Introduction to Geochemistry Workshop.pp.pptx

análisis de datos geoquímicos

1. Comparison of assay methods How “complete” is a 4 acid digest? What are the advantages and disadvantages of a 4 acid digest compared with a Li-borate fusion? Quartz-topaz-andalusite-fluorite breccia, Akulla , Sweden Scott Halley Consulting Geochemist, Mineral Mapping Pty Ltd, Hawley Beach, Tasmania, 7307. Email: [email protected] Web: www.scotthalley.com.au

2. Digest Methods Aqua regia Mixture of nitric acid and hydrochloric acid in a molar ratio of 1:3. Aqua regia readily dissolves many sulfide, oxide and carbonate minerals, and chlorite, as well as retaining mercury, a particularly volatile element. Four acid digest HNO 3 , HClO 4 , HF, HCl. Quantitatively dissolve nearly all minerals. Li Borate Fusion Flux for melting the sample into a glass which is then acid soluble. Strongest fusion technique in order to fully digest zircons, barite, rare earth oxides, and tin, tungsten, niobium and tantalum minerals.

3. What is a 4A digestion? Acids used are nitric HNO 3 , perchloric HClO 4 , hydrofluoric HF, and hydrochloric HCl Sample is digested in nitric and perchloric first and heated. HF is added, solution is heated and evaporated to some extent. Max temperature over 150°C Lastly, HCl is added. No two lab companies perform their sample decomposition methods exactly the same way.

4. 4A Digestion: Compromises Not fully dissolved: Zircon! Monazite, sphene Cassiterite, rutile, xenotime, barite, corundum, chromite Scheelite Columbite -tantalite Massive/refractory sulfides Not fully retained: Si is fumed off as SiF 6 All Hg is lost due to the temperature of digestion Semi-volatiles (As, Se, Sb, Tl, Ge, etc.) may be lost depending on exact methodology Some elements may re-precipitate as insoluble salts depending on exact methodology

5. “Research Quality Assays” By combining several methods into one package, a complete sample characterization is obtained. These packages combine whole rock analysis, trace elements by fusion, aqua regia digestion for the volatile trace elements, carbon and sulfur by combustion analysis, and several detection limit options for the base metals. Costs are typically around $100/sample This assay package is not routinely used because of the high unit cost. ALS Schedule of Services and Fees, Australia 2019

6. Exploration Assays In exploration it is common to use a 4 acid digest and a combination of ICP-AES and MS assay methods. This gives a broad suite of major elements, trace elements and metals, with exceptionally low detection limits Costs are typically around $42/sample; cheap enough to be used routinely. Resource Definition Assays In a resource drill out where the very low detection limits are not important, it is common to use a 4 acid digest with just ICP-AES assays. Costs are typically around $22/sample. ALS Schedule of Services and Fees, Australia 2019

7. Comparison of Methods The following presentation compares 4 acid digest ICP results with LiBorate fusion assays, These plots compare standard immobile trace element scatterplots that are used to categorise rock compositions and plots that are used to track magmatic fractionation. Also presented are plots of 4 acid digest results versus Complete Characterization results for elements that might be volatilised during digestion, or elements that are difficult to dissolve.

8. Comparison of Methods 4 acid digest ICP-MS These data are from calc-alkaline volcanics in a VMS province. The rocks include organic-rich black shale, siltstone, pumice breccia and a porphyry unit. These plots show a classification of the assay data from plots of Sc vs Th, Ti , V, Zr , Nb and Cr. Note these do not form distinct populations because of the heterogeneous nature of the rocks. These assays are from a 4 acid digest ICP-MS method. The same samples were also assayed with a LiBorate fusion ICP-MS. The same plots are presented in the next slide:

9. Comparison of Methods “Complete Characterization” These are immobile trace element plots from LiBorate fusion ICP-MS. These assays are more accurate than the 4 acid digest since it is a more complete digestion process, but the precision of the analyses is poorer.

