7/28/2019 CS and Cyanide

1/8

~J . S . A f r. I ns t . M i n . M e t a ll . , v o l . 9 3 , n o . 4 .A p r . 1 99 3 . p p . 1 0 5 -1 12 .

The behaviour of mild steel in aqueous cyan ide solu tions athigh temperatures b y J . R . d e Wet* and R.F. Sandenbergh *

SYNOPSIST he be havio ur o f m ild s te el in a qu eo us cy an id e s olutio ns sim ilar to ca ustic c yan id e elutio n so lu tio ns w asin ve stig ate d a t h ig h te mp era tu re in e le ctro ch em ic al e xp erim en ts . T he a no dic p ola riz atio n d ia gra ms o f m ilds te el in dic ate d th at s te el m ay u nd erg o a ctiv e d is so lu tio n u nd er th es e c on ditio ns . T he ra re o cc urre nc e o f g oldp la tin g in p ra ctic e is th e re su lt o f th e lim itin g o f th e a no dic d is so lu tio n o f th e s te el b y th e p re se nc e o f a p as siv ela ye r o n th e s te el. T his p as siv e la ye r is s ta ble o nly in a c er ta in p ote ntia l r an ge , a nd th e fre e c orro sio n p ote ntia la tta ine d d ete rm ine s w he th er th e ste el w ill be p as siv e or active . F or co nd itio ns typica l of e lu tion co lu mn s,s eve ra l co rro sion p ote ntials are p oss ib le , de pen ding p rim arily o n the in itial co nd itio n o f th e ste el a nd thec on ce ntra tio n o f o xyg en in the s olution . If th es e co nd itio ns re sult in the fo rm atio n of a p as siv e la ye r o n th es te el, c ha ng es in c on ce ntr at io n a nd te m pe ra tu re w ill n ot n ec es sa ri ly r es ult in p la tin g s in ce t he la ye r m a y e xis ti n a m e ta st ab le f or m .SAMEVATTINGD ie g ed ra g v an s ag te s ta al i n w at er ig e s ia ni ed op lo ss in gs s oo rt ge ly k a an h oe t em p er at uu r e lu ee ro pl os sin gs ism e t b eh ulp v an e le ktr oc he m ie se t eg nie ke o nd er so ek . A no die se p ola ris as ie dia gr am m e d ui d aa ro p d at a kti ew eo plo ss in g v an d ie s ta al m oo ntlik is o nd er h ie rd ie to es ta nd e. D ie s eld sa me v oo rk om s v an g ou dp la te rin g in d iepraktyk kan toegeskryf w ord aan die beperking van die anodiese oplossing van die staal deur diet ee nw oo rd ig he id v an 'n p as sie we la ag o p d ie s ta al. D ie p as sie we la ag is s la gs in 'n s ek ere p ote ns ia alg eb ie dstabiel en die korrosiepotensiaal sal bepaal of die staal passief of aktief is. V ir toestande tipies vane lu ee rk olo mm e is 'n a an ta l k orro sie po te ns ia le m oo ntlik , a fh an ge nd v an d ie in is ie le to es ta nd v an d ie s ta al e nd ie s uu rs to fk on se ntra sie in d ie o plo ss in g. In die n h ie rd ie to es ta nd e le i to t d ie v or min g v an 'n p as sie we la ag o pd ie s ta al s al d ie v era nd erin g v an k on se ntra sie e n te mp era tu ur n ie n oo dw en dig to t p la te rin g le i n ie a an ge sie nd ie l aa g in 'n m e ta st ab ie le to es ta nd k an b es ta an .

I N T R O D U C T I O NThe successful use of mild steel as a construction materialfor gold-leaching plants can largely be attributed to itsgood resistance to corrosion in aerated aqueous cyanidesolutions. The reason for the good performance of mildsteel lies in its ability to form passive layers thateffectively shield the steel substrate from attack by thesolution. How ever, it is well known that gold w ill becomeplated onto mild steel under certain conditions, and thatthis may cause problems with the plating of gold onto iron-rich particles I. The carbon-in -pulp (CIP) processintroduced new and potentially m ore aggressive processingconditions to gold-recovery circuits in the form of the acidtreatment of carbon and the elution of loaded carbon withhigh-temperature caustic cyanide solutions. Goldcementation can indeed take place on elution columns, ashas been confirmed by experience in gold-plant elutionoperations2,3. The w ork described here on the behaviour ofm ild steel in aqueous cyanide solutions at h igh tem peraturewas initiated by Genmin, and was undertaken to moreclearly define the conditions under w hich gold plating w illoccur and to evaluate the resistance of various steels to thisphenomenon.

