Canadion MinerqlogistVol. 28, pp. i0148 0990)

ABsrRAcr

Grains of placer gold from several localities in thesoutheastern United States exhibit a high-purity gold rim.The individual rims on the various grains range from ( Ito 60 rm in thickness and have silver contents of 3.3 to 0wt.9o (average of 0.9 wt.Yo), even though the core com-positions range from 45.1 to 2.8 wt.qo silver. This gold rim-ming most likely is responsible for the commonly cited casesof gold from placer deposits assaying at higher values offineness than the gold in the corresponding source lode.A gold-rich rim apparently forms by precipitation of goldfrom the surrounding solution, because simple leaching ofsilver from the electrum surface is an ineffective mechan-ism for the enrichment of gold. Diffusion of silver to thesurface of a placer gold grain to expose it to oxidizingmeteoric waters, and thus create a diffusion-enhanced leach-ing process, proceeds far too slowly to produce the observednatural rim thicknesses; furthermore, this mechanism failsto produce the sharp gradients in concentration observedin natural grains. Comparison of the complexation effica-cies of 49 different ligands indicates CN-, OH-, NH3, Cf,I-, Br, and HS- to be the ligands most capable (in decreas-ing order) oftransporting gold in ordinary stream environ-ments. Self-electrorefining of placer electrum grains is alikely process of forming gold-rich rims and probably oper-ates in tandem with dissolution-precipitation (cementation)to produce the observed phenomena.

Keywords: gold (Au), silver (Ag), secondary gold enrich-ment, Au(CN)2-, electrum, gold rims, gold transport,electrorefining, aqueous complexes, gold-species ther-modynamics.

SoN,rMeRs

A plusieurs endroits dans le sud-est des Etats-Unis, lespafticules d'6lectrum des graviers aurifdres montrent unebordure en or trbs pur. Ces bordures peuvent atteindre 60pm en epaisseur, et contiennent entle 3.3 et 090 en poidsd'Ag (en moyentre, 0.990), malgr6 le fait que le noyau deces particules en contient de 45, I d, 2.80/0. La pr€sence d'untel liser6 d'or serait responsable de la teneur accrue en ordes particules d'6lectrum des gtraviers aurifbres, compar6esi leurs sources min6ralis6es. La bordure se forme par pr6-cipitation de I'or d partir des solutions environnantes; unsimple lessivage de I'argent de la surface des particulesd'6lectrum est inefficace comme m6canisme d'enrichisse-menl, La diffusion d'atomc d'Ag vers la surface d'un graind'or dans un sable, pour les mettre en contact avec I'eaum6t€orique oxyg6n6e, et ainsi crder un m6canisme de lessi-vage qui d6pend de la diffusion, proc€de beaucoup troplentement pour expliquer les dpaisseurs observdes. De plus,

GOLD.RICH RIM FORMATION ON ELECTRUM GRAINS IN PLACERS

JOHN C. GROEN, JAMES R. CRAIG AND J. DONALD RIMSTIDTDepartment of Geological Sciences, Virginia Polytechnic Inslitute and State University,

Blacksburg, Virginia 24061 -0420, U.S.A.

ce m€canisme ne peut expliquer la pr€sence de gradientsabrupts en composition. Une comparaison de l'efficacit6relative de complexation de 49 ligands diff€rents montreque CN-, OH-, NH3, Cf, I-, Br- et HS- (en ordre d€crois-sant) pourraient transporter I'or dans un tel mitieu. Un pro-cessus d'autoelecfio-affinage des grains d'6lectrum dans ungravier serart impliqu6 dans I'origine de ces bordures, etagirait avec un ph€nombne de dissolution et pr€cipitation(cimentation) pour produire les liser6s.

(Traduit par la Rddaction)

Mots-clds; or (Au), argent (Ag)t enrichissement secondaireen or, Au(CN);, 6lectrum, bordures d'or, transfert deI'or, 6lectro-affinage, complexes aqueux' thermodyna-mique des complexes aurifbres.

INTRoDUcTIoN

Gold, since before recorded history, has been themost prized of mineral possessions. Because of itsgreat value, economic concentrations of gold can bephysically small in size and can contain as little asl-3 ppm gold so dispersed that the deposits aredifficult to locate and recognize. Consequently, mostof the great gold rushes of history, and indeed manyrecent discoveries, resulted not from initial locationof primary gold-bearing deposits, but rather fromthe detection and recovery of placer gold, which wasconcentrated in streams after erosion of the sourcelode deposits. In many areas, only the placer depositshave proven to be economical to exploil. In 1497,Ulrich Rulein von Kalbe stated that "The gold gener-ated in river sand is the purest and most exalted kindbecause its matter is most thoroughly refined by theflow and counterflow ofthe water and also becauseof the characteristics of the location where such goldis found, that is, the orientation of the river in whichsuch placer gold is made" [translation by Sisco &Smith (1949)1. Since at least as early as von Kalbe'stime, reports have indicated that the gold in placerdeposits is generally purer than that in the lodedeposits from which the placers were derived (Lind-gren l9ll, Boericke 1936, Emmons 1937, Mackay1944, Fisher 1945, Koshman & Yugay 1912, Matn1984). The purity of gold is expressed in lerms offineness, which is the concentration of gold in analloy (in wt.9o) divided by the concentrations of, goldand silver, the quantity then multiplied by 1000.

207

208 THE CANADIAN MINERALOGIST

Frc. l Sample localities of the gold from this investiga-tion (from north to south): SM4 Manassas quadrangle,Virginia; JMB Storck quadrangle, Virginia; JMA Chan-cellorsville quadrangle, Virginia; BC Brush Creek, Vir-ginia; SM2 Cabamn County, North Carolina; SMIMontgomery County, North Carolina; NC85 Lilesvillequadrangle, North Carolina; SM3 White County,Georgia.

Grains of lode and placer gold are very rarely pure.They invariably contain significant amounts of sil-ver, and occasionally minor amounts of copper(Stumpfl & Clark 1965), palladium elanison &Fuller 1987), selenium (Dilabio et al. 1985), mer-cury (Nysten 1986), and 45 or more other elementsincluding: Fe, Pb, Pt, Bi, Cd, Ni, Te, As, Sb, Rh,and Al @rasmus et ol. 1987). Gold that containsmore than 2090 silver is generally referred to as elec-trum, but there is considerable variation in thisarbitrarily set compositional definition. In this study,"electrum" will be used only as a qualitative termto emphasize the fact that at least minor silver ispresent in the alloy. Craig & Rimstidt (1985) com-piled analytical data for nearly 6500 samples ofnatural gold, and found that virtually all the sam-ples fall in the range of 500 to 1000 fine, with themode at approximately 920 fine.

The present study examines the nature of placergold grains from some localities in the southeistern

United States and evaluates the processes that maycontribute to the formation of the gold-rich rimfound on many of these grains. Inasmuch as theprocesses active in this study area are representativeof those operating in most other localities, the con-clusions of this study are believed to have generalapplicability.

The samples examined in this study were obtainedfrom several localities in the southeastern UnitedStates (Fig. 1), an area with a moist, temperate cli-mate and oak-hickory-predominant forests. Exten-sive weathering has occurred in the area owing toprolonged subaerial exposure and a current annualprecipitation of 100-200 cm. These environmentalconditions subject the gold grains to substantialchemical attack, with local variations manifested indiffering water and gold chemistries, and the extentof human influences. All streams from which sam-ples for this study were collected have streambedsthat contain mud to pebble- or cobble-sized sedi-ment, and generally abundant organic matter (hencea relatively low Eh) in the finer sediment fractions.Gold grains from Brush Creek in southwest Virginiareceived the most attention.

CnanacreRrsrrcs oF THE GRATNS oF PLACER GoLD

The morphology of the grains is influenced bynumerous factors, including character ofthe original lode particles, stream energetics, nature of thestream channel material, time spent in the stream,distance of transport, and chemistry of the streamwater. More than 300 gold grains from Brush Creekwere closely examined to identify relationshipsbetween morphology and some of these parameters.The grains were divided into the following fivegeneral morphological gtoups: irregular (-4la/0,spherical to semispherical ( - 3390), wafer-shaped(width-to-thickness ratio s5) (-]4s/o), flake-shaped(width-to-thickness ratio >5) (-ll9o), and cylin-drical (-l9o). The grains recovered from BrushCreek during this study range in size from less than0.01 mm to 2 mm (maximum dimension); weightsrange from less than 0.01 to nearly 10 mg.

The surficial features of many of the gold grainswere investigated under reflected light and by scan-ning electron microscopy. Althouefr the surficial tex-tures represent a wide and probably continuous spec-trum, they have been broadly classified into the sixtypes shown and described in Figure 2. The generalrelationships between grain morphologies and sur-face textures are quantified in Figure 3. The mostprevalent correlation exists between irregular grainmorphologies and smooth, unpitted surfaces. Manyof the other relationships are not as universal, buta general diagonal trend of compatible morpholo-gies and surface textures can be seen from the upperleft to the lower right of Figure 3, where a good

COLD.RICH RIMS ON ELECTRUM IN PLACERS

Frc. 2. SEM phoromicrographs of the six most readily identifiable textures observed in this study. (a) Euhedral iso-metric electrum crystals on solid, fairly flat substrates of electrum. (b) Well-rounded, srnooth surfaces with littleif any evidence of chemical attack. (c) Smoothly rounded grains with pitted surfaces. (d) Smoothly rounded surfaceswith large pits containing branching-coral type features. (e) Hackly, friable-looking texture similar to that of foun-dry slag orweathered, gritty sandstone. (f) Inegular, lobate or bulbous, "stromatolite-like" texture with occasion-ally stepped but usually smooth individual lobe surfaces.

209

correlation exists between flake-shaped grains andlobate to hackly or grainy surface-textures.

An apparent relationship exists between grain size,

grain shape, and distance of transport for the goldgrains from Brush Creek. Near the headwaters ofthe creek, where the lode deposit is located, irregu-

" . { : ' *

\% of the mor-

\ Rholoslcal

\ sroup

"/o of \the surface \texture group \

GRAIN MORPHOLOGIES

l rregular( ' 4 1 o / o l

Spherlcal( - 3 3 o / o )

Cyllndrlcal( ' 1 o l o l

Waler( - 1 4 o / o l

Flake( ' 1 1 o / o l

aUIE3FxlJlFIIJC)ILtrfa

Srnootlunpitted( - 3 6 % )

7O-BOVI sufficlenl

Smmttpitted

( ' 3 2 o / o l

't 5-250/o

g0-400/"

60-7O"/o

Pitted withbranching-clral typefeatures( - 1 6 % ) 5 0 - 6 0 %

90'4Oo/o 1 5 - 2 5 o / o

Hacklyto

gralny( - 1 0 % ) ' l O-2Oo/" 3O-4Oo/o 4 5 '55o/"

l r regularlobate(' 7 "/"1 25-55% x

2t0 THE CANADIAN MINERALOGIST

Frc. 3, Correlation diagram indicating the predominance of relationships between morphologies of electrum grains andsurface textures, The value in the upper-right corner of each box indicates the percentage of grains (with the particu-lar morphology given at the top of the column), which possess the surface texture listed to the left. Similarly, thevalue in the lower left corner of each box indicates the percentage of grains (with the particular surface texture givento the left), which possess the morphology listed at the top of the column.

larly shaped grains up to I mm are not unconmon,and the grains average about 0.2 mm in maximumdimension. Downstream 19 km, the average size ofthe gold grains is less than 0.1 mm. Grain morphol-ogy evolves in a predictable trend from the originalirregular form through to semispherical and wafer-like, ultimately yielding flake shapes.

