John Cottle on the Petrochronology Revolution
Transcript
Oliver Strimpel
This is Geology Bites with Oliver Strimpel. To address many important geological questions, we often need to determine the history of a crystalline rock when it initially formed out of a melt, and what temperature and pressures it may have been subjected to between its formation and the present day. One important way of doing this is to look within the rock to see if it contains minerals that are especially useful for this purpose. In previous episodes, we've talked about zircon crystals and how we can use the decay of uranium to lead to obtain their crystallization ages. But in addition to their value as geochronology, iters, zircons, and other crystals such as those of monazite titanite and rutile contain a number of elements in trace amounts. This makes them valuable as targets for geochemical analysis as well. By measuring these trace abundances, we can learn quite a bit about the environment in which the crystal formed. This contributes to our understanding of the rock’s history, both as to the origin of the melt out of which the crystals formed, as well as the temperatures and pressures that prevailed when the crystal was forming. Two challenges have long bedeviled geochronology and geochemistry. The first is the sheer amount of labor involved in obtaining a corpus of trustworthy radiometric dates or accurate trace-element abundances. The second is the difficulty of ensuring that the geochronology and the geochemistry analysis actually apply to the same part of a crystal and are not biased by variations within the crystal. This can be especially difficult when a crystal contains several tiny domains, each reflecting a different phase of the crystal's history. John Cottle and his team have developed a technique that has largely overcome these hurdles. This has led to a revolution in our ability to unravel metamorphic histories and has given rise to a new field called petrochronology. John Cottle is a professor in the Department of Earth Science at the University of California, Santa Barbara. John Cottle, welcome to Geology Bites.
John Cottle
Thank you, Oliver. It's a real pleasure to be here and to talk with you about this really exciting topic.
Oliver Strimpel
I said in my introduction that one of the challenges faced by geochronology and geochemistry is the sheer amount of work involved in getting to a result. I have first-hand experience of dating Luka granites from the Himalaya this way. The method I use, called thermal ionisation mass spectrometry, was certainly laborious. I started by smashing up the rock samples, separating out the rare zircons, sometimes using highly toxic heavy liquids and kneeling them in a furnace, dissolving them in hydrofluoric acid, and then mounting the dissolved material as a gel onto a filament to be inserted as the cathode inside the mass spectrometer. All the while, I had to pay obsessive attention to cleanliness, as at the end of the day one is just measuring mere picograms of the lead decay products of uranium. I presume obtaining accurate abundances of trace elements within the crystal can also be very labor-intensive.
John Cottle
That's right, it's a very similar process to obtain trace elements in these minerals; you need to take the mineral and dissolve it and then separate out the individual elements using complex colon chemistry. Or you need to use some in situ technique that you might be able to avoid. Some of those are complicated processes and really get directly at individual parts of the crystal.
Oliver Strimpel
So, can you tell us how the field has advanced since then?
John Cottle
Yeah, so initially we would take parts or whole crystals and dissolve them up and separate out the elements of interest, and really we've moved on then to try and make in situ measurements of parts of crystals down to the Micron scale, and it really started with electron probe microanalysis, which are relatively old technology but useful for measuring trace-element concentrations down to about 100 or so parts per million. Subsequent to that, a couple of other techniques were developed, primarily laser ablation, inductively coupled plasma mass spectrometry, and also secondary ionisation mass spectrometry were both used. The advantages of those are that they can measure trace elements down to the parts per million, and in some cases even down to parts-per-billion concentrations of trace elements.
Oliver Strimpel
OK, let's concentrate on the laser-ablation technique, since that's the one that you've refined. What is the role of the laser-ablation system?
John Cottle
A laser-ablation system really is just a way of sampling the material; it just takes the mineral itself and disaggregates it into relatively small or relatively fine particles. And then we take those fine particles, and we transport them to the argon plasma. So you can think of the laser-ablation system as a way of sampling or drilling out material and disaggregating it into very small crystals. Small parts of crystals do then separate those using the plasma source itself.
