THE AMERICAN MINERAI,OGIST, VOL. 41, JULY AUGUST, 1959

THE I\,IAGNETIC SEPARATION OF SOI'IE ALLUVIALI,IINERALS IN I'IALAYA*

B. H. FnNrant, Minerals Eramination Diaision, Geological SurtteyD epartm ent, F' ederotion of M al,ay a.

Assrnlcr

This paper presents the results of a series of magnetic separations which have been in-

vestigated {or a number of minerals occurring in X{alayan alluvial concentrates. The pur-pose of the investigations was to establish, by the isolation of individual mineral species, areproducible and reliable method for the identification and quantitative estimation of

minerals in alluvial concentrates examined by the Geological Survey in Malaya In par-

ticular was sought the isolation of columbite from ubiquitous ilmenite. All the separations

were made on the small, highly sensitive Frantz Isodynamic Model L-1 laboratory separa-

tor, The minerals which have been successfully separated include ailanite, anatase, andalu-

site (and chiastolite), arsenopyrite, brookite, cassiterite, columbite, epidote, gahnite, garnet

(pink), ilmenite, manganese oxide (51.6/e Mn), monazite, pyrite, rutile, scheelite, siderite,

staurolite, thorite, topaz, tourmaline, uranoan monazite, wolframite, xenotime, and zircon.

PnocBpunp

When using an inclined feed, the Frantz Isodynamic separator (seeFigs. 1(A) & (B)) has three inherent variables. These are the field strengLh(current used), the side slope, and the forward slope.

( A ) ( B )

Irc. 1. Diagrammatic representation of side slope and forward slope.

The field strength is increased by means of a rheostat which raisesthe current from zero in stages of 0.05 amps. to 1.4 amps. Early in theinvestigations it was decided that steps of 0.1 amp. would be sufficientlygradual.

The side slope can be varied continuously from 0 to 90o and can alsobe given a reverse slope, as much as 90", for the diamagnetic separationof minerals such as zircon. Results of a few separations made by my

+ Published by permission of the Director, Geological Survey, with the authority of the

Minister of Natural Resources. Government copyright is reserved.

5 ;6s $ l opewdrd 511)Pe

738

MAGNETIC SEPARATION OF ALLUVIAL MINERALS 739

coileagues when the separator was firsl installed showed that, for thepurpose of these investigations, side slopes varying by less than 5o wouldproduce no significant differentials, and further, that side slopes greaterthan 15o showed more promise, as regards the separation of columbitefrom ilmenite, than did slopes of less than 15o. Accordingly a series oftests was carried out at side slopes of 15o, 20",25", and 30o.

The forward slope and a fourth, independent, variable, the rate of feed,control the rate of f low of the sample. The steeper the slope, the fasteris the flow. The rate of f low of a sample is not crit ical provided that it isslow enough to allow the mineral to remain long enough in the magneticfield for the magnetic pull exerted by the separator to be fully effective.A forward siope of 15o was chosen for, and maintained throughout, theinvestigation.

Much more crit ical is the uniformity of grain size of the sample, asthe larger grains tend to "suppress" the finer material. For the purposeof simple identif ication and the estimation of mineral percentages in analluvial concentrate a grain size of 0.4 mm. or 0.0166 inch (36 meshB.S.S.) is considered a satisfactory upper l imit. With larger grains diff i-culty is experienced in distinguishing amongst the opaque minerals andbetween the opaque minerals and the translucent minerals such asgahnite, cassiterite and ruti le. The grain size of 0.2 mm. or 0.0083 inch(72 mesh B.S.S.) was selected as a suitable lower l imit. This gave uni-formity to the sample and provided finer material for a parallel series oftests to be conducted, thereby ascertaining whether a difference in grainsize gives rise to any significant difference in results.

For each mineral the same sample was used for all the separations,thus eliminating variations independent of the change in side slope.Each sample, averaging 2 to 3 gms., was dried before being passedthrough the separator. After separation the fractions were examinedunder the microscope and the weight percentages recalculated to allowfor any impurit ies, which were estimated visually.