10. Comparison of Methods There are some specific plots we can use to track fractional crystallization of particular minerals. In this suite of samples, the pumice breccias show evidence of fractional crystallization of plagioclase, magnetite and eventually biotite and zircon. V/Sc vs Sc is a plot designed to show the degree of fractional crystallization of magnetite; decreasing V/Sc with decreasing Sc (increasing SiO2). This is very obvious in the 4 acid digest data, and not at all obvious on the Li Borate data. 4 acid digest ICP-MS Complete Characterization Magnetite fractionation Organometallic V

11. Comparison of Methods There are some specific plots we can use to track fractional crystallization of particular minerals. In this suite of samples, the pumice breccias show evidence of fractional crystallization of plagioclase, magnetite and eventually biotite and zircon. Ta/ Nb vs Ti is a plot designed to show the degree of fractional crystallization of biotite; Increasing Ta/ Nb with decreasing Ti . This is very obvious in the 4 acid digest data, and scattered from the Li Borate data. 4 acid digest ICP-MS Complete Characterization Biotite fractionation

12. Comparison of Methods; Aluminium Li-borate fusion produces more accurate analyses of the major elements. 4-acid digest produces higher precision analyses of the trace elements because it results in lower inference effects during analysis. Choose a method that is fit for purpose! Can you accept lower accuracy on the major elements in order to get greater information from the trace elements? If too much liquid is retained during the HF/heating stage, Al is precipitated as a fluoride complex. Al2O3_pct _LiBorate fusion Al_pct 4 acid digest

13. Comparison of Methods; Potassium Test how “complete” the 4 acid digest is by plotting 4acid vs Li Borate results from the same samples. With a 4 acid digest, some K may reprecipitate as potassium perchlorate. K2O_pct _LiBorate fusion K_pct 4 acid digest

14. Comparison of Methods; Titanium Ti is a real test of the efficiency of the 4 acid digestion. Zr will usually underreport quite significantly even with the best 4 acid procedure. If a 4 acid digest is done well, >90% of the Ti will dissolve. Ti is more difficult to dissolve in samples that have had advanced argillic alteration. If Al under-reports, then it is likely there will be significant under-reporting of Ti . TiO2_pct _LiBorate fusion Ti_pct 4 acid digest

15. Comparison of Methods; Arsenic Arsenic under-reporting at higher values; volatilization during digest? As_ppm _aqua regia ICP-MS As_ppm 4 acid digest

16. Comparison of Methods; Antimony It is common for antimony to be incompletely dissolved in aqua-regia. Sb_ppm _aqua regia ICP-MS Sb_ppm 4 acid digest

17. Conclusions Good quality 4 acid digest ICP-MS trace element data has significant advantages over Li Borate fusion “research quality” assays. With the exception of Zr , in most data sets the digestion is very close to total. The major elements are measured with less accuracy than a Li Borate fusion, but the results are good enough for quantitative mineralogy calculations. The calculated mineralogy estimates are significantly improved by adding SiO2 analyses by portable XRF to the assay suite.

18. Comparison of Methods ALS has developed a proprietary ICP-MS method with a modified sample introduction method and a collision/reaction cell. This method has an order of magnitude reduction in detection limits and increase in precision due to fewer oxide and polyatomic interferences, and low carry over effects. The difference between this and the standard ICP-MS is as great as the difference between ICP-MS and LiBorate fusion ICP-MS. ALS is now working to introduce the same proprietary method for LiBorate samples, which will give complete digest AND high precision. Costs are typically around $50/sample. ALS Schedule of Services and Fees, Australia 2019

19. Comparison of Methods Sample fractional crystallization diagrams from ME-MS61L with ultra-low detection limit 4 acid digest ICP-AES/MS . The effect of the higher precision in these assays is obvious. Hf versus Zr, ME-MS61L Hf versus Zr, Li-borate fusion

20. BUT,…. Not all 4 acid digests are the same!! Each assay company has different procedures for the digestion process . The following example shows samples analysed twice, by Li-borate fusion ICP-AES, and 4 acid digest ICP-AES/MS. Some laboratories DO get close to a complete dissolution with a 4 acid digest, but some DO NOT.

