Enrichment of Rare Earth Elements from Coal and Coal By-Products by

Physical Separations

Ronghong Lin a,*, Bret Howard a, Elliot Roth a, Tracy Bank a,b, Evan Granite a and Yee Soong a

a
U.S. Department of Energy, National Energy Technology Laboratory, 626 Cochrans Mill Road,

P.O. Box 10940, Pittsburgh, PA 15236, United States.

b
AECOM, Pittsburgh, PA 15236, USA.

* Corresponding author: 1-412-386-5064; ronghong.lin@netl.doe.gov

1
Abstract

Rare earth elements (REE) are of strategic importance because they find numerous applications

in various sectors of the global economy. The concern about the REE supply challenge has led to

increasing interest and research in the recovery of REE from end-of-life products and secondary

sources such as coal and coal by-products. The work reported here was focused on examining the

technical feasibility of physical separation techniques for the enrichment of REE from coal and

coal by-products. Size, magnetic and density separations were performed on clean coal, coal ash

and clay-rich shale samples. It was found that the samples responded to size separation

differently. For all ash samples, higher REE concentrations were found in the finer fractions,

indicating that REE may be more concentrated on the surface of particles. For the shale samples,

however, the REE concentrations decrease as the particle size reduces possibly because RE

minerals were not effectively released by grinding. Magnetic separation showed that REE are

enriched in non-magnetic fractions for all ash samples. All samples responded similarly to

density separation. Among the three methods, density separation showed the highest enrichment

of REE. A combination of these methods is recommended. Finally, correlations between

elements were demonstrated, which leads to the classification of three groups containing mainly

Al/Si, Fe and Ca, respectively. REE are strongly associated with the Al/Si group..

Keywords: Rare earth elements, coal, fly ash, physical separation, magnetic separation, float-

sink separation, density separation

2
 1. Introduction

Rare earth elements (REE) are a group of seventeen chemical elements of the periodic table

including fifteen lanthanides plus yttrium and scandium. The latter two are included because they

are usually found in the same ore deposits as the lanthanides and exhibit similar chemical

properties. REE are of strategic importance because they find numerous applications in various

sectors of the global economy such as clean energy production, health care, oil refining and

electronics. They are the key ingredients for magnets for wind turbines and other devices,

batteries for vehicles, phosphors for lighting, catalysts for petroleum refining and many others

[1-2]. Due to their critical roles in modern economy, both the European Commission [3] and the

U.S. Department of Energy [1] consider REE as critical raw materials. In the global market,

however, China dominates the REE supply chain of mining, separation, refining and

manufacturing, producing more than 85% of the REE raw materials and oxides [2, 4]. The

concern about the REE supply challenge has led to increasing interest and research in the

recovery of REE from end-of-life products such as permanent REE magnets, nickel metal-

hydride batteries and lamp phosphors [5] and from secondary sources such as coal and coal

mining and combustion by-products [6,7].

It is well known that coal contains REE in addition to many other trace metal elements and

that the majority of REE are associated with inorganic minerals in coal. The worldwide average

total REE concentrations (lanthanides and yttrium) are 68.5 ppm and 404 ppm for coal and coal

ash, respectively [7,8]. Recent studies [9-11] showed that the beneficial recovery of REE from

coal requires a minimum rare earth oxide (REO) concentration of 800–900 ppm in the coal ash.

Based on this datum, Zhang et al. [7] further estimated the cut-off grade of REE in coal which is

115-130 ppm on the whole coal basis and 677-762 ppm on the ash basis. An analysis of the U.S.

3
Coal Quality Database showed that greater than 700 ppm (ash basis) REE are distributed in the

U.S. coal basins [7]. Also, a study by Rozelle et al. [12] suggested that U.S. coal byproducts may

be technically suitable as REE ores. Advantages of utilizing coal and coal by-products as

feedstocks for REE production include: (1) large and reliable sources, (2) already mined

materials (no need for new mining permits), (3) potential environmental and health benefits from

no new mining, and (4) utilizing potential waste materials. Therefore, coal and coal by-products

are promising alternative sources for REE recovery. Development of processes and technologies

for the economic recovery of REE from coal and coal by-products would be game-changing and

beneficial, particularly to countries like the U.S. that rely heavily on REE imports. Incorporation

of the REE recovery process into the coal mining, processing and combustion processes, on the

other hand, would also benefit the society by using coal by-products and thus reducing the

environmental impacts of these by-products. .

Physical separation methods have been widely applied in the traditional processes for the

beneficiation of RE minerals from conventional RE ore deposits. These methods mainly include

size separation, magnetic separation, density (or gravity, float-sink) separation, electrostatic

separation and flotation. Limited applications of these methods in the recovery of REE from coal

and coal by-products have been reported in recent studies [13-16]. Dai et al. [13] investigated the

abundance and distribution of minerals and elements in high-alumina coal fly ash from a Chinese

coal power plant. The fly ash was first separated into three fractions, namely magnetic, MCQ

(mullite + corundum + quartz) and glass phases. It was found that REE were enriched in the

glass phase but depleted in the magnetic and MCQ phases and heavy REE (HREE) in the glass

phase were more enriched than light REE (LREE). In the magnetic and MCQ phases, however,

the LREE were more enriched than HREE despite of the overall depletion. They further

4
examined elemental distribution in size-fractioned fly ash and found that there was a significant

increase in the REE concentrations with the decrease in particle size. They also noticed that the

enrichment factors for LREE were greater than those of the HREE as the particle size decreases.

Studies by Dai et al. [14], Hower et al. [15] and Blissett et al. [16] confirmed that the REE

concentrations in coal ash increase as the particle size decreases. Hower et al. [15] also found

that LREE/HREE showed a minimum in the 160 × 300 mesh fraction and a maximum in the

finest fraction. Blissett et al. [16] proposed and analyzed a pilot-scale multi-stage REE

enrichment scheme composed of froth flotation, magnetic separation and hydrocyclone

separation. They found a slight enrichment of REY (lanthanides + yttrium) from 419 to 529 ppm

and suggested that the technical feasibility of using such a process for enrichment and extraction

of REE should be further evaluated.