AssBsslmNr on RBsur,rs

On the basis of the separations carried out the minerals tested can bedivided into two groups, the "magnetic" and the "non-magnetic." The"non-magnetic" minerals are those that are not attracted at a value of1.4 amps. giving the maximum field strength of the separator. They areanatase, andalusite (and chiastolite), arsenopyrite, brookite, cassiterite,pyrite, rutile, scheelite, topaz, and zircon. The "magnetic" mineralsare given in Tables 1 to 8.

A study of these tables shows that, for the purpose for which theinvestigations were efiected, side slopes of 20' (Tables 3 and 4) and 30"

B. H. FLINTER

IN

. o ' @ o n N

f l ' o i r o N c o F

, . \ O D , H O O F O O

i F :d i - ; Dd f i . ;

N r H < r . < i N c O . O h

- o R 3 F $ H $ S F - S

f i F i r o" a @

\ i a q :* r O \ N O < 1N 6 i 4

- c Q 4

F O N

< 1 N . \ o O D. . !

j o r ' 8 8 :

<'

!

F

o

r

o

N

r'i

Fla

a

zm-

: i <z &

H <

> aei u)

v f'l

f f>t s N< NO I

z f r

A Z^ o

< o& A4 . 4

,n41 3 F

Z ot 4 4c) i&

FI

14

-;

Ll

€< 1 d r .

' k H

\ o F 4 F r F

6 X b O

, i < nv e A

MAGNETIC SEPARATION OF ALLUVIAL MINERALS

F

r; , -r

t r < rN

s

r

\o

r'i

F]a

(n

2a

H

v 4d H

a ?

z>E aF. rn

O p

O NF s2 l

Er nDZ d

(J

t4

Fl

N

i

F

O \ \ O i N

. j ; ; . r , r J c i- 4 @ - -

N € b €j - i * d ; * , ;- H 6 r

\ o N N 4 a * o c ' )* o ' l . - + - d ; . ; . o d ;a a r - € - O \ €

s o € r N < l c O A

d . - . d d * - J o . iH - c ) c | 1 < r

cod o

a c o N 4 a oN o o . | o oN 6 - O

r

O n rr . l J J6 - i

T ?€ i

.g3 Ng o d o

E ' a o " . ' . E F . ! o

:€;; !:! :5;5*E:5€3 + . \ 4 E ; i 6 * H 9 : F 6 E? c i r . r - c i o i > > i ; 6 F F 5 B x

6 X b o

, : < . Ee v a

B. II , FLINTER

- lI

z lI

@

r

\o

O

N

dU)r'l

a

N

z,f-

a -

Z t )1 4 eg <

A A? az a* .

F ] N

t+E ' ^H OZ l

7 Z

Irl a

< t4 A

( ; s. ^ 5

i r rF r ^Z i hl i d(,

a

rd

tsl

F

o

6 0 0 \ o N

H O < t O

q 4 0 4 0 \ < 1

- i ' o o .6c i 6b €

N < r r o a o @o - o . N o o o n- F < r + r 4

o\ \orO \ r

O N r i

4 $ N

< ) C O \ O O \ \ O O \ O \ i . O

g 'S" i :XS i . i i * ;

c a i * o r O , o

c d , L € o @ o o o .- F € r o o -

@ \ o N

J.- J <,- N a

? ? \ " ? a nO \ \ O N O O Ni o 4 q

r ^ E

F E.E, - ,< Fv v a

MAGNETIC SEPARATION OF ALLUVIAL MINERALS 743

: a \ a . o l no o 4 N . . 1 € O

@ i - H

r- l

Z iI

aIv l

i l

I

o

N O €

6 d ; i -

a N a€; < , ,d . t . i Ig t - - t -

< f O \ N N N

N @ * c ) c )@ N

o \ q

@ 4

O € O

@ € H

; p S $ - + J i

+ @ a

d o ; i

€

<i

GI

? €g 'X * " , , .EH. : "

€€i:! EE s+icE:S€J = ' = = = t r g o + d : < t r o i

ad r fdd i>>d ;F i5> i

E X b o

e - A

Ail

a

a

N

za

a . ,

z ;

c t 4t s F

a <r .E7 aF Ag F

P H

< N( J rE Iz -

A :

? a

4 . eA 3H B

a &a - OF] JT

j ; ozQ

ts

F

!

