Coal recovery from coal fines cleaning wastes by agglomeration withvegetable oils: effects of oil type and concentration

Marta I. Alonso, Adolfo F. Valde´s, Rosa M. Martı´nez-Tarazona, Ana B. Garcia*

Instituto Nacional del Carbo´n, CSIC, Apdo 73, 33080-Oviedo, Spain

Received 28 September 1998; received in revised form 30 November 1998; accepted 1 December 1998

Abstract

The aim of this work was to obtain high calorific value products from coal fines cleaning wastes by agglomeration with vegetable oils.These residues are mainly being disposed of in dumps, causing important economic and environmental problems. Three Spanish coal fineswastes from different coal cleaning plants were agglomerated with crude and refined sunflower and soybean oils over a wide range of oilconcentrations. The response of these fines wastes to agglomeration with the oils, was evaluated by the percentages of coal matter recovery,ash rejection and efficiency index. Speaking in terms of products quality, the best results were attained at the lowest oil concentrations,especially when the refined ones were used. In these cases, the agglomeration with vegetable oils allowed the recovery from coal fines wastesof a ready to burn fine coal fuel.q 1999 Elsevier Science Ltd. All rights reserved.

Keywords:Coal fines wastes; Oil agglomeration; Cleaning; Vegetable oils

1. Introduction

The mining and cleaning of coal generates large amountsof solid wastes, the latter being up to 90% of the totalproduction. The Spanish coal industry produces about sixmillion tons of solid wastes annually [1]. Although coalwastes are used in mining, civil engineering, construction,agriculture and fluidized combustion [2–4], most must bedisposed of in dumps and slurry ponds, causing majoreconomic and environmental problems. Among coal clean-ing wastes, the refuse from the coal fines cleaning circuit,which represent a sizable fraction, may have considerablecoal content depending on the method used for cleaning aswell as on the size and distribution of the mineral matter incoal. As an example, calorific values in the range 1800–2100 kcal/kg were determined for the coal cleaning fineswastes studied in this work. Owing to their higher coalcontent and small size which increases the surface arealiable to be wetted and oxidized, the disposal of coal finescleaning wastes in coal dumps aggravates the risk of spon-taneous combustion leading in turn to the emission ofnoxious gases. In addition, a lot of energy is wasted.These adverse effects can be overcome by recovering coal

from coal fines cleaning wastes before disposal. Moreover,coal recovery from waste reduces the need of coal to bemined and the cost of handling and storage of coal wastes.A future-orientated study realized in 35 coal cleaning plantsof Illinois has clearly demonstrated the economical feasi-bility and environmental benefits of coal recovery from coalfines cleaning wastes [5].

The aim of this work is to obtain high calorific valueproducts from coal fines cleaning wastes by agglomerationwith vegetable oils. This objective complies with theEuropean Union policy to use some vegetable oils forpurposes other than food. In the present European agricul-tural situation, vegetable oils are candidates for alternativeuses such as biodiesel production [6–8]. Oil agglomerationwas found to be one of the most effective methods for clean-ing coal at very fine particle size [9,10]. By contrast, vege-table oils were previously used in this laboratory as coalagglomerants with results comparable or even better [11]than those obtained employing light hydrocarbons, whichwere recognized to provide cleaner coal products [12–15].However, the price of the hydrocarbons make their commer-cial use uncertain. Vegetable oils are also light compounds(density , 0.9 g cm23), less expensive than hydrocarbonsand even more important, unlike mineral oils, they arerenewable and nonpolluting energy resources.

In the work described here, the first of a series to bepublished, three Spanish coal fines wastes from different

Fuel 78 (1999) 753–759

0016-2361/99/$ - see front matterq 1999 Elsevier Science Ltd. All rights reserved.PII: S0016-2361(98)00218-X

* Corresponding author. Tel.:1 349-98-528-0800; fax:1 349-98-529-7662.

