Instructions and Tips for Sample Preparation
General Notes
Below, you'll find information about how to prepare samples for Ar-Ar and U-Th/He dating in the Lehigh noble-gas lab. If you have any questions or doubts, please feel free to talk to Peter or Bruce. There are some general procedures and principles that are important to follow for most samples, but many samples have their idiosyncracies, and often you will need to compromise depending on your goals and the material you have available. More than anything, it is important to understand the reason for a particular procedure or sequence of procedures, and whether the procedure is essential or just one of the many possible ways to get the job done.
Generally, we expect all users to carry out their own mineral separations. We are happy to comment and advise on sample quality and purity. We strongly prefer to carry out final sample preparation and loading for irradiation here at Lehigh, and will either do this ourselves or guide you in these final steps. Any costs you have been quoted for analyses do not include the cost of mineral-separation supplies or technician time, and in general we will do this work for you only under special circumstances. Quoted costs do cover the materials and supplies involved in final sample preparation, loading for irradiation, and the irradiation itself.
General Work Flow for Ar-Ar Analysis
Sampling.You can save yourself heaps of time and possibly some woe by spending a few extra moments when sampling. For most Ar-Ar work, a fist-sized specimen is a good amount, allowing you the chance to make a thin section, do some geochemistry, get the minerals you need (including apatite and zircon), and still leave you with a small hand specimen. Of course, you need to look at the rock and be sure it has unaltered minerals of the sort you are after. The piece you bring back should be fresh, trimmed of weathering rinds that complicate separations and lichens and creepy crawlies that might complicate importation. Be sure the hunk you bring back will fit in your crusher; if not, break the sample at the outcrop and bring extra pieces, so that any losses occur where you have the world's supply of your sample in front of you, and not after the hassles and costs of transport. Do not bring back large melon-sized hunks!.
Mineral Separation.We can't give you a guide to this art, but we do have a few tips and there are a few technical issues that are important for you to understand. Given that we now only analyze milligram quantities, keep in mind that little tricks like "papering" micas and hand picking can often get you a good separate very quickly. Not all samples have to go the full route through heavy liquids, magnetic separator, etc. In general, your life will be much easier if you wash off any crushing dust and sieve away fine material.
There are three things you must keep in mind as you plan a separation. First, try to work with as coarse a size fraction as is consistent with clean monomineralic separation and analytical goals. Fine-grained samples are harder to get pure, are nasty to handle once irradiated, and potentially are more prone to recoil artifacts in the reactor. Somewhere between 40 and 60 and 60 and 80 mesh is a nice range. Second, never expose your sample to HCl, sometimes used by people to remove iron staining that comes from the rusting of crusher scrap after washing. Any traces of HCl that survive on your separate can create serious background problems in the mass spectrometer (free hydrogen is always present and combines with 35-Cl to make a mass 36 interference on 36-Ar). You can probably wash this acid from your sample but it is better to be safe. Third, if you use heavy liquids it is essential that you thoroughly clean your samples acetone, particularly if you are dating micas. Adsorbed heavy liquids can be tenacious and lead to hydrocarbon backgrounds inferering with Ar isotopes.
Sample Purity. An essential part of obtaining useful age data is to know what you are dating, and you can't do this with an impure separate. Attempts to obtain a precise age depend on having samples which obey closed-system behavior, and attempts at gettting thermal histories require that you are doing a diffusion experiment on a known system. Depending on the nature of the impurity, even one percent can be sufficient to compromise an analysis to the point where it is a waste of time and money. Clearly K-bearing phases can be an interference (think of how biotite could alter results from relatively lower-K amphibole), but even "inert" material like quartz can cause problems by containing extraneous Ar, or by changing the melting point of samples in the furnace (e.g., you can't count on Kspar breaking down at the incongruent melting point if quartz is present). Always strive for a separate that is 100% pure, and if you simply cannot achieve this, do some point counting to document how pure the sample is and what the impurities are. Because we now date very small aliquots, it is not very hard to hand pick or high-grade a good separate. If you find most grains are altered or dodgy looking, you might want to reconsider why you are dating the sample in the first place. There is nothing the mass spectrometer or subsequent high-power number-crunching can do to rescue you from a lousy sample.
Final Preparation and Irradiation Generally we will handle these final steps at Lehigh, although if you are a regular user you might end up doing some of this yourself. Your separate needs to be washed in distilled water in a sonifier until the water is completely clear. The sample is then washed and sonified twice in 100% purity ethanol. At this point, after drying at 50C, it is usually advisable to sieve away any grain fragments or fines that have been generated during cleaning and sonification. An aliquot of the appropriate size is then weighed and packed in tin-foil or aluminum-foil tubes for irradiation. Sample size will depend on the mineral type and estimated age, and the experiment you want to run (step-heating, total fusion, laser, etc.). The samples are then vacuum-sealed into quartz tubes along with age standards, the tubes are placed into an irradiation jig, and at the reactor the jig is lowered into the core for the specified duration. You will want to consult with us about irradiation duration, as samples of differing ages require different durations, and you may have other factors to consider, like the need for sufficient 38-Ar production to monitor Cl in fluid inclusions, or the need to minimize neutron-generated interferences in samples with low K/Ca.