Recent projects funded by the U.S. Department of Energy (DOE) National Energy

Technology Laboratory (NETL) explored the technical feasibility of using a variety of physical

separation techniques for REE enrichment from coal and coal by-products [17-20]. The

University of Kentucky and Virginia Tech (UK&VT) team [17] tested six different separation

techniques including table concentration, enhanced gravity concentration, wet high intensity

magnetic separation, electrostatic separation, froth flotation and agglomeration on coal and coal

by-product samples. They found that froth flotation was the only method that could be used for

REE enrichment. Density-based separators were the least effective methods due to ultrafine grain

size of RE minerals. Higher REE concentrations were found in the magnetic fractions of all

samples. However, this method was not effective considering extremely low mass recovery. The

University of Utah team [19] examined magnetic, flotation and float-sink separations on Western

U.S. coal samples. They found that the wet high intensity magnetic separation method was

5
effective for the removal of the pyritic content of the coal, but they concluded that this method

could not be used for the efficient recovery of REE from coal for the same reason as the UK&VT

team determined. They found that the froth flotation method was ineffective in concentrating

REE from fine coal of less than 100 mesh due to the poor liberation of RE minerals which are

less than 5 µm and uniformly dispersed in the coal matrix. They also found that density-based

mineral concentration methods could be used for concentrating REE and the recoveries could be

increased by decreasing particle sizes. The Pennsylvania State University team [20] evaluated

float-sink, magnetic and electrostatic separation methods for the enrichment of REE from coal

and coal by-products including roof rock, pit cleanings and coal refuse samples. They found that

the roof rock samples showed a best response to gravity separation, while the pit cleaning sample

and one of the refuse samples responded best to magnetic separation and electrostatic separation,

respectively. From these findings obtained by different research groups, it can be concluded that

different REE feedstocks respond to physical separation methods differently. Furthermore, there

are some disagreements among these studies that need to be further addressed.

The work reported here is part of the REE from Coal and Coal By-Products R&D Program,

NETL’s new initiative aiming to address the current global REE separations market and process

economics and to demonstrate the techno-economic feasibility of domestic REE separation

technologies [21]. This work was focused on examining the technical feasibility of physical

separation techniques including size, magnetic and density separations for the enrichment of

REE from clean coal, coal combustion by-products and shales.

2. Experimental

2.1 Materials

6
 Hundreds of coal and coal by-product field samples have been collected and analyzed at

NETL. Detailed information on these samples is available elsewhere [22]. Seven representative

samples were selected from them and studied in this work. They include one clean coal sample

(240), four coal ash samples (251, 339, 345 and 357) and two clay-rich shale samples (1July15-2

and 1Oct15-9). Table 1 lists a brief description of each sample, and Figure 1 shows a photo of

all seven samples. Samples 251 and 345 are fly ashes, sample 339 contains mostly fly ash with

lesser amount of bottom ash, while sample 357 contains mostly bottom ash with lesser amount of

fly ash. All samples but two fly ashes 251 and 345 were ground prior to physical separations.

Size and density separations were performed on all samples, while magnetic separation was

performed on ash samples only because the coal and clay samples are essentially non-magnetic,

as indicated in Table 1.

Lithium metatungstate, used as a separation medium in the float-sink separation experiments,

was purchased from Geoliquid, Inc., Illinois, USA. It contains by volume 79.2% lithium

metatungstate and 20.8% water and has a density of 2.95 g/ml, as provided by the supplier.

2.2 Physical Separations

Physical separation methods employed in this work for the enrichment of REE from coal and

coal by-products include size, magnetic and density separations. A brief description of the

methods follows.

Size separation of coal and coal by-products was performed using a Gilson Performer III 3-

inch sieve shaker (Gilson Company, Inc.) equipped with an electromagnetic vibrator. The shaker

is capable of separating particles down to 20 µm using a 635 mesh sieve. Sieves of desired mesh

sizes (100 – 635 mesh or 150 – 20 µm) were selected and installed, and known amounts of dry

7
samples (typically 10-30 g) were loaded onto the top sieve. The sieving operation generally

lasted 30-90 min depending on particle size and the amount of sample loaded. After separation,

samples were collected and weighed.

Magnetic separation of coal and coal by-products was carried out using a simple setup as

shown in Figure 2. An electromagnet was secured and attached to a metal stand and placed

above an adjustable lab jack. Samples were loaded on a white paper placed on the lab jack. The

magnetic field strength was manually adjusted by changing the distance between the lab jack

surface and the magnet surface as indicated by “h” in the figure. The magnetic field strength

applied to samples increases as the h value decreases. The h values used in this work were 2.0,

1.5, 1.0, 0.5 and 0.3 cm, and the corresponding magnetically-separated samples were denoted by

mag-1 (2.0 cm), mag-2 (1.5 cm), mag-3 (1.0 cm), mag-4 (0.5 cm), mag-5 (0.3 cm) and non-mag.

Density separation of coal and coal by-products was done using the traditional float-sink

method as follows. Lithium metatungstate was used as a separation medium. Known amounts of

sample were placed in 50-ml plastic centrifuge tubes and then 40 ml lithium metatungstate was

added to each tube. The tubes were shaken by hand to ensure thorough dispersion of sample

particles in the heavy liquid. Samples were then centrifuged to achieve complete float-sink

separation. Heavier solids on bottom of the tubes were carefully transferred to new 50-ml

centrifuge tubes using disposable Pasteur pipettes. Deionized water was added to the remaining

samples to adjust the liquid density to a next lower level. This procedure was repeated to obtain

several different density fractions. Density-fractioned samples were washed with deionized water

and centrifuged at least three times to minimize lithium metatungstate residue in the samples.

Finally, samples were dried at 120 oC overnight prior to elemental analysis and characterization.

8
 2.3 Analytical Methods

2.3.1 Elemental Analysis

Major, minor, and trace element concentrations were measured in the dried solid samples in

order to quantify the REE enrichment that occurred after each physical separation step. Samples

were dried at 107 oC under N2 for 2 hours and then ashed at 550 oC for 4 hours under air. Drying

and ashing was typically completed in a LECO thermogravimetric analyzer (TGA). Occasionally,

the sample size was too small for the TGA (< 0.5 g). Those samples were dried and ashed in a

muffle furnace using the same conditions. Approximately 50 mg of ashed sample was mixed in a

platinum crucible with 400 mg of lithium metaborate (LiBO2) and fused at 1100 oC for 5 minutes

in a microwave fusion oven. The fused glass was then digested in 5% HNO3 with low heat and

stirring. The platinum crucibles were rinsed in triplicate with 5% HNO3 and diluted to a final

volume of 100 ml to ensure a complete digestion of the fused glass. For quality control and

quality assurance, certified reference solids that most closely matched the starting material were

processed along with each batch of samples. Element concentrations were measured by

Inductively Coupled Plasma - Mass Spectrometry (ICP-MS) and Inductively Coupled Plasma -

Optical Emission Spectroscopy (ICP-OES). The precision and accuracy of the ICP-MS technique

and the drawbacks and limitation of this method were discussed elsewhere [23]. Carbon and

sulfur concentrations were determined on the dried but unashed solid using a 2400 CHNS/O

Series II Elemental Analyzer (Perkin Elmer).

