Jan Svoboda. Magnetic Techniques for the Treatment of Materials

Подождите немного. Документ загружается.

410 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS

Figure 5.87: Oxidizing roasting of ilmenite: magnetic susceptibility of ilmenite

as a function of roasting time and temperature (adapted from

[C21]).

used routinely in the magnetic separation of mineral sands. For instance, il-

menite concentrates from heavy mineral deposits in northern Kwazulu-Natal

(South Africa) and Mozambique often contain as much as 0.3% Cr

,which

would contaminate the ﬁnal high-titanium slag. While the magnetic suscepti-

bility of iron-poor chrome-rich spinel is similar to that ilmenite, it is not signif-

icantly aected by oxidizing roasting [N11]. However, six-fold increase in the

magnetic susceptibility of ilmenite, by oxidizing roasting, enabled the removal

of chromium-bearing spinel from ilmenite concentrates by magnetic separation

[N11]. This process is practiced at Richards Bay Minerals (South Africa) where

the ilmenite concentrate from WHIMS is roasted in a ﬂuidized bed at a tem-

perature of 800

C for 40 minutes. The mass magnetic susceptibility increases

from 2.4×10

6

/kgto15×10

6

/kg.

A detailed laboratory investigation of the roasting behaviour of ilmenite

(St. Urbain, Quebec, Canada) was conducted by Ciu et al. [C21]. The test

results indicated that oxidizing roasting of ilmenite (" =2.18×10

6

/kg)

could increase its magnetic susceptibility within limited temperature and time

ranges. As is evident in Fig. 5.87, at 500

C, even extended heating (up to

2 hours) did not change the magnetic susceptibility of ilmenite. At higher

temperatures between 650

C and 700

C, the susceptibility started to increase

after about 30 minutes of roasting. After reaching its maximum at " = 5.5×10

5

/kg, the susceptibility dropped sharply to 4×10

7

/kg, lower than that of

the original ilmenite.

However, reduction roasting of ilmenite exhibited a considerably more sig-

5.3. WET MAGNETIC SEPARATION 411

Figure 5.88: Magnetic susceptibility of ilmenite as a function of temperature, for

oxidizing and reduction roasting. Roast time: 30 min. (adapted

from [C21]).

niﬁcant increase in magnetic susceptibility, approaching that of magnetite (" =

2×10

3

/kg), as is illustrated in Fig. 5.88.

Allen [A31, A32] conducted a thorough experimental study of the eect of

oxidizing roasting on the magnetic properties of ilmenite. He observed that

the ilmenite particles behaved as though they contained discrete ferromagnetic

components within a paramagnetic matrix. Magnetic changes at the roast tem-

perature up to 450

C were ascribed to small changes in size and shape of

small magnetic cells. Although magnetic susceptibility was found to increase

in the range around 500

C, the dierences were not su!ciently convincing.

At temperatures above 550

C, the growth of the magnetic cells and magnetic

susceptibility appeared to be time and temperature-dependent.

The de-coppering of galena and molybdenite concentrate by the removal of

chalcopyrite can be signiﬁcantly improved by roasting the chalcopyrite. It was

reported by Makarow et al. [M30] that by heating chalcopyrite in an oxidizing

atmosphere to a temperature between 250

C and 500

C for 10 minutes, the

magnetic susceptibility of the product increased by three orders of magnitude.

5.3.10 The selectivity of separation

It was shown in Section 1.2 that a separation process will be selective only when

the magnitudes of the magnetic and competing forces are of comparable mag-

nitude, compatible with conditions expressed by eqs. (1.1). The relative values

of the magnetic and competing forces are thus critically important in determin-

412 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS

ing the selectivity of magnetic separation. Because some of these forces have

dierent dependence on particle size, their relative importance will vary with

particle size, as has been outlined in Section 1.2.2. It is also important to recall,

as shown in Section 1.2.2, that particle size is a more important discriminating

factor in magnetic separation than magnetic susceptibility and that a magnetic

separator is also a particle size classiﬁer.

Some investigators have tried to determine the threshold ratio of magnetic

susceptibilities of the components of a material mixture, for which their sep-

aration will be e!cient. Such a threshold ratio is generally of limited value

because of the signiﬁcant eect of particle size distribution on the balance of

forces. Classiﬁcation or desliming can alleviate the negative eects of wide size

distribution. In addition, magnetic susceptibilities of numerous materials de-

pend on the magnetic ﬁeld strength, to which they are exposed, as discussed in

Section 1.4.7.