Compositional inhomogeneity within the placergold grains at Brush Creek has been known since asearly as 1882, when Fontaine noted: "The gold con-tains about 32 per ct. of silver. Most of it has a rich

gold color, and, judged by the color of the surface,would seem to contain over 90 per ct. of gold. Thiscolor, however, is confined to a thin, external film.When this is removed by the file or knife the coloris quite white. It would seem that the silver had beenremoved from the surface by solution in some natur-ally formed solvent, and the gold thus was concen-trated in the external film. When particles areobtained from the interior of quartz in the veins, theyoften show the white color."

Additional note of the rims on grains from Brush

GOLD-RICH RIMS ON ELECTRUM IN PLACERS 2ll

FIc. 4. Back-scattered electron images of two placer gold grains from Brush Creek, Virginia. (a) Grain number BC 2-6,showing partially complete formation of a gold-rich rim (briebtest, high-electron-yield areas), with preferential develop-ment in and around embayments and pits on the surface. (b) Grain number BC 2-15, showing a much more highlydeveloped gold-rich rim with many voids (obate texture).

Creek was made by Solberg & Craig (1981); we havesubsequently found that most of the grains of placergold from Brush Creek exhibit some degree of gold-rich rim development, as illustrated in Figure 4.Electron-probe microanalysis of 29 randomly chosenand representative grains from Brush Creek (Table1) indicates a compositional range of core electrumfrom 549 to 849 fine. This range is quite large rela-tive to those of other localities considered in thisstudy (Table 1), and other southeastern U.S. locali-ties (Craig et al. 1980). Microprobe analysis of thegold-rich rim on grains from all the sample locali-ties shows that 95Vo of the rims have compositionsof >985 fine. The remaining 590 of the rims rangebetween 967 and 985 fine (Table l). It should benoted that some of the optically distinct rims thatanalyze at finenesses lower than about 995 may actu-ally be purer, because the analytical volume of exci-tation may overlap or penetrate lower-fineness coreelectrum. When viewed in cross section, these rimsare not uniform in thickness, but vary from < I to60 pm. The boundary between the individual coresand rims generally is sharp, and is especially strik-ing in reflected light if the surface has been lightlyetched using a potassium cyanide plus ammoniumpersulfate solution. Grains with an incomplete rimtend to have gold-rich "pockets" wherein the enrich-ment areas occur in small embayments on the grainsurfaces (Fig. 4a). In other cases, lhe rim is so fullypervasive that it completely encloses the core (Fig.4b). These pervasive rims commonly exhibit a "swisscheese" texture, in which numerous small roundedvoids are found in the gold-rich rim.

A few distinct relationships were observed betweenthe character and prevalence of gold-rich rims and

the morphologies of the gold grains. Firstly, polishedsections of gold in vein quartz show no signs ofdevelopment of a gold-rich rim @riscoll et al. 1990),implying that the rim forms after liberation from thehost rock. Furthermore, irregularly shaped grains ofplacer gold (relatively new to the stream environ-menl) exhibit little or no development of a gold-richrim. Flat, well-worn, flake-shaped grains, on theother hand, exhibit the most extensive developmentof a gold-rich rim. Spherical, semispherical, andwafer-shaped grain morphologies show intermedi-ate degrees of rim development.

The surface features observed by scanning elec-tron microscopy also show a general relationship torim development. There is an increasing progressionin the degree of rim development corresponding tothe following respective surface textures: smooth,smooth pitted, pitted with branching features, irregu-lar hackly, and lobate (Figs. 2b-0. The "euhedralcrystals on gold substrate" texture was observed onone grain only, #SM105364; these crystals and theirunderlying substrate both have fineness valuesbetween 807 and 820. This compositionalhomogeneity suggests that the crystals are most likelyresidual lode-gold crystals, as all indications suggestthat gold remobilized in the weathering environmentin this study area results in virfually pure gold precipi-tates.

Consideration of these observations permits therecognition of a sequential trend @ig. 5) that relatesformation of the gold-rich rim to the morphology,surface texture, and distance of transport of theplacer gold grain. Most of the observed morpholo-gies and surface textures can be roughly correlatedinto this trend of increasing development of a rim

2t2 THE CANADIAN MINERALOGIST

GrainNumber(l) (mn)O) (ng)

lilass shape€) FinenessCore(4) Rim(4)

7n:r298.' 98995(3)5rG596(4) -re90(4)6550) ryr0).4u4r4(4t 9r&9r9e)iT?-JtgQ) .4y2-9CaQ'r0--yr8(3) .sn 0)e-"667(3.) 9fi-r88e)qr7rc@, 9fl!9rrc2)s53 (l) 996 (l)6n:6[/.c4 9y2(l)55&562@) 90(l)6500) sr-.st 8)6t3-6,/.@.) 9r5(l)930) 9769nC2'66t 0) 9eL%9 C2)59r (l) s? o)5490) 9rr(r)e,5{3,9Q) -9rr94(3)6E5(l) 988(l)54(l) e9l (1)55s0) 99l (l)56ao) 9rro)78r 0) e5(l)lt4 (l) 993 0)6150) W (r)6010) I't (r)7130) 4)4-gt Q)t49(l) gel (r)n6|J, 9890)n4(r) 9n$,94r@) 930)-953-955(2) ND.(o9v,-,958Q) ND.866 0) 995 O)8o(l) w2(r)994 (l) ND.e45O) ee(l)E69O) 991_99r(2)966 (l) ND.sl 0) #-s7(3,cn-&ge) ND.e6l o) 998 o)9U-943 @) 99(r'9D@' 980)9n 8' r{J.e4? 0) 96 O)947 (r) l{D.83c$oe) ND.957-959(3) ND%1.9!9|.J) ND.tt44-946cj) 965 (l)ctog,ro, 990)p$uz,(4' -!fi-9 '.8t3)e390) s7--99t2tlme) un(l)r6!8drcj) 9rr-996€)961@) 9t2-lmg3)m9tr2(3) ND.uG8r8€) NJr.&r(t) ,.qc2-93@)rr{20(o ND.

Cnim witi "Jlll" prefix wcrc gaerously dcccd by Mr. L R.I\dcClord of Rpderlc&shr8, VA" (hsfursrrudngwnh.Slf rnml@nfi@ thcunited Srd€sNstiorll uamof Nsnml lilsrysnd

-. arc identified with LANMNH ^Fnpb mnbcrr follosirg tbo ,SI\ll.Q) vatu aasnined by nm of ofr'cal c*innb.t', $apeabhuddm: S<t'raqV'*dcr,I-trogulat,Ff,rlo,R-lod,

CEcsysElline, TE{hin ph!c.

TABLE 1. CIIEIdICAL AND PEYSICAL DATACONCERNING PLACER, ELPSTR,UM GRAINS

with time spent in the stream. The above relation-ship suggests that the physical characteristics ofplacer grains can give clues about the potentialpresence and nature of a gold-rich rim without goingthrough lengthy sample preparations.

The contact between the gold-rich rim and theIower-fineness core in the grains from Brush Creekis very sharp and well defined, facilitating immedi-ate recognition of the rim, even where it may be thinand irregular (Fig. 4a). The rim-core boundary ingrains from most other localities is equally sharp,though generally less apparent oq/ing to the higheraverage fineness of their core electrum. The composi-tional characteristics of two representative rims asdetermined by electron-microprobe traverses areshown in Figure 6. These iraverses were especiallyuseful in the quantification of the steep composi-tional gradients in such rimmed grains.

The compositions of coexisting rim-core pairs inthis study, as well as from other sources (data on681 grains from 32 localities) are presented in Figure7. The average core composition is roughly equiva-lent to that given by Craig & Rimstidt (1985) for goldgrains from all over the world. This similarity, andthe fact that all of these grains have a gold-rich rim,suggest that there is no strong relationship betweenthe composition of the core and the potential for rimdevelopment. The plotted rim-core compositionalpairs from Brush Creek (solid circles in Fig. 7) illus-trate the unusual composition of this gold. The scat-ter of rim compositions shown in Figure 7 also indi-cates that there is no real relationship between thepurity of the gold-rich rim and the composition ofthe underlying electrum core.

Mineral inclusions observed in the gold grainsfrom Brush Creek are primarily present in the rim.The occasional mineral inclusions found within coreelectrum tend to be considerably larger than thosein the rims. The majority of inclusions, both in therims and in the cores, appear to be complex silicates(e.9.; amphiboles and pyroxenes), as they usuallycontain substantial amounts of Si, Al, Fe, and Ti+ subordinate K and Ca, as determined by energy-dispersion X-ray (EDX) analysis. Roughly 20Vo ofthe inclusions in the rim are believed to be iron oxides

(4) iatrsinpareorbeseshdhatcthcnrmbsd@lytcsmdotodenerninetlrvalue crogc sboun, od'-' irrlicrlisvrhlt rtbktluve malyical onr dcriating fro lm ry >4%.

(0 gC"-groi* = plaw gold, Bnrsh Creek, VA.(O \.lt'=notmeund.o) te-g*it"- paleplrcgold,Lilesvillcqd.,NC.(E) 'X.O." - nc rtacaed.(9) 'nr{.l'-gtatr - plag gol4 CheeIwtlb qurd-, vA.OA .fMg"-gaiu - placer gold, Sbrct qud., VA.(ll) Sft,t C-126 = Flag gold (?) frm urloom fotity b NC.(u) S[\,t 6?994 - gpld fro rbc Etdmalo Mno, ftmgmcry Ca, NC thrr

ls well wom ed rorded oggscing oeu tmspon in iu hnnry.(13) SM 969S2 - plag gold, Silvq Cree&, Cabqnrc Co. NC04) sM r/398 (l-3) = vom systals, Dugr Gold ltf,r, whib ci, GA-05) $U 105364 - plEc.r gol4 Bull Rrn, Fairfrx Co, VA.