2.3.2 XRD and SEM

Powder X-ray diffraction (XRD) analysis of the bulk mineral composition of the samples

was carried out using a PANalytical X’Pert PRO diffractometer equipped with a Cu anode

9
operated at 45 kV and 40 mA. Phase identification was verified by comparison to the

International Centre for Diffraction Data (ICDD) inorganic compound data base. Scanning

electron microscopy/energy dispersive x-ray analysis (SEM/EDS) of select samples was carried

out using an FEI Company Quanta 600 field emission scanning electron microscope equipped

with secondary and backscatter electron detectors and an Oxford Inca Energy 350 X-act energy

dispersive x-ray analyzer (EDS) to investigate minor minerals not observed by XRD and to

determine mineral associations and crystal size distribution. Samples were not coated so were

analyzed in low vacuum mode to avoid charging. A backscattered electron detector was used so

that elemental contrast could be used to more easily locate the minerals of interest.

3. Results and Discussion

3.1 Raw material characteristics

Physical and chemical characteristics of the raw materials were analyzed by TGA, ICP-MS,

ICP-OES and other methods as described in section 2.3.1. Moisture, dry ash, major, minor and

trace element contents of the samples are listed in Table 2. It should be noted that the as-

collected field samples with the exception of fly ash sample 345 and shale sample 1July15-2

have higher as-collected moisture contents than listed in Table 2. These samples were dried

overnight at 120 oC for easy grinding and processing. The ash contents vary significantly

depending on the type of the samples. The coal combustion products contain 95-98 % dry ash

except sample 251 which contains 85%. The dry ash contents of both shale samples are almost

the same at about 93%. The clean coal sample has much lower mineral content resulting in only

7.8% dry ash. The dry ash-basis REE concentration (hereafter referred to as the sum of

lanthanides concentration) also vary significantly among samples. The clean coal sample has the

10
highest ash-basis REE concentration of 1658 ppm. The REE concentration of the coal ash

samples vary from 312 to 623 ppm with sample 251 having the maximum. The two shale

samples have REE concentrations of 296 and 414 ppm, respectively. The LREE/HREE ratios are

in the range of 8.1 - 9.5 except for the shale sample 1July15-2 which has an LREE/HREE ratio

of 15.9. This shale sample, accordingly, has a low content of the most critical REE at 28%

compared with 36-39% for all other samples. Figure 3 plots the CI-chondrite-normalized REE

concentrations showing smooth patterns with the exception of Eu anomalies for all samples.

The starting materials were analyzed by XRD to determine the primary crystalline mineral

compositions of the matrix. The results are listed in Table 3 along with a qualitative estimate of

the amount of each identified mineral. As expected, the compositions fall in three general

categories; shales containing quartz, kaolinite and muscovite/illite; ashes containing mullite,

quartz, iron oxides and a large amount of amorphous silica-rich glassy material; and cleaned coal

containing quartz, kaolinite, muscovite/illite and a large amount of amorphous carbon-rich

material. Diffraction patterns for these materials are shown in Figure 4A (ash materials) and

Figure 4B (shales and cleaned coal) with the primary diffraction peaks of the major crystalline

components labeled.

SEM/EDS was used to examine the particle morphologies and attempt to identify minor

minerals of interest in a representative sample from each sample category. The samples

examined were ash 251 (as received), shale 1Oct15-9 (ground) and the clean coal sample 240 (as

received). Backscattered electron imaging was used to more easily locate REE-containing

minerals as well as other heavy element-containing minerals. Secondary electron imaging was

used to determine the mineral matrix morphology.

11
 Ash sample 251 has a particle size distribution and morphology typical of a fly ash. It

contains a wide range of particle sizes from around 50 µm to less than 1 µm and particle

morphologies from jagged and irregular to smooth and spherical. The RE minerals monazite and

xenotime were identified based on EDS. Monazite appears much more prevalent than xenotime.

In addition, the RE minerals were commonly but not always embedded in the glassy ash matrix

as can be observed by comparing backscattered versus secondary electron images of the same

particle (Figure 5 A and B).

The shale sample, 1Oct15-9, had been ground and sieved prior to examination. The minus

100 mesh (-150 µm) fraction was examined so, as would be predicted, the particles were all less

than 150 µm and ranged to less than 1 µm. Most particles consisted of agglomerates of smaller

individual mineral particles. Monazite was the only RE mineral identified in this screening

(Figure 5 C and D). Pyrite was the most prevalent heavy mineral observed in this shale. Other

minerals observed include zircon and barite.

The cleaned coal sample, 240, consisted of large coal particles up to nearly 500 µm with smaller

coal and mineral particles intermixed. Monazite was the only RE mineral identified in this

sample (Figure 5 E and F). Other heavy minerals observed include pyrite, gypsum and barite.

3.2 Size separation

Size separation was performed using a combination of 5-7 sieves. The sizes used were 100,

200, 270, 325, 400, 500 and 625 mesh (or 150, 75, 53, 45, 38, 25, 20 µm). REE concentrations

of size-fractioned samples are plotted in Figure 6 along with the weight percentage of each size

fraction. For the two fly ash samples (251 and 345), more than 97 wt% of the ash particles are

less than 150 µm and more than 60 wt% are less than 45 µm, as shown in Figure 6A&B. The

12
size distributions of both fly ash samples are very similar and show a distinct peak at 38-25 µm,

although the 25-20 µm fraction was not separated for the fly ash 345. This characteristic peak is

also observed for the other two ash samples (339 and 357) containing both fly ash and bottom

ash (Figure 6C&D). This result suggests that regardless of their sources, fly ashes have a very

similar size distribution, which may further imply the same fly ash formation mechanism. All

samples except fly ashes 251 and 345 were manually ground using a mortar and pestle prior to

physical separations. The grinding procedure resulted in very similar size distribution for the

resulting samples. For the two ash mixture samples (339 and 357), the majority of >45 µm

fraction consisted of ground bottom ash particles and the <45 µm fraction consisted primarily of

the fly ash part as indicated by the 38-25 µm peak (Figure 6C&D). The percentage of size

fractioned ground bottom ash decreases as the particle size decreases. For the coal and shale

samples, as shown in Figure 6E-G, the weight percentage of size fractioned samples increases

from the first to the second fraction and then decreases. This size distribution pattern indicates

that the grinding process applied in this work may not be able to effectively release RE minerals

from the matrix as the size of the RE mineral particles are typically much smaller based on SEM

observation.

Despite some variations, the REE concentrations of size fractioned fly ashes generally

increase as the particle size decreases (Figure 6A&B), showing moderate enrichment of REE in

the smaller ash particles. This result is in good agreement with previous findings [13-16].

