidence for the formation of post-combustion secondary REE minerals in some fly and bottom ash samples, supports the observation that trace REE-bearing minerals (e.g., rhabdophane, monazite, zircon, and ilmenite) may be encapsulated within the glass phases of the ash materials. The encapsulation hinders observation and quantification of

Fig. 4. SEM-BSE image of particles from bottom ash sample 357 in grain mount. (a) Gypsum (Gyp) with Ti, Fe oxide crust and Al, Si glass spheres. (b) Unburnt carbon particle with iron, sulfur, oxygen mineralization. 368

International Journal of Coal Geology 195 (2018) 362–372

S.N. Montross et al.

Fig. 5. SEM-BSE images of REE-phosphate minerals monazite in matrix muscovite from reject coal sample 508. Mineral abbreviations are Qz = quartz, Ms. = muscovite, Chl = chlorite, Chm = chamosite, Rbd = rhabdophane, Mnz = monazite, Sd = siderite, and Zrn = zircon.

thermal decomposition during coal combustion could act as a nucleation site for the agglomeration of mixed element slag or glass during condensation and particle formation outside of the boiler (Meij, 1993). This process could lead to encapsulation of the REE mineral or other mineral phase with a melting point greater than ~1400 °C in the neoformed glass (as shown in Figs. 6 and 7) and prevent the extraction of REE without total dissolution of the slag and/or aluminosilicate glass. While refractory-type minerals can persist in the post-combustion ashes, there remains a possibility for the redistribution of REE from authigenic phases into secondary phases (e.g., glass spheres, slag, coatings on particles). This process likely occurs after REE bound to clay and carbonates is released during combustion. The analyses presented indicate that secondary (Al, Si, and Fe) oxides in the ash do not contain measurable REE; however, the results presented via SHRIMP-RG ion microprobe analyses in Kolker et al. (2017) support the potential for this mechanism. Our SEM-EDS results using REE standards are comparable to results from SEM-WDS or EMPA. SEM-EDS analysis has identified Ce, La, and heavy REE (e.g., Y and Yb) in low concentrations (< 2 wt. percent) associated with Ca-oxide in fly ash samples 251 and 345. Since REE has a higher exchange rate with Ca, it is plausible that amorphous Ca-oxides formed during the combustion process could incorporate REE. Hower et al. (2013) reviewed the formation and distribution of elements in ash particles produced during combustion. Following this work, we describe key mineral reactions and transformations that likely control the chemical composition of the ash material and provide insight into the occurrence of REE in coal utilization byproducts:

such grains, if present, unless FIB-SEM is performed to capture this phenomenon. It may also be that in some samples, REEs are diffusely distributed within the ash matrix and glass phases (as described in Kolker et al., 2017; Thompson et al., 2018), which is also difficult to detect using the current microanalytical techniques described. In the samples that contained identifiable REE phases (samples 443, 508, and 345), rhabdophane, monazite, and xenotime were the predominant rare earth mineral phases present. These results are consistent with prior studies of domestic coal and coal utilization byproducts (Brownfield et al., 2005; Hower et al., 2013). Fly ash, coal reject, and roof rock samples all have HREE:LREE ratios between 0.4 and 0.6. This is critical, as heavy REE are higher in demand and economic value and are therefore a primary interest to REE recovery. The structure and thermal resistance of the source material (and potential corresponding REEs) could have implications for the viability of extraction processes. Rare earth minerals are embedded within the pore space of phyllosilicate minerals in rock samples, and REE tends to be co-located with clay grains (See Figs. 5 and 8). On the other hand, monazite is present in ash samples as individual crystals in the ash matrix or fully encapsulated in glass (Figs. 6 and 7). Based on the SEM results, REE-phosphate mineral phases, specifically monazite and xenotime, appear to be unaffected by the combustion process in these samples and remain as crystalline minerals in the ash matrix. The melting point of natural monazite ranges from 1916 to 2072 °C depending on which lanthanides are present (Sm < < La) (Hikichi and Nomura, 1987). Combustion of pulverized coal typically occurs at 1400–1600 °C (IEA Clean Coal Center, 2010); therefore, it is plausible to assume that a large proportion of monazite and xenotime will remain in the ash as unaltered mineral solids. Minerals resistant to

1. Quartz- crystalline quartz persists in the byproducts from the

Fig. 6. REE mineral phases in fly ash samples. Mineral abbreviations: Ilm = Ilmenite, Zrn = zircon, Rbd = rhabdophane, and Mnz = monazite. 369

International Journal of Coal Geology 195 (2018) 362–372

S.N. Montross et al.

Fig. 7. (Left) SEM-BSE image of fly ash particles from sample 345. The area within the green box was used for initial pixel/object classification. (Right) Mineral phases of fly ash (segmented image) are Al/ Si-rich particles (blue), Fe-oxide (red), Rhabdophane/monazite (yellow)), and CaO-rich (pink). The montaged full field of view analyzed is x = 985 μm, y = 850 μm.

