SWIR_SEG

SEG   NE  N EWSLETTER  OCTOBER 1999

HSOCIETY OF ECONOMIC GEOLOGISTSH

NUMBER 39

Alteration Mapping in Exploration: Application of Short-Wave Infrared (SWIR) Spectroscopy ANNE J. B. THOMPSON (SEG 1990)

PHOEBE L. HAUFF

AUDREY J. ROBITAILLE

PETRASCIENCE CONSULTANTS INC. 3995 W. 24TH AVENUE VANCOUVER, B.C. • CANADA V6S 1M1 EMAIL, [email protected] [email protected]

SPECTRAL INTERNATIONAL INC. P.O. BOX 1027 ARVADA, COLORADO 80001

PETRASCIENCE CONSULTANTS INC. 3995 W. 24TH AVENUE VANCOUVER, B.C. • CANADA V6S 1M1

ABSTRACT  Alt  A lt er at io n mi n er al as se mb la ge s a re im p or ta n t to t he understanding of and exploration for hydrothermal ore deposits. Conventional mapping tools may not identify fine-grained minerals or define important important compositional compositional variations. variations. Field portable portable short wave infrared (SWIR) spectrometers spectromete rs solve some of these problems and provide a valuable tool for evaluating the distribution of  alteration assemblages. assemblages. Spectrometers such as the PIMA-II allow allow rapid identification of minerals and mineral-specific variations at a field base. Mineral assemblages, assemblages, integrated integrated with other exploration exploration data, are then used to target drill holes and guide regional exploration exploration programs. Data collection collection must be systematical systematically  ly  organized and carried out by a trained operator. Analysis of data sets requires the use of spectral reference libraries from different geological environments and may be aided in some cases by  computer data processing processing packages. packages. Integration Integration of results with field field observations, petrography, and X-ray diffraction analysis is necessary  for complete complete evaluation. evaluation. The PIMA (portable (portable infrared infrared mineral mineral analyzer) has been used successfully in the high-sulfidation epithermal, low-sulfidation epithermal, volcanogenic massive sulfide (VMS) and intrusion-rel intrusion-related ated environments. environments. Case studies from from these systems demonstrate the ability to rapidly acquire and process SWIR  data and produce produce drill logs logs and maps. maps. The resulting resulting information information is critical for targeting.

INTRODUCTION Field portable SWIR spectrometers are becoming increasingly  important to exploration. Spectrometers typically are are employed to determine the mineralogy of altered rocks and hence assist in classifying ore systems, identifying alteration patterns, and locating

ore. In addition to to its early use in remote remote sensing, sensing, development of  the PIMA-II in 1991 allowed direct use of SWIR on rocks, greatly  enhancing enhancing its practical practical application application to exploratio exploration. n. SWIR  spectroscopy detects minerals such as phyllosilicates, clays, carbonates, and selected sulfates and is also sensitive to variations in individual mineral species. SWIR field spectrometers are used in numerous deposit environments, including high- and low-sulfidation epithermal, porphyry, mesothermal, sediment-hosted gold and copper, uranium,  VMS , and an d kimb ki mber erlilite te depo de posit sit s (Tab (T able le 1, page pa ge 16). 16 ). In addi ad ditition on,, spectrometers aid regolith mapping, both for determination of  bedrock composition and for differentiation of residual and transported transported regolith. regolith. Publications Publications on the the results of SWIR  SWIR  spectroscopy are sparse, reflecting the confidential nature of most companies ’ programs and the lack of academic work applied to field mapping. mapping. A selection selection of recent papers papers and abstracts, abstracts, however, however, highlight the work that is currently underway: underway: Stewart and Kamprad, 1997, and Shen et al., 1999 (regolith mapping); Zhang et al., 1998 (uranium); Passos and de Souza Filho, 1999 (Archean greenstone); Denniss et al., 1999, and Huston et al., 1999 (VMS); Martinez-Alonso et al., 1999, and Kruse and Hauff, 1991 (epithermal clays); and Crowley, 1996 and 1999 (evaporites). SWIR field spectrometers fill an important gap in exploration data by helping to map alteration consistently throughout a mineralized system. Determining alteration mineralogy routinely during during an exploration program aids rapid evaluation and therefore increases ef ficiency. Alteration Mapping 

Determination of the type and distribution of  ... alteration minerals is a routine part of exploration to page  16

SEG

16

...

from 1

S W I R

S P E C T R O S C O P Y , 

NEWSLETTER

CONT .

for hydrothermal mineral deposits and is useful in the assessment of  exploration properties and the construction of deposit models. Typically, alteration maps are based on macroscopic field observations, supported by limited petrographic or X-ray diffraction studies. Alteration studies at the deposit scale are limited, or reliant on detailed but necessarily restricted sample suites. Lithogeochemistry is used in some environments to evaluate alteration but only works well where lithologies and their petrology  are well understood. Lithogeochemistry is difficult to apply to extensive areas of clay alteration, where protoliths are hard to identify during the exploration program. Fine-grained alteration minerals commonly are grouped as “ argillic ” or “ phyllic ”   (Thompson and Thompson, 1996). Such descriptions ignore the mineralogy and lose valuable information regarding the nature of the alteration. The importance of using minerals and mineral assemblages was noted by Rose and Burt (1979) and subsequent authors, but this approach is not always applied during exploration. Classification of alteration by mineralogy  involves field observations that may be aided by SWIR spectroscopy  (Table 2). The use of SWIR spectrometers at a field base allows mineralogy to be mapped or placed on cross sections. The resultant interpretation can be applied in real time to guide drilling, and ultimately can be integrated with other data to develop targets, models, and regional guides. Field observations must be made in a careful and systematic manner. Care is needed in determining the relationship among minerals prior to assigning them to a single assemblage or interpreting their relationship to other types of alteration. A series of  steps should be followed in order to make realistic interpretations of  the observed hydrothermal alteration. These steps are: 1. Determine the minerals present, based on field observations; 2. Determine their distribution at the outcrop and hand specimen scales; 3. Employ SWIR analysis carefully, analyzing a variety of locations on each sample and using systematic sampling techniques;

Nº 39 • OCTOBER ’99

4. Use the above data to establish the relationships among the main minerals; 5. Outline the distribution at the map scale; 6. Use petrography of selected samples to further define relationships of minerals; 7. Augment with X-ray diffraction (XRD) analysis if necessary; 8. Use scanning electron microscope (SEM) with energy dispersive system (EDS) to determine variations in individual minerals and assist with interpretation in fine-grained material; 9. Refine and reevaluate, continually, the interpretation and integrate the results with other geologic, geochemical, and geophysical data sets. SWIR analysis aids exploration from regional to property scales. For example, in complex zoned intrusive systems, alteration mineralogy determined routinely during mapping helps to define  vertical and horizontal zoning and related ore environments. Within each environment, alteration mineralogy can define local zoning, providing vectors to mineralization. SWIR spectroscopy is most helpful where alteration mineralogy is not easily identified in hand specimen because of grain size or weathering. Even where field mapping of alteration minerals is effective, SWIR spectroscopy  allows recognition of subtle mineralogic and compositional  variations; these can be important in locating ore.

