215
 Open Access. © 2019 Janc B., Hann D., published by Sciendo.  This work is licensed under the Creative Commons 
Attribution-NonCommercial-NoDerivatives 4.0 License. 
Received: Jan 27, 2020 
Accepted: Jan 29, 2020
 DOI: 10.2478/rmzmag-2019-0020
Original scientific paper 
Abstract
In this paper, research on the possibilities of sodium 
sulphate (Na
2
SO
4
) separation from other substances in 
the filter ash sample is presented. The research materi-
al contains six components that differ in chemical com-
position and density. The possibilities of Na
2
SO
4
 sepa-
ration using dry and wet methods were studied. The 
dry method was based on separation with a centrifugal 
air classifier at four cut size limits. The wet method was 
based on the dissolution of water-soluble components, 
filtration of insoluble components, and drying the 
products. The sulphur content of the individual prod-
ucts was determined using both methods. The aim of 
the research was to determine which method is more 
suitable for separation of the material in a way that 
most of the material would contain as little sulphur as 
possible and the rest of the material would contain con-
centrated sulphur. The wet method proved to be more 
successful. The product with mass fraction 33.1% of 
the total mass, obtained from the aqueous solution, 
contained 8.39% sulphur after filtration and drying. 
The water-insoluble component, with mass fraction 
66.9% of the total mass, contained 0.56% sulphur. The 
dry method with the centrifugal air classifier proved to 
be less successful in comparison with the wet method. 
The particles containing Na
2
SO
4
 are very similar in size 
and density to the other components of the material, so 
the separation to the desired extent was not achieved.
Keywords: sulphur, centrifugal air classifier, filtration, 
filter ash, waste material.
Removal of Na
2
SO
4
 from a Filter Ash
Odstranjevanje Na
2
SO
4
 iz filtrskega prahu
Blaž Janc*, Damjan Hann
University of Ljubljana, Faculty of Natural Sciences and Engineering, Department of Geotechnology, Mining and Environment, 
Aškerčeva 12, Ljubljana, Slovenia
* blaz.janc@ntf.uni-lj.si
Povzetek
V članku je predstavljena raziskava možnosti ločenja 
Na
2
SO
4
 od ostalih snovi v vzorcu filtrskega prahu. Prei-
skovan material vsebuje šest komponent, ki se med se-
boj razlikujejo po kemijski sestavi in gostoti. Preučevali 
smo možnost ločenja Na
2
SO
4
 s suhim in mokrim po-
stopkom. Suh postopek je temeljil na ločevanju s cen-
trifugalnim zračnim klasifikatorjem pri štirih mejah lo-
čenja. Mokri postopek je temeljil na raztapljanju v vodi 
topnih komponent, filtraciji netopnih komponent in 
sušenju produktov. Po obeh izvedenih metodah smo 
določili vsebnost žvepla v posameznih produktih. Cilj 
raziskave je bilo ugotavljanje, s katero metodo doseči 
ločenje materiala tako, da bi večina materiala vsebova-
la čim manj žvepla, preostanek materiala pa bi vseboval 
koncentrirano žveplo. Kot bolj uspešna se je pokazala 
metoda z mokrim postopkom. Produkt z masnim de-
ležem 33.1% od celote, pridobljen iz vodne raztopine, 
je po filtraciji in sušenju vseboval 8.39% žvepla. V vodi 
netopna frakcija, katere delež je bil 66.9% od celote, je 
vsebovala 0.56% žvepla. Suhi postopek z zračnim klasi-
fikatorjem se je izkazal kot manj uspešen. Delci Na
2
SO
4
 
so namreč po velikosti in gostoti dokaj podobni ostalim 
komponentam materiala, zato nismo dosegli ločenja 
v želenem obsegu.
Ključne besede: žveplo, centrifugalni zračni klasifika-
tor, filtracija, filtrski prah, odpadne snovi.

