Minerals Engineering 23 (2010) 350358

Contents lists available at ScienceDirect

Minerals Engineering
journal homepage: www.elsevier.com/locate/mineng

Application of dielectrophoresis for the separation of minerals

G.R. Ballantyne *, P.N. Holtham
The University of Queensland, Julius Kruttschnitt Mineral Research Centre, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Comminution is an energy intensive process, accounting for approximately 29% of total mining energy
Received 29 July 2009 (Tromans, 2008). Better liberation at coarser size ranges (0.52 mm) has been identied by the minerals
Accepted 3 September 2009 industry as an opportunity for energy conservation. In order to capitalise on future developments in com-
Available online 9 October 2009
minution, an effective coarse particle separation method is required for liberated particles which are too
coarse for otation. This paper reviews the application of dielectrophoresis as a potential technique for
Keywords: the removal of coarse liberated gangue, thereby signicantly reducing the volume of ore continuing onto
Sulphide ores
ne grinding. Dielectrophoresis is the translational motion of neutral matter caused by polarisation
Electrostatic separation
Sorting methods
effects in a non-uniform electric eld.
Mineral processing A simple apparatus has been constructed, in which the deection of a particle in a non-uniform elec-
trical eld is used to quantitatively measure the dielectrophoretic force on quartz and galena particles of
various sizes. The results show that the force increases as a function of voltage squared and particle
radius cubed. This technique can potentially be used to measure the effective permittivity of single min-
eral particles and quantify the effect of particle composition on separation.
2009 Elsevier Ltd. All rights reserved.

1. Introduction thereby reducing energy in ne grinding. Dielectrophoresis is a

separation technique that uses non-uniform electrical elds to sep-
The current energy environment has resulted in efforts to de- arate particles of differing permittivity. The permittivity (or dielec-
crease the amount of energy that is consumed by comminution. tric constant) of a material is a bulk, volumetric property that
A study by the National Materials Advisory Board of America determines its polarizability in an electric eld. Since the surface
(1981) estimated that comminution constitutes approximately area to volume ratio decreases as the particle gets larger, dielectro-
1215% of Australian and 34% of worldwide energy usage. A more phoresis should be more efcient for larger particles than the sur-
recent study revised this gure to approximately 1.8% of Australias face forces used in processes such as otation and electrostatic
national energy (Tromans, 2008). Tromans study used data from separation. Applying a stress by a force eld is also one of the most
the Department of Energy, which estimated that beneciation uses efcient separation techniques (Andres, 1980). In some cases, per-
39% of the total energy used in mining and minerals processing, of mittivity may be less useful as a separation property than density,
which 75% is accounted for by comminution DOE. Therefore, com- magnetic susceptibility, wettability, etc.; however, it adds another
minution accounts for approximately 29% of the total mining en- useful dimension to the art of separation (Pohl, 1978).
ergy in the USA (Tromans, 2008). Thales of Miletus was the rst to record the effects of dielectro-
Size reduction from 1 mm to 75 lm uses about 75% of the total phoresis, in about 600BC (Pohl, 1978). His experiment involved the
comminution energy input. Therefore, separating out and remov- rubbing of amber to charge it and attract insulated (dust and uff)
ing 50% of waste at any size above 1 mm offers a 37% energy saving particles. Hateld (1924a) was the rst to apply dielectrophoresis
(Powell, 2008). Since conventional otation utilises surface proper- to minerals processing. The basic principle involved using a liquid
ties, coarse particle (above about 0.5 mm) otation of metal sulp- with a permittivity between the two components in order to push
hides is not currently possible, therefore an alternative coarse the low permittivity elements towards the weaker region of an
separation process is required that utilises a bulk, volumetric electrical eld, while at the same time attracting the high permit-
property. tivity elements towards the stronger region of the eld. Although
It is possible that dielectrophoresis can efciently remove many changes have since been made in the design of dielectric sep-
coarse (between 0.5 mm and 2 mm) liberated gangue particles arators, the fundamental principles of dielectrophoresis have re-
mained constant since Hatelds early experiments.
The design of the dielectric separator has gone through many
* Corresponding author.
phases from precipitation-type, to isomotive, to screen-type and
E-mail address: [email protected] (G.R. Ballantyne). most recently the travelling-wave separators. Contemporary

