B. TROPENAUER et al.: SUSTAINABLE WASTE-TREATMENT PROCEDURE FOR THE SPENT POTLINING (SPL) ...
277–284
SUSTAINABLE WASTE-TREATMENT PROCEDURE FOR THE
SPENT POTLINING (SPL) FROM ALUMINIUM PRODUCTION
TRAJNOSTNI POSTOPEK OBDELAVE ODPADKOV SPL IZ
PROIZVODNJE ALUMINIJA
Bla` Tropenauer
1
, Du{an Klinar
2
, Niko Samec
3
, Janvit Golob
4
, Jo`e Kortnik
5
1
TALUM d.d. Kidri~evo, Tovarni{ka cesta 10, 2325 Kidri~evo, Slovenia
2
Scientific Research Centre BISTRA Ptuj, Slovenski trg 6, 2250 Ptuj, Slovenia,
3
University of Maribor Faculty of mechanical engineering, Smetanova ulica 17, 2000 Maribor, Slovenia
4
Univerza of Ljubljana, Faculty of Chemistry and Chemical Technology, Ve~na pot 113, 1000 Ljubljana, Slovenia
5
University of Ljubljana Faculty of Natural Sciences and Engineering, A{ker~eva ulica 12, 1000 Ljubljana, Slovenia
Prejem rokopisa – received: 2018-07-13; sprejem za objavo – accepted for publication: 2018-11-29
doi:10.17222/mit.2018.147
Some different possible solutions to the problem of industrial waste originating from primary aluminium production (also
known as SPL) have been investigated. The latest techniques to treat SPL are presented, with the highlight being flotation and
chemical treatment technology. Laboratory tests confirmed that the SPL treatment model is not feasible and viable within partial
solutions. Only the systematically achieved solution can be in accordance with a wide range of contemporary demands. Partly,
the cause is the allotropic form of the carbon cathodes produced and partly also the impurities in the waste material that prevent
an effective one-phase treatment. Laboratory investigations of reactive extraction open up the possibilities for wide SPL material
utilization. The preliminary inquiry confirms that the interdisciplinary approach leads to open solutions leading to possible
market interest or generated value from the waste as the main goal of a circular economy.
Keywords: spent pot line, waste treatment, flotation, alkaline–acidic treatment
Raziskane so bile nekatere mo`ne re{itve za problem industrijskih odpadkov, ki izvirajo iz primarne proizvodnje aluminija
(znani tudi kot SPL). Predstavljene so najnovej{e tehnike za predelavo SPL s poudarkom na tehnologiji flotacije in kemi~ne
obdelave. Laboratorijski testi so potrdili, da model predelave SPL ne bi bil izvedljiv z delnimi re{itvami. Sistemati~no dose`ena
re{itev lahko ustreza le {irokemu obsegu sodobnih zahtev. Delno je vzrok alotropna oblika proizvedenih ogljikovih katod in
delno tudi ne~isto~e v odpadnem materialu, ki onemogo~ajo u~inkovito enofazno obdelavo. Laboratorijske preiskave reaktivne
ekstrakcije v veliki meri odpirajo mo`nosti za materialno izrabo SPL. Preliminarna poizvedba potrjuje, da interdisciplinaren
pristop vodi k odprtim re{itvam, ki omogo~ajo tr`ni interes oziroma ustvarjeno vrednost iz odpadka, kot glavni cilj kro`nega
gospodarstva.
Klju~ne besede: izrabljene katodne obloge, obdelava odpadkov, flotacija, alkalno kisla obdelava
1 INTRODUCTION
Throughout the history of aluminium production, the
industry has searched for a feasible and environmentally
friendly solution for the treatment of spent potlining
(SPL) as a waste product of the primary aluminium
production process. SPL represents a basic part, i.e., the
pot where melt of electrolyte and molten aluminium are
kept during the electrolysis process, where the pot can
last from 6 to 9 years. An SPL viable treatment is a
global challenge for the primary aluminium industry, as
is declared as a mixed hazardous waste that contains a
first cut as mainly carbonaceous and a second cut as
mainly refractories materials.
