![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
Opening lecture of day 2! (Which is a Tuesday. This is relevant later.) The guy who gave this talk is my academic grandfather and an absolute mensch - seriously, he is funny enough that even if this is not your field I think it is worth reading for enjoyment value.
(First brown session moderator!)
M TIM ELLIOT -- WHERE ON EARTH DOES THE MOON COME FROM?
(He's wearing a t-shirt. I think he might be the first presenter to have done that. Good. I like him.) [NB for those of you reading this later -- when I say Bristol is a place I am considering for postdoc, what I meant was "this dude's reputation is stunning, he runs the best isotope geochemistry lab in the world"]
Ti is more-or-less identical to the Earth: "chip off the old block"
Today trying to focus on mass dependent rather than mass independent (which we saw a lot of).
Majeur Elemente, Majeur Probleme (why is this funny? everybody laughed!):
- mass-independent ratios trace source
- mass-dependent ratios also reflect process
- mass dependent isotopic compositions of major Earth/Moon forming elements
- uses Mg - solely lithophile, not influenced by core formation, unlike Si
-- claims no del26Mg fractionation during melting - so we don't see this in the badger (oh dear already?)
-- arc basalt, OIB, continental basalt, peridotite all... plot in pretty much the same place (measured via lots of different methods, whoo)
-- ... or crystallisation (Teng et al 2007)
Layered Earth before Terminal Impact:
"I'm putting up some geochemist's ideas of dynamics, which I expect to be shot down in flames, but at least it will liven up the morning"
- crystallisation of earlier magma ocean from last planetesimal collision [sounds convinced we didn't have whole-planet magma ocean]
- xlisation from bottom/middle up, w/ high P phases first
- isotopic consequences of extensive Mg-perovskite crystallisation... different coordination numbers
- looks at Mg isotope fractionation between Forsterite and MgPv (Fo kiiiiiinda looks like melt, actually doing this because melt is hard)
-- getting significant fractionations even at high T
-- over PT range get a fairly constant factor because they compensate one another (obviously not WHOLE range, but one can pull out An Area, e.g.)
"I think there's a lot of SPH evident with geochemists...." [what? again, laughs!]
- argues that we get perhaps a fractionated upper mantle, with MgPv cumulate; then Moon removes fractionated bit, and we get the two convectively remixed post-impact
"ETH, Harvard and Arkansas, the three glittering spires of research"
- data is UTTERLY SCATTERED for Moon/Earth for low vs high, melts vs mantle
- "of course this is an invitation for geochemists to go measure the same bloody thing all over again"
-- so Bristol DID. ... and their data is fairly convincing - low Ti like Earth, high Ti light -- agrees with Arkansas, so is amused! "But everyone clearly did their best, so it's hard to say why we got different answers. High-precision measurement with LCIPMS"
Should we believe 50ppm variations? Bracketed assessments... "Well, my nose may get longer, that would be really entertaining."
- pure standard measured against bottle you've tried to get out of your matrix
- or you can use double spiking (for 4-isotope systems) (
-- explicitly correcting for mass bias
- ... but Si and Mg only have 3 isotopes
- invented "critical mixture double spiking"
-- with 3 isotopes unique composition for which trajectory of sample-spike mixing and instrumental mass bias is identical
-- uncertainty in instrumental mass-bias hidden in error in sample-spike mixing; accuracy of natural mass fractionation unaffected
- when slightly off critical mixture, inaccuracy in knowledge of instrumental mass bias greatly [crap he changed slide]
- ooh, get nearly the same answers as standard-bracketed!
- mean low and low Ti from Moon both match Earth
- high Ti is still WAY the hell off (v light - mixing ilmenite back in? suspects ilm isotopically light, wants to check Moon ilm, since high-Ti ilm is isotopically light)
-- and now we've got lolololol double-spiking we've got precision AND accuracy
Note: that the Moon is very similar to the Earth, even in mass-dependent, where process ought to take over
- so from his layered mantle model, we sample both fractionated upper mantle and Mg-Pv cumulate for Moon and both... OR the Earth recrystallises in EXACTLY the same mantle pattern and we... only see the upper mantle
- now explains that if geochemists have a problem, we shove it in the "deep mantle" - because the Earth is special, it has high-pressure phases, and we compare it with things like chondrites that... don't have high-pressure phases
Testing impact scenarios:
- lunar impact samples whole mantle, currently terrestrial mantle chonritic?