Oliver Strimpel
What is an inductively coupled plasma, and what is its role here?
John Cottle
So an inductively coupled plasma is if you take an argon gas and you heat that argon gas to around 11,000 Kelvin using an electrical source; you can ionize that argon and turn it into a plasma. If you take then solid pieces or liquid material, you can pass that into that plasma, and you can essentially take that material and ionize it in that argon plasma, and then you can produce ions of individual elements, and then you can use a mass spectrometer to separate those individual ions and measure their relative abundances.
Oliver Strimpel
OK. I see. So the inductive coupling of the argon is the way you heat up the argon gas to the extremely high temperatures of 11,000 Kelvin, when all the electrons are stripped off and it becomes a plasma.
John Cottle
That's right, so you take your solid pieces of zircon or monazite and you have those as very small particle size, and you take those particles and you pass them into a plasma, and essentially you strip electrons from those particles, and eventually you strip enough electrons out of them, and so you produce individual ions. And so you're going from a solid piece of mineral all the way through to individual ions, and essentially the laser does the initial part of that process by breaking into small particles. But the ions themselves are made actually inside that argon plasma.
Oliver Strimpel
OK, so you get a spot of a laser onto a sample. Get a fairly good spatial resolution because of the size of the laser beam, which is pretty small, 10 to 30 microns. And then you can get it into a mass spectrometer. So why did these instruments that already existed not adequately overcome the challenges that I mentioned earlier, namely, the amount of labor involved and the problem of making geochronology and geochemistry results consistent with each other.
John Cottle
So typically these laser instruments have been in use since around 1995 or so. But those early instruments, and actually up until relatively recently, would either measure the age of a mineral, or they would measure the trace-element concentrations. Only relatively recently have we figured out how to measure both the ages as well as the trace-element concentrations together on the same volume of material, and really the advances come from new generations of instruments that are more sensitive and have a better ability to separate out those different ions, and so better sensitivity allows us to use smaller Outsize, but also measure those trace elements that are in sub PPM concentrations. Things that we simply couldn't detect because of the sensitivity of those early instruments.
Oliver Strimpel
From what I gather, those instruments had to be used as separate phases. You did the two separate analysis in two separate phases, and so you still weren't quite sure if you were actually dealing with the same actual sample.
John Cottle
That's right, and so typically when I did my PhD and I did a lot of laser-ablation analysis, the first thing I would do is I would take my sample to an electron microprobe and I. Measure the major and trace elements in those minerals and I would then take my sample to a laser ablation system and measure the age and so we could never be entirely sure that we were measuring exactly the same composition or the same domain within a crystal, and so there's always this mismatch between the different data that you generate from the different instruments.
Oliver Strimpel
Yeah, and as I mentioned in the introduction, that can be quite a big problem if there are variations across a crystal on a fairly small spatial scale, which is often the case, especially in the kind of complex situations you’re trying to unravel, I imagine. So how did you overcome this problem?
John Cottle
So instead of doing sequential measurements, what we did is we took two mass spectrometers, and we took the stream of particles that comes from a laser, and we took that stream of particles and we split that stream of particles, and we sent them simultaneously to two mass spectrometers. So one mass spectrometer would measure the A. Which and the other mass spectrometer would measure the trace-element concentrations or other isotopic traces, for example, like hafnium and zircon or neodymium in monazite things like that, but in isotopic or elemental tracer, along with the age of the mineral. The real advantage of that, then, is that you can be certain that the age you're getting is coupled directly to some form of trace element or isotopic tracer, and so when you get an age you get direct correlation between those other crucial indicators of what the age actually means in a geologic context.
Oliver Strimpel
That's brilliant. You have one laser. And then what comes off you split into two streams, so it's coming from the same spot, so there's absolutely no difference between what's going into the age determination and what's going into the geochemical trace-element determination. When I put it like that, it seems kind of obvious. Why did nobody do it before? Was it very?