The Pourbaix diagram for the iron-water system shownin Figure 1 indicates that iron will dissolve actively over awide range of potentials at low pH values, but only over a. D epartm ent of M aterials S cience and M etallurgical E ngineering,U n iv er sit y o f P re to ri a, P re to ri a 0 00 2.@ T he S ou th A fric an In stitu te o f M in in g a nd M eta llu rg y, 1 9 9 3 . S A IS S N0 0 3 8 - 2 2 3 X / 3 . 0 0 + 0.00 . P ap er rec eive d S ep te mbe r 1992; revisedp ap er r ec ei ve d F eb ru ar y 1 9 9 3 .

sm all range of potentials at high p H values. A t interm ediatepH values, oxides of Fe(II) are formed at low potentials,and Fe ( Ill) at higher potentials. These oxides will inhibitthe kinetics of iron dissolution depending on the solubilityand retention of the oxidized species on the surface.Insoluble compounds may lim it the corrosion ratesufficiently for steel to become effectively corrosion-resistant, i.e. passive. Iron (Ill) species such as a-FeOOH,a-Fe203' an d y-Fe203 h av e lo w e le ctrica l co nd uctiv ities,while iron species with a greater number of Fe(II) siteshave greater conductivities5 and are considered to be lesseffective passifiers. O nce a passive species is present on thesurface of iron, further oxidation is slow. The goodcorrosion resistance of steel that is usually obtained inalkaline gold pulps indicates that very protective oxidesfo rm u nd er th ese c on ditio ns.

A n increase in tem perature leads to a significant increasein the solubility of iron oxide at high pH values owing tothe formation of soluble HFeOz, as indicated in thePourbaix diagram for this system at 90C (Figure 2). Thepossibility that iron will corrode in alkaline solutionsthrough the formation of HFeOz is therefore increased athigh temperatures, as has been found by a number ofinvestigators6-9. T his is in contrast to the behaviour at low ertemperatures, where active dissolution seldom occurs ath ig h p H v alu eslO .

Cyanide in aqueous solution m ay change the behaviour ofiron by the formation of soluble iron cyanide complexes inpreference to insoluble oxides, as is indicated in thePourbaix diagram for the system Fe-CN-H2O at 25C(Figure 3). However, little is known about the combinedinfluence of cyanide and high temperature on the

J o u r n a l o f T h e South A f ric a n I ns ti tu te o f M in in g a n d M e ta llu rg y 10 5PRIL1993

7/28/2019 CS and Cyanide

2/8

2000

16001200 """"

~ 800~J; 40 0en

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -corrosion

-800

~ 0 """'"~ -400 "'" ..........w

-1200 ../'corrosion-1600 immunity

-2000 -2 6pH

1210 142 4

F igure 1-P ourb aix dia gram4 fo r the system F e-H 2O a t 25 edissolution of iron, and this was one of the aspects dealtw ith in the present investigation. In C IP elution circuits, thepresence of cyanide will favour the dissolution of iron toform a ferrous cyanide complex while also stabilizing thegold in solution as the aurocyanide com plex. T em peraturesof up to 75C do not have a significant influence on thestability of the aurocyanide complex at constant cyanidelevels, as shown in Figure 4. However, a decrease in thecyanide concentration, as could result from the hydrolysisof cyanide at high temperaturesl3, or from elutionoperations without cyanide additions (as is now typical formany South African operationsI4), will make the auro-cyanide complex less stable. The driving force for goldplating is determ ined by the stabilities of both theaurocyanide and the soluble iron species, such as the ironcyanide complexes and hypoferrite, and will depend on

1200"'"

"""""'"

"""..............

"""",

80 0

40 0w:I:en

~ 0.. Fe +,~ ..........w -400 "'"

Fe,O3

-800 ". HFeO,"" "

""Fe

-1200 0 12104 14-log (aH + ) at 90Ge

F igu re 2-P ou rb aix dia gram6 fo r the system F e-H 2O a t 90 C

how these are influenced by changes in the solutionconditions.In the present work, the influence of hydroxide andcyanide in aqu eous solu tio n o n the d issolution o f iron w asin ve stig ate d fo r p ro ce ss c on ditio ns ty pic al o f e lu tio n. Theaim w as to more clearly define th e conditio ns u nder w hichgold plating w ill take place so that these can be avoided inthe operation of C IP elution plants. A com parison of theperform an ce of different steels in cyanide and hyd rox ides olu tio ns a t h igh tempe ra tu re s was a ls o made .

E X P E R I M E N T A L

16

The dissolution of iron in cyanide and hydroxidesolutions at high temperatures was investigated by the useof electrochem ical polarization techniques and m easure-ments of corrosion potential. All the experiments wereconducted in an autoclave m ade of type 316 stainless steel,as shown in Figure 5. A silver-silver chloride (SSC)reference electrode maintained at room temperature andconnected to the cell through a pressure-reducing Luggintube was used for the potential m easurem ents. The Luggintube was filled with saturated potassium chloride, and athread of asbestos wool was passed through the tubing toensure continuity of the electrolyte. A counter-electrode ofplatinum was used in the polarization experim ents, whilethe working electrode consisted of the material underinvestigation. The temperature in the autoclave wascontrolled by an electrical-resistance elem ent fitted roundthe outside of the autoclave, a thermocouple inside theautoclave, and a tem perature controller. The polarizationstudies were conducted under isothermal conditions, andthe gold-plating experiments were conducted while thetem perature w as steadily increased from room tem peratureto 110C at a fixed rate. Sam ples of low-carbon steel, type304 stainless steel, and 3CR12 (nominal compositionsshown in Table I) were used. The samples were ground to a600-grit finish, and w ere w ashed in de-ionized w ater beforebeing immersed in the solution. T he preparation w as carried

20001600 F e ( C N ) 3.

Fe+3 FeOOH FeOOH1200"""

80 0 ................

w

". ................:I: 400enCl) 0. . Fe+,~ ............... . -400 Fe (CN ) 4 .......................-..