Of signiﬁcant importance but particular di!culty is the selective separation

of ﬁne materials. Large amounts of ﬁnes and ultraﬁnes are generated during the

mining and milling of large tonnages of low-quality ores. A direct consequence

of this situation is an increasing need for e!cient technologies to beneﬁciate

the ﬁnes. It has been shown in Section 3.6 and in [S1] and [S31] that the

complexity of selective beneﬁciation of ﬁnely dispersed materials rests in the fact

that volumetric forces present in a separator, viz. magnetic, gravitational and

erosion forces, have to compete with surface forces. The question of whether a

quasi-colloidal particle will be captured by a matrix element covered by particles

previously deposited is thus determined, inter alia, by the surface characteristics

of the particles.

In practice, therefore, ﬁne particles to be separated should be colloidally

unstable, i.e. they should not repel each other and thus prevent contact. On

the other hand, particles of the gangue minerals should not coagulate with those

of the valuable mineral lest they become entrained in the magnetic concentrate.

The eectofpHoftheslurry

An experimental study of the eect of pH on magnetic separation of weakly

magnetic iron ores [S38, S72] has shown that in order to achieve high recoveries

of iron and a high grade of the magnetic concentrate, the pH of the slurry must

be adjusted in such a way as to correspond to the point of zero charge (PZC) of

the oxide mineral. This behaviour was ascribed to a combination of ionic and

magnetic ﬂocculation in the primary and secondary minima, and to a higher

retention probability of particles on the matrix [S72, S1].

Similar observations were made in the magnetic separation of uranium and

gold from cyanidation residues [S73]. As shown in Fig. 5.89, the grade of the

magnetic concentrate, as a function of pH, reaches its maximum at a pH of

1.2, and decreases continuously to the alkaline region. It can be seen that pH

adjustment from the usual operating value of about pH 9 can result in a gold

content increase by 25%.

Control of surface interactions can also result in selective coating of mineral

5.3. WET MAGNETIC SEPARATION 413

Figure 5.89: The grade of the magnetic concentrate from gold and uranium

cyanidation residues, as a function of pH (adapted from Svoboda

et al. [S73]).

and metallic particles with magnetite, followed by magnetic separation. It can

be used as a method of particle separation, as demonstrated, for example, by

Parsonage [P13].

Energeticandkineticcriteriaforselectiveseparation

In order to correlate quantitatively the rules for selective separation with the

results of experimental investigations, energetic and kinetic criteria, based on

DLVO theory, for selective magnetic separation of a useful mineral from a system

of containing several mineral species, were derived by Svoboda et al. [S76]. It

was shown that the onset of destabilization of a suspension and, consequently,

the critical condition for selective magnetic separation, is determined from the

energy condition Y

=0,whereY

is the total energy of interaction of two

particles and is given by eq. (3.123). At the same time, the residence time

w of particles in the magnetic ﬁeld must be such that wAw

1@2

,wherew

1@2

the half time of coagulation, given by eq. (3.126). This condition imposes a

restriction on stability ratio Z , given by eq. (3.124), for a given concentration

of particles in a slurry. The stability ratio for a suspension containing l and m

mineral species can be re-written as:

exp(

)

+ e

+ k)

gk (5.19)

where e

and e

are the radii of particles of species l and m, respectively, and k

is the distance between the surface of the interacting particles.

414 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS

Figure 5.90: Total potential energy of interacting uraninite and quartz particles

(2e =5m), as a function of pH, at E

= 1 T (adapted from [S76]).

If a particle of a useful mineral is denoted by 1, and the gangue particle as

2, high recoveries will be achieved if Y

? 0andZ

? 1. Simultaneously, for

the grade of the magnetic concentrate to be high, one must ensure that valuable

and gangue particles do not coagulate mutually, and thus Y

A 0andZ

1. Furthermore, to avoid the mechanical entrainment of large clusters of gangue

particles, it would be preferable to work under the condition Y

A 0andZ

A 1. These conditions for selective magnetic separation of ﬁne particles can be

summarized as [S76]:

High recovery: Y

? 0; Z

? 1 (5.20)

High selectivity: Y

A 0; Z

AA 1andY

A 0; Z

A 1 (5.21)

The validity of these conditions was conﬁrmed experimentally by magnetic

separation of uranium-gold cyanidation residues [S1, S73, S76]. The curves of

total potential energy of interaction of uraninite and quartz particles are shown

in Fig. 5.90. It can be seen that selective separation can be expected when Y

A 0, which occurs at pH 1.2. With further increase of pH, Y

decreases and

the selectivity is reduced. This trend is conﬁrmed by the experimental results

shown in Fig. 5.89. Poor selectivity of separation in the alkaline region is a

consequence of heterocoagulation into a deep secondary minimum [S76].