0892-6875/$ - see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mineng.2009.09.001
 G.R. Ballantyne, P.N. Holtham / Minerals Engineering 23 (2010) 350358 351

dielectric separation may be enhanced by the adjustment of the through build up of space charges in the volume or of surface
frequency, temperature or the electrical conductivity of the liquid charges at the interfaces (Von Hippel, 1954).
dielectric (Andres, 1996). However, the major issue standing in There are many different equations that can be used to describe
the way of dielectrophoresis is the absence of a suitable medium how the permittivity of a solid mixture changes with the volume
with a variable permittivity, low conductivity and a high break- fraction and permittivity of each phase (see Table 3). The Lichte-
down threshold which is also safe and environmentally friendly. necker, Looyenga (LandauLifchitz) and Brown equations can all
be used to model binary mixtures of minerals in a particle. Sharif
(1995) used the Looyenga equation to predict the permittivity of
2. Theory dust with various compositions. Zheng et al. (2005) used the Lich-
teneckers mixture formula to calculate the permittivity of mineral
2.1. Permittivity powders in a polythene matrix. Briggs (1992) modelled the permit-
tivity of mixtures of mineral sands using the Looyenga and Brown
Before discussing the theory of dielectrophoresis, an under- models and found that the Looyenga equation tted the data bet-
standing of the underlying physical property, permittivity, is re- ter. Fig. 2 shows a comparison between the Lichtenecker, Looyenga
quired. The permittivity of a substance is an electrical property and Brown mixing equations applied to the composition of galena
that determines its polarizability in an electrical eld. It can also (permittivity 190) in a quartz (permittivity 4) matrix. The Looy-
be referred to as the dielectric constant, dielectric permeability, enga and Lichtenecker mixing models show a gradual increase in
electric susceptibility or electrical polarizability, but for consis- permittivity as the galena ratio increases. This gradual increase
tency, permittivity will be used throughout this paper. Dielectro- means that a large volume of high permittivity material is required
phoresis uses differences in permittivity in order to separate to change mixture permittivity signicantly. The Looyenga model
particles. has a relationship that is closest to linearity. The Brown model is
The main factors that determine the permittivity of a material able to t mixtures with smaller permittivity differences, however,
are the density and ease of charge displacement which in turn de- for galena and quartz the mixture permittivity is negative for com-
pend on the atomic weight, coordination number, etc. There is, positions of less than 20% galena. Further research is required to
therefore, a relationship between mineralogy and permittivity investigate which of these models best describes the permittivity
(see Fig. 1). Sulphide minerals in general have much higher permit- of semi-liberated mineral particles.
tivities (5230) than all other mineral groups due to the combina- In order to understand the effect of composition on dielectroph-
tion of the sulphur anion with a highly polarisable cation oretic force, the ClausiusMossotti function (Eq. (3)) was evaluated
(Parkhomenko, 1967). Oxides have intermediate permittivities, as a function of medium permittivity (Fig. 3). It can be seen theo-
with a few exceptions such as rutile. All the other mineral group retically that particles of different galena to quartz ratio can be
tend to have permittivity in the low to very low range, averaging separated from each other simply by altering the medium permit-
between 6 and 9. Textural factors (i.e. matrix vs. mixture) are very tivity. It also shows that as the medium permittivity is increased,
important in determining both the permittivity and resistivity of the dielectrophoretic force increases until a maximum is reached,
rock samples (Parkhomenko, 1967). at which point it decreases, intercepting the x-axis when the med-
Permittivity is difcult to measure, especially for sulphides, be- ium permittivity equals the particle permittivity. Experimental
cause it depends not only on the chemical makeup, but also the fre- work is currently investigating whether these theoretical results
quency and temperature at which the measurement is made, as can be reproduced in practice.
well as the crystallographic direction. Low frequency permittivity
measurements compiled from the literature are given in Tables 1 2.2. Dielectrophoresis
and 2.
When two miscible liquids are combined, the permittivity of Dielectrophoresis can be dened as the translational motion of
the mixture is a simple weighted average of the two liquids per- neutral matter caused by polarisation effects in a non-uniform
mittivities. Solid mixtures, however, are more complex and show electric eld. When a neutral particle is polarised by a non-uniform
a non-linear relationship. This is due to interfacial and space- electric eld, the two charges on the body are equal, but the elds
charge polarisation which causes large-scale eld distortions operating on the two regions are unequal creating a force impelling

Permittivity as a function of mineral group

250
Permittivity measurements

200

150

100

50

-
carbonate element halide hydroxide oxide phosphate silicate sulfate sulfide

Fig. 1. Box plot of permittivity as a function of mineral group (Berg, 1936; Church et al., 1988; Cumbane et al., 2008; Goossens and Van Biesen, 1989; Keller, 1981;
Parkhomenko, 1967; Salsman, 1991; Tarjan, 1986; Tickell, 1931; Young and Frederikse, 1973; Jordan and Sullivan, 1985).
352 G.R. Ballantyne, P.N. Holtham / Minerals Engineering 23 (2010) 350358

Table 1
Compilation of mineral permittivity measurements from literature.

