GOLD PROSPECTING ECUADOR: ARTICLE-1

Microchemical

Characterization of

Alluvial Gold GrainsMicrochemical

Characterization of

Alluvial Gold Grains

as an Exploration

Tool

Robert Chapman, Bob Leake and

Mike Styles*

School of Earth Sciences, University of Leeds, Leeds

LS2 9JT, United Kingdom

*British Geological Survey, Keyworth, Notts NG12 5GG,

United Kingdom

E-mail: aedrjc@LUCS-03.NOVELL.LEEDS.AC.UK

Received: 15 October 2001

There is considerable variation in the composition of

native gold and the nature of minerals co-existing with

it, and this reflects differences in the geological

environment and chemistry of ore-forming processes.

In areas where gold-bearing mineralization is subject

to active fluvial erosion, especially in temperate

climatic regimes, any discrete grains of native gold

pass into alluvial sediment with little modification.

The chemical characteristics of alluvial grains and the

nature of preserved mineral inclusions provide a

signature which points back to the type of source

mineralization. This signature may be established

using electron probe microanalysis and scanning

electron microscopy and can be interpreted to provide

information about the original bedrock mineralization.

Identification of the type of source mineralization

using the technique at an early stage in regional

exploration can help focus attention on targets with

the most potential economic importance.

In nature, gold occurs predominantly as the native metal,

although it is commonly alloyed with highly variable amounts

of other metals; primarily silver, but also mercury, copper and

palladium. It may also occur within common sulphur-rich

minerals such as pyrite and arsenopyrite either as submicroscopic inclusions of native gold or as a minor

component within the lattice of these minerals (1). In a few

occurrences gold may be present primarily in combination

with tellurium in such minerals as calaverite ((Au,Ag)Te2)

rather than as the native metal. There are a wide variety of

types of gold mineralization, influenced by differences in

their geological setting, the chemistry of the ore fluids, and

the nature of their reactions with rocks into which they

penetrate. Characterization and classification of ore deposits

has long been based on assessment of the geological

environment of formation as inferred from structures, and

mineralogical and chemical features observable in the field,

augmented by chemical data. More recently, genetic models

of major deposit types have been formulated as a result of

combining descriptive information with increased

understanding of the physics and chemistry of mineralization

processes obtained from experimental and theoretical work.

The classification of gold deposits has developed

progressively by the application of genetic models and as

new types have been discovered as a result of exploration,

such as the Olympic Dam Cu-Au-U deposit of South Australia

(2). Comprehensive reviews of gold deposit classification

follow after periods of extensive exploration and mineral

deposit research, such as that of Foster (3) in response to the

proliferation of exploration activity in the previous decade.

Variation in chemical parameters, both in environments of

ore fluid generation and ore precipitation, result in a

considerable variety in the mineralogy of gold-bearing ores.

This manifests itself both in the composition of the native

gold alloy and the associated minerals. This variation may

also be seen in comparisons of gold deposits of the same

type, for example Hedenquist et al (4) indicates the range of

minerals associated with some economically important

epithermal gold deposits.

Commonly, minerals which co-exist with native gold in the

source mineralization also occur as microscopic inclusions

within the gold (Plates 1, 2). Grains of native gold are

chemically stable within most, but not all, environments on

the Earth’s surface, and thus gold grains liberated from the

hypogene ore are normally unchanged on passing from

bedrock into superficial sediments as a result of weathering

and erosion. Evidence for the ability of gold to form an

effective barrier between inclusions and the atmosphere is

provided by the case of an alluvial grain from Ecuador which

contained large (200µm) complex multiple telluride

inclusions. The inclusions were observed to have suffered

marked oxidation and hydration in less than 3 weeks54 Gold Bulletin 2002 • 35/2

following preparation of the polished block and exposure to

the atmosphere in the laboratory.

Some grains, especially those rich in silver, may show

compositional modification of the outer part but even then a

relict grain core that retains its original composition and

inclusions is commonly recognizable (5 – 9). However, in

areas which have suffered long and intensive periods of

lateritic weathering, native gold may have completely

recrystallized, so as to reflect this environment rather than

that of the original source mineralization (10).

The search for bedrock gold deposits has historically

involved systematic searches for alluvial gold in river gravels

and the sampling of drainage sediment remains an

important activity within the suite of modern exploration

techniques. Gold concentrations can be determined directly

by chemical assay of various size fractions or other

components, including a heavy mineral fraction obtained by

panning. Panning allows the visual detection of grains of

native gold directly in the field which in many cases leads to

the location of bedrock mineralization. However, there have

also been fruitless searches for the source of alluvial

concentrations of native gold. This paper describes how the

microchemical characterization of a number of alluvial gold

grains from a given site can provide information at an early

stage in the exploration process that permits informed

speculation about the type or types or mineralization from

which the alluvial gold is derived.

Many workers have investigated the link between the

composition of alluvial native gold grains and potential

sources, examples being studies of Witwatersrand gold (11,

12), and gold from the Yukon (7). Desborough (13) first

suggested the potential of mineral inclusions within the gold

as an aid to distinguishing between alluvial gold from

different sources. However, the work of the present authors,

which is summarised here (and augmented by previously

unpublished data for localities in North America and

Australia), represents the first studies to systematically record

the mineral inclusion assemblages and to generate

classifications of gold grain chemistry which combine this

information with that of the gold alloy composition. The

resulting ‘microchemical signature’ provides a more powerful

technique for interpreting the origin of alluvial gold than gold

alloy composition alone, because the composition of native

gold grains can vary from point to point within the same

mineralized structure (9). This work has been undertaken for

a period of over 15 years during which over 20,000 gold

grains from 314 localities in Great Britain and Ireland,

together with other sites in North America, South America,

southern Africa, Australia, south east Asia, and Fiji, have been

studied. A wide range of compositional variation in alluvial

gold grains has been observed reflecting differences in the

Plate 1

Alluvial gold grain from the Crediton Trough, Devon containing many

microscopic inclusions of selenide minerals

Plate 2

Alluvial gold grain from Ximena area of Ecuador containing relatively

large complex inclusion of CuAg sulphotelluride (dark grey), Bi metal

(pale grey) and a Bi+Cu+Au alloy (pink)

Figure 1

Worldwide locations of sites discussed in the text. 1: Klondyke District,

Yukon Territory, Canada, 2: Ximena Province, Ecuador, 3: Mberengwa

District, Zimbabwe, 4: Lubuk Mandi, Malaysia, 5: Victoria goldfield,

Australia, 6: Wainmanu River, Fiji, 7: Localities in Great Britain and Ireland

detailed in Figure 2Gold Bulletin 2002 • 35/2 55

style and geological environment of the host mineralization.

