Jan Svoboda. Magnetic Techniques for the Treatment of Materials

Подождите немного. Документ загружается.

320 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS

5.1 Selection of magnetic separation technique

There are numerous types of magnetic separators, each one being suitable for

only a limited range of applications. The choice is inﬂuenced by a wide variety

of factors and the successful operation of a magnetic separator depends, to a

great degree, on practical experience. The results of separation cannot generally

be accurately predicted.

As with most other physical techniques of material treatment, magnetic

separation cannot be successfully performed simply on the basis of theoretical

models and experience is still the foundation. At present, there is no branch of

magnetic separation where theoretical treatment of the subject could be con-

sidered entirely adequate.

Each magnetic separation technique has its merits as well as limitations; no

single technique is applicable to all applications of magnetic separation. The

evaluation of magnetic separation as a means of beneﬁciation, or generally of

recovering solids, liquids or gases, is often based solely on the data collected

from a few brief tests, not always carried out in a suitable magnetic separator.

The likelihood of disappointing results and even of rejection of the technique

is high, in spite of adaptability and large scope of the magnetic methods of

materials manipulation.

The choice of magnetic equipment, and the selection of appropriate operat-

ing conditions dier widely between these processes. The mode of operation,

namely wet or dry, and the throughput requirements impose additional com-

plicating restrictions on the selection of the most suitable magnetic technique.

Since particle size of the material to be treated is the most important variable

that determines the selection of the magnetic separation technique, the general

guidelines of the process selection are based on particle size distribution, as

shown in Fig. 5.1.

If the size of particles to be manipulated by magnetic separation is greater

than, for example, 75 m, we have an option of a dry or wet process. If, how-

ever, the particles are ﬁner than 75 m, the wet technique is more appropriate,

although this threshold particle size can be smaller or greater than 75 m, de-

pending upon the amounts of the ﬁnes present in the feed and on the physical

properties of the feed.

The subsequent choice of the magnetic separation technique depends on the

magnetic properties of the material. If a strongly magnetic component is present

in the feed, low-intensity drum magnetic separators can be used, either in wet

or dry modes of operation.

Coarse, weakly magnetic materials can be treated by wet roll magnetic sep-

arators, rare-earth drum separators and possibly by wet high-intensity high-

gradient magnetic separators. In the latter technique, there is an upper limit

imposed on the particle size by the type of the matrix and particles greater than

1 mm usually cause clogging, even of a coarse matrix.

In a dry mode, coarse weakly magnetic particles can be recovered by a variety

of high-intensity magnetic separators, including induced magnetic roll, cross-

5.2. DRY MAGNETIC SEPARATION 321

Particle

size

> 75 µm

< 75 µm

Wet

Dry

Wet

Strongly

magnetic

Weakly

magnetic

Strongly

magnetic

Weakly

magnetic

Strongly

magnetic

Weakly

magnetic

LIMS, Chute

Wet roll MS, WHIMS,

HGMS, REDs

LIMS, Suspended,

Pulle

Grate

Roll MS

Roll, REDS, Cross-belt,

IMR, OGMS, HGMS

LIMS

REDS, WHIMS,

HGMS

Figure 5.1: Classiﬁcation of magnetic separation processes (adapted from Svo-

boda [S1]).

belt separator, permanent magnet roll separator, rare-earth drum separator

and open-gradient magnetic separator.

If the particles are ﬁner than 75 m, the best approach is to use wet magnetic

separation. Fine, strongly magnetic particles are usually recovered by wet low-

intensity drum separators, while for weakly magnetic materials the wet high-

intensity or high-gradient magnetic separation is appropriate.

The decision making process for selection of the magnetic separation tech-

nique is schematically shown in Fig. 5.2.

5.2 Dry magnetic separation

Dry magnetic separation is a technique dating back to the beginning of the last

century. It has been practiced successfully in the removal of tramp iron from

process streams, in beneﬁciation of strongly magnetic minerals, as well as feebly

magnetic minerals such wolframite, cassiterite or ilmenite. Puriﬁcation of nu-

merous industrial minerals, such as andalusite, glass sand, ﬂuorspar, feldspar,

diamonds and others, by removing ﬁne strongly magnetic impurities, is very of-

ten carried out in a dry mode. More recently developed eddy-current separators

also operate in a dry mode.

For dry magnetic separation to be successful, the material must be com-

322 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS

Magnetic

separation

Particle

< 75 um

Wet MS

Strongly

magnetic?

