a review of gold particle size and recovery methods wc-97-014

7/27/2019 A Review of Gold Particle Size and Recovery Methods WC-97-014

1/34


2/34

British Geological SurveyTECHNICAL REPORT WC/97/14Overseas Geology Series

A REVIEW OF GOLD-PARTICLE-SIZEAND RECOVERY METHODSCJ Mitchell, EJ Evans & M T StylesTh is docum ent is an output of a project funded by the UK Overseas DevelopmentAdm inistration (OD A) for the benefit of developing countries. The views ex pressedin this documen t are not necessarily those of the ODA.

ODA Classi$cationSubsector: GeoscienceTheme: G2, Identify and ameliorate minerals-related and other geochemical toxic hazardsProject Title: Mitigation of mining-related mercury pollution hazardsReference number: R6226Bibliographic reference:C J Mitchell, E J Evans & M T Styles,A review of gold particle-size and recovery methodsBG S Technical Report WC/97/14Keywords:Gold, mineral processing, M alaysia, Zimbabwe, New Zealand, Papua New Guinea, CanadaFront cover illustration:Processing of gold ore using a sluice box, Malaysia0 ERC 1997

Keyw orth, Nottingham, British Geological Survey , 1997


3/34

Mitigation of mining-related mercury pollution hazards

BRITISH GEOLOGICAL SURVEYMineralogy & Petrology Group

Technical Report WC/97/14

A REVIEW OF GOLD PARTICLE-SIZE AND RECOVERY METHODSCJ Mitchell, EJ Evans & MT Styles

1. INTRODUCTION

This report reviews published literature concerning the particle-size distribution ofgold occurring in alluvial and bed-rock deposits and the various methods used for itsrecovery. The aim of this review is to recommend m ethods for the recovery of fine-grained gold (generally less than 100pm in size) as an alternative to theenvironmentally damaging use of mercury amalgamation. This work has been c arriedout as part of an ODA / BGS T echnology Development Research (TDR ) projectR6226 "Mitigation of mining-related mercury pollution hazards".2. PARTICLE-SIZEOF GOLDInformation on the particle-size distribution of gold is sparsely scattered throughoutthe literature and presented in various different forms. Many gold assays are carriedout on sieved heavy m ineral concentrates (HM C's). This is of limited use, to thecurrent review , as coarse grained gold present in the ore will not be includ ed in theanalysis. This is simply due to the practicalities involved, often samples were pre-sieved on site to remove coarse material (for example >2 mm) and only the finematerial was evaluated. Often laboratory evaluations will be geared toward the use ofspecific gold recovery techniques. Many of these techniques have their effectivenesslimited to specific size ranges. Therefore the evaluation will often only be applicableto gold w ithin that size range, regardless of the presence of finer (or coarser) gold.Even where a com plete gold particle-size distribution is quoted the informationshould be treated w ith caution. Gold generally occurs in m inute am ounts whichrequires (for statistical accuracy) large bulk samples to be collected for size analysis,especially to determine weight percentages of coarse gold grains (>500pm). Variousphysical m ethods are used in order to separate gold for analysis, most comm onlylaboratory-scale gravity separators.

British Geological S urvey , February 1997 1


4/34

Mitigation of mining-related m ercury pollution hazards

How ever for particles below approximately 75 pm the processing efficiency of mostgravity sep arators starts to fall. Therefore it is impo rtant that the gold conten t of al lseparation produ cts, middlings and tailings (waste products) as well as concentrates,are determined to ensure that an accurate gold particle-size distribution is producedfor the whole samp le not just for certain size ranges.The size at which gold is considered to be "fine-grained" varies. The followingclassifications have been devised (W ang & Poling, 1983):

Russia Canada (British C olumbia)Fine gold -315 +100 pm Fine gold -1.5 mm +350 pmMinute gold -100 +37 pm Flour -350 pmDispersible gold -37 pmOthers suggest that 75 pm or 100 pm (Stewart & Ramsay, 1993) should be the topsize of fine-grained gold. This report will use 100 pm.Th e particle-size distributions quoted for gold d o not reflect the particle-sizedistribution of the ore. For example 50%of the gold by w eight may exist in a sizefraction that only represents 10%of the ore by weight. Also the variation in th eweight d istribution of gold with particle-size depends upon the nature of the ore ( an dfor alluvial gold the source area). Some alluvial gold deposits may contain asignificant amount of fine grained gold whereas others will contain virtually none.2.1. Alluvial Gold

Generally the particle-size of alluvial gold falls into the range 5 mm to 100 pm .Several exam ples are given in the following Tables. The information given in Tab les1 to 6 is from a B GS study of alluvial gold. The samples were sieved in the field at 2mm and hence will not contain gold grains larger than that, although it is unlikely thatthey are present. Grain size was measured by image analysis of grains hand pickedfor subsequent polishing and m icroanalysis. The minimum size is governed by thatwhich co uld be m anipulated and w as considered large enough for polishing. M anysam ples actually contained a sm all proportion of finer grains that were not measured .

British Geological Survey, February 1997 2


5/34


Sam ples purchased from artisanal alluvial miners and are shown with *. They oftencontained less fine gold than the samp les collected by BGS staff and counterparts.This is due to both poor technique and, in some cases, poor equipment; Africanminers were seen using broken washing-up bowls and old ice cream tubs as pans.This ex plain s the variation in data obtained for samples collected by different peop lein different ways an d makes establishing the true natural variation very difficult.As can be seem from most of the tables little information is available for gold finerthan 100 pm . In part this may be due to the inefficiency of fine gold recovery in thestudies referred to.

