Interpreting Multielement

Geochemistry dataScott Halley

July 2016

Mineralogical Patterns in Hydrothermal Systems. A seminar presented by;

ALS ME-MS614 acid digest uses a combination of HCl (hydrochloric acid), HNO3 (nitric acid), HF

(hydrofluoric acid) and HClO4 (perchloric acid). Because hydrofluoric acid dissolves

silicate minerals, these digestions are often referred to as 'near-total digestions'.

48 elements by ICP-MS and ICP-AES analysis.

ALS ME-ICP614 acid digest uses a combination of HCl (hydrochloric acid), HNO3 (nitric acid), HF

(hydrofluoric acid) and HClO4 (perchloric acid). Because hydrofluoric acid dissolves

silicate minerals, these digestions are often referred to as 'near-total digestions'.

33 elements by ICP-AES analysis.

Applied Lithogeochemistry

3 objectives

•Identify Rock types• Immobile trace elements

• Sc, Ti, V, Zr, Hf, Nb, Th, La, Ce

•Quantify Alteration• Major elements

• Al, K, Na, Ca, Fe, Mg, (Ba, Rb, Cs, Sr)

•Pathfinder patterns• As, Sb, W, Mo, Bi, Te, Tl, Ag, Au

Consider the importance of ;

•Digest Method

•Assay Method

•Detection Limit

•Mineralogical Controls

What is “background”?

Importance of Detection Limits

Outcropping Porphyry Cu

deposit

Blind Porphyry Cu deposits

300m below surface

Bismuth assays by ICP-MS, detection limit 0.01ppm

Outcropping Porphyry Cu

deposit

Blind Porphyry Cu deposits

300m below surface

Bismuth assays by ICP-AES, detection limit 5ppm

Must use ICP-MS to get useful detection limits for most elements

Importance of Detection Limits

Accuracy versus Precision

Accuracy versus PrecisionComparison of Scandium assays from 4 acid digest versus Li-borate fusion. Sc is hosted in Fe-silicates; easily dissolved in mixed acid.

Li-borate analyses might have better ACCURACY, but the 4 acid digest provides better

PRECISION. This is true for many of the immobile trace elements.

4 acid digest

Lit

hiu

m B

ora

te F

usio

n

This is a data set from a rhyolite-hosted VMS system; samples have been assayed twice; once with a lithium borate fusion AND

once with a 4 acid digest.

Accuracy versus PrecisionThis is an example of why precision is important. The chemical differences between units in this data is very subtle. With a fusion digest method, the results would have been more accurate, but

the different units could not be distinguished.

Sc versus Zr

4 acid digest

Aq

ua

Reg

ia d

igest

TungstenComparison of assays from 4 acid digest versus aqua regia digest. Is W

insoluble as H2WO4 or is it in silicates?

This is a data is from soil geochem over porphyry Cu project; samples have been assayed twice; once with an aqua regia digest

AND once with a 4 acid digest.

Mineralogical Controls

•What are the host minerals for each element?

•Are these sparsely distributed accessory minerals, or

ubiquitous, homogeneously distributed alteration

minerals?

•This has a big impact on assay “variance”.

•Pathfinders hosted in homogenously distributed minerals

allow a small sample size and a very broad sampling

pattern; important when designing sampling programs.

Average crustal AbundancePathfinder Chemistry

Zinc; blue<50ppm, red>500ppm Antimony; blue<1ppm, red>10ppm

Hellyer, TasmaniaPathfinder chemistry from exploration holes surrounding the massive

sulfide; Cu-Pb-Zn confined to small veinlets, but pyrite is pervasive.

Trace elements from pyrite give a much more consistent near-miss

signal!

Pyrite is a host for a wide variety of pathfinder elements. This soil geochem survey

gives no indication of the proximity of a VMS deposit, but it very clearly maps a

pathfinder element signature in the footwall pyrite! If just Cu-Pb-Zn had been

analysed, this target would have been missed.

VMS stringer zone. Cu is in chalcopyrite;

restricted to sparse veinlets.

Sb is in the lattice of the pyrite.