Is dry

process

required?

Dry MS

LIMS

WHIMS

HGMS

Wet MS

Strongly

magnetic?

LIMS

Plate

Chute

Strongly

magnetic?

IMR

Roll

REDS

Cross-belt

Disc

OGMS

HGMS

LIMS

Pulley

Suspended

Grate

Plate

Roll

WHIMS

HGMS

REDS

Wet roll

Yes

Figure 5.2: Selection of magnetic separation techniques (adapted from Svoboda

[S1]).

5.2. DRY MAGNETIC SEPARATION 323

pletely dry, which sometimes involves an expensive heating stage. For particles

smaller than a certain threshold size (for instance 75 m), the attractive elec-

trostatic and van der Waals forces will cause sticking and the formation of

aggregates. Selectivity of separation can thus be severely aected.

Dry processes often require the material to be sized into several rather narrow

size fractions, each of which must be spread in a monolayer over the separating

surface. It is a well-established fact that when magnetic separation is performed

on materials sized into, for example, three or more fractions, the e!ciency of

separation is usually much improved in comparison with unsized samples.

The loading of the separating surface, such as belt or drum, inﬂuences the

throughput and the e!ciency of the separation, and a compromise must be

found between these contradictory requirements. As a rule of thumb, when the

particles are introduced onto a belt in a monolayer, a safe belt loading (portion

of the belt covered with feed particles) should be approximately 25%, even when

the feed contains a high percentage of the magnetic material.

As a rule, a larger proportion of dry ﬁnes necessitates higher rotating speeds

of a roll or drum. Although increase in speed will improve the grade of the

magnetic product and the throughput of the separator, it will also increase the

losses of the magnetic component. The recovery can be improved by re-treating

the non-magnetic fraction in a second stage. On the other hand, when the

valuable product is the non-magnetic fraction, high speed of rotation will result

in high recovery and poor selectivity of separation. The grade of the concentrate

can be improved by cleaning the non-magnetic fraction in the second pass.

5.2.1 Electrostatic phenomena in dry separation

During dry material handling operations, the dielectric particles invariably be-

come electrostatically charged. The dominant charging mechanisms are tribo-

electriﬁcation (or frictional charging) and contact charging. These mechanisms

can happen during any dynamic contact of surfaces, such as sliding, rolling and

vibration of surfaces at contact. When two dissimilar dielectric materials make

and then break contact, charge transfer from one of the materials to the other

takes place. This is not to say that conductors will not charge by contact elec-

triﬁcation - they usually charge - but since they are conductors, they often lose

their charge quickly.

While physical mechanisms of contact charging are well understood, tribo-

electriﬁcation, the oldest form of charging, is poorly understood and explana-

tions are of phenomenological nature. One of the controversial rules of thumb

often used for predicting the sign of the surface charge resulting from contact

electriﬁcation is Coehn’s rule, which states that "when two dielectric materials

are contacted and separated, the material with the higher dielectric constant

becomes positively charged" [L12]. Coehn’s rule was quantitatively formulated



=15×10

6

(

 

) (5.1)

where 

is the surface charge density (in [F/m

]) and 

and 

are dielectric

324 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS

Table 5.1: Dielectric constants of selected materials.

Material Dielectric constant 

Water 80.4

TiO

Barium titanate 1200

Carbon 5.7

Ilmenite 33.7 to 81

Chromite 11

Rutile 89 to 173

Zircon 12.6

Glass 4to6

Quartz 4.3

Hematite 25

Limestone 6.1

Fluorspar 7.4

Bakelite 4.9

Perspex 3.4

constants (relative permittivities).

Although Coehn’s rule is reported to have been veriﬁed for more then 400

substances [L12], it is of limited value to the metallurgist because of the dif-

ﬁculty in determining the relative permittivities of the surface layers of two

materials that are to make contact. Typical values of dielectric constant of

selected materials are shown in Table 5.1.

Sometimes several materials can be arranged into the so-called triboelectrical

series such that each material is charged positively when it is rubbed against

the previous one. Examples of such series are:

• zinc-glass-cotton-ﬁlter-paper-silk-zinc [B26] and

• teﬂon-metals-resins-wood-mica-quartz-glass [H19].

The triboelectric charging behaviour of materials is generally sensitive to

an atmospheric environment. At high values of relative humidity, considerable

negative charging takes place. At low values of relative humidity, although ﬁne

particles still charge negatively, the coarse particles show a change of charging

polarity such that the net charge conveyed by the powder material is positive

[B27].

