Bioleaching of high grade Pb–Zn ore by mesophilic and moderately thermophilic iron and sulphur oxidizers

Hydrometallurgy 97 (2009) 1–7

Contents lists available at ScienceDirect

Hydrometallurgy

j ourna l homepage: www.e lsev ie r.com/ locate /hydromet

Bioleaching of high grade Pb–Zn ore by mesophilic and moderately thermophilic ironand sulphur oxidizers

Moazur Rehman, Munir Ahmad Anwar, Mazhar Iqbal, Kalsoom Akhtar,Ahmad Mukhtar Khalid, Muhammad Afzal Ghauri ⁎Bioprocess Technology Division, National Institute for Biotechnology and Genetic Engineering, P. O. Box 577, Jhang Road, Faisalabad, Pakistan

⁎ Corresponding author. Tel.: +92 41 2550814; fax: +E-mail address: [email protected] (M.A. Ghauri).

0304-386X/$ – see front matter © 2008 Elsevier B.V. Adoi:10.1016/j.hydromet.2008.12.008

a b s t r a c t

Article history:

Keywords:Acidithiobacillus ferrooxidansBioleachingSulfolobus thermosulfidooxidansLeadZincSulphide ore

as to test the feasibility of using iron and sulphur oxidizing bacteria for the acidleaching of a high grade Pb–Zn ore material. Three strains (ATCC 13661, NCIMB 13537 and C2-TF) of Acidithio-bacillus ferrooxidans and two strains (MT-TH1 andMT-13) of Sulfobacillus thermosulfidooxidanswere tested in thisstudy. Thebioleachingwasmonitoredbymeasuring thedissolvedmetals andbyX-raydiffractionanalysis of leachresidues. The bioleaching efficiency varied between 0.014 and 0.35. The maximum dissolution of lead wasachieved with the mesophilic At. ferrooxidans (NCIMB 13537) at 30 °C. The maximum recovery of zinc wasachieved with moderately thermopilic S. thermosulfidooxidans (MT-TH1) at 45 °C.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

In nature, lead sulphides such as galena (PbS) are often associatedwith zinc sulphides such as sphalerite (ZnS). The separation of zinc andlead is traditionally achieved with selective flotation, but the process isdifficult to control when used for complex sulphide ores that containpyrite or copper sulphides. Bioleaching may be utilized as an alternativeprocess for theextractionofmetals fromcomplexores/concentrates thatare otherwise difficult to be treated by conventional extractionprocesses(Gilbertson, 2000;Miller et al.,1999). Bioleachingof sulphideminerals isbased on the capability of acidophilic bacteria to oxidize ferrous ironand/or reduced sulphur compounds. Biological oxidation of zincsulphides (sphalerite and wurtzite) and galena has been investigatedin many studies using samples of low-grade and high-grade sulphideores and concentrates as well as research-grade mineral specimens(Boon et al., 1998; Fowler and Crundwell, 1998; Garcia et al., 1995a,b;Groudev, 1980; Khalid and Ralph, 1977; Torma and Subramanian, 1974).

Pakistanhashuge reservesof zincassociatedwith leadand ironpyrite(Aslam, 1995). Although metals from these reserves may be retrievedthrough conventional processes (i.e., flotation and smelting), complica-tions are likely to arise due to the complexmineralogical nature of theseores. The purpose of the present study was to evaluate a bioleachingprocess, as an alternative to recovermetals froma zinc–lead sulphide ore

92 41 2651472.

ll rights reserved.

as an alternative to flotation and smelting. Three acidophilic mesophilicand two moderately thermophilic strains of Fe- and S-oxidizers isolatedfrom geographically different locations were chosen for this study.

2. Materials and Methods

2.1. Zn–Pb ore

A high-grade Zn–Pb ore sample was obtained from the Duddardeposit at District Lasbela in Balochistan province of Pakistan. Dry

Fig. 1. X-ray diffractogram of the untreated high-grade Zn–Pb ore. G, galena; S,sphalerite; W, wurtzite. The vertical bar indicates the relative peak intensity.

Fig. 2. Changes in pH during the bioleaching of the high grade Pb–Zn ore. A. Mesophilicconditions: ·, control; ●, ATCC-13661; ▲, NCIMB-13537; ■, C2-TF. B. Moderatelythermophilic conditions: ·, Control; ●, MT-13; ▲, MT-TH1.