Although samples 339 and 357 are mixtures of fly ash and bottom ash and the REE

concentration of each of these components was not independently determined, the analyses of the

size fractioned samples suggest that the REE concentrations of fly ash are higher than those of

the bottom ash from the same coal combustion process (Figure 6C&D). The REE concentration

13
of size fractioned clean coal generally decrease as the particle size decreases except for the 45-38

µm fraction which has the highest REE concentration (Figure 6G). For both shale samples, the

REE concentration generally decreases as the particle size decreases, despite some small

fluctuations (Figure 6E&F), which indicates that RE minerals were not effectively mobilized by

grinding. If grinding did release the small REE-containing mineral particles from the matrix, the

REE concentration would be expected to be enhanced in the smallest size fraction.

3.3 Magnetic separation

Magnetic separation was performed only on the coal ash samples as preliminary tests

indicated that the clean coal and two shale samples are essentially non-magnetic. Fly ash 345

was tested first and separated into six fractions. The ICP-MS analysis showed that the REE

concentrations of the mag-1, mag-2 and mag-3 fractions are significantly lower than those of the

mag-4, mag-5 and non-mag fractions. Therefore, mag-1 and mag-2 fractions were not separated

for the rest of coal ash samples, and only four fractions were obtained for them: mag-3

(containing mag-1 and mag-2), mag-4, mag-5 and non-mag. The REE concentrations and weight

recoveries of magnetically separated coal ash samples are plotted in Figure 7 showing that the

REE concentrations of magnetically separated fractions increase as the magnetic susceptibility

decreases from mag-1 to non-mag fractions and the non-mag fraction has the highest REE

concentrations for all samples. The REE concentrations of mag-4 and mag-5 fractions are very

close except for fly ash 251 of which the mag-4 fraction contains much less REE. Fly ash 251 is

also different from others in terms of the mass distribution of magnetically separated fractions. It

contains more than 95 wt% non-magnetic ash while the other three contain less than 50 wt%.

The iron content of magnetically separated ashes are plotted in Figure 8 showing, as expected, a

14
decrease in iron concentration from mag-1 to non-mag fractions. All these results suggest that

REE are enriched in the non-magnetic fraction and likely associated with non-iron minerals. This

result agrees well with the previous finding by Dai et al. [13] but shows a disagreement with the

work by Honaker et al. [17] where higher REE concentrations were found in the magnetic

fractions of all samples. This discrepancy possibly resulted from the differing magnetic field

intensities used in each study leading to different magnetic properties of the resulting magnetic

and non-magnetic fractions. Additional research is required to address this issue.

3.4 Density separation

The fact that different minerals have different densities forms the foundation for density-

based separation. Typical minerals in coal and coal by-products reported in literature [14, 24, 25]

and observed by SEM/EDS in this work are shown in Table 4 along with their chemical

formulas and density ranges [26]. The densities of three common RE minerals, monazite,

bastnasite and xenotime, are in the range of 3.9-5.5 g/ml. It was thus anticipated that RE minerals

would concentrate in the heaviest fraction (>2.95 g/ml) because their densities are greater than

that of the heavy liquid used in the experiment. Figure 9A-D shows REE concentrations and

weight percentages of density fractioned fly ashes. The REE concentrations of fly ashes 251 and

345 reach maximum values in the density ranges of 2.95-2.71 g/ml and 2.71-2.45 g/ml,

respectively, while those of the other two ash mixtures reach maximum values at the same

density range of 2.71-2.49 g/ml. As the density of ash particles decreases, for ashes 251 and 357,

the REE concentrations increase to maximum values and then decrease, while for ashes 339 and

345, they similarly increase, reach maximums, and then decrease, but finally they increase

slightly again. This result disagrees with the hypothesis that REE would be enriched in the

15
heaviest fraction and suggests that RE minerals are incorporated with other lower density

minerals and dispersed inside fly ash particles. Previous studies have also indicated that REE are

disbursed throughout the entire glassy fly ash particles [27]. As a result of this incorporation, the

densities of particles containing REE are reduced and density separation becomes less effective

than would be predicted. To improve the effectiveness of this method, grinding is necessary to

reduce particle size and release RE minerals from their ash particle matrix, which, however,

would require a substantial increase in energy consumption. Magnetic separation tests on the

density fractioned samples indicated that the heavier fractions (>2.95 g/ml for 251 and >2.71

g/ml for 339, 345 and 357) contain mainly magnetic fractions, which was further confirmed by

the Fe concentration seen in the ICP-MS analysis. Furthermore, Figure 9A-D also shows

different weight distributions of density fractioned ashes, which suggests that these combustion

products might have been derived from different coal sources.

Figure 9E&F shows the results for density separation of the two shale samples. These

samples have different mineral compositions as indicated by their different weight percentage

curves. The majority of the shale 1July15-2 is within the narrow density range of 2.62-2.56 g/ml.

The shale 1Oct15-9, however, contains three major fractions within the density range of 2.71-

2.49 g/ml. The highest REE concentrations were obtained from the 2.95-2.71 g/ml fraction for

both shale samples. The highest values are 1340 and 412 ppm for the shales 1July15-2 and

1Oct15-9, respectively. A negligible amount was collected at >2.95 g/ml for the shale 1July15-2.

For the shale1Oct15-9, the REE concentrations of the >2.95 g/ml fraction is much less than that

of the second heaviest fraction. This result is also an indication that RE minerals were not

effectively released from the rock matrix by the grinding method used in this study. In addition,

16
the REE concentrations generally decrease as density decreases which indicates that there is

potential for improvement in the density concentration of the RE minerals.

Figure 9G&H shows the results for density separation of the clean coal sample. The clean

coal contains only 7.8 wt% dry ash (Table 2), and as expected, the density is much lower than

coal ash and shale samples. As the density drops from >1.45g/ml to 1.34-1.30 g/ml, the ash

content reduces dramatically from 49% to 4% and the dry ash-basis REE concentration increases

significantly from 1143 ppm to 2056 ppm. As the density drops further, the ash content remains

stable at 3-5% and the REE concentration reduces slightly with some fluctuations. The highest

REE concentration (dry ash basis), 2056 ppm, was obtained from the 1.34-1.30 g/ml fraction

which makes up 30% of the coal. Figure 9H also shows a strong correlation between the REE

concentration and the ash content. The dry ash-basis REE concentration generally decreases as

the ash content increases, which can be explained by the dilution effect of non-RE minerals. The

whole coal-basis REE concentration, however, increases as the ash content increases.

3.5 Enrichment factors and LREE/HREE distribution

We have observed the enrichment of REE from coal and coal by-products by physical

separations from the above discussion. To further quantitatively characterize REE enrichment by

each separation method, we define an enrichment factor (EF) as:

∑ ( )
(1)

Where REEi is the REE concentration of the ith fraction, Wi is the weight percentage of the ith

fraction, and n is the total number of fractions. The maximum EF values for each sample by each

method are presented in Table 5 along with the corresponding REE recoveries. The REE

recovery (R) is defined by:

17
 ( ) ∑ ( )
(2)

The results indicate that density separation has an overall better performance than size and

magnetic separations. The maximum enrichment factors obtained for the size and magnetic

separations are 1.19 and 1.16, respectively, both for the ash sample 339, while that of the density

separation reaches 3.58 for the shale sample 1July15-2. Increased enrichment factors are

anticipated through a combination of these physical separation methods. Additional research is

ongoing to address this issue and results will be reported in future publications.