When a particle undergoes charging during handling, the charge transferred

at the particle contact point will re-distribute itself over the particle surface by

electrostatic forces. The rate at which re-distribution proceeds depends on the

relaxation time 

(in [s]) given by



= 



( (5.2)

5.2. DRY MAGNETIC SEPARATION 325

Figure 5.3: The eect of moisture on the grade of the magnetic concentrate

obtained from - 2 mm magnetite and martite ores treated by cross-

belt and induced magnetic roll separators (adapted from [D4]).

where ( is the resistivity of the material (in [m]) and 

its relative permittivity

(dielectric constant) and 

=8.85×10

12

F/m is permittivity of vacuum.

The time dependence of the surface charge density, after completion of the

charge transfer, is given by:



= 

exp(w@

) (5.3)

where 

is the surface charge density at w =0.

As a general rule, if electrostatic phenomena arise in a material handling

situation, they are likely to become increasingly signiﬁcant as particle size de-

creases. Particles charged triboelectrically, either positively or negatively, may

adhere strongly to surfaces. Moisture contributes to the strength of adhesion

and Fig. 5.3 illustrates the eect of moisture on selectivity of dry magnetic

separation. It can be seen that the grade of the magnetic concentrate is seri-

ously impaired by the presence of moisture and the permissible moisture content

decreases with decreasing particle size. For instance, if an ore with size distribu-

tion - 20 mm can tolerate 4% of moisture, then the maximum moisture content

in a ﬁne ore - 2 mm should not exceed 1% [D4].

The eect of air humidity

The moisture concentration in an ore is also inﬂuenced by the relative humidity

of the air. This is a consequence of the capillary attraction between particles,

which results in adsorption and subsequent capillary condensation of the water

vapours in the particle contact area. If the force of capillary attraction I

326 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS

Figure 5.4: The eect of air humidity on dry magnetic separation of magnetite-

quartz mixtures, for dierent values of particle radii (adapted from

[V10]).

between a non-magnetic particle and a magnetic particle is greater than its

weight J, the non-magnetic particle will report in the magnetic fraction. The

force I

is the sum of of the surface tension and the Laplace pressure force:

=8u

 +4u

(



)

where u

(l = 1 and 2) are the radii of the curvatures of the meniscus between

the contacting particles 1 and 2, and  is the surface tension (in N/m) of the

ﬂuid.

Experimental investigation [V10] of the eect of the air humidity on dry

magnetic separation of magnetite-quartz mixtures, as a function of particle size,

conﬁrmed that the e!ciency of separation varies linearly with the ratio J/I

It can be seen from the results illustrated in Fig. 5.4 that with increasing hu-

midity, the separation e!ciency decreases steeply. The eect of the air humidity

becomes particularly pronounced with decreasing particle size.

5.2.2 Density and magnetic tracers

The use of the Tromp curve, or partition curve, is a well established practice

for the evaluation and representation of gravity concentration processes. The

partition curve can be conveniently constructed with the aid of density tracers

[N15]. The density tracers are manufactured as cubes in sizes ranging from 0.8

5.2. DRY MAGNETIC SEPARATION 327

to 20 mm. They are colour-coded according to the magnitude of their density

and are commercially available, for example from Bateman Engineering (Pty.)

Ltd. (South Africa), in the density range from 1300 kg/m

to 4000 kg/m

with density interval of 50 kg/m

. Partition Enterprises (Pty.) Ltd. (Australia)

supply density tracers that are strongly magnetic for easy recovery, by magnetic

means, from density separation products.

The partition number, or the percentage recovery U into the concentrate,

for each size and density fraction, is usually determined as

U = 100

+ q

where q

and q

are the numbers of tracers in the concentrate and in the tailings,

respectively.