John Cottle
There's a couple of things that prevented people doing this earlier. One is simply the cost involved of having two mass spectrometers in the same laboratory space. There's a bunch of infrastructure issues. The amount of gas that gets used. The power supplies that are necessary. The other part of that is that only relatively recently in the last decade or so that we’ve actually been able to get the instruments with sufficient sensitivity such that when you reduce the amount of material by half when you split it to two mass spectrometers. That you still get sufficient amounts of signal to actually measure those concentrations, and those isotopic ratios at a precision that's good enough to solve the geologic problem.
Oliver Strimpel
Is this stream coming from the laser split up as soon as it comes off the sample?
John Cottle
So the way the laser system works is the sample is at atmospheric pressure and so the laser samples the material at atmospheric pressure. But it's in a helium atmosphere, so we have an enclosed cell or chat. And we pump that chamber with helium gas and that helium gas essentially takes that aerosol material that's being ablated and extracts it from the cell. And then we mix that helium and that sample with some argon gas, and that drives it into the mass spectrometer. So the useful thing about that process is that it's relatively quick in the sense that every time the laser hits the sample, it excavates a small amount of sample, and we immediately remove that sample or that material from the sample site. And so you can actually get relatively good spatial resolution by rapidly removing that material as you go down a hole, or as you go along a crystal, and so when I say rapidly, I'm talking about much less than a second between the material being ejected from the sample site and it reaching the mass spectrometer. So you get relatively good depth resolution as well as good spatial resolution in the XY direction.
Oliver Strimpel
I'm impressed that you can go from an atmospheric type pressure, albeit filled with first helium, and you introduce Salgan, but inside the mass spectrometer itself you have a very hard vacuum, don't you?
John Cottle
That's right, so inductively coupled plasmas are unique in that sense that the sample starts at atmospheric pressure, and then there's a rapid decrease in pressure across that sample interface, and that plasma interface to get down to 10 to the minus 9 or 10 to the minus 10 millibars, to actually be able to get those ions to travel to the detector. Part of the thing we really focused on was trying to understand how those vacuum systems can be improved, and we spent a lot of time working with mass spectrometry companies to actually make those vacuum systems a lot better, and the better you make the vacuum, the better the sensitivity of the instrument. Alternatively, if you want to think about it another way, if you have enough signal, you can then actually use a smaller spot size. In other words, use less material for the same amount of precision, so depending on what you're trying to do, we really isolated the biggest improvement we could make really was to that vacuum system, and so we spent a lot of time trying to develop that.
Oliver Strimpel
What minerals are amenable to this laser-ablation split-stream technique, and do you use different minerals? Because sometimes a rock will only have one of those, and so you just have to use what you've got, or because each mineral can provide information on something different about the rock’s history.
John Cottle
We really tried to tailor the mineral we measure to the problem we're trying to answer. For example, in igneous systems we'll often use zircon, because that's the primary carrier of uranium. But in other geologic systems, for example, metamorphic rocks will often measure the monazite and the apatite the xenotime and maybe the titanite as well. And the reasons for that are primarily that they may tell us about different parts of the metamorphic history of that rock, but also they tend to take in trace elements at different concentrations because of the compositions of those minerals, and so different trace elements might tell us different things about how those rocks evolved. If the mineral has about half to about 1 PPM uranium, and it has some radiogenic lead in it, and that's not swamped by common lead, then we can usually measure an age, and if the mineral has greater than about half a PPM of any trace element, we can usually measure that to relatively good precision as well.
Oliver Strimpel
So this new term that we're using in connection with the simultaneous measurement of the age and the traceelement abundances is petrochronology. So I understand the chronology bit because we've talked a lot on the podcast earlier about uranium lead dating of zircons, and you get a date from the ratio of the daughter nuclides to the parent nuclides roughly. And with uranium lead you have two clocks, so that makes it even better. But on the petro side, if you like, or on the trace-element abundance measurement side. What exactly does that tell you about the crystal and about the rock within which it formed?