Janc B., Hann D. RMZ – M&G | 2020 | Vol. 66 | pp. 215–222
216
Introduction
The construction industry in Slovenia is cur-
rently in expansion. Major repairs, reconstruc-
tions and renovations of residential, public and 
industrial buildings are underway. In connec-
tion with the increased number of buildings 
under construction, the production of con-
struction materials is also increasing. Due to 
the demolition and reconstruction of buildings, 
the amount of construction waste is increasing, 
most of which have the potential of secondary 
use [1]. In the circular economy concept, waste 
is considered as raw material and could be re-
used in the production process. This avoids 
waste disposal and the associated environmen-
tal problems and reduces the need for new raw 
materials, in which sources are limited. The 
composition of construction waste has changed 
over the years. The construction waste whose 
use has spread in recent decades is increasing. 
Dry prefabricated materials such as gypsum 
boards and insulating materials which are the 
result of the energy efficiency concept and the 
introduction of new building materials and con-
struction methods are included in this group.
In the European Union, 40.9 million tons of 
waste materials were generated in the field 
of construction and demolition in 2016, rep-
resenting 36.4% of all waste, compared with 
10.3% of industrial waste. The construction 
sector is thus the largest producer of waste in 
comparison with other economic sectors [2].
The European Commission Waste Directive 
also addresses waste hierarchy, which makes 
prevention of the top waste management prior-
ity, followed by preparation for reuse, recycling, 
recovery and landfill [3]. As a result, industrial 
companies are looking for new approaches to 
the waste management which are consistent 
with the environmental, social and economic 
sustainability [4].
As with most production processes, the pro-
duction of building materials, whether using 
exclusively primary raw materials or adding 
secondary raw materials into the process, 
waste materials are produced as a by-product. 
Their quantity and composition depend on the 
type of technological process.
In the case of thermal treatment of materials, 
flue gases are mainly produced as by-products, 
carrying dusty ash particles with them. Mass 
fraction of heat treatment by-products may be 
significant in the intense production process. 
Products are frequently waste materials but 
can represent the potential for reuse in case of 
providing technological and environmental re-
quirements. Some of these materials have one 
or more hazardous properties and may repre-
sent an environmental risk. If the material is 
characterised as hazardous waste, it represents 
a much higher cost to the waste producer or the 
payer to properly treat it than if the waste were 
considered to be a non-hazardous waste. Such 
materials may be prepared by mechanical and/
or other processes to the extent that they are 
environmental friendly and reusable.
This study presents some methods for minimis-
ing sulphur content with mechanical processes 
in handling thermal treatment products, more 
specifically when handling filter ash containing 
a certain content of sulphur. By minimising the 
sulphur content, it is possible to achieve the 
removal of hazardous properties from waste 
which transforms waste from hazardous to 
non-hazardous type, thereby opening the pos-
sibilities of reuse or at least significantly reduc-
ing the cost of waste disposal.
Materials and Methods
The filter ash that is a subject of this research is 
a mixture of:
 ― Sodium bicarbonate (NaHCO
3
),
 ― Sodium sulphate (Na
2
SO
4
),
 ― Coke,
 ― Limestone,
 ― Basalt (rock wool fibres),
 ― Ash.
In order to minimise the sulphur content of the 
filter ash, the aim of the research was to sep-
arate Na
2
SO
4
 from the other components or 
to concentrate sodium sulphate. One option 
was to perform the dry process separation us-
ing a centrifugal air classifier. Considering the 
fact that the research deals with the particu-
late matter, which was extracted from the flue 
gas by filters, it is logical that the maximum 
particle size in the sample taken was limited 
to about 50 μm. If the particles of the Na
2
SO
4
 