Table 1. Ecuador (Styles et a l , 1992)

Sample Size range (pm) Average (pm)

EQ1*EQ2*EQ3*EW*EQ5*EQ6*EQ7*

153 - 2911186 - 2868980 - 329960 8 - 3199544 - 3146555 - 4066123 - 3280

139316072288155214 3118981347

Table 2. Mersing, Johore, Malaysia (Styles et al, 1994)

Site Size range (pm) Average (pm)Kuala Mayang 312 - 89 3Air Merah 208 - 48 1Telok Bangka 110 - 727Pulau Belanak 225 - 658

63134 238 5401



6/34


Table 3. Lubuk Mandi area, Terengganu, Malaysia (Henney et a l , 1994)

Site Size range (pm) Average (pm)

Lubuk MandiMA13* 56 0- 1477MA14 136 - 87 9MA 16* 13 6- 1599MA17* 476 - 3624

93 94429041332

Table 4. Zimbabwe (Styles et a l , 1995)

Site Size range (pm) Aversg:, (pn)

MazoezM5 21 2- 1603ZM89 30 - 550ZM90 121 -4 51Chegutu-Chakari" 171 - 2486KwekweKadoma* 217 - 2067ZM28* 19 6- 1961ZM29* 365 - 2237zM35 217 - 476ZM73* 234 - 2428Silobelam 3 3 * 165 - 1944m 3 4 * 25 6- 1025Zvishavane * 21 0- 1239

464249222126495886510472421070

313450585



7/34


Table 5. Penjom area, near Kuala Lipis, Pahang, Malaysia (Henney et a l , 1995)

Site Size range (pm) Average (pm)MA 105 283 - 719MA 106 197 - 55 1MA 107 241 - 82 2MA108 222 - 74 5MA 109 356 - 827MA 1 10 281 - 804MA1 11 120- 518

43930 638336 452 553 424 5

Table 6. Raub area, Pahang, Malaysia (Henney et al, 1995)~~~ ~~

Site Size range (pm) Average (pm)Luie RiverRaub valleyMA 86MA87MA88MA89MA9020 km S of RaubMA92MA93MA9410km N of RaubMA100MA101MA 102

215 - 51 9149 - 78 7132 - 306208 - 72 6197 - 57 5110 - 844

215 - 394146 - 308132 - 46 1266 - 575405 - 723253 - 390

38 633 220 340435936 9

27 022 527341352830 8



8/34


Table 7. Mount Kare, SW of Porgera, Papua New Guinea (Bartram et al, 1991)

Particle-size Weight Percentage~

+5 mm 0.4-5 mm +800 pm 29.6-800 +250 pm 66.2-250 +75 pm 3.6-75 pm 0.2

Other placer deposits in Papua N ew G uinea have up to 50% fine gold 4 0 0 pm(Subasing he, 199 1).Table 8. Average placer gold, Yukon, Canada (Clarkson, 1994)

~ ~~~

Par ticle-size Weight Percentage

+2.36 mm 10-2.36 + 1.4~ll~ll 20-1.4 mm +600 pm 30-600 +300 pm 30-300 +15 0 pm 10-150 pm Not quoted

Table 9 Placer gold, 'Middle Asia' (Solozhenkin et al , 1993)

Particle-size Sample 1 Sample2W t % Wt %

+1 mm - --1 m m + 4 0 0 p m eo.1 0.22-400 +200 pm


9/34


In New Zealand placer dep osits contain gold up to 95% 200 pm (Braithwaite & Jury, 1993).

2.2. Bed-rock (& miscellaneous) gold

The particle-size distribution of bed-rock gold varies considerably with the type ofgold deposit and is too diverse to generalise. Mu ch of the information gathered hererepresents go ld associated either with sulphide mineralisation or altered volcanicrockdm etasediments. Gold occurs in a w ide range of particle sizes from ultrafineinclusions in pyrite (dow n to


10/34


2.2.5. Granulite, Ren co Reefs gold mine, Zimbabwe (L eroy, 19 95). The gold is finegrained (65% e15 pm) and 28% occurs as free gold associated with native bismuth.

2.2.6. Gold in glacial till overlying Archean greenstone, Ontario, Canada (Sh elp &Nichol, 1987) is up to 90% 4 0 0 pm , mostly e125 pm.2.2.7. Gold in streams draining gneiss & plateau basalts, Harris Creek, BritishColumbia, Canada (Day & Fletcher, 1989) is mainly cl00 pm.

2.3. Summary of gold particle-sizeTh e following table gives a summ ary of the overall range and average of the availableparticle-size data fo r alluvial and bedrock go ld (the figures in brackets represent theparticle-size which most of the data falls into).

Table 11.Summary of alluvial and bed-rock gold particle-size

Gold type Particle-sizeOverall range Average

Alluvial 30 pm to 4 mm 222 to 2.3 mm(100 pm to 3.5 mm) (300 pm to 1.5 rnm)

Bedrock e3 pm to >500 pm e 100 pm

3. GOLD RECOVERY METHODS

This section w ill briefly outline the various mineral processing m ethods co mm onlyused to recover gold, giving an indication of the size-range of m aterial processed andtypical gold recovery figures. Panning is not considered as no 'hard data' was found .Th e methods considered m ainly involve the physical separation of gold from 'gan gue'(which ranges from vein m aterial in bed-rock deposits to sand and silt grade m aterialin alluvial deposits) using gravity-based processing methods. Go ld has a high specificgravity (19.3 g/cm 3) in relation to most comm on gang ue minerals (ranging from 2.65to 3 g/cm3) and is therefore em inently suitable for gravity processing. C onsidering theminute quantity of gold normally present in even the most auriferous ores (dow n to



11/34


lgltonn e in bed-rock and 0.25 g/tonne in alluvial deposits) gravity processing isvirtually the only method effective at producing the concentration ratios required,especially with the high volume throughputs associated with alluvial mining.Table 12. Particle-size range and typical ef'ficiencies of gravity and chemical

gold recovery methods

Method Effective size range Recovery efficiency

Sluice boxes(+ Reichert cones)

