Waste Management 27 (2007) 17471754

www.elsevier.com/locate/wasman

Recovery of PET from packaging plastics mixtures by wet shaking table

M.T. Carvalho *, E. Agante, F. Durao
CVRM Centro de Geosistemas, Instituto Superior Tecnico Av. Rovisco Pais, 1049-001 Lisbon, Portugal

Accepted 24 August 2006

Available online 21 November 2006

Abstract

Recycling requires the separation of materials appearing in a mass of wastes of heterogeneous composition and characteristics, into
single, almost pure, component/material ows. The separation of materials (e.g., some types of plastics) with similar physical properties
(e.g., specic gravity) is often accomplished by human sorting. This is the case of the separation of packaging plastics in municipal solid
wastes (MSW). The low cost of virgin plastics and low value of recycled plastics necessitate the utilization of low cost techniques and
processes in the recycling of packaging plastics.
An experimental study was conducted to evaluate the feasibility of production of a PET product, cleaned from PVC and PS, using a
wet shaking table. The wet shaking table is an environmentally friendly process, widely used to separate minerals, which has low capital
and operational costs. Some operational variables of the equipment, as well as dierent feed characteristics, were considered. The results
show that the separation of these plastics is feasible although, similarly to the mineral eld, in somewhat complex ow sheets.
 2006 Elsevier Ltd. All rights reserved.

1. Introduction The application of mechanical separation to these mate-

Recycling of solid wastes is a common practice in coun-
tries committed to sustainable development. The European
Parliament and Council Directive 94/62/EC, of 20 Decem-
ber 1994, on packaging and packaging waste requires that
a minimum of 15% of plastics be recycled by the end of
2005 while the Directive 2004/12/EC of the European Par-
liament and of the Council, of 11 February 2004 states that
a minimum of 22.5% must be recycled by the end of 2011.
In Portugal, the primary materials used in plastic pack-
aging are PE (polyethylene), with high, medium or low
density, PP (polypropylene), PET (polyethylene terephthal-
ate), PS (polystyrene) and PVC (polyvinyl chloride) (see
Table 1). In the municipal solid waste management sys-
tems, these materials are separated at the source as a mix-
ture of plastics, which are recovered by selective collection
(drop-o-points or centres or door-to-door collection), and
then are separated by material in sorting plants and sent to
the dierent recycling plants.
*
Corresponding author. Tel.: +351 218417247; fax: +351 218417442.
E-mail address: teresa.carvalho@ist.utl.pt (M.T. Carvalho).
rials is rather dicult due to the similarity of their proper-
ties. Separation of polyolens (PE and PP), which have
specic gravities less than 1, from the other three polymers,
which have specic gravities higher than 1, is quite simple.
The separation of PET, PS and PVC from each other is
commonly done manually or, in few cases, using very
sophisticated mechanical processes. These processes can
use some sort of radiation to identify the material and com-
plex equipment to sort each object from the waste ow.
These techniques are, however, very expensive.
The automatic/mechanical separation of these materials
with minimal manual intervention is an objective that has
been pursued for several years. For decades, many
researchers around the world have studied the separation
of plastics, applying techniques and processes used in the
minerals separation technology. Numerous experimental
studies have been reported on the application of mineral
processing techniques like froth otation (Carvalho et al.,
2004; Drelich et al., 1998; Fraunholcz, 1997; Le Guern
et al., 2000; Pascoe and OConnell, 2003; Shen et al.,
2002a,b; Shibata et al., 1996); gravity methods like jigging
and wet shaking table; or mixtures of processes (Agante

0956-053X/$ - see front matter  2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.wasman.2006.08.015
1748 M.T. Carvalho et al. / Waste Management 27 (2007) 17471754