Fig. 4. Distribution of free-swimming and attached bacteria in ore suspensions duringthe bioleaching of the high grade Pb–Zn ore.●, attached MT-13; ▲, attached MT-H1; ■,free-swimming MT-13; +, free-swimming MT-TH1.

2 M. Rehman et al. / Hydrometallurgy 97 (2009) 1–7

classification of particles was carried out. The ore sample was groundand separated to various particle size fractions using ASTM sieves.The −100 to +270 mesh size fraction (150–53 µm) was used in thebioleaching experiments. This size fraction contained 10% Zn, 15% Pband 7% Fe (S content was not determined). The main sulphideminerals were galena and sphalerite (cubic) (Fig. 1). The sample mayhave contained some wurtzite (hexagonal) and small amounts ofpyrite and anglesite (Pb-sulfate), but these could not be properlyresolved by X-ray diffraction (XRD) because of their low levels.

Fig. 3. Changes in redox potential during the bioleaching of the high grade Pb–Zn ore. ·,control; ●, MT-13; ▲, MT-TH1; ■, ATCC-13661; ♦, NCIMB-13537; □, C2-TF.

2.2. Bacteria and culture conditions

The test cultures were three strains of Acidithiobacillus ferroox-idans, ATCC-13661, NCIMB-13537 and C2-TF (a local isolate) and twolocal isolates of moderate thermophiles Sulfobacillus thermosulfidoox-idans (MT-TH1 and MT-13). Sulfobacillus thermosulfidooxidans MT-13has been characterized on the basis 16S rDNA (Ghauri et al., 2003)while MT-TH1 resembles with Sulfobacillus thermosulfidooxidans inmany aspects like pH, nutritional requirements, temperature etc(Ghauri and Johnson, 1991). These bacteria were grown in a mineralsalts medium that contained (per liter) 0.5 g MgSO4.7H2O, 0.15 g(NH4)2SO4, 0.05 g KCl, 0.05 g KH2PO4, and 0.01 g Ca(NO3)2, pH 1.8. Themedium for the moderate thermophiles was supplemented with 2 gyeast extract or tryptic soy broth per liter. The cultures were grown in2000 ml shake flasks (80 rev/min), each containing 750 ml mediumat 30 °C for mesophiles and 45 °C for moderate thermophiles. Growthwasmonitored by the oxidation of ferrous iron to ferric iron and visualobservations. Cells were harvested by centrifugation at 15000 ×g for20 min at 4 °C.

2.3. Bioleaching experiments

Bioleaching experiments were carried out in duplicate cultures in500 ml shake flasks containing 200 ml ore suspension using three

Fig. 5. Changes in bacterial counts during the bioleaching of the high grade Pb–Zn ore.●, ATCC-13661; ▲, NCIMB-13537; ■, C2-TF; ▼, MT-13; ♦, MT-TH1.

Fig. 6. Solubilization of iron from Pb–Zn ore by mesophilic chemolithotrophic bacteria.■, Fe2+; ▲, Fe3+; ●, total dissolved Fe.

Fig. 7. Solubilization of iron from Pb–Zn ore by moderately thermophilic chemolitho-trophic bacteria. ■, Fe2+; ▲, Fe3+; ●, total dissolved Fe.

3M. Rehman et al. / Hydrometallurgy 97 (2009) 1–7

strains of Acidithiobacillus ferrooxidans (ATCC 13661, NCIMB 13537and C2-TF)and two strains of Sulfobacillus thermosulfidooxidans (MT-TH1 and MT-13), separately, to select the most efficient strains forbioleaching of zinc/lead ore. 10 g (5% pulp density) of zinc/lead orewas added having mesh size ranged from 100–270 to each flask. Forthe first 24 h, the initial acid demand was satisfied with sulphuric acidat pH 2 and the flasks were then sterilized and inoculated(6×107 cells/ml) with the test bacteria. Uninoculated media wereused as chemical controls. The flasks, inoculated with Acidithiobacillusferrooxidans, (ATCC-13661, NCIMB-13537 and C2-TF) were put onshaker at 30 °C at 120 rpm while flasks inoculated with Sulfobacillusthermosulfidooxidans (MT-13 and MT-TH1) were subjected to shakingat 45 °C on 120 rpm speed. At intervals, samples (6 ml) were takenaseptically for cell counts and chemical analysis. Total cell counts wereestimated with a counting chamber under phase-contrastmicroscope.

