Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 73, 1943-44
This text is also available in PDF
(421 KB) Opens in new window
– 76 –

Some Features of Heavy Mineral Separations.

[Read before the Wellington Branch, February 1, 1943; received by the Editor, February 2, 1943; issued separately, September, 1943.]

Experiments have shown that unless special methods are adopted for heavy mineral separations from fairly fine-grained sediments, anomolous results will be obtained. Data obtained using the normal cylindrical separating funnel and the centrifuge are compared. These results appear to justify the suggestion that when heavy mineral residues are required from sediments of fine-sand grade or finer, the centrifuge should always be used.

Recently, while the writer was carrying out heavy mineral analyses of some Tertiary sediments, chiefly of fine-grain size, it became apparent that the residues prepared by flotation in bromoform of specific gravity 2.85 in the ordinary cylindrical separating funnel and by centrifuge procedure, were far from identical. In fact, in a few cases, it was possible to obtain a residue of only 10–20 grains by the standard separating funnel procedure, although 20 gramme portions were used. Quantitative experiments were then carried out and residues were prepared from the sediments by the two methods. For grain-counts for correlative purposes, screened samples are usually employed as recommended by Dryden (1932), Rubey (1933), Gogen (1935), and Russell (1936), but for the purposes of this investigation the total “heavies” from unscreened sediments were used.

Only little need be said here regarding technical procedure. Sedimentary petrologists are well aware of the value of the centrifuge when the separation of heavy fractions from very fine-grained sediments is involved and with the procedures described by Brown (1929), Schroeder (1930), Kunitz (1931), and Berg (1937) for such separations. Recently, however, Moshev (1935) has designed a special centrifuge for rapid and accurate, uninterrupted separation of heavy mineral residues using considerable quantities of sediments. Several types of centrifuge tubes have been proposed, such as those of Schroeder (1930), Kunitz (1931), Taylor (1933) and Bertholf (1940). The writer favours a modification of the Taylor tube in which the lower portion of the tube below the constriction is blown to a diameter equal to that above the waist. Each tube has a total volume of approximately 100–110 ccs., while the bulb below the waist has a volume of about 6 ccs. This modification enables a large residue to be accumulated before it is necessary to wash out the heavy fraction, and as the writer also uses the centrifuge extensively for the separation of minerals in the pure state from rocks for chemical analysis, the large capacity bulb is of considerable value.

All the samples for this investigation were centrifuged for 3 to 4 minutes at a speed of 3,500 revolutions per minute, the centrifuge arm having a radius of 85 mm. The light fraction was then poured off and as much as possible of this material adhering to the sides of the centrifuge thimbles washed off with undiluted bromoform. The thimbles were then refilled with heavy liquid, the

– 77 –

plug in the waist of the tube being removed and the stoppered tube shaken, then re-centrifuged briefly for a second period; thus the very small amount of light fraction remaining was removed. The heavy fraction was then recovered. The procedure with the separating funnel merely involved careful stirring and shading of the column from bright light. Stirring was continued intermittently for a period of 20 minutes. Actually the time required to perform a separation from start to finish was about 10 minutes for the centrifuge and about 25–30 minutes for the separating funnel. It was possible to carry out an average of seven separations in the hour using the centrifuge and two thimbles, or 12 an hour when employing four thimbles, while an average of 4–5 separations per hour were possible with a battery of two cylindrical funnels.

Picture icon

Fig. 1–Cumulative–frequency curves of the sediments investigated.

– 78 –

The sediments were mainly fine-grained types with silt and clay fractions from as low as 0.5 to as high as 93 per cent. The coarse grading of these sediments is shown in text figure 1. From these fairly varied sediments the residues obtained by separating funnel and centrifuge procedure were weighed, and portions mounted in Canada balsam ready for grain-counts and study. A comparison of the weights of the heavy mineral residues obtained by the two methods is diagrammatically represented in text figure 2. When text figures 1 and 2 are compared it will be observed that there is a tendency for the residue obtained by the centrifuge methods to be very much greater than that obtained by the separating funnel in the case of those sediments that are composed chiefly of silt or finer grades. Sediments A, C, B, F, H, and I show this particularly well, although in the fine-grained sediments B, D, and G, this feature is not so marked. On the other hand, in the coarser sediments M, where the amount of silt-grade is negligible the weights of the residues obtained by either two methods are almost the same, while with sediment K, results

Picture icon

Fig. 2—Diagram comparing the percentages of heavy residues obtained by the separating funnel and centifuge methods.

– 79 –

are moderately good. Actually one would expect that the ratio R* would have become less and less concurrently with the decrease in the percentage of the finer grades (silt and clay) in the sediments. However, this is only approximately so, and probable reasons for this apparent anomaly may be dependent upon several causes, of which the following appear to be important:–

  • (1) Grain-size of the heavy residue minerals themselves is of considerable importance; if the average grain-size of the residue is fairly coarse, e.g., 0.2 mm. as in specimen K (see text figure 2), then a fairly representative residue may be obtained by ordinary methods. On the other hand, with specimen H, where the preponderance of the heavy mineral grains are about 0.1 mm. in diameter, the ratio R is high, while this ratio is even greater with E, where the maximum grain-size, excluding micaceous minerals, does not exceed 0.1 mm.