Sufficient native gold obtained from bedrock and alluvial

sources in Great Britain and Ireland has been studied to

provide country-wide perspectives of the different types of

gold mineralization present. Where possible, alluvial gold has

been studied from several other parts of the world where the

gold mineralization is of greater economic significance and

which differ from the British Isles both geologically and

Table 1 Location and References for Sites Described in the Text

Country/Region Detail of Locality Geological Setting of Geomorphological/Climatic Reference

Gold Mineralization Description of Locality

Great Britain

and Ireland

Southern Scotland Leadhills area Metasediments, Southern Upland area, cool temperate climate 31

Uplands Terrain

Southern Scotland Tweed headwaters Metasediments, Southern Upland area, cool temperate climate 14

Uplands Terrain

Southern Scotland Glengaber Burn Metasediments, Southern Upland area, cool temperate climate 16

Uplands Terrain

Central Scotland Borland Glen, Ochil Hills Acid-intermdiate volcanic Upland area, cool temperate climate 37

sequence

Northern Scotland Sutherland area, Cononish Metasediments Grampian Upland area, cool temperate climate 8, 14

R., Glengarry, West Water, Terrain

Calliacher Burn, Glen Lednock

Northern Ireland R. Bann, Mourne Mountains Metasediments, Southern Upland area, cool temperate climate 8

Uplands Terrain

Ireland Co Mayo Metasediments, Southern Upland area, cool temperate climate 8

Uplands Terrain

Ireland Balwoges, Co Donegal Brecciated pipe and Upland area, cool temperate climate 8

volcanic boss

England, Devon South Hams District Red bed associated Lowland area, cool temperate climate 15, 35

England, Devon Crediton Trough Red bed and alkali basalt Lowland area, cool temperate climate 36

associated

Wales, Dolgellau Gwynfynnydd Mine Metasediments Upland area, cool temperate climate 9, 38

Gold Belt

Zimbabwe

Mberengwa District ‘C’ Mine Greenstone Belt Deep tropical weathering, slow erosion, 24

semi-arid climate

Sebakwe River Greenstone Belt Deep tropical weathering, 24

slow erosion, semi-arid climate

North America

Klondyke District, Bonanza Creek Metasediments Upland area, sub arctic climate 7, 26

Yukon Territory Bear Creek

Hunker Creek

Australia

Ballarat Goldfield, Dolly Creek, Violet Town Metasediments Temperate rain forest 39

Victoria

Walhalla Goldfield, Jordan River, Woods Point Metasediments Temperate rain forest 39

Victoria

Malaysia

Lubuk Mandi Metasediments Steep relief, rapid erosion, tropical 25

rainforest with waterlogged soils

Fiji

Emperor Mine Alkali-epithermal Tropical rain forest 23

Wainmanu River Porpyry+ alkali epithermal Tropical rain forest 23

Equador

Catopaxi province Valetanga High altitude rain forest, no tropical 30

weathering

Perros Bravos56 Gold Bulletin 2002 • 35/2

climatically, but much more work is required to cover the

whole range of mineralization types.

The geographical distribution of the gold localities

referred to in the text are provided in Figures 1 and 2, and

descriptions of the host environments of the gold

mineralization are provided in Table 1.

Determination of the Microchemical

Signature of Alluvial Gold Grains

The size of the population of grains studied from each

locality is dependent upon the abundance of opaque

inclusions within the grain. The proportion of sectioned gold

grains containing identifiable opaque inclusions varies widely

but in Britain and Ireland, they are typically found in about

20% of the grains (8), so a population of 30 grains is usually

sufficient to generate useful information. Where the

incidence of inclusions is lower and where multiple sources of

alluvial gold contribute to the alluvial population, a

correspondingly greater number of grains is required. This is

rarely a problem in mining areas, but the collection of even

30 alluvial gold grains from some areas where mineralization

is sparse or remote from drainage may be difficult and

specialized field techniques have been developed for this

purpose (14).

Prior to the mounting of gold grains in epoxy resin for

grinding down and polishing, observations can be made on

their size and shape. In a few cases grains with intricate

shapes or surface textures may be present in alluvial

sediment. At a few sites in South Devon (SW England),

dendritic grains (Plate 3) were recorded in the alluvial

sediment (15), very similar in form and composition to gold

occurring in carbonate veins exposed on the coast at Hopes

Nose near Torquay (Devon, SW England). Dendritic grains

cannot survive more than a trivial amount of transport before

dendrite spikes are folded around the core of the grain. Thus,

their presence in alluvial sediment indicates very close

proximity to a bedrock source.

After polishing, grains are examined by scanning electron

microscopy (SEM) and electron probe microanalysis (EPMA) to

determine the concentrations of alloying elements (Ag, Cu,

Hg, Pd, Sn) within the gold and the nature of the inclusion

suite. Details of the experimental procedures are given in

reference 8.

Mineral inclusions are of two types: opaque minerals, such

as sulphides and sulpharsenides and translucent minerals,

most commonly quartz and carbonates. In general, the

opaque mineral assemblage is more useful in characterizing

the type of source mineralization, although in some cases

translucent minerals suggest the presence of specific types

of mineralization, eg calcium-rich garnets and wollastonite

may be indicative of skarn mineralization. The opaque

mineral inclusion suite is reported in terms of the various

mineral classes, ie sulphides, sulpharsenides and arsenides,

sulphosalts and antimonides, tellurides and selenides which

reflect ore composition in terms of Groups Vb and VIb

elements. The classification of the inclusion assemblage is

Figure 2

Locations of gold deposits in Britain and Ireland mentioned in the text.