1.1 Origins and challenges of SPL
At the end of the electrolysis cell life, the allowed
iron and silicon concentration limits in the electrolyte are
exceeded and the bottom part of the cell must be
renewed. The SPL bottom part is lifted and transported
to a dedicated concrete pit. In the disassembly process
five major parts appear: the cryolite cover, the aluminium
layer, the carbon conductor layer, the functional bricks
and the metal frame. A major part of the cryolite is taken
back to the process. The metal frame is renewed and new
cathode blocks with chamotte and insulation bricks are
lined up to form a new cathode. The spent carbon
cathode and the functional bricks part remain as a waste
product (Figure 1). The industrial naming for the carbon
waste part is "first cut" as this part is the first to be
stamped off. Analogically, after the first cut is collected,
the bricks are on the turn as the so-called second cut. All
the parts or cuts are toxic due to the impregnation with
molten electrolyte during the operating period, con-
taining leachable fluoride salts and cyanides.
The typical composition of the SPL waste is given in
Table 1. However, it needs to be emphasized that these
values can vary by up to 10 mass %.
The main challenges that the innovative solution is
responding to are the following:
Materiali in tehnologije / Materials and technology 53 (2019) 2, 277–284 277
UDK 620.1:669.712:628.54 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 53(2)277(2019)
*Corresponding author e-mail:
blaztropenauer@gmail.com

• A large volume of waste (problem of waste treat-
ment, landfilling) that needs special storage and
transportation requirements before landfilling and
incineration
• The explosive and fire-hazardous nature of the SPL
due to the potential development of explosive gases
(ammonia and C
2
hydro-carbon gases) and the high
burning potential (difficulty in extinguishing the fire)
• The emission of dust particles due to the high fra-
gility of the waste material, which contains hazardous
particles (cyanides, fluorides, etc.)
• The treatment of the hazardous leachates due to the
high level of hazardous soluble materials (approx.
25%( w/%) fluorides and approx.1%c yanides)
• The development of hazardous smoke in the case of
the fire
• The corrosion properties of SPL due to the high pH
(up to 13) due to the alkali metals and their oxides.
• The low social acceptance
1.2 Treatment methods summary
In principle, we can divide the SPL processing tech-
nologies into mechanical/physical, thermal and chemi-
cal. Mechanical refers to the separation of different
fractions of waste, grinding, screening and other mecha-
nical separation. For separation, specific properties of the
material, such as solubility, density, chemical structure
and surface properties (hydrophobicity-hydrophilicity,
charge of the molecules, etc.) are used. These techniques
are important primarily for the proper pre-preparation of
the input material, thus achieving better efficiency in the
main processes.
The best available technologies for non-ferrous
metals describe many options to treat SPL: the re-use of
SPL in cement production, as a carbonaceous substance
in ironworks, as a secondary raw material (glass wool,
salt slag) and as a substitute fuel.
1
Øye describes both the
thermal and chemical methods of SPL processing, em-
phasizing the challenges of salt separation, which
accounts for as much as 30 % of the total SPL content.
2
Mik{a compares the use of uncleaned SPL in the cement
industry and thermoelectric power plants.
3
In cement plants, the carbon part is used as a sub-
stitute for fuel, while the brick part is added to the
production of clinker.
4
Consequently, there is a notice-
able decrease in the process temperature in the clinker
production mass, which in turn reduces the consumption
of heat per kilogram of clinker produced.
5
In ironworks,
the carbon part is used as a fuel and a reducing agent.
Fluorides increase the flow rate of the slag and lower the
melting point, but at the same time they are also harmful
(act corrosively) to the furnace wall.