21. Potassium There can be significant under-reporting of K by 4 acid digest K for values above 4% due to precipitation of potassium perchlorate 4 acid digest ICP-AES vs Li-borate fusion K_pct 4 acid digest K 2 O_pct _LiBorate fusion

22. Aluminium There can be under-reporting of Al. This is a very common outcome where acid digests leave too much water during the heating process before addition of HCl. Al is retained as an insoluble fluoride. 4 acid digest ICP-AES vs Li-borate fusion Al_pct 4 acid digest Al 2 O 3 _pct _LiBorate fusion

23. Conclusions Whether or not 4 acid digest ICP-AES/MS analyses are fit for purpose depends on which laboratory provided the data. In a 4 acid digest ICP-AES/MS package, the major elements are assayed by ICP-AES. Therefore ICP-AES assays from a reputable lab will be sufficient for the calculated mineralogy estimates (at about 20% of the cost of complete characterization assays). SiO2 by pXRF can be added for around $3.50 sample to significantly improve the estimations. If/when the Li Borate fusion ultra-low detection limit method is available it will provide superior quality results than the current research quality assay at about half the unit cost.

24. Multielement Geochemistry data validation Just because your data passes the validation tests in Acquire or Datashed doesn’t mean it is correct! Try these simple tools as an extra check on your data quality. Scott Halley Consulting Geochemist, Mineral Mapping Pty Ltd, Hawley Beach, Tasmania, 7307. Email: [email protected] Web: www.scotthalley.com.au Native Sulfur in advanced argillic alteration, Puren Sur, La Coipa

25. Check the Elements and units in your database. ioGAS has useful data validation tools. Set the element name and units (ppb, ppm, %) for all data columns.

26. Check the Elements and units in your database. Does your data base have a mixture of ppb, ppm and pct values? It is very easy to make a series of scatter plots or cumulative frequency plots. It is obvious when the wrong units are in a table, or when elements have been imported to the wrong columns.

27. Have some idea of what ranges to expect. When you are using multielement data, you should have an idea about what ranges of analytical results are reasonable, and which are not. Typical Ranges Basalt Rhyolite Sc 30 - 50 2 - 10 Th 0.5 - 3 10 - 20 Ti 5000 - 10000 1000 - 3000 V 200 - 300 10 - 40 Zr 40 - 80 150 - 250 Nb 4 - 8 10 - 15 La 4 - 8 20 - 40 Ce 10 - 20 60 - 80 Cr 100 - 300 5 - 10

28. ALWAYS check that you have the correct locations in your database First thing; always plot a map of your data. Are the locations correct?

29. Check for data type Many software systems will have tools that identify non-numeric (text), >, <, - in data tables. Fix these before you start interpreting data. Make sure that below detection results are properly flagged.

30. Check the digestion method. A 4 acid digest will dissolve almost everything. Most, but not all of the zircons will dissolve, especially in younger rocks. Aqua regia will dissolve carbonates, sulfides, Fe oxide and chlorite, but it will not dissolve quartz, feldspar, illite or most accessory minerals. Volcanic rocks will always have around 6 to 9% Al. However with aqua regia , the Al results will be around 1 to 3%. A scatterplot of Al versus Zr shows 2 very distinct populations. This clearly separates 4 acid digest from aqua regia data. 4 acid digest Aqua regia digest

31. Check the assay method; ICP-AES or ICP-MS? Many historic data sets will contain a compilation of data from different assay methods. To check whether the ICP data is AES or MS, make a probability plot of a pathfinder like Te for example, which has a low natural abundance. The detection limit by MS is 0.01ppm. The detection limit by AES is usually 5ppm. If there are AES results mixed with MS, it will be obvious! ICP-AES results with a 5ppm DL Note; for assay results that are below detection, store these results as the negative value of the detection limit. That way, everybody will know that the result was below detection, and they will know what the detection limit was, without any ambiguity.

32. Check for batch errors. Plot maps of elements. Check to see if there are mis -matches between batches of data or level shifts from sample line to sample line. Note the higher background levels in the data on the western side. These types of errors are most obvious when you look at elements that have low crustal abundances (it is unusual to see an error like this in arsenic!). This data will still pass database import validation rules, but something is clearly wrong.

33. Check for contamination. Quartz-rich samples (veins) often produce around 100ppm Cr contamination from Cr-steel in an LM5 mill. This section shows drill holes with high Cr that look out of place. Zoom in. Every second sample has high Cr. 2 different ring mills with different grades of steel!