REE are often described as being a LREE or HREE. In this discussion, the LREE and HREE

are defined as the groups of elements of lanthanide through gadolinium (La-Gd) and terbium

through lutetium (Tb-Lu), respectively. The LREE/HREE ratios of size, magnetic and density

separated coal, coal ash and shale samples are presented in Figure 10. It was found that the

LREE/HREE ratios of all ash and shale samples generally decrease as the particle size decreases

(Figure 10A&C) and the lowest LREE/HREE ratios are found in the finest fractions. This result

disagrees with a previous study by Hower et al. [15] where the highest LREE/HREE ratio was

found in the finest fraction. For fly ash 251 (Figure 10A), this ratio drops as the particle size

increases from 75 µm to above 150 µm. This is because the larger particles contain mainly

magnetic fractions and the LREE/HREE ratios of magnetic fractions are much lower than those

of the non-magnetic fractions, as shown in Figure 10D&E. The LREE/HREE ratio of the clean

coal sample remains nearly constant over the entire size spectrum (Figure 10B). The

LREE/HREE ratios for all density fractioned samples generally reduce to minimum values and

then increase as the densities of the samples decrease, as shown in Figure 10F-I. This parameter

is directly related to REE concentrations, and as illustrated in Figure 11, it varies with REE

concentration depending on the separation method used. Particularly for the clean coal sample

18
(Figure 11G), it increases significantly as the ash-basis REE concentration decreases. However,

on the whole coal basis, it drops significantly as the REE concentration reduces. It can be

concluded that the LREE and HREE do not contribute equally to the enrichment of REE by

physical separations. The LREE contribute more than the HREE towards the enrichment. In

addition, compared with the LREE, the HREE favor finest particles and magnetic fractions. One

possible explanation of this observation is that although REE macroscopically tend to occur in

the same minerals, the chemical composition of individual RE mineral grain varies. This leads to

unequal distributions of LREE and HREE across RE minerals and thus the unequal enrichment

of LREE and HREE in physically separated coal and coal by-product samples. Additional

research is necessary to address these issues.

3.6 Elemental correlations

In addition to REEs, major (>0.1%), minor (100 ppm-0.1%) and trace (<100 ppm) elements

were also analyzed for all 146 original and physically-separated samples. The data provide useful

information on not only the concentration levels of these elements but also the correlations

between them. The Pearson product-moment correlation coefficient [24] is applied here to

characterize the correlations between elements. Results are plotted in Figure 12. According to

their correlations, elements can be classified under three groups. The first group includes Al, Si,

Ti, K, Na, V, REE, Nb, Ga, Th and U. As Al and Si dominate this group, the major minerals may

include aluminosilicates and quartz. REE are strongly associated with this group. The second

group includes Fe, Mn, Ni and Co. The elements of this group have negative correlations with

the elements in the first group, indicating that they belong to different minerals. The major

minerals in this group may include pyrite and other magnetic minerals. REE have weak positive

19
or negative correlations with elements in this group. This result agrees well with the observation

from magnetic separation that REE are enriched in non-magnetic fractions containing less

amounts of iron. The major element, Ca, has a negative correlation with Al and a weak positive

correlation with Fe, and therefore it forms the third group. This group may also include Ge, Mg,

Mo and Zn, which have relatively strong positive correlations with Ca. The major minerals of

this group may include gypsum and calcite. REE also have negative correlations with this group.

The above classification of elements agrees well with a previous study on mineralogy and

geochemistry of coals by Moore and Esmaeili except that they classified Ni as an additional

group resulting in a total of four groups [24]. Furthermore, the well-known strong correlation

between REE and Th is also observed. In addition REE also have strong correlations with U, Hf,

Sb, Cd, Nb, Zr, Ga, Ag, Cu, Cr, V and Be, as shown in Figure 12.

4. Conclusions

Physical separations including size, magnetic and density separations were performed on

seven coal and coal by-product samples to investigate the enrichment of REE in these samples

by these physical separation methods. The samples studied include one clean coal, two fly ashes,

two fly ash-plus-bottom ash mixtures, and two clay-rich shale samples. It was found that the

samples responded to size separation differently. For all ash samples, higher REE concentrations

were found in the finer fractions, indicating that REE may be more concentrated on the surface

of particles because smaller particles have larger surface area per mass. For the shale samples,

however, the REE concentrations decrease as the particle size decreases. One possible

explanation is that RE minerals were not effectively released from the matrix by grinding

because RE minerals are usually much smaller than the final particle size obtained in this work.

20
Magnetic separation showed that REE are enriched in non-magnetic fractions for all ash samples.

All samples responded similarly to density separation. The densities of RE minerals are higher

than that of the heavy liquid used in this work. They were thus expected to be concentrated in the

highest density fractions. However, the highest REE concentrations were not found in the highest

density fractions. This is another indication that RE minerals are trapped in other minerals and

not effectively released. Improved enrichment of REE would require substantial grinding to

mobilize RE minerals from coal and coal by-products. Among the three methods, density

separation showed the highest enrichment of REE, which was obtained with the shale 1July15-2

sample. The REE were enriched by a factor of 3.58 to 1341 ppm. The combination of these

methods would be beneficial to improve REE separation. However, different combinations

would be needed for the different material types. Research is ongoing at NETL to improve this

methodology. Finally, correlations between elements were demonstrated, which leads to the

classification of these elements into three groups. The strong positive correlation between REE

and Al indicates that REE in coal and coal by-products are likely associated with the Al/Si group.

Acknowledgements

This research was supported in part by appointments (R. Lin and E. Roth) to the National

Energy Technology Laboratory Research Participation Program, sponsored by the U.S.

Department of Energy and administered by the Oak Ridge Institute for Science and Education.

References

[1] U.S. Department of Energy. Critical materials strategy, DOE/PI-0009, December 2011.

[2] Humphries M. Rare earth elements: The global supply chain. December 16, 2013.

21
[3] European Commission. Report on critical raw materials for The EU. May 2014.

[4] U.S. Geological Survey. Mineral commodity summaries 2016. U.S. Geological Survey, 2016,

202 p., http://dx.doi.org/10.3133/70140094.

[5] Binnemans K, Jones PT, Blanpain B, Van Gerven T, Yang Y, Walton A, Buchert M.

Recycling of rare earths: a critical review. J Clean Prod 2013; 51: 1-22.