Scotland, N. Highland and Grampian Terriane: 1: Sutherland, 2: Glen

Garry, 3: Cononish River, 4: Glen Lednock, 5: Calliacher Burn, 6: West

Water, 7: Borland Glen, 8: Leadhills Region, 9: Tweed Headwaters, 10:

Glengaber Burn, 11: Gwynfynydd Mine, Dolgellau Gold Belt, 12: Crediton

Trough, Devon, 13: South Hams District, Devon, 14: River Bann, Mourne

Mountains, 15: Balwoges, Donegal, 16: Co Mayo

Plate 3

Dendritic gold grain from alluvium in South Devon. A thin rim made of

Pt-rich alloys coats the gold grainGold Bulletin 2002 • 35/2 57

made in terms of the rough proportions of individual

inclusions corresponding to the mineral classes. In a few

cases the presence of specific or unusual inclusions may

contribute significantly to the microchemical signature. For

example, gold from Balwoges, a locality in Northern Ireland,

contained a high proportion of copper sulphide inclusions,

which are generally very rare (8). In other cases a distinctive

inclusion suite may be restricted to populations of alluvial

gold grains from specific sites within an auriferous region,

and this may be indicative of a particular mineralizing event.

For example, a study of the Glengaber Burn area in Southern

Scotland showed that gold from the richest alluvial locality

contained inclusions of tetrahedrite and sphalerite which

were absent in alluvial gold from nearby rivers (16), despite

the gold alloy compositions being broadly similar throughout

the study area. A study of alluvial gold from three adjacent

auriferous rivers draining the Dolgellau Gold Belt in North

Wales also showed different inclusion assemblages in each

case. One assemblage comprised sulphides and

sulpharsenides whilst 40% of the inclusions from another

were tellurium-bearing minerals. The third population was

distinguished both by the presence of molybdenite inclusions

and by a mean silver content in the gold alloy 15% higher

than recorded for gold from the adjacent rivers (9).

The composition of a population of gold grains in terms of

their silver content (and where appropriate other metals) is

represented using a plot of the type presented in Figure 3.

Each gold grain is represented as a percentile and plotted

against increasing silver content. In this way, populations of

differing numbers of grains may be compared directly. Figure

3 compares silver contents of gold grains extracted from

mineralization intersected in drill core with grains extracted

from nearby alluvial sediment in the Lubuk Mandi area of the

Malaysia peninsular. The significance of the shapes of the

plots is discussed further below.

The microchemical signature of a population of gold

grains is derived from combining the chemical description of

the inclusion assemblage with the quantitative data

describing alloy composition. Broad differences between

gold formed in different geological environments are readily

apparent (8). Examples are presented in Table 2 which

summarizes the chemical characteristics of composite

populations of gold grains from Scotland and Ireland. Gold

from the Tweed headwaters and Glengaber Burn area in

Scotland is generally similar to that from Co Mayo in Ireland

both in terms of silver content and inclusion assemblage,

although approximately 200km apart, both lie within the

same geological terrain. However, gold from the Leadhills

region, also in the Scottish Southern Uplands terrain exhibits

a different signature, with a far higher proportion of arsenicbearing minerals in the inclusion assemblage. Gold from the

Grampian terrain in Scotland and Northern Ireland is

Comparison of Inclusion Assemblages from Gold Localities Table 2

Locality/ Localities Contributing No Median Inclusion Assemblage Characterized Diagnostic Minerals Within

Sample Sets to Composite Samples Grains Ag % by Mineral Class the Inclusion Assemblage

% % % % Gold hosted by

Sulphides Sulpharsenides Sulphosalts Tellurides meta sediments

Mayo, Ireland Cregganbaun Shear 233 8.9 90 6 4 2 Sphalerite, tetrahedrite

Zone, Croagh Patrick

Southern Scotland Glengaber area 335 7.0-7.8 71 17 6 0 Sphalerite, tetrahedrite

Tweed headwaters

Southern Scotland Leadhills 500 10.6 51 49 0 0

Northern Scotland Sutherland area, Cononish R., 465 9.7-28 81 19 0 4

Glengarry, West Water,

Calliacher Burn, Glen Lednock

Malaysian gold Lubuk Mandi, 1829-10.7 75 23 1 1

Yukon gold Bear, Hunker and 83 15-33 59 41 0 3

Bonanza Creeks

Australian gold Dolly Creek, Jordan River 70 3.2 -8.2 68 20 12 0

Gold associated

with volcanics

Fijian gold Waimanu River 46 5-18% 63 0 0 37 Bismuth tellurides

Central Scotland Borland Glen 50 6.4 60 0 0 40 Bismuth tellurides58 Gold Bulletin 2002 • 35/2

are unstable in an oxygenated fluvial environment and thus

would decompose if deformation of the gold allows contact

with air and water (17). However, for transport distances

typical of alluvial gold from Britain and Ireland the host grains

provide effective barriers to decomposition of inclusions as

witnessed by the wide range of mineral species which are

unstable in oxygenated surficial environments (eg galena,

pyrrhotite, and chalcopyrite).

The formation of mineral inclusions by mechanical

embedding into the grain surface is another mechanism by

which the signature of alluvial gold could be altered.

However, the typical size of inclusions (2-20µm) is much

smaller than grains of the same minerals in the host

mineralization and such grains would be particularly unstable

in oxygenated fluvial environments. If mechanical

incorporation of mineral grains into gold in the alluvial

environment was prevalent, the most common mineral in

alluvial sediment, generally quartz, would be expected to be

most abundant and yet in most of the 20,000+ alluvial gold

grains studied it is absent.