BCI- l ( t 0. r8BC l-2 0r5BCI-3 030BC?r

0.04 s0.06 wo m I4.09 S9.n w4J3 II2L FNrr. (6) wNrt sNJIA FNIt|. IN}L WN t . sNIrt FNiA INJ'{. ItNt{. sNn'[ FNJ\A INJ'l. WNIA SNLL FNtd. tNJtrL IN t . w0.E4 S026 W0.65 F0.m INJ,t FNiA FNJA INtr. sN.T'. FNJr[. SNt[. wNJ't, IN t . sNIA FNta sNI,t. INI[. INJrr. SNra sNra sNra wN}', INi,. I&58 R19.90 Rn5! Irt39 I2I2L It4.92 rr88.45 S231.68 W&25 r0.9J C1.96 C3,& C352 T

BC2-2 lr0BCz-3 1.30BC24 1.40BCts 0.65BC2-6 050BCz-? 0.95BC2-t 0.60BC2-9 055BC2-10 05)BC2-ll 055BC2-r2 0.60BCz-r3 055BC2-14 055BC2-15 o.srBC2-16 0.60BC2-r7 050BC2-r8 0.?0BC2-19 0.A)BCz-z, 0.65BCz-26 0.60BCz-n o.{tBC 1l 0.?0BC&l 050BCtl 0.90BC7-l 020NC85L86I(7) 1.r5NCE5IJ62 1.45NC85L8G3 LloNC85L86.4 0J0NC85L86-5 lr5NCt5L86J6 0.70NC85L8G7 LzANC85L868 0.90NCEsLE&9 t.mNC85L&ir0 LlsNC85L86ll 0.65NCE5L86I2 0:ISNC85r.J6l3 0.85NC85L8GI4 0.65NC85IJ&15 0.?0NC85L86I6 0,55NC85L8G17 0,65NCttLEGlE 0.60NC'&ilJ619 0.45JMA I (9) X35JMA2 435JMA3 3.10JMA4 5.05JMB l o0) 3.15JMB 5 1.95s[\4 c-126 0l) 3.95s[\,r 67994 02) 5.45s\{96982o3) 82A$vt 939&r (la) O.tOs[\49739&2 255sMt39&3 3.q)sN,t 105364(15) 0.75

,N.REA',NG r;;;#;#;;ff;^RY scALE)213

rL,b;r*.;jfIDz#3#a

Frc. 5. Schematic illustration depicting the chemical and mechanical alterations experienced by grains of placer gold

as they are transported downitream. The figure in each circle is representative of a sample's appearance in polished

section. The foliowing characteristics are typical for each grain's particular stage of development: (1) unrimmedgold grain in mineraliied quartz vein, (2) ir;;gdarly shaped grain of placer gold with little or no evidence of gold-

ich rimming, (3) spherical to wafer-shaped grain of placer gold vrith spotty to intense development of gold-rich

rimming, (4) flake-shaped grain of placer gold with thick, very well-developed gold-rich rim.

Gold contonl

SllYor contonl

0 5 1 0 1 5 2 0DISTANCE lllTo GRAIN (pm)

L----=,".**Sl|vsrconbm

1m0 tm0

8&t

a3frzl|l

54e A313

tr860tc!Tr9 a 0=okn

80Fz

EHdrg f i

2 t &EB

km0

89,

79

549

313

J 025

J 02A0 5 1 0 1 5

DISTANCE INTO GRAIN (pm)

Fro. 6. Illustration of the compositions of placer gold grains. (a) Back-scattered electron (BSE) image of a portion of

the polished section of grain number BC 5-l along with a plot of the change in major-element chemistry determinedby;lectron-microprobe analysis along the traverse path indicated by the black-on-white horizontal line on the BSEimage. O) BSE image of a portion of the polished section of grain number BC 2-7 along with a plot of the changein riajor-element c[emistry determined by electron-microprobe analysis along the traverse path indicated by the

black-on-white horizontal line on the BSE image.

g v r Y - a r r v r . \

M{€$#ewElectrum /

\ core- l

214 THE CANADIAN MINERALOGIST

E a i" :i l"?"fl:"'*EEl i " t - t " t " . . . "? :

E

l"q'"Ii ' ' f - o o g sI o F i li o o a '

- q s

E dE t t

o

E a&"i:E o B o " . i

o E o oo !

c o

o ; ( b o - c

o

oo

c

g E

o

c

o

cc

Og

S o oo o o

o $ o o o

c o o o oC O

c $

o o o o

c o

s O

o 3 o o

o

O o

o

Llne ot equal rlmand core tlnenoss

o clust&smtthto8fE Daboror{hlgm. Thlgstudy, Brus| C-r€okdaiaA Thlssludy,other localltls

g, gDt ,o5mg

t l

EEsso

940

7i80r-500 050 700 750 800 850 900 950 1000

Gore FlnenessFlc. 7. Scatter plot of rim versw core compositions of placer electrum grains with gold-rich rims from Brush Creek

(filled circles), other localities in this study (filled triangles), Desborough (1970) (open squares), and Giusti & Smith(1984) (open circles).

and hydroxides, as iron is the only major elementindicated by EDX analysis. Placer gold grains fromthe other localities commonly have more mineralinclusions than at Brush Creek, but usually showsimilar inclusion mineralogy and size distributionswith respect to rim and core. Inclusions in the coreelectrum of all the grains from this study seem tobe primary, relict phases, as described by Driscollet al, (1990), whereas those in the rim probably havebeen injected into the ductile, high-purity gold bymechanical working, or have been incorporated dur-ing a process of chemical growth.

D$CUSSIoN

Introduction

Several explanations have been proposed toaccount for gold-rich rimming, which gives rise togold grains in placer deposits having higher averagefineness than the grains in their parent lode deposits.The three principal models of rim formation consi-dered here are: (l) preferential dissolution of silverfrom placer grains during weathering and transpon,(2) precipitation of gold onto placer grains from

oxidizing, Au-bearing stream water as it encountersmore reducing conditions (cementation), and (3) aself-electrorefining process where the electrum at thegrain-solution interface dissolves, and the goldimmediately precipitates back onto the surface of theplacer grain.

Preferential dissolution of silver

The most commonly invoked mechanism for thedevelopment of gold-rich rims is a process ofpreferential dissolution of silver from the Au-Agalloy. The relative solubilities of Au and Ag in low-temperature solutions are discussed by Mann (1983,1984) and Xue & Osseo-Asare (1985). These studiesindicate that a model based on the preferential dis-solution of silver is chemically reasonable; however,such a model fails to explain how silver atoms fromdeeper than the outer few Engstrdms ofthe grain cancome into contact with the solution. One of the fewresearchers to address this problem was Desborough(1970), who suggested: "Oxidation of silver from themetal to the ion reduces its size and affords theneeded mobility for removal from its site in the alloy,and it is thus placed in solution. The porosity result-

GOLD-RICH RIMS ON ELECTRUM IN PLACERS 2t5

ing from continued silver removal provides a newinterface in the gold-silver alloy for a repetition ofthe process." This mechanism, however, is proba-bly ineffectual, as Fontana (1986) commented on ananalogous system wherein the "dezincification" ofbrass results in a copper-rich coating: "A strongargument against [zinc being dissolved, leavingvacant sites in the brass lattice structurel is that dezin-cification to appreciable depths would be impossible or extremely slow because of difficulty of diffu-sion of solution and ions through a labyrinth of smallvacant sites". In addition, during the commercialrefining of electrum, sufficient porosity for selectiveleaching of silver from electrum is only developedthrough a process known as "inquartation". Inquar-tation involves the dilution of molten natural elec-trum with two to four times its weight of silver. Theneed for inquartation is explained by Bowdish (1983):"When alloyed with less than about three times itsown weight of silver, gold will also protect the sil-ver from attack by nitric acid solutions." Placer goldwith a core fureness as hich as 972 has been observed todevelop a distinct rim of 998 fineness (Table l, Fig.7); such a small difference in silver content wouldafford very little increased porosity along which aleaching solution could encroach. Finally, grains withhigh silver contents in the core electrum do not ingeneral develop a thicker or more prolific rim.

Another mechanism for the removal of silver fromthe core of an electrum grain is that of silver diffu-sion from the interior of the grain to the surface,where it can dissolve. This was tested by modelingdiffusion profiles formed at 25oC in grains of thefollowing compositions: Au66Aga6, Au-rrAg25 andAuegAgle (equivalent to finenesses of 733,846 and

943, respectively). These calculations are based ona model of non-steady-state diffusion in a semi-infinite medium presented in Darken & Gurry (1953,pp. 441-445). In accordance with this model's req-uisite boundary-conditions, the instantaneous andconstant concentration of silver at the surface of anelectrum grain is taken as zero, and the initial con-centration at the semi-infinite distance remains fixedat the values stated above. The tracer diffusioncoefficients of silver in Au66Auoa, AurrAu2r andAueeAul6 al 25"C were extrapolated from the dataof Malaro et al. (1963) as 1.36 x 10-32, 3.39 x10-32 and 1.05 x 10-31 cm2ls, respectively. The morerecent diffusion-coefficient data of Cook & Hilliard(1969) were not used even though they may beslightly more accurate, as they do not change ourconclusions, and also cannot be easily adjusted toany desired composition of electrum, as can thoseof Mallard et al. (1963).

Diffusion profiles (Fig. 8) calculated for reasona-ble periods of geologic time show that silver diffusesmuch too slowly in such dlloys tq deplete silver frommore than about the outer 2-3 A of the gold grain.In fact, the number of years necessary to producea gold rim roughly equal in thickness to natural rims(-8 pm) (Fig. 8b) is between 1017 and l0r8 years(about one million times the age of the universe).Furthermore, this mechanism is incapable of yield-ing the extremely sharp rim-core contact observedin natural grains.

Thus, the only conceivable means by which selec-tive silver dissolution could produce a significantgold-rich rim would involve repeated cycles of0ngstrdmJevel leaching of silver followed bymechanical deformation (leading to re-exposure of

a. eoS' s5

733

772

8 1 0

846

880

9 1 2

943

733

772

8 1 0

846

880

9 1 2

943

972

1 000

oE 3 0og 2 s

i 2 0H 1 sz8 r 0r D 5

0

at@l!zIIJ

=IL

b.sg

AEf rgG==Fr r i

oI

972 rD

I O O O0 4 I 1 2 1 6

DISTANCE INTO GRAIN (A)0 4 I 1 2 1 6

DISTANCE lNTo GRAIN (Pm)

Frc. 8. Observed natural gradients of silver and calculated diffusion profiles of silver in gold. (a) Calculateddiffusion profiles of silver in Au75Ag25 for the geologically reasonable time periods of l, 10, 50, 300, 1@0, and3000 Ma Oeavy black lines 1-6, respectively) and the diffusion profiles of silver in Au60.Aga0 and Au96Ag1s after300 Ma (fine lines 7 and 8, respectively). (b) Observed natural gradient in silver concentration in grain number BC5-1, as determined by electron-microprobe analyses along a traverse (heaviest black line, #l); calculated diffusion-profiles of silver in Au75Ag25 for time periods of 1014, 1015, 1016, 1017, and 1018 years, necessary to simulate thenatural gradient in Ag conceritration (heavy black lines 2-6, respectively); and, for comparison of the effect of com-position, the diffusion profiles of silver in Au66Agaa and Aue6Agls after 1016 years (fine lines 7 and 8, respectively).

