Madeleine

Question 1

(1) Not all hotspots meet criteria for a very deep mantle origin.
Three proposed types:

Answer 1

i. Narrow, primary plumes (Reunion, Hawaii)
ii. Secondary plumes emanating from top of ‘superswells’
i. Shallow tertiary plumes linked to tensile stress in lithosphere

Question 2

(1) Different plumes

Answer 2

Primary plume – from deepest part of mantle
Secondary plumes – from upper/lower mantle discontinuity
Tertiary plumes – lithosphere

Question 3

(1) Ocean island volcanism

Answer 3

• Oceanic intraplate volcanism is a product of mantle convection/ recycling processes
• Ocean islands produce isolation that allows for speciation and therefore diversity
• Ecological and evolutionary diversity – geographical isolation (e.g. Galapagos)
• Life systems closely linked to plate tectonics

Question 4

(1) Magma genesis in ocean islands:

Answer 4

• Ocean island magmas are generated by adiabatic decompression melting (no transfer of heat in or out)
• Isentropic adiabat = DT/DP for subsolidus upwelling, no conductive cooling (small temperature change with pressure).
• Mantle Tp is projected temperature of adiabatic trend at surface
• Solidus temperature increases more steeply with pressure
• Adiabatic gradient is steeper than the change in solidus temperature with pressure, so inevitable melting at low P
• Isentropic melting path has larger DT with pressure due to latent heat of melting

Question 5

(1) Melting compared to usual

Answer 5

No melting under normal conditions. Need to either increase mantle Tp (thermal anomaly) or decrease pressure (e.g. rifting)
• MORB – extreme rifting
• Mantle pot temp has a large effect on melting by adiabatic decompression
• Adiabatic decompression results from extension of the lithosphere (MORB), or plume upwelling. The extent of melting is controlled by the depth to which decompression can occur (crustal thickness).
• Relatively high melt production rates (e.g. on Hawaii) are consistent with anomalously high mantle potential temperatures (e.g. >1550 ℃)

Question 6

(1) Experimental constraints on melting

Answer 6

• High-P experiments give insights into conditions of mantle melting

Upper mantle is alumnus lherzolite - fertile, undepleted mantle

Melt extraction leaves residue depleted in clinopyroxene (harzburgite)

Question 7

(1) The basalt tetrahedron

Answer 7

The basalt tetrahedron can be used to understand and visualize the types of basalt formed in different conditions

Nepheline – a feldspathoid – like feldspar but no water
Silica undersaturated rocks to left, saturated to right
Normative components – on table column to right

Qz tholeiite = Si-oversaturated
Ol tholeiite = Is-saturated
Alkali ol basalt = Si-undersaturated

Question 8

(1) Conditions of formation

Answer 8

• In the plane Fo-Ne-Qz it is also clear that formation of alkali basalts is favoured by high-pressure melting
• Addition of CO2 shifts eutectic towards more Si-undersaturated alkali compositions
• Addition of CO2 means its easier to produce alkali basalts and towards nepheline
• Addition of H2O shifts eutectic towards more Si-saturated compositions
• Increasing extent of melting also pulls melt composition towards Si-saturation
• Fo-En cotectics as function of pressure (Kushiro 2007)
• Eutectic melt is in eqm with
• Fo + En + Ab
• Once Ab is consumed, melt composition moves up cotectic towards Si- saturated melts
• Increasing extent of melting also pulls melt composition towards Si-saturation
• Alkalis are incompatible so enriched in small degree partial melts

Question 9

(1) What happens during low-pressure crystallisation of those melts?