For chemical analyses, samples were centrifuged and the supernatantwas used for pH and redox potential measurements. Dissolved oxygen(DO) was measured using a DO meter (ORION model 386). Solubleiron species Fe2+, Fe3+, and total Fe were determined by o-phenanthroline method (Hiroyoshi et al., 1999). Dissolved Zn and Pbwere analyzed by atomic absorption spectrometry. Pb in solids wasdissolved by suspending the samples in 10% ammonium acetatesolution.

Solids were retrieved by filtering leach solution samples throughWhatman No. 42 cellulose filter paper. The residues were washed withdistilled water and dried at 110 °C, followed by grinding in a pestle andmortar for XRD analysis. XRD patterns were determined with a Rint2000 Rigaku Series X-ray diffractometer, courtesy of Central Hi-Technology Laboratory, University of Agriculture, Faisalabad, Pakistan.XRD analyses of top fill powder mounts were conducted using CuKαradiation and a wide-range goniometer equipped with diffracted-beammonochromator and a compensating slit. Step scans were conductedfrom 20 to 70° 2 in 0.5° 2 increments using a 4-s count time.

3. Results

3.1. pH, redox potential, and dissolved oxygen

After the initial acid demand, the pH values were adjusted to pH2.0 with H2SO4. In mesophilic cultures, the pH continued to increasefor seven days before the net reaction became acid producing. The pHchanges had a similar profile with the moderate thermophiles (Fig. 2).

Table 1Bioleaching of lead–zinc ore bymesophilic andmoderately thermophilic chemolithotrophs.

Strain Pb solubilization Zn solubilization Pb:Zn

mg/l(μmol) % Solubilized mg/l(μmol) % Solubilized

Mesophiles30 °C control 460 (2.22) 30 120 (1.9) 2.4 1.168ATCC13661 980 (4.73) 65 590 (9.32) 12 0.508NCIMB 13537 1100 (5.31) 73 425 (6.71) 8.5 0.791C2-TF 830 (4.0) 55 310 (4.89) 6.2 0.818

Moderate thermophiles45 °C control 370 (1.79) 24 154 (2.43) 3.1 0.737MT13 1050 (5.07) 70 2190 (34.6) 44 0.146MT-TH1 1060 (5.12) 71 2370 (37.44) 47 0.136

4 M. Rehman et al. / Hydrometallurgy 97 (2009) 1–7

After an initial decline, the redox potential values increased from~300 mV to N600 mV in the test cultures, indicating progressiveoxidation (Fig. 3). The redox potential showed very little change in theparallel control flasks.

Fig. 8. X-ray diffractograms of leach residues after 70 days of contact in mesophilic culturesD, C2-TF. A, anglesite; G, galena; J, jarosite; S, sphalerite; W, wurtzite. The vertical bar indic

The DO concentrations fluctuated between 4.0 and 6.0 mg O2/l forthe first five weeks, followed by an increase to 7–8 mg O2/l. Thedissolved oxygen level was higher in the control flasks than in themesophilic and moderately thermophilic cultures.

3.2. Cell counts

In order to determine the extent of rate of mineral dissolution interms of cell attachment, the number of free-swimming and attached,moderately thermophilic bacteria were monitored as a function oftime in the Zn–Pb ore suspensions (Fig. 4). The number of free-swimming bacteria decreased with time and pseudo-equilibriumconditions were attained within 40 h. Cell attachment was compar-able with MT-TH1 and MT-13. With mesophiles, cell attachmentoccurred within 10 minutes (data not shown).

The number of bacteria in the ore suspensions decreased from6×107 cells/ml to b106 cells/ml for the first 2–3 weeks of incubation.Subsequently, cells counts increased at least two orders of magnitudebefore tailing off. The ATCC 13661 strain produced the fastest increase

and the corresponding chemical control. A, control; B, ATCC-13661; C, NCIMB-13537;ates the relative peak intensity.

Fig. 9. X-ray diffractograms of leach residues after 70 days of contact in moderately thermophilic cultures and the corresponding chemical control. A, control; B, MT-13; C, MT-TH1.A, anglesite; G, galena; J, jarosite; S, sphalerite; W, wurtzite. The vertical bar indicates the relative peak intensity.