Maps the footprint of the stringer zone.

Projected position

of massive sulfide

VMS horizon

Hangingwall

stratigraphy

Footwall

stratigraphy 1km

What is the background?

• For some elements, the background is strongly lithology-

dependent. There are ways to deal with this.

• For example, in unaltered rocks, Cu, Zn, Mn, V are highly

correlated with Sc.

• Cu, Zn, Mn, V are much more mobile than Sc.

• Use the Sc values to “normalise” the other elements.

Pathfinder Ranges

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!

ALS ME-MS61; Immobile trace elements to map Lithology.From the ME-MS61 package, the following elements are considered to be the most

immobile. Use these to classify rock compositions. Note that the immobile trace

elements have an ionic charge of +3 or +4.

ALS ME-MS61; Major elements to map Alteration.Major elements track changes in the abundance of rock-forming minerals and alteration

minerals.

Lithogeochemistry Workflow; Rock types

1. xY plots Sc vs Cr, Mg, Al, Zr (to pick ultramafic rocks)

2. xY plots Sc vs Ti, Th, V, Zr, Nb, P

3. xY plots Ti vs Sc, Th, V, Zr, Nb, P

4. Check Sc vs Cr, Al, La, Ce

5. Plot Sc vs V to check for magnetite fractionation

6. Plot Zr vs Hf to check for zircon fractionation

Lithogeochemistry Workflow; Alteration

1. K/Al (molar) vs Na/Al (molar) to pick sericite and advanced argillic alteration

2. Ca-K-Na ternary plot to pick hydrothermal feldspars

3. Al-K-Mg ternary plot to pick Mg metasomatism

4. Fe vs S to pick sulfidation

5. Cu-Fe-S ternary to pick Cu-sulfide mineralogy

6. Ca vs Mg to pick carbonate mineralogy

Lithogeochemistry Workflow; Alteration2

1. Plot Xy plot of Sc vs Cu, Zn, Mn, Co, Ni, In to look for evidence of enrichment

or depletion of divalent transition element metals

2. Plot K vs Tl (thallium) to look for low temp Tl-bearing pyrite

3. Plot K vs Cs to look for low temp, disordered illite

4. Plot K vs Ba to look for evidence of barite

5. Plot Sc vs V to look for evidence of organic carbon (in black rocks) or extreme

oxidation (in red rocks)

6. Plot V vs Mo, U, Cr, As to look for evidence of organic carbon

Lithogeochemistry Workflow; Pathfinders

1. Split cumulative frequency plots, coloured by mineralogy to look fro correlations

between pathfinders and alteration; Au, Cu, Mo, Sn, W, Se, Bi, Te, As, Sb, Tl, etc