Table 1 Wills (1997) dened a concentration criterion to evaluate

Plastics packaging consumption and recycling in Portugal the diculty of separation. This criterion, CC, is dened
Consumption (2003) Recycling of domestic as
plastics packaging
Domestic Domestic + industrial CC d h  d f =d l  d f
(2004) (tons)
(tons) (tons)
PE 53,500 126,000 6600 where dh, dl, and df are the specic gravity of the heavy spe-
PET 55,000 55,000 4000 cies, light species and uid, respectively.
PS 31,000 31,000
PVC 5000 5500 137
The eciency of the separation increases with CC. Wills
PP 12,500 41,000 0 (1997) denes an indicative value of 2.5 to characterize a
Source: www.plastval.pt.
relatively easy separation. However, it should be pointed
out that other variables besides specic gravity, such as
grain size and particles shape, can play an important role
et al., 2004; Fedriga et al., 2002; Dodbiba et al., 2002, 2003; in the separation.
Pascoe and Hou, 1999; Tsunekawa et al., 2004). A typical wet shaking table (see Fig. 1) has a slightly
The present paper reports an experimental work con- inclined rectangular deck on which a sheet of water is in
ducted to evaluate the feasibility of separation of PET from motion. Longitudinally, a vibrating movement, generated
PS and PVC, observing the specications of the Portuguese by a mechanism that superimposes a slow forward stroke
recycling industry. The Portuguese Green Dot system and a rapid return, causes the movement of particles along
includes products made out of PET. With respect to con- the deck. Wash water is introduced longitudinally along
tamination by other types of plastics the limits are as fol- the feed side by a shower. The table can have ries, gener-
lows: 0.5% PVC, 2% PS and 15.25% polyolens. ally placed parallel to the deck motion, that generate the
pulsation of particles between them (Burt, 1984).
2. Wet shaking table The feed is introduced at the upslope corner of the table
(see Fig. 2). The separation between particles is a conse-
The wet shaking table is a process widely used in mineral quence of the combined eects of the asymmetrical motion
processing and in coal cleaning. It is a process with a low of the deck, water movement and turbulence ow between
cost per tonne of throughput, requiring low installed power the ries. Due to the gravity force, to the eld of asymmet-
with minimum environmental impact. It makes use of sev- rical accelerations, static friction and resistance to the
eral mechanisms and is such a complex process that an water ow, the particles will move diagonally across the
empirical approach usually is utilized for its understanding. deck, from the feed end, spreading over the deck. They will
The wet shaking table is one among the gravity concen- attain dierent exit positions over the concentrate and tail-
tration methods that separate minerals or other bulk mate- ings edges depending upon their characteristics (e.g., den-
rials of dierent specic gravity by their relative movement sity, grain size and shape). The net result is a spread of
in response to gravity and other forces. A minimum density particles in bands. One or two cutters (that split the con-
dierence between the materials is required for separation. centrate, tailing and middling products) are visually

Fig. 1. Wet shaking table equipment.

 M.T. Carvalho et al. / Waste Management 27 (2007) 17471754 1749

Wash Water Feed

C o n c e n t r a t e E dg e
riffles

Heavier particles Motion

Middling particles
Lighter particles
Tailings Edge

Lighter particles

Fig. 2. Shaking table scheme.