34. Check for calibration errors. There are some pairs of elements which are VERY highly correlated in nature, eg La versus Ce Low precision High precision, but…..

35. Check for calibration errors. There are some pairs of elements which are VERY highly correlated in nature, eg Hf versus Zr ? Low precision High precision

36. Check for calibration errors. There are some pairs of elements which are VERY highly correlated in nature, eg Ta versus Nb Low precision High precision, but a few Ta values are clearly wrong

37. Check for out of range values. Make a K/Al versus Na/Al molar ratio plot. It is highly improbable that any valid assays plot above the albite-orthoclase tie-line. This is a common outcome in samples where the 4 acid digest was not taken to incipient dryness. Al remains in the test tube as an insoluble fluoride complex; samples plot with impossibly high ( K+Na )/Al ratios.

38. Database management Set up your database to export ICP-MS data in a format that is ready to use without the need to manipulate data in excel.   Just one column per element; export the data as ppm of the element to avoid any ambiguity. Even where results are reported by the lab as % or oxides, convert to ppm of the element for the export.   Below detection results should be exported as the negative of the detection limit value. That way, the user knows explicitly that the result was below detection, and what the DL value was.   Each row should also include a field for the Assay Lab name and a field for the lab assay method (code, eg ME-MS61). That way we can check the digest and assay method.   Ideally, the export should also include a field for the logged lithology and logged alteration from that interval. This is a nice to have, but not a have to have, since these can very easily be merged using Leapfrog. X,Y,Z coordinates are also nice to have for downhole data, but these similarly can be added in Leapfrog.

39. Making the most of multi-element geochemistry A work-flow for interpreting ICP assays Scott Halley Consulting Geochemist, Mineral Mapping Pty Ltd, Hawley Beach, Tasmania, 7307. Email: [email protected] Web: www.scotthalley.com.au

40. Advantages of 4 acid digest ICP Exploration Better identification of rock compositions, including magmatic fertility . Better characterization of alteration. Map the full extent of pathfinder footprints and metal zonation patterns.

41. Advantages of 4 acid digest ICP Resource Drilling Consistent logging of lithology and alteration mineralogy. Better geological model Map the variability in ore types; improved metallurgical test work sample selection. Proxies for rock hardness domains Clay mineralogy distribution map sulfide mineralogy and abundance. Map deleterious elements Impact on concentrate quality and environmental management

42. Equilibrium Constant K D = X i solid / X i liquid Incompatible element; size and/or charge of the cation is unsuitable for inclusion in the crystallizing mineral phases. Partition coefficient between rock-forming minerals and melt much smaller than 1 Two groups of incompatible elements; LILE , or large-ion lithophile elements have large ionic radius, such as rubidium, caesium , strontium, barium; HFSE , or high-field-strength elements have large ionic valences (or high charges), such as zirconium, niobium, hafnium, rare-earth elements (REE), thorium, uranium and tantalum. Trace elements; Compatible versus Incompatible

43. Compatible elements have decreasing abundance with increasing SiO2 Example: Ti is a compatible element.

44. Compatible elements have increasing abundance with increasing SiO2 Example: Th is an in compatible element.

45. Accessory mineral phases control HFSE distribution patterns Example; zircon crystallization begins at around 800 o C; Zr changes from incompatible to compatible behavior at around 70% SiO 2 Onset of zircon crystallization

46. Classifying rock compositions from 4 acid digest ICP-MS analyses 4 acid digest ICP-OES/MS reports all the major elements except SiO2. Without SiO2, plot immobile trace elements against 2 different immobile compatible elements; Sc substitutes for Fe in silicates. Ti is an immobile compatible element; mostly in opaque oxide minerals. As a guide; Basalt 30 to 50ppm Sc Andesite 20 to 30ppm Dacite 10 to 20 ppm Rhyolite <10ppm Sc