5M. Rehman et al. / Hydrometallurgy 97 (2009) 1–7

and the highest count of 2×108 cells/ml. For moderate thermophiles,the counts were below b105 cells/ml for the first sevenweeks of incu-bation. Subsequently, the counts increased to about 1.2×107 cells/ml(Fig. 5).

3.3. Iron solubilization

The Fe3+/Fe2+ ratio increased progressively during the incubationand the total iron reached a level of 830–990 mg/l (Fig. 6), which was8–9 times higher than in the control flasks. With the moderate ther-mophiles, strain MT-TH1 had a shorter lag period as compared withMT-13 preceding the active iron oxidation phase (Fig. 7). The finallevels of total iron were comparable to those in the mesophiliccultures.

3.4. Lead solubilization and recovery

Because lead precipitates in the presence of sulfate, dissolved leadwas not analyzed in leach solutions in these experiments. At thetermination of the experiments, solid residues were retrieved andwashed with 10% ammonium acetate, which converted PbSO4 towater soluble lead acetate. PbS is not dissolved in this protocol. Leadacetate thus dissolved accounted for 55–73% of the initial Pb contentin the mesophilic cultures and about 70% in the moderatelythermophilic cultures (Table 1). Mineralogical changes in leachresidues were also evaluated by XRD (Figs. 8 and 9). A decrease inthe peak intensity of galena was observed in the leached samples.

3.5. Zinc solubilization

With the mesophilic cultures, zinc solubilization ensued after athree-week lag phase (Fig. 9). The most efficient strain was At.ferrooxidans ATCC-13661 with maximum zinc solubilization of590 mg/l (12% yield). Dissolution of zinc in the NCIMB 13537 andC2-TF cultures was 310–425 mg/l and the control flask contained120 mg Zn/l (2.4%) (Fig. 10). The rate of zinc leaching was found to be7.7 mg Zn/l day for the ATCC-13661 culture (Table 1).

With theMT-13 andMT-TH1 cultures, the yield was 2190–2370mgZn/l (44–47% yield) (Table 1). Zinc solubilization in the correspondingcontrol amounted to 3.1% yield. These two isolates had superiorleaching rates i.e., 40–48 mg Zn/l day — when compared to themesophiles (Table 1).

3.6. Bioleaching efficiency coefficient

Bioleaching efficiency is a key parameter to check the most effi-cient bioleaching system. It is a combination of rate and percentrecovery comparable to optimum conditions. The zinc dissolutiondata were used to calculate the bioleaching efficiency coefficient EB inan effort to develop a quantitative comparison of the results obtainedunder different environmental parameters. The coefficient EB isdimensionless and defined as a function of metal recovery and kineticefficiency:

EB = Rm � Ek

Table 2Bioleaching efficiencies of mesophilic and moderately thermophilic chemolithotrophs.

Strain Bioleaching efficiency (EB)

Mesophiles30 °C control 0.00037ATCC13661 0.014NCIMB 13537 0.0078C2-TF 0.0035

Moderate thermophiles45 °C control 0.00098MT13 0.272MT-TH1 0.35

6 M. Rehman et al. / Hydrometallurgy 97 (2009) 1–7

where Rm=metal recovery (%) and Ek=kinetic efficiency (%), definedas the percent ratio of the rate of metal dissolution obtained/themaximum rate attainable under optimal conditions. The numericalvalues of EB, Rm, and Ek vary between 0 and 1. The coefficient EB for themesophilic conditions ranged between 0.0035 and 0.014 and for themoderately thermophilic conditions 0.35–0.27 (Table 2). Thus thecoefficients indicated that themoderate thermophiles were at least 20times more effective in leaching zinc from the ore material.

