Mantle geochemistry: How geochemists see the deep Earth

Don DePaolo/Stan Hart

CIDER - KITP Summer School

Lecture #1, July 2004

Geochemistry 50 years ago dealt with fewer questions and parameters, e.g. Birch (1952)

• How does meteorite chemistry compare with seismic properties of Earth’s interior

• Is it Olivine+Pyroxene or other phases ?• How much Fe in the mantle ?• How much Al,Ca,Na,K (“sialic components”) is in

the mantle ?• 11 elements of interest:

O,Mg,Si,Fe,Ni,Al,Ca,S,Na,K,P

What can geochemistry do in 2004?

• The earth is made of 90 or so chemical elements, about 30 w/isotopic variations

• Chemical/isotopic characteristics can be tied to geological processes - mantle isotopic chemistry is a tracer

• We can tell where a particular piece of mantle has been in the past and/or what has happened to it

• Radiogenic isotopes provide clocks as well as tracers

Questions for geochemistry• How deeply does near surface material circulate into the mantle? On what

time scale?• Does the mantle have large scale chemical structure (layering?)• Does the core exchange material with the mantle? (Do plumes come from

the CMB?)• What are the characteristics of mantle convection in terms of its ability to

stir and homogenize heterogeneous materials?• What features of mantle seismic heterogeneity are thermal and which are

chemical?• What aspects of mantle structure are congenital?, of recent origin?; steady-

state features?

Components of geochemistry• Petrology of the mantle (proportions of minerals or rock

types - e.g. lherzolite, harzburgite, eclogite, pyroxenite)• Melting of the mantle• Trace element composition of the mantle (doesn’t affect

mineralogy, but can be indicative of history)• Trace element composition II (water and CO2) - affects

melting behavior.• Isotopic composition of the mantle (from radioactive

decay, input from surface reservoirs, input from core?)• Sampling of the mantle (scale of sampling by magmatism;

sampling biases, invisible reservoirs)• Material balance - the sum of the parts must equal the

whole Earth for every element and isotope

(200)

Lower mantle

Upper mantle

(300)

(100)

(2)

(1)

Mass in units of 1025g

(200)

Lower mantle

Upper mantle

(300)

(100)

(2)

(1)

Mass in units of 1025g

Oceanic lithosphere ≈ 10Continental lithosphere ≈ 5

Consider this: (1) “Heterogeneities” are introduced from the top and the bottom(2) Magmatism samples only the top and the bottom

There are key elements of the system where chemistry is done.(Most of what we infer about the mantle depends on how well

we understand the processes.)

Choose one:

There are...(a) too few(b) too many(c) just the right

number

...of isotopic tracers

There are stable isotopes too !

Why the crustal reservoirs matter...

Depleted mantle

Why the crustal reservoirs matter...

10-4

10-3

10-2

10-1

100

101

102

103

0 500 1000 1500 2000

Age (Ma)

Laminarshear

Turbulentshear

Chemical Diffusion

ε = 10-14 sec-1. MORB

Otherlavasequences

UMrocks

Geochemically AnomalousLayer (Thickness = λh )

Background mantle that must be averaged withanomalous material (effective anomaly thickness):

Concentration = Ch

Isotopic contrast = ∆R h

Background mantleconcentration = Cb

Ch ∆Rh

Cb ∆Raλb = λh

Thickness of geochemical anomalies

Making heterogeneity at a mid-ocean ridge...

Incipientlydepletedlherzolite

stronglydepletedlherzolite

harzburgite

basalt, gabbrosediment

unmodifiedlherzolite

H2O-enhanced melting region

Mid-ocean ridge factoryHydrothermally altered

Incipientlydepletedlherzolite

stronglydepletedlherzolite

harzburgite

basalt, gabbrosediment

unmodifiedlherzolite

-20 0 +20εNd -20 0 +20

εNd

0.1 b.y. later 1.0 b.y. later

Anything systematic about distribution of heterogeneities?

Anything systematic about distribution of heterogeneities?

Bulk Earth

Younger cont.crust

Older cont.crust

Lower cont.crust

Upper cont.crust

Anything systematic about distribution of heterogeneities?

Bulk Earth

Chondrites

Distribution of isotopic ratios among ocean islands is not entirely random

Al Hofmann’s analysis, 2003

0

20

40

60

80

100

120

-8 -4 0 4 8 12 16

εNd

- Mid Atlantic Ridge( )data from PETDB

AverageMantle

Material balance for Sm-Nd...

BulkEarth

-10

-5

0

5

10

15

-60 -40 -20 0 20 40 60 80

LATITUDE

Mid-Atlantic Ridge(data from PETDB)

An example of heterogeneity on various scales - Nd isotopes in MAR basalts. 5 to 10 units of variation can be found over 10km or 10,000km along the ridge. The entire range of values observed worldwide (in all types of oceanic basalts) is found along the ridge

“PM”

“DM”

AM

MORBRecycled Primitive

The helium problem

0

5

10

15

20

25

100 200 300 400 500 600Model Age (ka)

HSDP Mauna Kea(Kurz, 2002)

MORB Range

6.0

6.5

7.0

7.5

8.0

8.5

9.0

100 200 300 400 500 600

εNd

( )Model Age ka

HSDP Mauna Kea( & , 2002)Bryce DePaolo

edge center

20 km

N

Melt Supplymax = 5 cm/yr

0.001

0.1

0.3

0.5

0.7

0.9

1.0

0.05

Present

200 Ka

400 Ka

600 Ka

800 Ka

Kohala

Hualalai

Mauna Loa

Loihi

Kilauea

Mauna Kea

Magma CaptureArea

HSDP

Mahukona

Melting region

Sampling issues: Pt. 1

5 10 15 20 25 30 35 40

0.7025 0.7035

0

500

1000

1500

2000

2500

3000

3He/

4He

87Sr/

86Sr

Sr

He

He-3 anomaly(Pb anomaly?)

Sr anomaly

Core or core-mantling dense layer

Width of melting region

Things may get even more interesting when we model the melting in the context of the flow - (M. Jull, unfinished, 2003)

Melting versus tracers...

(Modeling from Jull & Ribe, 2002)

Partial melt zone(φ , )w

W

∆z

Plume

Lithosphere

Storage (Chamber Vr)

z

Erupted Lava

= 0z

εi

Sampling issues, Pt. 2:Over what vertical distance are isotopic ratios averaged?

DePaolo, JGR, 1996

101

102

103

104

0 5 10 15 20

HePbSrNd

w/W

Mauna Kea (380-1000m)

4

6

8

10

12

14

200 400 600 800 1000Depth (meters)

18.40

18.45

18.50

18.55

18.60

18.65200 400 600 800 1000

Depth (meters)

Estimating dispersivity in the Hawaii melting region

From DePaolo, JGR, 1996

1 km

Spiegelman et. al, 2001 (JGR, 106, 2061-2077)

For MOR’s the channeling instability may apply; makes for very large vertical dispersion - i.e. lots of averaging. May not be the case for plumes (?)

OK, so what do we think we know......?

Where we are going in the next 2 weeks....

Geochemistry Tutorials....