positioned in appropriate locations in order that the tions: 02, 24 and 45.6 mm. As the composition of
desired products can be obtained. MSW from drop-o points changes with time (there is even
This process can eectively separate particles with simi- an overall trend for reduction in PVC grade) and feed com-
lar densities if a narrow size range is used. In this case, in position may inuence table performance (Manser et al.,
rough terms, the heavier particles will travel longitudinally 1991), a study of the eect of feed composition was con-
and lighter ones will be transported by wash water ow. It ducted. For each size fraction, two types of mixtures com-
should be pointed out that a middling product can bining the three plastics were prepared:
occur, even when liberated materials are used. In this case,
this product is composed of coarse heavy particles and ne  Samples with actual proportion of plastics (AP): these
light particles. samples were prepared according to the actual relative
The performance of the wet shaking table can be proportion in which these plastics appear in drop-o-
inuenced by a number of machine parameters, such as points: 75% PET, 15% PVC and 10% PS. This propor-
the frequency and amplitude of the tables movement, tion was calculated in a sampling study carried out in
longitudinal and transversal slope of the table, feed ow a sorting plant of MSW from four towns in the metro-
rate and pulp density, wash water ow rate and positions politan area of Lisbon (Agante, 2003).
of the cutters. The eects of feed characteristics, such as  Samples with equal proportions (EP) of each plastic in
particle grain size and shape and feed composition, on mixture (33%).
the separation results can be important as well.
Design parameters, such as table shape, rie pattern
and deck surface material, also have signicant eects 3.2. Shaking table
on the separation, although they are xed for a specic
table. The laboratory shaking table used in the study is a typ-
ical one, commonly used to concentrate minerals (see
3. Experimental approach Fig. 1). It has one deck of rectangular form with
350 1000 mm, with linoleum as surface material. It is par-
3.1. Samples tially ried, with ries parallel to the deck motion. The rif-
es have 5 mm height on the feed end with decreasing
Samples of virgin packages of PET, PVC and PS were height from the feed edge to the concentrate edge. The
received from several companies, already shredded, with gap between ries is 12 mm. The movement frequency
maximum grain size of 5.6 mm. The characteristics of the and amplitude are constant, equal to 200 cycles/min and
plastics are detailed in Table 2. 10 cm, respectively. The longitudinal deck inclination is
The samples were classied by conventional sieving xed while the transversal one is adjustable.
with square aperture openings 2, 4 and 5.5 mm, in frac-
3.3. Experimental procedure