A population of gold grains can be augmented or even

replaced by the addition of authigenic gold, that is gold

precipitating from solution in the surficial environment. This

can result in the formation of additional discrete gold grains

or coatings on the original grains. The degree to which

authigenic gold contributes to a population of gold grains is

governed by the prevailing climate and geomorphology of

the location (18). Authigenic gold may be important in some

tropical environments where chemically aggressive

groundwaters can alter the profile of gold distribution over

long periods of time. This process is important in Australia

(19), Ghana (20, 21) and various other tropical localities

referred to by Nichol et al (18). Complete re-mobilization of

gold will obliterate the textures and chemical signature of the

original hypogene gold, however textures characteristic of

authigenic gold such as crystalline form or extensive

modification around the grain core are identifiable by SEM

methods (40). Consequently an evaluation of the degree of

modification to the population of alluvial gold is possible as

part of the analytical procedure.

Studies of gold from Britain and Ireland (8, 9), North

America (5 – 7), and New Zealand (41) failed to identify any

evidence for augmentation of the alluvial population by

authigenic processes, and the internal characteristics of the

gold grains were considered consistent with those from the

hypogene source. These results give a strong indication that

for alluvial gold in temperate climates the contribution of

authigenic gold is negligible, although many grains exhibit

gold rich rims typically to a thickness of about 10 µm, which

have been attributed to a process of silver depletion (22).

In studies of gold from Fiji (23), Zimbabwe (24), Malaysia

(25), and the Victoria goldfield in Australia, the full range of

generally of higher silver content than gold from the

Southern Uplands and exhibits a tellurium signature in the

inclusion assemblage absent in gold found to the south.

The technique is highly reproducible in terms of both

sampling and analysis (9). The microchemical signature of

gold from an area in the Leadhills district of Southern

Scotland is independent of the field sampler, the date of

collection and the exact location of the sampling site. In

addition, the same microchemical signature was obtained

using very similar analytical facilities but different analytical

procedures in two separate laboratories.

Factors Modifying the Microchemical

Signature of Alluvial Gold

The success of the microchemical characterization technique

as an exploration tool depends on a close relationship

between the microchemical signature of alluvial gold and the

characteristics of the source mineralization. There are two

ways in which the microchemical signature of a population of

gold grains can be altered by post-liberation processes in the

secondary environment: either by modification within the

fluvial environment or by the addition of authigenic gold, and

both are considered below.

The opaque mineral inclusion assemblage may become

more difficult to identify with increased transport in the

fluvial environment as most of the opaque mineral inclusions

20 40 60 80 100

Lubuk Mandi core 1

Lubuk Mandi core 2

Lubuk Mandi alluvial

% Ag

Cumulative %

Figure 3

Comparison of silver contents of gold grains from drill core and nearby

alluvial sediment Lubuk Mandi, MalaysiaGold Bulletin 2002 • 35/2 59

Relationship between Microchemical

Signature of Alluvial Gold and Source

Bedrock Gold

The close correspondence between minerals coexisting with

hypogene gold in bedrock and the nature of inclusions within

alluvial gold from nearby rivers was reported for all nine Irish

and Scottish sites studied by Chapman et al (8) and for gold

from Malaysia (27). In each case there is good agreement

between the inclusions in alluvial gold, inclusions in gold

extracted from bedrock (5 examples) and the nature of

minerals coexisting with hypogene gold in the bedrock

mineralization revealed by mineralogical study. Minor

differences probably reflect the presence of additional

sources of mineralization in the drainage catchment.

Populations of gold grains from both drill core of the

bedrock source and alluvial sediment some 1-2 kms downhill

in the Lubuk Mandi area of Malaysia are compared in Figure 3

(data from Henney et al, reference 25). The gold extracted

from the core exhibits two main silver contents (8.5% and

10.5%) both of which are well within the range of silver

contents of the alluvial gold. Such a simple relationship

between bedrock gold and adjacent alluvial gold indicates

that the mineralization is probably relatively uniform in

composition with the alluvial site close to the source.

There are commonly differences between the

composition of gold grains extracted from hand specimens

20 40 60 80 100

Soil sample

Table concentrate

Alluvial

Mine specimen

% Ag

Cumulative %

Figure 4

Comparison of silver contents of gold grains extracted from a hand

specimen of ore from the ‘C’ mine in Zimbabwe, a mine table

concentrate, a soil sample and nearby alluvial sediment

inclusions found in alluvial gold from Britain have been

observed, showing that preservation of the original features

of gold grains as they pass into the fluvial environment is also

possible in other climatic regions. However in some cases

there is evidence for some post-depositional alteration of

gold grains. Plate 4 shows a gold grain from the Mazowe area

of Zimbabwe which shows the alteration of primary silver-rich

gold to pure gold. Gold from Hunker Creek in the Klondike

district of North America differs from grains from other

Klondike localities studied by the authors in that they exhibit

relatively large (up to 200 µm) rims typically containing up to

5% silver (Plate 5) with single or multiple cores containing 15-

25% silver. Knight et al (26) also observed similar features in

gold from this locality and interpreted the texture as

indicative of silver depletion rather than the deposition of

new authigenic gold.

Plate 4

Alluvial gold grain from Mazowe area of Zimbabwe showing alteration

of primary silver-rich gold (pale yellow) to secondary pure gold (deep

yellow)

Plate 5

Back-scattered electron image of alluvial gold grain from Hunker Creek,

Yukon showing irregular rim and patches (pale) of gold-rich gold

around electrum60 Gold Bulletin 2002 • 35/2

Gold grains from Hunker and Bear Creeks exhibit similar

ranges of silver contents (Figure 5) with small numbers of low

silver grains possibly representing separate populations,

whereas the Bonanza Creek sample shows a completely

different type of plot with a break in slope suggesting two

contributing populations. The composition of Bonanza Creek

gold is consistent with its derivation from two known lode-gold

sources upstream (Lone Star and Pioneer), while that from

Hunker and Bear Creeks is far closer to the composition of gold

from the Virgin Lode in the Bear Creek area. This geographical

correlation implies that gold in the Klondyke valleys is generally

close to source. Knight et al (7) reported that the vein

mineralogy of lodes throughout the region is dominated by

base metal sulphides and arsenopyrite and during this study

inclusions of galena, pyrite, chalcopyrite and arsenopyrite were

observed in alluvial gold with silver contents of over 20% which

suggests that this type of gold is derived from a number of

related sources of mineralization. An inclusion of hessite

((Ag,Au)2Te) was recorded in the population of low-silver (15%

Ag) gold from Bonanza Creek which raises the possibility of an

additional unrecorded bedrock source to that from the Lone

Star lode. These data suggest that a more detailed knowledge

of the inclusion signature of the alluvial gold in the region in

addition to the gold alloy composition would greatly help in

establishing the relationship between bedrock and alluvial gold.