Removal of Na2SO4 from a Filter Ash
217
component were present in a specific granu-
lometric interval, there exists a possibility of 
separation with the centrifugal air classifier. Al-
ternatively, a wet process could be performed 
in which the water-soluble Na
2
SO
4
 is separated 
from the other components by dissolving, fil-
tration of insoluble components and drying the 
products.
A centrifugal air classifier is a device for sep-
arating particles according to their size and 
density. During the classification operation, 
the particles are under the influence of centrif-
ugal force (F
c
), drag force (F
d
) and gravity (F
g
) 
force. The centrifugal force field is generated 
by a rotor, which accelerates the particles to-
wards the periphery of the classification zone 
where a coarse fraction is deposited. The air 
enters the classification zone tangentially, flows 
through the middle of the classification zone 
and removes finer particles on which the drag 
force of the air resistance is greater than the 
centrifugal force. The fine particles carried by 
the air stream are separated from the gas phase 
in the air cyclone. The material enters the clas-
sifier using a uniform dosing mechanism (dos-
ing screw). The device principle is explained in 
Figure 1.
The definitions of the forces acting on particles 
are as follows [5]:
𝐹𝐹𝐹𝐹 𝑐𝑐𝑐𝑐 =
4
3
⋅ π ⋅ 𝑟𝑟𝑟𝑟 𝑝𝑝𝑝𝑝 3
⋅ 𝜌𝜌𝜌𝜌 𝑝𝑝𝑝𝑝 ⋅
𝑣𝑣𝑣𝑣 2
𝑟𝑟𝑟𝑟  (1)
𝐹𝐹𝐹𝐹 𝑑𝑑𝑑𝑑 = 𝑐𝑐𝑐𝑐 𝐷𝐷𝐷𝐷 ⋅ 𝜌𝜌𝜌𝜌 ⋅ π ⋅ 𝑟𝑟𝑟𝑟 𝑝𝑝𝑝𝑝 2
⋅
𝑣𝑣𝑣𝑣 𝑎𝑎𝑎𝑎 2
𝑎𝑎𝑎𝑎  (2)
𝐹𝐹 𝑔𝑔 = 𝑚𝑚 ⋅ ( 𝜌𝜌 𝑝𝑝 − 𝜌𝜌 ) ⋅ 𝑔𝑔 
 (3)
where r
p
 is particle radius, ρ
p
 is particle density, 
v is the peripheral velocity of the rotor , r is rotor 
radius, c
D
 is drag coefficient, ρ is the air density, 
m is mass of particle and g is gravitational con-
stant.
The cut size of the classifier is the particle size 
limit at which the material is separated into 
a fine and coarse fractions. It depends on the 
rotor revolutions per minute (RPM), which cre-
ates the centrifugal field, and on the air volume 
flow through the classification zone. The re-
quired rotor RPM and the airflow for the select-
ed cut size are determined from the diagram 
provided by the classifier manufacturer.
For the purpose of determining the operating 
conditions of the centrifugal air classifier, the 
sample density was measured using a pycnom-
eter method. Since two components of the ma-
terial (NaHCO
3
 and Na
2
SO
4
) are water-soluble, 
in order to determine the density, we used iso-
propanol in which those components are insol-
uble or slightly soluble. First, we determined 
the density of isopropanol with a pycnome-
ter (ρ = 784.8 kg/m
3
). Following this, we per-
formed three measurements of sample density, 
which are presented in Table 1.
The calculated mean density of the sample was 
2,521.6 kg/m
3
. In the next step, four different 
cut sizes were selected, namely 10, 20, 30 and 
40 μm. For each cut size, the required rotor 
RPM and airflow quantity in accordance with 
Figure 1. Principle of operation of the centrifugal air classifier.
Table 1. Density measurements with a pycnometer
Test no. Density (kg/m
3
)
1 2,445.8
2 2,579.5
3 2,539.5

Janc B., Hann D. RMZ – M&G | 2020 | Vol. 66 | pp. 215–222
218
the operating diagram of the device were deter-
mined (Table 2).
Na
2
SO
4
 is soluble in water, so it is possible to re-
move it from insoluble components by mixing 
the material in water at the appropriate tem-
perature, and then filtration process is used to 
remove the insoluble residue, which is followed 
by the elimination of water-soluble substances, 
including Na
2
SO
4
, for which drying or a reverse 
osmosis process may be used.
Of the components contained in the input ma-
terial, NaHCO
3
 is also soluble in water, whose 
solubility at 35°C is approximately 120 g/L. The 
solubility of sodium sulphate in the water rises 
to 32.4°C (497 g/L), and decreases slightly at 
higher temperatures [6].
Sampling of Material
The material was sampled with a spoon. It was 
poured on a flat surface; a rectangle 1–2 cm 
high was formed and 5 × 6 even squares were 
drawn into it. Next, we took a spoonful of ma-
terial from each quadrant to get four samples 
with about 250 g each for classification purpos-
es. Sampling is shown schematically in Figure 2. 
The sulphur content of the samples was mea-
sured with an X-ray fluorescence (XRF) spec-
trometer and S = 3.38% for the initial sample.
Results and Discussion
Removal of Sulphur from Filter Dust 
by a Dry Process with Classification
The classification of the sample was performed 
four times, for each cut size individually. After 
each classification, the weight of the material of 
the coarse and fine fraction was determined by 
weighing. The data are presented in Tables 3–6.
Classification of sample 1 at the cut size of 
10 m m yielded 21.8% fine fraction and 78.2% 
coarse fraction. This means that by weight 
21.8% of the particles are smaller than 10 m m 
and 78.2% are larger than 10 m m.
When classifying sample 2 at the cut size of 
20 m m, we found that about half of the particles 
were smaller than 20 m m and half were larger 
than 20 m m.
Classification of sample 3 at the cut size of 
30 m m yielded 75.1% fine and 24.9% coarse 
fraction.
When classifying sample 4 at the cut size of 
40 m m, the share of the fine fraction was 78.3% 
and the share of the coarse fraction was 21.7%.
The mean particle size of Na
2
SO
4
 can be, accord-
ing to the sulphur content results, estimated to 
be around 20 m m. The most favourable result 
with air classification was obtained at a cut size 
of 10 m m, where the sulphur content in the fine 
fraction was 0.55%, but the mass share of the 
fine fraction was only 22.4% at this separation 
size. In all other cases, the sulphur content was 
not minimised.
The particles of the Na
2
SO
4
 component are quite 
dispersed in size and at the same time they are 
not significantly different in density from the 
other components in such a way that they can 
Table 2. Determination of operating conditions for the 
centrifugal air classifier using the operating diagram of the 
classifier
Cut size d
TV
 