Table 2 A total of 36 experiments were carried out in accordance

Characterization of the plastics used in the study to a factorial design with four variables; two were related
Composition Colour Measured densitya (g/cm3) Source to equipment: wash water ow rate (Q) and transversal
PET Blue 1.32 Sulpet deck inclination (inc), which are the only parameters that
PVC Yellow 1.38 Vitor Guedes are adjustable in this table; and two related to feed charac-
PS White 1.05 Logoplaste teristics: particles grain size (U) and feed composition (see
a
Measured at Institute of Polymer and Composite Materials of Uni- Table 3). Other factors, such as feed ow rate and pulp
versity of Minho. density, were kept constant.
1750 M.T. Carvalho et al. / Waste Management 27 (2007) 17471754
Table 3
Changed variables
Level Feed characteristics Operational conditions
Feed composition U (mm) Q (l/min) inc ()
1 AP (75% PET + 15% PVC 02 2 5
+ 10% PS)
2 24 4 upgraded in a subsequent stage of concentration). In this
3 EP (33% PET + 33% PVC 45.6 6 12.5
+ 33% PS)
4. Results and discussion
According to the above mentioned concentration crite-
rion, CC, the separation between the three plastics should
be easy in the cases of the PVC/PS mixture (CC = 6.4) and
PET/PS mixture (CC = 7.6). In the case of PET/PVC mix-
ture (CC = 1.19), the eciency of the separation should be
more dicult than in the other two cases.
In the experimental study performed, three bands were
formed with fuzzy boundaries between them. Although
the measured density of the PVC sample used was slightly
higher than measured density of the PET, the experimental
tests showed that PET behaved as a heavy product.
Roughly, the three plastics arranged themselves in the fol-
lowing way: PET mainly in the heavy product band loca-
tion, PS in the light one and PVC in between. In the tests
carried out with the AP samples (with grade in PET higher
than in PVC and PS), PET appeared in all of the bands
formed and PS rarely appeared as a heavy product.
It should be pointed out that the concentration crite-
ria, CC, an indicator variable that has been validated in
the concentration of particulate solid material of mineral
origin, only takes into account density to evaluate the dif-
culty of separation by gravity concentration processes.
However, besides density, the primary property of particu-
late material, other secondary properties, such as particle
shape and size can have a signicant eect on gravity sep-
aration. Moreover, Pascoe and Hou (1999) showed that the
interfacial tension can play an important role in the behav-
ior of plastics.
In this case, the samples were sieved in close size inter-
vals; therefore, in one size range the particle size can be
assumed as constant. However, the shape of the particles,
due to the dierent origins of the plastics, was quite dier-
ent from plastic to plastic and even for the same type of
plastic. This secondary property along with density and,
very likely, interfacial tension (that could not be measured)
may explain the obtained results.
As shown in Fig. 2, when the particles leave the shaking
table, they spread over part of two edges of the table (con-
centrate and tailings edges). Usually, there is not a sharp
boundary between the bands of the dierent plastic materi-
als. Consequently, the cutters are empirically positioned
(adjusted on-line) according to the purpose of the opera-
tion. If a clean/nal PET concentrate is desired, the cutters
should be located such that a high grade of the high-value
material is attained. Conversely, if the shaking table has a
roughing (use of a moderate separating force to remove
fully liberated particles) duty, the goal of the operation is
to maximize the recovery of the high-value material irre-
spective of the concentrate grade (because it can be
case, the cutters are generally positioned dierently from
the former case. In the experimental work conducted, the
cutters were positioned in each test in a way that an
upgraded product of PET could be produced.
The criteria used to evaluate the performance of the pro-
cess were the usual component grade (purity degree) and
component recovery (percentage of component contained
in the feed that is recovered in one product, concentrate
or tailing) of polymer in each product.
Contrary to what happens in the mineral processing
eld, where quick and accurate analysis of product grade
is possible, there is not yet available, as far as the authors
know, an adequate procedure to measure plastic grade in
granulated mixtures of plastics, which is currently a
research topic. While this problem has not yet been solved,
in research studies where the feasibility of separation is
addressed, the common practice is manual sorting of the
plastics. To achieve this, the plastics in the mixtures must
exhibit dierent colours. In the present study, the plastics
with dierent colours (see Table 2) were separated by
hand-sorting and the grade and recovery were calculated
from the resulting weights.
Figs. 3 and 4 show, respectively, the PET grade and
PET recovery in PET concentrate versus feed grain size
for dierent settings of table slope and wash water ow
rate, when AP samples were used. In all tests it was pos-
sible to obtain a PET concentrate with grade in PET over
84%. A pure PET concentrate, with recovery higher than
20%, cleaned from PS and PVC could be produced only
when the smallest grain size samples were used. In this
inc=5; Q=2 inc=5; Q=6
inc=5; Q=4 inc=12.5; Q=2
inc=12.5; Q=4 inc=12.5; Q=6
100.00
PET grade (% )
80.00
60.00
40.00
20.00
0.00
0-2 2-4 4-5.6
(mm)
Fig. 3. PET grade in PET concentrate as a function of feed grain size for
dierent transversal deck inclinations and wash water ow rates AP
samples.
 M.T. Carvalho et al. / Waste Management 27 (2007) 17471754 1751

inc=5; Q=2 inc=5; Q=6 inc=5; Q=2 inc=5; Q=6

inc=5; Q=4 inc=12.5; Q=2 inc=5; Q=4 inc=12.5; Q=2
inc=12.5; Q=4 inc=12.5; Q=6 inc=12.5; Q=4 inc=12.5; Q=6
100.00 100.00

PET recovery (% )
PET recovery (% )

80.00 80.00

60.00 60.00

40.00 40.00

20.00 20.00

0.00 0.00
0-2 2-4 4-5.6 0-2 2-4 4-5.6
(mm) (mm)

Fig. 4. PET recovery in PET concentrate as a function of feed grain size Fig. 6. PET recovery in PET concentrate as a function of feed grain size
for dierent transversal deck inclinations and wash water ow rates AP for dierent transversal deck inclinations and wash water ow rates EP
samples. samples.

