Fusion is a technique used to prepare inorganic samples, with a view to analyze them by x-ray fluorescence (XRF), inductively coupled plasma (ICP),
atomic absorption (AA) or any traditional wet chemistry method. Typical samples include: cements, ores, slag, sediments, soils, rocks, ceramics, pigments,
glasses and even metals.
A fusion can produce either a small, homogenous glass disk (or "bead") for XRF, or an acid solution for other analytical methods.
The process of fusion as a sample preparation method exhibits many advantages
over other methods, as it does not produce mineralogical, grain size or orientation effects and the result is perfectly homogenous. Fusion is also capable of
preparing samples that are very refractory, and otherwise hard to digest.
Temperature and flux
In sample preparation by fusion, the sample never actually melts. It is merely dissolved into a solvent. This solvent, generally a lithium borate flux, is
solid a room temperature and must be liquefied to dissolve anything. This is the only reason the process requires heat. Therefore, the actual fusion temperature
depends almost exclusively on the flux blend.
Lithium tetraborate (LiT) melts at 920°C and has the highest melting point of common fluxes. Lithium metaborate (LiM) melts at 845°C and LiT/LiM blends melt
at lower temperatures.
The specific LiT/LiM blend, in turn, depends on the sample type: alkaline oxides [MnO, MgO, FeO] call for acidic fluxes [more LiT], while acidic samples
[TiO2, SiO2] call for alkaline fluxes [more LiM] and neutral samples [Fe2O3, Al2O3] call for 50/50 mixes.) Also, the addition of a sample can significantly
reduce the melting point of a given flux. This is why one should never exceed 1050-1100°C, as it is not required but more importantly because flux and some
samples evaporate faster with higher temperatures.
Temperature and competition
We just do not think that claiming that a given fusion machine can reach 1600°C is serious, because the Pt-Au crucibles would melt ! This figure is often
the flame temperature, and not the actual temperature inside the crucible. Our 1200°C is the actual temperature in the crucible.
There are several reasons that explain why some fluxer makers claim their instrument can reach 1600°C, but the most obvious one is the impression of power.
The higher the temperature, the more flux volatilizes, and the more inaccurate the measurements. If you check on the web site of any serious fusion specialist,
they will actually recommend temperatures lower than 1100°C, to minimize flux and sample volatilization.
Keep in mind that fusion involves the melting of the flux, which then acts as a solvent for the sample.
If temperature is very high, a little more sample can dissolve in the flux, but this excess will precipitate upon cooling (and possibly crack the bead). Also,
in some cases, the dissolution kinetics are fairly slow, and users are tempted to increase temperature. We prefer to make a slightly longer fusion, and obtain
Non-wetting agents (NWA) act as surfactants that make the melt less prone to sticking to platinumware. Non-wetting agents are halogen compounds (generally
containing Iodine, Bromine or Fluorine) and typical formulations include KI, LiI, LiBr and NaI. Only a few milligrams are required.
We strongly recommend to use such a non-wetting agent, to lengthen the mold’s life expectancy.
General directions (typical fusion example)
- Grind your sample <100µm
- Accurately weigh about 0.5 g sample in a 95%Pt/5%Au crucible (actual mass will depend on sample solubility in flux)
- Accurately weigh about 7g of flux (actual total mass in crucible will depend on the mold diameter)
- Add a few milligrams of non-wetting agent (the NWA, typically LiBr, prevents the melt from sticking to the crucible)
- Place the mold and crucible in the fluxer
- Select a method from the instrument's memory press "OK".
If one wishes to make solutions only (for ICP or AA), the preference goes to straight LiM, which has a great dissolution power, but would lend crystallized
disks. For solutions, amounts of sample and flux are smaller (about the half), and fusion times are shorter.
In all cases, the best guideline is to always fuse at the lowest practical temperature. If some prefer higher temperatures, it is most certainly because
their recipe is not quite optimal.
When the sample to analyze contains metallic species, one must pay great care that those species be oxidized before the temperature reaches the point where
alloying between the crucible and the sample.
If metallic species remain in the crucible, a potentially irreversible damage to the platinumware may occur, as well as a loss of sample. Therefore, one must
select a suitable method to fully oxidize the sample:
- Natural oxidation. This is for samples that spontaneously oxidize rapidly at high temperatures. No need to add anything, as these samples will just
combine with ambient oxygen to form stable oxides.
- Liquid oxidation. This involves the addition of a liquid base or acid to oxidize the material to fuse. The choice depends on the sample type (two
reagents may be required in some cases). Typically, the sample is weighed in the platinum crucible, and a few milliliters of acid/base are added. The
crucible may be placed on a hot plate (or directly in the K1) to promote the reaction at temperatures below 100°C (not possible on gas-fired machines). Then, the
flux is added and the fusion is carried out in the fluxer.
- Solid oxidation. This generally involves the use of nitrate or carbonate compounds. They are mixed with the sample, and sometimes with just a little
flux. A first heating step is performed in the fluxer, and 250~750°C, depending on the oxidizer type. Then, if required, more flux is added, and the fusion is
- Calcination. This is normally performed in a regular lab oven, in ceramic crucibles, around 1000°C. This operation burns carbon to CO2. If there are
no other components to oxidize, the sample is transferred to a platinum crucible, the flux is added and the fusion is carried out in the fluxer.
- Make sure your samples are oxidized, or oxidize them using a suitable method (Katanax team can help you on that).
- Weigh components accurately (fusion will not likely introduce any analytical error, but weighing might).
- Use a non-wetting agent, for a better melt transfer from the crucible and an easier bead extraction from the mold.
- Fuse at the lowest practical temperature.
- Enjoy the results !