(mm)
RPM n  
(min
-1
)
Airflow Q 
(m
3
/h)
10 8,900 46.1
20 5,200 49.8
30 4,100 50.9
40 3,500 51.5
Figure 2. Sampling with a spoon [after 7].

Removal of Na2SO4 from a Filter Ash
219
be successfully separated from the other com-
ponents by the air classifier.
Removal of Sulphur from 
Filter Dust by Wet Process
Approximately 100 g of sample was mixed with 
800 mL of water. The suspension was stirred 
with a magnetic stirrer for 30 min, while it was 
heated to a temperature of about 35°C. The sus-
pension was then poured into a Sartorius filtra-
tion cell equipped with a 0.2 μm aperture filter. 
The filtration was initially carried out at atmo-
spheric pressure. The remaining fluid from the 
filtration cell was obtained using compressed 
air. The residue of the solid phase in the filtra-
tion cell (filter cake) was dried in an oven at 
105°C until dry. The solution containing Na
2
SO
4
 
was also processed to dryness in the oven. The 
process is schematically shown in Figure 3. Af-
ter drying, both masses were determined. The 
weight of the filter cake of the water-insoluble 
solid phase was 72.5 g, while the mass in the 
water-soluble material was 35.9 g. The mea-
sured sulphur content of both samples is given 
in Table 7.
Table 7 shows that the sulphur content of the 
insoluble material is 0.56%.
Table 3. Classification at a cut size of 10 μm
Sample 1 (S1)
Mass (g) Mass per cent (%) Sulphur content (%)
258.4 100.0 3.38
Fine fraction (1F) 56.3 21.8 0.55
Coarse fraction (1C) 202.1 78.2 4.19
Table 4. Classification at a cut size of 20 μm
Sample 2 (S2)
Mass (g) Mass per cent (%) Sulphur content (%)
251.7 100.0 3.38
Fine fraction (2F) 128.9 51.2 3.30
Coarse fraction (2C) 122.8 48.8 3.44
Table 5. Classification at a cut size of 30 μm
Sample 3 (S3)
Mass (g) Mass per cent (%) Sulphur content (%)
260.2 100.0 3.38
Fine fraction (3F) 195.4 75.1 3.71
Coarse fraction (3C) 64.8 24.9 2.32
Table 6. Classification at a cut size of 40 μm
Sample 4 (S4)
Mass (g) Mass per cent (%) Sulphur content (%)
254.0 100.0 3.38
Fine fraction (4F) 198.8 78.3 3.49
Coarse fraction (4C) 55.2 21.7 2.81

