OLOGI C

GE A
D

DE
SOC I E DA

C HILE
un

F
da 6

2
la serena octubre 2015 d a e n 19

Geochemistry of Hydrothermal Alteration Associations in

Porphyry Copper Deposits: Applications to
Geometallurgical Modeling
1,2,3 3 3 2 1
Brian Townley *, Rodrigo Luca , Luis López , Marcia Muñoz y Pamela Castillo
1
Departamento de Geología, Universidad de Chile, Plaza Ercilla 803, Santiago, Chile
2
Advanced Mining and Technology Center (AMTC), Universidad de Chile, Av. Tupper 2007, Edificio AMTC, Santiago, Chile
3
Geoanalytical Vision S.A., Av. Ricardo Lyon No. 3547, Ñuñoa, Santiago

* email: [email protected]

Abstract. We discuss the value of multi-element above mentioned parameters in a quantitative manner at a
geochemistry in ore modeling which has important resolution scale equal to that of sample support, being of
applications to geometallurgical processes. In mine assistance in detailed determination of ore deposit geologic
production of porphyry copper deposits much effort is models, applied to exploration and/or geo-mineral
placed on geological modeling which is the base for mine
metallurgic models.
planning, applied for predictive mineral processing
behavior, from blasting, comminution, leaching or flotation,
down to waste disposal and environmental management. The potential application of multi-element geochemistry to
High confidence geological models are commonly based on mineral characterization in porphyry copper deposits, or
geological mapping of drill core, and surface/ underground any other type of deposit, requires data processing and
exposures, combined with mineral characterization interpretation, and final calibration of geochemical
techniques (e.g., petrography, QEMSCAN®, XRD, classification parameters to mapping and mineral
spectroscopy, etc.). Unfortunately these are expensive and characterization data. Classification parameters may be
can only be applied to limited numbers of samples thus determined as element combinations (e.g. resulting from
lacking sufficient representativity. In addition, geological
multivariate statistical analysis), element ratios or
mapping always presents a degree of uncertainty, based on
qualitative and semi-quantitative estimates. The combinations of additive variables and ratios, in some
combination of information usually provides geological cases, multiples. The objective of synthetic variables is to
models that are hard to combine with geometallurgical test mass transfer processes that occur during hydrothermal
data, much less with mineral processing predictive alteration and mineralization, and later, during supergene
behavior. The advent of multi-element geochemistry, at low alteration (if applicable). As mass transfer processes
cost, has allowed many companies to employ such tool, yet introduce geochemical modifications and ion exchange in
little use is given to the data. Geochemistry reflects the rocks, ratios may commonly discriminate alteration
mineral composition of rocks, thus allowing characterization processes that individual elements do not. The
of lithology, alteration types and intensities, and
compositional variance of an individual element will
mineralization. As an example of such applications we
present a generic classification based on aqua regia usually depend on lithology, alteration and mineralization,
geochemistry. hence the sources of variance do not respond to a single or
specific process. As opposed to single elements, synthetic
Keywords: Geochemistry, Geometallurgical Modeling, variables, in particular those that include ratios, will detect
Geological Modeling, Porphyry Copper and discriminate those rocks in which one or more
elements have been added or depleted as a result of a mass
transfer process, this, respect to those ions participating in
1 Introduction equilibrium reaction exchange during hydrothermal
alteration and mineralization. Another discrimination
A rock represents a mineral aggregate, these of primary variable that may be recognized from data processing and
(lithology), endogenous secondary (hydrothermal) and/or interpretation could result from element groups associated
of secondary supergene origin, among others. Minerals are with specific hydrothermal alteration assemblages, all
constituted by chemical elements, mineral associations, be elements associated to the same specific mineral
them of lithologic, hydrothermal alteration and association, having similar variance behaviors, as
mineralization and/or of supergene origin will be determined from multivariate statistics.
represented by specific geochemical compositions and
chemical element combinations, major, minor and trace. As testing of geochemical classification criteria may be
Multi-element geochemical data base processing and crossed respect to geological mapping and/or mineral
interpretation may allow characterization and characterization, the level of confidence obtained for
determination of lithology, hydrothermal alteration (type classification can be measured, hence confidence of any
and intensity) and mineralization, and supergene alteration. specific parameter, in the ability to discriminate, may be
The previous procedures allow characterization of all evaluated numerically. It must be noted that there are no

327
 ST 7 GEOMETALURGIA

specific recipes to determine geochemical characterization compared respect to core logging, allowing detection of
parameters, as these may vary from one deposit to another, those elements and/or variables that best discriminate one
depending on combinations of host rock lithology, type of alteration respect to another. Cross comparison of
hydrothermal alteration type and intensity, mineralization, all variables will finally determine those combinations that
and superposition of multi-stage hydrothermal events. best discriminate, be it on first pass, or resulting from data
sub-classifications and iteration of discrimination on data
subsets.
2 Hydrothermal Alteration Characterization
in porphyry copper deposits: A generic
example
The procedures followed for the determination of
geochemical mineral association classification parameters
are summarized on figure 1. Data processing usually
includes full statistical univariate, population and
multivariate analysis of all valid geochemical data.
Univariate statistics, depending on data distributions, may
be carried out in arithmetic or log normal base, depending
on nature of statistical distributions. Such nature must be
established during the univariate statistical analysis.