E-mail address:anabgs@muniellos.incar.csic.es (A.B. Garcia)

coal cleaning plants were agglomerated with crude andrefined sunflower oils and crude and refined soybean oilsover a wide range of oil concentrations. The response ofthese fines wastes to agglomeration with these oils wasevaluated by the percentages of coal matter recovery andash rejection. The influence of oil type and concentration onthe results was investigated. These results were comparedwith those previously obtained in the recovery of coal fromfines wastes by agglomeration with mineral oils such asdiesel oil [16–19].

2. Experimental

2.1. Materials

Coal fines wastes from Villablino, Camocha and Modestapreparation plants in north-west Spain were selected for thisstudy. The original samples were designated as TV(0-500),TC(0-500) and TM(0-500), and were obtained from therefuse of the coal fines cleaning circuit before they wereaccumulated in a slurry pond, TC, or in dumps, TV andTM. The production of these coal fines wastes amounted

to 175 000 tons in 1996. Samples denoted TV(20-500),TC(20-500) and TM(20-500) were obtained from theoriginal ones by screening out the# 20mm fraction. Theparticle size distributions of the three original fines wasteswere determined by wet screening and the results arereported in Table 1. The proximate and elemental analysesas well as the calorific value of the original and ‘‘deslimed’’samples are given in Table 2.

Crude (GIC) and refined (GIR) sunflower oils and crude(SOC) and refined (SOR) soybean oils were used as agglom-erants. The characteristics of these oils, which consistmostly of glycerides (95%–98% for crude oils and.98% for refined oils) with a low concentration of free fattyacids (2%–5% for crude oils and, 0.1% for refined oils)appear in Table 3. The oil viscosities were determinedaccording to ASTM method D445-86. The interfacialtension of the oil against water was measured by the ringmethod using a KSV tensiometer, model Sigma 70.

2.2. Agglomeration procedure

Oil agglomeration experiments were conducted in acommercial 7-speed Waring blender equipped with a

M.I. Alonso et al. / Fuel 78 (1999) 753–759754

Table 1Characteristics of size fractions of TV(0-500), TC(0-500) and TM(0-500) coal fines cleaning wastes

TV(0-500) TC(0-500) TM(0-500)

Size (mm) Weight (% dry) Ash (wt.% dry) Weight (% dry) Ash (wt.% dry) Weight (% dry) Ash (wt.% dry)

. 500 3.36 51.01 3.41 79.20 4.60 55.30355–500 2.83 53.09 3.85 80.97 4.55 65.96250–355 3.83 60.36 4.14 80.23 5.15 71.07180–250 4.69 68.42 4.37 77.94 4.94 72.25125–180 6.49 73.95 5.18 67.83 4.85 63.86100–125 4.41 73.76 4.01 51.44 4.73 52.8363–125 7.95 75.21 7.29 47.64 7.52 45.2445–63 4.88 73.33 5.53 54.64 6.16 49.1020–45 11.04 72.37 13.49 62.00 9.85 58.790–20 50.47 73.56 48.76 79.63 47.65 79.01

Table 2Proximate and elemental analyses, sulfur forms and calorific value of samples

Sample TV(0-500) TV(20-500) TC(0-500) TC(20-500) TM(0-500) TM(20-500)

Proximate analysis (wt.% db)Moisture 0.91 0.87 1.55 0.90 1.33 1.41Ash 71.67 69.26 71.84 64.20 68.45 57.95Elemental analysis (wt.% daf)Carbon 20.12 22.51 18.49 26.35 22.02 32.49Hydrogen 1.55 1.55 2.10 2.72 2.24 3.10Nitrogen 0.52 0.50 0.48 0.60 0.57 0.68Sulfura 0.64 0.63 0.65 0.96 0.81 1.06Sulfur forms (wt.% db)Pyritic 0.61 0.59 0.61 0.79 0.78 0.99Sulfate 0.02 0.04 0.02 0.02 0.03 0.04Organic 0.01 0.00 0.02 0.15 0.00 0.03Calorific value (kcal/kg) 1886 1944 1832 2479 2168 2939

a Dry basis.