Analysis Following irradiation and a delay to allow dangerous levels of radiation to decay away, your samples will return to Lehigh from the reactor. We will load your standards and other monitors into our laser system to determine correction factors for interferences, and J-factor curves for the sample tubes. At this point, analysis can begin, but how it does depends on the nature of the experiment you have in mind. You are welcome and encouraged to visit and participate in the analytical work, but there is little to see, and with the automated system, relatively little user intervention is required.
Packaging Your Samples for Shipment to Lehigh: If you are sending us purified separates, please be sure that your vials are tightly sealed. Taping down the lid may be a worthwhile precaution. Short, squat snap-top vials are vastly superior to the thin screw-top vials found in many labs (the tops of these vials tend to rotate off in shipping or snap off at the neck if they're dropped, and they always topple over at an inconvenient time). Send samples to the contact addressed listed on the lab's home page.
Preparing Whole-Rock Basalts for Ar-Ar Analysis
Whole rocks will in general be treated just like any other sample, but there are a few things to pay special attention to.
Sampling. Sample massive material from the interior of flow units. Avoid vesicular material, altered or weathered material, material with calcite veins, etc.
Amount of Sample. For total-fusion analyses, we run far less than a gram: rarely 0.5 g, sometimes 0.2 g. Now that we are performing all analyses on the electron multiplier, typical sample weights are well below 0.1 gram. We are limited by the capacity of our mass spectrometer's digital voltmeter to a signal of less than 12 V, and for most whole-rocks, our sensitivity is good enough that a few hundred milligrams gives us an amount of gas that we can handle easily. It is important to realize that what is limiting is the atmospheric-Ar content of the average basalt. In other words, even for a zero-age sample, more than a few hundred milligrams would probably give us too much gas. However, you will want to have several grams of purified separate, so that we can run multiple aliquots and you can run geochemical analyses, etc. It will probably take something like a fist-sized chunk of raw sample to obtain a few grams of perfect separate with no hand-wringing along the way.
Sample Assessment. Look at a thin section of your sample! Check for weathering or alteration, calcite veins, nature and size of phenocrysts and groundmass, presence of glass, presence of inclusions (xenoliths, or material entrained during movement of flow). Basaltic glass tends not to be a very reliable material for K-Ar dating. If your sample has more than a few percent glass, even if it is very fresh, dating may not be a good idea unless your sample(s) come from an excellent, ironclad stratigraphy, in which anomalous ages have some chance of being detected. Calcite produces CO2 in the extraction system and can cause serious analytical problems. Minor amounts can be treated; major amounts suggest that the sample has been too badly altered to be worth dating. In either case, the presence or absence of calcite must be noted. Phenocrysts, especially olivine and pyroxene, may contain trapped Ar components that do not have an atmospheric 40Ar/36Ar ratio of 295.5 (MORB can have elevated values). This can lead to massive errors when dating young and/or poorly radiogenic rocks. The freshness of mafic phenocrysts can also be a good indicator of the freshness of the specimen as a whole. Zeolites, when present in anything more than trivial amounts, suggest that your sample has been substantially altered. Finally, the minerals in xenoliths may contain significant amounts of trapped Ar (mantle xenoliths; see discussion under phenocrysts) or inherited Ar (e.g., incompletely outgassed feldspars in granitoid fragments).
Crushing, Sizing, Washing. Crushing should be to as coarse a grain size as possible for the purposes of radiation safety, handling under vacuum, and avoidance of 39Ar recoil artifacts. However, crushing must also be to a size equal to or below that of any phenocrysts. It is preferable not have any material crushed to below 100 microns. Material much coarser than 40 mesh causes us some problems in encapsulation for irradiation. Make sure your sieves are perfectly clean before using them!
Before going any further, you will have to thoroughly wash your sample until the rinse water runs clear. Preferably, make the final rinse in deionized water (municipal water is often chlorinated). Crushing dust will foul your heavy liquids and make other separations inefficient. Dry the sample at moderate temperatures of ~50 C.
