Friday, March 16, 2007

Electron probes vs ion probes

CJ asked me to explain what this so-called ultraprobe is all about, so here’s a brief rundown on the difference between electron probes and ion probes, and why it might be useful. The ultraprobe is an electron probe.

An electron probe bombards a sample with electrons with energies in the low tens of KeV. At these energies, these electrons can dislodge inner shell electrons in the material being analyzed. When an outer shell electron then decays to fill that vacancy, an Xray is produced, and the energy of that X-ray is characteristic to each element. So by attaching a bunch of X-ray spectrometers to an electron gun, you can determine the elemental composition of the target by analyzing the X-rays produced by electron bombardment. X-ray diffractometers basically use crystals with known lattice parameters to diffract the X-rays according to the X-ray wavelength.

Since X-rays are photons, the wavelength and energy are related by the equation e=k/lambda, where e is energy, lambda is wavelength*, and k is a constant, 1243 eV/nm. This constant can be derived by multiplying the speed of light (in nm) by planks constant (an obscure, extremely small physics number that relates fundamental physics properties to each other), and dividing by the number of joules per eV (not many).

Electron probes are great for determining the major and minor elemental composition of minerals, but they give no information on isotopic composition, since isotopes all have the same electron configurations. Since the only things being moved are electrons, this technique is non-destructive, unless you turn the power up too high and melt your sample by overheating it.

Ion probes are Secondary Ion Mass Spectrometers. They bombard a sample with ions, and those ions then ionize the target. The ions from the target are then accelerated into the mass spectrometer, sorted by mass. The ionization efficiency of ions is different for different elements, but the same for isotopes of the same element. Since the ions are sorted by mass, isotopes can be separated and their rations are easy to determine. Elemental ratios can also be determined if a standardization method to determine relative ionization efficiencies is used.

U/Th/Pb geochronology determines radiometric ages by using one or more of the following decay chains: 232Th -> 208Pb, 235U -> 207Pb, and 238U -> 206Pb. Comparing the results of two or more of these chains, allows a geochronologist to determine whether or not the geologic material being dated has lost or gained U, Th, or Pb since crystallization, since elemental loss or gain will disturb one or more of these systems. We can also measure 204Pb, which is not a decay product, to estimate how much initial Pb was in the material. Thermochronic may have explained this in more detail.

Because the electron probe cannot measure isotopes, it cannot determine whether or not a sample has lost or gained U, Th, or Pb, and it cannot identify the presence of common Pb. Thus electron probe dating has to assume closed system behavior and low common Pb. It is also limited to minerals with a fairly high U, Th, and Pb content, since the detection limits on the electron probe are generally in the tens of ppm. On the other hand, they are easier to use, more common, and non-destructive. And if you choose your geologic question such that Pb loss and common Pb are unlikely to occur, and high precision is not necessary, you can get useful numbers quickly and easily.

The article mentions a potential snowball earth application, so my guess is that they plan to use this thing for sedimentary survey work. As Brian previously demonstrated, sediments contain datable minerals (e.g. zircon, rutile, or monazite) of various ages, but the sediment cannot be older than the youngest grain it contains.

If some future civilization wanted to determine if the Mississippi delta sediments were Quaternary (that’s our current geologic epoch), they could dig out a whole lot of grains, and analyze them. The youngest monazites in the sediment will be from the Yellowstone Hot Spot, which is a quaternary rhyolitic volcano. Trouble is, the grains from that volcano are only a tiny portion of the total sediment load of the Mississippi, so in order to find them, you need to survey a huge number of grains.

The electron probe can be used to eliminate grains that can’t possibly be the right age- a grain with too much Pb is either too old or contains common Pb, while a grain with not enough Pb has either lost Pb or is too young. Since electron probe work is non-destructive, the grains identified with the electron probe can then be retrieved an analyzed using a more precise method, to get a high-confidence, precise date. So I suspect that is the Neoproterozoic application that this machine will have.

Neither probe technique possesses the precision of isotope-dilution mass-spectrometry, but isotope dilution cannot be performed in-situ.

It’s too late to proofread, so if there are any glaring fuck-ups, comment and I’ll fix them later.

* Usually, lambda is the decay constant, but scientists like to use the same Greek letters for a million different unrelated quantities, just to piss everyone off.


Chris R said...

Thanks for the explanation. So even if you can isolate and date these 'domains' which correspond to different alteration events, you're going to have some uncertainties - in fact, to my unexpert thoughts a period of alteration is likely to promote U/Th/Pb loss or gain...?

C W Magee said...

Alteration to what? The grain, or the rock? If you have a hydrothermal or metamorphic event that breaks down a major phase that contains a little bit of Ce (like cpx, or plag), then you have the potential for monazite growth, but that event will not necessarily upset the pre-existing monazite, which is fairly resistant to low grade events.

What you need to do is measure it and see what you get.