2500 to 100 pm As low as 20% for 4 0 0 mgold to 96% for 40 pm

3000 to 75 p m 65 to 80 %

6000 to 30 pm u p t o 9 9%1500 to 70 p m As low as 65% for


12/34

Mitigation of mining-related mercury p ollution hazards

organic layer o r iron oxide coatings and so me are leached free of impu rities (such assilver) leaving a rim of pure gold, all of these render the surface hydrophobic (Wang& Poling, 1983).The mineralogical character of the gold is often not considered when planning aprocessing plant, especially if the gold respond s well to standard gravity andcyanidation processes. How ever, if the gold recovery is poor ( 430%) he ore istermed refractory and a detailed mineralogical investigation becom es nece ssary.This will involve the determination of the mode of occurrence of minute gold grainsand the proportion of invisible gold. Gold usually occurs as native gold. A solidsolution exists w ith many heavy metals including electrum (Au, Ag), argentian gold(Au, Ag), cuprian gold (Au, Cu), palladian gold (Au, Pd), mercurian gold (amalgam )(Au, Hg) and Au-Ag-Hg alloy. Other gold-bearing minerals occur only in very smallamounts including gold tellurides, gold selenides, gold sulphides and intermetalliccom pounds such as ama lgam (Au, Hg), aurostilbite (Au, Sb) and maldonite (Au, Si )(Petruk, 1989).3.1. Sluice boxes

3.1.1. Riffled sluice boxesThe sluice box is by far and away the most com monly used means of concentratinggold from alluvial gravels. They are generally cheap to make, easy to operate andrequire minimal technical knowledge to maintain (Hancock, 1991). Essentially asluice box consists of a sloping open rectangular flume with regularly spacedtransverse bars, or riffles, hrough w hich a dilute slurry of water and alluvial g ravelflows. Heavy minerals, including gold are usually captured in the upstream side of theriffles (or the dow nstream side depending upon the design of sluice box). These areregularly removed by raking or cleaning-out the sluice box riffles.Sluice boxes are effective for the recovery of gold with particle-size from 25 mm to100 pm. The efficiency of gold recovery varies from 80 to 100% in well operatedsluices run by modern com mercial companies to less than 50% in makeshift sluicesrun by small-scale m iners. A typical sm all-scale operation would involve 2 to 3people, with 1 digging and feeding the sluice, the second picking out stones and thethird m onitoring the wash w ater (Hancock, 1991). In New Zealand gold recoveriesare up to 80%, the remainder is mainly finer than 250 p m and is physicallyrecoverable by gravity separation (Fricker, 1986). In Yukon recoveries up to 98% are

British Geological Su rvey, February 1997 10


13/34


reported (C larkson, 1994). The efficiency of gold recovery is influenced by a host ofparameters:i) The design of the sluice box, including the slope, width and length, andthe type of riffles incorporated.

In Papua New Guinea sluice boxes consist of a wooden launder tilted 5 to 15" withtransverse wooden slats (Subasinghe, 1991). In Yukon ex panded metal riffles areincorporated into sluices operating at a relatively shallow slopes (7-12") with feedrates of 20 m3/h r and water flows of 40 l/s/m are used for recovering gold finer than 1mm. Angle iron riffles used in sluices with steeper slopes (12-14"), faster feed rates,40 m 3/hr and faster water flows, 80 l/s/m are used for recovering gold coa rser than 1mm. Flat bar riffles ('nugget trap') may be suitable for recovering go ld coa rser than 6mm (Clarkson, 1994).A porous ma tting is often used to line the floor between riffles, particularly in thelower part of the s luice box, in order to enhance the recovery of gold d uringoperation. Oscillation of the sluice box during operation may improve gold recoveryfrom alluvial gravel with a high proportion of heavy minerals or clay. Howev eroscillation leads to reduced gold recovery from other alluvial gravels (Clarkson,1994).The recovery of gold increases with sluice length and particle-size as indicated here(Fricker, 1986):Table 13. Gold recovery with increasing sluice length and particle-size

Sluicelength -100pm -250+100 pm -1 mm +250 pm

Gold recovery for a given particle-size

wt % w t% wt %0.5 m 4 341.0 m 5 501.5 m 6 582.0 m 7 59

50687880



14/34


ii) The particle-size range of the gold and the accompanying sediments.

Th e recovery of gold present in sand grade as compared to gravel grade material hasbeen determined as follows (Fricker, 19 86):Table13. Recovery of sand- and gravel-grade gold

Gold Gold recoverysize -4mmsand -16 mm gravel

0.1 mm 74 nd0.2 mm 84 650.5mm 92 851.0 mm 97 952.0 mm 100 99

Wt % W t %

Fine gold (e 10 0 pm) is a problem for sluices and up to 60%of the total gold contentreports to the tailings (i.e. is not recovered). Go ld recovery efficiencies of up to 50%are achievable for material finer than 200 pm (Wang & Poling, 1 983) and up to 40 to50% for material finer than 100 pm (Subasinghe, 1991; Stewart & Ramsay, 1993).The 'slimes' content (this can mean m aterial less than 10p m in size to less than 75p m in size) of the feed can adversely affect gold recovery. Scrubbing and screeningprior to sluicing can improve gold recovery (by up to 20%) by freeing gold trapped inclay-boun d and weakly cemented material (Clarkson, 1 994). If no coarse-grainedgold is present the sluice can be operated using conditions app ropriate for therecovery of fine-grained gold. However if coarse-grained gold does exist the ore canbe split into coarse and fine grained material. Gold can then be recovered from thecoarse-grained material by using a second sluice operated using conditionsappropriate for coarse-grained gold recovery. The treatment of coarse and finematerial separately will increase overall gold recovery.iii) The sluice feed and wash water rate

Sluices are often operated in an attempt to recover coarse- (nugget) and fine-grainedgold in a single pass (Subasinghe, 1991). However these require distinctly d ifferenthydrodynam ic conditions. Coarse gold is best separated using high feed rates, withsteep slu ice tilt and high riffles - the coarse gold drops out and finer gold rem ainsBritish Geological Survey, February 1997 12