[6] Zhang W, Rezaee M, Bhagavatula A, Li Y, Groppo J, Honaker R. A review of the occurrence

and promising recovery methods of rare earth elements from coal and coal by-products. Int J

Coal Prep Util 2015; 35: 295-330.

[7] Seredin VV, Dai S, Sun Y, Chekryzhov IY. Coal deposits as promising sources of rare metals

for alternative power and energy-efficient technologies. Appl Geochem 2013; 31: 1-11.

[8] Ketris MP, Yudovich YaE. Estimations of clarkes for carbonaceous biolithes: world average

for trace element contents in black shales and coals. Int J Coal Geol 2009; 78: 135–148.

[9] Seredin VV, Dai S. Coal deposits as potential alternative source of lanthanides and yttrium.

Int J Coal Geol 2012; 94: 67–93.

[10] Dai S, Jiang Y, Ward CR, Gu L, Seredin VV, Liu H, Zhou D, Wang X, Sun Y, Zou J, Ren

D. Mineralogical and geochemical compositions of the coal in the Guanbanwusu Mine, Inner

Mongolia, China: Further evidence for the existence of an Al (Ga and REE) ore deposit in the

Jungar Coalfield. Int J Coal Geol 2012; 98: 10–40.

[11] Seredin, VV. The Au-PGE mineralization at the Pavlovsk brown coal deposit, Primorye.

Geol Ore Deposit 2004; 46: 36–63.

[12] Rozelle PL, Khadilkar AB, Pulati N, Soundarrajan N, Klima MS, Mosser MM, Miller CE,

Pisupati SV. A Study on Removal of Rare Earth Elements from U.S. Coal Byproducts by Ion

Exchange. Metall Mater Trans E 2016; 3: 6-17.

22
[13] Dai S, Zhao L, Peng S, Chou C-L, Wang X, Zhang Y, Li D, Sun Y. Abundances and

distribution of minerals and elements in high-alumina coal fly ash from the Jungar Power Plant,

Inner Mongolia, China. Int J Coal Geol 2010; 81: 320–332.

[14] Dai S, Zhao L, Hower JC, Johnston MN, Song W, Wang P, Zhang S. Petrology, mineralogy,

and chemistry of size-fractioned fly ash from the Jungar power plant, Inner Mongolia, China,

with emphasis on the distribution of rare earth elements. Energ Fuel 2014; 28: 1502−1514.

[15] Hower JC, Dai S, Seredin VV, Zhao L, Kostova IJ, Silva LFO, et al. A note on the

occurrence of yttrium and rare earth elements in coal combustion products. Coal Combust

Gasificat Prod 2013;5:39–47.

[16] Blissett RS, Smalley N, Rowson NA. An investigation into six coal fly ashes from the

United Kingdom and Poland to evaluate rare earth element content. Fuel 2014; 119: 236-239.

[17] Honaker R, Hower J, Eble C, Weisenfluh J, Groppo J, Rezaee M, Bhagavatula A, Luttrell

GH, Bratton RC, Kiser M, Yoon R-H. Laboratory and bench-scale testing for rare earth elements.

December 30, 2014. (Available at https://edx.netl.doe.gov/ree/?page_id=1587, accessed

2016.04.21).

[18] Akdogan G, Ghosh T. Identification of REE in some Alaskan coal and ash samples.

December 2014. (Available at https://edx.netl.doe.gov/ree/?page_id=1587, accessed 2016.04.21).

[19] Miskovic S. REE recovery from western coals. January 2015. (Available at

https://edx.netl.doe.gov/ree/?page_id=1587, accessed 2016.04.21).

[20] Soundarrajan N, Pulati N, Klima M, Ityokumbul M, Pisupati SV. Separation of rare earth

elements from coal and coal products. May 7, 2015. (Available at

https://edx.netl.doe.gov/ree/?page_id=1587, accessed 2016.04.21).

23
[21] The U.S. Department of Energy National Energy Technology Laboratory. NETL Rare Earth

Elements Program 2016 Project Portfolio. April 2016. (Available at

http://www.netl.doe.gov/research/coal/rare-earth-elements/publications, accessed 2016.04.21).

[22] The U.S. Department of Energy National Energy Technology Laboratory, Integrated

Sample Databases - Collected Samples Spreadsheet v051515. (Available at

https://edx.netl.doe.gov/ree/?page_id=1587, accessed 2016.09.14).

[23] Bank T, Roth E, Tinker P, Granite E. Analysis of rare earth elements in geologic samples

using Inductively Coupled Plasma Mass Spectrometry. US DOE Topical Report - DOE/NETL-

2016/1794, April 14, 2016.

[24] Moore F, Esmaeili A. Mineralogy and geochemistry of the coals from the Karmozd and

Kiasar coal mines, Mazandaran province, Iran. Int J Coal Geol 2012; 96-97: 9–21.

[25] Rao RB, Chattopadhyay P, Banerjee GN. Removal of iron from fly ash for ceramic and

refractory applications. Magnetic and Electrical Separation 1999; 10: 21-27.

[26] Anthony JW, Bideaux RA, Bladh KW, Nichols MC, Eds., Handbook of Mineralogy,

Mineralogical Society of America, Chantilly, VA 20151-1110, USA.

http://www.handbookofmineralogy.org/.

[27] Hower JC, Groppo JG, Joshi P, Dai S, Moecher DP, Johnston MN. Location of cerium in

coal-combustion fly ashes: implications for recovery of lanthanides. Coal Combust Gasificat

Prod 2013; 5: 73-78.

24
Figure 1. A photograph of coal and coal by-product samples. 251
- fly ash, 339 - fly ash & bottom ash, 345 - fly ash, 357 - bottom
ash & fly ash, 240 - clean coal, 1July15-2 - shale, 1Oct15-9 -
shale.
 h

Figure 2. A photograph of the magnetic separation unit. The

letter “h” indicates the distance between the magnet and the
sample loading surface.
 10000
240 251 339 345

357 1July15-2 1Oct15-9

CI-Chondrite Normalized REE

1000
Concentrations

100

10
La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 3. Chondrite-normalized REE concentrations of all

raw samples.
3000
Q M = mullite
A Q = quartz
2500 G = magnetite or substituted magnetite
H = hematite
S = silicon (added as a standard to 345)
2000
M
M M
S
M Q G
1500 357 M
S S
345
1000

339
500
251

0
5 10 15 20 25 30 35 40 45 50 55 60 65 70

6000
B Q Q = quartz
5000 K = kaolinite
I = muscovite/illite
K Q G = gypsum
4000 K
I
3000
I Q Q Q Q Q 240 Q
2000
I
KI 1July15-2
1000 G
I 1Oct15-9
0
5 10 15 20 25 30 35 40 45 50 55 60 65 70

Figure 4. XRD analysis of ash (A), coal and shale samples (B)
 Backscattered electrons