Magnetic tracers, which have become commercially available only recently,

can be used, in a fashion similar to the way density tracers are used in grav-

ity concentration, to determine and optimize the separation e!ciency of mag-

netic separators. A magnetic tracer is a simulant designed to have the same

size and magnetic susceptibility as the material which is being recovered or

removed in a magnetic separator. Magnetic tracers of mass magnetic suscepti-

bility ranging from 2.5×10

7

/kg to 125.5×10

7

/kg (20×10

6

/g to

1000×10

6

/g), in steps of 2.5×10

7

/kg, for low-susceptibility tracers,

and 12.5×10

7

/kg for high susceptibility tracers, are available from Bate-

man Engineering (Pty.) Ltd., South Africa [D12]. These tracers, whose size

ranges from 1 mm to 20 mm, are colour- and shape-coded, according to the

magnitude of their magnetic susceptibility for easy evaluation of the separation

process. Tracers smaller than 1 mm are impractical as they cannot be easily

identiﬁed among the mineral particles. The magnetic partition curve can be

constructed using the partition numbers, or percentage recovery into magnetic

or non-magnetic fractions. In terms of the recovery of the tracers into the mag-

netic fraction, for example, the partition number, for a given size and magnetic

susceptibility, is expressed as:

U = 100

pdj

+ q

qrqpdj

where q

pdj

and q

qrqpdj

are the numbers of tracers in the magnetic and non-

magnetic fractions, respectively. The probable error of separation is given by

eq. (4.58).

While magnetic tracers are ideally suited for quantiﬁcation of the e!ciency of

a magnetic separation process, their practical use is restricted to the separation

of materials coarser than 1 mm. Although they can be used to evaluate the

e!ciency of separation of even ﬁner materials in dry separators such as roll

and drum, the particle-size dependence of the magnetic susceptibility cut point

can result in considerable inaccuracy of the audit. In wet separators, such as

low-intensity drum separators or WHIMS or HGMS, even the smallest 1 mm

magnetic tracers would interfere with the separation process

328 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS

Figure 5.5: Comparative metallurgical e!ciencies of various dry magnetic sep-

arators as applied to apatite recovery (adapted from Roux et al.

[R19]).

5.2.3 Comparison of dry magnetic separation techniques

The minerals industry has always had a great need to beneﬁciate coarse, weakly

magnetic minerals and a wide spectrum of dry magnetic separators has been

developed over the years. The range of various types of separators, based on

dierent principles of the generation of the magnetic ﬁeld, makes a correct choice

of the most suitable technique a di!cult task. Comparisons of the separation

e!ciencies of dierent techniques are usually not published and if they are,

caution should be exercised when taking these comparisons at face value. Every

comparison is only conditional since the relative merit of each technique is a

function of the mineral to be treated and its physical properties. Sometimes

comparisons are oered by manufacturers of magnetic separators who might be

biased towards their own product.

Comparison of dry separators for beneﬁciation of the pyroxenite ore

One of the rare comprehensive comparative production-scale exercises was con-

ducted by Roux et al. [R19]. In order to beneﬁciate pyroxenite ore to yield ap-

atite concentrate, several dry magnetic separators were tested at Foskor (South

Africa). Figure 5.5 summarizes the metallurgical e!ciencies of dierent types

of separators, while Fig. 5.6 shows the relationship between phosphate recovery

and particle size for IMR, HGMS and OGMS machines. A thorough practi-

cal experience with these magnetic separators was gained at Foskor and their

5.2. DRY MAGNETIC SEPARATION 329

Figure 5.6: Phosphate recovery as a function of particle size for various magnetic

separators (adapted from Roux et al. [R19]).

comparison is summarized in Table 5.2.

A gradual deterioration in the results was observed with IMR, caused by

abrasive wear on the epoxy ﬁller between the steel discs of the rolls. Further-

more, because the separation distance between the magnetics and non-magnetics

is very small, a precision setting of a splitter was required. As the mineralogy

of the ore varies, constant attention to the optimization of roll speed and split-

ter settings was required, which made it di!cult to maintain good metallurgy

[R19].

A Sala HGMS 120-15 operated reliably and well for 1000 hours and it was

found that single-stage separation yielded metallurgical results equivalent to

that of two stages of IMR separation. The expanded metal matrix showed no

measurable wear, while the metallurgy was shown to be adversely aected by

oversize material, which gradually accumulated in the matrix. In order to keep

coarse particles out of the feed, and to avoid the onerous task of cleaning the

matrix chambers, it became necessary to ﬁt a second screen which could be kept

under continuous visual observation.

An extensive series of experiments with SmCo permanent roll separators

showed that the feed rate, roll speed and thickness of the magnet and steel

rings must be optimized for each separation stage.

The open-gradient, split-coil superconducting magnet separator was tested

in two modes, shown in Chapter 2 in Fig. 2.85(a) and (b). The free-fall system

gave acceptable metallurgical results only at very low feed rates. A much more

satisfactory ore presentation system was found to be the one in which the ore

slid down a ramp, as shown in Fig. 2.85(b). The ramp imparts an outward