Carrier magnetic separation

E!cient recovery of ﬁne particles can be aided by promoting mutual interaction

between ﬁne and coarse particles. In carrier (or piggy-back) magnetic sepa-

ration, coarse particles serve as carriers for ﬁne particles. Carrier magnetic

5.3. WET MAGNETIC SEPARATION 415

Table 5.29: Magnetic separation of mixtures of dierent size fractions of

uranium-gold leach residues [S69].

Mixture Head Theoretical Experimental

5:1 [m] U

[ppm]

Grade

[ppm]

Recovery

[%]

Grade

[ppm]

Recovery

[%]

-17+(-53+25) 185 820 30 1033 52

-17+(-75+53) 173 700 29 850 40

-17+(-150+106) 159 616 27 600 22

separation makes use of the fact that the frequency of particle collisions is much

greater when there are particles of dierent sizes, which is the result of dierent

trajectories in the accelerating ﬂuid. The rate of adhesion of ﬁne particles (e 

2 m) to coarse ones (e  30 m) is 10

to 10

times higher than the rate of

adhesion between ﬁne particles [S77].

The results of the investigation of the possible eect of carrier magnetic

separation on the e!ciency of separation are summarized in Table 5.29 [S69].

It can be seen that the presence of fractions - 53 + 25 mand-75+53m

considerably enhanced the recovery of uranium from the ﬁne (- 17 m) fraction,

and the grade of the magnetic fraction.

The presence of the coarse fraction (- 150 + 106 m) neither improved the

selectivity of the process nor increased the recovery of uranium from the ﬁne

fraction. Moreover, no dependence of recovery or grade on pH was observed, in

contrast to the mixtures of ﬁner fractions, where pH played an important role.

Since the value of pH did not aect the e!ciency of separation of individual size

fractions, it is obvious that pH inﬂuenced mutual interaction of ﬁne and coarser

particles.

A detailed investigation of carrier magnetic separation was carried out by

Yanmin Wang et al. [W25] and the results conﬁrmed the trends observed by

Svoboda [S69, S1]. A puriﬁed chromite concentrate and chromite ore slime were

used in that investigation. A very ﬁne fraction (- 10 m) was mixed with coarse

fractions, - 53 + 38 m, - 75 + 53 m and - 106 + 75 m. It was found that the

ﬁne particles coagulated onto the coarse ones and the cluster thus formed was

recovered in the matrix of a magnetic separator. The fraction - 53 + 38 mwas

found to be the most eective in improving the process e!ciency, as shown in

Fig. 5.91. The process of carrier separation was also found to be dependent on

pH and the highest e!ciency was achieved at a pH near the point-of-zero-charge

of the valuable mineral.

The matrix vibration and slurry pulsation

The vibration of a matrix, either by mechanical or electromagnetic means, was

proposed by several investigators, to assist in the ﬂushing of the matrix, to

propel the ore through the matrix in dry high-gradient magnetic separation and

416 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS

Figure 5.91: The eect of pH and carrier size on the mass recovery of chromite (-

10 m) by magnetic separation (adapted from Wang et al. [W25]).

Figure 5.92: Vibration-assisted magnetic separation of wolframite-quartz mix-

ture as a function of the vibration current (adapted from [S78]).

5.3. WET MAGNETIC SEPARATION 417

to improve the e!ciency of separation [K5, L14]. Mechanical vibration can be

provided, for instance, by an eccentric rotating shaft driven by a variable speed

motor. Longitudinal electromagnetic vibration is provided by a coil mounted

on the chamber and energized with alternating current. A vibratory force is

exerted on the coil by the interaction of the d.c. ﬁeld generated by the magnet

of the separator, and of the a.c. ﬁeld in the coil.

The vibration of the matrix reduces the mechanical entrainment of gangue

particles in the magnetic product and facilitates their desorption from the matrix

by ﬂuid ﬂow. For instance, matrix vibration was found to improve the selectivity

of wet high-gradient magnetic separation of wolframite-cassiterite slimes and of

wolframite-quartz mixtures, as is shown in Fig. 5.92 [S78], and of Cu-Pb and

Cu-Mo ﬂotation concentrates [Y3]. The application of matrix vibration also

resulted in an increased recovery of kaolin in wet [L15] and dry [Y4] high-

gradient magnetic separation.