The potential to correlate the inclusion assemblage of alluvial

gold grains with their range of silver and other metal contents

Hunker Creek

Bonanza Creek

Beer Creek

Lone Star Load,

Bonanza Creek

area

Virgin Lode,

Bear Creek area

Single analysis for

Pioneer Lode

Bonanza Creek

area

Data from

Knight et al

(1994)

% Ag

Cumulative %

0 2 0 40 60 80 100

Figure 5

Comparison of silver contents of alluvial gold grains from different

localities in the Klondike, Yukon

of ore and from mine shaking tables. At the ‘C’ mine

Mberengwa district near Zvishvane in Zimbabwe, gold occurs

in shear zone hosted quartz veins. Coarse, easily visible gold

is rare and analysis of grains extracted from a sample

provided by the mine manager (24, 28) shows little variation

in composition about a 10% Ag level (Figure 4). Most of the

gold in the producing veins is much finer grained and a

sample of the production from the shaking table showed a

much greater range of silver contents (10-20% Ag, Figure 4).

Similarly, Chapman et al (9) who showed that gold from hand

specimens of ore from the Gwynfynydd Gold Mine in North

Wales are not necessarily representative of the gold from

throughout the mineralization.

Gold grains were also extracted from a soil sample

collected from the area above the main vein system at the ‘C’

mine and from alluvial sediment in a small river approximately

2 km from the mine, though this had some other small mines

also within its catchment. The soil gold shows two populations

(Figure 4), a small one with silver around 10% and thus similar

to the hand specimen, and a larger population with between

20 and 30% silver, outside the range in the shaking table

sample of deep mine ore. The alluvial sample shows multiple

populations that reflect all the samples collected at the mine,

a dominant group which corresponds to the mine shaking

table gold, and a minor population with higher silver which is

similar to that shown by the soil sample. The alluvial sample

thus provides a representative sample of gold grains in

environments where there is a considerable range in

composition of the native gold in the mineralization.

In some areas where alluvial gold is of significant economic

interest the precise relationship between placer gold and the

source mineralization may be more difficult to establish. Knight

et al (7, 26) undertook an extensive study of the chemical

composition of both alluvial and lode gold from the Klondike

district in an attempt to clarify the discrepancy between the

amount of alluvial gold won from the region with the potential

of known lode gold occurrences. They determined the

chemical composition of the cores of 2,700 gold grains in terms

of silver, mercury and copper content and were able to

characterize the populations of alluvial gold in the various

valleys according to the characteristics of different types of

known lode gold. In addition they predicted that some types of

gold in the alluvial populations were derived from unknown

bedrock sources. In the present study, a more modest number

of alluvial gold grains from this region have been analysed, but

inclusion assemblages have been identified in addition to

concentrations of the minor alloying elements. Figure 5 shows

silver plots for alluvial gold obtained from Bonanza Creek,

Hunker Creek and Bear Creek together with information on the

composition of local lode gold from Knight et al (26). Further

comparative information is presented in Table 3 which shows

close agreement between the two data sets.Gold Bulletin 2002 • 35/2 61

Northern Ireland, distinct silver-rich populations can be

recognized on cumulative silver content plots (8, 14). In each

case an inclusion suite in the silver-rich population is completely

different from inclusions in the lower silver populations.

A similar scenario with potentially important economic

significance is provided by alluvial gold grains from the

Waimanu River in Fiji. Two distinct populations can be

recognized on the basis of silver content (Figure 6) which also

differ in their inclusion suites (23). The lower silver population

(< 13% Ag) is probably derived from a porphyry-type source

which is known within the catchment. The higher silver

population contains a range of telluride inclusions which are

absent in the lower silver gold but are similar to those in gold

from a mine table concentrate from the Emperor gold mine.

As the range in silver content is also similar it is possible that

this type of alluvial gold in the Waimanu River is derived from

undiscovered alkali-epithermal style mineralization similar to

that worked at the Emperor gold mine.

Applications of the Technique to

Exploration for Gold Mineralization

The Search for a Specific Style of Gold Mineralization

Alluvial gold is widespread in Northern Ireland and in one part

the local geology suggested the presence of skarn

mineralization (9). However, interpretation of the

microchemical signatures of alluvial gold samples from the

area shows that all the gold grains correspond to the two

major types of mineralization further south, neither of which

is of the skarn type. In addition, the microchemical signatures

is important in the identification of multiple sources. Whilst

there may be overlap in silver contents between two

contributing types, in most cases the inclusion assemblages are

distinct. In alluvial gold from Shortcleugh Water, Leadhills,

Scotland and from the River Bann in the Mourne Mountains of

Whole

population

Grains

containing

Te-bearing

Inclusions

% Ag

Cumulative %

0 2 0 40 60 80 100

Figure 6

Two populations of alluvial gold grains from the Waimanu River, Fiji

revealed by silver content and associated inclusions

Table 3 Comparison of Gold Samples from The Klondike district, Yukon Territory

Locality %Ag Range %Hg Inclusion Assemblage*

%Sulphides %Sulpharsenides %Tellurides

Bonanza Creek

Alluvial * 1. 12-16 1. 0 1. 80 1. 0 1. 20

2. 20-46 2. 0 2. 67 2. 33 2. 0

Alluvial

14-24 0.3

Lodes

Lone Star 1. 14-17 0.008

2. 14-19 0.016

Pioneer 23.4 0

Bear Creek

Alluvial * 20-36 0.37 100 0 0

Alluvial

Upper: 30-420.39

Lower 28-36 0.49

Lodes

Virgin 28-34 0.54

Hunker Creek

Alluvial * 28-38 0.2 38 62 0

Alluvial

18-28 0.03

* Data from present study

Data from Knight et al (reference 26)62 Gold Bulletin 2002 • 35/2

micrometres thick. Second, microanalysis of the cores of the

grains showed that there were large differences in silver

content of gold between the two areas (Figure 7).