15/34


entrained in the flow of material down the sluice. Fine gold recovery requires lowerfeed rates, shallower tilts and smaller riffles. Attempting to recover both coa rse- andfine-grained go ld at the sam e time w ill result in poor overall recoveries. A slower ratewill improve recove ry of fine gold but decrease the remov al of gangue and reduceoverall throughput. Th e effect of varying sluice feed and wash wa ter is similar to thatof varying sluice tilt. In many cases sluices are operated with water diverted fro mlocal streams and it is more practical to vary sluice tilt than the wash w ater rate. Thesteeper the tilt the faster the wash rate.iv ) The time interval between cleaning out the riffles

Th e sluice box is essentially a batch, or semi-continuous, process. In order to recovergold from the sluice box riffles the operation is halted. The sluice boxes op erated inPapua New Guinea are operated for approximately a week between clean-out. Theexperience of the operator is important in deciding when to clean. To o long aninterval will allow an excessive am ount of solids to pack up against the riffles. Thiswill sto p gold being trapped and it will be displaced to the tailings. A shorter intervalwill result in improved recov eries but this has to be balanced ag ainst the time spen tcleaning and decreased throughput (Subasinghe, 1991)3.1.2. Pinched sluices (including Reichert cones)A pinched sluice is similar to a riffled sluice box, minus the riffles, that tapers to thedownstream end (F igure 1). During operation the minerals present in the feed slurrysegregate with heavy m inerals adjacent the floor of the sluice and progressivelylighter minerals occurring at the top of the flow. The heavy mineral concentrate isseparated by the use of an inclined plate over which the slurry flows from the en d ofthe sluice.A Reichert cone has a sim ilar separating principle to that of a pinched sluice. Itcon sists of a squat cone, as if the base of a pinched sluice had been stretched out intoa circle. A slot near the base of the cone removes the concentrate, whereas thetailings, which are entrained in the flow, pass straight over.Pinched sluices and Reichert con es are effective in the processing of material in thesize range 3 mm to 30 p m. T ypically used as a primary rougher (preconcentration ofgold prior to gold recovery) or scavenger (reprocessing of waste to improve overallgold recovery) where high capacity, low operating cost is required. For exa mple goldrecoveries of 50 to 60%are achieved with Reichert cones at New C elebration goldBritish Geological Survey, February 1997 13


16/34


Mine in Western Australia (Martins et a l , 1993). Used (to remove coarse go ld) aheadof expensive stages such as froth flotation or chem ical leaching, which work better onfine gold, they can save costs. The recov ery of coarse gold requires ore to be groundto a size suitable for flotation or leach times to be extended. Up to 60% gold can berecovered by Reichert cones this way (Feree, 1993).3.2. Jigs

A jig is described as a 'hindered settling device used to concentrate heavy m ineralsfrom feed material'. It consists of a shallow , flat tray with a perforated bottom platethrough which water is pulsed up and down (Figure 2). This causes the heavyminerals to m igrate downw ards and the lighter minerals to remain at the top of thepulsing liquid. The heavy m inerals are drawn through the base plate and the lightminerals pass over the top as tailings.Jigs are effective in the processing of material in the size range 25 mm to 75 pm.Material is best pre-screened and processed as separate coarse and fine fractions.Typically for gold below 100 pm recovery falls to less than 50%. In a combinedsluicing-jigging operation gold recovery may be as high as 89 to 95 % (Wang &Poling, 1983). In an Alaskan op eration conversion of a sluice plant to a jig plantincreased gold recovery by 37% (Fricker, 1986) and in a New Zealand operation asimilar conversion increased gold recovery by 14% (Braithew aite & Jury, 1993). AKelsey centrifugal jig has been used to concentrate fine gold-bearing sulphides from aprocess tailings stream with recoveries of >70% (Brewis, 1995). The following tablecom pares gold recovery between a jig and a sluice at different grain sizes (Fricker,1986):Table 14.Recovery of gold using a sluice versus a jig

Particle Gold recovery-size Sluice Jig

Wt % Wt %0.1 mm 20 480.2 mm 56 740.5 mm 88 951.0m 96 98



17/34


3.3. Shaking tablesThe comm onest form of shaking table used is the wet table (the 'dry' form is knownas an air table, which uses air as the fluid separating medium ). It consists of a flattable (or 'deck') with parallel riffles to trap the heavy m inerals (Figure 3 ). The 'deck ' isvibrated longitudina lly and inclined laterally during operation. A perforated pipefeeds wash w ater from the upslope side . The slurried feed is introduced at the topupslope corner. The minerals in the feed segregate. The heavy m inerals sink to thedeck , migrate along the riffles and are discharged over the end of the deck . The lightminerals, entrained in the water, pass straight over the riffles and down to the bottoman d so to the tailings.Shaking tables are effective in the processing of xA eria l in the size range 3 mm to 15pm . Shaking tables have been used to recover 88 % of the gold present in aconcentrate produced on a spiral (Eltham, 1984). Wang & Poling (1983) record thatup to 90% of gold coarser than 40 pm can be recovered, whereas typically only 20%of 20 - 40 p m gold can be recovered. The efficiency drops greatly below 40 pm. Thefollowing table g ives the size distribution of gold present in a shaking table middlingproduct from a comm ercial mine in Malaysia i.e. material that had passed over thetable and was stockp iled for possible later reprocessing. Estimated assay of around 10ghon ne is higher than the m ined ore. There is considerable scope for increased yieldwith more effective processing.Table 15. Particle-size distribution of gold in a middlings product from a goldmine in Malaysia

Particle-size (pm) Yield (wt %)+250 pm 35.3-250+120 pm 42.11-120 +63 pm 21.81-63 pm 0.75



