INAA Services at Washington UniversityAdvantages and Disadvantages
| Services Provided | Our Requirements
Dr. Randy L. Korotev and colleagues in the lunar and planetary geochemistry laboratory analyze rocks and minerals by instrumental neutron activation (INAA) in conjunction with research on lunar samples and their terrestrial analogs. We also analyze samples for other researchers if those analyses to not conflict with our own research efforts and do not require that we deviate substantially from our standard procedures. For this service we charge a fee and relinquish any further interest in the results of the analyses. Services ProvidedFor most rock samples we encounter we are able to provide concentration values or upper limit values (if the concentration is below our detection limit) for the following elements: Na, Sc, Cr, Fe, Co, Sb, Cs, Ba, La, Ce, Sm, Eu, Tb, Yb, Lu, Hf, Ta, Th, and U with precisions of 1-10% (one standard deviation). Data for Ca, Ni, As, Br, Rb, Sr, Zr, Nd and W may also be obtained, but usually with poorer precision. In special cases, we also obtain signals for Zn (samples with a high Zn/Sc ratio), Se (5 µg/g), Ag (5 µg/g), and Ir and Au (10 ng/g). However, we do not routinely include standards for these elements. Thus if you are interested in obtaining data for Au, for example, let us know ahead of time and we will include a standard. This will allow us, at least, to obtain an upper limit estimate of the concentration. Note that because we do not do the first radioassay of the samples until the sixth day following the irradiation, we cannot provide data for some elements (e.g., Al, Cl, V, Mn, Dy) that are determined via isotopes with short half lives, as do some other INAA labs. The detection limit for an element varies greatly with composition of the sample and we can provide no universally applicable list of detection limits. However, for a list of typical results (with uncertainties and upper limits) for a variety of different rock types, click here. For certain nuclear reasons and consideration of laboratory contamination, we cannot provide data for certain types of samples. These include most ores, organic samples, samples containing more than several µg/g of boron (e.g., tourmaline), and samples with very high concentrations of rare earth elements (1000 times chondritic concentrations) or uranium. We are also reluctant to analyze synthetic materials because concentrations of some trace elements for which we are very sensitive may be unpredictably high. For a summary of the advantages and disadvantages of INAA, click here . Our RequirementsSamples should be provided to us as powders or as grains not exceeding 3 mm in longest dimension. We do not prepare powders from hand specimens. In INAA, no sample dissolution is required. We can analyze "as-is" any sample that fits in our silica sample tubes (usually 3 mm inside diameter and up to 3 cm long). For whole-rock analyses, it is usually best to provide us with powders for sampling reasons. However, for mineral separates or very fine grained or glassy rock samples, chips and grains are also acceptable and, in fact, easier to work with. (In our own lunar work, we analyze whole breccia clasts and soil particles and then make thin sections of them after INAA.) We have analyzed samples ranging in mass from 40 µg to 500 mg. For whole rock analyses, we prefer masses toward the high end of that range in order to minimize sampling problems. For powdered or granular whole-rock samples, the mass we analyze will typically be about 150-350 mg (depending on powder density). For handling ease, we prefer to be provided with 0.5 to 2.0 g of sample in a small glass or plastic vial (the excess will be returned). Each vial should be labeled with your sample number or designation (preferably not exceeding 12 characters in length). To minimize the possibility of contamination, samples are taken from the vial simply by pushing the open end of a 5 mm (outside diameter) silica tube into the powder; thus oversized vials or vials with small openings are a nuisance. We can analyze much smaller samples, but this is more costly because of the additional handling and longer irradiation and radioassay times required. If you know that some of your samples have unusual compositions or much higher than usual concentrations of some elements (e.g., Ba, REE, Ta, U), we would like to know this in advance. It is helpful to us if you include with the samples a document that verifies the sample numbers, identifies the rock types, and describes anything you know to be unusual about the sample compositions. Please include your address and phone number on the such documentation. We also appreciate knowing from you what it is that you expect from us. Are you particularly interested in the rare earth elements? Sb/Ta ratios? Ni concentrations? Knowing these things in advance sometimes helps us design our radioassay schedule and provides a focus during data reduction. For example, As and W have short half-lives and the most precise data for these elements are obtained on samples radioassayed the early in the schedule. Tungsten CarbideWe prefer not to analyze samples that have been ground in tungsten carbide for two reasons. First, such samples are grossly contaminated with W and significantly contaminated with Co and Ta, which are deliberate