REFLECTANCE SPECTROSCOPY Reflectance spectroscopy is an analytical technique used by  chemists and mineralogists since the early 1900s, with infrared data on minerals published between 1905 and 1910 by W.W. Coblentz of  the U.S. Bureau of Standards. Commercially available infrared spectrophotometers in the mid-1940s led to increased use of the technique for mineralogy. Early reviews of mineral spectra were published by Lyon (1962) and Moenke (1962). Farmer (1974) published a comprehensive book on theoretical and practical aspects and Marel and Beutelspacher (1976) compiled clay minerals. Kodama (1985) published spectra of minerals typically found in soils, including numerous hydroxides, oxides, phyllosilicates, carbonates, and sulfates.

Table 1. Examples of the Use of SWIR in Exploration

Mineral Identification

Alteration Interpretation

Exploration Application

 Alunite

Advanced argillic

— High-sulfidation environment, and zoning around high sulfidation — Steam-heated zones in low sulfidation

Dickite

Advanced argillic

— Zoning around high sulfidation — Sediment-hosted Au, with mineralization

Kaolinite

Advanced argillic and

— High-sulfidation

weathered rock

— Sediment-hosted Au, zoning

Dickite, pyrophyllite, diaspore

Advanced argillic

— Depth estimation

Chlorite

Propylitic, chloritic

— VMS zoning — Uranium zoning

Illite/smectite

Argillic

— High and low sulfidation, zoning — Uranium, zoning

Carbonate

Carbonate

— Mesothermal, zoning

SEG

OCTOBER ’99 • Nº 39

NEWSLETTER

Mineral spectra in the SWIR were first compiled by Hunt and Salisbury (1970, 1971), and Hunt et al. (1971a, b, c). Hunt’s database provided a basic reference for infrared active minerals in the SWIR  region that is still in active use. The work was expanded on by  Clark et al. (1990). Hauff (1993) published a commercial reference library. Workers at the Jet Propulsion Laboratory also added to the references available (Grove et al., 1992). Field Portable Spectrometers 

Geologists in the remote-sensing community drove the initial development of fi eld portable SWIR spectrometers that were particularly useful to the mineral exploration industry. The ability to easily produce laboratory-quality data in field situations enhanced

the ability to field check Landsat Thematic Mapper imagery. Several field spectrometers are available. These include the Geophysical Environmental Research, Inc. (GER-IRIS), Analytical Spectral Devices (ASD-FieldSpec) and Integrated Spectronics Pty. Ltd. (PIMA). The GER and ASD instruments provide data in the visible, near, and shortwave infrared wavelengths. The instruments are field portable, but require use of solar illumination. Early published papers include documentation of the GER instrument (Marsh and McKeon, 1983). The instrument was used in field checking of airborne spectroradiometer data in the Oatman district (epithermal veins), Arizona. Hunt and Ashley (1979) and Crowley (1984) linked to page  18 . . . SWIR spectroscopy to alteration mapping.

Table 2. Summary of Infrared-Active Minerals, with Distinctive Spectra in the SWIR

Environment of formation

Standard terminology

SWIR active mineral assemblage (key minerals are in bold)

Intrusion-related

Potassic (biotite-rich), K silicate, biotitic

Biotite (phlogopite), actinolite, sericite, chlorite, epidote, muscovite, anhydrite

Sodic, sodic-calcic

Actinolite , clinopyroxene (diopside), chlorite, epidote, scapolite

 

Phyllic, sericitic

Sericite (muscovite-illite), chlorite, anhydrite

 

Intermediate argillic, sericitechlorite-clay (SCC), argillic  Advanced argillic

 

Greisen Skarn

Low-sulfidation epithermal

Pyrophyllite , sericite, diaspore, alunite, topaz, tourmaline, dumortierite, zunyite Clinopyroxene, wollastonite, actinolite-tremolite, vesuvianite, epidote, serpentinite-talc, calcite, chlorite, illite-smectite, nontronite

 

Chlorite, epidote, calcite , actinolite, sericite, clay

 

Advanced argillic— acid sulphate

Kaolinite, dickite, alunite , diaspore, pyrophyllite, zunyite

 Argillic, intermediate argillic

Kaolinite, dickite, montmorillonite , illite-smectite

Propylitic

Calcite, chlorite, epidote, sericite, clay

“ Adularia” — sericite, sericitic,

argillic

Mesothermal

Sericite (illite-smectite), chlorite, kaolinite (dickite), montmorillonite, calcite , epidote

Topaz, muscovite, tourmaline

 

Propylitic

High-sulfidation epithermal

 

  Sericite, illite-smectite , kaolinite, chalcedony, opal, montmorillonite, calcite, dolomite

   Advanced argillic — acid-sulphate (steam-heated)

Kaolinite, alunite , cristobalite (opal, chalcedony), jarosite

Propylitic, zeolitic

Calcite, epidote, wairakite, chlorite, illite-smectite, montmorillonite

Carbonate  

Chloritic Biotitic

Calcite, ankerite, dolomite, muscovite (Cr-/V-rich), chlorite

 

Chlorite, muscovite, actinolite Biotite , chlorite

 

Kaolinite, dickite , illite

Sediment-hosted gold

Argillic

 Volcanogenic massive sulfide

Sericitic

 

Sericite, chlorite, chloritoid

Chloritic

 

Chlorite, sericite, biotite

Carbonate Sediment-hosted massive sulfide

Tourmalinite

Dolomite, siderite, ankerite, calcite, sericite, chlorite

   

17 

Tourmaline, muscovite

Carbonate

Ankerite, siderite, calcite , muscovite

Sericitic

Sericite, chlorite

 Albitic

Chlorite, muscovite, biotite

Minerals are grouped by assemblages of alteration minerals, and keyed to commonly used terminology; Complete assemblages are in  Thompson and Thompson (1996) 

SEG

18

...

from 17 

S W I R

S P E C T R O S C O P Y , 

NEWSLETTER

CONT .

