Geochemistry of Mafic Layered Intrusions Do s and Don’t s

GEOCHEMISTRY OF MAFIC LAYERED INTRUSIONS

DOS AND DON’TSJames D. Miller

Precambrian Research CenterDepartment of Geological Sciences

University of Minnesota Duluth

Workshop on Nickel -Copper-Platinum Group Element Mineralization Thunder Bay, Ontario

January 14, 2011

OUTLINE Geochemical Analyses for Exploration The Problem with Cumulates Major Element Chemistry Trace Element Chemistry Mineral Chemistry Assay Data for Cu-Ni-PGE Mineralized

Intrusions

GEOCHEMICAL ANALYSES FOR EXPLORATIONPURPOSE OF GEOCHEMICAL ANALYSES OF

MLI ROCKS IN EXPLORATION (IN ORDER OF IMPORT)

ESTABLISH GRADE OF ORE DEPOSIT EVALUATE THE POTENTIAL FOR MINERALIZATION EVALUATE THE COMPOSITION OF THE PARENTAL

MAGMA (SOURCE OF METALS) AND POSSIBLE CONTAMINANTS (COMMONLY THE SOURCE OF S)

EVALUATE THE CRYSTALLIZATION AND DIFFERENTIATION HISTORY OF THE MAGMA

GEOCHEMICAL ANALYSES FOR EXPLORATION

XRF+ICP-MSFull digestion>$60/smpl

ICP-MS/AESPart. digestion$20-25/smpl

Fire Assay$20-30/smpl

No Si

2009 Acme Analytical Lab Brochure

THE PROBLEM WITH CUMULATESThe Classic View of Cumulate Rocks is that they are

Mixtures of Primocrysts and a Liquid ComponentPrimocrysts are:

Enriched in high-T solid solution components (Mg in mafic phases, Ca in plagioclase)Enriched in compatible trace elements (e.g. Ni in Ol, Sr in Pl)

Liquid component is:Enriched in low-T solid solution components (Fe in mafic phases, Na,K in plagioclase)Enriched in incompatible minor and trace elements

THE PROBLEM WITH CUMULATES

The concentration (X) of any element (a) in a cumulate rock (WR) is dependent on:

• The relative proportions of primocrysts (PC) and the liquid component (LC)

• The compositions of those componentsXa(WR) = %PC1* Xa(PC1) + %PC2* Xa(PC2) + ..... + %LC* Xa(LC)

THE PROBLEM WITH CUMULATESWhat parts of this mass balance can we know?

Xa(WR) = %PC1* Xa(PC1) + %PC2* Xa(PC2) + ..... + %LC* Xa(LC)

Modes of Primocrysts? Problem is that cumulus phases continue to crystallize post-cumulus rimsPlagioclase – possible if zoning preserved, but painstakingOlivine and pyroxene –not precisely, zoning lost due to subsolidus re-equilibrationOxide - ????

THE PROBLEM WITH CUMULATESWhat parts of this mass balance can we know?

Xa(WR) = %PC1* Xa(PC1) + %PC2* Xa(PC2) + ..... + %LC* Xa(LC)

Compositions of Primocrysts? Problematic because most primocrysts are solid solutions phases

Plagioclase – possible, but cumulus cores can be very complexly zoned

Olivine and Pyroxene – Ease of re-equilibration leads to “trapped liquid shift”

Oxides – easily re-equilibrated and oxy-exsolved

THE PROBLEM WITH CUMULATESWhat parts of this mass balance can we know?

Xa(WR) = %PC1* Xa(PC1) + %PC2* Xa(PC2) + ..... + %LC* Xa(LC) Compositions of the “Trapped” Liquid Component.... and might this be representative of the Parent Magma? Problem is ...the amount of liquid in the cumulate changes over time due to compaction driven by bouyancy and crystal accumulation, and ....the composition of the liquid in the cumulate changes over time due to fractional crystallization of the intercumulus liquid.

From Tegner et al., 2009

THE PROBLEM WITH CUMULATESRare Solution to Estimating the Parent Magma of an MLI – Chill Zone

Tamarack Intrusion, Minnesota

(Goldner, UMD MS thesis, in progress)

Basal Chill Zone – 30% Ol phenocrysts in a fine gabbroic groundmass

~1cm

MAJOR ELEMENT CHEMISTRY

A B D

Fo 89Olivine

Chill Zone 4-334.2

4-334.2-30% Fo89

SiO2 39.0 47.8 51.5Al2O3 8.73 12.5FeOT 10.8 10.9 10.9MnO 0.16 0.17 0.17MgO 49.8 23.3 12.0CaO 0.25 5.66 7.98Na2O 1.14 1.63K2O 0.41 0.59TiO2 0.82 1.17P2O5 0.08 0.11