Answer 9

• Fo-Ab join is a thermal divide at low pressure so cannot cross this line during crystallisation. End up with alkaline and tholeiitic suites
• Join is a thermal divide that can’t be passed by fractional crystallisation – can’t move between
• Produces either alkaline basalt suites or tholiete basalt suites

Question 10

(1) Isotope geochemistry of the mantle

Answer 10

• OIBs are abundant and sample oceanic mantle far from crustal contamination – therefore a good way to investigate global mantle geochemistry
• OIBs are generally enriched in incompatible trace elements such as LILE (K, Rb, Cs), LREE (La, Ce), and HFSE (Nb, Zr, Hf)
• OIB magmas (e.g. Hawaii) have more radiogenic Sr and less radiogenic Nd than MORB source. Some even more radiogenic than primitive mantle (EMI, EMII)
• Source must be more enriched mantle reservoir. Significant mantle heterogeneity.
• Depletion in HREE relative to MORB – melting in presence of gt, >60 km

Question 11

(1) Mantle plume volcanism – Hawaii

Answer 11

• Hawaii represents the ‘classic’ mantle plume case
• Primary melts from Hawaii sample primitive, undegassed deep mantle (high 3He/4He)
• Geoid/ residual depth anomaly:
– Broad 1500km swell (buoyant uplift due to presence of hot plume)
– Localised (200km) trough due to load of volcanic edifice – flexure of elastic lithosphere
• These trends are ‘classic’ case – Hawaii – not all hotspots follow this
• Depression around islands – due to big load causing flexure of the lithosphere
• Broad swell – due to buoyant uplift of hot material under the islands
• Linear age progression away from active volcanism

Question 12

(1) Stages of volcano: Hawaii

Answer 12

Loihi seamount - alkali and transitional tholeiitic lavas – unusual for OIBs
Less than a million years to end transition

• Pre-shield stage: As the plume impinges, low-degree melts are formed at the margins of the plume. These high pressure, low degree melts are enriched in incompatible elements (including Na, K) and are alkaline.
• Shield-building stage: High degrees of melting over the plume axis result in formation of tholeiitic basalts.
• Post-shield stage: ‘Downstream’ from the plume head, again get lower degrees of melting and hence post-shield alkali basalts.

Question 13

(2) What is the evidence for magma storage beneath volcanoes?

Answer 13

– Passive degassing at active (but quiescent) volcanoes
– Magmas saturated with multiple low-P phases
– Geophysical evidence (e.g. seismic tomography, imaging of reflectors, surface deformation)
– Large volumes of magma in explosive eruptions
– Ancient magma chambers preserved in eroded crust
– Gravity anomaly across a hot body
– Surface uplift and movement

Question 14

(2) Thermobarometry

Answer 14

use mineral ± melt ± fluid reactions that are sensitive to pressure and/or temperature
– Gives equilibration P, T
– For example, there is a large volume change for formation of jadeite pyroxene (NaAlSi2O6) from melt (i.e., between reactants and products)
– NaO0.5 (liq) + AlO1.5 (liq) + 2SiO2 (liq) = NaAlSi2O6 (cpx) Large partial molar volume of Al and Na components in liquid Small partial molar volume of NaAlSi2O6 (jadeite component)
– Therefore, this reaction is strongly sensitive to pressure and the jadeite proportion of clinopyroxene can be experimentally calibrated as a function of pressure (and temperature)
– Using crystal geochemistry to understand reactions

Question 15

(2) What do the pressures mean?

Answer 15

• last conditions of equilibrium between cpx and melt
• suggests locations of prolonged magma stagnation and xlln – therefore main level of magma storage and fractionation is 19-26 km
• No evidence for shallow magma storage region under the edifice
• Crystal variability suggests many smaller reservoirs?
• Crystal compositions and zoning can tell us about the P, T, timing of crystallisation
• Results suggest that last itme the lava was stored and allowed to crystalize for a significant time was 19-26km below- below moho
• No evidence for shallow storage – it is suggested then that there are small plumes that from small sills and reservoirs that are then tapped off and come to the surface

Question 16

(2) Stokes’ law

Answer 16

Stokes’ law describes the velocity (Vs) of a spherical particle in a free fluid

• BUT crystal settling does not take into account magma convection; irregular particle shapes; particle aggregates; magma yield strength; particle-particle interactions (e.g. McBirney & Noyes 1979)

Stokes law doesn’t take into account convection currents – will interrupt settling

• Doesn’t take into account yield strength – will not move to settle
• Doesn’t take into account crystal aggregates – will have a different drag coefficient and move/settle much differently than a single spherical crystal

Question 17

(2) How effective is crystal mush compaction?