- OR lunar impact samples surface of proto-Earth, Earth secretly non-chondritic
- ... Earth is Mg-heavy relative to chondrites
-- ... so both Earth and Moon are non-chondritic in terms of Mg. But... Mg is major element.
Conclusions:
- Moon identical to Earth in del26Mg (+- 30ppm)
- Mg-Pv fractional xlisation imparts significant Mg isotopic fracntionations >50ppm in crystallised layered Earth
Either:
(a) lunar impact samples whole proto-Earth mantle and current Earth mantle well-mixed
(b) OR lunar impact samples outer mantle and current Earth reformed and preserved layered structure
del26Mg of chondrites hints ("rather alarmingly") at (b)...
Discussion:
- ilm contains light Mg; total fraction of Mg in ilm not enough to swing total mass balance - would this explain KREEP being slightly light/Earth? (it is post-ilm, residual liquid, but that doesn't necessarily do the trick)
- 1-1.5% Mg by mass in core? But that probably can't be the fractionation (HUGE factor). Aussie not convinced that [oh crap I've forgotten it]
- "right, so there's a role for experimental petrology?" "yeah, they're quite useful, those guys" - in re perovskite-??melt or was it ilmenite-melt?? something like that
- hahahaha worrying about getting 100% yield for double-spiking
- holy crap "I firmly believe what you've done is better" says Alex... Halliday, I think. But he says he thinks that perovskite can't be it because Vesta; points out that they've only got one Eucrite and it's got big error bars. "Erm, well," says Tim Elliot, having the good grace to look slightly embarrassed, "that's because we got it on Sunday night." THERE IS A ROUND OF DELIGHTED APPLAUSE. Seriously, this dude is great, he's just given a Royal Society talk on #overlyhonestmethods.
(First brown session moderator!)
M TIM ELLIOT -- WHERE ON EARTH DOES THE MOON COME FROM?
(He's wearing a t-shirt. I think he might be the first presenter to have done that. Good. I like him.) [NB for those of you reading this later -- when I say Bristol is a place I am considering for postdoc, what I meant was "this dude's reputation is stunning, he runs the best isotope geochemistry lab in the world"]
Ti is more-or-less identical to the Earth: "chip off the old block"
Today trying to focus on mass dependent rather than mass independent (which we saw a lot of).
Majeur Elemente, Majeur Probleme (why is this funny? everybody laughed!):
- mass-independent ratios trace source
- mass-dependent ratios also reflect process
- mass dependent isotopic compositions of major Earth/Moon forming elements
- uses Mg - solely lithophile, not influenced by core formation, unlike Si
-- claims no del26Mg fractionation during melting - so we don't see this in the badger (oh dear already?)
-- arc basalt, OIB, continental basalt, peridotite all... plot in pretty much the same place (measured via lots of different methods, whoo)
-- ... or crystallisation (Teng et al 2007)
Layered Earth before Terminal Impact:
"I'm putting up some geochemist's ideas of dynamics, which I expect to be shot down in flames, but at least it will liven up the morning"
- crystallisation of earlier magma ocean from last planetesimal collision [sounds convinced we didn't have whole-planet magma ocean]
- xlisation from bottom/middle up, w/ high P phases first
- isotopic consequences of extensive Mg-perovskite crystallisation... different coordination numbers
- looks at Mg isotope fractionation between Forsterite and MgPv (Fo kiiiiiinda looks like melt, actually doing this because melt is hard)
-- getting significant fractionations even at high T
-- over PT range get a fairly constant factor because they compensate one another (obviously not WHOLE range, but one can pull out An Area, e.g.)
"I think there's a lot of SPH evident with geochemists...." [what? again, laughs!]