additives in commercial tungsten carbide. Second, both W and Ta introduce a number of peaks in the gamma-ray spectrum that interfere with analytical peaks of other elements. As a result, precision for some elements degrades. In some cases, we are unable to obtain data for some elements in the presence of gross tungsten carbide contamination (100 µg/g W, which is not atypical for quartz-bearing rocks ground in tungsten carbide). Data reduction for samples ground in tungsten carbide is more complicated and we may charge a surcharge for such samples. If samples ground in tungsten carbide are provided to us, we wish to know of this fact ahead of time. We also recommend that you send us a sample of silica sand ground in the same grinder. This will provide you with the information necessary to apply corrections for Co and Ta contamination. (Co/W and Ta/W ratios are highly variable among samples of commercial tungsten carbide.) Routine AnalysesIn our routine analyses, samples are irradiated for 12 hours in a thermal neutron flux of 5x10 13 cm -2 s -1 in the University of Missouri Research Reactor . We radioassay each sample three times, first for about 10-15 minutes between 6 and 7 days following irradiation, second for 1.5-2 hours between 7 and 11 days following irradiation, and finally for at least three hours (at a closer sample-to-detector distance) sometime after 28 days following the irradiation. If you have some reason to believe that this schedule is inadequate for your purposes, e.g., that extra radioassays or longer counting times are necessary to increase precision (perhaps only for a subset of your samples), then please discuss this with us prior to the analysis. Doing three radioassays is essential to our quality control and we are generally not willing to provide data based only on one or two radioassays as a means of reducing the cost or providing data sooner. There are commercial laboratories available to provide this service. In INAA (actually, gamma-ray spectrometry) we receive signals for all elements simultaneously and our computer programs routinely work up data for all elements that we expect to observe. There is no substantial savings to us in time or effort to provide you with data for fewer elements than we routinely determine, particularly if we have to determine the usual suite of elements for other samples in the batch. Thus we generally do not offer a reduced rate for a subset of the elements we determine. However, for a large batch of samples for which you are only interested in a few elements, particularly if they are determined via isotopes with half lives exceeding several weeks, we would be willing to offer a reduced rate. A more complete description of our analytical method is given in Korotev (1991) . Time Required for AnalysisBecause INAA involves measurement of radioactive decay, it is inherently a slow technique. With no sample backlog, it takes about eight weeks from the time we receive samples until we have an analytical report. In practice, we are seldom able to provide data in less than three months from receipt of samples and it sometimes takes longer. The type and number of samples also affects the length of time required. We submit samples for irradiation in batches of 50-100 samples. We try to group samples of similar composition and mass in a batch. The analysis of a set of, e.g., four granite samples is likely to be delayed until we acquire more samples of similar composition. A 300-mg whole-rock sample will not be analyzed in the same batch as a 20-mg plagioclase separate from the same rock. We do not analyze batches of samples with small masses (<100 mg) as often as large-mass (200 mg) batches. In general, if you send us a batch of 50 samples they are likely to be analyzed sooner than a batch of 4 samples. If we have agreed to do a batch of samples, we highly encourage you to send the samples to us as soon as possible. Occasionally, we have scheduled irradiation dates for people who have promised to send samples and the samples do not arrive in time. Regardless of any commitments we make, the samples in our possession are the ones analyzed first. We make every effort to provide an accurate estimate of when the data will be available to you. However, we can make no promises because unforeseen circumstances may cause delays (equipment failure, unscheduled reactor shutdown). Secondary Standards and Precision ChecksWe are surprised at how seldom researchers test the quality of the data we provide. Seldom do researchers deliberately include samples of a well-analyzed material disguised as an unknown, although we encourage them to do so. We are seldom asked to analyze replicate subsamples. However, we are often asked, sometimes several years after the analysis, "How good are these numbers?" We provide an estimate of the analytical uncertainty with each datum we obtain. The value is an estimate of the precision (one standard deviation) that we would expect to obtain from multiple analyses under identical conditions of subsplits of the sample if the subsplits are actually identical in composition. The estimate is based both on counting statistics and our experience with analyzing each particular element. Actual precision can only be judged by analyzing replicate samples, however. Although we have great confidence in our overall