Answer 17

– Reorganisation of grain packing (initial mush framework) results in relatively small amounts of compaction (a few %)
– Viscosity of coherent mush could be very high (1017 Pa s?) – inhibits further compaction
– Compacting a crystal mush
– Reorganizing crystals to fit/pack together more – decreasing porosity
– Porosity does increase with time (shaking to reorganize) - but very low % changes

Question 18

(4) What is a glass?

Answer 18

• A glass is an amorphous solid whose atomic configuration is ‘frozen in’, equivalent to a super-cooled liquid/ melt
• Crystals have ordered, repeated (periodic) long-range 3D atomic arrays
• Melts/ glasses lack long-range structure
• But can have short-range structural components, similar to crystalline structures
• Similar building blocks – different patterns
• Oxygen dominates silicate structure
• – Large ions (O2-) and abundant

Question 19

(4) Glass/ melt structure

Answer 19

• 3D structure (“polymer network”). Basic building blocks are SiO4 tetrahedra, linked by corners
• A polymer is a large molecule formed from repeating sub-units that are covalently bonded, eg. Polythene, PVC, silicone
• Natural variety of silicate minerals is generated by linking by 1, 2, 3 or all 4 corners

Question 20

(4) NBO vs BO

Answer 20

• Bridging oxygen: oxygen atom that is shared between two network formers (Si, Al)
• Non-bridging oxygen: oxygen atom that is bonded to only one network former
• Parameter “NBO/T” correlates with physical properties (e.g. viscosity) as calculated from melt composition (practical)
• T = tetrahedrally coordinated cation (Si4+, Al3+)
• Simple parameter but correlates with physical properties, e.g. melt viscosity
• Non-bridging oxygens / tetrahedra cations = NBO/T – to a first order tells us about physical properties

Question 21

(4) Comparison of crystal and glass: single component melt (pure SiO2)

Answer 21

• Pure SiO2 crystal – tetrahedra linked at corners, structured and ordered at short range and long range. Rings have regular size and shape
• Fully polymerised, NBO/T = 0
• Pure SiO2 melt has disorder in ring shape, size, bond lengths, bond angles
• BUT still NBO/T = 0 (identical chemistry)

Question 22

(4) Multi-component melts

Answer 22

• Addition of network-modifying cations (e.g. Mg2+, Fe2+, Na+, K+) results in breaking of some Si-O-Si bridges
• Network-modifying cations disrupt (modify) the tetrahedral network
• These cations link to (alumino)silicate network by ionic bonding with ‘non-bridging oxygen’ (NBO)
e.g. Si-O-Si + M-O = 2Si-O- + M2+
• Coordination is higher
e.g. octahedral [6-fold]

• ‘Pyroxene’-like crystal has ordered lattice with regular cation sites. Chains of linked tetrahedra
• Two different tetrahedral network-formers (Si4+, Al3+), regularly ordered
• Charge balance Na+, K+ for Al3+
• Pyroxene melt can be disordered in several ways, e.g. random cation arrangement; varied tetrahedral groups; different coordination (Al)

Question 23

(4) Substituting for network-formers (Si4+)

Answer 23

• Natural silicate melts are aluminosilicates – most common substitute for Si4+ is Al3+
• Al3+ needs charge balance from adjacent cations (typically alkalis ± Ca) to keep overall 4+ charge.
• Any “excess” alkalis can act as network-modifiers
• Two possible structural roles for alkalis and alkaline earths (Ca)
(i) charge balancing cation
(ii) network modifier

• Peralkaline melts – more alkalis than are required to charge-balance Al3+
• But no alkalis are available to act as network-modifiers for peraluminous or metaluminous melts – all required for charge-balance
• In phonolites, ≤20% of alkalis are network-modifier
• In tholeiites, alkalis are never network-modifier
• Ca role varies, Mg is always network-modifier

Question 24

(4) What are volatiles?