- argues that we get perhaps a fractionated upper mantle, with MgPv cumulate; then Moon removes fractionated bit, and we get the two convectively remixed post-impact
"ETH, Harvard and Arkansas, the three glittering spires of research"
- data is UTTERLY SCATTERED for Moon/Earth for low vs high, melts vs mantle
- "of course this is an invitation for geochemists to go measure the same bloody thing all over again"
-- so Bristol DID. ... and their data is fairly convincing - low Ti like Earth, high Ti light -- agrees with Arkansas, so is amused! "But everyone clearly did their best, so it's hard to say why we got different answers. High-precision measurement with LCIPMS"
Should we believe 50ppm variations? Bracketed assessments... "Well, my nose may get longer, that would be really entertaining."
- pure standard measured against bottle you've tried to get out of your matrix
- or you can use double spiking (for 4-isotope systems) (
-- explicitly correcting for mass bias
- ... but Si and Mg only have 3 isotopes
- invented "critical mixture double spiking"
-- with 3 isotopes unique composition for which trajectory of sample-spike mixing and instrumental mass bias is identical
-- uncertainty in instrumental mass-bias hidden in error in sample-spike mixing; accuracy of natural mass fractionation unaffected
- when slightly off critical mixture, inaccuracy in knowledge of instrumental mass bias greatly [crap he changed slide]
- ooh, get nearly the same answers as standard-bracketed!
- mean low and low Ti from Moon both match Earth
- high Ti is still WAY the hell off (v light - mixing ilmenite back in? suspects ilm isotopically light, wants to check Moon ilm, since high-Ti ilm is isotopically light)
-- and now we've got lolololol double-spiking we've got precision AND accuracy
Note: that the Moon is very similar to the Earth, even in mass-dependent, where process ought to take over
- so from his layered mantle model, we sample both fractionated upper mantle and Mg-Pv cumulate for Moon and both... OR the Earth recrystallises in EXACTLY the same mantle pattern and we... only see the upper mantle
- now explains that if geochemists have a problem, we shove it in the "deep mantle" - because the Earth is special, it has high-pressure phases, and we compare it with things like chondrites that... don't have high-pressure phases
Testing impact scenarios:
- lunar impact samples whole mantle, currently terrestrial mantle chonritic?
- OR lunar impact samples surface of proto-Earth, Earth secretly non-chondritic
- ... Earth is Mg-heavy relative to chondrites
-- ... so both Earth and Moon are non-chondritic in terms of Mg. But... Mg is major element.
Conclusions:
- Moon identical to Earth in del26Mg (+- 30ppm)
- Mg-Pv fractional xlisation imparts significant Mg isotopic fracntionations >50ppm in crystallised layered Earth
Either:
(a) lunar impact samples whole proto-Earth mantle and current Earth mantle well-mixed
(b) OR lunar impact samples outer mantle and current Earth reformed and preserved layered structure
del26Mg of chondrites hints ("rather alarmingly") at (b)...
Discussion:
- ilm contains light Mg; total fraction of Mg in ilm not enough to swing total mass balance - would this explain KREEP being slightly light/Earth? (it is post-ilm, residual liquid, but that doesn't necessarily do the trick)
- 1-1.5% Mg by mass in core? But that probably can't be the fractionation (HUGE factor). Aussie not convinced that [oh crap I've forgotten it]
- "right, so there's a role for experimental petrology?" "yeah, they're quite useful, those guys" - in re perovskite-??melt or was it ilmenite-melt?? something like that
- hahahaha worrying about getting 100% yield for double-spiking
- holy crap "I firmly believe what you've done is better" says Alex... Halliday, I think. But he says he thinks that perovskite can't be it because Vesta; points out that they've only got one Eucrite and it's got big error bars. "Erm, well," says Tim Elliot, having the good grace to look slightly embarrassed, "that's because we got it on Sunday night." THERE IS A ROUND OF DELIGHTED APPLAUSE. Seriously, this dude is great, he's just given a Royal Society talk on #overlyhonestmethods.