Fig. 8. Left: TLD-BSE 3D reconstruction of monazite (green) in Flint Clay Roof Rock (#443) with associated organics (purple) and pore space (white). Right: SEM-BSE images of accessory minerals in sample 443. (a) Hematite, (b) xenotime with Gd, Dy, Er, Yb and secondary electron image of the mineral (inset of b), (c) pyrite with Ni/Co enrichment and La, Ga, Zr, and (d) zinc sulfide in the flint clay roof rock matrix. Mineral abbreviations: Qz = quartz, Hem = hematite, Ms. = muscovite, Clc = clinochlore, Kln = kaolinite, Py = pyrite, Ilm = ilmenite, Hem = hematite, Xtm = xenotime. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

glass material; 4. Slag- Combination of Al, Si, K, Fe, and Ca from the decomposition of feldspars and clay (K from illite, or K-silicates if present, Fe from the breakdown of pyritic minerals or siderite in the coal feed, and Ca is from calcite or Ca-rich aluminosilicate in the coal feed);

feedstock; 2. Mullite- formed from thermal breakdown of aluminosilicate minerals in the coal feed; 3. Aluminosilicate glass- Al and Si from the combustion of inorganic minerals in the coal, such as feldspars and clays, to form amorphous

Fig. 9. FIB-SEM images and 3D reconstruction of ash particles and mineral phases in fly ash (#345). (a) Segmented SE-ETD images, (b) 3D reconstruction of ash particles with zircon (ZrSiO4) mineral grains (in blue) segmented from the surrounding aluminosilicate glass (c) 3D reconstruction showing the glass sphere surface and zircon (in blue), (d) zircon grains segmented from the glass. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 370

International Journal of Coal Geology 195 (2018) 362–372

S.N. Montross et al.

for FIB-SEM sample prep and imaging. Lastly, we thank Michael Outrequin and Anne-Sophie Robbes from Ametek Inc./Cameca for electron microprobe analysis.

5. Rhabdophane, monazite, zircon, ilmentite, barite—Carried over from coal feedstock into the ash as monomineralic grains. Volume and mass calculations using data from FIB-SEM analysis constrains the density changes associated with aluminosilicate glass that has fully encapsulated mineral phases (e.g. zircon) persisting during coal combustion. A glass with heavy mineral encapsulation may have a density on the order of ~1.7× greater than glass with no encapsulation, depending on the size of the encapsulated grain. The collection and recovery of other refractory minerals would behave similarly. The density separation is applicable to monazite, xenotime, and rutile, all of which have melting temperatures that exceed the temperature for PC coal combustion. Therefore, it may be postulated that when density increases, a separation of glass containing encapsulated minerals could be achieved. As suggested by Kolker et al. (2017), the separation and recovery of the dense fraction of aluminosilicate glass may reduce the amount of material that has to be subjected to treatment for REE recovery. Our methods make it possible to deduce the average density of ash particles that contain a certain level of heavy mineral resource (e.g., mass of element of interest) that could be targeted and recovered via density separation.

Disclaimer This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. Appendix A. Supplementary data

5. Conclusions

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.coal.2018.06.018.

The microscopy and microanalysis results presented here provide a quantitative characterization of the types of REE minerals present and the association between REE mineral phases and other ash particles. These analyses pinpoint key mineralogical and geochemical relationships that aid in the interpretation of bulk analyses (e.g., bulk ICP-MS results) used for resource evaluation. The workflow established an efficient way to identify and quantify REE mineral grains. Automated SEM-BSE imaging at a resolution of 16.07 pixels/μm at magnifications up to 5000× with full elemental data produced detailed, high-resolution image stacks and montages. The image analysis techniques used in this study permit numerous (up to 1024) SEM fields of view to be simultaneously analyzed for specific minerals or elements of interest, and expands the total area of sample that can be scanned for REE minerals. In addition, analysis of FIB-SEM volumes yielded quantitative data on the volume of monazite grains in coal refuse. This technique demonstrates the ability to analyze and quantify the volume of REE mineral grains in a complex matrix, as well as how calculations of encapsulation of the REE-rich grains in glass may impact the density of the amorphous ash. The volumetric reconstruction in combination with elemental measurements for minerals of interest can be used to better constrain resource estimates in both natural and engineered materials. Monazite (LREE-PO4) and xenotime (HREE-PO4) coexist as isolated crystals (< 10 μm in size) with muscovite, kaolinite, and other phyllosilicates within parent rock, coal refuse, and glass/slag. The presence of crystalline monazite and xenotime in the ash indicates that combustion has a limited impact on these phosphates, thus CUBs may require rigorous extraction methods. Bulk chemistry (ICP-MS) indicates CUB samples contained high REE, which suggests that a large proportion of the REE may be encapsulated in amorphous glass or diffusely adsorbed to the surfaces of minerals or phases.