Answer 24

•	Elements that would be gaseous at atmospheric pressure, room temperature
•	H2O, CO2, sulphur (SO2, H2S), fluorine, chlorine, nitrogen, noble gases
•	Critical influence on magmas through:
-	melt density
-	melt viscosity
-	mineral phase stability
-	mineral compositions
-	magma density (vesiculation)

Question 25

(4) Volatile examples

Answer 25

• H2O and CO2 – most abundant; most significant for melt viscosity and vesiculation (density)
• Sulphur – important atmospheric contaminant; influence on climate and ore deposition
• Halogens – important environmental influence; act to mobilise metal cations (e.g. Cu, REE)
• Nitrogen – speciation poorly understood; important for development of earth’s early atmosphere
• Noble gases – clues for understanding large-scale mantle evolution; chemistry of melt generation

Question 26

(5) Pros and cons of melt inclusions?

Answer 26

POSITIVE
+ One of the only direct ways to access pre-eruptive volatiles
+ See past (inevitable?) shallow-level degassing
+ Time-lapse sequence may be possible

NEGATIVE

• May need sophisticated analytical techniques
• Post-entrapment modification likely
• Diffusive modification
• Boundary layer entrapment?

Question 27

(5) Post-entrapment modification of MI

Answer 27

1. Perfectly glassy – will tell you everything about melt at cooling
2. Crystalisation at crystal influence changes composition of residual melt + vapor bubble
3. Crystal is overpressured – cracks and leaks

Question 28

(5) Boundary layer entrapment of MI

Answer 28

• Boundary inclusion wont represent normal equilibrium melt

* Some inclusions may experience prolonged contact with external melt before being “sealed off”

Question 29

(5) Using MI – defining mode of degassing

Answer 29

• Evolution of volatiles depends on degassing style
– Closed-system degassing. The total exsolved gas remains in continuous equilibrium with the melt at each step. There may be a pre-existing gas phase.
– Open-system degassing. Melt and gas are in equilibrium at each step but gas is incrementally removed from the system
– CO2 solubility < H2O solubility so CO2 preferentially moves into gas phase

Question 30

(5) Using MI – magma decompression

Answer 30

• Melt inclusions give good evidence for decompression xlln due to H2O loss (degassing) as volatile- saturated magma ascends towards the surface (refer L17)
• H2O (± CO2) contents give saturation pressure (from solubility/ isobars)
• Use enrichment of an incompatible element to infer crystallinity
• Crystallisation accompanies magma ascent to surface

Question 31

(5) MIs record degassing during eruptions – e.g. Mount St Helens

Answer 31

• MI from Plinian phase record high pH2O, low T, low Rb
• MI from dome-forming eruptions record decreasing pH2O, increasing T, increasing Rb –> crystallisation drives release of latent heat into melt?
• Increasing in temp during crystallisation due to latent heat of crystallisation
• Crystalisating because of loss of water during ascention

Question 32

(5) Using MI – physical behaviour of volatiles

Answer 32

• Melt inclusions can reveal the physical behaviour of different volatile species during magma fractionation and ascent
• E.g. Soufriere Hills Volcano, Montserrat
– Sulphur – incompatible: remains in the melt
– Chlorine – initially incompatible and then lost into fluid
• Sulphur is incompatible – increases in conc during fractionation
• Chlorine is incompatible at first then lost as a vapour

Question 33

(5) The petrologic method

Answer 33

• Quantify gas emissions from past eruptions to define effect of volatile emissions (e.g. SO2) on regional or global climate
• Links between flood basalt episodes and biological extinctions?
• Columbia River basalts, Miocene, 175,000 km3
• Individual flows up to 1300 km3 volume (Scotland, 16 metres!), 600 km length (Brighton to Edinburgh)
• Principle:
• Compare “undegassed” volatile concentrations in melt inclusions with “degassed” matrix glass
• Assume that difference in concentration has been lost to gas phase
• ‘Scale up’ using estimates of magma volume erupted, to estimate total mass of gas transferred to atmosphere

Question 34

(5) Possible problems with the petrologic method?