The most widely used portable instrument in exploration is the PIMA, which collects data only in the SWIR region. The PIMA-II is a commercial field instrument manufactured by Integrated Spectronics Pty. Ltd. in Australia. The instrument has an internal light source, allowing collection of laboratory-quality data in the field by direct illumination of the rock sample. In addition, internal calibration results in reliable spectra not subject to variability due to the conditions under which they were measured. The instrument is capable of measuring a variety of sample types, including rocks, chips, core, powders, and liquids. An analysis typically takes less than 30 seconds. PIMA dominates current usage in the industry for the purpose of alteration mapping. Several PIMA-II instruments  were used in the collection of the data discussed in this paper.  All instruments require training for effective use, both in the interpretation of results and in instrument operation. Lack of training can result in broken and malfunctioning instruments or worse, misinterpretation of data. The limitations of the technique must be understood in order to utilize fully a powerful tool. Integration of  spectral data with geologic, geochemical, and geophysical information is also critical. Field use of short-wave infrared spectrometers has increased dramatically in the last fi ve years. The increased application of the tool is the result of several developments during the twentieth century. The milestones in the development of infrared spectroscopy for minerals include:

Nº 39 • OCTOBER ’99

Short-wave infrared spectroscopy detects the energy generated by vibrations within molecular bonds. These bonds have bending and stretching modes within the 1,300- to 2,500-nm region of the electromagnetic spectrum. The observed absorption features are manifestations of first and second overtones and combination tones of fundamental modes that occur in the mid-infrared region. SWIR is particularly sensitive to certain molecules and radicals, including OH, H2O, NH4, CO3, and cation-OH bonds such as Al-OH, Mg-OH, and Fe-OH. The positions of the features in the spectrum and their characteristic shapes are a function of the molecular bonds present in the mineral. Variations in chemical composition may be detected as the wavelength positions of features shift consistently with elemental substitution. SWIR spectroscopy is partly sensitive to crystallinity variations, but may not detect primary changes in the lattice structure. A typical spectrum consists of several absorption features. Figure 1 illustrates the various aspects of an absorption feature, including wavelength position, depth and width (full-height, half-width maximum). The outline of the hull or continuum is also shown.

• Early documentation (1905 – 1910); • Laboratory use, expansion of mineral reference databases (1940 – 1985); • Development of field portable instruments (1978 – 1991); • Real time processing of data (late 1980s); • Commercial availability of portable instrument with internal light

source, PIMA (1991); • Continued expansion of mineral reference data sets (1990s); • Dissemination of case histories and examples of the application of SWIR spectroscopy to mineral exploration (1990s); • Use of PCs in the field, allowing rapid data interpretation (1995 – present); • Heightened interest due to use of airborne hyperspectral scanners and increasingly sophisticated data processing software (1998 – present). SWIR spectrometers also are now employed in numerous other capacities beyond exploration. In particular, they are useful in mineral processing control procedures and evaluation of leach piles and tailings dumps. Continued development of applications will lead to uses in other environmental applications and the geotechnical fields. SWIR Spectroscopy 

Remote-sensing geologists use a variety of bands within the electromagnetic spectrum, including the visible-near infrared (VNIR), short-wave infrared (SWIR), and mid-infrared (MIR). Field portable instruments detect in the SWIR region, which is sensitive to molecular changes, and also in the VNIR, where color variations and changes in elemental oxidation states (e.g., iron and chromium) are observed. VNIR, however, does not relate directly to composition.

Figure 1. Detail of feature in kaolinite SWIR spectrum collected with PIMA-II spectrometer. The hull, feature depths, position and the full width half-maximum (FWHM) are shown.

OCTOBER ’99 • Nº 39

SEG

NEWSLETTER

19

Minerals can be distinguished not only on the basis of distinctive features and wavelength positions, but also by the character of  the profile (without hull subtraction). Examples of  common alteration minerals are shown in Figure 2. Mineral identification is based on wavelength positions, intensity and shape of  absorption troughs, and the overall shape of the entire spectrum. The short-wave infrared  wa vel en gt h re gi on is no t suitable for most anhydrous silicates. In addition, it is difficult to identify minerals present in amounts less than 5%, unless the sample is a simple mixture with quartz and the mineral is highly  reflective. Infrared reflectivity varies between mineral species. In mixtures of  infrared-active minerals, the dominant and typically most F i g u r e 2 . Stack plot with examples of spectra reflective mineral is easily  characteristic of individual minerals. Reflectance values are offset for clarity. Examples are from the SPECMIN™ identified; however, as a database. general rule, 10% or more of  a mineral must be present for positive identification.  Where low reflectance minerals are present, recognition may require 20% or more of  the mineral in the sample (e.g., carbonate, chlorite). Mineral chemistry : Variations in mineral chemistry  are typically detected by  shifts in wavelength posiFigure 4. Examples of K-, Na-, Ca-, and NH 4-bearing alunite group tions or changes to the hull minerals. An inset shows the positions of distinctive features for K and Ca in the Ca-dominant sample. Elements present in this sample were shape. The presence of iron Figure 3. Example of Fe- and Mg-bearing clinochlore. Note the steep slope on the Fe-rich sample from 1,300 confirmed by EDS analysis (scanning electron microscope). Examples in most minerals results in a nm to 1,900 nm. are from the SPECMIN™ database. strong positive slope from 1,300 to 1,900 nm. A commay be required in order to assign an observed variation to a parison of the spectra from Fe-rich and Mg-rich clinochlore is shown change in chemical composition. Mineral composition variations are in Figure 3. Chemical variation in the carbonate group of minerals is best evaluated from monomineralic samples; however, it may be gauged by a shift in the position of the major feature as a function possible to define variations in some mineral mixtures. of the cation present. The dominant feature varies widely, including magnesite (Mg) at 2,300 nm, dolomite (Mg, Ca) at 2,320 nm, calcite DATA COLLECTION AND ANALYSIS (Ca) at 2,330 nm and rhodochrosite (Mn) at 2,360 nm. Variations in alunite-group mineral chemistry are manifested by shifts in the Data Collection  1,480-nm position, with values ranging from ~1,461 (NH4), to ~1,478 Understanding the variables that affect spectra is critical to the nm (pure K) to ~1,496 (Na) to 1,510 nm (Ca). Examples of all four interpretation of spectral data sets. These variables include grain spectra are shown in Figure 4. Depending on the quality of  size, transparency, sulfide content, overall to page  20 . . . reference spectra, petrographic, SEM or electron microprobe data reflectivity, water content, heavy element