References American Coal Association, 2014. Coal Combustion Product (CCP) Production and Use Survey Report. https://www.acaa-usa.org/publications/productionusereports.aspx (accessed 10 Sept. 2017). Bank, T., Roth, E., Tinker, P., Granite, E., 2016. Analysis of rare earth elements in geologic samples using inductively coupled plasma mass spectrometry. In: US DOE Topical Report-DOE/NETL-2016/1794. Benson, S.A., Hurley, J.P., Zygarlicke, C.J., Steadman, E.D., Erickson, T.A., 1993. Predicting ash behavior in utility boilers. Energy and Fuels. 7, 746–754. Brownfield, M.E., Cathcart, J.D., Affolter, R.H., Brownfield, I.K., Rice, C.A., O'Connor, J.T., Zielinski, R.A., Bullock Jr., J.H., Hower, J.C., Meeker, G.P., 2005. Characterization and modes of occurrence of elements in feed coal and coal combustion products from a power plant utilizing low-sulfur coal from the Powder River Basin, Wyoming. In: U.S. Geological Survey Scientific Investigations Report 20045271, . http://pubs.usgs.gov/sir/2004/5271/36pp. Curtis, C.D., Spears, D.A., 1970. Diagenetic development of kaolinite. Clays and Clay Min. 19, 219–227. Dai, S., Finkelman, R., 2018. Coal as a promising source of critical elements: progress and future prospects. Int. J. of Coal Geology. 186, 155–164. Dai, S., Chekryzhov, I.Y., Seredin, V.V., Nechaev, V.P., Graham, I.T., Hower, J.C., Ward, C., Ren, D., Wang, X., 2016. Metalliferous coal deposits in East Asia (Primorye of Russia and South China): a review of geodynamic controls and styles of mineralization. Gondwana Res. 29, 60–82. Dai, S., Xie, P., Jia, S., Ward, C.R., Hower, J.C., Yan, X., 2017a. Enrichment of U-Re-V-CrSe and rare earth elements in the Late Permian coals of the Moxinpo Coalfield, Chongqing China: genetic implications from geochemical and mineralogical data. Ore Geol Rev. 80, 1–17. Dai, S., Ward, C., Graham, I.T., French, D., Hower, J.C., Zhao, I., Wang, X., 2017b. Altered volcanic ashes in coal and coal-bearing sequences A review of their nature and significance. Earth Sci. Rev. 175, 44–74. Ekmann, J.M., 2012. Rare Earth Elements in Coal Deposits – A Prospectivity Analysis, Search and Discovery Article #80270. In: Adapted from Poster Presentation Given at AAPG Eastern Section Meeting, Cleveland, Ohio, 22–26 September 2012. Engineering Toolbox Basics, 2009. Basics. https://www.engineeringtoolbox.com/si-unitsengineering-t_19.html Accessed 02 Feb. 2018. Franus, W., Wiatros-Motyka, M.M., Wdowin, M., 2015. Coal fly ash as a resource for rare earth elements. Environ Sci Pollut Res. 22, 9464–9474. Hikichi, Y., Nomura, T., 1987. Melting temperatures of monazite and xenotime. J. Am. Ceram. Soc. 70, 252–253. Hower, J.C., Ruppert, L.F., Eble, C.F., 1999. Lanthanide, yittrium, and zirconium anomolies in the Fire Clay coal bed Eastern Kentucky. Intl. J. Coal Geol. 39, 141–153. Hower, J.C., Groppo, J.G., Joshi, P., Dai, S., Moecher, D.P., Johnston, M.N., 2013. Location of cerium in coal-combustion fly ashes: implications for recovery of lanthanides. Coal Combustion and Gasification Prod. 5, 73–78. ICDD, 2016. PDF-4+ 2016 (Database). In: Kabekkodu, Soorya (Ed.), International Centre for Diffraction Data. Newtown Square, PA, USA. IEA Clean Coal Center, 2010. Pulverized Coal Combustion. http://www.iea-coal.org, Accessed date: 1 October 2017. Ketris, M.P., Yudovich, Ya.E., 2009. Estimations of Clarkes for carbonaceous biolithes: world averages for trace element contents in black shales and coals. Intl. J. Coal Geol. 78, 135–148.