Mineral Group Berg Tickell Parkhomenko Young and Jordan and Church Salsman Cumbane
(1936) (1931) (1967) Frederikse Sullivan (1985) et al. (1991) et al. (2008)
(1973) (1988)
Apatite Phosphate 3.54 10.5 4.2 8.859.97
Arsenopyrite (FeAsS) Sulde >81 5075
Barite Sulfate 3.54 12.2 6.27.9 9.149.69 58
Calcite Carbonate 33.5 7.8 7.58.7 8.09.15 8.3 8.669.16
Chalcocite Sulde >81 210275 818
Chalcopyrite Sulde >81 30100 35
Diamond Element 5.5 5.5 0.3
Dolomite Carbonate 3.54 8 6.3-8.2 7.227.62
Fluorite Halide 3.54 6.8 6.2-8.5 6.768.7 6.416.59
Galena Sulde >81 190 100225 68
Gypsum Sulfate 45 11.6 7.96.3
Halite Halide 3.54 6.2 5.66.4 5.455.92 4.815.06
Hematite Oxide 25 12.0 (a), 4.5 (c)
Llmenite Oxide >33.7 <81
Magnetite Oxide >33.7 <81 20
Orthoclase (feldspar) Silicate 3.54 4.56.2 4.344.56
Pyrite Sulde >33.7 <81 720
Quartz Silicate 33.5 5 4.414.6 4.3 3.854.00
Rutile Oxide 170 89173 115
Sphalerite Sulde 56 17 7.8 8.08.7 720
Zircon Oxide 12 3.65.2 12.5 11.8

Table 2
Compilation of liquid permittivities from literature and safety rating (Berg, 1936, and Chemwatch).

Liquids Permittivity Flammability Toxicity Body contact Reactivity Chronic

1propanol 20.44 3 2 3 0 2
Acetic acid 6.4 0 0 0 0 0
Amyl alcohol 15.8 2 2 2 0 2
Benzene 2.3 3 2 2 1 4
Ercnoform 4.5 0 3 2 0 2
Carbon tetrachloride 2.2 0 3 2 1 2
Chlorobenzene 5.62 2 2 2 2 0
Chloroform 5.2 0 3 2 1 3
Di-n-butyl phthalate 6.1 1 2 1 0 3
Dioxane 2.2 3 2 2 2 2
Ethyl alcohol 25.7 3 2 2 1 2
Furfural 42.0 3 4 2 2 2
Kerosene 2.0 2 2 2 1 2
Methyl alcohol 33.7 3 3 2 1 2
Nitrobenzene 36.0 1 3 1 2 2
Water 80.0 0 0 0 0 0

Table 3
Compilation of mixing equations from literature.

Name Function Application Reference

P
Lichtenecker  log emix n v i log ei Mineral assemblages Von Hippel (1954); Zheng et al. (2005)
Looyenga or Landau Pn Mineral assemblages Sharif (1995); Briggs (1992); Goossens
emix i1 v i e1=3
1=3
i
Lifchitz et al. (1992)
Brown  0 2  2
 0 3 Mineral assemblages Briggs (1992)
Power series: ee0 1  13 pq de0  13 p2 q 19 pq2 pck de0   
  2 
e.g. emix pe1 qe2 1  13 pq pee11 e2
qe2

Pn
i1 v i ei orientated parallel (k = +1) or perpendicular
Lichtenecker and ekmix k Laminated mineral grains Raju (2003); Parkhomenko (1967)
Rother (k = 1) to the applied eld in rock

the neutral body towards the region of stronger eld (Pohl, 1978). diate permittivity (see Fig. 5). Positive dielectrophoresis occurs
Fig. 4 shows the difference between dielectrophoresis of neutral when the polarizability (permittivity) of the particle is greater than
matter and electrophoresis of charged particles. Fig. 4 also shows the medium. The asymmetric eld forces accelerate the particle
that forces found in dielectrophoresis are independent of the direc- more than the medium, concentrating it in the highest eld. On
tion of the electric eld, and only requires a highly divergent eld the other hand, negative dielectrophoresis occurs when the parti-
(Pohl, 1950). It is therefore possible to use an AC eld with cle permittivity is less than the medium. The low permittivity par-
dielectrophoresis. ticles are pushed toward the region of weakest eld strength in a
Dielectrophoresis produces both attractive (positive) and repul- way that is analogous to the buoyancy of low density materials
sive (negative) forces when a liquid medium is used with interme- in higher density liquids.
 G.R. Ballantyne, P.N. Holtham / Minerals Engineering 23 (2010) 350358 353

Comparison of mixing equation

200

Permittivity of mixture
150

100

50 Brown

Looyenga
-
Lichtenecker

-50
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percentage liberation of galena in quartz

Fig. 2. Comparison of mixing equations for galena/quartz binary mixtures.