6
B. TROPENAUER et al.: SUSTAINABLE WASTE-TREATMENT PROCEDURE FOR THE SPENT POTLINING (SPL) ...
278 Materiali in tehnologije / Materials and technology 53 (2019) 2, 277–284
Table 1: Typical composition (in mass %) of SPL waste; Source: Talum d.d. Slovenia
Amorphous
carbon
Graphite
carbon
Mullite
Al
6Si
2O13
Cryolite
Na
3AlF 6
Sodium
fluoride
NaF
Silicon
dioxide
SiO
2
Cordierite
(K,Mg)
2Al 3
(Si 5AlO
18)
Alumina
Al
2O 3
Fluorite
CaF
2
Diaoyudaoi
te
Na
2Al22O34
Other (Ni,
Al, Mg)
24,7% 19.6 % 19.0 % 12.9 % 9.6 % 4.0 % 3.8 % 2.6 % 2.0 % 1.1 % 0.6 %
44.3 % 55.7 % – Mineral part
Table 2: Amounts of elements (in mass %) after various procedures in the carbon part of the SPL processed waste
Procedure F (%) Na (%) Al (%) Si (%) Ca (%) Fe (%)
Untreated SPL 7.68 5.15 5.22 5.46 0.68 0.55
Water extraction (2x) 4.85 2.41 5.48 5.88 0.81 0.62
Flotation (2xEkof) 5.78 1.87 4.44 0.71 1.24 0.25
Alkaline (FC) 2.70 2.56 2.11 0.94 2.46 0.27
Acidic (FC) 1.00 1.90 0.99 0.87 0.29 0.07
Figure 1: Typical repository of SPL waste in a warehouse: a) first-cut
carbon part, b) the second cut

Chemical methods allow the separation of compo-
nents and their purification in order to produce products
that could be used either in their own processes or in
other local industries. Some methods are based on the
neutralization of the filtrate with calcium compounds
(e.g., calcite CaCO
3
).
7
Others focus on reuse and mostly
predominate in the full exploitation of the raw-material
base. They tend to extract cryolite and carbon
8,9
and
useful fluoride compounds
10,11
by dissolving in bases,
acids or in solutions of aluminium salts. Some separation
technologies use a two-step technique. First, they use
aqueous purification to separate the soluble salts and
then use compounds (HF, Al(NO
3
)
3
, HNO
3
,H
2
SO
4
)t o
form aluminium cations, which have a high affinity for
fluorides to form stable and soluble fluoro-aluminium
complexes with the purpose being to dissolve insoluble
salts and form more useful types of filtrate com-
pounds.
10–13
Other techniques use three-stage separation,
i.e., water purification, alkaline or acidic cleaning.
8,14,15
Schönfelder reports on a flotation technique to
separate carbon and silicate parts.
16
The SPL waste is
firstly exposed to sulphuric acid. In this process, the
waste detoxifies, and at the same time H
2
SiF
6
acid forms
as a product. The detoxified SPL is then subjected to a
flotation process where products such as carbon and
silicates are obtained. The obtained H
2
SiF
6
acid, in
combination with the addition of Al(OH)
3
, produces the
desired product of AlF
3
, which is used as a valuable
reagent for the electrolysis process in aluminium produc-
tion. Flotation was also reported to be a technology
option in Chinese aluminium plants.
16,17
2 EXPERIMENTAL PART
The main target of our research was focused on
solving the old SPL waste problem, which requires the
additional separation of carbon from the mineral (brick)
part. The next goal is to create technologies for SPL
treatment for the already-separated carbon and brick
parts.
The proposed method does not include additional
energy for reaction purposes. It is performed under stan-
dard ambient temperature and pressure conditions
(SATP), i.e., 25 °C, 298.15 K and 101.353 kPa.
The following four research methods were chosen for
the experimental work:
• Mechanical/physical treatment of the sample
• Extraction of soluble salts
• Flotation separation process
• Alkaline and acidic chemical reactive leaching
The particle size of the material is important as it
increases the surface area of the waste and thereby
improves the hydrophilicity and promotes the leaching
reactivity. For good flotation results, the size of the
particles should range between 20 μm and 300 μm.
15
Graphite and sodium in the spent cathode make the spent
cathode sticky, slippery and thus difficult to crush.
18
The
effective crushing process for a proposed plant and its
economics are therefore of significant importance, so as
to avoid jams and damage to equipment in the future.
In the next step, the water-extraction process ensures
the dissolution of all the water-soluble salts (mostly
NaF). They account for as much as 30 % of the sample
mass. The resulting filtrate is detoxified with one of the
methods for cyanide destruction.
19
The clean slurry of the mixed carbon and bricks part
is further subjected to flotation. The flotation reagents
are oily and surface-active substances, which accumulate
on the surface of the carbon and increase the hydropho-
bicity. The hydrophobic particles attach to the bubbles
and are discharged as a foam.
15
After the first-stage of the flotation slurry, the carbon
part was subjected to a further alkali reactive extraction
process, in order to extract the insoluble salts from the
sample.
In the last step, the acid-reactive extraction process
dissolved the insoluble compounds like CaF
2
and
NaAl
11
O
17
, where the carbon content (C
fix
) rose to the
desired level.