-800-1200 Fe Fe(OHj-'4-1600 [Fe] = 10-4 M[CN] = 10-3 M

16 -2000 0 8 12 1410 166pH

Figure 3-Pourbaix diagram ll for the system Fe-C ~20 at 25C

1 0 6 A P R IL 1993 J o u r n a l o f T h e S o ut h A fr ic an I n s t i t u te o f M in in g a n d M e ta llu rg y

7/28/2019 CS and Cyanide

3/8

Material C Mn P S Si Cr Ni Al Ti MoType 304 S; S; S; S; S; 18,00-- 8,00-- - - -stainless 0,08 2,00 0,045 0,03 1,00 20,00 10,50steel

f--- ------M ild 0,08- 1,75- S; S; 0,05- S; S; S; - -steel 0,16 2,25 0,025 0,025 0,25 0,015 0,15 0,D33CRI2 S; S; S; S; S; 11 - S; - S; -0,03 1,5 0,04 0,03 1,0 12 1,5 0.6

3000

2000 Au(OHI,----------------~.~,., ", Au(CN}-, U_---

~1000

--~--';':~'::'~:'~'::'::'~:'~'-': '~:~':'~:'~~.:.::.::.~:.~...:'..'uii 0 - '-,"," ,-, ,-, '-,", -" -,,-, C-,-, ' -'C ;~- 'C-,- , C- ~C- ;C-,CO-;C-;C--w -1000

-2000 25 75 [AuJ = 10' M[CN] = 10' M

-3000 0 8 104 6 12pH

Figure 4-Pourbaix diagram for the system Au-CN-H2O,indicating the influence of tem perature on the stability of theaurocyan ide spec ies12out just prior to the experim ents to avoid variations due tothe form ation of different oxide layers on the surfaces ofthe sam ples. The solutions w ere m ade up from reagent-g rade chemical s and d is ti ll ed wa te r. The solut ions typ ical lyused in Z adra elution circuits contain about 0,5 m ass percent NaCN (O,IM NaCN) and I to 2 mass per cent NaOH(0,25 to 0,5M NaOH), and the concentrations of 0 M ,0,1 M , and 0,2 M NaCN and 0,05 M and 0,5 M NaOH usedin the experiments were chosen to approximate theseconditions and possible excursions. T he tw o hydroxidecon centra tio ns y ie ld th eo re tic al pH valu es o f 1 2,7 and 13,7respectively at 25c. Potassium sulphate w as added to ac on cen tra tio n o f 0 ,1 M K2SO4 to e nh anc e th e c on du ctiv ityo f th e so lu tion . T he v olume o f th e so lutio n u se d was 3 00 m l,which gave a ratio of volume to sample area of approx-imately 50 ml /cm2 for a ll t he exper iment s.In th e p ola riz atio n exp erimen ts , th e auto cla ve was fille dw ith the solution and w as then de-aerated by the bubblingof purified nitrogen through it to give an oxygenconcentration in the solution of less than 1 mg/l asm easured w ith an O xi-92 oxygen m eter. The electrodeswere then immersed in the solution, and the workingelectrode w as cleaned by being polarized to a potential of-1 V (SSC) for 5 minutes, after which the potential wasa llow ed to mov e fre ely . T he rep ro du cib ility o bta in ed w ith

Table INo min alc o mp o s it io n s of th e ma te ria ls u s e d ( in p e rc e n ta g e s )

TemperaturecontrolPotentiostat x- vrecorderAux R ef W

SS Celectrode

ThermocoupleCounterelectrode

14

P res su re g au gePt

N it ro ge n f ee dSample

S af et y v al ve

P re ss ur e v es se l

Inlet/outlet

Figure 5-Schem atic diagram of the experim ental set up forthe pola riza tion s tud iesth is tre atm ent w as fo un d to b e re la tive ly g ood , a s sh own inF ig ure 6 . T he au to clav e was h ea ted immed iate ly after thistr ea tmen t, a h ea tin g r ate o f a pp ro ximate ly 1 0C p er m in utebeing achieved. The experim ents w ere started after thetem perature had stabilized to w ithin 1C of the set value,wh ic h to ok app ro xima te ly 1 5 m inute s from th e time th at th eheating started. A m odel 176 P rinceton potentiostat w asu sed fo r th e po la riza tio n e xpe rim ents , w hic h s tarte d at th ecorrosion potential, the potential then being scanned at arate of 3,3 m V/s in the positive direction. The current-vo ltage curves were recorded cont inuously .As the validity of the results would be affected if as ign ific ant lo ss o f cy an id e o ccu rre d d ue to h ydro ly sis , th elo ss o f c yan id e was d eterm in ed du rin g the ex pe rim en ts. T odo this, a solution containing 0,2 M NaCN and 0,05 MNaOH was heated to a temperature of l30C andmain tain ed at th at temp eratu re fo r th e du ra tio n o f a n orm alexperim ent, w hich w as approxim ately 5 m inutes; afterc ooling , th e so lution was an alys ed . T he lo ss o f cy an ide wasfou nd to be ab ou t 1 8 mas s p er c en t. B ec au se the co nd itio nsin th ese e xpe rim en ts w ere co nsid ered to b e fa vou ra ble fo rth e hy dro ly sis o f cy an id e (h igh cy anid e an d low hy dro xideconcentrations), and because the norm al polarizationexp eriments n eeded sho rte r time s a t tempera tu re , th e lo ss o fc yan id e d urin g th e e xp erim en ts w as n ot co nsid ered to ha veseriously affec ted the resu lts.T he a ctiv ation b eha vio ur o f m ild ste el an d th e p latin g ofgold from solutions m ade up in the laboratory and fromZadra eluate were evaluated with the steel in differentinitial conditions, and with and w ithout oxygen in the

J o u r n a l o f T h e S o u th A f r i c a n I n s t i t u t e o f M i n in g a n d M e t a l lu r g y 10 7P R I L 1993

7/28/2019 CS and Cyanide

4/8

109 0.05 M N aO H110.CScan rat e = 3 .3 mV/s87

"E 6u

1 5 0,2 M NaCN432

0-900 -500 -400800 -60070 0E (mV V S. S .H .E .)