Effect of galena composition on Clausius-Mossotti function using Lichtenecker mixing

equation 100 % galena 0%
quartz
1.E-10 90 % galena 10%
quartz
Clausius-Mossotti function

8.E-11 80 % galena 20%

quartz
70 % galena 30%
6.E-11 quartz
60 % galena 40%
quartz
4.E-11 50 % galena 50%
quartz
40 % galena 60%
2.E-11 quartz
30 % galena 70%
quartz
0.E+00
20 % galena 80%
0 1 2 3 4 5 6 7 8 9 10 11 12 quartz
-2.E-11 10 % galena 90%
quartz
0 % galena 100%
-4.E-11 quartz
Medium permittivity

Fig. 3. Effect of galena composition on ClausiusMossotti function using Lichtenecker mixing equation.

Fig. 4. Neutral and charged particles in uniform and non-uniform elds and the effect of electrode polarity on dielectrophoresis (after Pohl, 1978).

Pohl (1958) derived an expression for the translational force (F) where a is the radius of spherical particle, eo is the permittivity of
on a dipole in a non-uniform eld (E): free space (8.854  1012 C/V m) and Km and Kp are the relative per-
  mittivities of the medium and particles, respectively.
~ Kp  Km
F e 4pa3   E  rE eo K m 1 Pohls dielectrophoretic force equation can be simplied to the
K p 2K m
  product of the spherical volume, electric eld strength, electric
Kp  Km eld gradient and degree of polarisation (Jordan and Sullivan,
or ~
F e 2pa3  eo K m  rE2 2
K p 2K m 1985). Unlike otation, the dielectrophoretic force is proportional
354 G.R. Ballantyne, P.N. Holtham / Minerals Engineering 23 (2010) 350358

tude of the force. Water and air are the most appropriate mediums
from an economic and safety point of view, however, water has a
permittivity that is too high (see Table 2) and dissolved ions create
a conductivity problem. Air has a permittivity that is too low and
spark-over is an issue. In past experiments, the permittivity of
the liquid was altered by changing the conductivity or by changing
the ratio of low permittivity and high permittivity liquids. It may
be possible to use a slurry of low permittivity ne particles in
water, however, no tests have been conducted as yet (Andres,
1979). Separating the particles in a medium which has a permittiv-
ity higher or lower than all the constituent particles is another pos-
sible solution to the problem (Aldaeus et al., 2006; Andres, 1982).

3. Pertinent past work

Fig. 5. Comparison between isomotive and cylindrical electrode congurations.
The design of dielectric separators has evolved from precipita-
tion-type, to isomotive, to screen-type and now travelling-wave.
to the volume of the particle, not the surface area; it is therefore The rst batch, precipitation-type designs were patented by Hat-
more suited to coarse mineral separation when the surface area eld (1924b) and were also the only separators ever tested on a
to volume ratio is smaller. Calculating the electric eld gradient plant-scale. Pohl (1950) designed the rst dielectric separator
is made extremely difcult, if not impossible, except for simple which used an isomotive electrode design. Both (Verschure &
eld distributions, because of the eld distortions created by the Ijlsts, 1966) dielectric sieve and Jordans (1980) rotating-drum
particles themselves (Abdel-Salam, 1993). Increasing the opera- dielectric separation used parallel plate electrodes and dielectric
tional space and displacement of low permittivity particles de- chaining to improve efciencies. More recently, travelling-wave
mands much higher potentials and electrical elds (Andres, 1980). separators have been used for the separation of colloidal particles
The degree of polarisation is the most important factor in the and in nanouidics (Hughes et al., 1996; Wang et al., 1997).
force equation and is given by the ClausiusMossotti relation: Although the design of the dielectric separators is an important re-
search question for future applications, this paper will concentrate
Kp  Km ep  em on the measurement of the dielectrophoretic force.
eo K m or em 3
K p 2K m ep 2em In order to validate Eq. (1), quantitative measurements of the
The ClausiusMossotti relation is important because it determines actual dielectrophoretic force on a real particle are required. These
whether the force will be positive or negative as well as the magni- force calculations can also be used to back-calculate the permittiv-
tude of the force. From the equation it can be seen that the force in- ity of a solid of some arbitrary shape (Pohl and Pethig, 1976). Hawk
creases with the medium permittivity and the difference between (1966) investigated the effect of medium conductivity in dielectro-
the particle and medium permittivity. Fig. 6 shows the theoretical phoresis with the aim of separating biological cells. The relative
dielectrophoretic and gravitational forces on a 1 mm galena forces involved were calculated by observing the deection per
(e = 190) particle as a function of the medium permittivity from unit voltage for particles of lead hafnate (PbHfO3), silicon dioxide,
Eq. (1). The dielectrophoretic force is greater than the gravitational lead, silicon and rutile particles in a non-uniform eld. Small
force with medium permittivities above about 5 and a maximum amounts of tap water were added to decrease the resistivity of
force is reached at a medium permittivity of approximately 70. the low conductivity water. The resistivity of the water was varied
Fig. 7 shows how the force varies with the particle permittivity with and the crossover between the attractive force to repulsive force
water as the medium. was determined both theoretically and practically by using the
The major difculty in the application of dielectrophoresis is the equation:
absence of a suitable medium with a variable permittivity and high jK 01  i=p1 xeo j jK 02  i=p2 xeo j 4
breakdown threshold (Plaksin and Olonsky, 1966; Andres, 1979).
The choice of medium is a very important one in dielectrophoresis Feeley (1969) investigated the effect of the applied eld frequency
because it not only determines the cut-point but also the magni- as well as the medium permittivity. He altered the resistivity of