SEG

20

...

from 19

S W I R

S P E C T R O S C O P Y , 

NEWSLETTER

CONT .

content, contaminants (e.g., oil, organic material), orientation of  minerals (e.g., micas), and mineral mixtures. Good data collection procedures that minimize the effect of these variables should be adopted. A minimum of 2 spectra per sample is necessary, both for reproducibility of data and to test for heterogeneity of the samples. By analyzing the groundmass, veins, phenocrysts, vug infill, fracture coatings, and weathered surfaces, the data gatherer can identify  several minerals in one sample. Clear descriptions of basic observations — e.g., color, texture, veining or fractures, vein envelopes, and weathering state — are important for high-quality  spectral interpretation. Sample types and sample processing may also affect the spectra. Hand specimens, powders, rock chips, liquids, and reject samples can all be analyzed, with some minor variations observed in the spectra. Samples that have been pulverized (e.g., analytical pulps) commonly yield extremely degraded patterns, where many of the spectra are similar in appearance. Pulps are typically made with ring mills that generate heat during the crushing process. Clay analysis, in particular, will be inaccurate if the structure of the mineral changes with heating. Standard XRD analysis for clays is also done  without the use of ring mills for the same reason. Instrument stability must be considered in evaluating spectra, particularly in variable field conditions. Without calibration,  wave length positions will shift as the instrument heats up or is moved. Analysis of a standard kaolinite with the PIMA-II shows a systematic wavelength shift of 2 nm downward as the instrument heats up from 22° to 44°C. Although small (within the spectral resolution of the PIMA spectrometer), this shift highlights the need for consistent recalibration of the instrument. Good laboratory  practice also includes use of standards and the saving of calibration files as references for future checks on instrument drift. Some  wo rke rs ha ve rep ort ed 2 nm va ri at io ns in as few as 10 measurements. SWIR spectroscopy is a useful tool for identifying minerals in individual samples; however, its greatest power comes from consistent collection of data in a systematic manner. Sample intervals may need to be as small as 1 to 2 m to evaluate gradients in alteration mineralogy and define boundaries when used for drill core logging or detailed traverses. Once the basic variations are described, spacing can be widened, depending on the area to be covered and the goal of the survey. For example, core logging may  be done on 5- to 10-m spacing, whereas mapping may be widened to 50 or even 100 m. Closely spaced sampling typically produces the most useful information. For mapping, the sample locations may  be laid out on a grid pattern and include soils or may be tied to outcrop patterns. Data processed and evaluated concurrent with mapping can have a direct impact on an exploration program. Data Analysis and Processing  Mineral identification is based on the use of reference data sets,  which are empirical records of each mineral’s characteristic spectra.

 Visual observation of a group of mineral spectra will quickly show  variations based on numerous factors, including mineral chemistry, temperature, and mode of formation (reflected in crystallinity), and other subtle changes. The user’s greatest asset, then, is a welldefined and large reference collection or spectral database created from samples representative of a wide variety of deposit environments and occurrences. Experienced users are ultimately 

Nº 39 • OCTOBER ’99

able to identify many spectra characteristic of minerals by eye, as data collection is carried out. Even novice users can quickly learn the basic minerals important in their study area. Only a few minerals typically are required to characterize an area, which also realistically  allows for the tracking of variability within a single mineral species. Reference databases are extremely useful in refining the ability of  data processing software to provide automatic identification.  Automatic identi fication may be helpful when working with large data sets on well-defined areas. In order to achieve high-quality  results, variations at the deposit scale must be observed and recorded by the user, using reference data sets created for that deposit. Deposit or region specific data sets appear to be critical in obtaining reliable results from attempts at unmixing (identifying mineral mixtures) using algorithms. Identification of complex mixtures will require geologic context, user experience, and establishment of reference samples with additional information, e.g., petrography, XRD, and SEM analysis, but may be difficult with currently available algorithms. Data processing software allows the subtraction of the hull (see Fig. 1), typically followed by extraction of feature positions, intensities, and widths. There are various methods to extract diagnostic derivative spectra from those that are measured. Two common methods are the hull quotient, which is a “rubber band” method of removing the effects of variable background slopes, and the first derivative that removes the effects of background by  emphasizing changes in response. A variety of software packages are available commercially, with the most flexible being those that allow the importing of data from a variety of sources (spectrometers or scanner data). Care must be taken when extracting data to ensure that the data is treated in the same manner, e.g., feature positions all based on the hull quotient, or on the first derivative. In some cases, use of a single feature position, depth, or ratio of  depths of two features may provide broad outlines of alteration zones. This style of data extraction with a computer program can be done extremely quickly; however, the data must be carefully  evaluated by an experienced user to confirm that the comparison is of similar material, with comparable mineral assemblages and features. Since wavelength positions for various minerals overlap, faulty results which are based on a single feature may be produced. For example, the Al-OH feature at ± 2,200 nm may represent alunite, pyrophyllite, kaolinite, dickite, illite, mixed-layer illite/smectite, smectite, or muscovite, which all obviously have different implications in terms of deposit modeling. Contouring of such data must also be carefully carried out. Individual analyses may falsely   weight the data, resulting in spurious features. Mineral percentages:  A common objective of SWIR analysis is to determine not only the minerals present in a sample but also their relative abundance. Many software programs attempt to provide mineral unmixing as part of the package. This task is challenging due to the lack of knowledge regarding the absorption coef ficients for the molecular bonds detectable in the SWIR range of the electromagnetic spectrum. The spectral data indicate that minerals apparently are not present in linear mixed configurations, but rather as a function of these unknown absorption coefficients. The intensities of the absorption features, therefore, cannot be used as a 1:1 correlation to relate directly with the amount of mineral present. For instance, an iron chlorite will absorb more energy at its diagnostic wavelengths and reflect less back to the detector than an aluminum-bearing mineral, a muscovite, which more accurately 

OCTOBER ’99 • Nº 39

SEG

NEWSLETTER

reflects the absorbed energies. In addition, there is a matrix effect problem when non-infrared active minerals are also present and absorbing, but not reflecting the excitation energy. In cases where the absorption coef ficients are likely to be similar for the minerals in question, the resulting spectrum can be treated essentially as a linear mixture. The key to accurate results lies in building calibration files from the sample suite under investigation. Monomineralic end members must be chosen for the models to succeed. The accuracy of the mixing algorithm may be as good as 4%; however, this will vary depending on the software (algorithm), the materials in the mixture, and their relative abundance. Instrumentation is also a limiting factor in producing accurate results. PIMA-II has approximately a 5- to 6-nm resolution and samples are collected in 2-nm steps. This over-sampling is done to improve the reproducibility of the method; however, it does not necessarily improve the accuracy and leads to an artificial perception of a 2-nm resolution. Therefore, the limit of the method for resolving the wavelength positions remains between 5 and 6 nm.