4. Discussion

The present study was aimed at elucidating the bioleaching of ahigh-grade Pb–Zn ore by mesophilic and moderate thermophilicchemolithotrophic bacteria. Some important environmental factorssuch as pH, dissolved oxygen and redox potential were also monitoredduring bioleaching of the ore. Unlike in previous Zn-bioleachingstudies (Deveci et al., 2004; Rodriguez et al., 2003) conducted atdifferent initial pH values, sulphuric acid was repeatedly added to thebioleaching medium until the pH stabilized at 2.0, followed by pHmonitoring periodically during the time course. In spite of initial acidconsumption, no zinc was leached from the ore during the first sevendays in the inoculated and sterile flasks, suggesting that carbonate-containing gangue was reacting during this initial phase with theadded acid. After the initial acid consumptionwas neutralized, the pHbecame stable and the concentration of dissolved zinc increased in theinoculated flasks, suggesting ensuing bacterial activity. The pH of themedium decreased because of acid producing reactions. The keyleaching reactions can be presented with the following equations,which involve direct bacterial attack or the chemical action of ferriciron generated by bacteria (Fowler and Crundwell, 1999; Sand et al.,2001).

ZnS þ Fe2ðSO4Þ3→ZnSO4 þ 2FeSO4 þ S0

PbS þ Fe2ðSO4Þ3→PbSO4 þ 2FeSO4 þ S0

The notion whether the mineral dissolution is due to the directattack by microbial cells or it takes place through some lixiviant, suchas ferric iron in this case, has been discussed quite a lot in previousyears. There exists a third option according to which both these modeof actions i.e., direct and indirect attack may operate concomitantlywhich in our view is highly likely for mineral dissolution.

Adsorption of bacterial cells, at mineral surface, gives informationabout the extent of mineral dissolution in solution. Rapid cellattachment to the Pb–Zn ore exhibited the similar pattern observed

Fig. 10. Zinc solubilization during the bioleaching of the high grade Pb–Zn ore. ·,control; ●, MT-13; ▲, TH-1; ▼, ATCC-13661; ♦, NCIMB-13537; ■, C2-TF.

by other investigators (Fowler and Crundwell, 1999; Rodriguez et al.,2003). However, in contrast to the results obtained by Rodriguez et al.(2003), the cell attachment rate for mesophilic bacteria was higherthan that for moderate thermophiles probably due to the difference inthe nature of the ore used in these experiments. The reason for thelower cell counts in the moderately thermophilic cultures is not clear,but they may reflect more tenuous cell attachment to ore particles,thereby rendering them undetachable for microscopic cell counts.Moreover, on the contrary at elevated temperatures the hydrophobicinteractions are stronger than the hydrophilic interactions renderingcells to remain as free swimmers.

Lead has a limited solubility in leach solution because it formsanglesite (PbSO4), a poorly soluble complex with sulfate (Forward andPeters, 1985). Under the bioleaching conditions, the anglesiteaccumulates in the residual solids (Gomez et al., 1995). Consequently,the peak intensities of galenawere diminished and the peak intensitiesof anglesite increased in solid residues, indicating the conversion ofPbS to PbSO4. The oxidative nature of these reactions is furthersupported by the increase in the redox potential, suggesting thatbacteria maintained a high Fe3+/Fe2+ ratio in the medium, and thedecrease in the dissolved oxygen concentration in inoculated media.Consistent with other reports (Deveci et al., 2004; Dew et al., 1999;Konishi et al., 1998;Witne and Phillips, 2001), the findings reported inthis paper indicate that moderately thermophilic bacteria werecomparatively faster than the mesophiles in the bioleaching process.

Conventional methods of treating lead/zn ore are very cumber-some, however, biohydormetallugical technologies present a veryfeasible and practical option of treating such ores where lead isconverted into insoluble lead sulfate and zinc is removed in solubleform. Previous studies have not attempted to work out ratios betweenlead and zincmetal taking into account the bioleachability of the targetore and stoichiometric perspectives are also lacking, which warrantfuture studies.

References

Aslam, K.M., 1995. Mineral Profile of the Muslim World. OIC Standing Committee onScientific and Technological Cooperation (COMSTECH), Islamabad.

Boon, M., Snijder, M., Hansford, G.S., Heijnen, J.J., 1998. The oxidation kinetics of zincsulphide with Thiobacillus ferrooxidans. Hydrometallurgy 48, 171–186.

Deveci, H., Akcil, A., Alp, I., 2004. Bioleaching of complex zinc sulphides using mesophilicand thermophilic bacteria: comparative importance of pH and iron. Hydrometallurgy73, 293–303.