F

B. II. FLINTER7M

<'- l6 iz l

* F

r q o i ' . . :-

4 o \ o 1 N \ o

s a p i N o

" i ; $ = iss

. $ + N o \ s o < 1 0

. ;dS$ < .F ! f r $F J

+ + 1 O c | , N o d N

- - + . 1 , i s o € - € o6 ' O - . F + o - +

o \ oN D \ O O \ O* o * , n < ,H O N d

oI

N

d €q ' a e " , o . E F . : q

. . . = " , . , - o . : J - . . : - ? i E E. ! t : : E : 5 S . :E . : H i F i! t *E!e ; !+ i ! i==;=S ' r .53:>=E, ;F€5**

a

a

N

za

i <

z t )H FF <

3 Az ap -

1 4 Ng sl +Y . Aci i6Z l

H I

> z

F ] A

H A^ t o

a SL \ &

< f r2 %f 4 H(J

ts

FI

trF

@

r

d1 \or4 r d r i

o r r \ o €

d iN

t r o - *

, i < . Ee v a

* . i -

MAGNETIC SEPARATION OF ALLUVIAL MINERALS

N < r

o . i ; ;f N i

rla

a

N

z

H . 1

z ;C r-1F i t s

a <

frnAFq

a F l

< NL ) ra l

, l tPE H

i hz '(,

HH

Fl

F

N \ O N

ro c i 6, ;-

@ @

o o N

s o

o

@

o c o + 1 r N

o 6 s i d J

r c o \ o o @ 4 N d* j * * . - c i + n J os N \ o - + o o . \ o

6 : - F

O \ r ro . " , € €- o o

D T O

O N N

-

D O

F

o 4

ssF- ; .1-o

r

-N

o

+ a E

6 r * *, i < iv z a

B. I-1. FLINTER

@ € o \ N n

; o * { , JN

o o + i \ o o N

, i d * ; J c i G + j- + + N e F

q ioo o \. l . f n r / : c o o o- - H o - o D

q q a a 9 0 tc ) + 1 N O N 4 N - N N

D N L '

\ o + o o s < r o N 4 \ O

6 N N N b o \ o + . a O O \4 O N i * D - N 4

r c Fr

N

@

r

Fla

a

o

z

a -

z <Z t')H q& <

a 4? az aD -

F ] Nt s r< - r9 . 15 :2 ;

H I

> z

r d a< i lH A

c b s, ^ ft--1 C

i f rF ^Z v )F J iQ

Fr

r

j

trF

4 r

- N O- o

r N

4 r 4 b a l \ o

+ . i . i J d JN \ O O N O

N 4 N

-

o

6 X b o= t r 9 ?

, ' , 4 Ev v a

TIAGNETIC SEPARATION OF ALLUVIAL MINERALS

- N. N

d lF iz i

I

I1i l

l

. . b . 4 . .

F <, .oF

f

oF,l

a

a

a")

za

t ' lz ;

H t sa <H >zi,i u)a ad c aP H

t4a< No NE IEz

Il trj i D

z

O

Fi

@

I

F

N N

C N H N

o 6 . i J ;$ $ r

o o o 4 c ..lo oi o. -i€;-+r o\

o c a 4 vo , : s . L \ o \ o r9 - s - - - € N

O O \ O O O . N \ O O O* - i ; . j n m o q i o o .$ a r c o \ o +

< j r o i c ) i o i 4

6 i d - j , :o - j . i . ' : o* ;- o - - N - - o q

N O + S O \ O

; F : 8 f r i ; r f i o

r €O \ @

N \ O

N o . l -

-

€

r

? . €.e 5. .e* e iE. :g

:€€+E.E dEfrE! FE E:: = E F F o t r ; i 5 : = t r = :

i8;s358 ggEsfi € Sg *

p ^ . E

6 K B

r i < i jv r a

748 B. H, FLINTER

Frc. 2. A plot of data from Table 5. Side slope 25', forwardslope 15"; grain size -36+72 mesh B.S S.