  • (2) The presence of minerals such as chlorites and micas appear to have a marked effect on the amount of residue obtained, the flaky or micaceous habit tending to hinder the easy sinking of the grains. Together with other factors, such as the general fineness of the sediments, the residues of specimens A and H (text figure 2) obtained by the funnel method appear to be affected by the abundance of biotite and chlorite; these two residues are particularly rich in these minerals and grain counts indicated 45 per cent. and 31 per cent., respectively.

Additional factors causing imperfect separations of heavy mineral residues and their apparent non-adherence to Stoke's law are elaborated by Krumbein and Pettijohn (1938, p. 331).

Apart from the considerable discrepancy that exists between the weights of heavy mineral residues obtained by separation in the ordinary funnel on the one hand and by the centrifuge on the other, a further important point arises. Grain-counts (300 grains per slide) have been made of preparations of the heavy residues obtained by the two methods, and it seems clear that, if the ratio R is high, then grain-counts of both residues are likely to differ considerably, i.e., the relative proportions of heavy minerals in the two residues will not be comparable. If the sediment is a rather coarse one, then, as shown previously, the residues obtained by the two methods are nearly of equal weight, and in addition the relative proportions of the different constituents are very nearly alike in the two residues. Three examples are illustrated in text figures 3, 4, and 5. In these figures the percentages of the different minerals in the residues obtained by the two methods have been plotted and the differences that exist in the residues will be clearly seen. Sediments A and F are both fine-grained types with a high percentage of material finer than very-fine-sand grade; here the heavy opaque ores predominate in the separating funnel residues, although these constituents do not exceed 10 per cent. of the total centrifuged residues. In both examples, A and F, the relative proportions of the various heavy constituents obtained by the two methods are quite distinct.

[Footnote] *R = Centrifuge residue/Separating funnel residue

– 80 –
Picture icon

Fig. 3–Comparison of the heavy mineral frequencies in sediment A, obtained, by the separating funnel (A), and the centrifuge (A′).

– 81 –
Picture icon

Fig. 4–Comparison of the heavy mineral frequencies in sediment F, obtained by the separating funnel (F), and the centrifuge (F′).

– 82 –

However, nearly true percentages of only high density minerals such as garent and zircon appear to be attained by the separating funnel method. In text figure 5 the relative proportions of the heavy minerals obtained from a coarser sediment by the two methods are compared and little variation will be observed.

Picture icon

Fig. 5-Comparison of the heavy mineral frequencies in sediment M, determined by the separating funnel (M) and the centrifuge (M′) methods.

– 83 –

To sum up then it is desirable to use the centrifuge rather than the separating funnel for practically all heavy mineral separations, and certainly essential when the investigation of sediments containing more than 20 per cent. of material finer than very-fine-sand grade is to be undertaken.

Specimens Used in This Work.
A = P. 8553 B = P. 8552 C = P. 8557
D = P. 8559 E = P. 8555 F = P. 8554
G = P. 8561 H = P. 8550 I = P. 8551
J = P. 8558 K = P. 8560 L = P. 8556
M = P. 8532

These numbers refer to specimens in the Petrological Museum of the Geological Survey.

Literature Cited.

Berg, E., 1937. A Method for the Mineralogical Fractionation of Sediments by Means of Heavy Liquids and the Centrifuge. Jour. Sed. Pet., vol. 7, no. 2, pp. 51–54.

Bertholf Jr., W. E., 1940. A New Centriifuge Tube for Heavy Mineral Separation. Jour. Sed. Petr., vol. 10, no. 2, p. 94. And for diagram see Jour. Sed. Petr., vol. 10, no. 3, p. 136.

Brown, I. C., 1929. A Method for the Separation of Heavy Minerals of Fine Soil. Jour. Palaeon., vol. 3, no. 4, pp. 412–414.

Cogen, W. M., 1935. Some Suggestions for Heavy Mineral Investigations of Sediments. Jour. Sed. Petr., vol. 5, no. 1, pp. 3–8.

Dryden, L., 1932. Heavy Minerals of the Coastal Plain of Maryland. Amer. Mineral., vol. 17, pp. 518–521.

Krumbein, W. C., and Pettijohn, F. J., 1938. Manual of Sedimentary Petrology, D. Appleton–Century Co., New York and London.

Kunitz, W., 1931. Eine Schnellmethode der Gravimetrischen Phassenanalyse mittels der Zentrifuge. Centralbl. f. Min., Abt. A., pp. 225–232.

Moshev, A., 1935. Separation of Minerals by Means of Heavy Liquids on the Principle of Uninterrupted Centrifugation. Mem. Soc. Russe Min., Ser. 2, vol. 64, Pt. I, pp. 118–129.

Rubey, W. W., 1933. The Size Distribution of Heavy Minerals within a Waterlaid Sandstone. Jour. Sed. Petr., vol. 3, no. 1, pp. 3–29.

Russell, R. D., 1936. The Size Distribution of Minerals in Mississippi River Sands. Jour. Sed. Petr., vol. 6, no. 3, pp. 125–142.

Schroeder, F., 1930. Schneidetrichter zum Elnsetzen in Zentrifuge beim Trennen von Mineralgemischen mit schweren Flussigkeiten. Centralbl. f. Min, Abt. A, pp. 38–46.

Taylor, G. L., 1933. A Centrifuge Tube for Heavy Mineral Separation. Jour. Sed. Petr., vol. 3, no. 1, pp. 45–46.