18/34


3.4. SpiralsTh e spiral concentrator is described as a 'low feed rate, low feed density' flowing filmgravity separator. It consists of a helical conduit of mo dified semi-circular cross-section, usually with between 3 and 5 complete 'turns' (Wills, 1992) (Figure 4).Ma terial is fed onto the top of the spiral as a slurry with typically 25 to 30% solids byweight. A s the material flows spirally downw ards the particles stratify due to factorssuch a s centrifugal force, differential settling, hindered settling and reverseclassification. Th ere is usually a density gradation ac ross the profile of the spiral withheavy minerals concentrating next to the axis and minerals of lower density beingswept to the outer edge. Concentrate, middling and tailing products are collected withthe use of adjustable splitter plates.Spira ls are effective in the processing of ma terial in the size range 3 mm to 75 p m(although up to 5 % 'slimes' can be tolerated w ith sufficient wash water). Spiralperformance is controlled by : the diame ter and pitch of the spiral; the pulp density(i.e. the so lids content of the slurry); the location of the splitters and take-off points;and, the volume and pressure of wash water. In one operation hydrocyclones are usedto deslime (removing particles nominally finer than 3 0 pm in this case) the feed tospirals (only rejecting gold finer than 14 pm ) and this leads to spiral recoveries of upto 65% (Eltham, 1984). At New Celebration gold mine in Western A ustralia a seriesof rougher (producing gold pre-concentrates) and cleaner (removing impurities fromgold con centrates) spirals consistently achieve gold recoveries of 70 to 80% (Martinset a l , 1993). Spirals are also known to be used for the recovery of fine flat free gold(recoveries up to 85%)and gold finer than 37 p m (recoveries up to 50%) (Feree,1993).3.5. Rotating conesRotating cones find application in arid and semi-arid countries where water is at aprem ium . These concentrators use the least amou nt of water of any wet gravityseparator and produce a high-grade concentrate in a single pass. They consist of asquat cone (included angles of 105" to 115") which is tilted and rotated. A h igh-density pulp is introduced at the top edge of the cone an d wash w ater at the bottomleading edge. A s the cone rotates the heavy minerals migrate to the centre of the coneand are drawn off. The light minerals ovefflow the lower edge of the c one as atailings.



19/34


Rotating co nes are effective in the processing of material in the size range 3 mm to 30pm . Th e concentrate grade is determined by : the slope of the cone; the grade of theore; and , the residence time of the material in the cone during operation.3.6. Bowl concentratorsComm only used m akes are the Knudsen and the Knelson bowl concentrators. A bowlconcentrator consists of a rotating cylinder that segregates heavy m inerals fro m lightminerals by a comb ination of centrifugal force and wash water action. The Knelsonbow l concentrator is claimed to recover "gold particles ranging from 6 mm to lessthan on e micron in a single pass" (sales brochure). Recovery is effective down toapproximately 30 pm.Typically Knelson concentrators have been retrofitted to process flotation tailings torecover gold coarser than 100 pm and also to replace mineral jigs as a means ofrecovering coarse gold ahead of flotation and cyanidation. A fully-automated Knelsonconcentrator was installed at the Dom e M ine, South Porcupine, On tario inreplacement for a jig circuit and exceeded jig gold recovery using o nly a tenth of thevolume of jig feed (Brewis, 1995). A similar concentrator was installed at the GoldenGiant m ine in north central Ontario which, accompanied by a single stage of tabling,accounted for up to 30% of the overall gold recovery. This was free-gold recovereddirectly from the grinding circuit prior to cyanidation (Brewis, 1995).Removal ofcoarser free gold ahead of cyanidation leads to savings from lowered carbon stripping(recovery of gold from solution) and a consequent reduction in the use of cy anideacid and o ther consumables. Also removal of free gold grains ahead of the grindingcircuit will ultimately improve flotation efficiency as there is a reduction in gold'smearing' onto other m inerals and effecting flotation properties.

3.7. Drum concentratorsThe M ozley M ulti-gravity separator (MGS) consists of a tilted drum that tapersslightly to the do wnslope end (Figure 5 ) . The d rum simultaneously rotates and isshaken longitudinally. Sample slurry is added to the upslope end and as it migratesdow n the dru m it segregates into heavy and light minerals. The h eavy minerals reportto the inside wall of the drum, where they are pushed to the up slope end by scrapers.Th e light minerals remain entrained in the wash w ater and report to the dow nslopeend. Th e MG S has been likened to a shaking table wrapped round itself into a drum .



20/34


The M GS can be used to process material in the size range 1 mm down to 1 pm. AnMGS has been used to recover 76% of the gold present in a zinc flotation concentrate(Mozley sales literature).3.7. Other gold recovery methods3.7.1. Non-gravity methodsMagnetic separation - Magnetic separation is the process whereby m inerals areseparated by ex ploiting differences in their mag netic susceptibility. Magneticseparators, using either permanent magnets or electromagnets, can be used to processmaterial in a we t or dry state at varying mag netic intensities (depending on theimpurities to be rem oved). Mag netite often concentrates with gold, this is easilyremoved by magnetic separation which makes subsequent gold recovery moreefficient (Wang & Poling, 1983). Magnetic sepa ration is effective in the processing ofmaterial in the size range 2 mm to 70 pm for dry magnetic separation and dow n to 5p m for wet magnetic separation.Electrostatic separation - Electrostatic separation is the process whereby mineralsare separated by exploiting differences in their electrical conductivity. Electrostaticseparators use elec trodes to create an electrical charge on the surface of minerals. Therate at which this charge dissipates controls the separation, conductive minerals losethe cha rge rapidly and non-conductive minerals slowly. If non-conductive heavyminerals occur with gold electrostatic separation can be used, leading to improvedgold recoveries (Wa ng & Poling, 1983 ). Electrostatic separation is effective in theprocessing of material in the size range 600 pm to 70 pm.Hydrocycloning - A hydrocyclone is a classification device used extensively in theminerals industry for the beneficiation of a wide variety of material and also for thedewatering of process products, for example ball mill discharge and flotation tailings.A hydrocyclone consists of a downward pointing cone through which a slurry is fedunder pressure to produce a whirlpool effect. A separation of coarse from finematerial is effected by the resulting centrifugal forces within the cone . A relativelycoarse-grained product is discharged from the apex, known as the underflow and arelatively fine-grained product is discharged from the base, known as the overflow.Gold occurring in the underflow product may possibly exist as grains of free gold andmay therefore be amenable to recovery by gravity processing. Hydrocyclones areeffective in the processing of material in the size range 3 mm to 40 pm.