(Tables 7 and 8) can be disregarded, because there is too much over-lapping of the mineral species, particularly columbite and ilmenite, andthat at a side slope of 30" (and therefore above) effective separationcannot be achieved because the magnetic attraction at the lower amper-ages is not strong enough to overcome the effect of gravity.

I lowever, side slopes of 15" (Tables 1 and 2) and 25'(Tables 5 and 6)give convenient breaks at amperages of 0.3, 0.5 and 0.8 and 0.4, 0.7 and1.0 respectively, and these separations are summarized in Fig. 2.

Aithough most of the minerals tested give rise to some variation dueto grain size, in nearly all cases these variations become negligible over asuitable range of amperages. For a side slope of 15" and amperages of0.3, 0.5 and 0.8, only four minerals (gahnite, staurolite, thorite andtourmaline) show a marked variation, and for a side-slope of 25o andamperages of 0.4,0.7 and 1.0, the only two minerals showing a markedvariation are gahnite and tourmaline. For this reason, and because ofthe better separation of coiumbite from ilmenite and monazite from xeno-time, a side slope of 25o is preferable to one of 15".

ExcrprroNs

Occasionally an anomalous result is encountered. The most commoncause is the presence of a partial coating of iron oxide, which can beremoved by treating the minerai with hydrochloric acid, when it willbehave normally. A less frequent cause is the presence of very fine in-

MAGNETIC SEPARATION OF ALLUVIAL MINERALS 749

clusions of a more magnetic mineral. Notable examples of this are zirconand monazite and, in columbite areas, cassiterite (due to inclusions ofthe columbite). Both these causes result in the mineral becoming moremagnetic.

It is known that, from certain localit ies in Malaya, cassiterite andilmenite wil l give anomalous results. Cassiterite from these localit ies isattracted at all fractions, from 1.4 amps. to 0.1 amps. and is also at-tracted by a horse-shoe magnet. This magnetic cassiterite is found inareas of major iron-ore development. No inclusions are discernible athigh power in transmitted l ight, but it is thought that this apparentlyinherent magnetism, which can be destroyed by heating to a tempera-ture of a few hundred degrees Centigrade, is due to micro-fine exsolutionlamellae of magnetite.

Ilmenite, on the other hand, becomes less magnetic. This is due toits alteration to "arizonrte." The greater the degree of alteration, theless magnetic is the ilmenite. The maximum current so far required toattract Malayan material (at a side slope of 25") is 0.7 amps.

Rutile is often quoted as being magnetic. Magnetic ruti le is commonlyencountered in Malaya. Exhaustive tests have shown that the magnetismis due to the ionic substitution of Ta/Nb (and therefore Fe). TheseTa/Nb-ruti les range, in Malaya, from non-magnetic materiai to materialattracted at a current strength of 0.7 amps. depending on the amountof substitution present. The increase in magnetism is accompanied byan increase in opacity. Pure rutile is non-magnetic.

TnB Esrrlr.qTroN ol M.qcNrrrc BBn.cvron rnolr KwowN RESULTS

The results given in Tables 1 to 8 represent the mass magnetic sus-ceptibil i t ies of the minerals tested. The mass magnetic susceptibil i ty(Kv) of a mineral is given as (1):

magnetic susceptibility

density

An approximate value for Ka can be obtained from the equation

a* : \Y:\ 1o-6 c.g.s.

whereo:the angle of slide-slope

and1:the current in amperes, provided that Kn is much less than 1 and

that determinations are made at amperages lower than 1.3, assaturation sets in above this value.

