1920.] Heskett. — Smelting Titaniferous Ironsands. 113 
0-91 per cent, silicon, 042 per cent, manganese, 0*56 per cent, phosphorus, 
0-04 per cent, sulphur, and 0-28 per cent, titanium. However, difficulty 
was experienced by the ferro-coke becoming friable at incandescence and 
clogging- the hearth with finely divided carbon, so the process was finally 
abandoned. 
The promising nature of this experiment led to briquetting one part 
of coal and three of concentrated ironsand with sodium silicate. These 
briquettes were then coked and charged with coke and limestone to 
the same small blast furnace (or cupola). This gave free-running slags 
and clean working, but a smaller yield of a white iron low in silicon 
(04 per cent.) was produced, and more carbon was used per ton of pig. 
This was no doubt explainable by the fact that in the first process some 
ore was reduced in the initial coking, and this being reoxidized at the 
tuyeres a higher temperature was evolved. The briquettes could only be 
coked for thirty minutes—too short a time for any reduction. 
The next step was the partial abandonment of the idea of “ direct ” 
reduction, and the building of a higher furnace, 45 ft. high, with hearth 
4 ft. diameter, bosh 8 ft. high, and bosh-angle 70°. Hot blast at 1,000° F. 
with a pressure of 20 oz. was used, with water-cooled bosh, tuyeres, and 
coolers. Eight parts of concentrated sand were briquetted with one part 
of finely pulverized coal, and the briquettes were carbonized and charged 
with coke and limestone. Five tons per day of high-silicon pig was 
produced (G.C., 2-89 per cent.; C.C., 0-7 per cent.; Si, 1-6 per cent.; Ti, 0-63 
per cent.) with fluid slags containing 30 per cent, of Si0 2 , 17J per cent, 
of A1 2 0 3 , 31 per cent, of CaO, and 10-2 per cent, of Ti0 2 . The titanic 
acid present seemed rather to lower the fusion-point and viscosity of the 
slag, and later on slags with 20 per cent, of TiO a were made without 
raising either. The coke charge was high—at least 35 cwt. per ton of pig. 
The trouble with this process arose from the fact that titaniferous 
accretions built up in the hearth and gradually prevented tapping. These 
accretions were of a greyish stony colour, metallic in fracture and magnetic 
in nature, and were composed of a mixture of ferro-titanium and slag. On 
crushing, about two-thirds proved to be magnetic, and on separation analysed 
66 per cent. Fe, 26 per cent. Ti, 3*7 per cent. Si, 24 per cent. C; while the 
non-magnetic portion was composed of 18 per cent. Si0 2 , 20 per cent. A1 2 0 3 , 
39 per cent. Ti0 2 , 12J per cent. CaO, 6 per cent. MgO. The slag was then 
tested by allowing it to settle in a sand well, when two well-defined layers 
resulted. The bottom layer was of a dark-blue tint, with metallic material 
freely intermixed and occasional small copper-coloured deposits about the 
area of a threepenny-piece. It had a specific gravity of 3-8, and rusted 
if allowed to weather. It was very mushy or pasty even at the highest 
temperatures (3,000° F.) and did not flow at all from the assay crucible. 
It flattened out on beating, and appeared very strong. The analysis was 
26 per cent. Si0 2 , 21 per cent. A1 2 0 3 , 13*7 per cent. Ti0 2 , 23J per cent. 
CaO, 1-3 per cent. S, 11-3 per cent. Fe. The top layer was sky-blue, with 
stony structure, crystalline interior and somewhat vitreous edges, and 
showed no signs of rusting. Its specific gravity was 3-2, and at 2,600° F. 
it was quite fluid and ran freely. The analysis gave 33J per cent Si0 2 , 
15J per cent. A1 2 0 3 , 10 per cent. Ti0 2 , 34J per cent. CaO, 24 per cent. 
MgO, 0-2 per cent. FeO, 14 per cent. S, and 2 per cent, alkalies. 
It appeared probable that a very viscous layer of titaniferous material 
was interposing between the fluid metal and fluid slag between each 
tapping, and that this layer was too mushy to flow with the slag and was 
gradually building up in the hearth and preventing the proper separation 
8—Science. 