21/34

Mitigation of mining-related m ercury pollution hazards

Compo und water cyclones (a form of hydrocyclone composed of m ulti-angle cones)have been tested on fine feed material. Gold recovery of 63% was achieved with asample containing 70%


22/34


are recovered. A recovery of up to 98% is quoted for one operation in Papua N ewGuinea (Eltham, 1984).Mercury is occasionally poured between the riffles on a sluice box in an attempt tocapture fine-grained gold. Howev er the contact time between the mercury a nd thegold is not sufficient to allow amalgam ation to occur. Often fine gold remainssuspended in the flow of material above the riffles and does not com e into contactwith the mercury. Up to 30% of the mercury used in sluices in Papua New Guineafinds its way d irectly into local rivers. Passing the tailings ov er ama lgamation unitsor through mercury filled columns has been recommended as a method of recoveringthis fine gold. Ho wev er these are ultimately un satisfactory as they still pose a threatto the environment (Subasinghe & Maru, 1994).Cyanidation - Cyanidation is the process whereby gold is recovered using a cyanidesolution. Gold is dissolved using the cyanide solution and the resulting com plex,Au(CN),, c an be removed from solution by various methods (Deschenes, 1986):i) The M erill-Crowe process, is used to remove the gold from the cyanide bycementation w ith powdered zinc.ii) Activated carbon absorption (otherwise known as C-I-P, carbon-in-pulp) is usedfor the processing of ores with a high slimes content which are difficult to treat by theMerrill-Crowe process. The absorption of g d d is either performed by:

- Carbon-in-columnfrom solutions typically from heap leaching which are virtuallyfree of suspended material,- Carbon-in-pulp (CIP) from leach pulps typically slimes, ground ores and c alcines.An alternative to C IP is RIP (Resin-in-pulp) which is easier to use a nd less sensitiveto the influence of naturally occurring carbon.,- Carbon-in-leach (CIL ) whilst leaching is still in progress. Typically w ith orescontaining carbonaceous material that could rob the gold from the pregnant(gold-bearing) solution.The carbon is reactivated by heating to 600 to 900C in a reducing a tmosphere.Electrowinning (deposition of gold by electrolysis) and zinc cementation is used torecover the gold from the eluate.

British Geological Surv ey, February 1997 20


23/34


Typical gold recovery efficiencies for cyanidation range from 95 to 99% (Marsden &Fuerstenau, 1993). Gold recove ry is reported to be 80% from refractory gold oreresidues from France (ground to 80% finer than 31 pm ) ( Lucion & Cuyper, 1993).Gold recovery is reported to be approximately 82.5% from the processing of uraniummill tailings at C luff Lake, Saskatchewan. It was found that gold loss to tailings wasassociated with natural carbon present in the uranium tailings (Melis & Rowson,1989). Cyanidation trials were carried out on m aterial from old tailing ponds. Arecovery of 90.5% was ach ieved using m aterial 60%


24/34


3.9. Typical large-scale gold processing flowsheets

3.9.1. Alluvial gold processingExtraction - Typically by hydraulic sluicing or dredging.Feed preparation - Drum scrubbing & trommel screening. Coarse m aterial ( A 0 mm)to tailings. Fines dewatered by hydrocyclone. Fine screening (e.g. 500 pm) optional.Primary processing - Riffled sluice or jig concentration. Coarse tailings remo ved.Secondary processing - Shaking table or bowl concentrator. Fine tailings removed.Fine processing - Spiral and shaking table concentration to process fine m aterial.Gold production - Amalgamation of concentrates with mercury to extract gold.3.9.2. Hard rock gold processingExtraction - Open cast bench m ining or deep mining.Feedpreparation - Primary jaw or gyratory crusher (e.g. to e1 50 m m) followed bysecondary milling (e.g. ball-, rod- or semi-autogenous milling). Size classification byscreening or hyd rocycloning .Coarse processing ( ~ 5 0 0m ) - Jig and shaking table or bow l concentration.Fine processing ( 4 0 0p ) Reichert cone concentration followed by spiral a n d o rshaking table concentration. Froth flotation of very fine material (e.g. 4 8 0 pm).Gold production - Coarse concentrate by am algamation. Fine concentrate by millingand C IP cyanidation.4. RECOMMENDATIONS FOR FINE GOLD RECOVERY

The review of separation techniques show s that several methods are effective for therecovery of gold coarser than approximately 100 pm but for gold finer thanapproximately 50 pm only chemical m ethods are very efficient. The efficientrecovery of gold coarser than 50 pm should be possible using simple gravity methodsand so me recomm endations for possible implementation are given below. If there isevidence that there is significant gold finer than 50 pm a cyanide treatment of the finetailings is necessary for recovery.These recom mendations apply to gold recovery by sluice box. Figure 6summarisesthe main recommendations made for their use in sm all-scale gold processing.

British Geological S urvey, February 1997 22


25/34


4.1. Improvements to current practiceOptimise clean-out

Th e time interval between clean-ou t of sluice box riffles is dependent upon thenature of the material processed and the operating conditions applied. It isrecommended that the time interval should be short enough to en hance therecovery of fine gold (that would otherwise be lost due t o solids packedaround the riffles). However this should be tempered by the requ irement tomaintain a relatively high throughput, which if reduced by too much cleaningwill ultimately reduce gold production.

Appropriate configurationIt is recommended that the tilt of the sluice box should increase withincreasing particle-size. Typically (in the Yu kon, Canada) for material finerthan 1 mm 7 to 12" is used and coarser than 1 mm 12 to 14".

Appropriate operating practiceAppropriate feed and wash water rates are dependent upon the nature of thematerial being p rocessed, especially the particle-size (and also the clay and /orheavy m ineral contents). It is recommend ed that the feed and w ash water ratesare high eno ugh to enable the efficient separation of coarse grained goldwithout excessive loss of fine grained gold.