The gold from Veletanga is largely high-silver gold while

that from Perros Bravos has much less silver. A division at 11%

Ag separates roughly 90% of the gold from the sites,

indicating that there are two different populations of gold.

There are also differences in the spread of values within each

site; Veletanga has a relatively narrow range with most in the

range 12-17% Ag, while Perros Bravos shows a much more

even spread of compositions.

There are also differences between the suites of microinclusions in the gold from the two areas. This can be

demonstrated by comparing the composition of the host

gold for the main varieties of inclusion on cumulative

frequency curves (Figure 8). This demonstrates that most of

the inclusion types, such as pyrite, galena and pyrrhotite occur

in both types of gold but bismuth-rich tellurides and copper

minerals are confined to the Perros Bravos type of gold.

The two different types of gold in the Estero Hondo

alluvial gold mine are associated with different parts of the

catchment area. A drilling programme showed that there

was a significant skarn deposit in the Veletanga area but a

bedrock source was not located in the Perros Bravos area. It

was concluded that this source had probably largely been

eroded away, which would also account for its greatest

abundance in the lowermost gravels of the mine.

suggest that the alluvial gold was derived from a series of

small sources of mineralization of these types, and not from

glacially dispersed material derived from a large source

further south. This result emphasizes that even a negative

result is of potential value in an exploration context, as

microchemical signatures of the alluvial gold allow informed

evaluation of the prospect relatively cheaply and quickly at an

early stage in the exploration.

Identification of Multiple Sources of Different Types

of Gold within a Region

The Estero Hondo alluvial gold mine is situated in western

Ecuador, at the foot of the Andes in Catopaxi province (30).

The mine was operated by Odin Mining in the 1990’s who

also carried out exploration for bedrock gold deposits in the

area. Odin Mining geologists had noticed that the alluvial

gold on one side of the valley (Valetanga) tended to be bright

and angular, whilst that on the other side (Perros Bravos)

tended to be more rounded, darker and duller. Odin Mining

wanted to know if this indicated two possible sources of

bedrock gold or if it was due to processes acting in the

secondary environment.

Styles et al (30) carried out a gold characterization study

of samples from the Estero Hondo catchment area. This

study showed first that the differences in the appearance of

the alluvial gold were not due to variable amounts of

secondary alteration as for both populations this was

restricted to thin rims around the grains generally only a few

% Ag

Cumulative %

0 2 0 40 60 80

type1

Veletanga

type2

Perros

Bravos

100

Figure 7

Comparison of silver contents of alluvial gold from two areas in the

Ximena area of Ecuador

Perros Bravos

galena

hessite

Bi-rich tellurides

Veletanga

pyrrhotite

altaite

pyrite

Cu minerals

type 1

type 2

20 0 40 60 80 100

Figure 8

Comparison of silver contents of alluvial gold grains from the Ximena

area of Ecuador containing different types of inclusionsGold Bulletin 2002 • 35/2 63

gold grains from the area show complex patterns of chemical

compositional variation (Plate 6). Palladium occurs as a minor

alloy within the gold up to 11.9%, as inclusions of palladiumbearing minerals such as potarite, (PdHg), within gold grains

and as discrete palladium-bearing minerals such as goldbearing potarite (15). The inclusion suites are dominated by

selenide minerals which are very common in gold grains from

some localities. Sulphide and sulpharsenide inclusions are

absent. In some cases it was possible to deduce from the

compositional variation how a grain had crystallized. The gold

grain chemistry and inclusion types seemed inconsistent with

transport of gold as a bisulphide complex and deposition

within the stability field of sulphide minerals as is postulated for

the majority of types of gold mineralization. However, the

observed features and mineralogy of the grains were

consistent with transport of gold and palladium as chloride

complexes in a solution of high Eh, similar to a model proposed

for the formation of the Coronation Hill U-Au deposits of

Northern Territory, Australia (32). In the case of the Devon gold

mineralization, Leake et al (15) proposed that precipitation of

gold and palladium would take place when Eh was reduced but

still in conditions too oxidising for the precipitation of sulphides

but allowing selenides to form (33, 34).

Although red beds of Permian age with which such

oxidizing solutions were likely to be associated are absent

from all but coastal areas of the South Hams district,

geological reasoning suggested that the original interface

between these rocks and the underlying Devonian rocks was

not far above the present erosion surface. The mineralization

which provided the source of the alluvial gold was thought to

Identification of Alluvial Gold that has been

Dispersed Comparatively Far from its Source

Most of Britain and Ireland has been glaciated. This has had

the effect of enhancing erosion in upland areas and causing

deposition of glacial sediments in more lowland areas,

particularly in valleys at the margins of uplands masses which

housed large glaciers carrying ice from a number of sources.

In the Leadhills region of Scotland, the Snar Water originates

in high ground and flows northwards into a basin filled with

glacial debris which it is currently eroding. Alluvial gold grains

from the upper part of the river are similar to those from the

rest of the Leadhills region and are locally derived. The gold

from the lower reaches of the river differs markedly from that

further upstream. The inclusions in typical Leadhills gold are

predominantly arsenopyrite, pyrite, pyrrhotite and base metal

sulphides with rare cobaltite and gersdorffite. In contrast, the

gold from the lower Snar contains a much greater range of

inclusions including platinum minerals, bismuth tellurides and

copper and other selenides which are not recorded

elsewhere in the Leadhills region (9, 31). In addition, the gold

shows a greater range of silver contents and there are several

grains with copper contents well in excess of those found in

the typical Leadhills gold. The range of gold compositions

and inclusions is not consistent with one single geological

environment of mineralization. It would appear that the

glacial till contains a range of different types of gold derived

from unknown sources, some probably from a considerable

distance away from the Leadhills region. An equally wide

range of gold grain compositions and inclusion type is also

recorded at a site in the upper Tweed basin about 15 km east

of Leadhills where the river is also excavating a basin of glacial

till down to bedrock (9).