The scheme of the described procedure is presented
in Figure 2.
Separately, a test to evaluate the extraction effective-
ness with HCl and AlF
3
solvents was performed.
The sample was extracted with the standard method
(SIST EN 1247-4) where the ratio of liquid to solids
(L:S) is kept in the range of 10:1.
2.1 Analytical support
All the samples for analysis were milled in Retsch
MM 400 device. The resulting material for the flotation
experiment weighed approx. 500 g with a 30-μm average
particle size. The sample was dried at 120 °C for 24 h to
measure the carbon content using the standard ASTM
3172/73 (800 °C for 5 h).
Qualitative and quantitative analysis with the Riet-
veld method. To obtain the mass fractions of the com-
pounds in the samples, the following method was used.
The samples were analysed using an X’Pert PRO x-ray
powder diffractometer with Cu–K
 1
radiation in the 2 B. TROPENAUER et al.: SUSTAINABLE WASTE-TREATMENT PROCEDURE FOR THE SPENT POTLINING (SPL) ...
Materiali in tehnologije / Materials and technology 53 (2019) 2, 277–284 279
Figure 2: Schematic presentation of the described experimental
procedure

range from 5° to 90° and a 2 step of 0.017°. Based on a
qualitative analysis of the X-ray powder diffractograms
with X’pert High Score Plus, the presence of several
crystalline phases was identified, including phases in
trace amounts. Through a quantitative analysis of the
XRD diffractograms with the Rietveld method (TOPAS
program), the mass proportions of the identified
crystalline phases in the samples were determined. The
estimated error of the calculation ranges from3%t o
10 % for the phases with amounts below 5 %.
Scanning electron microscopy (SEM). Samples
(approximately 5 mg) in the as-obtained form were
positioned on aluminium carriers equipped with a carbon
adhesive tape, inserted into the Zeiss SEM microscope
chamber and analysed at various magnifications and
acceleration voltages with three different detectors
(inlens, SE2, EDS).
X-ray fluorescence (XRF) was used for the quanti-
tative element analysis. The samples were measured on a
Thermo ARL Advant’X XRF with a copper tube
operated at 50 kV and 60 mA. Standardless data analysis
software Thermo Scientific Uniquant (v5,49; 2008) was
used for a semi-quantitative content assessment of the
elements (the carbon content, determined with a TGA,
was calculated for a more accurate assessment).
X-ray diffraction (XRD) was used for solid crystal
phase identification using a Panalytical Cubix3 X-Ray
Diffractometer equipped with an X’Celerator detector.
The copper tube was operated at 50 kV and 40 mA. The
conditions used were 60 min
–1
rotation, 10–80° 2-theta,
0.011° increment, 3 s/step time per step, 3.01 min scan
time, with a 0.5 deg fixed divergence slit. The resulting
diffractograms were processed using the Panalytical
HighScore+ 4.0 (2013). This software compared the
measured XRD pattern with the International Centre for
Diffraction Data (ICDD) database PDF-2 (2013) to find
matches with known crystalline compounds.
Carbon content (C
fix
) was measured on a Mettler
Toledo TGA1 with STARe software. The sample was
first heated to 950 °C under inert gas for a determination
of the volatiles and then burnt in the air to a constant
weight. C
fix
was calculated from the amounts of ash and
volatiles.
2.2 Samples preparation
The experiments started with familiarization of the
material. Many carbon-content and elemental analyses
were carried in our laboratory in the past. The sample
compositions were acquired with regards to the position
in the cell (corner, side and bottom).
The samples were gathered from the Talum d.d. alu-
minium production plant. Large pieces of material were
taken from the first and second cuts, in order to properly
quantity the mass distribution in the cathode of the
electrolytic cell (first cut 60 %, chamotte bricks 30 %
and isolating bricks 10 %). The chosen bulk samples
were not contaminated with cryolite crust. In the process
of crushing, some 70 % of the sample mass was obtained
by sieving. The sample was crushed in a metal moon,
then in a jaw crusher and finally manually in a mortar.
The sample was mixed in the mixing drum in order to
achieve homogeneity. The resulting material for the
flotation experiment had a weight of 500 g and an
average particle size of 30 μm.