Janc B., Hann D. RMZ – M&G | 2020 | Vol. 66 | pp. 215–222
220
In this case, it is the majority of the material 
(66.9%). Sulphur is concentrated in dissolved 
material, in which the sulphur content in-
creased to 8.39%.
Figure 4 shows that only the 5I and 1F samples 
represent the minimisation of sulphur content, 
with the 5I sample having a much larger mass 
fraction than the 1F sample, which means that 
the separation using wet process is more opti-
mal.
Conclusions
This article presents dry and wet methods of 
minimising the sulphur content in filter dust.
A dry method of sulphur minimisation would 
be a faster and more cost-effective method of 
sulphur removal in the industry but attempting 
to classify it at different cut sizes did not pro-
duce adequate results. With respect to the sul-
phur content, only a classification at the cut size 
of 10 m m would be appropriate, where the sul-
phur content in the fine fraction S = 0.55% was 
measured, but the share of this fraction was 
only 21.8%. The mass content of sulphur in the 
coarse fraction at this cut size was S = 4.19%. 
In the case of all other cut sizes (20, 30 and 
40 m m), the sulphur-containing particles are 
approximately evenly distributed between the 
fine and coarse fractions so that the sulphur 
content in both the fine and coarse fractions 
does not deviate significantly from the sulphur 
content of the starting sample.
Better results were obtained from the wet sul-
phur content minimisation process. In this part 
of the study, the sample was mixed with a suf-
ficient amount of water, and the suspension 
was heated to the temperature necessary for 
Figure 3. Dissolution, filtration and drying of material.
Figure 4. Sulphur and mass content for different samples.
Table 7. Results of an attempt to remove sulphur from filter ash by the wet procedure
Sample 5 (S5)
Mass (g) Mass per cent (%) Sulphur content (%)
108.4 100.0 3.38
Insoluble material (5I) 72.5 66.9 0.56
Soluble material (5S) 35.9 33.1 8.39

Removal of Na2SO4 from a Filter Ash
221
maximum solubility of sodium sulphate in wa-
ter. The water-soluble material, together with 
the insoluble residue, was filtered on a 0.2 m m 
aperture filter and then dried, and the sulphur 
content in the insoluble and water-soluble ma-
terial was measured. The insoluble material 
represented about two-third of the total sample 
and had a sulphur content of S = 0.56%, while 
the measured sulphur content in the rest of the 
material was S = 8.39%.
Considering that a lot of energy is consumed 
in drying the material, a suitable solution for 
the industrial removal of Na
2
SO
4
 would include 
mixing of water and material while heating at 
35°C, separation of the solid and liquid phase 
by filtration, elimination of Na
2
SO
4
 from the liq-
uid phase by reverse osmosis and drying of the 
products.
By using reverse osmosis, drying volumes and 
energy consumption would be greatly reduced.
References
[1] Aleksanin, A. (2019): Development of construction 
waste management. In: XXII International Scientific 
Conference on Advanced In Civil Engineering, Tash-
kent, Uzbekistan, Volkov, A., Pustovgar, A., Sultanov, 
T ., Adamtsevich, A. (eds.). E3S Web of Conferences.
[2] Eurostat [online]. European Commission [cited 
1/15/2020]. Available on: https://ec.europa.eu/eu-
rostat/statistics-explained/index.php?title=Waste_
statistics#Total_waste_generation.
[3] DIRECTIVE 2008/98/EC (2008). Directive 
2008/98/EC of the European Parliament and Council 
of 19 November 2008 on Waste and Repealing Cerain 
Directives. European Parliament. Official Journal of 
the European Union, pp. 3–30.
[4] Shahbazi, S., Kurdve, M., Bjelkemyr, M., Jönsson, C., 
Wiktorsson, M. (2013): Industrial waste manage-
ment within manufacturing: A comparative study of 
tools, policies, visions and concepts. In: Proceedings 
of the 11
th
 International Conference on Manufactur-
ing Research (ICMR2013), Shehab, E., Ball, P ., Tjahjo-
no, B. (eds.). Cranfield University: Cranfield, UK, pp. 
637–642.
[5] Altun, O., Toprak, A., Benzer, H., Darilmaz, O. (2016): 
Multi component modelling of an air classifier. Min-
erals Engineering, 93, pp. 50–56. DOI:10.1016/j.mi-
neng.2016.04.014.
[6] Bharmoria, P ., Gehlot, P .S., Gupta, H., Kumar, A. (2014): 
Temperature-Dependent Solubility Transition of Na-
2
SO
4
 in Water and Effect of NaCl Therein: Solution 
Structures and Salt Water Dynamics. The journal of 
physical chemistry B, 118, pp. 12734–12742. DOI: 
10.1021/jp507949h.
[7] Stražišar, J., Knez, S. (2001): Vaje in računski primeri 
iz mehanske procesne tehnike. Univerza v Ljubljani, 
Naravoslovnotehniška fakulteta, Oddelek za geoteh-
nologijo in rudarstvo: Ljubljana, 176 p.