accuracy ( Korotev, 1996 ), you can judge the accuracy of the data we provide for your samples only by inclusion of samples of one or more well-characterized materials of composition similar to your samples. Extra samples cost extra money, which researchers are usually reluctant to spend. We are similarly reluctant, after the fact, to speculate on the actual accuracy and precision of the data we provide when a better-designed experiment could have provided data that would have avoided the need for speculation. We always include with each batch of samples some GRS's (geochemical reference standards), such as basalt BHVO-1 or obsidian NIST SRM 278. We treat these GRS's as unknowns, except that we usually radioassay each GRS at least six times for longer durations than other samples, thus they are not "typical" analyses. The goal of these GRS analyses is to obtain high-precision data so that we can check the accuracy of a particular batch of samples against averages obtained over many years. We do not routinely provide these results to researchers, although the cumulative results of 11 years of data are available in Korotev, 1996 . Thus if you intend to submit samples from a single project to us to be analyzed in different batches, perhaps over several years, we encourage you to select homogeneous reference samples from the first batch that are reanalyzed in all subsequent batches as monitors of batch-to-batch precision. We also suggest, especially if you intend to submit a large batch of samples, that a few duplicate samples be included in each batch. This is particularly important for coarse-grained or highly evolved rocks for which 250-mg subsplits may actually differ from each other in composition. Feel free to include samples of GRS's. We prefer that you supply your own, but if you wish us to include a sample of a particular GRS with your batch of samples and have us treat it as a typical unknown, let us know ahead of time; we have a variety of such materials, e.g., NBS/NIST (U.S.), U. S. Geological Survey, GIT-IWG (French), ANRT (French), CRPG (French), Geological Survey of Japan (GSJ), South African Bureau of Standards (Mintek), Institute of Geophysical and Geochemical Exploration (IGGE, P. R. China). If it is important to you to know what subtle compositional differences may exist among a small set of related samples, identify these samples for us and we will attempt to radioassay the samples in a way that optimizes precision. Publication of ResultsIf you publish papers based on analyses we provide, we ask only that you (1) mention that the analyses were done at Washington University and (2) indicate in the tables or figures the magnitude of the uncertainties we include in our report to you. The best references for the analytical technique are Korotev (1991) and Korotev (1996). BibliographyLindstrom D. J. and Korotev R. L. (1982) "TEABAGS: Computer programs for instrumental neutron activation analysis." J. Radioanal. Chem. 70 , 439-458. Gromet L. P., Dymek R. F., Haskin L. A., and Korotev R. L. (1984) "The 'North American shale composite:' Its compilation, major and trace element characteristics." Geochim. Cosmochim. Acta 48 , 2469-2482. Korotev R. L. and Lindstrom D. J. (1985) "Interferences from fission of 235U in INAA of rocks." Trans. Amer. Nucl. Soc. 49 , 177-178. Korotev R. L. (1987a) "National Bureau of Standards coal flyash (SRM 1633a) as a multielement standard for instrumental neutron activation analysis." J. Radioanal. Nucl. Chem. , Art. 110 , 159-177. Korotev R. L. (1987b) "Chemical homogeneity of National Bureau of Standards coal flyash (SRM 1633a)." J. Radioanal. Nucl. Chem. , Art. 110 , 179-189. Korotev R. L. (1988) "Error in neutron activation analysis from recoil-implanted fission products from uranium in aluminum foil." Geochem. J. 22 , 133-137. Korotev R. L. (1991) "Geochemical stratigraphy of two regolith cores from the Central Highlands of the Moon." Proc. 21st Lunar Planet. Sci. Conf. (ed. V. L. Sharpton and G. Ryder), pp. 229-289. Lunar and Planetary Institute, Houston. Korotev R. L. (1996) "A self-consistent compilation of elemental concentration data for 93 geochemical reference samples." Geostandards Newsletter 20 , 217-245. Cost ScheduleCost of service is arranged prior to analysis on a case-by-by case basis based on the following guidelines. Base cost. Analysis costs are based primarily on the number of sample tubes irradiated. Our base cost for routine analyses is $140 per sample tube plus a setup fee of $200 per group of samples that can be analyzed in a single batch (i.e., similar mass and composition). Thus the base cost for a set of 10 basalt samples and 10 granite samples is $3200 (2x[$200 + 10($140)]). A single sample of any kind will cost at least $340. Duplicate samples each count as separate samples and reference materials treated as unknowns each count as a separate sample. To assure that samples are analyzed in one batch, they must be received in one shipment. Surcharge. Analysis costs are based secondarily on a number of factors:
For these services, we increase the base cost by a per-tube surcharge negotiated for each case. Normally, this will not result in a total cost exceeding $170/sample. Discount. For routine analyses of large numbers of samples of similar composition the base price may be discounted. Contact Information
|