H2O + CO2 1.05 1.50Total 100.0 100.0 100.0

mg# (.9FeOT) 89.1 81.0 68.5

PELE Crystallization Models (Boudreau, 2005)

Equilibrium Crystallization QFM buffer, P=1 Kb

Fractional Crystallization QFM buffer, P=1 Kb

Tamarack Parent Magma Calculation

(Goldner, UMD MS thesis, in progress)

MAJOR ELEMENT CHEMISTRYDO NOT PLOT CUMULATES ON PLOTS INTENDED FOR MAGMA COMPOSITIONS

PrimocrysticPlagioclase

Liquid Component

fractionation

MAJOR ELEMENT CHEMISTRY

0 10 20 30 40 50 60 70 80 90 1000.0

10.0

20.0

30.0

40.0

50.0

60.0

SiO2Al2O3FeOMgOCaO

0 10 20 30 40 50 60 70 80 90 1000.0

5.0

10.0

15.0

20.0

25.0

Al2O3FeOMgOCaO

Dunite Cumulate (O)

adcu

mul

ate

mes

ocum

ulat

e

orth

ocum

ulat

e

porp

hyrit

ic

liqui

d

oliv

ine

adcu

mul

ate

mes

ocum

ulat

e

orth

ocum

ulat

e

porp

hyrit

ic

liqui

d

plac

tioc

lase

+ o

livin

e

Troctolite Cumulate (PO)Fo77 olivine, An76 plagioclase, mg#50 liquidFo84 olivine, mg#60 liquid

MAJOR ELEMENT CHEMISTRY

Major and Minor Element chemistry can serve as approximate proxies for abundances of primocryst phasese.g.

Al – PlagioclaseMg – Augite + OlivineTi – Ilmenite and Ti-magnetite

From Joslin (2004)

MAJOR ELEMENT CHEMISTRYMajor and Minor Element chemistry (that includes accurate SiO2) can be used to calculate CIPW Norms.

Helpful for metamorphosed / altered cumulates (assuming a closed system)

Helpful for calculating average An content (Ca/(Ca+Na+K)) of complexly zoned plagioclase

TRACE ELEMENT CHEMISTRYLike major elements, the absolute concentration of a trace element in a cumulate rock is typically dependent on the relative proportions AND compositions of the primocrysts and the liquid component.

Primocrysts

Liquid

Cpx?

TRACE ELEMENT CHEMISTRYCompatibility – degree to which an element prefers to partition into the solid over the liquid phase .

Kd(i)1 – Mineral-Liquid Partition Coefficient for element i in mineral

1

Kd(i)1

= C(i)mineral 1/ C(i)

liquid (C(i) - concentration of element i in wt. %)

Kd(i)1

> 1 – Compatible, Kd(i)1

< 1 – Incompatible

D(i) – Bulk Rock Partition Coefficient for element i

D(i) = x1 Kd(i)1

+ x2 Kd(i)2

+ x3 Kd(i)3

+ .... (x1 – proportion of mineral 1)

TRACE ELEMENT CHEMISTRY

From Rollinson (1993)

Com

patib

leIn

com

patib

le

Bulk Rock Partition Coefficient of Ce,Yb, and Nifor Crystallization of:

1) Troctolite (70% Pl, 30% Ol)

D(Ce) = xPl Kd(Ce)Pl

+ xOl Kd(Ce)Ol

= .7*.103 + .3*.007 = 0.092

D(Yb) = xPl Kd(Yb)Pl

+ xOl Kd(Yb)Ol

= .7*.07 + .3*.065 = 0.069

D(Ni) = xPl Kd(Ni)Pl

+ xOl Kd(Ni)Ol

= .7*.01 + .3*25= 7.5

2) Olivine Gabbro (63% Pl, 12% Ol, 25% Cpx)

D(Ce) = xPl Kd(Ce)Pl

+ xOl Kd(Ce)Ol + xCpx Kd(Ce)

Cpx

= .63*.103 + .12*.007 + .25*.09 = 0.088

D(Yb) = xPl Kd(Yb)Pl

+ xOl Kd(Yb)Ol + xCpx Kd(Yb)

Cpx

= .63*.07 + .12*.065 + .25*.09 = 0.074

D(Ni) = xPl Kd(Ni)Pl

+ xOl Kd(Ni)Ol + xCpx Kd(Ni)

Cpx

= .63*.01 + .12*25 + .25*8 = 5

0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.000.100

1.000

10.000

100.000

Tr(Yb)Tr(Ce)Ce/Yb

0.000.100.200.300.400.500.600.700.800.901.000.100

1.000

10.000

100.000

OG(Yb)OG(Ce)Ce/YbOG(Ni)