Acknowledgements This technical effort was performed in support of the National Energy Technology Laboratory's ongoing research for the Department of Energy Rare Earth Elements Program. We thank Mary Anne Alvin (DOE Rare Earths Technology Manager) and Thomas Tarka (REE FWP Technical Portfolio Lead) for their support. We thank Evan Granite and Elliot Roth (NETL-Pittsburgh) for distributing samples. Peter Hsieh (NETL-Albany) for technical review and Amy Plechacek (NETL-Albany) for assistance with revisions and editing. We thank Theresa Sawyer (Oregon State University Electron Microscopy Facility) and Jan Goral 371

International Journal of Coal Geology 195 (2018) 362–372

S.N. Montross et al.

1–64. Rudnick, R.L., Gao, S., 2018. Composition of the continental crust. Treatise Geochem. 3, 1–64. Schatzel, S.J., Stewart, B.W., 2003. Rare earth element sources and modification in the lower Kittanning bed Pennsylvania: implications for the origin of mineral matter and rare earth element exposure in underground mines. Intl. J. Coal Geol. 54, 223–251. Schindelin, J., Arganda-Carreras, I., Frise, E., et al., 2012. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682. Seames, W.S., 2013. An initial study of the fine fragmentation fly ash particle mode generated during pulverized coal combustion. Fuel Process. Technol. 81, 109–125. Seredin, V.V., 1996. Rare earth element-bearing coals from the Russian far east deposits. Int. J. Coal Geol. 30, 101–129. Seredin, V.V., Dai, S., 2012. Coal deposits as potential alternative sources for lanthanides and yittrium. Intl. J. Coal Geol. 94, 67–93. Taggart, R.K., Hower, J.C., Dwyer, G.S., Hsu-Kim, H., 2016. Trends in rare earth element content of U.S.-based coal combustion fly ashes. Env. Sci and Tech. 50, 5919–5926. Thompson, R., Bank, T, Montross, S., Roth, E., Howard, B., Verba, C., Granite, E., 2018. Analysis of rare earth elements in coal fly ash using laser ablation inductively coupled plasma mass spectrometry and scanning electron microscopy. Spectrochim. Acta Part B. 143, 1–11. United States Energy Information Administration (EIA), 2016. Monthly Energy Review. http://www.eia.gov/totalenergy/data/monthly/#coal (accessed 2 February 2018). Vassilev, S.V., Vassileva, C.G., 2005. Methods for characterization of composition of fly ashes from coal-fired power stations: a critical overview. Energy Fuel 19, 1084–1098. Warwick, P.D., Crowley, S.S., Ruppert, L.F., Pontolillo, J., 1997. Petrography and geochemistry of selected lignite beds in the Gibbons Creek mine (manning formation, Jackson Group, Paleocene) of East Central Texas. Int. J. Coal Geol. 34, 307–326.

Kolker, A., Scott, C., Hower, J.C., Vazquez, J.A., Lopano, C.L., Dai, S., 2017. Distribution of rare earth elements in coal combustion fly ash, determined by SHRIMP-RG ion microprobe. Intl. J. Coal Geol. 184, 1–10. Kutchko, B., Kim, A., 2006. Fly ash characterization by SEM-EDS. Fuel 85, 2537–2544. Lin, R., Soong, Y., Granite, E., 2018. Evaluation of trace elements in U.S. coals using the USGS COALQUAL database version 3.0. Part II: Non-REY Critical Elements. 192, 39–50. Luttrell, G.H., Kiser, M.J., Yoon, ROE-Hoan, Bhagavatula, A., Rezaee, M., Honaker, R.Q., 2016. Concentrations of rare earth elements generated by U.S. coal preparation plants. In: Conference Paper: Coal Prep Conference, Louisville, KY. Mardon, S.M., Hower, J.C., 2004. Impact of coal properties on coal combustion by product quality: examples from a Kentucky power plant. Intl. J. Coal Geol. 59, 153–169. Meij, R., 1993. Trace element behavior in coal fired power plants. Fuel Process. Technol. 39, 199–217. Mooney, R.C., 1950. X-ray diffraction study of cerous phosphate and related crystals I Hexagonal modification. Acta Crystallogr. 3, 337–340. Nagy, G., Dragantis, E., Demeny, A., Panto, G., Arkai, P., 2002. Genesis and transformations of monazite, florencite, and rhabdophane during medium grade metamorphism: examples from the Sopron Hills, Eastern Alps. Chem. Geol. 191, 25–46. National Energy Technology Laboratory Energy Data Exchange (NETL-EDX), 2018. Rare earth elements from coal and coal by-products. https://edx.netl.doe.gov/ree/, Accessed date: October 2017. Rozelle, P.L., Khadilkar, A.B., Pulati, N., Soundarrajan, N., Klima, M.S., Mosser, M.M., Miller, C., Pisupati, S.V., 2016. A study on removal of rare earth elements from U.S. coal byproducts by ion exchange. Metallurgical and Materials Transactions E. 3, 6–17. Rudnick, R.L., Gao, S., 2003. Composition of the continental crust. Treatise Geochem. 3,

372