The McAllister Ta-Sn deposit, located in Coosa County, Alabama, consists of a series of complex pegmatite dikes and pipes which intrude a granite pluton formation. Wodginite ((Ta, Nb, Sn, Mn, Fe)16O32) is the primary tantalum mineral. The deposit was discovered in 1982 by Callahan Mining Corporation, of Phoenix, Arizona, through stream-sediment reconnaissance and soil geochemical survey methods.

Sample Description

The samples used for evaluation were supplied by O’Dell Construction Company from stocks of previously prepared concentrate. Sample evaluations included chemical analysis, x-ray diffraction (XRD), and scanning electron microscopy with energy dispersive spectroscopy (SEM-EDS).

When using cross-belt high-intensity magnetic separation equipment it was found that the cassiterite reported to the non-magnetic fraction and the tantalum minerals reported to the magnetic fraction. The effects of several variables were then studied: field strength, preliminary grinding, and multiple passes through the separator.

Field Strength

The effect of field strength was evaluated by repeatedly passing a quantity of Sample A through the cross-belt separator at successively increasing field strength. The non-magnetic fraction from using lower strength fields was reprocessed at higher field strengths. Figure 1 shows the cumulative amounts of tantalum, niobium, and tin recovered in the magnetic fraction. The spacing between the magnet head and the feed belt was kept constant at about 5 mm.

The magnetic product obtained at the highest field strength (0.72 Wb/m²) had a relatively higher level of tin contamination than the low field strength products. This material contained considerably more cassiterite than those products obtained at lower field strength. Much of this cassiterite was locked with magnetic minerals. A plot of the Sn/Ta weight ratio for individual products at each field strength emphasizes this jump in tin contamination (figure 2). The non-magnetic cassiterite byproduct, which passed through the separator at 0.72 Wb/m², had a Sn/Ta weight ratio of 44.

Multiple Pass Separation

Tests were performed to determine how many times the magnetic material should be passed through the separator to remove non-magnetic minerals. Quantities of sample A were dry ground in a rod mill to -589 µm and a portion was passed through the cross-belt separator at 0.38 Wb/m². The magnetic tantalum product was sampled by splitting, and the remaining magnetic product was reprocessed three additional times through the separator and sampled by splitting after each pass.

Effect of Particle Size

A test was performed to determine the effect of particle size on separation and recovery. Quantities of sample A were sized by screening at -1180 +589, -589 +300, -300 +147, and -147 µm. Each sample was passed through the cross-belt separator at 0.65 Wb/m². The non-magnetic cassiterite portion was again passed through the separator to maximize tantalum recovery. The products were assayed to determine tantalum and tin distributions.

Methods of Separation

Evaluations were conducted to compare magnetic separation methods involving no grinding, preliminary grinding, and grinding of a middling product. A description of each process flow sheet follows.

The first separation method (figure 4) involved the following steps:

1. The sample was passed through a cross- belt separator operated at 0.72 Wb/m² to obtain a magnetic tantalum concentrate. A scalper on this separator removed tramp iron and highly-magnetic minerals at about 0.08 Wb/m².
2. Magnetic material from step 1 was reprocessed and cleaned on a second separator at 0.72 Wb/m².
3. The non-magnetic material from steps 1 and 2 was reprocessed to clean the cassiterite and recover fugitive tantalum. The magnetic tantalum concentrates recovered from each of the three steps were combined into a final tantalum product.

 

 

Only two flow sheets were evaluated using sample B. Preliminary grinding to -589 µm followed by magnetic separation produced results nearly identical to those obtained when no preliminary grinding was used.

Sample C was tested using only the method that involved grinding of a middling product. While the amount of tin removed was quite high (78%), tantalum loss to the tin byproduct was somewhat more than desired (4.5%). The loss of tantalum to the tin byproduct was proportional to the amount of tin byproduct obtained. Tin byproducts using samples A and C assayed 4.6 and 5.0% Ta2O5, respectively.

Application of Results

Tests showed that Alabama tantalum concentrates with 22 to 29% SnO2 can be magnetically upgraded. As much as 78% of the tin was removed during magnetic separation. The resulting cassiterite byproduct contained 70 to 74% SnO2. The magnetic tantalum product contained 54 to 60% Ta2O5, 9 to 11% Nb2O5, and 10 to 12% SnO2.