Common Hydrothermal minerals

Mineral Composition Cations Ratio

Illite KAl3Si3O10(OH)2 K/Al 1/3

Orthoclase KAlSi3O8 K/Al 1/1

Albite NaAlSi3O8 Na/Al 1/1

Montmorillonite (Na,Ca)0.33(Al,Mg)2(Si4O10)(OH)2·nH2O

Kaolinite Al2Si2O5(OH)4

Pyrophyllite Al2Si4O10(OH)2

Dickite Al2Si2O5(OH)4

Calcite CaCO3

Dolomite CaMg(CO3)2

Biotite K(Mg,Fe)3AlSi3O10(OH)2 K/Al 1/1

Chlorite (Fe,Mg)5Al2Si3O10(OH)8 (Fe+Mg)/Al 5/2

Useful Conversion Factors

Al x 1.889 = Al2O3 Na x 1.348 = Na2O

Ba x 1.699 = BaSO4 Nb x 1.431 = Nb2O5

Ba x 1.117 = BaO P x 2.291 = P2O5

Be x 2.775 = BeO Rb x 1.094 = Rb2O

C x 3.663 = CO2 Si x 2.139 = SiO2

Ca x 1.399 = CaO Sn x 1.270 = SnO2

Ca x 2.497 = CaCO3 Sr x 1.183 = SrO

Cr x 1.462 = Cr2O3 Ta x 1.221 = Ta2O5

F x 2.055 = CaF2 Th x 1.138 = ThO2

Fe x 1.286 = FeO Ti x 1.668 = TiO2

Fe x 1.430 = Fe2O3 U x 1.179 = U3O8

K x 1.205 = K2O V x 1.785 = V2O5

Mg x 1.658 = MgO W x 1.261 =WO3

Mg x 3.468 = MgCO3 Y x 1.270 = Y2O3

Mn x 1.291 = MnO Zr x 1.351 = ZrO2

Examples of recommended scatterplots; selecting Ultramafic rocksStart by plotting Sc versus Cr, Mg, Al and Zr. It is not always reliable picking ultramafic rocks from the Mg content since Mg

can be mobile during alteration and it is easily stripped during weathering. Ultramafic rocks have >1000ppm Cr. They will

have a low Al content (since they have abundant olivine, but low plagioclase, and they will have low Zr. We are looking for

those samples that meet all of these criteria.

Examples of recommended scatterplots; mafic rocksxY plots Sc vs Ti, Th, V, Zr, Nb, P; This is an example that shows sequential lava flows (eruptions) from a fractionating magma

chamber. The ultramafics were picked from the previous Sc-Cr plot. Point density contour overlays were used here to

highlight the compositional clusters in the data. The mafic rocks typically have 30 to 50ppm Sc.

Examples of recommended scatterplots; mafic rocksxY plots Sc vs Ti, Th, V, Zr, Nb, P; This is the same plot as the previous slide with the point density contours removed. The red

arrows show fractionation pathways. Ti and V are fractionation indicators for titanomagnetite. V and Sc substitute for Fe in

silicate minerals. However, V can substitute into oxides, but Sc does not. When these melts start to crystallize significant

amounts of titanomagnetite, the Ti and V contents increase, until the melt becomes depleted in V (pink group). All the time,

the HFSE elements (eg Zr, Nb, Th, REE) are increasing during fractionation.

Examples of recommended scatterplots; mafic rocksxY plots Ti vs Sc, Th, V, Zr, Nb, P; Plotting Sc first is useful to pick broad compositional groups eg, mafic, intermediate, felsic.

Plotting Ti first is useful for picking fractionation sequences.

Examples of recommended scatterplots.If we plot Ti (which is hosted in early crystalizing Fe-Ti oxides) against Zr, Hf, Th, La, Ce, etc (which are hosted in late

crystallizing zircons and phosphates) then we will see an array of linear trends that project back towards the origin. Mafic

rocks will plot with a steep slope; felsic rocks with a shallow slope.

Examples of recommended scatterplots.Plot Ti versus Nb. There are a couple of scattered andesitic populations in purple and mauve.

Note the pale blue group; these are sample that are quartz-rich; Quartz-rich to the extent that all the other components are

diluted! Mantle melts plot with a high Ti/Nb ratio. Crustal melts and fractionated magmas plot with a low Ti/Nb ratio.

Examples of recommended scatterplots; magnetite fractionationPlot Sc versus V.

Most magmas have a Sc to V ratio of around 1:7. Both Sc and V substitute for Fe in amphibole and pyroxene, and they tend

to have very linear correlations. However, V can substitute into oxides, but Sc does not. There is a very high partition

coefficient of V into titanomagnetite. Calc-alkaline rocks begin fractional crystallization of magnetite early in the cooling

history. As the melts fractionate, V is depleted. These rocks show a very clear sequence of magnetite fractional

crystallization, indicated by the array of arrows.

Examples of recommended scatterplots, Zircon FractionationHf and Zr always plot with a near perfect straight line correlation; Hf can only substitute into the lattice of zircon crystals.

However, as zircon crystallizes, the melt very gradually evolves to higher Hf/Zr ratios. Zircons tend not to nucleate as new

crystals, rather they just form overgrowing rims, so the final Hf/Zr ratio remains constant. However, where there is

fractional crystallization of zircons, early formed zircons are left behind in a restite phase, and the separated melt has a

lower zircon content but a higher Hf/Zr ratio.