John Cottle
Yeah, so one of the challenges we've faced with metamorphic rocks, but also with igneous rocks as well, is if you have a date or a series of dates from a mineral or different minerals in a rock, the fundamental question is what does that date mean in the context of that rock and in the context of the broader analogy, and so in a metamorphic rock, if you were to date a monazite, for example, you might typically see ranges of ages that might be 5 to maybe 10 or 20 million years spread within an individual crystal or within an individual rock. For example, monazite is a great mineral because it takes in rare earth elements, and those rare earth elements can be used to tell you about what else is happening in that rock. For example, the mineral garnet has a big control on the rare earth element budget of a rock, and so if a garden is growing it will compete with monazite for rare earth elements. And so if we think about the way rare earth elements are partitioned between, say, garnet and monazite, then, if garnet is growing, the monazite that grows at the same time might well be depleted in certain trace elements. You can think about the trace elements in your mineral as being a reflection of what's going on in the geochemistry of the rock, or the petrology of that rock, and so being able to make a quantitative link between your date and your trace-element concentrations to what's happening in the rock allows you to tie it to, for example, things like deform addition, but also the pressure temperature conditions of metamorphism, and you can do a similar kind of thing in igneous rocks as well, if you understand something about the phase relationships amongst different minerals. So really, the key is to take it from a qualitative interpretation, which is what people were doing previously, to making some quantitative estimates of what the conditions of that age actually are relative to the history of that rock.
Oliver Strimpel
Oh, I see. So just to expand on the example you gave with Ghana, if you can identify that at the particular age there was a certain depletion in certain rare earth elements that you're measuring, you would infer that garnet was growing at that time in that rock, or from that melt, and therefore the fact that garnet is growing tells you something about where you are on the pressure temperature graph.
John Cottle
That's right, so one really good example is the element atrium. If you take a metamorphic rock, and as you metamorphose that rock, you increase the pressure, you increase the temperature. And as the rock gets to maybe about 500 degrees or so, monazite might well start to crystallize, and it might start to crystallize either at the same time, or potentially even earlier than garnet starts to nucleate any monazite that grows prior to any garnet forming will be relatively enriched in the element atrium, and so earlier formed knows it would be relatively high in it. But then if you continue to metamorphose that rock and you nucleate garnet, that garnet will tend to take up most of the atrium in the rock, and so any monazite that grows will be relatively depleted in atrium, and so you can track the growth of garnet in that rock by actually looking at the atrium concentration of the monazite. And often what you'll see is at the very end of the metamorphic cycle. Some of that garnet might actually start to break down, or, alternatively, the rock might start to melt, but in either case you might see breakdown of garnet, and so that liberates a bunch of atrium and any monazite that grows during that period will be relatively enriched in atrium; there's a bunch of atrium available in the system, and so the monazite will take that up. And so you can make a link between the atrium concentration or the other Earth elements between your monazite and your garnet. Now there are complications, because there are other minerals in the rock which also have atrium in them. For example, Zener team or appetite. But if you know something about the modal abundances of those other minerals, you can then start to calculate what the effect of those minerals might be, but the principle really is to try and get at whereabouts are you on the metamorphic path in both your monazite and as well as your garnet? And then you can calculate the pressure, temperature conditions. You can calculate the age and you can then get a rate, and that's fundamentally what most people really care about when they think about metamorphic rocks is the absolute conditions of metamorphism. But then also the rates of the process as well, because it's the rate that we dominantly care about.
Oliver Strimpel
How long does it take you to obtain a date and a trace-element abundance measurement for a particular location within a crystal?
John Cottle
An individual spot can take anywhere from 5 seconds to about 20 seconds to measure, so that would give you one analysis for a typical monazite that maybe has three or four different age means we might make 20 to 40 measurements on an individual monazite crystal. And for an individual rock, we might measure 10 to 12 monazite crystals. So we might end up with on the order of 100 to 200 analysis per sample, and that might take one to three hours by the time you measure some secondary standards and some primary calibration materials as well to make sure you're getting a precise and accurate point.
Oliver Strimpel
That's incredible, compared to weeks and weeks of work for one single thermal ionization mass spectrometry measurement. Totally different style of working.