t2

216 THE CANADIAN MINERALOGIST

TABLE 2. DISSOLUTION REACNIONS FOR METALLICGOLD

LIGAND AG'q316yO) REACTION

TRIVALENT GOLD COMPLEIGS

NOLIGAI{I}S

Au(s) + 0.7502 a ltt+ = ds3+ a t.lil{p

ONEIJGAI\D

Au(s) + 0.7502 1!tI* a Q[- = trgQl2f + l.5H2OAu(s) + 0.7502 + 2H+ = tuOF2+ + 05H2O

Au(s) + 0.7502 + 3H+ + 3Cl- = AuCtg(rq) + l.5H2OAu(s) +0.7502 +2H*+2Cl- = A@HCl2(sd +0.tH2OAu(s) +0.7502+H+ +05H2O+FA2- = A[(O[I)ZFA-Au(s) +0.7502 +H+ +05HZO +Cl- = Au(OHhgl(aq)Au(s) + 0.?5OZ + l5H2O = Ar(OI{h(sd

trIOURLIGAIYI}S

Au(s) +0.7502 +3H++4CN-= Au(CX\O4- + l.5H2OAu(s) + 0.7502 + 3H* + 4NH3 = Ag(NH3)a3+ .r 1.511rpAu(s) + 0.75O2 + 3H* + 4I- = AuI4- + l5H2OAu(s) +0.7502 +3H*+,$CN-=Au(SCN)4- + l5H2OAu(s) + 0.?502 + 3H+ + 4Br = AtEr4- + l5H2OAu(!) + 0.7502 + 3H+ + 4Cl- = AuCt4- + 15H2OAu(s) +0.?5OZ +2H* +,cl- = AuOHC!- + 05H2OAu(s) +0.7502 +H+ +0-5H20 +2Br =Au(OI!ZBr2-Au(s) + 0.7502 + H+ + 05H20 + 2Cl- = Au(OH)zClz-Au(s) + 0.7502 + l5H2O + Cl- = Au(OE)rCl-Au(s) + 0.7502 +25H2O = Au(Qt{)a- a t{+Au(s) + 0.75C)2 + 15H2O + Br- = Au(OH)gBr

FTYELIGAI\DS

Au(s) + 0.7502 + 3H+ + 5SCN- = Au(SCII)S2- + ISHZOAu(s) +0.7502 +35H2O = Au(OI{)*'+Ztt+

SIxLIGAT{I}S

Au(s) + 0.7502 + 3H* + 6SCN- = Au(SChf63- a 151119Au(s) +0.7502 +45H2O = A{0I{)63- +3H+

logK Sc(2) Ref.(3)

N.A"(4) 433.5 -11.47 NJ|" FL

ct-oH-

239.3184.9

{.45 u.E3-9.{I NJ{"

B LB L

cl-SO42-oH-

TWOI,IGAIII'S

1OJ Au(s) + 0.7502 + 3H+ + 2Cl- = AuCl2+ + l.5H2O-1@0.4 Au(s) + 0.7502 + 3H+ + 2SOa2- = Au(SOa)z- + l.5Hz0{3.6 Au(s) +0J5O2 +H+ +05H20 = Au(OtDz+

TEREELIGAI.IDS

6.10 Lg BL-537 8.38 MFL:1.{I N.A" BL

ct-oH-,cl-oH-,FA2-oH-,cl-oH-

-83.3-148.4*-(t-2t7.3-2C4.4

10.08 0.43 BL\95 t.22 AL1.33 -1.94 KBL:354 L93 AL-10.51 N-A L

cN- t7L4NH3 -11.92r -45.0scN- 561.6Br -161.3ct- -23s.1oH-, cl- -293.6OH-,Br -333.7oH-, cl- -349.7

oH-,cl- ,403.7oH- 455.4OH-. Br- -251.9

@.53 -t4.54 JLq.98 -9.15 rL36.2t 42t HL31.05 4.92 L20.92 -238 L'3,A {5E L5.38 0.m AL3.42 -2.4 cL-335 1.36 AL-1L42 5.81 AL-21.91 N.A" L-34.23 n.6 cL

scN-oH-

654.5{16.5

31.0t -3.93

a5.24 NA"LL

scN-oH-

747.O:163.2

31.05 a28-51.08 N.A"

L

L

GOLD.RICH RIMS ON ELECTRUM IN PLACERS 2I7

TABLE 2(contrd). DISSOLUTION REACTIONS FOR METALLICGOLD

LIGAND Acof(cnx)(r) REACTION

MONOVALEM GOLD COMPLEXES

logK SC(2) Ref.(3)

N/4"(4)

NOIIGAIIIIS

163.6 Au(s) +025021t{* = 6rr* 4-05H2O

ONEIIGAI{D

21,6 Au(s) + 02502+ [IS- = AuS- + 05H2O35.6 Au(s) + 02502 + H+ + IIS- = AuIIS(aq) + 0.5Hp-3933 Au(s) + 025021 l{* 1 $Q32- = AuSq- + 05Hp418.4 Au(s) + 0.2502 + H+ + SZOI2- = AuSZO:- + 05HZO5.0 Au(s) + 02502 + H+ + Cl- = ArC\eq) + 05H2O-54.4 .tr(s) +0.2502+ 0.5H2O = Au(OHXss)

2L.O 2Au(s) + 0502 + H* + 3HS- = AuZGIS}9- + HZO

116.8 2AuG) + 0JO2+ 2HS- = Au2S22- +H2O

:1.t7 N.A"

S2-HS-sq2-SZOI2-cl-oH-

t9.82 -2L22 MELn.n -l9:n GL5.16 :156 BL3.25 -5.65 BL-238 4.02 BL-10.52 N.A" BL

TWOLIGANIIS

cN- 2S5.8 Au(s) + 0.2502 + H* + 2Cl{- = Ar(CN)2- + 0.5H2O 31.82 -17.1 I L

CS(NH/2 -----(6) Au(s)+02502+H++2CS(NHdz=AulCS(NHz)elz+ +OSHZO Tl-91 -15.16 HL

HS- LSJ Au(s) + 02502+ H* +2IIS- = Au(HS)2- + 05H2O 22'n -1268 GL

Sq2- -9623 Au(s) +0.2502+ H+ +2SOg2-= Au(SOrlf- +0.5H2O 19.61 -ll.0l BL

SzQ2- -1030.2 Ar(s) + 0.2502 + r$ + 2SZOr2- = Au(S2O3)i3- + 05H2O 18.90 -10.65 J L

NH3 13 Au(s)+0.25O2+H*+2NH3=Au(NH3b+1g5H2O ll.yl :1.19 IL

I- 47.6 Au(s) + 0.2502+ H* + 2I- = AuIt +0.5H2O 1176 -7'08 HL

SCN- 251.9 Au(s) +0.25O2+H++2SCN-= An(SCND- +05H2O 9'E4 4'12 L

Br -115.0 Au(s) +0.2502+ H+ + 2Br= AuBr2- +O.5H2O 5'21 -3'81 L

Cl- -151.1 Au(s) + 0.2502 + H* + 2Cl- = AuClt + 05H2O l'9E -2'19 L

OH-, Br -lgg.g Au(s) + 0.2502 + 0.5H2O + Br = AuOlIBr -3'25 -5'15 DL

OH-, Cl- -2152 Au(s) + 0.2502+ 0.5H2O + Cl- = AuOHCI- -5'35 -3'05 DL

OH- -275.4 Au(s) + 0.5O2 + 1.5H2O = Au(OH)Z- + H* -13'34 N'A" DL

DI4OIJ)OOMPLEXES

HS-. 52-52-

45.& -15.72 MEL2I.35 -17.13 GL

(2) 5"=5fimgrhofcomple,x=thelogarirbmoftheactivityoftheligmdinequili-baiumwiththecorespmdingaqueous gold-complex solutim (P(O/ = 0.2 atm. pH = 6), yielding 0.1-ppb of gold'

(3) References: A - Baes & Mesmer 1974 B - Baranova & Ryzhenko 1981, C - Bjemm 19/1, D'IUPAC 1982'

E-Johnsonetal. 1978,F-Irtimer1952,c-Renders&Seward 1989,H-Scbmid 1985, I-SkibsFd&

Bjernrm 1974,J-*ibsted&Bjemrm l9n,K-"VarihaletaL l9t4,L-WagmanetaL 1982, M-Webster 1986.(4) N.a" = not applicable.(5) FA = Fulvic acid - stoichiometry unlno*r, therefore AGgundeteminable.(6) No AGf availablefcthe coinplexorfree ligand.

fresh electrum). This mechanism is, however, in no Cementqtionway supported by the textures of the observed sur-faces. Despite its chemical inertness, gold is transported

2t8 THE CANADIAN MINERALOGIST

in significant quantities as aqueous complexes at hightemperatures to form hydrothermal deposits. At lowtemperatures, gold is much less soluble, yet can becomplexed by a host of ligands (Table 2). Depend-ing on the concentrations of such ligands, andambient conditions of Eh and pH, several of theligands appe€u to be capable oftransporting signifi-cant amounts of gold in the supergene environment.The following observations have been offered as evi-dence for low-temperature aqueous transport ofgold: (l) placer gold grains with smooth, well-formedcrystal faces (Warren 1982, Crug&Callahan 1990),(2) high-purity gold crystals and dendrites on ironoxides in laterites (Mann 1984, Freyssinel et al. L9g9),(3) gold impregnations of originally carbonaceousmatter (Wilson 1984), and (4) secondarily enriched

concentrations of gold associated with recentlydeposited iron hydroxides (Lesure 1971).

At least two detailed accounts suggest that therehas been secondary enrichment of gold in depositsin the southeastern United States. Lesure (1971)reported on secondary enrichment at the Calhounmine in Lumpkin County, Georgia, where signifi-cant concentrations ofgold (2.9 ppm) occur in freshlimonite coatings of the walls of an adit. Cook (1987)compiled a series of historical accounts of gold minesof Georgia and Alabama, which document largeaccumulations of coarse gold in quartz veins just atthe water table. These same veins commonly showa significant decrease in particle size and grade ofgold below the water table.

The calculated equilibrium solubility of gold in

1 , 9 7 0

1e7 A

15.7

1 . 9 7

0 . 1 9 7 =

0 . 0 1 9 7

1 97 ,000

I 9 ,700

1 , 9 7 0 3

1 . 9 7

0 . 1 9 7

0 . 0 1 9 7

B -reo

'14$

trIggo(,0C)

-10.0

-11.0

-12n

a.

=goE

tro6

troC'coo3

tr-9a6

tro3'coo=,(

197

19.7

€

;l

€

-9

-10

.11

-12

-13

b.

ggoEcI6

co()E0C)

:'

ct)o

ptl

--t-tr]b

""""ubTAu(oH);

/Au(OH)n

Au(OH)r'

SEAWATER

nns

rtn S - ' lb 6 N -o o o ;

i .- .-ii,, j'j I i-.'il : 'i.';

F ON Oo o

MINE DRAINAGE

t

t:

@€

:

-!lt' 'l

i,.3

otso

a o

o o

.14

SURFACE WATER

t [ . ;t o

" l :9' l [ i

l=.' lrli . '

ll,l l't'i-ilfflilI

. N o o Fo l a @ F @o l o o o o

@@o

o a@ N@ o

GROTJNDWATER

l.I r o d' - - . 1

ittil .," ""';ti

r r t

@ N F oo o t s oo o o o

GOLD-RICH RIMS ON ELECTRUM IN PLACERS 2t9

1 ,970 3

l sz E|E

le .z Eo(,

1 . 9 7 CoC'

0 . 1 9 7 5

0 . 0 1 9 7

-8

-7

€

-9

-10

-11

-'12

.13

c .