1000 ml glass vessel. For each agglomeration test, 400 ml ofdistilled water and 16 g of sample were placed in the blenderand mixed at 11 000 rpm for 5 min to disperse the particles.A specific amount of oil, ranging from 5 to 50 wt.% ofsample, was then added and mixing was continued at thesame speed for 60 s to produce agglomerates. The resultantagglomeration product was separated from the refuse byfroth flotation using a flotation machine with a cell capacityof one liter. The products of two simultaneous agglomera-tion experiments were poured into the cell and mixed at2700 rpm for 30 s, followed by air introduction and collec-tion of the agglomerates during 3 min at the same impellerspeed. To evaluate the coal matter recovery and ash rejec-tion without the interference of oil, the agglomerates werefiltered, washed with ethanol, perchlorothylene and ethylether to extract the oil, water-washed, dried overnight at508C, and analyzed for moisture and ash. In an industrialscale, the solvents washing step of the agglomerates wouldnot be necessary and the coal/vegetable oil agglomeratescould be directly used as fuel. Agglomeration experimentswere carried out with ‘‘deslimed’’ wastes, TV(20-500),TC(20-500) and TM(20-500).

The results of the agglomeration process were evaluatedby (1) percentage of organic matter recovery (OMR) or coalmatter recovery, and (2) the percentage of ash rejection.These percentages were calculated as follows:

% OMR� 100(wt. agglomerate/wt. waste)[(1002ashagglomerate)/

(1002ashwaste)]

% AR� 100[ashwaste2(ashagglomerate × wt. agglomerate/

wt. waste)]/ashwaste

3. Results and discussion

3.1. Coal recovery and ash rejection from wastes

Figs. 1 and 2 illustrate the effects of vegetable oil type andconcentration on coal matter recovery and ash rejectionfrom the wastes. As can be seen, coal recoveries$ 80%from the wastes were obtained at even the lowest oilconcentration tested. In fact, the maximum coal recoveryfrom TV(20-500) was achieved at a SOC concentration of

5 wt.%, this value being 90%. Oil concentration does nothave a significant effect on the recovery of coal from thesecoal wastes. Thus, as the SOR concentration increased from5% to 20%, the coal recovery for TM(20-500) wastedecreased from 77% to 73%, for TV(20-500) wasteincreased from 80% to 83%, whereas the coal recovery

M.I. Alonso et al. / Fuel 78 (1999) 753–759 755

Table 3Physical properties of the oils

Oil SOR SOC GIR GIC

Density (g cm23) 0.92 0.92 0.92 0.92Viscosity (mPa s) 70.1 63.0 68.1 68.7Oil–waterinterfacial tension(mN m21)

26.5 14.4 21.2 9.1

Fig. 1. Effects of oil type and concentration on coal recovery from coalfines cleaning wastes.

from TC(20-500) remained unvaried (80%). However, theappearance of the agglomeration product changes remark-ably and it lost consistency with increasing oil concentra-tion. In fact, at oil concentrations $ 30 wt.%, theagglomerate appeared as an oily paste. Similar variationsof the aspect of the agglomeration product with vegetable

oil concentration have been observed previously [11]. Themaximum ash rejection percentages from the wastes wereachieved at SOR, SOC, GIR and GIC concentrations of5 wt.%. For example, by using SOR oil, these values were67%, 74% and 43% for TM(20-500), TC(20-500) andTV(20-500), respectively. That is, products with only37 wt.% of ash content were obtained from TM(20-500)and TC(20-500) wastes by agglomeration with 5 wt.% ofrefined soybean oil. Unlike coal recovery, ash rejectionfrom the wastes tends to decrease with increasing oilconcentration (Fig. 2). At low oil-to-waste ratios, theamount of oil is insufficient to adsorb on to all the coalparticles present and there is competition for oil wettingamong the different types. As a result, only the highesthydrophobic particles, that contain relatively low levels ofmineral matter, will be agglomerated; hence, higher ashrejection are achieved. If the amount of oil is raised, thenumber of oil droplets available for collision with wasteparticles to form the agglomerates increases, as does the

M.I. Alonso et al. / Fuel 78 (1999) 753–759756

Fig. 2. Effects of oil type and concentration on ash rejection from coal finescleaning wastes.