Figure 6-Polarization diagrams for mild steel in cyanidesolutions con taining 0 ,05 M N aOH at a tem perature o f noc.The sam ples w ere pre-polarized at -1 V ISSC ) for 5 m inutes,and potential scanning w as started from the open-circuitpotential at a rate of 3,3 m V/sso lu tio n. Gold was a dd ed to a ll th e s olu tio ns, in clud in g th eZ ad ra elu ate , to a c on ce ntra tio n o f 1 00 p.p .m o f po ta ssiumaurocyanide. A fter the sam ple had been immersed in theso lu tio n, h ea tin g was s tarted at a rate o f 1 2,5 C p er m in utewh il e the cor ro sion poten ti al was r ecorded con tinuously.

R E S U L T S A N D D IS C U S S IO NInfluence of Sodium Cyanide and Hydroxide on theD issolution of M ild SteelThe influence of cyanide and sodium hydroxide on theanodic polarization characteristics of mild steel at atemperature of 110C are shown in Figures 6 and 7. In allcases, higher concentrations of either h ydroxide or cyanidegave rise to an increase in the active peak current densities.In the case of the hydroxide, the low er potential boundaryof the active peak is most probably due to the hydrogenreaction, while the upper boundary is defined by theh yp ofe rrite-m ag netite o r h yp oferrite -h em atite b ou nd arie s,as indicated in Figure 2. The low current densitiesmeasured at higher potentials are due to the passivation ofthe mild steel by the formation of Fe304, FeOOH, orFe203' Consequently, no gold cementation would beexpected at these potentials. A t the higher concentrationsof hydroxide, the active region moved down to lowerpotentials, as would be expected in view of the lowerreversible potential for the hydrogen reaction at increasedpH values. The anodic peak current also increased,probably due to the increased stability of the hypoferritespecies (Figure 2) but, in view of this, it is surprising thatthe poten tial region w as not ex panded.Cyanide additions caused the active potential region toexpand, and increased the peak current densities

significantly. This effect was more pronounced at higherhydroxide concentrations, but increased current densitiesw ere also noted at low er pH values. This is in contrast to

109 0,2 M NaCN 0,5 M NaOHll0.eS ca n ra te = 3.3 m V/s8

-400

7

- 6Eu~ 5.s.- 432

0-900 -700 -600E (mV VS. S . H . E . ) -500

F igure 7-Pola riza tion diagrams for m ild steel in cyanidesolutions containing 0,5 M NaOH at a temperature of noc.The sam ples were pre-polarized at -1 V ISSC) for 5 minutes,and potentia l scanning was started from the open-circuitpotentia l at a rate of 3,3 m V/sthe results of Kenna et al.3, w ho found that cyanideinhibited the corrosion of steel at a pH value of 9. Thisdiscrepancy may have been due to the presence ofpotassium sulphate in the solution, w hich w as used in thep re se nt te sts to e nh an ce th e condu ctiv ity o f th e e le ctro ly te .T o investigate this, som e experim ents w ere conducted incy an id e-fre e s olu tio ns at pH 9 a nd 1 10 C w ith a nd w ith ou ta dd itio ns o f K2SO4 'The re su lt s a re shown in F ig ure 8 , fr omwhich the considerable anodic activity of m ild steel in thesolution containing potassium sulphate is immediatelyapp are nt, wh ile th e p as siv ity was ma in ta in ed in th e abs en ceof potassium sulphate. The potassium sulphate mayfa cilitate othe r rea ctio ns o n th e s urfa ce o f th e ste el, an d th ean od ic ac tiv ity m ay b e d ue n ot to iron d iss olu tio n b ut to th ed ecompos itio n o f th e e le ctro ly te . Howeve r, t he in ac tiv ityobserved for the type 304 stainless steel (F igure 8) m akesthis less probable and strongly suggests that the anodicbehaviour observed at pH 9 by Kenna et aP can beascribed to the dissolution of the m ild steel facilitated bythe potassium sulphate background electrolyte. A t theh ig her h yd ro xid e co nc en tra tio ns u se d in th is s tu dy , th erewas no in flu en ce o f sulp ha te o n th e re su lts .The in cre as ed anodic a cti vi ty w ith an in cre as e in cyanid eco nc entra tio n is p ro ba bly d ue to th e fo rm atio n o f a s olu bleiro n cyani de comp lex th at is s ta ble in th is p ote ntia l re gio n i fthe 25C diagram (F igure 3) is taken as a guideline. It w asco ns id ered th at th e an od ic pe ak c ou ld a ls o h ave be en d ue tothe oxidation of cyanide to cyanate, for which theequilibrium potential is close to that of hydrogen, andsh ou ld th ere fo re b e th ermod yn am ic ally p oss ib le at th es epotentials. H owever, the decrease in the rate at higherpotentials is difficult to explain on this prem ise, as is thelack of activity found for the stainless steel (Figure 9).A no th er a pp are nt fea tu re o f th e a no dic cu rv es is th e s plit inth e c urv e th at d ev elo ps at h ig her c ya nid e c on ce ntra tio ns.T here is no obvious explanation for this and, in the view ofthe l im i ted p ract ic al impl icat ions , i t was not pur sued .