An important application of this rule is that, once the Ka value for

750 B. H. FLINTER

Tesrc 9. Mass Mecvorrc Suscnprrsnrrv (K;a) rN 10-6 C.G.S.

Side Slope

25"15 '

0 . 10 . 20 . 30 . 40 . 50 . 60 . 70 . 80 . 91 . 01 . 11 . 2l . J

1 . 4

5r7 .64129.415 7 . 5 23 2 . 3 52 0 . 7 r1 4 . 3 81 0 . 5 68 . 0 96 . 3 9. ) . 1 6

4 . 2 83 .59J . t r o

2 . 6 4

684.041 7 1 . 0 176.004 2 . 7 527 .3619.00t 3 . 9 61 0 . 6 98 . 4 46 .84J . O . )

4 . 7 54 . 0 53 . + 9

845.2+2 r t . 3 l93.9252 .8333 .8123.4817 .2513.21t0.448 .456 .995 .875 .004 . 3 1

1000.00250.001 1 1 . 1 16 2 . 5 040.002 7 . 7 820.411 5 . 6 312.3510.008 . 2 66.945 . 9 25 . 1 0

any mineral is known, the current at which it is attracted can be calcu-lated for any other side slope or because of any compositional variationafiecting the specific gravity. In order that the information contained inthis paper can be so used, the Ku values for the various amperages,at side slopes of 15",20",25o and 30o, are given in Table 9.

The application of this relationship to the results given in the tablesshows that the most significant factor influencing the magnetic suscepti-bility of a mineral is the percentage of ferrous iron (Fe2+) present in itscomposition. Thus i lmenite and siderite, both with a high percentage ofFe2+ (and a fairly low density) have similar high values of Ky. The pinkgarnet, with its high mass magnetic susceptibil i ty, is a highly ferroanvariety (almandite). The increased amperage needed to attract i lmenitewhen it is altering to "arizonite" can be foreseen because of the de-creased value for K,44 resulting from a loss in Fe2+ content whereas thespecific gravity remains similar.

The columbite/tantalite mineral series is very interesting. Althoughno tantalite has been separated it would presumably be attracted athigher amperages than columbite because of its much higher specificgravity, and it should be possible to obtain, from the current used, thetantalum/niobium ratio of the mineral. Research in hand on the Ta/Nb-ruti le minerais indicates that a relationship exists between the ferrousiron content, the ratio of tantalum to niobium and the specific gravity.With the Ta/Nb-ruti le minerals, however, a variable amount of substi-

MAGNETIC SEPARATION OF ALLUVIAL MINERALS 751

tution is possible, with correspondingly variable amounts of ferrous iron,which give rise to an added complication which must be taken into ac-count.

Cor{cLusroN

The effective separation of minerals according to their mass magneticsusceptibility depends principally upon a fine balance between gravity(controlled by variation in the side slope), and field strength (controlledby the current used). Other factors such as forward slope, rate of feed,and grain size, are subsidiary provided they fall within certain fairlybroad limits.

For the routine examination and estimation of the approximate per-centages of minerals in Malayan concentrates it has been shown thatelectromagnetic separation is effective and that the best separation isachieved by using a side slope of 25o and current strength of 0.4,0.7and 1.0 amps. on material less than 36 mesh B.S.S.

In order of decreasing mass magnetic susceptibility the minerals sepa-rated are siderite, garnet (pink), i lmenite, wolframite, columbite,manganese oxide (51.6/6 Mn), allanite, xenotime, epidote, staurolite,tourmaline, gahnite, uranoan monazite, monazite and thorite, withanatase, andalusite (and chiastolite), arsenopyrite, brookite, cassiterite,pyrite, ruti le, scheelite, topaz, and zircon being non-magnetic.

RBrcnrlqcr

(1) H. H. Hnss, 1956 "Notes on operation olFrantz Isodynamic Magnetic Separator."Pamphlet published by S. G. Frantz Co., Inc., p. 8.

Manuscript receiled. October 15, 1958.