Medium permittivity variation

1.4E-03
1.2E-03
1.0E-03
Force (N)

8.0E-04
6.0E-04 Dielectrophoretic force
4.0E-04 Gravitational force
2.0E-04
0.0E+00
0 15 30 45 60 75 90
Medium permittivity

Fig. 6. Force variation with medium permittivity.

 G.R. Ballantyne, P.N. Holtham / Minerals Engineering 23 (2010) 350358 355

the medium with the addition of the salt tetra-iso-pentyl-ammo- Connected to AC voltage
nium nitrate, and varied the permittivity with different ratios of
dioxane and water. Six materials were studied: silicon, silicon diox-
ide, tin, stannic (tin) oxide, rutile and lead hafnate. The addition of
salt made most of the particles act as though they were in a liquid of
higher permittivity, however, lead hafnate and rutile acted errati-
cally and the force trend for silicon was opposite of what was ex-
pected theoretically. The response of the silicon and lead hafnate
particles deviated from theoretical expectations when the medium
permittivity was altered, a result attributed to ionic double layer
interactions for the silicon particles.
Chen investigated the electrical double layer assumption used Plate electrode Pin electrode
by Hawk and Feeley to explain their observed anomalies (Chen,
1969). His results showed that introducing cations of differing
valencies did affect dielectrophoresis and therefore, the anomalies
may be able to be explained by electrical double layer assump-
Teflon supports
tions. The effects were more striking with lead hafnate and quartz,
but only at high frequencies (Chen, 1969). The size of the crystal,
smoothness of the surface and attachment of the glass bre were
all major sources of error. Particle glued to
glass fibre

4. Experimental work and results

A simple initial experiment was carried out, which was similar

Fig. 8. Single particle test cell schematic including glass beaker with optically at
to that conducted by Hateld (1924a), using two stainless steel
viewing rectangle.
needles to create an electrical eld in distilled water. The needles
were held 1 mm apart and a 50 Hz electrical eld of 124 kV/m
was created between the needles. One millimetre galena particles that pyrite has a much higher permittivity than the reported liter-
demonstrated positive dielectrophoresis, being attracted to the ature values.
highest point in the eld, while quartz was repelled due to nega- In order to quantify the dielectrophoretic force acting on a sin-
tive dielectrophoresis. Positive dielectrophoresis was also observed gle particle, a test tube separator was built, based on apparatus
for pyrite, chalcopyrite and galena, and negative dielectrophoresis used in the work by Pohl, Hawk, Chen and Feeley (see Fig. 7).
for quartz, plagioclase and sphalerite. Magnetite and hematite The deection of an approximately 1 mm particle glued to the
showed inconclusive results most likely due to magnetite becom- end of a ne glass bre hanging in a non-uniform electrical eld
ing magnetically attracted to the electrodes even when the electri- was measured. The eld was created by a pin and a plate electrode
cal eld was removed, and the hematite being in the form of long separated by 3 mm (see Fig. 8). Ethanol was used as the dielectric
thin bres. This simple experiment demonstrated that dielectro- medium as it is safe and stable and has a permittivity of 25.7 (see
phoresis can be used to separate sulphide minerals from siliceous Table 2). The dielectrophoretic force was found by increasing the
gangue on the basis of permittivity. With the exception of pyrite, voltage in steps from 0 to 300 and measuring the subsequent
the results agreed with what would be expected based on permit- deection of the particle using a handheld digital microscope with
tivity values from the literature. Theoretically, pyrite should have 200 magnication. The measurement of deection was calibrated
shown negative dielectrophoresis because both of the permittivity to a force by tilting the apparatus and measuring the deection per
values given by literature were less than that of water (see Table degree of tilt. This had the added effect of removing the inuence
1). However, the literature values for permittivity are not compre- of the elasticity or stiffness of the glass bre from the measure-
hensive and vary quite substantially, so this might be an indication ments. A variable frequency, alternating electrical eld was created

Particle permittivity variation

1.5E-03
1.0E-03
5.0E-04
Force (N)

0.0E+00
-5.0E-04
-1.0E-03 Dielectrophoretic force
-1.5E-03 Gravitational force
-2.0E-03
-2.5E-03
0 50 100 150 200
Particle permittivity

Fig. 7. Force variation with particle permittivity.