10 8 A P R I L 1993 J o u r n a l o f T h e S o u th A fr ic an I n s t i t u te o f M i n in g a n d M e ta llu rg y

7/28/2019 CS and Cyanide

5/8

8,07,5

0,03

0,05 pH = 9110.CS can rate = 3,3 m V/s,06,5

304 S'a ln le ..0,04 0, lM K ,S O,No cyanide

6,05,55,0

0,020,010,00-4{)0 -300

"E 4.5u~ 4,0.s 3,53,0

O ,1 M K ,S O.0,1M N aCN. .. ,2. 52, 01,51,00,50,0 .' ""'. . """""-500 -400 -300 -200 .100

E (mV vs. S . H . E . )Figure 8-Polarization diagrams for mild steel in cyanidesolutions at Hoot and a pH of 9 (measured at 25CI. Thesamples w ere pre-polarized at -1 V (SSC I for 5 m inutes, andpotential scanning was started from the open-circuitpotential at a rate of 3,3 mV/sAn increase in temperature increased the active peak and

expanded the active potential region to more negativevalues, as shown in Figure 10. The split in the anodic peaknoted at the lower temperatures became more pronouncedat higher temperatures. The small increase in the anodicpeak with temperature indicates that the process isprobably under diffusion control, as confirmed by the

109

0,05M NaOH0,2M NaCN110OCS can ra te = 3 ,3 m V/s87

- 6Eu 5..- 4 M il d s tee l32

0-900 -600 -500 -40080 0 -700E (mV v s. S .H .E .)

Figure 9-Polarization diagrams for mild steel, type 304stainless steel, and 3CR12 in a solution containing 0,05 MNaOH and 0,2 M NaCN at a temperature of HOot. Thesam ples w ere pre-polarized at -1 V (SSC I for 5 m inutes, andpotential scanning was started from the open-circuitpotential at a rate of 3,3 mV/s

109 O ,O 5M N aO HO ,2M NaCNS ca n ra te = 3,3 m V/s8 130.C7

- 6NEu 5..- 432

0-900 -800 -700 -600 -500 -400E (mV vs. S . H . E . )

Figure 10-Polarization diagram s for m ild steel in a solutioncontaining 0,05 M NaOH and 0,2 M NaCN at varioustemperatures. The sam ples w ere pre-polarized at -1 V (SSCIfor 5 minutes, and potential scanning was started from theo pe n-circu it p oten tial at a rate o f 3 ,3 mV/srelatively low activation energy of 25 kJ/m ol calculatedfrom th e p eak h eig hts .

Comparison of M ild S teel, S tainless S teel, and3CR12The polarization characteristics of type 304 stainless steeland 3CRl2 in 0,2 M NaCN and 0,05 M NaOH arecompared with that of mild steel in Figure 9. Type 304shows much lower anodic current densities than mild steelunder these conditions, w hich w ill facilitate passivation andm ake gold plating less likely. The 3CR l2 appears to be m oreactive than type 304, but less so than the mild steel and,while the lower potential boundary w as close to that of type3 04 , th e m easu re d p ea k c urren ts w ere sig nifican tly lower.

The Plating of GoldSince gold is readily available in an oxidized form asaurocyanide throughout the extraction process, thedetermining factor in the plating of gold is the activedissolution of iron. For steel in cyanide and hydroxidesolutions at high tem perature, the active state is stable for awide range of conditions coinciding with normal elutionconditions. The peak currents measured for mild steel arealso relatively high, which makes the active dissolution ofmild steel likely from a kinetic point of view . However, theisolated reports of gold plating found in practice does notsupport this, and it appears that steel may be protected bysome or other coating of metastable oxide. The removal ofthis oxide, at least in part, appears to be necessary foractivation of the m ild steel.

To investigate this, the behaviour of mild steel and theplating of gold from solutions made up in the laboratoryand from Zadra eluates were evaluated for the steel indifferent initial conditions with and without oxygen in thesolution. Gold was added to all the solutions, includingthe Zadra eluate, to a concentration of 100 p.p.m . ofpot as si um aurocyan ide.