21

 Alteratio n was def ine d bas ed on mineral assemb lages, and the alteration patterns determined by spectral analysis were used to help predict zones of mineralization. It was found that dickite increased in areas associated with gold mineralization, thus providing a mineralogical pointer to areas favorable for gold. Figure 5 is a drill log from the Alumbre zone, showing the relationship of the alteration to mineralization. Spectral analyses were carried out every  2 to 3 m down hole. This hole was subsequently deepened and additional mineralization associated with to page  22 . . . dickite, alunite, and vuggy quartz was found.

CASE STUDIES The following case studies illustrate the use of SWIR in exploration. All of the analyses were collected with PIMA-II spectrometers. These examples emphasize the use of the technique as a mapping tool, for integration with other data types. The examples include alteration maps, detailed drill logs, and integration of geochemistry, petrography and spectral data. The mineral identifications and variations in feature positions are shown to define alteration zones and provide vectors toward mineralization. Blind application of digital data, however, may lead to false results based on inadequate sampling or misinterpretation of spectral  variations. High-Sulfidation Epithermal Gold Deposits  The general characteristics of high-sulfidation deposits are well

known and are summarized in Arribas (1995), who includes numerous examples of deposits around the world. These deposits are known for their extremely fine grained alteration minerals and typically homogeneous appearance. Minerals that are infrared active and form in these environments are shown in Table 2. Alteration mapped in the field typically relies first on varying degrees of silicic alteration ranging from leached, vuggy quartz to zones of  replacement quartz. Beyond the quartz-dominated areas, however, the alteration assemblages are commonly mapped simply as advanced argillic and argillic alteration during exploration programs. Identification of individual minerals, some of which are critical to zoning patterns, is extremely difficult. Use of SWIR spectroscopy  allows the major alteration minerals to be easily and rapidly  identified. Virgen : A comprehensive alteration study of the Virgen property   was com ple ted for Git ennes Exp lor ation Inc ., Vancouver. The  Virgen property is a gold prospect located 180 km east of Trujillo, Peru. Cretaceous sedimentary rocks and Tertiary andesites host mineralization. The aim of the work was to determine alteration zone patterns with respect to mineralization using the PIMA-II spectrometer. Data were collected from available drill holes, hand samples, and road cuts. A total of 22 drill holes were analyzed and over 900 spectra were obtained across the property. Alteration minerals observed using spectral analysis include alunite, dickite, pyrophyllite, diaspore, kaolinite, smectite, illite, and quartz.

Figure 5. Drill log from the Al umbre zone, Virgen deposit, Peru, showing variations in alteration mineral assemblage, lithology, and mineralization. Dickite in the advanced argillic alteration was used as an indicator for mineralization.

SEG

22

...

from 21

S W I R

S P E C T R O S C O P Y , 

NEWSLETTER

CONT .

Pamel : An alteration study of the Pamel prospect in the Western

Cordillera of Peru was conducted for Candente Resource Corp. The property is at an early stage of exploration. Geologists employed at the property selected hand and reject samples from the geochemical sampling program and submitted these to Vancouver for analysis  with a PIMA-II spectrometer. The results from approximately 128 samples were integrated with geologic information and outcrop patterns. Based on three days work, combined with the previous geologic data, an alteration map was created for the property. The results clearly showed distinct alteration zones and helped delineate zones of interest (Figure 6). The alteration varies, from silicification to alunite-dickite, to alunite-kaolinite, to kaolinite dominant, and outward to sericite, illite, and chlorite. Small amounts of diaspore, topaz, and tourmaline were also noted locally. A detailed study was conducted in the western portion of the map area. Samples from the

Nº 39 • OCTOBER ’99

zone contain two to three minerals in a single spectrum. Examples of these spectra are shown in Figure 7. The alteration and style of  mineralization is consistent with a high-sulfidation epithermal to magmatic-hydrothermal environment.

1000metres

Figure 7. Examples of spectra for mineral mixtures, from the Pamel prospect, Candente Resource Corp.

Low-Sul fid   ation Epithermal Gold Deposits   Alteration in low-sulfidation deposits is characterized by adularia

Figure 6. Alteration map based on outcrop patterns for the Pamel prospect, Candente Resource Corp.

and calcite-replacement textures within quartz veins that grade outward in the host rock to illite, illite-smectite, and illite/smectitechlorite zones. Calcite may also occur within the alteration envelope. The variation in clay alteration outward from mineralization is typically dif ficult to define in the field, but may be detected with SWIR. Zoning patterns are well described by  numerous authors, including White and Hedenquist (1990) and Sillitoe (1993). The widths of alteration envelopes vary from centimeters to meters. Zones of steam-heated, advanced argillic alteration may also cap or develop laterally from low-sulfidation

OCTOBER ’99 • Nº 39

SEG

NEWSLETTER

mineralization. Differentiation of these zones from high-sulfidation systems is critical to exploration in this environment. Table 2 lists minerals found in the low-sulfidation and steam-heated environment. Patagonia, southern Argentina: SWIR spectroscopy was carried out systematically on RC (reverse circulation) chips from a property  in Patagonia, southern Argentina. The area contains zones of highgrade mineralization associated with illite-dominant alteration. The use of spectral analysis allowed mapping of alteration patterns. Figure 8 is a drill log showing the distribution of illite, illite-gypsum, illite-smectite, and illite-chlorite. The illite-gypsum zones clearly  flank the mineralization.