Fig. 3. Effects of low concentrations of refined sunflower GIR oil on coalrecovery and ash rejection from coal fines cleaning wastes.

agglomeration of coal particles with much higher propor-tions of mineral matter. The fall of ash rejection withincreasing oil concentration is highly variable and for agiven oil, it depends on waste.

According to the mathematical model developed byPetela [20] to calculate the final size, the agglomeratediameter increases continuously as the amount of oil israised, the maximum size, and therefore the highestcombustible recovery [21] being achieved at the so-calledcritical oil concentration. After this concentration wasreached, the product diameter and consequently the recov-ery are expected to remain constant, until the amount of oilin aqueous suspension becomes excessive. From this point,the coal particles exist separately in an oily environment.Based on this model, under the experimental conditionsused in this work, concentrations of SOR, SOC, GIR andGIC vegetable oils up to 50 wt.% may still be considered notexcessive. However, the aforementioned experimentalobservation, that at high vegetable oil concentration theagglomeration products appear as an oily paste, togetherwith the fall of ash rejection with increasing oil concentra-tion, without any corresponding increment of coal recovery,as should be expected [12], suggest that (i) oilconcentrations. 5 wt.% are excessive in the agglomera-tion of these coal fines wastes, and (ii) the excessive amountof agglomerant acts as a collector during the further separa-tion of the agglomerates by flotation; waste particles with avery high proportion of mineral matter which were notpreviously agglomerated are therefore recovered togetherwith the agglomerates leading to lower ash rejection,whereas the coal recovery remains almost unaffected. Toconfirm these points, the three coal wastes studied wereagglomerated with GIR at oil concentrations, 5 wt.%.As can be seen in Fig. 3, by increasing GIR concentrationfrom 0.5 to 5.0 wt.%, the coal recoveries from TM(20-500)and TV(20-500) wastes were increased from 77% to 85%and from 65% to 84%, respectively; whereas the corre-sponding ash rejection percentages decreased, the maxi-mum value 68%–69% being achieved at the lowest GIRconcentration employed. The response of TC(20-500) toagglomeration with GIR at low concentrations was slightlydifferent. Thus, coal recovery from this waste increased to amaximum, 84%, as the amount of GIR was raised to 4 wt.%.However, a further increase in GIR concentration up to5 wt.% led to a coal recovery of 80%. Unlike coal recovery,the effect of GIR low concentrations on ash rejection fromTC(20-500) coal waste is comparable to that of TM(20-500)and TV(20-500) coal wastes (Fig. 3).

Little information is available in the literature on therecovery of coal fines wastes by oil agglomeration [16–19] or any other cleaning process [5]. In addition, owingto the different characteristics of the coal fines wastes,experimental conditions and type of oils used in the agglom-erations, comparison between the results of this work andthose previously reported by other authors may be mislead-ing. Even so, it is interesting to mention that a decrease in

ash rejection with increasing oil concentration, attributed tothe agglomeration of the less hydrophobic feed particles,was also reported by Shrauti and Arnold [16]. In addition,values of OMR of < 40% (calculated according to theexpression of OMR given in Section 2) from fines wastesby agglomeration with 5 wt.% of diesel oil were attained,the ash content being reduced from 54% in the waste to aminimum of 12% in the agglomeration product.

Regarding oil concentration of 5 wt.%, a comparisonbetween the results obtained with refined and crude vege-table oils in Figs. 1 and 2 reveals that ash rejections fromwastes were improved by using the refined ones; this effectwas especially remarkable with soybean oil. As expected,this ash rejection increase led to a decrease in coal recovery.For comparison, ash rejection from TC(20-500) was 54% inthe agglomeration with SOC, against 74% attained whenusing SOR, the values of OMR being 87% and 80%, respec-tively. Speaking in terms of products quality, the agglom-eration with SOR implies the recovery from TC waste of acoal with 37% of ash content; this amount increases up to49% with crude oil SOC. It seems that crude vegetable oilswith higher concentrations of free-fatty acids are able to wetwaste particles with higher mineral matter content in turnleading to higher recoveries. In this context, evidence wasfound that compounds with polar functional groups mayadsorb on the mineral matter portion of coal particles,rendering them hydrophobic and promoting their recoveryby oil [22].