47. R.R. Loucks (2014) Distinctive composition of copper-ore-forming arc magmas, Australian Journal of Earth Sciences: An International Geoscience Journal of the Geological Society of Australia, 61:1, 5-16 Richards, J.P., 2011. High Sr/Y arc magmas and porphyry Cu±Mo±Au deposits: just add water. Economic Geology , 106 (7), pp.1075-1081. Olson, N.H., Dilles, J.H., Kent, A.J. and Lang, J.R., 2017. Geochemistry of the Cretaceous Kaskanak batholith and genesis of the Pebble porphyry Cu-Au-Mo deposit, southwest Alaska. American Mineralogist: Journal of Earth and Planetary Materials , 102 (8), pp.1597-1621. Escolme , A., Berry, R.F., Hunt, J., Halley, S. and Potma , W., 2019. Predictive Models of Mineralogy from Whole-Rock Assay Data: Case Study from the Productora Cu-Au-Mo Deposit, Chile. Economic Geology, 114(8), pp.1513-1542. Halley, S., 2020. Mapping Magmatic and Hydrothermal Processes from Routine Exploration Geochemical Analyses. Economic Geology, 115(3), pp.489-503. Recommended references

48. Scandium; Immobile, Substitutes for Fe in silicate minerals. Plot Sc vs Ti , Th, V, Zr, P, Nb , Al, Cr Titanium; Immobile, Occurs in Fe- Ti oxides + biotite, amphibole. Plot Ti vs Sc, Th, V, Zr, P, Nb , Al, Cr Individual plots to pick specific signatures; V/Sc vs Sc; to pick signatures of high pressure melting of an amphibole-bearing source, or fractional crystallization of magnetite Ti vs Nb ; to characterize opaque oxide mineralogy Ta vs Nb ; to pick fractional crystallization of biotite Hf vs Zr ; to pick fractional crystallization of zircon Sr/Y vs Y; to pick high P melting of plagioclase in hydrous environment Sc vs Ni; to pick sulfur saturated magmas Procedure for Rock Compositions

49. Rock Compositions Plot Sc vs Al, Nb , P, Th, Ti , V, Zr, Cr. Look for clusters in the data. Assign each cluster to a category. Chemistry does not differentiate rock type, ie an intrusive, extrusive, volcaniclastic or epiclastic.

50. Rock Compositions Take the same suite of elements and plot Ti first; ie Ti vs Sc, Al, Nb , P, Th, V. This shows patterns related to the accessory mineral phases in each magmatic event, eg magnetite, rutile, titanite, apatite, monazite etc.

51. Classifying rock compositions from 4 acid digest ICP-MS analyses; Case study In this particular case study, some of the clearest distinctions between different eruptive products are shown by a plot of Sc vs Th. Chemical Classification

52. R.R. Loucks (2014) Distinctive composition of copper-ore-forming arc magmas, Australian Journal of Earth Sciences: An International Geoscience Journal of the Geological Society of Australia, 61:1, 5-16 Magmatic Fractionation Pairs of trace elements are sensitive indicators of magmatic fractionation. The order of fractionating minerals and extent of fractionation depends on source chemistry, temperature, pressure, redox and water contents. These factors determine which metals are concentrated in fractioning melts.

53. Magmatic Fractionation Trace element pairs indicate; Which mineral is fractionating Extent of fractionation Sulfur saturation Infer redox Infer Temp Infer H2O. This is nothing new, BUT modern analytical methods have broader suite of elements and precision 2 orders of magnitude better than 1990’s data Pine Hill Granite (Sn)

54. Magmatic Fractionation Geoscience Australia samples Kennedy Province, NE QLD State-of-the art analyses, 1980’s Same samples reanalysed ; Conventional 4 acid ICP-MS Suite of granites with fractional crystallization of zircons High prospectivity for magmatic Au

55. Redox sensitive elements Consider Sc and V; both are compatible elements. Sc substitutes for Fe in silicate minerals (behaves like the silicate-hosted component of Fe). V 3+ behaves the same way as Sc. V 4+ behaves more like Ti . Sc and V have the same trajectories in reduced magma, but different trajectories in oxidized melts.