All the following extractions were carried out using
the same methods. An L:S ratio of 10:1 was used for the
extractions, the samples were stirred at a speed of
500 min
–1
and at the end of the experimental time they
were dried for 24 h in order to measure the carbon
content. The sample was filtered with a blue-ribbon filter
paper and dried for 24 h at 120 °C to determine the
residual mass.
2.3 Extraction of soluble salts
As the SPL material is impregnated with soluble
salts, a leaching procedure was performed first, since the
ionized salts can disturb the effect of the flotation.
Technological water was chosen as a solvent to imitate
the fluoride and cyanide leaching in a real industry
situation. The leaching procedure was performed for a
maximum time of 24 h, with the length of the individual
steps after 10 min in the first hour, and then after hourly
readings of the sample values.
2.4 Froth flotation
Froth flotation is a process for the selective separa-
tion of hydrophobic and hydrophilic materials. The
flotation is normally performed in several stages to
maximize the recovery of target minerals and the con-
centration of those minerals in the concentrate, while
minimizing the energy input. In this work the flotation
was carried out in a Denver flotation cell with two diffe-
rent agents.
The sample used in the froth flotation contained an
average mixture of brick and carbon. It was washed
twice and placed in a 1500-mL metal container of the
flotation cell.
Flotation test with collector and froth reagent
For the first flotation experiment, a collector agent
was chosen (sodium ethyl xanthate (C
3
H
5
NaOS
2
)), used
in mining as a collector for the enrichment of pyrite from
quartz and galenite from sphalerite.
In 1000 mL of deionized water (L=liquid) 20 g of the
sample (S=sample) was added (S:L ratio = 1:50). With a
pipette, 6 droplets of collector agent were added (one
droplet for approx.3gofsample). In the flotation cell,
the slurry mixture (more properly called the pulp)w a s
stirred for approx. 10 min (Figure 3).
Then the froth reagent (sodium lauryl sulphate, SLS)
was added. When foaming, the resulting bubbles/froth
were instantly pulled into a separate dish. In the froth
flotation, hydrophobic particles flow to the top and sepa-
B. TROPENAUER et al.: SUSTAINABLE WASTE-TREATMENT PROCEDURE FOR THE SPENT POTLINING (SPL) ...
280 Materiali in tehnologije / Materials and technology 53 (2019) 2, 277–284

rate from the chamotte heavy particles. The process was
repeated until the bubbles of the foam were transparent
and lasted for approx. 1 h. Samples from the metal con-
tainer (carbon part and the remaining liquid sample)
were stored separately, filtered, dried for 24 h and ana-
lysed for carbon content and for elements with an XRF
instrument.
Flotation test with an Ekofol reagent
To further raise the carbon content of the remaining
carbon part, a second flotation test was performed. The
procedure was repeated with the difference that this time
more sample was processed (180 g). In addition, instead
of separate by used collector and froth reagents, the
Ekofol 452 GK100 collector/froth agent was used, as
performed by Schönfelder.
15
The process lasted 42 min
and 14 droplets of the agent were used. The sample was
again separated on the silicate part on the bottom and the
carbon part in the froth (Figure 3). After double drying,
the two samples were subjected to a carbon-content and
an XRF elemental analysis.
2.5 Alkaline reactive extraction
Afterwards, the test with the sample obtained from
the flotation tests with the Ekofol reagent followed. The
water washed SPL material contained almost no soluble
salts and no silicates (bricks part); therefore, no flotation
was needed anymore.
Finally, the alkali reactive leaching procedure was
performed in an 1-M NaOH solution for 3 h with a L:S
ratio of 10:1. Subsequently, the samples were filtered
and dried to define the carbon content.
Experiments were repeated with the "first cut" sam-
ples of SPL with rather a similar compound structure to
further analyse the elementary and compound compo-
sition.
2.6 Acidic reactive extraction
To reduce the remaining calcium content and further
increase the carbon content an acidic reactive extraction
was performed. The remaining first cut alkali-extracted
sample was washed with distilled water to avoid the
alkali species contamination. Then the sample was
filtered, dried and stirred at a speed of 500 min
–1
in a
2.7-M HCl solution for3hinanL:Sratio of 10:1. The
samples were again filtered, two-times water washed to
avoid contamination, filtered and analysed.
Using a scanning electron microscope (SEM)
analysis of the samples the morphology was determined.