Procedures
Rock geochemistry
Incorporates lithologic,
alteration and
mineralization (1ry & 2ry)
composition Multielement geochemical
data base processing and
interpretation
Calibration of geochemical
compositions and element
combinations to types of
Quantitative mineral
lithology, hydrothermal
characterization and
alteration types/intensity
assistance to high
and mineralization
resolution geologic ore
deposit models and/or
exploration models

Applications to Mineral
Exploration models Applications to Geo-Mineral
Metallurgic models

Figure 1. Work flow methodology followed for data processing

and interpretation and determination of geochemical mineral
association classification parameters.

Results of statistical analysis and interpretation of results

allows a first approximation for quantitative determination
and classification of mineral associations, indicating
lithology and hydrothermal environments. Discrimination
and classification parameters can then be tested against
geological mapping and any other mineral characterization
technique, allowing crossed comparison respect to any
Figure 2. Examples of element concentration comparisons for
particular geological feature, such as lithology, Al, Ca, and synthetic variable KxAl, respect to hydrothermal
hydrothermal alteration, mineralization, among others. alteration types. Pink, orange and green-brown colors represent
mostly hydrolytic alterations (quartz-sericite, sericite, sericite-
A generic example for the characterization of hydrothermal clay); green colors represent mostly chlorite and chlorite-sericite;
mineral associations is shown on figure 2, multi-element yellow represents silicic alteration; dark brown and reddish
geochemistry from an aqua regia digest ICP-MS data set. brown represent potassic alteration, the first biotite dominant, the
All individual elements and synthetic variables can be rest, K-feldspar dominant.

328
 AT 2 geología económica y recursos naturales

These concepts may be further tested by use of bi-variant

or ternary diagrams, such as those shown on figure 3 and
4. On these discrimination diagrams samples plot along
trends and fields that represent hydrothermal extremes and
superposition trends. Also, in some cases, a specific
mineral may cause a separate subset of data, parallel to a
main trend. Such case is determined on figures 3 and 4, on
which the presence of anhydrite in a portion of samples
separates a parallel alteration trend, discriminating sulfate-
rich environments from those sulfate-poor. Such
discrimination allows detailed determination of an
anhydrite sulfate roof, common in porphyry copper
systems, and of important implications in mine design,
mine planning, extraction, mineral processing and
production.

Figure 4. Example for ternary diagram data discrimination,

Mg-Fe-S. Potassic and chlorite alterations (brown and green
colors) are discriminated respect to hydrolytic alterations (quartz-
sericite, sericite-clay; orange and yellow). Trend from dark brown
to orange represents biotite dominant potassic, to K-feldspar
dominant (light brown) to phyllic (orange), a transitional trend;
on which phyllic overprint on potassic occur along the trend, as
well as chlorite-sericite and chlorite. On the right, a parallel trend
represents those samples that are anhydrite-rich, displaced respect
to the others due to elevated S provided by the sulfate.

These are some examples of geochemical discrimination

and classification parameters for hydrothermal alteration.
Study and combination of all geochemical parameters,
individual elements and synthetic, applied to massive data
sets, allows discrimination of classification extremes and
overprint trends, such as those shown on figure 5. It must
be noted that large data sets, including multi-element
geochemistry and combinations of synthetic parameters,
may allow much more discrimination criteria, this as
further data processing on multiple pass selection and
classification is conducted.

Once any amount of geochemical single and/or synthetic

parameters are defined, these, being quantitative, and
spatially defined, may be used for 2 and 3D spatial models,
using any geostatistical techniques desired, on any specific
ore deposit modeling software. If sample support for multi-
element geochemistry is similar to ore reserve estimate
Figure 3. Examples of bi-variant diagrams. Top, Mg vs. Fe, sample support, in addition to ore reserves, gangue
showing samples that plot along the biotite-chlorite trend mineralogy can be modeled with the same or similar
(brownish and green colors), clearly discriminated from those accuracy as metal concentrations and ore mineralogy. On
samples that plot along the pyrite trend representing hydrolytic further applications, if gangue mineralogy is of known
alterations (orange and yellow). Bottom, S vs. Fe, samples impacts to mineral processing, geometallurgical models
plotting along the pyrite trend (yellow-orange), discriminated may be derived as mineral characterization determinations
from potassic and chlorite (browns-greens). Bottom trend is
are crossed to metallurgical tests.
anhydrite absent; samples above the pyrite line are anhydrite-rich.

329
 ST 7 GEOMETALURGIA

Sericite-quartz +- clays

Chlorite-Sericite Potassic only

Potassic biotite

Figure 5. Example for massive data classification of

hydrothermal environments (data frequency density), based on
ScxV vs. (Al+K)/(Na+Ca+Mg).

3 Discussion and Conclusions

Once samples are classified by lithology, hydrothermal

alteration type and intensity, and mineralization, these may
be studied for spatial distribution. As a result, these
geological features will be mapped on a resolution
equivalent to sample support. These results may be applied
to 2 and 3D ore deposit models for exploration and/or geo-
mineral metallurgical models. Models may be based of
thousands to tens of thousands of samples with multi-
element geochemical analysis, which at present is usually
analyzed by digestion ICP combination techniques, but that
may soon migrate to analytical XRF, as these techniques
allow ever better resolution and access to larger suites of
elements. Based on the same concepts it is feasible to
determine elements that discriminate lithologic types,
especially immobile elements (e.g. Ti, Zr, Th, REE).
Classifications based on immobile elements allow
determination of lithologies, independent of alteration. It is
important to note that these are only tools to assist geology,
and by no means replace the basics, core logging and
mapping!

330