J o u r n a l o f T h e SouthA f ric a n In s ti tu te o f M in in g a n d M e ta llu rg y 10 9P R I L 1993

7/28/2019 CS and Cyanide

6/8

The variations in the potential of the samples withtemperature are indicated in Figure 11. For the sample inthe wet abraded condition that was exposed to a de-aeratedsolution containing 100 p.p.m . of potassium aurocyanide,the potential was initially at relatively positive values butdecreased as the solution was heated. The sample wasactivated at approximately 90C, and the potentialdecreased sharply to a value of -750 mV(SHE) and stayedat that potential for the duration of the experiment. Thesam ple was gold in colour after the experiment, and EDAXanalysis of the surface of the sample indicated that gold hadindeed been plated on its surface. The steel sample etchedin hydrochloric acid was active right from the beginning ofthe experim ent before the heating started, even though thesolution w as not de-aerated prior to the experim ent. A gain,gold plating was noted at the end of the experiment. In thecase of wet abraded specimens exposed to aeratedsolutions, the potential remained at noble values, and nogold plating was noted on the steel at the end of theexperiments. In the case of the Zadra eluate, the potentialmoved to a positive value and no gold plating was observedeven though the solution had been de-aerated. As theconcentration of hydroxide in the solution was fairly high(4,2 m ass per cent), the observed behaviour m ust have beendue to the low concentration of cyanide (0,015 mass percent), and this tentatively indicates that gold plating shouldnot be a problem in these eluates.These results can be explained by a consideration of the

polarization characteristics of gold and steel together. Forgold plating on iron, the active region of the iron mustcoincide with the stability region of metallic gold, as waspointed out in the Introduction. If the data of G irardi e t a /. l2are taken as a guideline, the transition from the cyanidecomplex to metallic gold occurs at approximately

-150 1.8tim e (m in i3.4 66.0

-200-250

0,05M NaDH0,2M NaCNNo de -a e ra t ion-300-350 0 ,0 5M N aDH0 ,2M N aCNDe-aerated40 0

WJ: -4500>..5. -500W -550

S pe nt e lu an ts ol ut io n s pi ke dto 100m g/l K Au(C N)2De-aerated

-600-650

0,05M NaD HO.IM NaCNNo de -a e ra ti onEtched in 10% HCI-700

-750-800 40 60 80

Temperature (aCI100

Figure 11-0pen-circuit potentials (corrosion potentials)obtained for samples of mild steel in the conditions andsolutions at the tem peratures indicated. T he tim e scale at thetop gives the approxim ate tim e it took for the sam ples to beh ea te d to th e re qu ire d tempe ra tu re s

-500 mV(SHE), which means that gold plating is possibleat the lower potentials measured for steel in thisinvestigation. T he kinetics of gold precipitation onto m etalshas been investigated in the past, and data are available forreagent concentrations sim ilar to those investigated here,although only at lower temperatures. If it is assumed thatsignificant gold plating w ill occur w hen the current densityfor the gold cathodic reaction is 0,1 m A/cm 2, the data of 00and Tran15 (500 p.p.m . of gold, 40C) indicate that thiscurrent density will be achieved at approximately-660 mV(SHE), which is close to the anodic peak for thedissolution of iron. The precipitation of gold is thus highlylikely, although the gold reaction may be depolarized tosuch an extent at the higher tem peratures that it m ay passifyth e iro n s urfa ce .This is certainly a possibility if the decrease in the

activation polarization found by 00 and Tran15 between 10and 40C is taken as a guideline, but passivation ispossible only if the anodic peak current density of the ironis lower than the cathodic current density for the gold-reduction reaction that is likely to occur at lowconcentrations of hydroxide and cyanide and at lowertem peratures. A t higher concentrations and tem peratures,the active peak for iron will be so high that depolarizationof the cathodic reaction will result only in higher rates ofgold deposition. This last statement must be qualified inthe sense that som e degree of m etastability may exist in thesystem , and that the free potential or corrosion potentialattained by the steel will depend very much on the initialconditions, as has already been pointed out. For exam ple,when the sample was etched in acid or heated in oxygen-free water, the potential stayed fairly negative for theduration of the experiment. If the heating was done in thepresence of oxygen, the potential moved in a positivedirection and stayed there. For these passified specim ens,the current density of the cathodic gold-plating reaction islikely to be higher than the anodic current density, and thepositive shift in the corrosion potential with an increase intemperature may be a result of the depolarization of thec ath od ic re ac tio n a t h ig he r tempe ra tu re s.The way in which the reactions possibly interact to cause

this behaviour can be illustrated by a schematic Evansdiagram , as shown in Figure 12. The important feature isthat three corrosion potentials are possible for this system ,of w hich the m ost negative and the m ost positive potentialsare likely to be fairly stable. In this type of system , thecorrosion potential depends on the surface condition of themetal and on the solution conditions during the initials ta ge s o f th e e xp erimen t.Where the conditions are such that the system will be at

the more positive potential initially, e.g. a stable filmpresent on steel or conditions conducive to film form ationbefore high tem peratures are reached, the m etal w ill tend tostay at that potential. Very little corrosion of steel and noplating of gold will take place, since the steel will bepassive and the aurocyanide complex will be stable at thatpotential. If the conditions change so that the system is nolonger stable at the positive potential, e.g. by a decrease inthe oxygen concentration, the system will tend to move tom ore negative potentials, at which the plating of gold willbe possible. A nother possibility is that the steel is initiallyin the active state, or becomes so by the removal of its

110 A P R I L 1993 J o u r n a l o f T h e S o u th A f r i c a n In s ti t u te o f M i n in g a n d M e t a l lu rg y

7/28/2019 CS and Cyanide

7/8

10O ,O 5M N aO H9 110CS ca n ra te = 3.3 m V/s

87

N" 6E,) ~5! 0,2M N aC N432

High oxygenconc.