356 G.R. Ballantyne, P.N. Holtham / Minerals Engineering 23 (2010) 350358

using a signal generator connected to an audio amplier. The low The coefcient from the line t in Fig. 11 can be used to deter-
voltage output of the audio amplier was transformed up to a mine the ClausiusMossotti function from Eq. (2) and therefore the
maximum output of 2000 V. particle permittivity. In order to calculate this more accurately a
Figs. 9 and 10 show the forces recorded for ve sizes of galena more precise and robust measurement technique is required for
and ve sizes of quartz in ethanol as a function of voltage. Each test the particle size and particle position in the electrical eld. The
was repeated three times, trend lines were then tted to the three particle radius used in this analysis was the average of the height
sets of data by minimizing the sum of squared errors at each point. and width measurements.
For each particle size, the force was seen to be proportional to the In order to verify the measurements taken by the single particle
voltage squared, a result which aligns with the theoretical force test cell, the theoretical force was calculated, however, problems
equation (Eq. (2)). The force also increased with particle size. were encountered with modelling the electrical eld. If the eld
The theoretical force equation (Eq. (2)) suggests that the in- created by the pin-plate electrode conguration is assumed to be
crease in the force should be proportional to the particle size cubed similar to a cylindrical, coaxial electrode conguration then the fol-
(volume). Fig. 11 uses the tting coefcients from the various sizes lowing equation for the eld gradient can be used, in conjunction
in Figs. 9 and 10 in order to verify the relationship between force with Eq. (2), to calculate the force (Pohl, 1960, 1978):
and particle size. Although, the results from galena t this relation-
2r o V 21
ship well, the results from quartz are more varied. This may be due rjEcyl j2 2
5
to quartz having an irregular particle shape created from crushing, r 3 lnr 1 =r 2 
while galena has a very regular cubic crystalline structure.

Positive Dielectrophoresis Galena in Ethanol

1.4E-06
-11 2
Fg5 = 5.81x10 V
Particle Radius
1.2E-06
1.17mm
Dielectrophoretic force (N)

1.0E-06 0.90mm
0.65mm
-11 2
0.52mm Fg4 = 2.32x10 V
8.0E-07
0.47mm
6.0E-07

4.0E-07 -11 2
Fg3 = 7.84x10 V
-12 2
Fg2 = 7.12x10 V
2.0E-07
-12 2
Fg1 = 5.84x10 V
0.0E+00
0 20 40 60 80 100 120 140 160 180 200
Applied voltage (V)

Fig. 9. Positive dielectrophoresis: galena in ethanol.

Negative Dielectrophoresis Quartz in Ethanol

Applied voltage (V)
0 50 100 150 200 250 300 350 400

-1.0E-07

-3.0E-07
Dielectrophoretic force (N)

-12 2
Fg1 = -2.74 x10 V

-5.0E-07 Fg3 = -7.36 x10

-12
V
2

-7.0E-07 Particle Radius

0.73mm
-11 2
-9.0E-07 0.92mm Fg2 = -1.01 x10 V
1.14mm
-1.1E-06 1.17mm
1.38mm
-11 2
-1.3E-06 Fg3 = -1.15 x10 V

-1.5E-06 -11 2
Fg5 = -3.13 x10 V

Fig. 10. Negative dielectrophoresis: quartz in ethanol.

 G.R. Ballantyne, P.N. Holtham / Minerals Engineering 23 (2010) 350358 357

Fig. 11. Force calibration with particle size.

is strongly dependant on the position of the particle from the cen-

tral pin, varying inversely as the third power of the radial position
(see Fig. 13). This creates problems in measuring the force as the
particle occupies a large volume in the eld and the position cannot
be dened by a point. Either a theoretical mathematical model or an
empirical model tted experimentally is necessary to be able to
determine the eld gradient term (rE) in the force equation more
precisely.