Crustal melts Mantle melts

Fractional Crystallization of zircons

Project X Regional Lithogeochemistry, Fractionation plotsUsually these trends are easily recognised with resorting to ratios or log plots, but here is an example where it works really

well; log Hf/Zr versus log Ti/Nb

Crustal melts Mantle melts

Fractional Crystallization of magnetite

Project X Regional Lithogeochemistry, Fractionation plotsUsually these trends are easily recognised with resorting to ratios or log plots, but here is an example where it works really

well; log Sc/V versus log Ti/Nb

Alteration ClassificationWith a 4 acid digest method, the changes in whole rock chemistry due to hydrothermal alteration reactions can be

investigated. Consider a rock that is totally sericitized. The mineralogy of the rock might be muscovite-quartz-carbonate-

pyrite. All of the K and Al in that rock will be within sericite. Muscovite has a composition of KAl3Si3O10(OH)2. Therefore the

ratio of K:Al in the sericitized rock is 1:3. Similarly, a totally K feldspar (KAlSi3O8) altered rock will have a K:Al ratio of 1:1. In

the same way, albitisation can also be tracked. Albite is NaAlSi3O8: Na:Al =1:1.

Relatively unaltered dacite

Albite

Illite

Alkali Feldspar

Adularia

Chlorite

K/Al versus Na/Al molar ratio plot

3

6

Anorthite

Albite

Ca

NaK

Alteration ClassificationCa:K:Na ternary plot with point density contour overlay; useful for mapping hydrothermal feldspar compositions.

Sodic-Calcic Alteration

(Oligoclase)

Albite Alteration

Potassic Alteration

Kspar or Muscovite

Least-altered andesite

Alteration ClassificationAl-K-Mg ternary plot to pick Mg metasomatism.

Alteration ClassificationAl-K-Mg ternary plot to pick Mg metasomatism.

Alteration Classification; Extent of sulfidationPlot Fe versus S. On this plot, the pyrite line shows Fe to S ratios that match the stoichiometry of pyrite, ie, Fe:S (molar) =

1:2

In the vast majority of hydrothermal systems, pyrite precipitation is just a sulfidation process; the reaction just utilizes Fe

that is already present in the rock. The trend of increasing Fe and S up the pyrite line maps samples that contain additional

pyrite in veins. Points that plot of the Sulfur-rich side of the pyrite line contain sulfates.

sulfidation

Alteration ClassificationPlot Ca versus Mg to map carbonate mineralogy.

For example, this plot shows a limestone (40 wt% Ca) being partially replaced by dolomite.

Sulfide MineralogyPlot Cu-Fe-S ternary to map sulfide mineralogy.

Alteration ClassificationPlot Xy plot of Sc vs Cu, Zn, Mn, Co, Ni, In to look for evidence of enrichment or depletion of divalent transition element

metals. This plot shows that samples with relatively acid alteration mineralogy have depleted levels of Zn, Mn, Ni and Co

relative to less-altered samples of the same lithology. This also provides the basis for identifying samples with enrichment

of these metals relative to background.

Alteration ClassificationPlot K vs Tl (thallium) to look for low temp Tl-bearing pyrite. Thallium is a most unusual element as it can reside in either

silicates of sulfides. In the vast majority of cases, Tl substitutes for K in silicate minerals. This produces a linear trend on a K

vs Tl plot. At low temperatures, Tl can substitute into the lattice of pyrite. These points plot on the Tl-rich side of the silicate

trend.

Thallium in sulfides

Organic Carbon SignaturePlot Sc vs V to look for evidence of organic carbon (in black rocks) or extreme oxidation (in red rocks).

When all the V and Sc are hosted in silicate minerals, the V:Sc ratio is typically around 7:1 (along the trend of the arrow)

The orange and dark green points are mildly enriched in Vanadium; the red points are highly enriched.

Organic Carbon SignaturePlot V vs Mo, U, Cr, As to look for evidence of organic carbon. The Vanadium-enriched samples from the previous figure

show a correlation with elevated Mo, U, (Bi, As, Sb). The V-Mo-U signature in particular in a classic organometallic signature.