4.2. Process modificationsWashing prior to sluicing

It is recom mended that gold ore with a significant proportion of clay-boundand weakly-cemented m aterial be washed (' scrubbed') and screened prior tosluicing. This will enable the liberation of gold trapped in clay.

Processing coarse and fine separatelyOperating conditions appropriate for the recovery of coarse-grained gold aredifferent to those for fine-grained gold. Typically material containing coarseand fine grained gold w ill be processed on the same sluice. This willinvariably lead to a loss of gold. Ideally coarse gold should be recovered on asteep sluice using high feed and w ash rates, whereas fine gold should berecovered on a shallower sluice using lower feed and wash rates. It isrecommended that ore is screened prior to sluicing and the resulting coarseand fine streams be diverted down different sluices. This will improve overallgold recovery.



26/34


Expanded metal and angle iron sluice rimesThe use of alternative riffles will enable a higher recovery of gold. Expandedmetal riffles are recommended for gold finer than 1mm and angle iron rifflesfor gold coarser than 1 mm.

4.3. Efficient alternativesReplace sluices with jigs

Jigs are efficient in the recovery of gold down to 75 pm and have been used toreplace sluices, with a resultant increase in gold recovery.

Introduce shaking tables and bowl concentrators to concentrate fine goldSha king tables and bowl concentrators are efficient in the recovery of golddown to at least 40 pm and are used to process fine gold-bearing ore. It isrecommended that these are considered for the recovery of fine go ld that isnot recovered by sluice box.

5. CONCLUSIONS

5.1. Alluvial go ld has a particle-size range typically between 5 mm and 100pm(how ever there is little information ab out the amount of gold finer than 100 pm).Hard rock gold has a wide range of particle-size depending upon the nature of the ore,and in some cases down to less than 5 pm.5.2. Gold is usually recovered by a combination of physical and chem ical processingmethods. Com monly riffled sluice boxes and jigs are used to pre-concentrate h eavyminerals (including go ld) from gold-bearing ore. Shaking tables, spirals, rotatingcones and bo wl concentrators are used to upgrade this to a heavy m ineral concentrate.Gold is comm only extracted from fine-grained concentrates by Carbon -in-pulpcyanidation and from coarse-grained concentrates by m ercury amalgamation.5.3. Th e efficiency of gold recovery depends upon the processing method and theproficiency of operation. A well operated modern sluice box can potentially recoverup to 98% of gold coarser than 100 pm . However gold finer than 100 pm will be lostunless a fine processing m ethod (e.g. shaking table or bow l concen trator) is used toreprocess the fine tailings.



27/34


5.4. If small-scale operators can be m ade to work at efficiencies close to that achievedby large com panies, and in the laboratory, this will be as good as, and often supe rior,to mercury amalgamation.5.5. Recom mendations for the recovery of fine gold include improving currentoperating practice, modifying the process used, and introducing more efficientalternatives.6. REFERENCESBartram, JA, Heape, JMT & Caithness, SJ (1991) The M ount K are gold project.

Proceedings of PN G geology, exploration and mining conference, Rabaul,Australasian Institute of M ining and M etallurgy, p.96- 100.

Braithewaite, JC & Jury, AP (1993) Alluvial gold mining and treatment in NewZealand. in Woodcock, JT & Ham ilton, JK (eds.) Australasian Mining &Metallurgy, Monograph 19, Volume 2, Australasian Institute of Mining &Metallurgy, p. 1074-1079.

Brewis, T (1995) Gravity separation. Mining M agazine, May 1995, p.279 - 292.Clarkson, R (1994) The use of nuclear tracers to evaluate the gold recovery efficiency

of sluiceboxes. CZM Bulletin, 87. No. 979, p.29-37.Cristovici, MA (1986) recovery of gold from old tailings ponds. CZM Bulletin, 79,

NO. 895, p. 27 - 33.Davies, RH (1992) Peak gold mine - A different approach to underground m ining.

Proceedings of AusIMM annual conference, 1 7-21 May 1992, Broken H illNSW, Australasian Institute of Mining & Metallurgy, p. 127-135.

Day, S J and Fletcher, W K (1989) Effects of valley and local channel morphology onthe distribution of gold in stream sediments from Harris Creek, B ritishColumbia, Canada. Journal of geochemical exploration, 32, p. 1 16.

Deschenes, G (1986) Literature survey on the recovery of gold from thioureasolutions and the com parison with cyanidation. CZM Bulletin, 79, No. 895, p.76 - 83 .



28/34


Eltham, JA (1984) Mining and processing of a low grade gold ore body in the W auValley, Papua New G uinea. Proceedings of AusIMM c onference, Darwin,Austr alasian Institute of Mining & Metallurgy, p. 427-434.

Erceg, MM , Craighead, GA , Halfpenny, R & Lewis, PJ (1991) The explorationhistory, geology and metallurgy of a high sulphidation epithermal gold depositat Wafi river, Papua New Guinea. Proceedings of PNG geology, explorationand m ining conference, Rabaul, Australasian Institute of Mining andMetallurgy, p.58-64.

Ferree, TJ (1993) Application of MD L Reichert cone and spiral concentrators for theseparation of heavy minerals. CIA4 Bulletin, 86, No. 975, p. 35-39.

Fricker, AG (1986) The processing of placer gold. in Kear, D (ed.) Mineral depositsof

New Zealand, Monograph 13,Australasian Institute of Mining andMetallurgy, p.155-158.

Hancock, G (199 1) An econom ic appraisal of small to medium scale alluvial goldmining systems in Papua New Guinea. Proceedings of PNG geology,exploration and mining conference, Rabaul, Australasian Institute of Miningand Metallurgy, p.96- 100.

Henney, PJ, S tyles, MT , Bland, DJ & Wetton, PD (1994) Characterisation of goldfrom the Lubuk M andi area, Terengganu, Malaysia. British Geo logical SurveyTechnical Report WC/94/21,28p.

Henney, PJ, Styles, MT , Wetton, PD & Bland, DJ (1995) Characterisation of goldfrom the Raub area, Pahang, Malaysia. British Geological Survey TechnicalReport WC/95/20,26p.

