The choice of the furnace strongly depends of the exact compositions you plan to melt. Here are some general remarks and advises:

(1) In silicon alloys you always have to take intermetallic phase formation into account. These phases are normally very brittle and make related alloys difficult to handle. So carefully check phase diagrams (in case of binary alloys) or simulate the phase diagram, e.g. by ThermoCalc.

(2) Silicon has a high affinity to e.g. Oxygen or Nitrogen, so you need to do your melt either in vacuum or in inert gas atmosphere. The use of standard air furnaces and/or ceramic crucibles is not advisable. Use Zirconia crucibles if possible.

(3) If you have alloying elements like Magnesium or Manganese (both have low evaporation temperatures), vacuum furnaces are not optimal as the alloy composition is not easy to control.

(4) Good opportunity for alloy melting with minimised contamination is the cold-hearth technique. Here, the alloying elements are placed in a water-cooled copper crucible. Heating is possible by e.g. induction (induction furnace with cold crucible) or plasma beam (PB-CHM). Induction melting is easy, PB melting needs some experience.

(5) In induction melting, the typical quantities are relatively large for lab scale (> 1kg), PB-CHMs are available from 10g. We have two PB-CHMs (one for 10g, approx. 50k € and one for 400g, approx. 350k €).

(6) If you only want to produce few alloys, it might be better to simply look for a subcontractor instead installing a furnace.

If you need further help, just let me know.

The choice of the furnace strongly depends of the exact compositions you plan to melt. Here are some general remarks and advises:

(1) In silicon alloys you always have to take intermetallic phase formation into account. These phases are normally very brittle and make related alloys difficult to handle. So carefully check phase diagrams (in case of binary alloys) or simulate the phase diagram, e.g. by ThermoCalc.

(2) Silicon has a high affinity to e.g. Oxygen or Nitrogen, so you need to do your melt either in vacuum or in inert gas atmosphere. The use of standard air furnaces and/or ceramic crucibles is not advisable. Use Zirconia crucibles if possible.

(3) If you have alloying elements like Magnesium or Manganese (both have low evaporation temperatures), vacuum furnaces are not optimal as the alloy composition is not easy to control.

(4) Good opportunity for alloy melting with minimised contamination is the cold-hearth technique. Here, the alloying elements are placed in a water-cooled copper crucible. Heating is possible by e.g. induction (induction furnace with cold crucible) or plasma beam (PB-CHM). Induction melting is easy, PB melting needs some experience.

(5) In induction melting, the typical quantities are relatively large for lab scale (> 1kg), PB-CHMs are available from 10g. We have two PB-CHMs (one for 10g, approx. 50k € and one for 400g, approx. 350k €).

(6) If you only want to produce few alloys, it might be better to simply look for a subcontractor instead installing a furnace.

If you need further help, just let me know.

Thanks, indeed the idea of let some else do the job is an option.

Following Carsten Siemers suggestion, copper heathed TIG button furnaces are relatively common for making small experimental heats of alloys. There is some art involved making good homogenous heats and grains tend to be large. Folks like http://www.thermaltechnology.com/arc-melting-furnace.html or

http://www.mrf-furnaces.com/acrmelt_sa200.htm or

http://www.vacuum-furnaces.com/laboratory-furnaces.htm

make them in the US. For the cheaper approch there is

http://blog.modernmechanix.com/experimental-arc-furnace-melts-anything/

If you are only making a few, see if you can get someone else to do it.