=6

6Ecg6

cootroo:t

EDo

1 97 ,000

1 9 , 7 0 0

Frc. 9. Gold concentration plots. (a) Log Au concentration (molality and ppt) verszs pH for gold-hydroxide complexes

and free gold ions twhere ]D(dj : -O.Zl

in the absence of other ligands. The predominant univalent gold species

(indicated by the upper t*o UoiO contreiting straight-line segments) fro_g.l-ow to high pH are Aur- and Au(OH);,

with the concentration of the neutral AuOA' complex constant a1 lQ-r1'426. The predominant trivalent gold spe-

cies (indicated by the lower four, bold, connecting, straightJine segments) from low to high pH include: Au(OH)2+,

Au(OH)i, eu(Oila, and Au(OHft. Plotted circles represent measured natural gold concentrations of stream waters'

wtrereas iriangtes represent grou.tdiuaters, and squares iepresent mine waters; open symbols are for-data Yt-tich included

specific.on.Jrrtrutioo-pffiairs, whereas solid symboli and their associated range bars are for data which specified

only t*g"r of pH and ioncentration. (b) Log Au concentration (molality and ppt) versus theyear of d€termination

for seawiter and mine drainage; dates'along bottom axis indicate the first, last, and several intermediate dates for

the corresponding Au-concentration determinations. Symbols used to plot concentration data are as follows: small

open circle: single concentration value, Iarge solid circle: average concentration'.lalue for.a.number of analyses'

xj fimit of deteclion, 1("x" with solid lower quadrant): sample(s) with no gold detected, X ("x" with solid upperquadrant): sample(si $ith a trace of gold detected, solid bar: range of reported gold concenlrations,-dashed bar:

.'*g. *ittt no lowei'limit specified, numerals above bars, open circles, and sotid circles: number of individual ana-

lysei within the set. Each set of concentrations generally represents analyses from a particular locality, except for

several of the seawater data sets, and some gfouped sets of individual analyses. (c) Log Au concentration (molality

and ppt) versas the year of determination for suiface water and groundwater. Conventions in part (c) are the same

as foipart O). Reported values of gold concentration have declined over time, probably as aresult of improved

analyd;al te;hniques and the greateicare taken in preconcentration steps to fully filter the solutions to remove fine

particulate mattei and colloidil gold. It should be noted that virtually all of the measuremeuts shown on this figure

were made on samples from regiins where gold deposits occur, or where gold was being explored for and expected

to occur (except ofcou.se, for ieawater samples). Data for Figures 9a-c were taken from the following sources and

tabulations within them: Borovitskiy e/ a/. 1966, Brooks et al. 1981, Chernyayev et al. 1969, Crocket 1978' Goleva

et al. 1970, Gosling et al. lg7l,Hallet al. 1986, Hamilton et al. 1983, Jones 1970, Kaspar et al. l91,2,,Koide et

a/. 1988. Lomonosov et d/. t985, McHugh 1988, Turekian 1968, Turner & Ikramuddin 1982, Volkov & Shakhbazova

t975.

therefore, may or may not be important in account-ing for the total dissolved contents of gold in naturalsystems, depending upon the relative influences andcontributions of the other gold-complexing ligands.

Table 2 lists the transporting efficacy of a widevariety of naturally occurring gold-complexingligands. The complexing strength (Ss) of the vari-ous gold complexes is defined as the logarithm ofthe activity of the ligand in equilibrium with metal-lic gold and a corresponding gold-complexed solu-tion [pH : 6.0, P(Ot : 0.2 atm.] vielding 0.1 ppbgold. Note that an 56 value refers to the concentra-

pure water as a function of pH is shown by the boldlines in Figure 9a, constructed for P(O) = 0.2 atm.At the near-neutral pH values found in most streamwaters, the equilibrium concentration of gold ionsand hydroxy complexes would be expected to rangefrom -0.04 ppb at pH 4.5 to - 12 ppb at pH 7.0.These values are considerably higher than the onesreported in the literatwe for various fypes of naturalsolutions (plotted points and ranges in Fig. 9), prob-ably as a result of slow kinetics of dissolution or,more likely, insufficient interaction of solution andgold-bearing sediment. Gold-hydroxy complexes,

220 THE CANADIAN MINERALOCIST

tion of a free ligand remaining after gold complexa-tion; it does not refer to the total concentration ofthe ligand. The smaller the value of Ss, therefore,the more effective the ligand is at complexing gold.A concentration of dissolved gold of 0.1 ppb waschosen for the Sq parameter, as it represents themiddle to upper range of recent determinations ofthe dissolved gold content ofvarious natural environ-ments (Figs. 9a-c). This concentration is assumedto be a reasonable estimate of the minimum valueat which sufficient transport occurs to yield theproducts commonly ascribed to secondary enrich-ment of gold (Lesure l97l).

The value of Sq can thus be used as a .,yard-stick" to eliminate from consideration all comDlex-ing agents that could in no way be capable of produc-ing significant supergene transport of gold. Ligandswith values of Sq greater than -1.65 (: 100 x theaverage concentration ofthe most abundant naturalligand, chlorine) are eliminated from further cou-sideration because such ligand concentrations are vir-tually never observed in typical, fresh surface watersor shallow groundwaters, where gold grains withgold-rich rims are frequently found. Gold-hydroxycomplexes were chosen for further considerationif they are capable of yielding concentrations greaterthan I x L}-tt m (1.97 ppt) at normal values of pHfor streams, since the 56 parameter does not applyto these complexes. Because solutions in theseenvironments are normally quite dilute, all activitycoefficients were assumed to equal one. Several ofthe ligands listed in Table 2 are metastable in solu-tions with an Eh buffered by equilibrium withP(O, = 0.2 atm, but since their metastable persis-tence is not quantitatively known, they have beenassumed to be completely stable for the ,,S"', con-sideration. The choices ofpH and Eh are consistentwith values commonly found in fresh surface waters@aas Becking et al,1960). As can be seen from Table2, 28 of the gold complexes have 56 values thatindicate they are at least strong enough to warranta more thorough evaluation. The only ther-modynamic data available for complex organicligands that might dissolve gold pertain to fulvic acid(Table 2), and although the concept ofgold dissolu-tion by micro-organisms and complex organicmolecules has been substantiated by various sources(Lakin et al. 1974, Mineyev 1976), sparse ther-modynamic data prevent a further quantitative evalu-ation of them. Colloidal and very fine particulategold in natural waters also is possible, but tlte resulrsof workers who have considered this problem(Golevaet a|.1970, Gosling et al. 1971, Lomonosovet al. 1985) are not in agreement over their impor-tance. Clearly, there is a great need for further workin this field.

Of the ligands that can form strong Au complexes,only a few occur at sufficient concentrations in

natural waters to produce significant transport ofgold. The average concentration of each ligand form-ing a complex with an S" value from Table 2 of-1.65 or more was estimated for stream waters andslow-moving sediment solutions from literature data.These data were then used in combination with theequilibrium constants given in Table 2 to calculatethe expected concentrations of the gold complexesin equilibrium with elemental gold (Table 3). Theconcentrations of gold complexes given in Table 3were calculated using both mass-action and mass-balance considerations. The pH value of 6.0 used incalculations for both stream water and sediment solu-tion (Table 3) was estimated from the extensive com-pilation of field measurements by Baas Becking elol. (1960). The activity of dissolved oxygen used forthe stream water in calculations in Table 3 is thatfor air-water equilibrium. The value of a(O) inorganic-matter-rich sediment solutions was estimatedto be 2.51 x 10-6e (Barnes 1979), by assuming thatthe solution exists at the H2S-SO?- boundary at DS= 0.001 m and 25oC. Stream sediments tend todevelop a sharp redox boundary where HrS existsjust below the stream bed - stream water interface,whereas SO!- predominates in the stream water.

The CN- concentration used for stream waters inTable 3 (3.84x l0-8 m or 1.0 pg/L) was estimatedfrom the values given by Krutz (1981), who meas-ured the.yearly variation in cyanide content of theWehebach Brook, which drains a wooded andpastured area in Germany. I3utz (1981) also statedthat "The sediment fraction with a particle diameterless than I mm contains about 0.2 me/kg of totalcyanide, only a small amount of it (0.04 mglkg)being easily set free." Accordingly, the CN- con-centration used in Table 3 for stream sediments was40 p.e/L (1.54x 10-6 m).

Various studies indicate a wide variety of naturalsources of cyanide, including certain plants (Lakinet al.1974, Hughes 1981, Leduc 1981,'Warren 1982),and algae (Vennesland et al. l98l). Seigler (1981)described some of the natural sources of cyanogenicglycosides and cyanolipids, which upon hydrolysisliberate hydrogen cyanide; they include apricot,peach, and almond trees, as well as other importantfood plants, ferns, gymnosperms, monoco-tyledonous and dicotyledonous angiosperms, fungi,bacteria, and several insects. Altogether, cyanide isproduced by more than 2050 plant species. Krutz(1981) noted that man-made sources of cyanide, suchas steel and galvanic industries, fertilizers, anddomestic sewage, can elevate the cyanide content ofstreams to levels as high as 200 pg/L, The samerespective estimates were used for the concentrationsof SCN- as for CN- in Table 3, as it was assumedthat cyanide would be present in similar concantra-tions whether or not the solutions were sufficientlysulfur-rich to convert the CN- to SCN-.