Henney, PJ, S tyles, M T, Wetton, PD & Bland, DJ (19 95) Characterisation of goldfrom the Penjom area, near Kuala Lipis, Pahang, Malaysia. British GeologicalSurvey Technical Report WC/95/21,24p.

Lecomte, P and Colin, F (1989) Gold dispersion in a tropical rainforest weatheringprofile at Dondo M obi, Gabon. Journal of geochemical exploration , 34,p.285-301.



29/34


Leroy , JC (19 95) New mines spring from old workings in Zimbabw e. The Northernminer, February 6 , p. B6-B7.

Lins, FF & Adam ian, R (1993 ) Som e chemical aspects of gold particles flotation.Proceedings of XVIII International Mineral Processing Co ngr ess, Volume 5 :Gold processing, hydrom etallurgy & dewatering and m iscellaneous,Australasian Institute of Mining and Metallurgy, p. 1 19-1122.

Lucion, Ch & Cuyper, J de (199 3) Preprocessing of cyanided goldcalcine. Proceedings of XVIII International Mineral Processing Congress,Volume 5 : Gold processing, hydrom etallurgy & dewatering andmiscellaneous, Australasian Institute of Mining and Metallurg y, p. 1 129-1134.

Marsden, JO & Fuerstenau, MC (1993 ) Com parison of Merrill Crowe precipitationand carbon absorption for precious m etal recovery. Proceedings of XVIIIInternational Mineral Processing Cong ress, Volume 5 : Gold processing,hydrometallurgy & dewatering and m iscellaneous, Australasian Institute ofMining and Metallurgy, p. 1189-1194.

Martins, V, Dunne, R & Delahey, G (1993) New Celebration tailings treatment plant- 18 months later. Proceedings of XVIII International Mineral ProcessingCongress, Volume 5 : Cold processing, hydrometallurgy & dewatering andmiscellaneous, Australasian Institute of Mining and Metallurg y, p. 1215- 1222.

Melis, LA & Rowson , JW (1989 ) Operation of a gold extraction circuit for recoveryof gold from uranium m ill tailings at Cluff Lake, Saskatchewan. CIM Bulletin,82, NO. 93 1, p.40 - 46.

Moyle, AJ, Doyle, BJ, Hoog vliet, H & W are, AR (1991) The geology andmineralisation of the Ladolam g old deposit, Lihir island, Papua N ew G uinea.Proceedings of PNG geology, exploration and mining conference, Rabaul,Australasian Institute of Mining and M etallurgy, p. 101-111.

Petruk, W (1989 ) Recent progress in m ineralogical investigations related to goldrecovery. CIM Bulletin, 82, No. 93 1, p.37 - 39.

Shelp, SG and Nichol, I (1987) Distribution and dispersion of gold in glacial tillassociated with gold mineralisation in the Canadian shield, Journal ofgeochemical exploration, 27 , p. 3 15-336.



30/34


Solozhe nkin, P, Vo lkov, SV, Shaluhina, ZA Fokina, ZA Zinchenko & Emelyanov,AF (1993 ) Th e technology of gold recovery from black sands by the m ethodof liquid phase chlorination. Proceedings of XVIII International MineralProcessing Congres s, Volume 5 : Gold p rocessing, hydrom etallurgy &dewatering and miscellaneous, Australasian Institute of Mining andMetallurgy, p. 1147-1152.

Stewart, DF and R amsay, PW (1993) Improving the simple sluice box. Internationaljournal of mineral processing , 39, p, 119-136.

Styles, MT , M Perez-Alvarez, Bland, DJ, Wetton, PD & Naden, J (1992)Characterisation of gold from Ecuador. British Geological Survey TechnicalReport WG l92148, 13p.

Styles, M T, Bland, DJ & W etton, PD (19 94) Characterisation of gold fromMersing, Johore, M alaysia. British Geological Survey TechnicalReport WGl9417, lop.

Styles, MT, Wetton, PD & Bland, DJ (1995) Ch aracterisation of go ld fromZimbabwe: Part 2 . Alluvial and soil gold. British Ge ological Survey TechnicaZReport WC /95/5,2 lp.

Subasing he, GKN S (1991) Drawbacks in sli ice box operating practices in P apuaNew Guinea. Proceedings of PNG geology, exploration and m iningconferenc e, Rabaul, Australasian Institute of Mining and M etallurgy, p.132-136.

Subasinghe, GKN an d Maru, Y (1994) The use of mercury in small-scale gold miningoperations: effectiveness and pollution aspects. Proceedings of PN G g eology,exploration and m ining conference, Lae, Papua New Guinea , AustralasianInstitute of Mining and Metallurgy, p.264-270.

W alsh, DE & Rao, PD (1988 ) Study of the compound water cyclone's concentratingefficiency of free gold from placer material. CIM Bulletin, 81,No. 919, p. 53-61.

W ang, W and Poling, GW (198 3) Methods for recovering fine placer gold. CIMBulletin, 76, No. 8 60, p.47-56.



31/34

L o n g i u d i n o I s e c t i o n

Plon

Figure 1. Schematic outline diagram of a pinched sluice

To i l i n q sove r f tow--0 0 0 0 0

/Hutch1 I Co n c e n t r a t e1 d i s c h a r g e sp igo t

Figure 2. Schematic outline diagram of a mineral jig


32/34


33/34

Figure 5. Schematic outline diagram of an MGS (Multi-gravityseparator)


34/34

Trammel s c r e e n i n gat 500 pmi

-c\.SIu i c e box

to recover coarsegold (>500

Slu icc b o x \ Fine tailingsto recover fine gold(c500 pm)

Coarse tailings(>500pm)

(e500 pm)\I 1

Screen4 0 0 pm tailings

on 200 pm

Fine concentrate(containing gold 4200 - 40 pm)

Fine tailings(4200pm; gold

a review of gold particle size and recovery methods wc-97-014

Documents