A regional study of auriferous drainage in Zimbabwe

included alluvial gold from the Sebakwe river. The population

of 41 grains contains at least 17 different types of opaque

inclusion (28), including, a range of complex copper

minerals, bismuth tellurides and silver minerals, in addition to

pyrite and base metal sulphides. The large variety of mineral

inclusions is significantly greater than normally associated

with gold formed during a single mineralizing event and

suggests that either a single source containing multiple

phases of mineralization or a number of discrete sources

contribute to the alluvial gold in the drainage sediment. The

latter is considered more likely given the large size of the river

and the presence of a number of gold mines with contrasting

mineralogy within its catchment.

Development of a Model for Mineralization Controls

and a Strategy for Exploration

Identification of the inclusion signature of alluvial gold grains

from the South Hams district of Devon, England, proved crucial

in working out a model to explain its origin (15). Many of the

Plate 6

Microchemical map showing distribution of gold in dendritic alluvial

gold/platinoid grain from South Devon. Rainbow scale represents

increasing gold concentration64 Gold Bulletin 2002 • 35/2

originate by reaction of the gold-bearing oxidizing

solutions circulating within the Permian red bed sequence

and the more reduced rocks below or by reaction with

more reduced solutions associated with these rocks. This

model accounted for decrease in the intensity of

mineralization with depth in structures within the

Devonian strata and also for the complex pattern of

growth zonation in many grains which reflects successive

pulses of mineralizing fluids with differing physical and

chemical properties. In particular the rapid precipitation of

the gold due to abrupt changes in Eh also explained the

highly crystalline and dendritic nature of many of the gold

grains (Plates 3, 6). On the basis of this model, exploration

effort was switched to the Crediton Trough further north in

Devon which is filled with Permian red beds and associated

alkaline basalt and lamprophyric lavas and where there are

many kilometres of sub-cropping contact between the

Permian red beds and underlying more reduced

Carboniferous rocks. It was predicted that gold of the same

type as that found in South Devon should also be in the

Crediton Trough and at its contacts with the surrounding

rocks. Subsequent exploration showed this to be the case

(35) and interpretation of the inclusion assemblage in the

alluvial gold from this area suggested an association with

the alkali basalts which was subsequently proved as a result

of drilling (36).

On the basis of these results, further exploration was

carried out in Scotland in and around Permian red bed basins

containing alkali basalts and this resulted in the discovery of

widespread alluvial gold with palladium enrichment and a

suite of selenide mineral inclusions generally similar to those

found in Devon gold (29).

Regional Surveys of Alluvial Gold

Alluvial gold has been studied from 314 sites in Britain and

Ireland and a regional pattern of the distribution of different

types and the nature of the potential controls of the source

mineralization has been established (8), or is in the process

of being established. It is clear that, although there are

several types of gold present, each can be equated with a

particular geological environment. Moreover, similar gold

is associated with similar geological environments even

when geographically separated. Such a comprehensive

data base is potentially of use, in association with other

information in focusing exploration on a particular type of

gold deemed to have been derived from mineralization

with most potential to be of economic interest. However,

this would only be applicable to mineralization containing

grains of native gold large enough to be isolated and

mounted for analysis.

Though the scale of investigations into alluvial gold from

other countries is minor compared with that of Great Britain

and Ireland, it is clear that comparable data can be obtained.

Microchemical signatures of alluvial gold from Palaeozoic

metasedimentary sequences in Malaysia, south east Australia

and the Klondyke region of Canada are presented in Tables 2

and 3 and are similar to one of the types identified in Britain

and Ireland. The inclusion suite is dominated by base metal

sulphides and sulpharsenides with minor contributions from

either antimony or tellurium-bearing minerals. In contrast,

the inclusion assemblage found in gold from Fiji (Table 2),

(derived from alkali epithermal and porphyry mineralization)

is distinctly different, but similar to that of gold from Borland

Glen in Central Scotland, which is associated with a volcanic

sequence of intermediate composition (37).

Conclusions

1. Different styles of gold mineralization produce gold grains

with different microchemical signatures.

2. The microchemical signature of a population of alluvial

gold grains reflects the mineralogy of the source

mineralization.

3. The technique of microchemical characterization permits

assessment on the nature of source mineralization even

before sources are discovered.

4. Interpretation of the microchemical signature(s) of alluvial

gold help focus attention on types considered to be

derived from mineralization with most potential

economic importance. Information can be obtained at an

early stage in the exploration process.

5. Extensive use of microchemical signatures has provided

new insight into the origins of alluvial gold and the

controls of source mineralization throughout Britain and

Ireland. The same approach has been successfully applied

to problems of origin of alluvial gold in more scattered

areas in many other parts of the world.

Acknowledgements

Professor Bruce Yardley of the School of Earth Sciences, the

University of Leeds is thanked for providing access to

analytical facilities. We are also indebted to Mr Peter Gower

and Mr John Krenc who kindly donated samples of Australian

and North American gold from their private collections, and

without whose assistance this paper would not have been

possible. We would also like to thank the three anonymous

referees for their helpful suggestions during the review

process. This paper is published by permission of the Director,

British Geological Survey (NERC). Gold Bulletin 2002 • 35/2 65

About the authors

Rob Chapman graduated as a Minerals Engineer and worked

on the gold mines in South Africa before returning to the UK

to study for a PhD. He joined the University of Leeds in 1990

and began collaboration with Bob Leake on the

characterization of indigenous gold in 1994. He remains an

enthusiastic gold prospector.

Bob Leake graduated from Durham University in 1965

and obtained a PhD after working at the Universities of Oslo

and Durham. He joined the British Geological Survey in

1968 and until retirement in 1997 worked mainly on the

Mineral Reconnaissance Programme and, since 1987,

mostly on gold.

Mike Styles graduated from Swansea University in 1971

and obtained his PhD from Manchester University for regional

studies in northern Norway. He joined the British Geological

Survey in 1976 and has worked for the Mineralogy and

Petrology section on a wide range of mineralization projects.

He is in charge of the BGS electron microprobe laboratories

and started the development of the gold characterization

technique in collaboration with Bob Leake in 1989. Since

then he has applied the technique to many studies of gold

from a wide range of localities.