0-800 -400 -300E (m V v s . S .H .E . )

Figure 1 2 - E v a n s d iagram f or s te el i n s ol ut ion s con ta in ingNaO H, N aC N, and oxygensurface layer. This will sim ilarly result in more negativepotentials and gold plating, with the difference that thepotential will tend to move to more positive values once thedefect is covered with gold, which will then once againdissolve. The only negative aspect of this w ill be that someiron will be dissolved during each of these cycles, whichmay result in some thinning of the w all of the vessel.

C O N C L U S I O N S-----The tendency of gold to be cemented onto steel isdetermined by the thermodynamic stabilities of both thesteel and the gold. These indicate that gold cem entation w illbe more probable if either or both the hydroxideco nc en tratio n an d th e temp eratu re are in cre ase d. A n in creasein cyanide concentration w ill favour the dissolution of iron,but w ill also stabilize the aurocyanide ion.The kinetics of gold cementation are determined by the

rate of diffusion of the aurocyanide ion to the reactioninterface and/or by the anodic activity of the steel. Theanodic activity of the steel increases with increasingh yd ro xid e an d cy an id e co nc en tratio n in th e h ig h-tem pera tu reso lu tio ns ty pic ally fo un d u nd er elu tio n co nd itio ns.In solutions at high temperature, the potential region in

which the steel is in the active state is such that the platingof gold from solutions containing aurocyanide is possible.However, the steel will passify at higher potential values,where the aurocyanide complex is stable and plating willbe un like ly .The characteristics of the Au-Fe-CN-H2O system are

such that the system may be metastable at variouspotentials, depending on the initial condition of the steelsurface and the exposure conditions. The initial presence ofstrong oxidants, such as oxygen, in solution will move thepotential into the passive region, and lim ited dissolution ofiron and no gold plating w ill take place, even if the systemlater m oves to m ore aggressive conditions. C onsequently,the cementation of gold onto the iron will take place only if

the passive layer on the steel is removed by a reducingenvironment such as de-aerated solution, or by pickling inacid, or by m echanical dam age.The nature of the gold plated onto the steel during

excursions to negative potentials, for example because ofdamage to the surface or a decrease in the concentration ofoxygen, will determine whether the cementation of goldw ill continue, or w hether the system w ill start to oscillatebetween gold plating and dissolution. Dendritic depositsthat typically form at low concentrations w ill favourcontinued dissolution, w hile com pact deposits w ill favoura lte rn atin g p la tin g a nd d iss olu tio n.Type 304 is much less reactive than mild steel in

solutions containing cyanide and caustic soda, and goldplating onto its surface is unlikely. The sam e can generallybe expected for 3C R12, although it is slightly m ore reactivethan type 304.From a practical point of view , gold plating on steel can

be avoided by. keep ing the solutio n satu rated w ith ox ygen. avoiding damage to the passive layer on the steel as aresult of contact w ith acid or m echanical abrasio n. avoiding high concentrations of either cyanide orcaustic in the solution , o r avoiding high tem peratures;these factors will not give rise to plating in itself, butwill favour plating in the case of damage to the passivelayer on th e steel.

A C K N O W L E D G E M E N T ST he financial and other assistance provided by G enm inP rocess R esearch, especially D r D .A . H oltum , during thecou rs e o f th is wor k is g ra te fu lly a ck now ledg ed .

R E F E R E N C E S1. DAV ID SON , RJ., BROW N, G .A., SCHM IDT, e.G ., DU NCA NSON, D.,

and TAYLOR, J.D . The intensive cyanidation of gold-plant gravityconcentrates. J. S. Afr. Inst. M in. Metal/., Jan. 1978.p p. 1 46 -1 65 .

2. BAILEY, P.R. Application of activated carbon to gold recovery.The extractive metallurgy of gold in South Africa. S tanley, G .G .(ed.). Johannesburg, The South African Institute of M ining andM etallurgy, 1987. pp. 379-614.

3. KENNA, e.C ., RITCHIE, I.M ., and SINGH, P. The cementation ofgold by iron from cyanide solutions. Hydrometallurgy, vo!. 23.1 99 0. p p. 2 63 -2 79 .

4. POU RBAIX, M . Atlas of electrochemical equilibria in aqueoussolutions. N ACE, 1974. pp. 307-321.

5. STRATM ANN , M . The atm ospheric corrosion of iron-a discussionof the physico-chemical fundamentals of this omnipresent corro-sion process. Invited review. Ber. Bunsenges. Phys.C hem ., vo!.94. 1990. pp. 626-639.

6. ASH WORTH,V., and BoD EN, PJ. Potential-pH diagrams at elevatedtemperatures. Cor ro sio n S cie nc e, vo!. 10. 1970. pp. 709-718.

7. BROOK, P.A. A computer method for calculating potential-pHdiagrams. Cor ro sio n S cie nc e, vo!. 11. 1971. pp. 389-396.

8. MACDONALD, D .D. The electrochem istry of metals in aqueoussystem s at elevated tem peratures. Modern aspects of electro-chemistry, no. 11. Conway, B .E., and Bockris, J.O 'M . (eds.).Plenum , 1974. pp. 141-197.

9. TowNsEN D, H.E. Potential-pH diagram s at elevated temperaturefor the system Fe-H2O. C orrosion Science, vo!. 10. 1968.p p. 3 43 -3 58 .