An additional acidic test was carried out to compare
the efficiency of the different types of acids. The first cut
sample was subjected to the above-described water and
alkali extraction. The alkali-leached sample was exposed
to a 0.74-M AlCl
3
solution, in order to compare its
effectiveness with respect to HCl. Finally, both samples
were water washed to avoid acidic contamination and
analysed.
3 RESULTS
Water extraction tests
I na1h(time) period, almost all the soluble salts
(mostly NaF) were extracted, representing about 20 to 30
mass % of the entire sample. An additional 24 h of
extraction did not increase the extracted mass signifi-
cantly. In one leaching step, about 40 to 50 mass % of
fluorides and cyanides were extracted. Repeated extrac-
tion removed the residual soluble components (about 4
mass % of the whole soluble salts) and decreased the pH.
The XRF elemental analysis showed the greatest reduc-
tion in elements like sodium and fluoride (Table 2).
Flotation results
Double-dried samples after froth flotation with the
collector and froth reagent were subjected to a car-
bon-content and an XRF elemental analysis. The carbon
content of the silicate (brick) part was 4.06 mass % with
the main compounds quartz (SiO
2
) and mullite
(Al
4
Si
6
O
10
). The carbon part contains 70 mass % of
carbon and other compounds, merely quartz (SiO
2
) with
some minor content of cryolite (Na
3
AlF
6
) and calcium
B. TROPENAUER et al.: SUSTAINABLE WASTE-TREATMENT PROCEDURE FOR THE SPENT POTLINING (SPL) ...
Materiali in tehnologije / Materials and technology 53 (2019) 2, 277–284 281
Figure 3: Denver flotation cell in operation (on the left) and sample
flotation process with good adherence of carbon in the bubble-froth
(on the right)

fluoride (CaF
2
). The results are presented in Figure 4
(picture of the separated sample) and Figure 5 (diffracto-
gram measurement results).
The reagent Ekofol 452 GK100, on the other hand,
proved to be very effective because of the significantly
better adherence of carbon (graphite and anthracite) to
the surface of the bubbles. The Ekofol reagent has an
about 4-times lower consumption and a 13-times shorter
process time with respect to the collector and froth
reagent.
The gained carbon part weighted 71 g and was
subjected to a second-step flotation with the same proce-
dure, using the Ekofol reagent. The process lasted
35 min and 4 Ekofol droplets were used. The second-
step flotation did not noticeably increase the carbon
content. The problematic compounds are the following:
cryolite (Na
3
AlF
6
), calcium fluoride (CaF
2
), gibbsite
(Al(OH)
3
) and diaoyudaoite (NaAl
11
O
17
). The XRF
analysis gives an elemental content of problematic
species masses as follows: fluorine by 6 %, sodium by
2 % and calcium by1%( Table 2).
Alkaline extraction tests
The carbon-content analysis (C
fix
) after the alkaline
reactive extraction procedure shows a significant rise
from 72 % to 82 %. The diffractogram analysis of the
first-cut sample showed that with NaOH reactive extrac-
tion, removal of most of the cryolite in the waste sample
was achieved (Figure 5). In addition, an elemental
analysis made with the XRF clearly shows that elements
like F, Na, and Al, that constitute cryolite, did reduce
significantly. On the other hand, the calcium fraction
increased (Table 2).
B. TROPENAUER et al.: SUSTAINABLE WASTE-TREATMENT PROCEDURE FOR THE SPENT POTLINING (SPL) ...
282 Materiali in tehnologije / Materials and technology 53 (2019) 2, 277–284
Figure 4: Obtained carbon part (left) and brick part (right) after water
wash and flotation process
Figure 5: XRD diffractograms of an effective cryolite extraction:
pixels: 29, 32, 32.5 (left picture) and 57.8, 59.6 (right picture) have a
lower value on the red line (alkaline extraction step). Crystallographic
data from ICCD: carbon 00-008-415, CaF2 03-065-0535, cryolite
00-025-0772
Figure 6: SEM analysis of water washed (left) and acid extracted
sample (right) of the first cut

Acidic extraction tests
The results after acidic extraction showed a signifi-
cant reduction of the calcium content. The sample com-
position obtained by qualitative and quantitative analyses
with the Rietveld analysis method is as follows (w/%):
3.7 % NaF, 3.5 % aluminium oxide hydrate, 1.4 % CaF
2
and 0.7 % Sodium aluminium oxide (NaAl
11
O
17
). The
remainder was 90.7 % of graphite. The cryolite was
removed completely, while the calcium fluoride and
sodium aluminium oxide were significantly reduced.