John Cottle
That's right.
Oliver Strimpel
So when you want to unravel, say the history of an entire outcrop, how many mineral grains would you typically use and how many spot measurements would be used?
John Cottle
For a typical outcrop, you might date maybe four to five rocks. We would date the metamorphic rocks present, and if there are cross-cutting igneous rocks, you might date those as well. And so typically within a day or so, a 12 hours day on the instrument, you might be able to date five to six rocks. Really, the limitation of this technique is how long are you prepared to run the instrument for. We have people who come to our lab, and they'll be essentially awake for 24 hours a day for seven days a week, and in that time you can get literally thousand to several 1000 analyses off a range of different rocks. Now it's great because the more data you have, the bigger picture you see, but it comes at a cost and you mentioned thermal ionization mass spectrometry, so I think it's really important to note that both of those techniques have really important parts to play. If you are the ultimate precision date. Then you will never beat a thermal ionization mass spectrometry date. They're the most precise. They're the most accurate, and instead what we do is we trade some of that accuracy. We trade some of that precision for number of analyses, and for that spatial resolution we can measure individual parts of a crystal. And so I hear a lot of people in the community, they have one or the other, but to me, they're both tools and depending on the question you want to answer, you should choose one or the other. Or ideally you might use both.
Oliver Strimpel
Can you give us some examples of how the use of this bit-stream method has enabled us to deconvolve some metamorphic histories that were out of reach with the prior method?
John Cottle
Yeah, most of my research over the last 20 years or so has been focused on the Himalaya and really trying to understand the metamorphic history of the Himalaya, but also the magmatic history as well, and the spatial scales of metamorphism and melting as well. And so when I first started out in my PhD, we really considered the mid-crustal rocks in the Himalaya these rocks the Greater Himalayan series. These rocks make up the highest peaks of the Himalaya. And we always imagined that there was this 20-to-40-kilometre-thick sequence of these high-grade metamorphic rocks that formed a pretty coherent package. But subsequent to that, we've dated literally hundreds to thousands of these metamorphic rocks with collaborators across the orogenic front, and we realized that actually the internal structure of that series of mid crustal rocks, they’re a lot more complicated, and I think we only really were able to figure that out by literally dating 10s of rocks on an individual transect and then doing multiple transects across different parts of the Himalaya. And you start to see that the timing of metamorphism is subtly different. But also the conditions of metamorphism are subtly different, and so not only can you see differences in when these rocks are metamorphosed, but where in the crust they were metamorphosed, and how they were juxtaposed against one another. And if I think about my PhD I dated, I think something like 15 rocks, and it took me 3 or 4 years of really hard work to do that. And then we would go back five or ten years later with this new technique, we could date that number of rocks in an afternoon on the instrument, and so with a really concerted effort, you can get large amounts of data. And only then do you actually start to see that really fine scale resolution of how that mid crust was constructed. It really changed the way I thought about how you actually make the mid crust in the Himalaya. It's not really this homogeneous package of metamorphic rocks. It actually has a significant amount of variation in it that you only really are able to see when you actually get these really large date sets that link the geochemistry and therefore the conditions of metamorphism to those ages.
Oliver Strimpel
Do we have an understanding yet of what caused that heterogeneity within the Great Himalayan series?
John Cottle
One way of considering it is when you make the mid crust or when you alter the mid crust, there might actually be a series of stacks of thrust slices that get accumulated relative to one another, and when you take those thrust slices and you put them together, you then take all of that material and you start to extrude it southward towards the Indian front. We imagined that that process of taking that mid crust and moving it southward over the South Camilla happened as a coherent block, whereas actually I think we see now is that's happening at a finer scale. There are individual thrust sheets that are actually being moved relative to one another to then accumulate together to produce that mid-crustal package. And so, to me, it's a question of scale. We thought it was one giant package of rocks, and now I think we realise that actually it's smaller packages with their own individual histories which we can start to tease out are there any?