23

Volcanogenic Massive Sul fid   e Deposits 

 Alteration mapping is an important aspect of VMS exploration in areas where metal distribution may provide limited information. Typically, distribution of Fe and Mg chlorite and sericite (muscovite) zones are used as a vector toward ore lenses. Clays may also be an important part of some systems. The style of alteration varies, depending on the setting of the deposit; Franklin (1993) gives several models. Lithogeochemistry is commonly employed to differentiate alteration types, but use of SWIR spectroscopy may also provide valuable, direct information regarding the alteration mineralogy. Kidd Creek: Distribution of chlorite and sericite is outlined by  Koopman et al. (1999) for the Kidd Creek deposit, western Abitibi Subprovince, Canada. Variations in proportions of the two minerals appear to reflect proximity to ore and help to outline mineralized zones. The initial work was based on field observations, X-ray  diffraction and petrographic techniques. Seventy-six of the XRD samples were analyzed with a PIMA-II spectrometer. A variety of  processing techniques were applied to the data set, including automatic mineral unmixing using commercially available software; comparison of ratios of feature depths in the Al-OH region to the FeOH (as described by Huston, 1999); and simple comparison of the data to a set of artificially (linear) mixed spectra. The linear mixed spectra used site-specific end members, and 10% increments (Figure 9, page 24). Results from this comparison were the most consistent and correlate well with the estimated mineral abundance based on peak intensities from the XRD analysis (Table 3, page 26). The attempts at semiquantification using the other techniques, however, did not produce reliable results. The effect of different reflectivity  and of mixed mineral assemblages may hinder the use of the automatic techniques in determining mineral percentages in this environment until better data processing is possible. Intrusion-Related Deposits 

Figure 8. Drill log shows the distribution of alteration with respect to gold assays. Example of low-sulfidation system, Patagonia, southern Argentina.

Many deposit types occur within the intrusion-related environment. Mapping of alteration has application to both broad and local zoning. As shown in Table 2, a wide variety of alteration minerals may be present in this environment and can be used to focus exploration on a specific target type. Red Mountain, British Columbia : A detailed study of the relationship of alteration and mineralization was completed at the Red Mountain project in northern British Columbia and is described by Rhys et al., 1995. Gold mineralization in the area is spatially  related to a porphyry Cu-Mo stockwork. Hydrothermal alteration is pervasive throughout the pre-Tertiary  rocks on Red Mountain, including all phases of the main intrusions. Several shallow-dipping alteration zones are stacked above a quartz stockwork/molybdenum zone with pervasive sericite alteration. These zones include: (1) sericite-quartz-pyrite alteration (pyritedominant), (2) chlorite-K feldspar-sericite-titanite alteration with disseminated and vein pyrrhotite, (3) brown to black tourmaline  veins, and (4) K feldspar-titanite-actinolite alteration. Anomalous gold values are associated with the transition from pyrite to pyrrhotite, and sericite to K feldspar alteration. High-grade zones are focused below areas of abundant tourmaline, in pervasive sericite alteration. Details of the alteration zones are given in Table 4 (page 26), along with a summary of major element to page  24 . . . geochemistry. The deposit is structurally 

SEG

24

...

from 23

S W I R

S P E C T R O S C O P Y , 

NEWSLETTER

CONT .

disrupted by several faults. The distinct alteration zones provide a  way to reconstruct relative locations within the system where fault boundaries are crossed in drilling. Data gathered by the PIMA clearly outlines the major alteration zones, based on the presence of key minerals for each assemblage. The results of SWIR analysis on drill core samples are compared to geochemistry and petrography in Table 4, and representative spectra from the deposit are included in Figure 10. The complete log for a drill hole from the Marc zone is shown in Figure 11. Integration  with the previous petrographic, geologic, and geochemical data

Nº 39 • OCTOBER ’99

provides a framework for application of SWIR spectroscopy to the deposit area and exploration within the district.

CONCLUSIONS SWIR spectroscopy is a tool to assist field mapping in mineral exploration. The ability to rapidly differentiate fine-grained alteration minerals in the field allows for an enhanced understanding of the property under investigation, and the results can be applied immediately to the exploration program. Further refinement of the alteration assemblages, including the use of other analytical techniques, will yield data important for the development of deposit models and regional exploration programs. Several components are critical to a successful survey. These are:

Figure 9. Artificial linear mineral mixtures determined from actual chlorite and muscovite end members for the Kidd Creek footwall rhyolite. The percentage of chlorite increases down the plot. Examples of the comparisons of real spectra with selected mineral mixtures are shown on the right. Mineral mixtures were generated with SPECWIN©, Spectral International Inc.

OCTOBER ’99 • Nº 39

SEG

NEWSLETTER

•  A syste matic, well-planned sampling program, with consistent

sample spacing depending on the purpose of the survey; •  Alteratio n mapping concur rent with a mapping or a drilling

• • • •

program, to allow rapid incorporation of information and effective application of data; SWIR data collection by a trained operator with a geologic background; Use of mineral-dominant assemblages for preliminary mapping; Subsequent data processing of wavelength positions or other spectral characteristics, to further evaluate alteration; Selected petrography and XRD for complete alteration assemblages that can be related to the spectral data.

25

······················· ACKNOWLEDGMENTS ·······················

 We would like to thank the many geologists and consultants in exploration for their contributions to the use of SWIR spectroscopy. The successful application of the technique would not have been possible without them. Data in this paper was published with the permission of Jerry Blackwell (Gitennes Exploration Inc.), Joey  Freeze (Candente Resource Corp.), Mark Hannington (Geological Survey of Canada) and Jacques Houle (Royal Oak Mines) who are all thanked for their support. The paper benefited from reviews by  Noel White and Charles Tarnocai. In particular, Noel White is thanked for his enthusiasm, wellto page  26 . . . timed reminders, and critical review.

 Wit h the increased use of SWIR spectrometers in the field, distribution patterns over large areas can be delineated. The use of  field spectrometers has provided a vast new database that not only  is aiding exploration, but also will contribute ultimately to the understanding of these systems.

Figure 10. Representative spectra used to determine alteration zones at Red Mountain, British Columbia.

Figure 11. Drill log from the Marc zone, Red Mountain, British Columbia. Log shows major alteration zones, lithologies, and the location of Au-Ag mineralization. The SWIR active minerals identified in each alteration zone are: A. actinolite dominant: actinolite + chlorite (Fe > Mg) + datolite + prehnite + axinite, B. tourmaline dominant: chlorite (Mg > Fe) + muscovite + schorl + axinite + calcite, C. Tourmaline stockwork: schorl + Mg chlorite + muscovite, D. auriferous pyrite-pyrrhotite stockwork: muscovite + chlorite, and E. pyrite dominant: muscovite + chlorite + clinozoisite.