As compared with refined sunflower oil, GIR and refer-ring to an nonexcessive oil concentration of 5 wt.%, theagglomeration of wastes with soybean refined oil SORresults in larger ash rejection, especially from TM(20-500)waste. In this case, the ash content of the agglomerationproduct ranged from 43% with GIR to 37% with SOR.Provided that the chemical nature, density and viscosity ofrefined soybean and sunflower oils are similar, the latter,with little lower oil–water interfacial tension which is corre-lated with the reversible work needed to disperse oil intowater, will be emulsified more easily into fine droplets(Table 3). However, the attachment of these oil droplets towaste particles to form the agglomerates depends on mutualsurface chemistry. The results obtained suggests that GIRoil droplets may adsorb on to waste particles with highermineral matter content because of the larger polarity of thisoil, as shown by its lower oil–water interfacial tension.

3.2. Efficiency of agglomeration with vegetable oils

The goal of any process used for recovering coal wastes isto achieve maximum ash rejection with virtually quantita-tive coal matter recovery. However, the need to grind thecoal waste to extreme fineness to totally liberate mineralmatter from coal, producing particles approaching ‘‘purecoal’’ and ‘‘pure mineral matter’’, as well as the efficiencyof the process used at that particle size make this objectiveunattainable. In practice, as was clearly seen in Fig. 3,

M.I. Alonso et al. / Fuel 78 (1999) 753–759 757

greater coal matter recovery is obtained at the expense ofash rejection. Therefore, both percentages of OMR and ashrejection must be taken into account to evaluate the overallefficiency of the agglomeration with vegetable oils in reco-vering coal from coal cleaning fines wastes and to comparethe results from different coal wastes. An efficiency index(EI) was calculated as EI� OMR 1 AR 2 100, whereOMR and AR are the percentages of OMR and ash rejectionrespectively, as defined in Section 2. The EI values corre-sponding to the agglomeration of the three coal wastes withvegetable oils at an oil concentration of 5 wt.% appear inFig. 4. As expected, refined oils led to higher EI values. Thehighest EI of 54 was achieved in the agglomeration ofTC(20-500) with SOR oil. The overall efficiency of theagglomeration with vegetable oils to recover coal from thecoal fines cleaning wastes studied followed the sequenceTC(20-500). TM(20-500)@ TV(20-500). As an example,the EI values obtained in the agglomerations of these coalwastes with 5 wt.% of SOR were 54, 44 and 24, respec-tively. Microscopic examination of TC(20-500), TM(20-500) and TV(20-500) coal wastes showed that the amountof the number of particles with coal matter-mineral matterassociations were 60.1%, 63.2% and 68.4%, respectively.From this data, it can be concluded that the efficiency of theagglomeration with vegetable oils to recover coal from coalwastes decreases rapidly with increasing number of thattype of particles.

To explore a possible linear correlation between EI andthe number percentage of particles with coal matter-mineralmatter associations, linear regression analysis was appliedto the data obtained in the agglomeration experimentscarried out with 5 wt.% of vegetable oils (Fig. 4). Thistype of analysis provided equations with correlation coeffi-cients of 0.998, 0.989, 0.998 and 0.972, respectively, forSOR, GIR, SOC and GIC vegetable oils. Based on thesevalues, a linear relationship appeared to fit the data quitewell. However, the results are only tentative as only threepoints which correspond to the three coal wastes studiedwere available for regression.