Oliver Strimpel
Other examples of places where you've applied these methods/
John Cottle
There are. Working on subduction related igneous rocks in Antarctica. Antarctica is an incredibly difficult place to get to, but there is a really large number of rocks available from people who have previously been to Antarctica, and many of those rocks were dated by isotope dilution techniques.
Oliver Strimpel
Is that the same as the thermal ionization method we were talking about earlier?
John Cottle
That's right, yes. And the problem with those minerals is particularly the Antarctic granites are relatively high in uranium. And so prior to the mid 2000s, one of the main limiting factors of thermalization technique was that if your zircon has lead loss, if it has high uranium and the crystal lattice is damaged, then that accumulated radiogenic lead can leak out of the crystal. And so when you measure the uranium lead ratio, the date is younger than you might expect, because it's lost that daughter product, and so a lot of those dates were really imprecise or they were subject to lead loss. Now the useful thing about the laser technique is that when you take a look at your crystal, you can actually isolate domains that have undergone lead loss and domains that have not undergone lead loss. You can see it when you do multiple spots in individual crystal, and so you can essentially eliminate any of the domains that have undergone lead loss and only select the parts of your crystal that are concordant or where there's been no disturbance to the uranium lead system. And so in Antarctica, we were able to date several hundred rocks across the original front and essentially eliminate all of that lead loss and really get at the true timing of igneous intrusion. But also the duration of that metamorphism. We kind of were able to redefine what the history of magnetism in Antarctica looks like. The durations, the timescales of magnetism, and link that to things like hafnium isotopes in the same analyzed volume that would allow us to understand the sources of those magmas, but also the processes that might modify those magmas. So from an individual spot we would get the age but also the isotopic composition. And we could then say something about these rocks so they're dominated by mantle components, or they're dominated by recycling of old crust. People argue a lot about subduction zones, whether they represent sites of crustal recycling or crustal growth, but unless you have those really precise age constraints coupled to some isotope tracer, it's really difficult to get at that. So that's another area. They've really done this campaign-style geochronology of literally hundreds of dates across thousands of kilometers of section, to really get at what that magmatic history looks like.
Oliver Strimpel
What period are we talking about in Antarctica?
John Cottle
The magmatic rocks and Antarctica ranged from about 560 million years all the way through to about 490 million years. Most of the magnetism ranges between about 515 and about 400 and 20, but one of the interesting things we're actually able to show is that magnetism is much older and much younger than we previously thought, because much of that variation is hidden by the lead loss that was in those older analyses. And so with the laser technique, you can see older magnetism that's concordant, and you can see younger dates that are concordant as well. But yeah, roughly that kind of camera or division.
Oliver Strimpel
Are you working on any other parts of the world with these techniques?
John Cottle
We're trying to think about other places on the planet where we see metamorphic rocks that either have more than one phase of metamorphism and have previously been either interpreted in the light of dates or ages that may be mixtures of those multiple Metamora events and places where we need a large amount of data to actually see trends in metamorphic history, so in particular, I've had PhD students working in New Zealand to try and understand the Cretaceous minimal history of Zealandia as it's breaking up from parts of Gondwana and particularly in places like New Zealand. It's really challenging because there are older and younger phases of metamorphism that you might also have to contend with, and so that high spatial resolution technique is really useful for Deacon solving the different scales of metamorphism, the different imprints of metamorphism on those rocks.
Oliver Strimpel
It's fantastic. Several people I've spoken to have said this is really creating a revolution in the field. I can imagine that all kinds of people will be falling over themselves to be able to use your methods and your instruments.
John Cottle
I think we're really actually only at the beginning of our thinking about how do you best take that information, and how do you really get the most out of that in terms of a geologic interpretation. That requires people who are interested in more the data science side of things, as well as the isotope geochemistry, as well as the metamorphic petrology. So for me, that's the most exciting thing is, I think we've got more people involved in the chronology community to bring their expertise to these problems.
Oliver Strimpel
John Cottle, thank you very much.
John Cottle
Thank you very much. It's been a real pleasure.
Oliver Strimpel