56. Compare and contrast typical crustal melts with porphyry Cu magma Mount Read Volcanics vs Southern Peru, Eocene porphyries Magnetite Fractionation; V depleted more rapidly than Sc Hornblende ( Cpx ) Fractionation V/Sc starts at higher values, Sc depleted more rapidly than V, especially in more oxidized systems. Reduced magmas; V/Sc constant with fractionation. Ilmenite fractionation Magnetite fractionation Fe-silicate fractionation Mount Read Volcanics , VMS district Cyril Chelle-Michou , 2013 Tintaya , Porphyry Cu district

57. Porphyry Cu fertility Without SiO2 analysis, plot V/Sc versus Sc Fertile porphyry Cu field Fractional crystallization of magnetite Ilmenite Fractional crystallization of Cpx Hbl

58. Most magmas have Ni/ Sc between 1 and 2, Cu/ Sc around 2.5 Magmas that fractionate magnetite become S-saturated. Immiscible ISS first (Fe-Ni-Co-S), MSS second (Fe-Cu-Au-S) Sulfide Saturation

59. In-situ fractionation in the Golden Mile Dolerite Top of sill Base of sill Magnetite Zone Golden Mile Dolerite Same bulk chemistry as the Lunnon Basalt Same as 90% of Archean Basalt Fractionates to magnetite instead of ilmentite . Sulfur saturated at the onset of magnetite crystallization.

60. Zircon fractionation Fractional Crystallization of Zircons Commonly observed in magmatic hydrothermal orebodies, eg Sn granite, Intrusion-related Au deposits. Reduced total Zr content, but high Hf / Zr ratio.

61. Ti vs Nb ; to characterize opaque oxide mineralogy Contrast magnetite versus titanite fractionation Mount Read Volcanics , VMS district Tintaya , Porphyry Cu district Increasing SiO2 Increasing SiO2 Cyril Chelle-Michou , 2013

62. Fractional Crystallization of Biotite Commonly observed in fractionated granites. Strongly depleted Ti but increasing Ta/ Nb ratio.

63. Classifying rock compositions from 4 acid digest ICP-MS analyses; Case study All of these melts plot on a spectrum from high Sr-low Y to low Sr – high Y. This maps a transition from hydrous melting at high pressure where garnet is stable in the source region and all the plagioclase melts, to low pressure melting where plagioclase is fractionating. High pressure melting of HBL-bearing source Mid-crustal melting Field of porphyry Cu magmas

64. Fractional Crystallization of Plagioclase REE’s have 3+ valency. Eu can be 2+; Ce can be 4+. In reduced to moderately oxidized melts, Ce 2+ substitutes for Ca 2+ . Plagioclase fractionation removes Eu from a melt relative to other REE’s. Plagioclase fractionation

65. Example; Lithology model from assays

66. Conclusions It is difficult to consistently log lithology, especially in very altered rocks. Immobile trace element chemistry provides quantitative data that can significantly improve the accuracy and consistence of logging; note, this is an aid to logging, it does not replace logging. Consistent logging will lead to better geological interpretations, more reliable resource models, and better classification of ore-types.

67. Alteration Mineralogy from Assays A work-flow for interpreting ICP assays Scott Halley Consulting Geochemist, Mineral Mapping Pty Ltd, Hawley Beach, Tasmania, 7307. Email: [email protected] Web: www.scotthalley.com.au

68. Barker, S.L., Hood, S., Hughes, R.M. and Richards, S., 2019. The lithogeochemical signatures of hydrothermal alteration in the Waihi epithermal district, New Zealand. New Zealand Journal of Geology and Geophysics , 62 (4), pp.513-530. Alteration Mineralogy from Assays Case Study from Waihi low sulfidation epithermal deposit, New Zealand

69. Common alteration minerals include; white mica, chlorite, carbonate, feldspar, pyrite, Smectite, sulfates. Think of the mineral compositions, then design major element scatter plots to show the composition and abundance of these minerals.

70. Common alteration minerals include; This process can be captured as a standard workflow. K/Al vs Na/Al molar ratio plot to map sericite alteration Al-K-Mg ternary plot to map relative proportions of sericite vs chlorite Ca vs Mg to map calcite vs dolomite Ca-Fe-Mg ternary plot to pick carbonate compositions Ca-K-Na to map feldspar compositions Fe vs S to map the degree of sulfidation Ca-Fe-S ternary plot to map anhydrite Rb vs K to map alunite

71. Carbonate; Ca vs Mg Plot Ca versus Mg to identify samples that might have a high abundance of calcite or dolomite.

72. Sericite; K/Al vs Na/Al molar ratio plot Note the proportion of samples that are strongly depleted in Na. Na-depleted samples could contain dominantly adularia, illite, kaolinite or chlorite Relatively unaltered rocks should plot around here

73. Relatively unaltered rocks should plot around here Plagioclase destruction, Na removal Calcite removal Feldspar Ca-K-Na ternary Sodium has been removed. So what happens to the Ca? Initially Ca is retained as carbonate, but then the calcite is dissolved.