Answer 34

– were there any volatile-bearing crystals (e.g. amph/ sulphide)?
– Was the magma already carrying a free vapour phase?
– When did the magma vesiculate (reach gas saturation) relative to MI entrapment?
– Has there been any post-entrapment modification of the melt inclusions?
– Is degassed melt ÷ erupted melt?
– Free vapour phase – dissolved volatiles not in inclusion

Question 35

(5) Volatiles as trace elements – CO2 and H2O

Answer 35

A major problem in Earth Sciences is the flux and cycling and volatiles through the solid earth and mantle.
• H2O content of Earth’s mantle controls:
– Extent and depth of partial melting
– Composition of melts produced
– Rheology of mantle and hence mantle circulation, dynamics
• OIB geochemistry – the mantle is heterogeneous. Is this reflected in volatile contents?
• H2O and CO2 are both incompatible so affected by degree of mantle melting and magmatic differentiation. Can normalise out this effect by ratioing to trace elements with similar partition coefficient (D)
• H2O/Ce, CO2/Nb (MORB)

• Straight line - good correlation of CO2 with Nb
• Indicates similar partitioning behaviour (highly incompatible)

Question 36

(4) Looking at the chemical compositions, which elements do you think have the strongest effect on NBO/T, and why?

Answer 36

The traditional ‘network modifying’ cations have a significant effect on “NBO”, e.g. Mg, Fe, Ca. “T” varies less strongly as the most SiO2-rich sample (rhyolite) also has less Al2O3.

Question 37

Plot NBO/T against log(viscosity) on Figure 1 (over), as documented in Table 2. What do you notice? How does the effect of changing composition compare with the effect of decreasing experimental temperature by 100 °C or 200 °C?

Answer 37

There is a systematic increase in melt viscosity with decreasing NBO/T. Decreasing the temperature by 100 °C increases the viscosity by ~ 1 order of magnitude. This is comparable with the change from basalt to andesite but less important than increasing levels of melt evolution (e,g. to rhyolite).

Question 38

How do you interpret the change in viscosity, in terms of the underlying structure of the silicate melt?

Answer 38

The decreased melt viscosity with higher NBO/T results from increased disruption of te polymer network. At the molecular scale, this means there are more ionic bonds between shorter molecules, and this results in less molecular-scale resistance to flow.

Question 39

Pressure, depth and density

Answer 39

Pressure = density x gravity x depth

Question 40

Assumptions made using pressures from entrapped volatiles

Answer 40

The isobars mark the range of melt volatile compositions that could be in equilibrium with CO2-H2O vapour at the specified pressure. Therefore we are assuming vapour saturation. If the magma were vapour-undersaturated at the time of entrapment the pressure could be higher, i.e. these are minimum estimates of entrapment pressure.

We are also assuming that the inclusions have not re-equilibrated with exsolved vapours of different composition during ascent, e.g. could have become dehydrated (or potentially enriched) in H2O during crustal storage; or enriched in CO2 if the magma system were affected by fluxing of CO2-rich vapour from a different (deeper) magmatic source.

Question 41

Assumptions made estimating volatile contents from inclusions of a source using the batch melting equation:

Answer 41

these are likely lower estimates as we have assumed that D = 0. Also assumed that the melts with the highest H2O, CO2 have not fractionated, mixed, assimilated crustal material or undergone any other processes that might change their compositions, i.e., they are reasonable candidates for primitive mantle melts.