(updated 03/24/2008 by Rick A. Knurr)
Lithium Metaborate Fusion
This technique is useful for major, minor, trace, and REE's with the exception of volatile elements, which might be lost in the fusion process, and Li and B, which of course are the major components of the fusion flux.
The fusion flux is a high purity lithium metaborate (LiBO2) of at least five 9's (99.999%) if trace elements are to be analyzed. This is generally in the form of a coarse powder. Fine powder LiBO2 can only be used for major element analyses due to its lower purity (~ 99.9%).
The procedure consists of taking approximately 0.15 grams of rock powder (<100 microns) weighed to the nearest 0.0001 grams and mixing it with 0.750 grams of LiBO2 in a disposable 10 ml polypropylene beaker cup or a graphite crucible. The two powders must be adequately mixed to ensure that all of the sample will fuse properly. For convenience, the LiBO2 is usually weighed out first into all the needed beaker cups. The samples are then weighed onto the flux. This allows the user to avoid unnecessary cross contamination between the bottle of flux and the samples themselves. The sample and flux mixture is then transferred quantitatively to a 10-15 ml graphite crucible (high purity graphite is highly recommended for trace element analysis). You can skip the cup mixing and do everything in the crucible if you are careful. In this case, weigh the flux into the crucible and then weigh the sample onto the flux. For high carbon containing samples you may refrain from mixing the flux and sample. The sample will be easily oxidized in the furnace while sitting on top of the flux. Extra swirling will be required, however, to ensure a homogeneous glass for pouring.
The crucibles are then placed in a furnace at 1000-1050 degrees C for 10-12 minutes to allow the fusion process to occur. For samples where the flux and sample is poorly mixed, remove the crucibles one by one from the furnace, swirl them to mix the liquid and pick up any stray molten particles along the sides, and then replace them promptly into the furnace until the temperature comes back up to at least 1000 deg C.
The crucibles are then removed, carefully swirled, and decanted into a pre-weighed 60 ml wide mouth polypropylene or polyethylene bottle containing 61.50 grams of a solution of 1 N HCl *(please see note below). Any molten material that doesn't transfer should be allowed to cool and then chipped off of the graphite with a spatula and added to the bottle. The bottles should then be capped tightly and placed on a wrist action shaker for approximately 20 minutes. After shaking the solution should be inspected for un-dissolved material. If material is present, this indicates the process has not be quantitative and should be repeated with less sample (same amount of flux) or another procedure should be employed. This solution is stable for about a week without dilution after which there may be appreciable loss of SiO2 as a gel. The fusion procedure should be limited to 8 samples or less at a time in the oven to prevent severe temperature changes as crucibles are removed and the door is repeatedly opened and closed. The final step is to reweigh the bottle and solution and thereby determine a total solution weight. This solution weight divided by the original sample weight is the sample dilution factor.
The HCl (and HNO3, HF) used in the solvent preparation should be trace metal grade or better and the deionized water should be 18 meg-ohm. Bottles should be acid washed and deionized water rinsed and dried before use. The graphite crucibles may be reused if there is adequate height to its sides and the crucible has been scraped with a spatula, polished with a paper lab wipe cloth, and blown clean of debris. For critical applications fusing a blank SiO2 and LiBO2 mixture should be used to clean the crucible.
To analyze on the ICP-OES the fusion solution must be diluted 100 fold with the addition of Yttrium as an internal standard element. The resulting solution should contain approximately 1 ppm Y and be 0.1-0.2 N in acid concentration. Standards and blanks should be prepared using the same LiBO2 and acids used for the samples and should be spiked with NIST traceable single and multi-element standards. The whole procedure should also be checked with USGS (or other certified rock powders) rock standards or their equivalent to ensure accuracy.
* The use of HCl for the solvent may be substituted for by HNO3. This may be advantageous for ICP-MS analysis since it eliminates the ClO interference on Cr and V. The use of HCl, however, eliminates the interference of NO on Si and tends to give the resultant solutions a little longer stability. For high SiO2 samples (>60%) it is highly recommended that 35 g of conc. HF per 1000 grams of solution be added to stabilize the Si numbers. This will also help the refractory elements Nb, Mo, V, and Ta. There is, however, danger involved in adding HF to the solutions. Extreme care must be taken in handling the acid (due to its highly corrosive nature to skin and bone) and all bottles must be extremely clean. The danger of contamination is much greater with the addition of HF since it readily leaches the refractory elements from dirty plastic. I highly recommend using new, cleaned bottles if HF is used.