Dew, D.W., van Buuren, C., McEwan, K., Bowker, C., 1999. Bioleaching of base metalsulphide concentrates: a comparison of mesophilic and thermophilic bacterialcultures. In: Amils, R., Ballester, A. (Eds.), Biohydrometallurgy and the Environmenttoward the Mining of the 21st Century: Part A. Elsevier, Amsterdam, pp. 229–238.

Forward, F.A., and Peters, E., 1985. Leaching principles. In: Weiss, N.L. (Ed.), SMEMineralProcessing Handbook, vol. 2. SME, New York, pp. 13, 6–12.

Fowler, T.A., Crundwell, F.K., 1998. Leaching of zinc sulphide by Thiobacillus ferrooxidans:experiments with controlled redox potential indicate no direct bacterial mechan-ism. Appl. Environ. Microbiol. 64, 3570–3575.

Fowler, T.A., Crundwell, F.K., 1999. Leaching of zinc sulphide by Thiobacillus ferrooxidans:bacterial oxidation of sulphur product layer increases the rate of zinc sulphidedissolution at high concentrations of ferrous iron. Appl. Environ. Microbiol. 65,5285–5292.

Garcia Jr., O., Bigham, J.M., Tuovinen, O.H., 1995a. Oxidation of galena by Thiobacillusferrooxidans and Thiobacillus thiooxidans. Can. J. Microbiol. 41, 508–514.

7M. Rehman et al. / Hydrometallurgy 97 (2009) 1–7

Garcia Jr., O., Bigham, J.M., Tuovinen, O.H., 1995b. Oxidation of sphalerite by Thiobacillusferrooxidans and Thiobacillus thiooxidans. Can. J. Microbiol. 41, 578–584.

Ghauri, M.A., Johnson, D.B., 1991. Physiological diversity amongst some moderatelythermophilic iron-oxidising bacteria. Microb. Ecol. 85, 327–334.

Ghauri, M.A., Khalid, A.M., Grant, S., Heaphy, S., Grant, W.D., 2003. Phylogenetic analysisof different isolates of Sulfobacillus spp. isolated from uranium-rich environmentsand recovery of genes using integron-specific primers. Extremophiles 7, 341–345.

Gilbertson, B., 2000. Creating value through innovation: biotechnology in mining. IMMTrans., C 109, 61–67.

Gomez, C., Roman, E., Blazquez, M.L., Ballester, A., Gonzalez, F., 1995. SEM and AESstudies of a lead sulphide bioleaching in presence of catalytic ions. Miner. Eng. 8,1503–1512.

Groudev, S.N., 1980. Leaching of sphalerite with different strains of Thiobacillusferroxidans. C.R. Acad. Bulg. Sci. 33, 11–19.

Hiroyoshi, N., Hirota, H., Hirajima, T., Tsunekawa, M., 1999. Inhibitory effect of ironoxidizing bacteria on ferrous-promoted chalcopyrite leaching. Biotechnol. Bioeng.64, 478–483.

Khalid, A.M., Ralph, B.J., 1977. The leaching behaviour of various zinc sulphide mineralswith three Thiobacillus species. In: Schwartz, W. (Ed.), Conference BacterialLeaching. Verlag Chemie, Weinheim, pp. 165–173.

Konishi, Y., Nishibura, H., Asai, S., 1998. Bioleaching of sphalerite by the acidophilicthermophile Acidianus brierleyi. Hydrometallurgy 47, 339–352.

Miller, P.C., Rhodes, M.K., Winby, R., Pinches, A., van Staden, P.J., 1999. Commercialisa-tion of bioleaching for metal extraction. Miner. Metall. Process. 16 (4), 42–50.

Rodriguez, Y., Ballester, A., Blazquez, M.L., Gonzalez, F., Munoz, J.A., 2003. Newinformation on the sphalerite bioleaching mechanism at low and high temperature.Hydrometallurgy 71, 57–66.

Sand,W., Gehrke, T., Jozsa, P.G., Shippers, A., 2001. (Bio)chemistry of bacterial leaching—direct vs. indirect bioleaching. Hydrometallurgy 59, 159–175.

Torma, A.E., Subramanian, K.N., 1974. Selective bacterial leaching of lead sulphideconcentrate. Int. J. Miner. Process. 1, 125–135.

Witne, J.Y., Phillips, C.V., 2001. Bioleaching of Ok Tedi copper concentrate in oxygen- andcarbon dioxide-enriched air. Miner. Eng. 14, 25–48.