TRACE ELEMENT CHEMISTRY

F (fraction of liquid remaining)

Rayleigh Distillation: CL/Co = F(D-1)

Conclusions: Fractional crystallization of mafic magmas gradually increases the concentrations of similarly incompatible elements, but has a minimal effect on their ratios; and strongly decreases the concentrations of compatible elements

F (fraction of liquid remaining)

CL/Co CL/CoTroctolite Olivine Gabbro

TRACE ELEMENT CHEMISTRY

From Rollinson (1993)

Fractional crystallization increases the REE abundance, but has a neglible effect on the REE pattern

Since incompatible elements are 2-3 orders of magnitude greater in abundance than primocrysts, the REE pattern of cumulates (especially orthocumulates and adcumulates) will approximate that of their parental magmas and the magma source

Fractional crystallization of olivine from a komatiitic melt

From Jirsa and Miller (2006)

TRACE ELEMENT CHEMISTRYSa

mpl

e/Pr

imiti

ve M

antle

(Goldner, UMD MS thesis, in progress)

Proposed Tamarack Parent Magma Trace Element Composition compared to Early MCR Volcanics

TRACE ELEMENT CHEMISTRYSpidergrams Tectonic Discrimination Diagrams

rock

/cho

ndrit

ero

ck/c

hond

rite

Increasing compatibilityFrom Bedard (2001)

MINERAL CHEMISTRY

From Miller (2004)

Stratigraphic variations in the compositions of solid-solution cumulus minerals generally reflect the progressive differentiation of the parental magma and the occurrence of recharge event...but not exactly.

MINERAL CHEMISTRYPLAGIOCLASE

Zoning is preserved and records a history of

cumulus and postcumulus crystallization

White (2009)

Strategy 1: Compare only An of cumulus cores...Problem:Cores are commonly complexly zoned

Strategy 1: Calculate An from CIPW normProblem:Integrates cumulus and postcumulus components

MINERAL CHEMISTRYOLIVINE AND PYROXENE

Zoning is NOT preserved and thus

integrates cumulus and postcumulus compositions

Overcoming the Trapped Liquid Shift

Solution: Compare only like types of cumulates (ortho, meso, ad)

Problem: Evaluating the type of cumulate is qualitative

MINERAL CHEMISTRYEvaluating the Trapped

Liquid Shift

Gradual increase in incompatible

elements;can assume nearly

constant over limited stratigraphic

thicknessMagma Recharge

POcfcumulate

From Miller (2006)

TLS

MINERAL CHEMISTRYEvaluating the Trapped

Liquid ShiftPOcfcumulate

From Miller (2006)

TLS

Postcumulus mineral abundance are general proxies for amount of trapped liquid

MINERAL CHEMISTRYEvaluating the Trapped

Liquid ShiftPOcfcumulate

From Miller (2006)

TLS

Assuming that well foliated cumulates have lower porosity – i.e. lower volume of trapped liquid

From Meurer & Boudreau (1997)

MINERAL CHEMISTRYMineral chemistry also allows the estimation of the magma composition in equilibrum with that mineral

A procedure for calculating the equilibrium distribution of trace

elements among the minerals of cumulate rocks, and the

concentration of trace elements in the coexisting liquids I

Jean H. BedardChemical Geology 118 ( 1994 ) 143-153

KD = (XFeOol/XFeO

liq)*(XMgOliq/XMgO

ol) = 0.3 (Roedder and Emslie, 1970)

which translates in determining the mg# of the liquid as:

mg# liq = 100 / (3.333(FeO/MgO)ol + 1)

This assumes no trapped liquid shift. Therefore, one should apply this only to adcumulates

ASSAY DATA

Precious Metals Zone (PMZ)

Meters above Cu-Au break Sonju Lake IntrusionVariation in the Cu/Pd is one

of the best monitors of sulfide saturation in magmatic systems, but need high precision Pd analyses (<2 ppb DT)

From Joslin (2004)From Miller (2004) Greenwood Lake Intrusion

ASSAY DATADsulf/sil~104-108

Dsulf/sil~102

Pd is several orders of magnitude more compatible in sulfide melt relative to Cu

ASSAY DATA

after Barnes and others, 1987

R = Xsil/Xsulf

ASSAY DATA

Pd (ppb)

Cu/Pd

From Joslin (2004)From Jirsa & Miller (2006)

ASSAY DATA

From Jirsa & Miller (2006)

Quadrant w/Best Potential

Quadrant w/Ore Grade

Quadrant w/No Potential

Quadrant w/Potential at Depth

Indicates duration since initial saturation

ASSAY DATA

From Jirsa & Miller (2006)