SEM results (Figure 6) showed that there were
almost no impurities left on the surface of the cleaned
sample (after the acidic leaching step). There were very
little traces of spherical and needle-shaped impurities in
comparison to the first-cut sample, which was water
washed only.
HCl and AlCl
3
efficiency comparison
The carbon-content results in the HCl and AlCl
3
efficiency comparison test showed approximately the
same values (C
fix
around 90 mass %). In addition, the
XRD diffractogram showed little changes in the species
left in the sample. Both results also showed a significant
drop in the CaF
2
species.
4 DISCUSSION
A four-step treatment of SPL waste has been demon-
strated as a suitable approach to obtain commercial
products in the form of a high content of carbon and the
silicate part. Water extraction is the most effective and
simplest process to dissolve the soluble NaF impurities.
These are part of the filtrate, which transforms to CaF
2
and NaOH when exposed to lime (Equation 1). Both are
desirable products in the metallurgical industry.
2NaF + Ca(OH)
2
= 2NaOH + CaF
2
(1)
A comparison between different flotation reagents in
the same conditions shows a remarkable efficiency gap
between the two reagents. After the second flotation the
silicon-based components like SiO
2
are significantly
decreased as the insulation and firebrick parts are separa-
ted (Table 2). The froth-flotation efficiency is deter-
mined by a series of parameters: particle-bubble contact,
particle-bubble attachment, transport between the pulp
and the froth, and froth collection into the product
launder. In a conventional mechanically agitated cell
(Denver flotation cell), the void fraction (i.e., the volume
occupied by the air bubbles) is low (5 to 10) % and the
bubble size is usually greater than 1 mm. This results in
a relatively low interfacial area and a low probability of
particle-bubble contact. Consequently, several cells in
series are required to increase the particle residence time,
thus increasing the probability of particle-bubble contact
and also improving the froth flotation efficiency.
The alkali extraction was successfully performed as
NaOH was effectively decomposing the sparingly sol-
uble cryolite salt. An additional reactive extraction with
acid decomposes the calcium fluoride and sodium alumi-
nium oxide and raises the carbon content of the sample
up to 90 % (Figure 7). The SEM images (Figure 6)
show a clear difference in the quantity of the impurities
on the surface of the treated and untreated samples.
The results showed that there is no significant
difference between the acids HCL and AlCl
3
extraction
as the AlCl
3
is changed to HCl. Since AlCl
3
is the more
expensive compound, it is economically more appro-
priate to use HCl in the final extraction step.
5 CONCLUSIONS
The goal to produce purified carbon and silicate
fractions from the SPL waste was achieved. The flotation
results confirmed the optimal reagent choice as the
separation of the silicate and carbon part was in both
cases successful. The brick part is a usable product in
firebrick production as it contains a reasonable part of
mullite, alumina and silicates. The carbon part, on the
other hand, has a high carbon content and is, therefore, a
desirable product in the metallurgical industry.
Reactive extraction is proven to be an indispensable
part of the process as higher purities of carbon were
demanded. The by-product, i.e., the filtrate from water
and alkali extraction, contains predominantly NaF, which
can be transformed to usable metallurgy compounds.
A former dangerous waste is now transformed into a
safe and useful product at the plant site.
Acknowledgements
The authors would like to express their sincere
gratitude to the University of Maribor, Faculty of Mecha-
nical Engineering for the support and managing to unite
interdisciplinary professions, and the University of
Ljubljana, Faculty of Natural Sciences and Engineering,
for providing the flotation tests and Talum d.d. (JSC)
Kidri~evo for providing the SPL samples and also for
financial support for the analytics.