A C E

Secondary electrons

B D F

Figure 5. SEM/EDS images of ash 251 (A&B), shale 1OCt15-9

(C&D) and clean coal 240 (E&F). The bright areas indicated by
the white arrows are monazite minerals.
 A 251:FA REE Wt% B 345:FA REE Wt%

700 40 400 60
650 30 380 50
REE (ppm)

Weight %

REE (ppm)

Weight %
360 40
600 20 340 30
550 10 320 20
500 0 300 10
280 0
>150 150-75 75-45 45-38 38-20 <20
Size (micron)
Size (micron)

357:BA+FA REE Wt%

C 339:FA+BA REE Wt%
D
330 30 390 50
320 25 40
380
REE (ppm)

REE (ppm)
Weight %

Weight %
310 20 30
300 15 370
290 10 20
360 10
280 5
270 0 350 0

Size (micron) Size (micron)

E 1July15-2: shale REE Wt% F 1Oct15-9: shale REE Wt%

500 60 350 50
40
REE (ppm)

REE (ppm)
Weight %

Weight %
400 40 300 30
300 20 250 20
10
200 0 200 0
>150 150-75 75-53 53-45 45-38 38-25 150-75 75-53 53-45 45-38 38-25 25-20
Size (micron) Size (micron)

G 240: clean coal REE Wt%

1900 40
1800 30
REE (ppm)

Weight %

1700
20
1600
1500 10
1400 0
>150 150-75 75-53 53-45 45-38 38-25
Size (micron)

Figure 6. REE concentrations and weight percentages of size-

fractioned clean coal, coal ash and shale samples. FA - fly ash,
BA - bottom ash.
 A 345:FA REE Wt%
B 251:FA REE Wt%

500 50 800 100

400 40 600 80

REE (ppm)
REE (ppm)

Weight %
Weight %
300 30 60
400
200 20 40
100 10 200 20
0 0 0 0
Mag-3 Mag-4 Mag-5 Non-mag

REE Wt%
C 339:FA+BA REE Wt%
D 357:BA+FA
500 50 600 60
400 40
REE (ppm)

Weight %
REE (ppm)

Weight %

400 40
300 30
200 20 200 20
100 10
0 0 0 0
Mag-3 Mag-4 Mag-5 Non-mag Mag-3 Mag-4 Mag-5 Non-mag

Figure 7. REE concentrations and weight percentages of

magnetic-separated coal ash samples. FA-fly ash, BA-bottom
ash.
 710000
345
610000
251
Fe Concentration (ppm)

510000
339

410000 357

310000

210000

110000

10000
Mag-1 Mag-2 Mag-3 Mag-4 Mag-5 Non-mag

Figure 8. Iron concentrations of magnetically separated coal ash

samples
 A 251:FA REE Wt%
B 345:FA REE Wt%

1200 80 500 40
1000 400
REE (ppm)

REE (ppm)
Weight %

Weight %
800 60 30
300
600 40 20
400 200
200 20 100 10
0 0 0 0

Density (g/ml) Density (g/ml)

REE Wt% 357:BA+FA REE Wt%

C 339:FA+BA D
400 25 500 40
20 400
REE (ppm)

REE (ppm)
Weight %
300

Weight %
30
15 300
200 20
10 200
100 5 100 10
0 0 0 0

Density (g/ml) Density (g/ml)

E 1July15-2: shale REE Wt% F 1Oct15-9: shale REE Wt%

1500 120 500 40

100 400
REE (ppm)

Weight %
30
REE (ppm)

Weight %

1000 80 300
60 20
200
500 40 100 10
20
0 0 0 0
2.95-2.71 2.71-2.62 2.62-2.56 <2.56
Density (g/ml)
Density (g/ml)
G REE Ash Wt% H Dry ash basis Whole coal basis
2500 60 2500
2000 50
REE (ppm)

Weight %

40 2000
1500
REE (ppm)

30 1500
1000 20
500 10 1000
0 0
500
0
0 10 20 30 40 50 60
Density (g/ml) Ash Contents (wt%)

Figure 9. REE concentrations and weight percentages of density-

fractioned ashes (A-D), shales (E&F) and clean coal (G&H). FA-fly
ash, BA-bottom ash.
 A 251: FA B 240: clean coal C 1July15-2: shale
10 10 20

LREE/HREE
LREE/HREE

LREE/HREE
18
9 9
16
8 8
14
7 7 12

Size (micron) Size (micron) Size (micron)

D 251: FA E 357: BA+FA F 240: clean coal

8.5 10 15
LREE/HREE

8.3
LREE/HREE

9.8

LREE/HREE
8.1 9.6 10
7.9 9.4
7.7 9.2 5
7.5 9
0

Density (g/ml)

G 251: FA H 345: FA I 1 July 15-2: shale

10.5
10 12 22
LREE/HREE

9.5
LREE/HREE

11 LREE/HREE
9 10 17
8.5
8 9
7.5 8 12
7 7
7

Density (g/ml)
Density (g/ml) Density (g/ml)

Figure 10. LREE/HREE ratios of physically separated shale, ash

and clean coal samples.
 A 251: FA Magnetic Size Density B 345: FA Magnetic Size Density

11 12
10 11
LREE/HREE

LREE/HREE
10
9
9
8
8
7 7
6 6
0 200 400 600 800 1000 1200 0 100 200 300 400 500
REE (ppm) REE (ppm)

Magnetic Size Density Magnetic Size Density

C 339: FA+BA D 357: BA+FA

11 12
10 11
LREE/HREE

LREE/HREE
10
9
9
8
8
7 7
6 6
0 100 200 300 400 500 0 100 200 300 400 500 600
REE (ppm) REE (ppm)

E 1Oct15-9: shale Size Density F 1July15-2: shale Size Density

11 20
10 18
LREE/HREE

LREE/HREE

9 16
8 14
7 12
6 10
0 100 200 300 400 500 0 500 1000 1500
REE (ppm) REE (ppm)

Size, whole coal basis Density, whole coal basis Size, dry ash basis Density, dry ash basis

15
13 G 240: clean coal
LREE/HREE

11
9
7
5
0 500 1000 1500 2000 2500
REE (ppm)

Figure 11. Correlations between LREE/HREE ratios and REE

concentrations of ash, shale and clean coal samples.
 1 Al
0.5

-0.5

-1 Be Na Mg Al Si P K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Y Zr Nb Mo Ag Cd Sn Sb Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Tl Pb Th U

1
Pearson Product-Moment Correlation Coefficients

Fe
0.5

-0.5

-1 Be Na Mg Al Si P K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Y Zr Nb Mo Ag Cd Sn Sb Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Tl Pb Th U

1 Ca
0.5

-0.5

-1 Be Na Mg Al Si P K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Y Zr Nb Mo Ag Cd Sn Sb Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Tl Pb Th U

