FUSION USING INDUCTIVE HEATING - FUSION TECHNOLOGY AND DEVELOPMENT

5.3 FUSION TECHNOLOGY AND DEVELOPMENT

5.3.1 FUSION USING INDUCTIVE HEATING

Early development work with fire assay involved the investigation of induction heating. In this application a graphite crucible was used to perform the fusion. With an induction generator it was possible to heat the crucible and perform a conventional fire assay fusion. This fire assay innovation was pioneered by Mr. Willie Brits, a former chief chemist at Anglo Platinum’s Potgietersrus laboratory, who later patented the process [149]. Considerable development work was done in conjunction with Mr. Brits [150]. There were two basic designs for induction furnaces [151]:

• Ajax-Wyatt (cored type)

• Leeds-Northrup furnace (uncored type)

The induction heating for fire assay applications was best performed with the second type of furnace. In that design the crucible was located inside a water-cooled induction coil. Eddy currents were induced in the conductive crucible to heat it. Essentially the inductive heating was a special form of resistive heating.

The heat generated in the conductor is proportional to its resistance and to the square of the current induced. This is given in Equation 5.1

c 2 1 2 2 2R R P=Ι = Ι _{Equation 5.1}

The currents Ι₁and Ι₂are the currents in the primary and secondary circuits. Similarly the R2 is the resistance in the secondary circuit and Rc is the effective coupled resistance the load causes in the primary circuit. With some mathematical manipulation, Equation 5.2 is derived.

( )

ρµ Ι π =4 r n l f P 2 0 2 1 2 _{Equation 5.2} Where:

P Power (W) µ Effective permeability of charge (Ω.cm.Hz-1₎ r0 Radius of cylindrical charge (cm) n Turns of coil per cm (cm-1)

f Supply frequency (Hz) l Length of charge (cm)

ρ Resistivity of charge (Ω.cm-1×10-9) Ι Current (A)

The basic formula and its derivation are not important but the concepts denoted by it are. The induction of a current in the charge/conductor is dependent on the geometry and the material of the charge. It also depends on the geometry of the induction coil itself.

About 90% of the heat is generated within a thin layer on the surface of the conductor, the rest of the charge is heated by conduction. The depth at which the heat is generated is dependent on the frequency used, the higher the frequency the thinner the layer. Also the layer will be thinner for ferromagnetic materials that have a high effective permeability and thick for non-magnetic materials. The induction depth is also dependent on the resistivity. Resistivity changes with temperature and typically for a conductor like iron, the resistivity will increase with temperature and hence the inductive depth will increase. An induction furnace must be built for a specific purpose with regard to size, geometry and the type of load that will be used.

The first trials with an induction furnace, a high frequency generator operated at 6.5kHz with a maximum power output of 10kW was used. It had a cylindrical induction coil 15cm tall, 15cm in diameter with 6 turns. The coil was encased within a ceramic insulator to protect it from mechanical wear. During operation the coil was cooled with continuously flowing water at about 5-10L per minute supplied with an evaporative cooler.

The induction furnace was used to heat a conductive graphite crucible. The crucible then provided heat for a fire assay fusion.

The initial test work involved a conventional fire assay flux without a reductant, and the pot provided the carbon necessary to reduce the litharge. Due to the fact that there was only one crucible, the test work was performed with a preheated crucible. The crucible was simply loaded with the charge from the top in situ. The loose flux mixture was either poured into the pot or alternatively introduced inside a thin polyethylene plastic bag.

The fusion was visually inspected from above and the fusion took between 15 to 25 minutes to get to a quiescent state. This initial observation agreed with the time line that was shown in Section 2.4 where the first 20-30 minutes of a standard fire assay fusion was required to get the crucible and charge up to the fusion temperature. Therefore, by using a pre-heated crucible the fusion time could be considerably reduced.

Once the fusion was complete, the carbon pot was raised with a lever system and removed with a pair of tongs. The melt was poured into a conventional iron mould. The slag was manually removed and the buttons were analysed using cupellation or lead dissolution methods.

Some of the key findings in this early work were:

• The addition of the charge to a preheated crucible could reduce the fusion time by a half or more with a standard fire assay flux.

• Removing the crucible from the top reduced heat exposure of the operator.

• There was a small amount of spitting and physical loss of the charge.

• Degradation of the carbon crucible caused graphite particles in the slag with subsequent losses of lead collector droplets due to entrainment.

• Results were reasonably accurate and precise and the method showed promise.

Further development of the induction heating method involved more sophisticated and dedicated fusion machines. The initial prototype was a mechanical piston type machine that simply loaded and unloaded the induction furnace.

The machine used a conventional induction generator at 13kHz with a maximum power output of 10kW. There was a similar coil construction to what was used in the manual operation. The coil was in a fixed position approximately 1.2m from the ground. The mechanical system operated with a crankshaft that was used to raise and lower the carbon pot. The entire assembly was rotated for loading and pouring of the melt. A ceramic sheath was cast around the carbon pot. This had a dual function, firstly to protect the outside of the pot from oxidation in air and secondly to give the pot better mechanical resistance. The instrument poured into a standard cast iron mould.

The second generation of induction machines were manufactured locally by a company Hiprom [152] under licence from Mr. Brits. The basic principle remained the same but the entire fusion head with the crucible was turned to cast the melt. A diagram of the second induction furnace melting head is given in Figure 5.5.

For automation purposes the flux was put into a plastic “pop-top” container where the sample could be weighed into the flux and the container sealed. The container could then be placed directly into the melting head. The problem was that the plastic produced residual carbon after combustion. This caused entrainment problems of the lead collector in the slag. Changes to the plastic by using calcium carbonate as a filler, proved ineffective in controlling the carbon particles in the slag.

A significant development was the use of a secondary slag separator. This volumetric separator worked reasonably well and is discussed a little later in Section 5.4.2. This first machine was purchased by Anglo Platinum and tested extensively. There were some significant flux developments and again these will be discussed separately Section 5.5.

Yet the technology turned out to be premature. The materials of design and engineering were inadequate. The machine could not be reliably operated for long enough periods to evaluate it properly. Many of the failures were associated with the furnace on the slag separator.

Figure 5.5 Second generation induction furnace

The induction furnace concept was licensed to Spectro Analytical Instruments who developed the third generation machine called the Immafuse [153].

The volumetric slag separator was incorporated within the fusion head. The melting pot design was improved and boron nitride inserts were placed into the carbon pot for better chemical resistance.

Sample container Suction cup and loading arm

Furnace retainer lid