5. Conclusion

Dielectrophoresis has the potential to separate mineral particles

of different permittivities; however, there are still issues that have
to be overcome before it can be employed at plant-scale. The major
technological issue is the absence of a suitable medium with a var-
Fig. 12. (a) View of electric eld lines created by a polarisable particle suspended iable permittivity, low conductivity and high electrical breakdown
within a pin-plate electrode system (Hughes, 1999) and (b) picture of quartz
particle suspended by glass bre between the pin-plate electrodes of the single
threshold which is also low cost, is safe and environmentally
particle test cell. friendly.
One of the major technological issues for dielectrophoresis is
the availability of an appropriate medium for the separation. Obvi-
Fig. 12 shows the electrode conguration and denes the values of ously, water would be an ideal liquid medium for environmental
r1, r2 and r. Eq. (5) shows that the electric eld gradient, and there- and economic reasons. However, the permittivity is very high, ions
fore the dielectrophoretic force in a coaxial electrode conguration dissolved in the water increase the conductivity and therefore the

Particle position variation

1.4E-03

1.2E-03 Dielectrophoretic force

1.0E-03
Gravitational force
Force (N)

8.0E-04

6.0E-04

4.0E-04

2.0E-04

0.0E+00
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
particle position from inner electrode (mm)

Fig. 13. Force variation as a function of distance between particle and pin electrode.
358 G.R. Ballantyne, P.N. Holtham / Minerals Engineering 23 (2010) 350358