4. Conclusions

Important amounts of energy that are currently wastedcan be recovered from TC(20-500), TM(20-500) andTV(20-500) by agglomeration with a low concentration ofsoybean and sunflower oils, especially when the refinedones are employed. For example, by using 5 wt.% of refinedsunflower oil GIR, a product of 39.5 wt.% of ash and4647 kcal/kg as calorific value could be obtained fromTC(20-500) coal fine waste of 64.2 wt.% of ash contentand a calorific value of 2479 kcal/kg. The product qualitycould be improved by decreasing oil concentration below5 wt.%; in this case, recovered coals with 29, 4 wt.% of ashand calorific value of 5735 kcal/kg were attained. Theseresults clearly demonstrate that the use of agglomerationwith vegetable oils allows the recovery from coal fineswastes of a ready to burn coal fine fuel. Moreover, theweight of the coal fines waste to be disposed of in dumpsor slurry ponds was reduced by at least 40% or much higherdepending on product quality.

Acknowledgements

This research was carried out with financial support fromDGICYT (Research Project AMB95-0182) and FICYT(Research Project PB-MAS96-01). The authors thank Dr.I. Suarez-Ruiz for her microscopic study of the samples.M.I. Alonso and A.F. Valde´s were supported by postgradu-ate grants provided by FICYT and CSIC/MSP (MineroSiderurgica de Ponferrada), and MEC, respectively, duringtheir work at Instituto Nacional del Carbo´n. Thanks are alsodue to the Instituto de la Grasa, CSIC, Sevilla (Spain) forproviding the vegetable oils.

References

[1] Mayo JM. Ministerio de Industria, Direccio´n General de Minas,personal communication, 1994.

[2] Leonhard J, Schieder Th. Aubereitungs-Technik 1990;31:89.[3] Leonhard J. Aubereitungs-Technik 1992;33:73.[4] Fenandez Valcarce JA, Gonza´lez Canibano J. Utilizacio´n de los este´r-

iles del carbo´n, Jornada Te´cnica, E.T.S.I.M., Oviedo, Spain, 1991.[5] Bhagwat SB, Khan LA, Curran LM, Baxter JW, Rice RJ. National

Symposium on Mining, University of Kentucky, Lexington, 1990, pp.253–258.

[6] Ozaktas T, Cigizoglu KB, Karaosmanoglu F. Energy Sources1997;19:173.

[7] Staat F, Vallet E. Chemistry and Industry 1994; p. 863.[8] Ahn E, Koncar M, Mittelbach R, Marr R. Separation Science and

Technology 1995;30:2021.[9] Couch GR. Advanced coal cleaning technology, IEACR/44, IEA Coal

research, London, 1991.[10] Capes CE, Jonasson KA. Interfacial phenomena in coal technology.

In: Botsaris GD, Glazman YM, editors. Surfactants science, 32. NewYork: Dekker, 1989 pp. 115–155.

[11] Garcia AB, Martı´nez-Tarazona MR, Vega JMG. Fuel 1996;75:885.[12] Garcia AB, Vega JMG, Martı´nez-Tarazona MR. Fuel 1995;74:1692.[13] Keller DW, Burry WM. Coal Preparation 1990;8:1.

M.I. Alonso et al. / Fuel 78 (1999) 753–759758

Fig. 4. Effect of oil type on efficiency of the agglomeration in recoveringcoal from coal fines cleaning wastes. Oil concentration 5 wt.%.

[14] Venkatadri R, Markuszewski R, Wheelock TD. Energy Fuels1988;2:145.

[15] Kim SS, Morsi BI, Chiang S-H. Coal Preparation 1994;15:51.[16] Shrauti SM, Arnold DW. Fuel 1995;74:459.[17] Kasperski KL, Briker Y, Desphande DP, O¨ zum B. Energy Sources

1996;18:43.[18] Blaschake Z. Proceedings of 64th CIC Coal Symposium. Al Tanseel

AM, editor, C.S.Ch.E., Otawa, 1982, pp. 229–234.

[19] Capes CE, Hazlett JD, Ignasiak BL. Proceedings of 11th InternationalCoal Preparation Congress, Mining and Materials Processing Instituteof Japan, Tokyo, 1990, pp. 355–360.

[20] Petela R. Fuel 1991;70:509.[21] Darcovich K, Capes CE, Talbot FDF. Energy Fuels 1989;3:64.[22] Labuschagne BCJ. Coal Preparation 1986;3:1.

M.I. Alonso et al. / Fuel 78 (1999) 753–759 759