case, the PET recovery in the concentrate is less than

30%.
The increase in grain size reduces the selectivity of the
separation, that is, the grade in PET is only slightly higher
than 80% (8386%) although a higher recovery of PET can
be achieved. The eects of other operational variables,
besides cutter position, on separation results were not
important.
Figs. 5 and 6 show the corresponding results for the
tests carried out with mixtures of equal proportions of
the three plastics (with a lower feed grade in PET than
the former ones). In this case, a high concentration ratio
(PET concentrate grade divided by PET feed grade) was
attained. It was also possible to obtain pure PET concen-
trates when a ner feed grain size was used although with
slightly less recovery rates than in the case where AP sam-
ples were used. This was obtained when a higher deck
slope was used.
inc=5; Q=2
inc=5; Q=4
inc=12.5; Q=4
100.00
80.00
PET grade (% )
5. Outline of a ow sheet for PET concentration by wet
shaking table
Fig. 7 shows the recovery versus grade curve obtained
from the experimental results corresponding to tests carried
out with AP samples (with dierent experimental condi-
tions, grain sizes and wash water ow rates). The data
are reasonably interpolated by a quadratic polynomial.
The interpolation was subjected to one constraint: the
curve must include the point (75% PET grade, 100% PET
recovery). Two particular points are represented: feed
grade (75% PET grade, 100% PET recovery) and the utopic
point (100% PET grade, 100% PET recovery).
The results show that it was possible, with a single sep-
aration stage, to obtain a nal PET concentrate within the
recycling industry specications, although with a very low
PET recovery. Alternatively, it is possible to produce a
pre-concentrate of PET. It should be emphasized that
grade (purity degree) and recovery (yield) are antagonist
criteria: if a higher grade is desired a lower recovery will
be obtained and vice-versa. Therefore, a decision must be
inc=5; Q=6
made on which one of the criteria has a higher priority of
inc=12.5; Q=2
fulllment.
inc=12.5; Q=6
Fig. 8 depicts a ternary diagram with the grade of the
three polymers in PET concentrate. As shown, in most test
cases PS was totally removed from the PET concentrate,
and PVC was responsible for the lower PET grade values

60.00 in PET concentrates (in cases when a pure PET concentrate

was not produced).
40.00 Mineral processing is rarely performed in only one stage
due to, among other reasons, the ineciencies of the unit
20.00
operations. Frequently, the recovery of one mineral is
0.00 achieved after the raw ore processing on a very complex ow
0-2 2-4 4-5.6 sheet with roughing (use of a moderate separating force to
(mm) remove fully liberated particles), scavenging (use of strong
Fig. 5. PET grade in PET concentrate as a function of feed grain size for separating force to remove the remaining valuable species)
dierent transversal deck inclinations and wash water ow rates EP and cleaning (use of a low separating force to upgrade the
samples. rougher concentrate by removing misplaced non valuable
1752 M.T. Carvalho et al. / Waste Management 27 (2007) 17471754

Fig. 7. PET recovery versus PET grade in PET concentrate tests carried out with AP samples.

species material) operations are used (Kelly and Spottis-
wood, 1982). It is, also common to recycle all or part of
the stream being processed and to regrind the materials dur-
ing the separation stages. Even when the species to separate
are fully liberated, multiple stages can be used.
In the case under study, one possibility is to produce a
pre-concentrate of PET using a coarser grain size, which
should be cleaned/upgraded, after regrinding, to achieve
the specications. An alternative is to grind all of the mate-
rial to the nest grain size, to recover a pure PET concen-
trate and to recycle, as many times as needed, the tailings to
separate misplaced particles. Cleaning using dierent pro-
cessing methods can be also foreseen.
For a proper decision, it is necessary to evaluate alterna-
tive potential circuit arrangements. In the absence of exper-
imental pilot scale tests, the evaluation can be made by
circuit simulation. In the present work, a simulation study
was made by considering the classic/conventional arrange-
ment: rougher + scavenger + cleaner stages with recycling
(Fig. 9).