References

1 R.W. Boyle, ‘The Geochemistry of Gold and its Deposits’. Geological

Survey of Canada, Bulletin 280, 1979

2 D.E. Roberts and G.R.T. Hudson, Economic Geology, 1983, 78 (5), 799

3 ‘Gold Metallogeny and Exploration’ ed. R.P. Foster, Blackie and Son,

1990, 432pp

4 J.W. Hedenquist, E. Izawa, A. Arribas and N.C. White, ‘Epithermal Gold

Deposits: Styles, Characteristics and Exploration’, Resource Geology

Special Publication Number 1, Society of Resource Geology, 1996

5 J.S. Loen, Economic Geology, 1994, 89, 91

6 I.M. Lange and T. Gignoux, Economic Geology, 1999, 94, 375

7 J.B. Knight, J.K. Mortensen and S.R. Morison, Economic Geology, 1999,

94, 649

8 R.J. Chapman, R.C. Leake, N.R. Moles, G. Earls, C. Cooper, K. Harrington

and R. Berzins, Economic Geology, 2000, 95, 1753

9 R.J. Chapman, R.C. Leake and N.R. Moles, J. Geochemical Exploration,

2000, 71, 241

10 C.R.M. Butt and H. Zeegers, ‘Handbook of Exploration Geochemistry,

Vol.4, Elsevier, Amsterdam, 1992, 607pp.

11 M. Shidlowski, Mineralium Deposita, 1968, 3, 344

12 R. Saager, Mineralium Deposita, 1969, 4, 93

13 G.A. Desborough, Economic Geology, 1970, 65, 304

14 R.C. Leake, R.J. Chapman, D.J. Bland, E. Condliffe and M.T. Styles, Trans.

Institution of Mining and Metallurgy (Section B Applied Earth Sciences),

1997, 106, B85

15 R.C. Leake, D.J. Bland, M.T. Styles and D.G. Cameron, Trans. Inst. Mining

and Metallurgy (Section B, Applied Earth Sciences), 1991, 100, B159

16 R.J. Chapman, R.C. Leake and J. Floyd, Scottish Journal of Geology,

2000, 36(2), 165

17 J.S. Loen, Exploration Mining Geology, 1995, 4, 335

18 I. Nichol, M. Hale and W.K. Fletcher, ‘Drainage Geochemistry in Gold

Exploration, in ‘Handbook of Exploration Geochemistry’, 1994, ed,

G.J.S. Govett, Vol. 6, ed. M. Hale and J.A. Plant, Elsevier, 1994, p 499

19 L.M. Lawrance and B.J. Griffin, Mineralium Deposita, 1994, 29, 391

20 R.J. Bowell, Mineralogical Magazine, 1992, 56, 545

21 R.J. Bowell, R.P. Foster and P. Gize, Economic Geology, 1993, 98, 999

22 J.C. Groen, J.R. Craig and J.D. Rimstidt, Canadian Mineralogist, 1990, 12,

293

23 J. Naden and P.J. Henney, British Geological Survey Technical Report

WC/95/41, 1995

24 M.T. Styles, P.D. Wetton and D.J. Bland, British Geological Survey,

Technical Report WC/95/5, 1995

25 P.J. Henney, M.T. Styles, D.J. Bland and P.D.Wetton, British Geological

Survey, Technical Report WC/94/21, 1994

26 J.B. Knight, J.K. Mortensen and S.R. Morison, ‘Shape and Composition

of Lode and Placer Gold from the Klondike District, Yukon, Canada:

Exploration and Geological Services Division, Yukon, Indian and

Northern Affairs, Canada, Bulletin 3, 1994, 142pp

27 R.C. Leake, M.T. Styles, D.J. Bland, P.J. Henney, P.D. Wetton and J. Naden,

British Geological Survey Technical Report WC/95/22, 1995

28 M.T. Styles, British Geological Survey, Technical Report WC/95/38, 1995

29 R.C. Leake, D.G. Cameron, D.J. Bland, M.T. Styles and N. Fortey, ‘The

Potential for Gold Mineralization in the British Permian and Triassic Red

Beds and Their Contact with Underlying Rocks’, British Geological

Survey Mineral Reconnaissance Report 144, 1997

30 M.T. Styles, R.C. Leake, P.J. Henney and D.J. Bland, British Geological

Survey, Technical Report WC/96/13, 1996

31 R.C. Leake, R.J. Chapman, D.J. Bland, P. Stone, D.G. Cameron and M.T.

Styles, J. Exploration Geochemistry, 1998, 63, 7

32 T.P. Mernagh, C.A. Heinrich, J.F. Leckie, D.P. Carville, D.J. Gilbert, R.K.

Valenta and L.Wyborn, Economic Geology, 1994, 85(5) 1053

33 J.E. Hulbert, Open File Report, Geol. Surv. Canada 1440, 1988, 148pp

34 S. Grigore, S.E. Kesler, and E.J. Essene, Economic Geology, 1997, 92 (4),

468

35 D.G. Cameron, R.C. Leake, R.C. Scrivener, D.J. Bland, and S.H. Marsh,

British Geological Survey, Mineral Reconnaissance Program Report, No.

133, 1994

36 R.C. Scrivener, M.T. Shaw, T.J. Shepherd, C. Smith and M.T. Styles,

Proceedings of the Ussher Society, 1999, 9, 396

37 J.S. Coats, M.H. Shaw, M.J. Gallagher, M. Armstrong, P.C. Greenwood,

B.C. Chacksfield, J.P. Williamson and N.J. Fortey, British Geological Survey

Mineral Reconaissance Programme Report Number 116, 1991, 93pp

38 S.C. Dominy, R.F.G. Phelps, M.A. Bussell, and C.L. Guard, British Mining,

1996, 57, 70

39 G.N. Phillips and M.J. Hughes, Ore Geology Reviews, 1996, 11(5), 255

40 J.H. Youngson and D. Craw, Economic Geology, 1995, 90, 731

41 J.H. Youngson and D. Craw, Economic Geology, 19