Pulsation of a slurry proved to be a useful approach to improving selectivity

of separation and reducing matrix blockage in HGMS. When applied to a low-

intensity drum magnetic separator, the slurry pulsation resulted in an improved

quality of the magnetic concentrate [K22]. Slurry pulsation, developed by Liu

Shuyi et al. [Y2] is being successfully used, on production scale, in SLON high-

gradient magnetic separators [X1].

5.3.11 Comparative tests with wet high-intensity magnetic

separators

As discussed in Chapter 2, the development and commercialization of high-

gradient magnetic separation as a viable technique for mineral beneﬁciation

resulted in the design and manufacture of a wide range of wet high-intensity

magnetic separators.

Most wet high-intensity magnetic separators dier in those technical and op-

erational characteristics that can signiﬁcantly aect their performance and cost-

eectiveness. Any comparison of these separators should, therefore, take into

consideration, in addition to the feed properties and the product requirements,

the size of the operation and the capability of the machine to be scaled-up, if

necessary.

Magnetic separation equipment, and particularly WHIMS and HGMS ma-

chines, cannot be readily miniaturized for bench-scale testing. The bench-scale

testing is, therefore, often focused on characterization of the feed and of the

separation products, as a function of those variables that can be controlled in

pilot-scale testing. Without proper understanding of these characteristics it may

be impossible, except by luck, to end up with a successful process.

On the other hand, pilot-scale tests are usually carried out using commercial

equipment, at full feed rates and over an extended time scale. Reliable technical

and economic assessment of the operation can then be made.

Comparative pilot-scale tests of several wet matrix high-intensity magnetic

separators were described by Forssberg and Kostkevicius [F22]. Five models,

418 CHAPTER 5. PRACTICAL ASPECTS OF MAGNETIC METHODS

Figure 5.93: Recovery as a function of particle size for assorted wet matrix high-

intensity magnetic separators (adapted from [F22]).

representing three distinct principles of magnetic circuit design, with three dif-

ferent types of matrices were tested. A mixture of hematite and quartzite, with

97% of the particles being smaller than 300 m and 21% smaller than 15 m,

was used as a test material. The mean magnetic susceptibility of the mixture

was approximately 2.5×10

6

/kg.

The technical data of the separators used in the tests are summarized in Ta-

ble 5.30. The eect of particle size on iron recovery is given in Fig. 5.93. For ﬁne

particles smaller than 50 m, the Sala HGMS was most e!cient, followed by the

Jones and Reading machines. The e!ciency on coarser material was comparable

among these separators, although the Jones separator showed a fall-o in the

recovery in coarse fractions associated with the middlings ﬂushing. The energy

consumption was lowest in the Jones separator (0.5 kWh/t), followed by the

Reading and Boxmag-Rapid machines, although the value of this information is

limited as the rating is likely to change for large-scale machines. The optimum

metallurgical results obtained with these separators are shown in Table 5.31.

Although these results indicate the potential these various machines have,

their direct translation into large-scale applications must be done with caution.

Dierent separators have dierent potential for scale-up and the metallurgical

performance, consumption of energy and water, and resistance to matrix clog-

ging, will vary with the size of the machine. The assessment of the technological

and economic potential of each machine must be based on a detailed analysis

of their performance with a wide range of ores of dierent magnetic suscepti-

bilities and size distributions, taking into account the anticipated magnitude of

the operation.

5.3. WET MAGNETIC SEPARATION 419

Table 5.30: Speciﬁcations of magnetic separators used in tests [F22].

Parameter Jones

P71

Boxmag

SHW1

Reading

16 pole

Sala

HGMS

MkII

Matrix Grooved

plates

Triangular

bars

Salient

plates

Expanded

metal

Rotor dia. [mm] 710 800 1800 1720

Capacity [t/h] 1-2.5 1.5 - 6 15 - 30 1-3

No. of separation

zones

2 1 8 1

Matrix width [mm] 75 25 66 120

Matrix length [mm] 220 150 205 140

Mass [t] 7.1 3.8 7.1 6

Table 5.31: Representative metallurgical results obtained with dierent wet ma-

trix high-intensity magnetic separators [F22].

Separator Grade [% Fe] Recovery [% Fe]

Sala HGMS Mk II 66.0 94.0

Jones P71 66.0 92.0

Reading WHIMS-16 65.0 91.5

Boxmag-Rapid SHWI 64.5 85.5