Fig. 8. Ternary diagram of polymers grade in PET concentrate. Fig. 9. Classic processing stage arrangement.
 M.T. Carvalho et al. / Waste Management 27 (2007) 17471754 1753

Table 4
Mass balance obtained by simulation of the classic stages arrangement
S1 S2 S3 S4 S5 S6 S7 S8
PET 75.0 220.6 44.1 176.5 4.4 39.7 105.9 70.6
PVC 15.0 32.4 24.3 8.1 14.6 9.7 7.7 0.4
PS 10.0 10.2 10.1 0.1 10.0 0.1 0.1 0.0
Weight (%) 100.0 263.2 78.5 184.7 29.0 49.5 113.7 71.0
PET recovery (%) 100.0 294.1 58.8 235.3 5.9 52.9 141.2 94.1
PET grade (%) 75.0 83.8 56.2 95.6 15.2 80.2 93.1 99.4
PVC grade (%) 15.0 12.3 31.0 4.4 50.3 19.6 6.8 0.6
PS grade (%) 10.0 3.9 12.9 0.1 34.5 0.2 0.1 0.0

The assumptions/approximations taken into account
were:
1. The plant feed composition contains, basically, three
components (PET, PVC and PS);
2. The component partition numbers (equivalent to com-
ponent recoveries in dense product) are independent of
the feed stage composition;
3. The partition number values assumed below took into
account the lab tests results:
3.1 Rougher partition numbers are based on PET,
PVC and PS recoveries, in dense product, for
region I (pre-concentrate), with assumed values of
0.8, 0.25 and 0.01, respectively;
3.2 Cleaner partition numbers are set based on PET,
PVC and PS recoveries, in dense product, for
region II (nal concentrate), but considering
slightly more optimistic values, because this stage
is expected to receive a PET richer feed, with values
of 0.4, 0.05 and 0.001, respectively;
3.3 Scavenger promotes recovery, so the partition
numbers, for all components, are greater than the
corresponding rougher values for PET, PVC and
PS recoveries, in dense product, with values of
0.9, 0.4 and 0.01, respectively.
Based on these assumptions, the simulation consisted in
a simple steady state (linear) mass balance, consisting of
the formulation and solution of a system of linear balance
equations.
The simulation results (Table 4) show that it may be
possible, using the considered arrangement under the sta-
ted approximations, to recover 94.1% of PET in a concen-
trate with a grade of 99.4% in PET. To achieve the specied
99.5% purity in PET, an extra cleaning stage would be
eventually needed.
As a nal remark, it should be pointed out that the
scope of the experimental work was limited, as it was car-
ried out with plastics coming from only a few sources. The
same type of plastic can appear in MSW with dierent
physical properties depending on the source. For instance,
the density of each type of plastic can change. PET and
PVC density variation intervals can even overlap. As
referred before, the particle shape and the interfacial sur-
face tension can have an important role in plastics behavior
too. Previous contact with contaminants can also inuence
the separation. Future work will focus on these variables.
It should be emphasized, as well, that the industrial
application of separation processes to granulated mixtures
is limited today by the lack of adequate procedures for the
analyses of plastics to both control and evaluate process
performance. Research on this subject also is under
development.
6. Conclusions
The experimental work carried out showed that it is pos-
sible, in a single operation, by controlling feed grain size, to
obtain:
 a pure PET concentrate within the recycling industry
specications, although with low recovery; or
 a PET pre-concentrate that must be re-treated for recy-
cling industry specications achievement.
Due to the ineciency of the processes, gravity concentra-
tion, and particularly wet shaking table, is usually operated
in multiple stages. Simulation results show that it should
be possible to attain the desired results with a simple ow
sheet. Further experimental work will consider the study of
dierent ow sheet congurations. The study will continue,
as well, with the experimentation with a larger set of samples
from diverse sources and dierent characteristics.