10. UHLIG , H.H. C orro sio n h an db oo k. WHey, 1948. pp. 129-135,636-638.

J o u r n a l o f T h e S o u th A fr ic an In s t i t u te o f M i n in g a n d M e ta llu rg y 1 1 1P R I L 1993

7/28/2019 CS and Cyanide

8/8

11. OsSEo-AsARE, K., XUE, T., and CIM INELLI, V.S.T. Solution chem -istry of cyanide leaching systems. Precious metals, mining,extraction and processing. Kudryk, V., Corrigan, D.A., andLiang, W.W . (eds.). TM S-AIM E, 1984. pp. 173-197.

12. GIRARDI, S.C., ALRUIZ, O.M ., and ANFRUNS, I.P. Pressurecyanidation: An option to treat refractory gold. Productivity andtechnology in the m etallurgical industries. Koch, M ., and Taylor,I.C. (eds.). The Minerals, Metals and Materials Society, 1989.p p. 6 27 -6 37 .

13. ADAM S, M .D. The chem ical behaviour of cyanide in the extractionof gold. Part 2. M echanisms of cyanide loss in the carbon-in-pulpprocess. J. S. Afr. Inst. M in. M etall., vo!. 90, no. 3. Mar. 199O.p p. 6 7- 73 .

14. DAVEY, LM . Elutions without cyanide. MMM A C ircular No.2/9I,M ine Metallurgical M anagers' Association of South Africa, Dec.1 99 1. pp . 8 5-9 3.

15. 00, M .T., and TRAN, T. The effect of lead on the cementation ofgold by zinc. Hydrometallurgy, vo!. 26.1991. pp. 61-74.

Estimator reduces gold lo sses *An instrument developed by M intek for the estimation ofparticle size in m illing and classification circuits w ill assistgreatly in minimizing the costly losses of insoluble goldduring the processing of W itw atersrand ores.The system is centred on a m icroprocessor-controlledparticle-size estimator (PSE) that uses the signal from

Mintek's hydrocyclone-underflow meter (HUM) tocalculate the particle size of the cyclone overflow in termsof the percentage passing a certain size.Advantages of the PSE over comparable commercialinstrum ents in clu de its. h ig h a cc ura cy a nd re lia bility. low capital and m aintenance costs. ability to read directly the size fraction of interest.Unlik e o th er o n-lin e meth od s fo r p artic le -siz e mea su re -m ent, w hich rely on the taking of a representative sam ple

* I s s u e d b y M in t e k , P riv a t e B a g X3015, R a n d b u r g , 2 1 2 5 .

from a slurry, the PSE uses a signal taken from the entirestream. This avoids inaccurate readings caused by non-representative samples, as well as the use of complexsamp ling mechan isms.The cost of the PSE is very competitive and, owing to

its sim plicity, the system needs very little m aintenance.Once it has been set up, routine calibration can easily bedone by plant personnel. Another major advantage is thatthe PSE can be calibrated for any particle size; forexample, as most losses of insoluble gold occur in thefraction larger than 150 Ilm , it is advantageous to have adirec t re ad in g o f this p arame te r.The PSE system, which forms part of M intek's

hydrocyclone-underflow meter package, can be used as astand-alone particle-size system , or can be incorporated inM intek's m ore com prehensive m ultivariable-control(MVC ) package for the optim ization of m illing circuits.

Engineering Faculty calls for more bursaries*The Faculty of Engineering at the University of theW itwatersrand has expressed concern at the markedd ec lin e in th e n umb er o f b ursa rie s b ein g o ffe re d to stu de ntsby industry .'Bursaries are generally in very short supply', saysP ro fe sso r Roy Adams, D ea n o f th e Eng in ee rin g F ac ulty . 'Inparticular, m iddle-class stud ents w ho se p are nts can no tafford to send them to university are losing out becauseburs ar ie s a re no l onge r b ei ng o ff ered to t hem '.H e adds that the shortage of bursaries is contributingtowa rd s a n a la rm in g d ec lin e in stu de nt n umb ers. 'In du stry* I s s u e d b y L y n n e H a n c o c k C o m m u n ic a t io n s , P .G . B o x 3 7 1 2 ,H o n e y d e w 2 0 4 0 .

should be taking a longer-term view of the engineeringprofession. The students we have in 1993 will be the staffyou have---or do not have if num bers continue to decline-in 2001.'Bearing in mind the cash-strapped position in whichindustry finds itself, Professor Adams suggests thatcom panies need not consider the all-inclusive bursaries thatwere expected in the past. All that is needed is a bursary tocover fees and an allowance for books. If necessary, thebursar could then raise a loan to cover additional expenses.He adds that, although employment of the bursar by thesponsoring com pany on com pletion of his or her studies hasoften been obligatory in the past, there is no need forthe bursary to be linked to a job.

1994. XV CONGRESS OF THE COUNCIL OF MINING ANDMETALLURGICAL INST ITUT IONS24-29 A pril 1994Bill Em mett, C ongress M anager, 15th C MM I C ongress,PO Box 809, Johannesburg 2000, South Africa.Telephones: ( 27 ) ( 11 ) 8 38 -8 211 (office), ( 27 ) ( 11 ) 7 88 -2 51 8 (home)Fax: ( 27 ) ( 11 ) 8 34 .1 884.XV$0 (; C ONGRES S \C ~1'lfERN A . f ~1 1 2 A P R IL 1 9 9 3 J o u r n a l o f T h e S o u th A f r i c a n I n s t i t u t e o f M i n in g a n d M e ta llu rg y