SEG

26

...

from 25

S W I R

S P E C T R O S C O P Y , 

NEWSLETTER

Nº 39 • OCTOBER ’99

CONT .

Table 3. Comparison of SWIR and XRD Results for Selected Samples from the Kidd Creek Footwall Felsic Units

Spectra

Lithology

SWIR

X-ray diffraction

ah09667b

Rhyolite

Chlorite

ah09682b

Quartz porphyry

Chlorite > muscovite

Quartz

ai00510b

Quartz porphyry tuff

Chlorite

Quartz >> albite, clinochlore, calcite > muscovite

ai00557a

Felsic tuff

Chlorite > muscovite

Quartz

>>

clinochlore, albite, calcite > muscovite

ai00671b

Rhyolite

Chlorite > muscovite

Quartz

>>

clinochlore > muscovite

ai00724b

Rhyolite

Muscovite > chlorite

Quartz

>>

albite, clinochlore > muscovite

ai01511b

Quartz porphyry tuff

Chlorite > muscovite

Quartz

>>

albite > clinochlore, muscovite

ai01532b

Quartz porphyry tuff

Chlorite > muscovite

Quartz

>>

clinochlore, calcite > muscovite

ai01538a

Quartz porphyry tuff

Chlorite = muscovite > kaolinite?

Quartz

>>

clinochlore > muscovite

ai01545a

Rhyolite

Muscovite

Quartz

>>

albite, muscovite > clinochlore

ai01550b

Felsic tuff

Muscovite = chlorite

Quartz

>>

albite > clinochlore, muscovite

>>

>>

muscovite

Quartz >> clinochlore > muscovite

muscovite

>>

chlorite > kaolinite?

>>

clinochlore, muscovite

SWIR analysis identified the chlorite as Fe-bearing 

Table 4. Distribution of Alteration Zones at Red Mountain, B.C.

Alteration Thickness Zone

Geochemistry

Veins

Petrography

SWIR

 Actinolite

>150m

Na2O > 3.3%,; K 2O < 0.5%; CaO > 2.8%; Sr > 400 ppm

Chlorite + pyrite + actinolite + calcite

K feldspar + actinolite + chlorite + titanite + albite + pyrite ±

 Actinolite + chlorite (Fe>Mg) ± axinite* ± datolite ± muscovite ± prehnite

Tourmaline stockwork 

100-300m

Na2O > 3.3%; K 2O < 0.5%; CaO > 2.8%; Sr > 400 ppm

Tourmaline + pyrite + chlorite + pyrrhotite

K feldspar + chlorite + titanite + pyrite + tourmaline + pyrrhotite

Schorl (tourmaline) + chlorite (Mg) ± muscovite ± carbonate ± axinite*

Pyrrhotite

100 – 200m

Na2O > 3.3%; K 2O < 0.5%; CaO >2.8%; Sr > 400 ppm

Pyrrhotite + pyrite ± chalcopyrite ± chlorite ± calcite ± quartz ± sphalerite ± galena

K feldspar + sericite + pyrrhotite + pyrite + chlorite ± tourmaline

Muscovite + chlorite (Mg) axinite ± carbonate ± prehnite

 Auriferous Py-Po stockwork 

10-50m

Na2O < 1.5%; K 2O > 5%; high values in Au (>0.5 ppm); Ag, As, Sb and locally Cu, Zn correspond with ore zones

Pyrite ± pyrrhotite ± quartz ± chlorite

Intense sericite + pyrite; mantled by disseminated and veinlet sphalerite + pyrrhotite + pyrite

Muscovite + chlorite (Mg)

Pyrite

100-200m

Na2O < 3.5%; K 2O > 4%; CaO < 3.3%; Sr < 100 ppm

Pyrite ± calcite ± quartz ± chlorite

Sericite, pyrite ± calcite ± chlorite ± tourmaline

Muscovite + chlorite

Gypsum stockwork 

<5 – 100m

Similar to pyrite-dominant alteration

Gypsum + pyrite + calcite ± quartz

Sericite + pyrite ± quartz ± K feldspar

Gypsum*

Quartz stockwork, Mo-Cu

>200m

Cu > 300 ppm; Mo > 30 ppm; SiO2> 55% and similar values to pyrite dominant alteration for Na2O, K 2O, CaO, and Sr

Quartz + pyrite ± chlorite Sericite + quartz + pyrite ± epidote ± magnetite ± + chlorite + K feldspar ± molybdenite ± chalcopyrite epidote ± tourmaline ± magnetite ± hematite

Chlorite (Mg), muscovite ± clinozoisite*

* Denotes vein mineralogy  Zones are listed from the top down, with their associated geochemistry, vein mineralogy and alteration mineralogy (petrography and SWIR  analysis); after Rhys et al (1995) 