GOLD.RICH RIMS ON ELECTRUM IN PLACERS

TABLE 3. PREDICTED GOLD.COMPLEX CONCENTRATIONS

221

Au-corytx(r) hgttlgliwe) hgtcmlr.(3) lt ttUtLs(O bgtorrl"r(9 Ref.(O

TRIVALENT GOLD COMPLEXES

FOIIR,LrcANI}S

A(cNL-Au(NH3ia3+AuI+-A(oIIbB12Au(SCN)a-AuBr4-

:1.42-851:1.42-E.00, -5.35:1.42{.35

-5.E1{51:1.42{.m,6.35-5.81{.35

-8.02-9.11-t4.14-fl.96-$a-x.15

-8.15-55.49-6L91-66:13-61.64:t3.93

DFABGACACADACA

Au(SCPrz- 4.4 -x:14

FTgD,LIGN{I}S

SIXLIGANI}S

-5.81 -67.49

-7t25

MONOVALENT COLD COMPLEXES ..--...-.-

ONELIGA}ID

Au(SCN)63-

AuS-AuIIS(q)AUSOr:AuS2O3-

--(r) :1,99:l.gg-t325-13.$

-tt.32-t3rl-3t.4-3tcl

EFAEFAEFAEFA

TIVOLIGANIIS

A'{cNb-Au(OII)2-AuOIICI-Au12-AuOHBfAu(SCN)2-Au(NHfirAuCt2-AuB12-Au(HS)2-^Au(SO!2r-Au(S2Oj25-

:1.42{.00{.00, -3.65:7.42-E.00 -6.35:1.42-851-3.65

=

-5.Et-E.(x){.00, -3.55:7.42-8.00,5.35-5.81-851-3.654.35:1.9-r3,25-t3.47

:732-823-9.89-9.n-1050-11.88-11.93-tLzl

:,

-6.ll-24.49-26.15-mB-m75-2l,.14-25.19-8.47-30.64-1616-30.04-31.19

CEAAcCACADABGACACAEFABFAEFA

DI4OLI' COMILEXES

Au2@S)2S2-Au2S2z-

:l.gg:L99

-lE.62-2tl.3

BFABFA

( l )

a)cl)(4)(5)(o

CI

Cmentalim daa is not pedicrcd fc AI{OII)2FA- or Au[CS(NH)212+, even though their "Sg" prmctfrr frmTable 2regreaerthan -1.65, because cdc€ntratim,{rtafothe ligmds inneal systrrns aerptavailable.logtliglsw = log activity (* cmemado) of tbe ligd in average strem wrt€r&log[onx]r* = log activity (- e6s€n6alirl of the Au-coqrlex in average crean waters.logtltglrs = bg activity (- cmcentatim) of lbe ligard in avemge redrcing solutios in agmicomer*ich sediment&logtcrtxlrc = log activity (= cmc.) of the Au-coplex in average redrcing solutims in ugmic-moer-rich sedims.References: A - Bam l9q B - F€rh 1966: c - Fuge l978a,b,c; D - Knnz l98l; E - Psr l98l;F-Wagmmetal. l9ee C-mcrzka 198.Redrcedsulinspecieseenor$ableina€ra[odstremwrt€rs. Althoght@rnaybereleasedfrmcgmic-rnd€r-richsedin€nts ed persist rretasubly in tte strean waters, lhere ue ro repqts of measrable cfic€otratims of S2-, HS-,SO3z-, or StOrz' in urpolftned srem waers.

222 THE CANADIAN MINERALOGIST

The NH3 concentration used in Table 3 wasdetermined using a pK value of -9.25 for the reac-tion NHf, - NH: * H*, and by assuming that the"ammonia" concentrations of natural waters(repofied inconsistently as either NH, or NHf)actually represents the total of both species. Thenumber of measured natural concentrations is notlarge, but a range of -5 ppb to -5,000 ppb appearsto be common for ocean water, river water, rainwater, and groundwater. River water and ground-water tend to have concentrations in the middle toupper portion of this range, but as a conservativeestimate, 100 ppb was chosen as the total ammoniaconcentration, yielding an activity of NH3 for usein Table 3 equal to 3.12x l}e m.

The concentrations of the HS-, S2Of, and SOr2-ligands in sediment solutions were calculated byassuming chemical equilibrium with H2S for the

appropriate Eh and pH conditions of stream sedi-ments. The HrS concentration value used for thesecalculations was taken from Pan (1981), who meas-ured the concentration of this species in a neutralpaddy soil that was adjusted to different pH values.The reported concentrations are likely to be close tothose found in water-logged stream sediments con-taining some organic matter. This approach is likelyto underestimate the concentrations of the sulfoxyligands, as these species are formed by partial oxi-dation of HrS and may persist metastably at higherconcentrations (Giggenbach 197 4, Hoffman 1977).The discovery of numerous small crystals of coppersulfide growing on the surfaces of recent coins foundwhile panning in Brush Creek subslantiates thepresence of significant amounts of sulfide in thesestream sediments (Groen et al. 1986). The crystalshave two basic types of morphology, roughly

Conoding DewbpingEecrum Gol4Rich Rim

Ie' + 0.2502 +05H2O>OH'

DifrrionBoundary

|.apr

3 L2'

BulkSolulion

Ftc. 10. Self-electrorefining model for the formation of a gold-rich rim. (a) Schematic illustration of the self-electrorefining process occurring on the surface of a placer electrum grain entailing the following steps (see cor-responding small circled numerals on diagram): (l) transport of Au- and Ag-complexing ligands (LzJ to the cor-roding surface of the electrum, (2) dissolution of the electrum by the complexing agents, with coincident liberationof three electrons (e-), (3) loss of the more soluble Ag-complex to the bulk solution, (4) dissociation of the Au-complex by reaction with the free electrons at the pure gold surface bringing about the precipitation of pure gold,(5) consumption of extra free electrons through reduction of O2 and H2O to OH- (diagram patterned after Fig.I of Strickland & Lawson 1973), (b) Back-scattered electron image of a potshed placer gold grain from Brush Creekshowing a gold-rich rim (brightest, high-electron-yield areas), with textural features nearly identical to those seenin the schematic of part (a).

GOLD.RICH RIMS ON ELECTRUM IN PLACERS 223

orthogonal prisms and trigonal to hexagonal plates,the latter being more common. The hexagonal platesare believed to be covellite, and the orthogonal crys-tals, chalcocite. Based on the log a(S) versus loga(O) diagram for copper (Garrels & Christ 1965, p.l@), covellite is only stable at log a(S) greater thanabout -21.0 [og a(S2) = -33.2]. These estimatesare consistent with those given by Pan (1981).

The concentrations of the gold complexes aregenerally much higher for stream waters than for thesediment solutions, because the more oxidizing con-ditions in the stream waters promote oxidation ofelemental gold to gold ions. The most likely com-plexes for significant transport ofgold in oxidizingstream waters indicated by Table 3 are Au(CN)j,Au(OlI)j, Au(NH3)l+, AuOHCf, AuOHBT- andAu(I)j, whereas in the stream-sediment solutionsAu(CN)! and Au(HS)z are potentially important,although Au(CN)! is by far predominant. Thesecomplexes are the main ones likely to yield concen-trations of dissolved gold in typical fresh waters, nearor above the upper limit of natural measured con-centrations (Fig. 9).

Table 3 indicates that cyanide is by far the mosteffective complexing agent of gold in both oxidiz-ing and reducing conditions; unfortunately, the smallnumber of measured CN- concentrations in naturalwaters makes it difficult to obtain a reliable estimateof its average concentration.

Thiosulfate has been credited as a likely trans-porter of gold in acidic, oxidizing sulfide ore deposits(Goleva et al.1974, Stoffregen 1986). Table 3 indi-cates that this ligand is not especially important innormal stream waters, however, where total sulfuractivity is considerably lower and Eh-pH conditionsare markedly different.

Precipitation of gold complexes from solutiontakes place by the reduction of the Aul* or Au3*ions as they encounter solutions with a lower Eh. Asteep redox boundary exists at the interface betwpenstream waters and organic-matter-rich stream sedi-ments. In Brush Creek, for example, crystals of cop-per sulfide grew on coin surfaces only one or twocentimeters below oxidizing stream waters. Thisredox boundary is a very likely site of gold precipi-tation. Sharp redox boundaries also occur at thewater table, and the large gold nuggets found downto, but not below, the water table in the southeasternUnited States (Cook 1987) may have grown at sucha redox boundary.

The precipitation of aqueous gold complexes wasexamined by Mann (1984) and Stoffregen (1980. Theprecipitation mechanism suggested by Mann (1984)uses Fd+as a reducing agent. The Fd* ion, liber-ated from weathering rocks, migrates in solutionuntil it encounters and releases its available electronto a gold complex. This reaction simultaneouslyprecipitates native gold and ferric hydroxide:

Au(CN)! + Fd+ + 2H2O :A u o + F e O O H + 2 H C N + H + .

Strong evidence supporting this type of process isprovided by the common occurrence of iron hydrox-ide coatings on many of the recovered placer grains,and the presence of iron hydroxide inclusions in thegold-rich rims, as noted above. There are also numer-ous reports of the common and intimate associationof the two phases @esborough 1970, Lesure 1971'Mann 1984, Wilson 1984, Dilabio et ol. L985,Freys'sinet et a|.1989). A precipitative origin for gold-richrims is furthermore supported by the fact that thetextures observed on natural grains are commonlyindistinguishable from those texture$ resulting fromthe hydrometallurgical process of cementation (Fig.10). Strickland & Lawson (1973) described "cemen-tation" as: "...the electrochemical precipitation ofa metal, usually from an aqueous solution of its salts,by a more electropositive metal." In the present case,however, the cementation process is driven by theoxidation of ferrous iron, rather than by the oxida-tion of a solid metal substrate.

Se(-electrorefining

Electrorefining is a hydrometa[urgical process inwhich a multicomponent alloy is electrochemicallydissolved, with the subsequent precipitation of agenerally pure phase of the most noble dissolvedmetal. The driving force for this process is an elec-trochemical potential. This potential can be broughtabout in two ways: (1) with a battery, (2) as a resultof the electromotive force @MF) between two dis-similar metals in a solution whose Eh is higher thanthat in which the starting alloy is stable. Our con-cerns are with natural systems, where batteries arenot involved, and so the term "self-electrorefining"will hereafter be used to refer to electrorefiningbrought about by the second process listed above.

Fontana's (1986) description of the dezincificationcorrosion proce$s shows that the dissolution of thebrass requires oxidizing solution conditions and thatthe EMF formed by the copper-brass couple causespure copper to precipitate. According to Fontana(1980: "The commonly accepted mechanism con-sists of three siep$, as follows: (l) the brass dissolves,Q) the zinc ions stay in solution, and (3) the copperplates back on." For our concerns then, dezincifi-cation is a form of self-electrorefining, even thoughit involves man-made materials, because it proceedswithout an external source of power.

Mann (1984) performed electrochemical experi-ments, which showed that in 0.03 la HCI - I ru NaClsolutions, a distinct increase in Au-Ag alloy stabil-ity occurs with increasing fineness as measuredagainst a pure gold electrode, demonstrating lowerreduction potentials for lower fineness alloys.

224 THE CANADIAN MINERALOGIST

Mann's (1984) results with synthetic alloys indicatethat placer electrum grains can oxidize, releasing dis-solved Ag and Au into solution @g. l0a). Once insolution, the less electropositive (and now isolatedor "refined") dissolved Au species accepts the avail-able electrons at the gold-rich areas on the electrumsurface, and immediately precipitates to yield the tex-tures described below. The more easily oxidized sil-ver is washed away. The resulting electrolytic reac-tion is self-perpetuating as a result of the EMFbetween the gold-rich areas and the unrefined elec-trum. This process does not require a sharp Eh gra-dient to precipitate a gold-rich rim, just an ambientsolution that is oxidizing enough to initiate dissolu-tion of the primary electrum. This process would beaccelerated by the presence of ligands such aS CN-, OH-, NH3, Cl-, I- Br, or HS-; as they increase thesolubility of the electrum components. Rudimentaryforms of this idea were suggested more than 50 yearsago by Fisher (1935).