REE
1

0.5

-0.5

-1 Be Na Mg Al Si P K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Y Zr Nb Mo Ag Cd Sn Sb Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Tl Pb Th U

Figure 12 Pearson product-moment correlation coefficients between elements

Table 1. List of samples and physical separations performed

Physical Separations
Sample ID Description
Size Magnetic Density
240 Clean coal x n/a x
Ash disposal site at power plant. Ponded fly
251 ash. Landfill fly ash. Central Appalachian x x x
Coal.
PC power plant. Ash pond sample containing
339 mostly fly ash with lesser amount of bottom x x x
ash. Coal seams(s) unidentified.
PC power plant. Fly ash. Coal seams(s)
345 x x x
unidentified.
PC power plant. Dry ash pond sample
containing mostly bottom ash with lesser
357 x x x
amount of fly ash. Coal seams(s)
unidentified.
1July15-2 Clay-rich shale sample. x n/a x
1Oct15-9 Clay-rich shale sample. x n/a x
Table 2 Characteristics of coal, coal ash and shale samples
Sample ID 240 251 345 339 357 1July15-2 1Oct15-9
Moisture (%) 1.36 0.36 0.28 0.22 0.04 1.43 0.79
Dry Ash (%) 7.80 84.95 97.15 95.85 98.34 92.60 93.09
Be 59 16 142 9 BDL BDL BDL
Na 3485 2364 4003 3785 2817 1122 884
Mg 3535 5422 4509 4723 4777 2503 4206
Al 203790 112112 157321 113596 147470 157595 102878
Si 284430 274087 285408 270463 228665 245488 320062
P 1218 BDL BDL 1057 BDL 611 BDL
K 13944 17428 18540 16586 12786 22075 20295
Ca 15107 9902 36613 21767 22966 1199 16328
Sc 123 83 108 86 89 74 83
Ti 13304 8325 7943 7043 5662 10368 5863
V 424 246 328 385 258 182 108
Cr 212 156 204 192 138 150 107
Mn 172 154 399 384 420 19 135
Fe 29421 31559 158558 144108 186456 14404 31034
Co 65 54 51 47 36 5 24
Ni 131 101 172 146 121 35 70
Cu 221 177 92 97 74 BDL 86
Zn 146 86 211 298 145 17 167
Ga 167 112 67 67 46 62 45
Ge 45 18 63 47 25 3 BDL
As 37 60 106 59 12 22 51
Se 12 BDL BDL BDL BDL BDL 10
Y 336 132 95 66 70 44 62
Zr 1175 272 302 262 210 324 188
Nb 75 29 29 25 20 29 19
Mo 15 4 48 43 13 BDL 4
Ag 5 2 2 2 BDL 2 BDL
Cd 29 7 6 8 6 8 4
Sn 15 8 14 5 9 25 7
Sb 16 6 10 6 3 1 3
Ba 1202 1400 418 551 534 486 511
La 319 115 87 57 69 93 47
Ce 662 246 184 124 142 178 112
Pr 79 29 22 14 18 20 14
Nd 300 113 85 58 72 71 59
Sm 62 24 18 11 15 13 14
Eu 7 4 4 3 3 3 3
Gd 62 25 20 12 15 11 15
Tb 10 4 3 2 2 2 2
Dy 63 26 17 13 14 9 14
Ho 12 5 3 2 3 2 3
Er 36 14 10 7 8 5 7
Tm 5 2 1 1 1 1 1
Yb 34 13 8 7 7 5 5
Lu 5 2 1 1 1 1 1
Hf 29 10 8 7 6 9 7
Tl 2 2 BDL 4 BDL 1 BDL
Pb 137 92 34 114 19 43 26
Th 105 42 23 19 20 24 23
U 43 17 13 27 18 6 7
REE 1658 623 463 312 371 414 296
REY 1994 754 558 378 441 457 358
REE/Th 15.8 14.8 19.8 16.4 18.3 17.5 12.9
LREE/HREE 9.0 8.5 9.5 8.7 9.3 15.9 8.1
Most Critical (%) 36.0 37.0 36.6 37.3 36.9 28.0 39.2
BDL-below detection limit. Unit: ppm.
Table 3 XRD analysis results. Approximate magnitudes of phases estimated.
Amorphous
Sample ID Description Quartz Hematite Mullite Magnetite
/glass
251 Ash Major Major Trace Intermediate
339 Ash Major Major Intermediate Intermediate Intermediate
345 Ash Major Major Intermediate Intermediate Intermediate
357 Ash Major Major Intermediate Intermediate Intermediate

Muscovite Carbon/
Sample ID Description Quartz Kaolinite Pyrite Gypsum Calcite
/Illite amorphous
240 Coal Major Minor Trace Major
1 July 15-2 Shale Major Intermediate Intermediate
1 Oct 15-9 Shale Major Intermediate Intermediate Minor Minor Trace
Table 4 Chemical formula and density of typical minerals found in coal and coal by-products
[26]

Mineral Chemical Formula Density (g/ml)

Albite Na1.0-0.9Ca0.0-0.1Al1.0-1.1Si3.0-2.9O8 2.6-2.65
Bastnasite (Ce,La)(CO3)F, (La, Ce)(CO3)F, Nd(CO3)F, Y(CO3)F 3.9-5.2
Biotite K(Mg,Fe2+)3(Al,Fe3+)Si3O10(OH,F)2 2.7-3.3
Calcite CaCO3 2.71
Gibbsite Al(OH)3 2.40
Gypsum CaSO4 • 2H2O 2.31
Hematite α–Fe2O3 5.26
Kaolinite Al2Si2O5(OH)4 2.61-2.68
Magnetite Fe2+Fe3+2O4 5.18
Monazite (Ce,La,Nd,Th)PO4 4.98-5.43
Mullite Al6Si2O13 3.11-3.26
Muscovite KAl2(Si3Al)O10(OH,F)2 2.77-2.88
Orthoclase KAlSi3O8 2.55-2.63
Pyrite FeS2 5.01
Quartz SiO2 2.59-2.63
Spinel MgAl2O4 3.6-4.1
Tridymite SiO2 2.25-2.28
Xenotime YPO4 4.4-5.1
Table 5 Maximum REE enrichment factors and recoveries for physical separations of coal and
coal by-products

Size Magnetic Density

Sample ID
EFmax R (%) EFmax R (%) EFmax R (%)
240 1.12 8.24 n/a n/a 1.21 40.68
251 1.05 36.85 1.01 96.19 2.14 3.49
339 1.19 20.15 1.16 40.79 1.18 24.70
345 1.04 13.14 1.13 49.79 1.27 37.15
357 1.05 20.61 1.15 55.18 1.25 30.29
1July15-2 1.18 5.53 n/a n/a 3.58 0.88
1Oct15-9 1.05 34.49 n/a n/a 1.71 13.74