The Art of Mineral Separation

UPDATE - My next post is a more complete discussion of mineral separation.

When many people in the geochronology/thermochronology community talk about new gadgets and gizmos on their wishlist they seem to focus almost entirely on the sample analysis side of things; particularly multi-collector noble gas mass spectrometers and various high end lasers. I'll admit, these machines are impressive and could be potentially very exciting, but if I had a pot of money to spend to increase the quality and quantity of the data my labs produce, they would not be the first thing I'd look into.

In a broad sense, doing meaningful thermochronology requires 4 basic techniques.

1. You must be able to identify tectonic and/or geochemical problems that can be be at least partly addressed thermochronologically. This starts with the big picture, but includes consideration of available lithologies and access to the necessary samples.

2. You then need to collect and process the samples. This means turning a 5 kg sack of rocks collected carefully in a very specific location to individual mineral separates ready for your lab.

3. Once you have mineral separates, you need to analyze them in a lab. Although actually collecting data in a lab is fairly trivial, running a lab well enough to insure that your data actually means something is not.

4. Once you have the data, you need to interpret it, again in order to answer the original tectonic and/or geochemical problem you set out to solve in the first place.

Steps 1 and 4 are probably the most complex, and in my opinion are the hardest skills to develop. To design and interpret good projects you need a strong background in basic geology and need to consult all of the experts that relate to the study. In my own work I need enough background to understand what the petrologists, sedimentologists, geophysicists, geomorphologists, structural geologists, and geochemists think. This requirement is not unique to thermochronologists. I'd argue that any geologist who considers tectonic questions is necessarily broad in scope. So steps 1 and 4 I see as general considerations for any earth science study.

Step 3 receives a great deal of attention. I've been a thermchronologist for less than a decade, but even in that time the number of new and expensive machines and techniques has ballooned. I've been involved with building and maintaining labs, and therefore have paid a lot of attention to these advances. As I've gone on in my career, I've started maying more attention to who gets what lab upgrade funded, or what people get with their start-up packages, or what they negotiate for when they have leverage. Right now the flavor of the day seems to be multi-collector noble gas magnetic sector mass spectrometers; these allow for the simultaneous measurement of all of the different isotopes you need to measure for whatever technique you are involved in, thereby cutting down the uncertainty and time lags of changing magnet power, yada yada yada. I won't get started on that.

What I do want to talk about is step 2, sample collection and preparation. In particular I want to talk about turning a rock into an individual mineral, a process called mineral separation. Mineral separation fascinates me, but what really amazes me is how many people either ignore or do not understand the process.

Here is the problem: for almost all analyses you need to analyze pure mineral samples. Techniques which work on small single crystals (fission-track, (U-Th)/He, U-Pb) typically require minerals that are small (100-200 microns in length) and not overly abundant in the average granite (maybe form a 5 kg sample I'll get a few milligrams of apatite). Techniques that work on multiple crystals (biotite, muscovite, and k-spar Ar/Ar), typically require a few milligrams of very pure separates. They both, therefore, require methods of separating a rock into piles of individual minerals. Mineral separation is a blanket term that describes the various ways to turn a rock into a sample. The first step is almost always reducing the rock into individual mineral grains. This is typically done by crushing and grinding the sample, trying to get the minerals to break apart along grain boundaries.

In my experience, the next step is to run your sample over a rogers or gemeni table. These are basically large gold pans that concentrate the denser minerals (apatite and zircon in particular) into a smaller pile. This is then washed and dried, and run through a magnetic separator, basically a large magnet where you can vary the power and separate minerals based on their magnetic susceptibility. This is done is a series of steps, and a skilled mineral separator can obtain almost pure concentrates of the various "magnetic minerals" such as biotite, hornblende, and monazite. When you are done, you are left with a pile of non-magnetic mineral grains, including apatite and zircon.

If you need to get apatite and zircon, you must then enter the world of heavy liquids. Heavy liquids are exactly what they sound like, liquids with very high densities, anywhere from water (1.0 g/cm³) to 4.4 g/cm³. Because minerals have fairly specific densities, they will either sink or float in different heavy liquids. Zircon is very dense (4.6-4.7 g/cm³), and will sink in a liquid like MEI (Methylene Iodide density=3.33 g/cm³), while apatite (density 3.2 g/cm³) will float. Heavy liquids have been used in geology for a long time, but the particular liquids and their methods of use have changed significantly. Many of these liquids are toxic, and therefore kind of a pain to work with. Two of the nastier liquids I have fortunately never worked with, those are Clerici's Solution (Thallium Malonate density=4.36 g/cm³) and Bromoform (Tribromoethane, density=2.89 g/cm³). Clerici's Solution and Bromoform are not all that common anymore, mainly because there are now less toxic alternatives. TBE (tetrabromoethane density=2.95 g/cm³) is also fairly nasty, but is still in use in many labs, primarily because it has a lower viscosity than its non toxic alternatives SPT or LMT (Sodium Polytungstate or Lithium Metatungstate, density 2.85 g/cm³).

I was lucky in my graduate education to learn from one of the masters of mineral separation. While many people have used the same techniques and materials they learned on decades ago, my min sep teacher has continually improved and refined his techniques, trying as best as possible to increase cleanliness, and efficiency, and reduce unnecessary exposure to toxic liquids. He tells me he will soon have a web resource of his methods available, which will be advertised heavily on this blog. I am presently trying to implement some of his methods in my new lab. This is the first time I have had to work with TBE, or with large quantities of MEI, both of which, in my opinion, are completely avoidable.

But what fascinates me is how little attention this necessary step in thermochronology, or geochronology, typically receives. Would anyone dream of asking NSF Facilities for money to upgrade a mineral separation lab? The amount of time and money wasted in mineral separtion is really astonishing. I think though, that one of the reasons these facilities rarely receive the attention they deserve has to do with the hierarchy of the average thermochronology lab. One of the first jobs you delegate with increasing seniority is mineral separation. Right now we have a fleet of undergrads working for us helping crush, grind, and separate minerals. The drive to streamline the procedures is reduced when those of us in charge no longer have to do them. My main reason for trying to improve the set up is primarily because I don't like the idea of an 18 or 19 year old handling liters of TBE on my behalf, especially when there are good alternatives. Although old ways die hard, I think I have convinced a critical mass to support my efforts, and was even able to put in an order with our glassblower this week.

Aside from heavy liquids, I think the most exciting (albeit expensive) recent advancement in the art of mineral separations is the introduction of commercially available electric pulse disaggregators (EPD's). Instead of physically crushing and grinding rocks, EPD's send a pulse of electricity through your rock, which causes minerals to break apart along grain boundaries. The technique has the advantage of retrieving the crystals from the rock intact, that is, you don't run the risk of physically breaking them apart (a huge advantage for separating apatite.) The method was developed originally to work with small and relatively expensive lunar samples, but is amazing in what it can retrieve from a rock. The link above includes a movie showing how quick and easy the process can be. Although it would increase the quality and throughput of samples, they right now are pretty pricey (well, from a geologists point of view, from a college athletics perspective it would cost about 0.33 D1 college football coach yearly salaries, and in my current situation would have only 2 fewer wins.)

I believe wholeheartedly in the garbage in, garbage out philosophy. This is one reason that I think a great deal about mineral separation. I think when it comes to bang for the buck, this could be one of the best ways to improve lab productivity. I'd be interested in other people's experiences with mineral separation, especially if you have used something safer or less toxic that the heavy liquids I described.