power draw, electrolysis can create turbulence and it is immiscible Chen, K.W.L.L., 1969. Dielectrophoresis of Solids in Aqueous Solutions. Stillwater,
Oklahoma State University.
with most hydrocarbons that can be used to lower the permittivity.
Church, R.H., Webb, W.E., Salsman, J.B., 1988. Dielectric Properties of low-loss
The best way to alter the mediums permittivity is to mix propor- minerals. Report of Investigations United States, Bureau of Mines.
tions of polar alcohols with non-polar hydrocarbons. However, it is Cumbane, A.J., Miles, N.J., Lester, E., Kingman, S.W., Bradshaw, S.M., 2008. Dielectric
important to choose chemicals that are safe and environmentally properties of sulphide minerals. Processing Microwave Technologies.
Feeley, C.M., 1969. Dielectrophoresis of Solids in Liquids of Differing Dielectric
friendly. Constant and Conductivity. Stillwater, Oklahoma State University.
The minerals industry is facing increased comminution costs, as Goossens, K., Van Biesen, L., 1989. Electrophoretic effects in ac dielectrophoretic
a result of decreasing ore grades and increasing energy costs. solidsolid separations. Separation Science and Technology 24, 5162.
Goossens, K., Van Biesen, L., Boven, A., Castelain, P., 1992. Theoretical and
Coarse liberation and removal of gangue has the potential to re- experimental analysis of the dielectric properties of powders. In: IEE
duce the quantity of rock proceeding to ne grinding, thereby, Conference Publication, 363 ed.
greatly reducing the energy requirements for comminution. Pres- Hateld, H.S., 1924a. Dielectric separation: a new method for treatment of ores.
Institution of Mining and Metallurgy Bulletin 233, 335342.
ently, otation of sulphides is generally not effective for particles Hateld, H.S., 1924b. Means and process of separating substances one from another.
larger than approximately 150 microns. Therefore, a new technol- United States Patent Ofce.
ogy is required to remove the liberated gangue particles between Hawk, I.L., 1966. Effects of Nonuniform Electric Fields on Real Dielectrics in Water.
Stillwater, Oklahoma State University.
0.52 mm. Flotation is not effective at coarser size ranges because Hughes, M.P., 1999. AC electrokinetics: applications for nanotechnology. Seventh
it relies on surface forces and as particle size increases, the surface foresight conference on molecular nanotechnology. Santa Clara, California.
area to volume ratio decreases. A force that acts not only on the Hughes, M.P., Pethig, R., Wang, X.B., 1996. Dielectrophoretic forces on particles in
travelling electric elds. Journal of Physics D: Applied Physics 29, 474482.
surface, but on the particle as a whole is therefore required in order
Jordan, C.E., Sullivan, G.V., Davis, B.E., Weaver, C.P., 1980. Report of Investigations
to separate particles at these coarse size ranges. Dielectrophoresis A Continuous Dielectric Separator for Mineral Beneciation. University of
is a good example of a volumetric force and therefore may be an Alabama, United States Department Of The Interior Bureau Of Mines.
appropriate technology that can be used in coarse removal of gan- Jordan, C.E., Sullivan, G.V., 1985. Fundamentals of continuous dielectric separation
of minerals. In: SME-AIME Fall Meeting. Society of Mining Engineers of AIME,
gue. However, it is imperative that the recovery of the valuable Albuquerque, New Mexico.
component to ne grinding is not reduced; therefore, the effect Keller, G.V., 1981. Electrical properties of rocks and minerals. In: Carmichael, R.S.
of composition on dielectrophoresis must also be quantied. (Ed.), Handbook of Physical Properties of Rocks and Minerals. Boca Raton, CRC
Press.
The results of this study show the potential of dielectrophoresis National Materials Advisory Board (1981). Comminution and Energy Consumption,
for coarse mineral separation. However, they also highlight the Washington, D.C, NMAB-364, National Academy Press.
ability to use dielectrophoresis as a measurement tool. If the mea- Parkhomenko, E.I., 1967. Electrical Properties of Rocks. Plenum Press, New York
City.
surement of the dielectrophoretic force on a single particle can be Plaksin, I.N., Olonsky, N.F., 1966. Review of Electrical Separation Methods in
calibrated to the theoretical force equation (Eq. (2)), the permittiv- Mineral Technology. Institution of Mining And Metallurgy.
ity of a single particle can be measured. This technique can poten- Pohl, H.A., 1950. The motion and precipitation of suspensoids in divergent electric
elds. Journal of Applied Physics 22, 869871.
tially be used to quantify single particle composition and Pohl, H.A., 1958. Some effects of nonuniform elds on dielectrics. Journal of Applied
liberation. To achieve this, a better understanding of the particle Physics 29, 11821188.
size and electrical eld gradient measurements is required. Pohl, H.A., 1960. Nonuniform eld effects in poorly conducting media. Journal of the
Electrochemical Society 107, 386390.
Pohl, H.A., 1978. Dielectrophoresis The Behaviour of Neutral Matter in
Acknowledgements Nonuniform Electric Fields. Cambridge University Press, Oklahoma.
Pohl, H.A., Pethig, R., 1976. Dielectric measurements using non-uniform electric
The nancial support of the Australian Research Council Link- eld (dielectrophoretic) effects. Journal of Physics E: Scientic Instruments 10,
190193.
age Scheme, AMIRA International, and State Governments of South Raju, G.G., 2003. Dielectrics in Electric Fields. Marcel Dekker, Inc., Windsor, Ontario,
Australia and Victoria, is gratefully acknowledged. Canada.
Financial support was also provided as a Student Scholarship by Salsman, J.B., 1991. Measurement of dielectric properties in the frequency range of
300 MHz to 3 GHz as a function of temperature and density. Ceramic
the Centre for Sustainable Resource Processing, which is estab- Transactions 21, 203214.
lished and supported under the Australian Governments Coopera- Sharif, S., 1995. Chemical and mineral composition of dust and its effect on the
tive Research Centres Program. dielectric constant. IEEE Transactions on Geoscience and Remote Sensing 33,
353359.
Tarjan, G., 1986. Electrical separation. Mineral Processing, Concentration, Flotation,
References Separation Backup Processes 2, 376411.
Tickell, F.G., 1931. Preparation of Specimens Dielectric Separation. The
Abdel-Salam, M., 1993. On the dielectric separation of solid particles. IEEE Examination of Fragmental Rocks. Stanford University Press, Stanford
Transactions on Industry Applications 29, 268273. University, California.
Aldaeus, F., Lin, Y., Amberg, G., Roeraade, J., 2006. Multi-step dielectrophoresis for Tromans, D., 2008. Mineral comminution: energy efciency considerations.
separation of particles. Journal of Chromatography A 1131, 261266. Minerals Engineering 21, 613620.
Andres, U.T., 1979. Separation of solid particles in liquid dielectrics. Powder Verschure, R.H., Ijlst, L., 1966. Apparatus for continuous dielectric-medium
Technology 23, 8597. separation of mineral grains. Nature 211, 619620.
Andres, U.T., 1980. Mechanics of Separating Particulate Solids in Liquid Dielectrics Von Hippel, A.R. (Ed.), 1954. Dielectrics and Waves. John Wiley & Sons,
and Magnetics. Department of Mineral Resources Engineering, Royal School of Massachusetts.
Mines, Imperial College of Science and Technology, London, England. Wang, X.B., Huang, Y., Gascoyne, P.R.C., Becker, F.F., 1997. Dielectrophoretic
Andres, U.T., 1982. Method for electrostatic separation of solids. UK Patents Ofce. manipulation of particles. IEEE Transactions on Industry Applications 33,
Great Briton, De Beers Industrial Diamond Division. 660669.
Andres, U.T., 1996. Dielectric separation of minerals. Journal of Electrostatics 37, Young, K.F., Frederikse, H.P.R., 1973. Compilation of the static dielectric constant of
227248. inorganic solids. Journal of Physical Chemistry Reference Data 2, 313405.
Berg, G.A., 1936. Notes on the dielectric separation of mineral grains. Journal of Zheng, Y., Wang, S., Feng, J., LI, X., 2005. Measurement of the complex permittivity of
Sedimentary Petrology 6, 2327. dry rocks and minerals: application of polythene dilution method and
Briggs, C., 1992. The Application of the Theory of Mixed Dielectrics to Compositional Lichteneckers mixture formulae. Geophysics Journal International 163, 1195
Analysis within an Electrostatic Mineral Separation Plant. University of 1202.
Queensland.