The electrorefining process frequently producessurfaces bearing lobate protrusions of the depositedmetal (crystalline growths also have been reported).These protrusions form because the surface of themore electropositive metal must maintain sites thatremain exposed to the aqueous medium so that dis-solution and subsequent charge-transfer c:rn occur;hence "channelways" develop between the growingmetal deposits on the surface @ig. 10). Thebranching-coral texture of Figure 2 and the magni-fied view of the lobate gold-rich rimming of Figure4a (shown in Fig. 10b) are strikingly similar to theschematic representations of cementation tgxtures inStrickland & Lawson (1973) (after which Fig. lOa ispatterned), and suggest that these rims were formedby precipitation from an electrorefining process. Thesize of the protrusions and the ratio of gold-coatedsurface area to surface area of the dissolving elec-trum core are highly variable on natural placergrains, as would be expected for grains of differingfineness of the core electrum in solutions of differ-ing ambient composition (yielding variable reactionrates and availabilities of gold for precipitation). Wedo believe, however, that the same basic processesare responsible for the range of microtexturesobserved in the gold-rich rims.

Gold-rich rims and their associated surface tex-tures appear to be the only likely products ofextended exposure of electrum grains to stream andsediment waters. Gold-rich rims are very commonon placer gold grains from the streams throughoutthe present study area, as well as from other locali-ties that we, and other investigators @esborough1970, Giusti & Smith 1984, Freyssinet et al. 1989),have sampled. These rims appear to form by elec-trorefining where sharp Eh gradients do not exist (inflowing stream water), and by cementation wheresharp Eh gradients do exist (e.9., the interface

between fine, organic-matter-rich sediment andoxidizing stream water, or at the water table). Wheregold precipitation resulting from an Eh gradient isthe dominant process, it is possible that the gold-richrim is an incipient stage of a large "grown" nuggeqthus gold grains that happen to reside in a soilhorizon at the water table, for example, might growto the remarkable and enigmatic sizes reported fromvarious areas in the southeastern United States(Crayon 1857). The validity of this idea could betested by analysis of large gold nuggets. If theyformed by an overgrowth process, the fineness ofthe outer portion of such grains should be very closero 1000.

CoNcrusroNs

Gold grains undergo both physical and chemicalchanges as they are transported by streams. Newlyliberated gold grains are usually irregularly shaped,and as a result of progressive stream transport theybecome semispherical, wafer-shaped, and finallyflake-shaped. Their surfaces evolve from smooth andclean to pitted, hackly, and eventually, to lobate-textured, although variations in the stream's com-position, energy and sediment type somewhat modifythis general trend. Grain size decreases with increas-ing distance of transport.

During transport, many grains of placer golddevelop an outer rim of nearly pure gold on the moresilver-rich electrum core. The rim has a very sharpcontact with the core, and appears to be the resultof electrochemical processes active in the stream orstream sediments (or both). The rim generally isthickest on flake-shaped (most transported) grainsand thinnest or absent on irregular (east transported)grains.

Three proposed mechanisms of gold-rich rimdevelopment were evaluated on the basis of obser-vational evidence and theoretical calculations. Selec-tive leaching of silver from the margin of gold grainsappear$ to be impossible, because there is no meansby which the solutions can physically extract the sil-ver from more than the outer few 0ngstrdms of thegrain. Diffusion of silver through the electrum tomake the silver available to solution cannot aid thisprocess, owing to the extremely slow rates of diffu-sion at low temperatures. Gold dissolution, trans-port, and cementation are steps in a viable processof growth of the gold-rich rims. There are sufficientamounts of CN-, OH-, NH3, Cl-, I- Br-, and HS-in average stream and sediment solutions to accountfor the highest gold concentrations reported forstreams. Several sulfur-bearing ligands are poten-tially capable of significant transport of gold,provided they can maintain their concentrations inoxidizing waters through metastable persistence.Some organic complexes also show potential for gold

COLD-RICH RIMS ON ELECTRUM IN PLACERS 225

transpofi; however, there simply are not enough dataregarding these species for a quantitative evaluation.Sharp redox boundaries, such as the stream water- sediment interface and the water table, would causegold precipitation. Self-electrorefining of placer elec-trum grains is also a process that can produce gold-rich rims. The required conditions are present in moststream environments, and the textures of the gold-rich rims match those reported from controlledhydrometallurgical experiments with other alloys.Thus, a combination of self-electrorefining andcementation can explain the formation of gold-richrims, as well as observed surface textures on thegrains of placer gold.

ACKNoWLEDGEMENTS

The authors express sincere gratitude to RobertF. Martin, Scott A. Wood, and an unnamed refereefor their insiebtful reviews of this manuscript. Sam-ples of placer gold were graciously donated for usein this study by the U.S. Museum of Natural His-tory and Mr. James R. McCloud. This research wassupported by the Department of the Interior'sMineral Institutes program administered by theBureau of Mines under allotment grant numberG l l 8415 l .

RSF'BnsNcss

Baes BscrrNc, L.G.M., Karrax, I.R. & Moone, D.(1960): Limits of the natural environmenl in termsof"pH and oxidation-reduction potentials. J. Geol.68, A3-284.

Bars, C.F., Jn. & MesNaen, R.E. (1976): The Hydrol-ysis of Cations. John Wiley and Sons, New York.

Banarove, N.N. & RvzHer.rco, B.N. (1981): Computersimulation of the Au-Cl-S-Na-H2O system in rela-tion to the transport and deposition of gold inhydrothermal processes. Geochem. Int. j:E(A), 4640.

Banms, H.L. (1979): Solubilities of ore minerals. /nGeochemistry of Hydrothermal Ore Deposits (2ndedition; H.L, Barnes, ed.). John Wiley and Sons,New York.

Bnmuu, J. (1971): Stability Constants of Metal-IonComplexes, Supplement No. 1, Part I: InorganicLigands (L.G. Sill€n, compiler). The Chem. Soc.,London.

Bornrcre, W.F. (1936): Prospecting and OperatingSmall Gold Placers (2nd ed.). John Wiley and Sons,New York.

Bonovrrsrrv, V.P., Mrllrn, A.D. & Ssnvryarn, V.N.(1966): Determination of minute amounts of goldin natural waters of the Aldan reeion. Geochem.3.371-377.

BoworsH, F.W. (1983): Fire assaying for gold and sil-ver. .Iz Proceedings: International Gold/Silver Con-ference III (Y.S. Kim, ed.). Nevada Inst. Technol-ogy, Reno, Nevada.

Bnoors, R.B., CnerrEnrrr, A.K. & RvaN, D.E.(1981): Determination of gold in natural waters atthe parts-per-trifion (pg cm-i) level. Chem. Geol,33, 163-169.

Cr:rnrvavnv, A.M., CnrruwevEva, L.E., YEREluEYEve.,M.N. & ANDReyEv, M.I. (1969): Hydrogeochemis-try of gold. Geochem. Int. 6,348-358.

Coor, H.E. & Httueno, J.E. (1969): Effect of gra-dient energy on diffusion in gold-silver alloys. ./.Appl. Phys. N, 219l-2197.

Coor, R.B., Jn. (1987): Notable gold occurrences ofGeorgia and Alabama. Mineral. Rec. lE, 65-74.

Cnarc, J.R. & CanaHar, J.E. (1990): Paleoplacergold in the Lilesville sand and gravel deposits, NorthCarolina. Georgia Geol. Sum., Bal/. (in press).

& Rrrr,rsrnr, J.D. (1985): Gold: compositionalations of naturallv occurring allovs. Gaol. Soc.variations of naturally occurring alloys.

Am., Abstr. Programs 17, 555.

SoLrEnc, T.N, & Snans, C.E. (1980):Appalachian gold. Placer deposits of MontgomeryCounty, Virginia. Geol. Soc. Am., Abstr. Programst2 .175 .

CnevoN, P. (1857): North Carolina illustrated. IV. Thegold region. Harper's New Monthly Mag. 15,289-300.

Cnocrnr, J.H. (1978): Gold - abundance in naturalwaters and in the atmosphere. In Handbook ofGeochemistry (K.H. Wedepohl, ed.). Springer-Verlag, Berlin.

DnnrnN, L.S. & Gunxv, R.W. (1953): PhysicalChemistry of Metals. McGraw-Hill, New York.

Drsnoroucr, G.A. (1970): Silver depletion indicatedby microanalysis of gold from placer occurrences,western United States. Econ. Geol.65, 304-311.

Dnanro, R.N.W., Nnwsotrm, J.W. & Mclvon, D.F.(1985): Gold spheres in surficial sediments..Episodes8, 39.

Dnrscor,r, A,J., Herl, D.L. & Cnatc, J.R. (1990):Placer gold: an unrealized potential for characteriz-ing lodes. Georgia Geol. Sun., Bzl/. (in press).

EuuoNs, W.H. (1937): Gold Deposits of the World:with a Section on Prospecting. McGraw-Hill, NewYork.

Enasvrus, C.S., SBtt-scsop, J.P.F. & WetrrnsoN,J.I.W. (1987): New evidence on the composition ofmineral grains of native gold. Nucl. Geophys. l, l-23.

226 THE CANADIAN MINERALOGIST

FnrH, J.H. (1966): Nitrogen compounds in naturalwater - a review. lV'ater Resources Res. 2, 4l-58.

Frsnrn, M.S. (1935): The origin and composition ofalluvial gold, with special reference to the Morobegold field, New Guinea. Trans. Inst, Min. Metall.44,337-420.

Frsuen, N.H. (1945): The fineness of gold, with spe-cial reference to the Morobe Goldfield, New Guinea.Econ. Geol, 40, M9-495, 537-563.

For.nArNE, W.M. (1882): Notes on Virginia geology,the Brush Creek, Virginia Gold District. The Vir-ginias 3, 108-109.

FoNrana, M.G. (1986): Corrosion Engineering (3rd,ed.). McGraw-Hill, New York.

Fnrvssrxe"r, P., Zrrcrns, H. & Tarov, Y. (1989): Mor-phology and geochemistry of gold grains in lateriticprofiles of southern Mali. ,f. Geochem. Expl,32,t7-31.

FucE, R. (1978a): Bromine - Abundance in naturalwaters and in the atmosphere. In Handbook ofGeochemistry (K.H. Wedepohl, ed.). Springer-Verlag, Berlin.

(1978b): Chlorine - Abundance in naturalwaters and in the atmosphere, In Handbook ofGeochemistry (K.H. Wedepohl, ed.). Springer-Verlag, Berlin.

- (1978c): Iodine - Abundance in natural watersand in the atmosphere. .ln Handbook of Geochemis-try (K.H. Wedepohl, ed.). Springer-Verlag, Berlin.

GARrrLs, R.M. & Cumsr, C.L. (1965): Solutions,Minerak, and Equilibria. Freeman, Cooper & Com-pany, San Francisco.