Removal of phenocrysts (if they are present) will probably require the use of a modified heavy liquid (a mixture of methylene iodide and bromoform for mafic phenocrysts). Note that these two liquids cannot be separated after mixing. Taking sodium polytungstate to the higher densities needed for mafic phenocrysts probably is not feasible due to the high viscosity of this liquid at these densities. You can cut methylene iodide with acetone to get a lighter density, although differential evaporation of the acetone makes this a little less convenient then MI-Bromo mixtures (but this is the way we do it at Lehigh). It is sometimes possible to obtain plagioclase that can be dated as an independent mineral separate, providing something of an internal cross-check on the whole-rock dating attempt. You may, alternatively, have success separating phenocrysts using a magnetic separator. In either case, the separation will not be crisp, since the groundmass of your sample will be polymineralic, and you will have to proceed by trial-and-error (taking three cuts from your sample--end-members, plus a middle one--is usually the best approach).
It is essential that you thoroughly rinse your samples in acetone and remove any traces of heavy liquids! Organic solvents carried into the mass spectrometer are a disaster: you will get a lousy date and the machine may require extensive bakeout.
It is worth trying to remove any magnetic material using a hand magnet. This may be magnetite from the sample, but could also be crusher scrap from your crusher or pulverizer.
Acid treatment If calcite is present in the sample, it must be treated with acid to remove the carbonate. Use dilute nitric acid (a few minutes, especially in an ultrasonic bath, should be sufficient). DO NOT USE HYDROCHLORIC ACID. One of the isotopes of chlorine is mass 35, and this will combine with the hydrogen always present in vacuum systems to make 1H35Cl; this could result in a serious isobaric interference on the 36Ar peak (this will cause an overestimate of the 36Ar content of your sample, and if the Cl contamination is significant, possibly damage the mass spectrometer by raising its background at m/e=36).
Whole rocks will in general be treated just like any other sample, but there are a few things to pay special attention to.
Sampling. Sample massive material from the interior of flow units. Avoid vesicular material, altered or weathered material, material with calcite veins, etc.
Amount of Sample. For total-fusion analyses, we run far less than a gram: rarely 0.5 g, sometimes 0.2 g. Now that we are performing all analyses on the electron multiplier, typical sample weights are well below 0.1 gram. We are limited by the capacity of our mass spectrometer's digital voltmeter to a signal of less than 12 V, and for most whole-rocks, our sensitivity is good enough that a few hundred milligrams gives us an amount of gas that we can handle easily. It is important to realize that what is limiting is the atmospheric-Ar content of the average basalt. In other words, even for a zero-age sample, more than a few hundred milligrams would probably give us too much gas. However, you will want to have several grams of purified separate, so that we can run multiple aliquots and you can run geochemical analyses, etc. It will probably take something like a fist-sized chunk of raw sample to obtain a few grams of perfect separate with no hand-wringing along the way.
Sample Assessment. Look at a thin section of your sample! Check for weathering or alteration, calcite veins, nature and size of phenocrysts and groundmass, presence of glass, presence of inclusions (xenoliths, or material entrained during movement of flow). Basaltic glass tends not to be a very reliable material for K-Ar dating. If your sample has more than a few percent glass, even if it is very fresh, dating may not be a good idea unless your sample(s) come from an excellent, ironclad stratigraphy, in which anomalous ages have some chance of being detected. Calcite produces CO2 in the extraction system and can cause serious analytical problems. Minor amounts can be treated; major amounts suggest that the sample has been too badly altered to be worth dating. In either case, the presence or absence of calcite must be noted. Phenocrysts, especially olivine and pyroxene, may contain trapped Ar components that do not have an atmospheric 40Ar/36Ar ratio of 295.5 (MORB can have elevated values). This can lead to massive errors when dating young and/or poorly radiogenic rocks. The freshness of mafic phenocrysts can also be a good indicator of the freshness of the specimen as a whole. Zeolites, when present in anything more than trivial amounts, suggest that your sample has been substantially altered. Finally, the minerals in xenoliths may contain significant amounts of trapped Ar (mantle xenoliths; see discussion under phenocrysts) or inherited Ar (e.g., incompletely outgassed feldspars in granitoid fragments).
Crushing, Sizing, Washing. Crushing should be to as coarse a grain size as possible for the purposes of radiation safety, handling under vacuum, and avoidance of 39Ar recoil artifacts. However, crushing must also be to a size equal to or below that of any phenocrysts. It is preferable not have any material crushed to below 100 microns. Material much coarser than 40 mesh causes us some problems in encapsulation for irradiation. Make sure your sieves are perfectly clean before using them!
Before going any further, you will have to thoroughly wash your sample until the rinse water runs clear. Preferably, make the final rinse in deionized water (municipal water is often chlorinated). Crushing dust will foul your heavy liquids and make other separations inefficient. Dry the sample at moderate temperatures of ~50 C.