6 REFERENCES
1
BREF Non – Ferrous Metals Industries, EC directorate-general JRC,
Final Draft 10, NFM, 2014
B. TROPENAUER et al.: SUSTAINABLE WASTE-TREATMENT PROCEDURE FOR THE SPENT POTLINING (SPL) ...
Materiali in tehnologije / Materials and technology 53 (2019) 2, 277–284 283
Figure 7: Fixed carbon (C
fix
) values in different treatment solutions

2
J. Hop, A. Støre, T. Foosnæs, H. A. Øye, Chemical and physical
changes of cathode carbon by aluminium electrolysis, VII
International Conference on Molten Slags Fluxes and Salts, The
South African Institute of Mining and Metallurgy, 2004, 775–782
3
D. Mik{a, M. Hom{ak, N. Samec, Spent potlining utilisation possibi-
lities. Waste Manag. Res. 21 (2003) 467–473, doi:10.1177/
0734242X0302100509
4
R. P. Pawlek, Spent potlining: An update, Light Metals (2012),
1313–1317, doi:10.1007/978-3-319-48179-1_227
5
M. L. G. Renó, F. M. Torres, R. J. da Silva, J. J. C. S. Santos, M. de
Lourdes Noronha Motta Melo, Energy analyses in cement production
applying waste fuel and mineralizer, Energy Conversion and
Management (2013), doi:10.1016/j.enconman. 2013.05.043
6
B. Meirelles, H. Santos, Economic and environmental alternative for
destination of spent pot lining from primary aluminum production,
Light Metals (2014), 565–570
7
D. B. Turner, P. J. Binning, S. W. Sloan, A calcite permeable reactive
barrier for the remediation of Fluoride from spent potliner (SPL)
contaminated groundwater, Journal of Contaminant Hydrology 95
(2008), 110–120
8
Z. Shi, W. Li, X. Hu, B. Ren, B. Gao, Z. Wang, Recovery of carbon
and cryolite from spent pot lining of aluminium reduction cells by
chemical leaching. Trans. Nonferrous Met. Soc. China, 22 (2012),
222–227, doi:10.1016/S1003-6326(11)61164-3
9
X. Cao, Y. Shi, Z. Shung, X. Xue, Recovery of valuable components
from spent pot lining of aluminium electrolytic reduction cells,
Journal of Northeastern University (Natural Science), 35 (2014), 12
10
L. D. F. Lisbona, K. M. Steel, Recovery of fluoride values from spent
pot-lining: Precipitation of an aluminium hydroxyfluoride hydrate
product, Separation and Purification Technology, 61 (2008), 182–192
11
N. Ubong, S. Tait, E. T. White, K. M. Steel, The precipitation and
solubility of aluminium hydroxyfluoride hydrate between 30 and
70 °C, Hydrometallurgy, 155 (2015), 79–87
12
T. K. Pong, R. J. Adrien, J. Besida, T. A. O’Donnell, D. G. Wood,
Spent potlining – A hazardous waste made safe, Trans IChemE, Vol
78, Part B, 2000, 204–208
13
D. F. Lisbona, C. Somerfield, K. M. Steel, Leaching of spent
pot-lining with aluminum anodizing wastewaters: Fluoride extraction
and thermodynamic modelling of aqueous speciation. Ind. Eng.
Chem. Res., 51 (2012), 8366–8377, doi:10.1021/ie3006353
14
D. Pankaj Indurkar, Optimization in the treatment of spent pot lining
– A hazardous waste made safe, M. Tech Thesis, Chemical
Engineering at National Institute of Technology, 2014
15
I. Schönfelder, E. Gock, E. Tochtrop, S. Rieder, n.d., Recycling von
Elektrolyseausbruch (SPL) aus der Primäraluminiumgewinnung
2014, 385–404
16
W. Li, X. Chen, Development status of processing technology for
spent potlining in China, Light Metals, (2010), 1064–1066
17
G. Holywell, R. Breault, An overview of useful methods to treat,
recover, or recycle spent potlining, JOM, 65 (2013), 1441–1451
18
X. Chen, W. Li, Research on crushing character of spent cathode,
Light Metals, 2007, 853–855
19
M. Cen~i~, I. Kobal, J. Golob, Thermal hydrolysis of cyanides in
spent pot lining of aluminium electrolysis, Chem. Eng. Technol., 21
(1998), 523–532, doi:10.1002/(SICI)1521-4125(199806)21:6<523::
AID-CEAT523>3.0.CO;2-P
B. TROPENAUER et al.: SUSTAINABLE WASTE-TREATMENT PROCEDURE FOR THE SPENT POTLINING (SPL) ...
284 Materiali in tehnologije / Materials and technology 53 (2019) 2, 277–284