Acknowledgements
The authors acknowledge FCT (Fundacao para a Cien-
cia e Tecnologia) for nancial support of the study (project
number POCTI/ECM/40799/2001) and Logoplaste, Sulpet
and Vitor Guedes companies for the supply of samples, as
well as the Institute of Polymer and Composite Materials
of University of Minho for the plastics density analysis.
References
Agante, E., 2003. Characterization of solid wastes from packaging drop-
o points. Internal Report, SEP-RSU project, CVRM, Instituto
Superior Tecnico.
Agante, E., Carvalho, M.T., Durao, F., Bartolo, T., 2004. Separation of
PET, PVC and PS using gravity concentration and froth otation. In:
1754 M.T. Carvalho et al. / Waste Management 27 (2007) 17471754
Global Symposium on Recycling, Waste Treatment and Clean
Technology (REWAS04), 2629/2004, Madrid, Spain, pp. 16911700.
Burt, R., 1984. Gravity Concentration Technology. Elsevier Science.
Carvalho, M.T., Durao, F., Felcio, A., 2004. Separation of plastic from
municipal solid wastes. In: Carvalho, M.G., Afgam, Naim (Eds.), New
and Renewable Technologies for Sustainable Development. Swets &
Zeitlinger Publishers, Netherlands.
Dodbiba, G., Haruki, N., Shibayama, A., Miyazaki, T., Fujita, T., 2002.
Combination of sinkoat separation and otation technique for
purication of shredded PET bottle form PE or PP akes. Interna-
tional Journal of Mineral Processing 65, 1129.
Dodbiba, G., Shibayama, A., Sadaki, J., Fujita, T., 2003. Combination of
triboelectrostatic separation and air tabling for sorting plastics from a
multi-component plastic mixture. Materials Transactions 44 (12),
24272435.
Drelich, J., Payne, T., Kim, J., Miller, D., 1998. Selective froth otation of
PVC from PVC/PET mixtures for the plastics recycling industry.
Polymer Engineering Science 38 (9).
Fedriga, M., Bevilaqua, P., Fraunholcz, N., Rem, P., 2002. Settling
behavior of shredded plastic particles in jigging separation. In: 6th
World Congress on Integrated Resources Management Proceedings,
Geneva, Switzerland.
Fraunholcz, N., 1997. Plastics otation. Ph.D. thesis, Delft University of
Technology.
Kelly, E.G., Spottiswood, D.J., 1982. Introduction to Mineral Processing.
John Wiley & Sons.
Le Guern, C., Conil, P., Houot, R., 2000. Role of calcium ions in the
mechanism of action of a lignosulphonate used to modify the
wettability of plastics for their separation by otation. Minerals
Engineering 13 (1), 5363.
Manser, R.J., Barley, R.W., Wills, B.A., 1991. The shaking table
concentratorthe inuence of operating conditions and table param-
eters on mineral separationthe development of a mathematical
model for normal operating conditions. Minerals Engineering 4 (3/4),
369381.
Pascoe, R.D., Hou, Y.Y., 1999. Investigation of the importance of particle
shape and surface wettability on the separation of plastics in a
Larcodems separator. Minerals Engineering 12 (4), 423431.
Pascoe, R.D., OConnell, B., 2003. Development of a method for
separation of PVC and PET using ame treatment and otation.
Minerals Engineering 16, 12051212.
Shen, H., Pugh, R.J., Forssberg, E., 2002a. Floatability, selectivity and
otation separation of plastics by using a surfactant. Colloids and
Surfaces A 196, 6370.
Shen, H., Forssberg, E., Pugh, R.J., 2002b. Selective otation separation
of plastics by chemical conditioning with methyl cellulose. Resources,
Conservation and Recycling 35, 229341.
Shibata, J., Matsumoto, S., Yamamoto, H., Kusaka, K., Pradip, E., 1996.
Flotation separation of plastics using selective depressants. Interna-
tional Journal of Mineral Processing 48, 127134.
Tsunekawa, M., Hori, K., Hirajima, T., Hiroyochi, N., Ito, M., 2004.
Application of jig separator to recycle plastics used in electrical
appliances. In: Global Symposium on Recycling, Waste Treatment and
Clean Technology (REWAS04), 2629/2004, Madrid, Spain, pp.
16511660.
Wills, B.A., 1997. Mineral Processing Technology. Pergamon Press.