OCTOBER ’99 • Nº 39

SEG

NEWSLETTER

·····························  REFERENCES  ·····························

Arribas, A., Jr. , 1995, Characteristics of high-sul fidation epithermal deposits, and their relation to magmatic fluids: Mineralogical Association of Canada Short Course v. 23, p. 419 –454. Clark, R.N., King, T., Klefwa, M., Swayze, G.A. , and Vergo, N. , 1990, High spectral resolution reflectance spectroscopy of minerals: Journal of Geophysical Research, v. 95, no. B-8, p. 12,653–12,680. Crowley, J.K. , 1984, Near infrared reflectance of zunyite: Implications for field mapping and remote-sensing detection of hydrothermally altered high alumina rocks: Economic Geology, v. 79, p. 553 –557. ——1996, Mg- and K- bearing borates and associated evaporites at Eagle Borax Spring, Death Valley, Califonia: A spectroscopic exploration: Economic Geology, v. 91, p. 622 –635. ——1999, Association of magnesium-rich clay minerals and borate deposits: A potential remote sensing guide for borate mineral exploration?: International Conference on Applied Geologic Remote Sensing, 13 th, Vancouver, B.C., March 1 – 3, 1999, ERIM International Inc., Proceedings, p. I-255 –262. Denniss A.M., Colman, T. B., Cooper, D.C., Hatton, W.A. , and Shaw, M.H. , 1999, The combined use of PIMA and VULCAN technology for mineral deposit evaluation at Parys Mountain mine, Anglesey, UK: International Conference on Applied Geologic Remote Sensing, 13 th, Vancouver, B.C., March 1 – 3, 1999, ERIM International Inc., Proceedings, p. I-25 –32. Farmer, V.C., 1974, The infrared spectra of minerals: Mineralogical Society of London, Darking, Surrey, Adlard and Son Ltd., Monograph 4, 539 p. Franklin, J.M. , 1993, Volcanic-associated massive sulphide deposits: Geological Association of Canada Special Paper 40, p. 403–417. Grove, C.I. , Hook, S. , and Paylor, E.D. , 1992, Laboratory reflectance spectra of 160 minerals, 0.4 to 2.5 micrometers: Pilot Land Data System, Jet Propulsion Laboratory, Pasadena, California, JPL Publication 92-2. Hauff, P.L., 1993, SPECMIN ™  Mineral Identification System and Spectral Library, v. 1 and 2.: Arvada, Colorado, Spectral International, Inc., 600 p. Hunt, G.R. and Ashley, R.P., 1979, Spectra of altered rocks in the visible and near i nfrared: Economic Geology, v. 74, p. 1613–1629. Hunt, G.R. , and Salisbury, J.W. , 1970, Visible and near-infrared spectra of minerals and rocks: I. Silicate minerals: Modern Geology, v. 1, p. 23 –30. ——1971, Visible and near-infrared spectra of minerals and rocks: II. Carbonates: Modern Geology, v. 2, p. 283 –300. Hunt, G.R., Salisbury, J.W. , and Lenhoff, C.J., 1971a, Visible and near-infrared spectra of minerals and rocks: III. Oxides and hydroxides: Modern Geology, v. 2, p. 195 –205. ——1971b, Visible and near-infrared spectra of minerals and rocks: IV. Sulphides and sulphates: Modern Geology, v. 3, p. 1 –14. ——1971c, Visible and near-infrared spectra of minerals and rocks: VI. Additional silicates: Modern Geology, v. 4, p. 85 –106. Huston, D.L., Kamprad, J., and Brauhart, C. , 1999, Definition of high-temperature alteration zones with PIMA: An example from the Panorama VHMS district, central Pilbara Craton: Australian Geological Survey Organisation Research Newsletter, v. 30, p. 10–12. Kodama, H., 1985, Infrared spectra of minerals; reference guide to identification and characterization of minerals for the study of soils: Research Branch Agriculture Canada, Technical Bulletin 1985-1E, 197 p. Koopman, E., Hannington, M., Santaguida, F., and Cameron, B., 1999, Petrology and geochemistry of proximal hydrothermal

27 

alteration at Kidd Creek, Economic Geology Monograph 10, in press. Kruse, F.A. and Hauff, P. L. , 1991, Identi fication of illite polytype zoning in disseminated gold deposits using reflectance spectroscopy and X-ray diffraction — potential for mapping with imaging spectrometers: IEEE Transactions on Geoscience and Remote Sensing, v. 29, p. 1101 –1104. Lyon, R.J.P., 1962, Minerals in the Infrared — A critical bibliography: Stanford, California, Stanford Research Institute, 76p. Marel, H.W. and Beutelspacher, H. , 1976, Atlas of infrared spectroscopy of clay minerals and their admixtures: Amsterdam, Elsevier, 396 p. Marsh, S.E. and McKeon, J.B. , 1983, Integrated analysis of highresolution field and airborne spectroradiometer data for alteration mapping: Economic Geology, v. 78, p. 618–632. Martinez-Alonso, S.E., Goetz, A.F.H., Atkinson, W.W., Kruse, F.A. and Eberl, D.D. , 1999, Multiscale study of infrared spectra of c l ay m i n er a ls : f r om a b i n i ti o q u a nt u m c al c ul a ti o ns t o hyperspectral remote sensing: International Conference on Applied Geologic Remote Sensing, 13 th, Vancouver, B.C., March 1 – 3, 1999, ERIM International Inc., Proceedings, p. I-174 –181. Moenke, H., 1962, Spektralanalyse von mineralien und gesteinen; eine anleitung zur emissions- und absorptionsspektroskopie: Leipzig, Akademie Verlagsges. (in German). Passos, R.V., and de Souza Filho, C.R. , 1999, Spectro-mineral mapping using field spectroscopy and the geometry of hydrothermal alteration zones associated with mesothermal gold mineralization: A case study in the Brumal Deposit, Quadrilatero, Ferrifero, Brazil: International Conference on Applied Geologic Remote Sensing, 13 th, ERIM International Inc., Proceedings, p. I73–80. Rhys, D.A., Sieb, M., Frostad, S.R., Swanson, C.L., Prefontaine, M.A., Mortensen, J.K. and Smit, H.Q. , 1995, Geology and setting of the Red Mountain gold-silver deposits, northwestern British Columbia: Canadian Institute of Mining Metallurgy, Special Volume 46, p. 811 –828. Rose, A.W. and Burt, 1979, Hydrothermal alteration, in  Barnes, H.L., ed., Geochemistry of hydrothermal ore deposits: New York, Wiley-Interscience, p. 173 –235. Shen, X.C., Dunlop, A.C., and Cohen, D.R. , 1999, Characterization of residual and transported regolith profiles using the PIMA-II (abs): The Association of Exploration Geochemists, International Geochemical Exploration Symposium, 19 th,Vancouver, B.C., April 1999, Program and Abstracts, p. 137 –138. Sillitoe, R.H. , 1993, Epithermal models: genetic types, geometrical controls and shallow features: Geological Association of Canada Special Paper 40, p. 403–417. Stewart, A. and Kamprad, J. , 1997, Piercing the regolith veil; identifying parent rocks from weathered equivalents, Eastern Goldfields, Western Australia: Australian Geological Survey Organisation Research Newsletter 26, May 1997, 6 p. Thompson, A.J.B. , and Thompson, J.F.H. , 1996, Atlas of alteration: A field and petrographic guide to hydrothermal alteration minerals: Geological Association of Canada, Mineral Deposits Division, 119 p. White, N.C., and Hedenquist, J.W., 1990, Epithermal environments and styles of mineralization: Variations and their causes, and guidelines for exploration: Journal of Geochemical Exploration, v. 36, p. 445 –474. Zhang, G., Pan, Y. and Chen, N., 1998, Short wave infrared (SWIR) reflectance spectroscopic study of kaolinite-group minerals at the Athabasca basin, Saskatchewan [abs]: Geological Association of Canada and Mineralogical Association of Canada Abstract Volume, Quebec 1998, Abstracts, p. A-203. 1