74. Ca versus C This data set contains Carbon analyses. Most of the Ca in altered rocks is retained in calcite. Ca:C (molar) = 1:1 ie Calcite

75. Al-K-Ca ternary plot Most of the Ca in these rocks can be accounted for by calcite. Note that the kaolinite, most of the strong illite and some of the adularia samples are Ca-depleted.

76. Smectite Adularia Chlorite Illite Chlorite; Al-K-Mg ternary plot Samples on the K-rich side of the illite-chlorite tie line are mixtures of illite, adularia and chlorite. Samples on the K-poor side of the illite-chlorite tie line are mixtures of illite, chlorite and smectite (+/- relict albite). Na-K-Mg poor samples are kaolinite rich.

77. Disseminated pyrite Utilizing host-rock Fe Pyrite Veins Fe versus S Plot Map the degree of sulfidation .

78. Conclusions It is REALLY difficult to consistently log alteration. Much of the alteration mineralogy can be guesstimated and quantified from major element geochemistry . Identification of the clay mineralogy is vastly improved if SWIR data is collected along with the assay information. Consistent logging will lead to better geological interpretations, more reliable resource models, and better classification of ore-types.

79. Pathfinder chemistry Plot pathfinder elements as a factor of average crustal abundance levels for each element. A coherent footprint (multi-point anomaly) of >10 x average crustal abundance is a significant anomaly!

80. Outcropping Porphyry Cu deposit Blind Porphyry Cu deposits 300m below surface Bismuth assays by ICP-MS , detection limit 0.01ppm Bismuth assays by ICP-AES , detection limit 5ppm Pathfinder chemistry; importance of detection limits

81. Calculated Mineralogy from Assays Calculation of weight% gangue minerals from 4 acid digest ICP analyses + SWIR Scott Halley Consulting Geochemist, Mineral Mapping Pty Ltd, Hawley Beach, Tasmania, 7307. Email: [email protected] Web: www.scotthalley.com.au

82. Quantitative Mineral estimates from chemistry alone are non-unique; there are always more variables than constraints. For example, a point here could be a mixture of orthoclase-muscovite-chlorite It could be a mixture of orthoclase-muscovite-biotite Or We could use extra information from qXRD , QEMSCAN or CoreScan to provide extra constraints, Or we can use thermodynamics to predict the permissible mineral assemblage.

83. Calculated Mineralogy Anhydrite

84. Calculated Mineralogy Feldspar Muscovite

85. Calculated Mineralogy Mineral percentages can be calculated from ICP analyses and treated like assay fields. This will change the way we model orebodies.

86. Summary. Rock Compositions To classify compositional groups , Plot Sc vs Ti , Th, V, Zr, P, Nb , Al, Cr Plot Ti vs Sc, Th, V, Zr, P, Nb , Al, Cr Individual plots to pick specific signatures; V/Sc vs Sc; to pick signatures of high pressure melting of an amphibole-bearing source, or fractional crystallization of magnetite Ti vs Nb ; to characterize opaque oxide mineralogy Ta vs Nb ; to pick fractional crystallization of biotite Hf vs Zr; to pick fractional crystallization of zircon Sr/Y vs Y; to pick high P melting of plagioclase in hydrous environment Sc vs Ni; to pick sulfur saturated magmas

87. Summary. Alteration categories K/Al vs Na/Al molar ratio plot to map sericite alteration Al-K-Mg ternary plot to map relative proportions of sericite vs chlorite Ca vs Mg to map calcite vs dolomite Ca-Fe-Mg ternary plot to pick carbonate compositions Ca-K-Na to map feldspar compositions Fe vs S to map the degree of sulfidation Ca-Fe-S ternary plot to map anhydrite Rb vs K to map alunite