GrccrwaecH, W.F. (1974): Kinetics of the polysulfide-thiosulfate disproportionation up to 240". Inorg.Chem. 13. 1730-1733.

Grusrr, L. & Svnn, D.G.W. (1984): An electronmicroprobe study of some Alberta placer gold.Tschermaks Mineral. Petrogr. Mitt. 33, 187-202.

GoLnvA, G.A., Knrvrurov, V.A. & Glnz, Z.G. (1970):Geochemical trends in the occurrence and migrationforms of gold in natural water. Geochem. Int.7,518-529.

GosLrNc, A.W., Jrxre, E.A. & Cuao, T.T. (1971):Gold content ofnatural waters in Colorado. Econ.Geol. 66.309-313.

GnonN, J.C., Cnarc, J.R. & Rrrvrsrror, J. (1986):Gold-amalgam and other secondary phases inplacers. Geol. Soc. Am., Abstr. Programs lE,62l.

Helr, G.E.M., VAtvE, J.E. & Bar-raNrvr.m, S.B.(1986): Field and laboratory procedures for deter-

mining gold in natural waters: relative merits ofpreconcentration with activated charcoal. /.Geochem. Explor, 26, l9l-202.

HautltoN, T.W., Elr,rs, J., Floneucs, T.M. & Fenov,J.J. (1983): Analysis of gold in surface waters fromAustralian goldfields: an investigation into directhydrogeochemical prospecting for gold. Econ. Geol.7t, t335-t34t,

HanrusoN, S. & Furlrn, J. (1987): Gold from Hope'sNose, Torquay, Devon, Eng)and. Mineral..Rec. lt,85-88.

HoFrrr4AN, M.R. (1977): Kinetics and mechanism ofoxidation of hydrogen sulfide by hydrogen perox-ide in acidic solution. Environ. Sci. Tech.11. 6l-66.

Hucurs, M.A. (1981): The genetic control of plantcyanogenesis. 1z Cyanide in Biology @. Vennesland,E.E. Conn, C.J. Knowles, J. Westley & F. Wiss-ing, eds.). Academic Press, London.

IUPAC (1982): Stability Constants oJ Metal-Ion Com-plexes. Part A: Inorganic Ligands @. Htigfeldt,compiler). Pergamon Press, Oxford.

JornsoN, P.R., Pnerr, J.M. & Trr.lnv, R.I. (1978):Experimental determination of the standard reduc-tion potential of the gold(I) ion. J. Chem. Soc.,Chem. Comm. 1978, 606-607.

Jores, R.S. (1970): Gold content ofwater, plants, andanimals. U.S. Geol. Sun. Circ.625.

Kasean, J., Htnrc, I., Scnrlrnn, P., Coor, G.8., Krr-zrNcER, A. & Wdrrr, E. (1972): A contribution tothe migration of gold in the biosphere of the humidmild zone. Chem. Geol. 10,299-305.

Komr, M., Honcr, V., Golornnc, E.D. & BeRTnrE, K.(1988): Gold in seawater: a conservative view.Applied Geochem. 3, 237 -241,

Kosuuex, P.N. & Yucav, T.A. (1972): The causes ofvariations in finenesss levels of gold placers.Geochem. Int, 9, 481-484.

Knutz, H. (1981): Different origins of cyanide con-tamination in small rivers. .& Cyanide in Biology@. Vennesland, E.E. Conn, C.J. Knowles, J. West-ley & F. Wissing, eds.). Academic Press, London.

LarrN, H.W., CunrN, G.C. & HuBntt, A.E. (1974):Geochemistry of gold in the rveathering cycle. U.S.Geol. Sum. Bull. 1330.

LnrrMsn, W.M. (1952): The Oxidation States of theElements and Their Potentials in Aqueous Solutions(2nd ed.). Prentice Hall, Inc., Englewood Cliffs,New Jersey.

Leouc, G. (1981): Ecotoxicology ofcyanides in fresh-water. In Cyanide in Biology (8. Vennesland' E.E.Conn, C.J. Knowles, J. Westley & F. Wissing, eds.).

GOLD-RICH RIMS ON ELECTRUM IN PLACERS 227

Academic Press, London.

Lrsunn, F.G. (1971): Residual enrichment and super-gene transport of gold, Calhoun mine, LumpkinCounty, Georgia. Econ. Geol.66. 178-186.

LrrocnrN, W. (l9ll): The Tertiary gravels of theSierra Nevada of Califomia. U.S. Geol. Sun. Prof.Pap.73.

LortroNosov, I.S., Konwrva, A.R. & LouoNosova,T.K. (1985): The origin of the lithogeochemical andgeochemical dispersal aureoles of gold-bearingplacers. Geochem. Int, 22(4), 59-65.

Macrev, R.A. (1944): The purity of native gold as acriterion in secondary enrichment. Econ. Geol.39,56-68.

Mauano, W.C., Genornn, A.8., Bass, R.F. & Srtr-rn, L.M. (1963): Self diffusion in silver-gold solidsolutions. Phys. Rev, 129(2). 611 -625.

MarN, A.W. (1983): Gold - nuggets and solutions. InCSIRO Div. Mineral. Research Review 1983 (W.E.Ewers, ed.). Inst. Energy Earth Resources, CSIRO,Canberra, Australia.

(1984): Mobility of gold and silver in lateriticweathering profiles: some observations fromWestern Australia. Econ. Geol.79, 38-49.

McHucn, J.B. (1988): Concentration of gold innatural waters. J. Geochem. Explor.30, 85-94.

MrNrvev, G.G. (1976): Organisms in the goldmigration-accumulation cycle. Geochem. Int. l3(2),t6+t68.

NysrnN, P. (1986): Gold in the volcanogenic mercury-rich sulfide deposit L8ngsele, Skellefte ore district,northern Sweden. Miner. Deposita 21, 116-120.

PaN Snu-ZnnN (1981): Chemical equilibria of sulfidesas studied with a hydrogen sulfide sensor. /rz Proc.Symp. Paddy Soil (Institute of Soil Science, Acade-mia Sinica, eds.). Springer-Verlag, Berlin.

RElnEns, P.J. & Spwano, T.M. (1989): The stability ofhydrosulphido- and sulphido-complexes of Au(I)andAg(I) at2!oC. Geochim, Cosmochim. Acta53,245-253.

ScHuro, G.M. (1985): Gold. ft Standard Potentials inAqueous Solutions (A.J. Bard, R. Parsons &J. Jor-dan, eds.). IUPAC, New York.

SErclEn, D.S. (1981): Cyanogenic glycosides andlipids: structural types and distribution. .Iz Cyanidein Biology (B, Vennesland, E.E. Conn, C.J.Knowles, J. Westley & F. Wissing, eds.). AcademicPress, London.

Srsco, A.G. & Surrn, C.S. (1949): Bergwerk-und-Probierbilchlelz (trans.). Am. Inst. Min. Metall.

Eng., New York, as cited in Bonr, R.W. (1987):Gold: History and Genesis of Deposits, Van Nos-trand Reinhold, New York.

Srmsreo, L.H. & BERnuu, J. (1974): Studies on goldcomplexes. II. The equilibrium between gold(I) andgold(IID in the ammonia system and the standardpotentials of the couples involving gold, diam-minegold(I), and tetramminegold(Ul). Acta C hem.Scand. 428,764-770.

&-OnT: Studies ongold complexes. III.The standard electrode potentials of aqua gold ions.Acta Chem. Scand. 431, 155-156.

Solssnc, T.N. & Cnaro, J.R. (1981): Chemical varia-tions in gold from the central Appalachians of Vir-gpna. Microbeam Analysis lgEl, 163-166.

Srorrnrcru, R. (1986): Observations on the behaviorof gold during supergene oxidation at Summitville,Colorado, U.S.A., and implications for electrumstabitty in the weathering environment. AppliedGeochem. 1, 549-558.

SrnrcxLeNo, P.H. & Lawsort, F. (1973): The measure-ment and interpretation of cementation rate data.Iz Int. Symp. Hydrometallurgy @.J.I. Evans & R.S.Shoemaker, eds.). Am. Inst. Min. Metall. &Petroleum Engineers, Inc., New York.

Srurrrcrl, E.F. & Cnnr, A.M. (1965): Electron-probemicroanalysis of gold-platinoid concentrates fromsoutheast Borneo. Trans. Inst. Min. Metall. 74,933-946.

TunnrnN, K.K. (1968): Oceans. Prentice-Hall, Inc.,Englewood Cliffs, New Jersey.

TunlrEn, L.D. & knanruoorN, M. (1982): Electrother-mal atomic absorption determination of gold, sil-ver, and arsenic in stream waters and their relation-ship to gold-silver occurrences in the RepublicGraben, N.E. Washington. In Precious Metals inthe Northern Cordillera (A.A. Levinson, ed.). TheAssociation of Exploration Geochemists, Rexdale,Ontario.

Varsual, G.M., ValwrHANovA, T.K. & BanANova,N.N. (1984): The geochemical role of gold (III) ful-vate complexes. Geochem. Int,27(3), 139-146.

VsNNpsr-ANp, B., Ptsrontus, E.K. & Grwtrz, H.S.(1981) : HCN production by microalgae. .tz Cyanidein Biology (8. Vennesland, E.E. Conn, C.J.Knowles, J. Westley & F. Wissing, eds.). AcademicPress, London.

Vor- ov, G.A. & Srnrunazova, L.V. (1975): Somedata on geochemistry of gold in natural waters ofthe Beskessk ore field. Int. Geol. Rev.77,lM-148.

WAcMaN, D.D., Evaus, W.H., Panr<rn, V.B., Sctttnau,R.H., Harow, I., Barr,sv, S.M., CnunNev, K.L. &Nurrarr, R.L. (1982): The NBS Tables oJ Chemi-

228 THE CANADIAN MINERALOGIST

cal Thermodynamic Properties. Am. Chem. Soc. &Am. Inst. Phys. for Nat. Bur. Stds., Washington.

WannEN, H.V. (1982): The significance of a discoveryof gold crystals in overburden. .Iz Precious Metalsin the Northern Cordillera (A.A. Levinson, ed.). TheAssociation of Exploration Geochemists, Rexdale,Ontario.

WEnsreR, J.G. (1986): The solubility of gold and sil-ver in the system Au-Ag-S-O2-H2O at 25oC andI atm. Geochim, Cosmochim. Acta 50,1837-1845.

Wrrsoll, A.F. (1984): Origin of quartz-free gold nug-gets and supergene gold found in laterites and soiis

- a review and some new observations, Aust, J,Earth Sci.3l, 303-316.

Wr-orzre, F. (1978): Nitrogen - Abundance in naturalwaters and in the atmosphere. In Handbook ofGeochemistry (K.H. Wedepohl, ed.). Springer-Verlag, Berlin.

Xue, T. & Ossro-Asanr, K. (1985): Heterogeneousequilibria in the Au-CN-H2O and Ag-CN-H2Osystems. Metall. Trans. 16B, 455-463.

Received July 5, 1989, revised manuscript occeptedFebruary 28, 1990.