Removal of phenocrysts (if they are present) will probably require the use of a modified heavy liquid (a mixture of methylene iodide and bromoform for mafic phenocrysts). Note that these two liquids cannot be separated after mixing. Taking sodium polytungstate to the higher densities needed for mafic phenocrysts probably is not feasible due to the high viscosity of this liquid at these densities. You can cut methylene iodide with acetone to get a lighter density, although differential evaporation of the acetone makes this a little less convenient then MI-Bromo mixtures (but this is the way we do it at Lehigh). It is sometimes possible to obtain plagioclase that can be dated as an independent mineral separate, providing something of an internal cross-check on the whole-rock dating attempt. You may, alternatively, have success separating phenocrysts using a magnetic separator. In either case, the separation will not be crisp, since the groundmass of your sample will be polymineralic, and you will have to proceed by trial-and-error (taking three cuts from your sample--end-members, plus a middle one--is usually the best approach).
It is essential that you thoroughly rinse your samples in acetone and remove any traces of heavy liquids! Organic solvents carried into the mass spectrometer are a disaster: you will get a lousy date and the machine may require extensive bakeout.
It is worth trying to remove any magnetic material using a hand magnet. This may be magnetite from the sample, but could also be crusher scrap from your crusher or pulverizer.
Acid treatment If calcite is present in the sample, it must be treated with acid to remove the carbonate. Use dilute nitric acid (a few minutes, especially in an ultrasonic bath, should be sufficient). DO NOT USE HYDROCHLORIC ACID. One of the isotopes of chlorine is mass 35, and this will combine with the hydrogen always present in vacuum systems to make 1H35Cl; this could result in a serious isobaric interference on the 36Ar peak (this will cause an overestimate of the 36Ar content of your sample, and if the Cl contamination is significant, possibly damage the mass spectrometer by raising its background at m/e=36).
Sample Preparation for U-Th/He Analysis
At some point we'll be posting a more detailed and illustrated page that focuses on all the steps needed for U-Th/He analysis. For now, here's a brief overview as a placeholder.
Getting Zircon and Apatite. Zircon is easy to obtain. Often even a rock the size of a thin-section blank yields sufficient grains. After getting, washing, and drying a size fraction in the range 60 to 200 mesh, one can go directly to the heavy liquid methylene iodide (MI) to sink zircon, and then use the magnetic separator to pull the fully non-magnetic fraction, which should be mostly zircon. Some people get good results using a gold pan; have fun if you've got this knack!
Apatite is a little trickier, as its lower density makes its separation more subtle, and as a bare minimum you will probably need a little more sample than a thin-section blank, more like a fist-sized piece. We get best results if we place a more tightly sieved size fraction (say 60 to 140 mesh) into a lower-viscosity heavy liquid like tetrabromethane (TBE); our results using the nicely non-toxic but hassle-ridden polytungstate solutions have been mixed. If you have a large sample, you can also use the Wilfley table, but to obtain apatites reliably it is important to pre-sieve and pre-wash the sample so that the table can work efficiently, and you should try to focus on a narrower size range (because large qtzofeldspathic grains can act hydraullically like smaller and denser apatites). Once you have TBE heavies or the equivalent, you will probably need to iterate between MI and TBE until you are sure you have only grains that are light in MI and heavy in TBE, with no entrained grains that have been dragged down or left stuck on the liquid meniscus. Often we find we need to go through the sequence TBE - MI - TBE for a cleanup. The nonmagnetic fraction of this product should be apatite.
Picking, Loading, and Analysis. Once you have your separate, you need to carefully pick inclusion-free, unbroken grains of regular shape, and load them into the appropriate container (small niobium tubelets for both apatite and zircon). Before loading the grains, you need to photograph them with our calibrated digital camera in order to get a record of their size, which is used in determining the alpha-loss correction. You should try to photograph the grains in two orientations, both with the C (elongated) axis parallel to the stage, so that you measure both the width and depth of the grain (few real grains are symmetric this way, and because grains will tend to fall with their two largest dimensions showing, you really want an estimate of the third, stage-normal dimension: this smallest dimension could be most critical for the alpha correction). Only one grain of zircon is required, whereas for young apatites less than 10 Ma you may need 5 or so grains. The sealed capsules are loaded into Nb foil carriers that have been scribed with uniques marks so that they can later be identified. These foil packets are then loaded in the helium line's sample dropper (up to 16 at a time), affixed to the vacuum system, and baked overnight at 70 to 100 C to reduce the hydrogen background. After heating to either 950 C for 10 minutes (apatite) or 1350 C for an hour (Zircon), the sample packets are retrieved from the crucible liner, and weighed to identify them. The actual sample capsules are then sent for U and Th analysis by isotope-dilution analysis using solution ICP-MS.
