Chemical Engineering Communications

ISSN: 0098-6445 (Print) 1563-5201 (Online) Journal homepage: https://www.tandfonline.com/loi/gcec20

Characterization and purification of waste

phosphogypsum to make it suitable for use in the
plaster and the cement industry

Yassine Ennaciri, Ilham Zdah, Hanan El Alaoui-Belghiti & Mohammed

Bettach

To cite this article: Yassine Ennaciri, Ilham Zdah, Hanan El Alaoui-Belghiti & Mohammed
Bettach (2019): Characterization and purification of waste phosphogypsum to make it suitable
for use in the plaster and the cement industry, Chemical Engineering Communications, DOI:
10.1080/00986445.2019.1599865

To link to this article: https://doi.org/10.1080/00986445.2019.1599865

Published online: 11 Apr 2019.

Submit your article to this journal

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=gcec20
CHEMICAL ENGINEERING COMMUNICATIONS
https://doi.org/10.1080/00986445.2019.1599865

Characterization and purification of waste phosphogypsum to make it

suitable for use in the plaster and the cement industry
Yassine Ennaciri , Ilham Zdah , Hanan El Alaoui-Belghiti , and Mohammed Bettach
Laboratory of Physical Chemistry of Materials (LPCM), Faculty of Sciences, Chouaib Doukkali University, El Jadida, Morocco

ABSTRACT KEYWORD
The phosphoric acid production in Morocco generates a large amount of phosphogypsum Characterization; Fluorine;
(PG) which is a significant source of environmental contamination. In this work, PG has Phosphogypsum;
been characterized by different analysis techniques that allow to study the effect of the Phosphorus; Purification
impurities on its physical and chemical properties. Thus, a method to purify the PG from
phosphates and fluorides was proposed in order to improve its quality and to make it
usable in plaster and cement. The purification process consists to sieve PG under 200 lm
and to treat it by sulfuric acid to dissolve the insoluble impurities. The obtained results
show that the treatment of PG by sulfuric acid 20 and 50% at 60
C for 2 h is sufficient to
make it suitable for used in the plaster and the cement respectively.

Introduction can substitute the natural gypsum. However, the

Phosphogypsum (PG) is a synthetic by-product
created during the commercial manufacture of
phosphoric acid by the wet process. Morocco
produces every year about 15 million tons of an
acid residue which is PG. All of this quantity is
discharged into the Atlantic Ocean. Pollution
problems resulting have been raised by several
authors, and reuse solutions or recovery of such
waste have also been proposed in other fields.
Only a small part of PG is used industrially or
commercially, the rest is stored or dumped into
water bodies. Although PG is especially constituted
of the CaSO4.2H2O dihydrate form, it contains
various impurities such as fluorides, phosphates,
clay minerals, organic matter, trace elements, and
radioelements. Hence, the presence of these impur-
ities at higher than natural levels, would be consid-
ered as environmental contamination (Saadaoui
et al., 2017; Tewo et al., 2017; El Zrelli et al., 2018).
The major domains of PG valorization are
agriculture, cement, and plaster industry. PG is
also employed as a remedy for saline soils and as
a fertilizer for several plant species (Ghazel et al.,
2018). In the cement and the plaster industry, PG
impurities contained in PG affect the physical
and chemical characteristics of this product,
which require an adequate treatment before using
it (Singh, 2002; Moalla et al., 2017). Several tech-
niques have been suggested to reduce the content
of impurities in PG including simple washing
with water, wet sieving, thermal decomposition
to hemihydrate and anhydrite, neutralization with
basic solution (Ca(OH)2, NH4OH …), or its treat-
ment with acid (sulfuric, citric …) (Singh et al.,
1996; Singh, 2002; Kandil et al., 2017).
Thermal or wet decomposition of PG have also
been investigated, which allows converting it to
high-value commercial products such as sulfuric
acid and alkali sulfates (Ennaciri et al., 2013, 2016;
Tewo et al., 2017; Ennaciri and Bettach, 2018).
The main objective of this paper is the charac-
terization of Moroccan PG in order to determine
the effect of impurities (especially phosphorus
and fluorine) on its properties. Thus, a procedure
for treating PG by sulfuric acid has been pro-
posed to reduce the maximum phosphorus and
fluorine content in order to adapt it to be used
in the plaster and cement industry.

CONTACT Yassine Ennaciri yassinemaster@gmail.com Laboratory of Physical Chemistry of Materials (LPCM), Faculty of Sciences, Route Ben
Maachou, El Jadida, 24000, Morocco.
Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/gcec.
ß 2019 Taylor & Francis Group, LLC
2 Y. ENNACIRI ET AL.

Material and methods

PG sample is recuperated from the fertilizer plant
Maroc Phosphore (dihydrate processes) settled at
Jorf Lasfar. The gypsum used in this research was
from reagent grade chemicals (Riedel de Ha€en),
with the purity superior than 99%. This type of
gypsum contained a chemical composition well
define (CaSO4.2H2O) which can be used to com-
pare its reactivity with that of PG.
Physical and morphological characterizations
were performed on PG and on the compounds
Figure 1. XRD pattern of PG sample.
produced in this work by the following techni-
ques: X-ray diffraction (XRD BRUKER D8 with PG sample. The change in the slope of the curve
Cu Ka radiation) and scanning electron micros- indicates the saturation of the solution and con-
copy X-ray analysis (SEM Environmental FEI sequently allows solubility value determination
Quanta 200). The particle size distribution was (expressed in g/l). The measurements were per-
determined by laser granulometry (CED center El
formed using a conductometer (Mettler toledo
Jadida). Concentrations of chemical elements
LE703) at room temperature after stirring the
were obtained by X-ray fluorescence (XRF S4
solution for 2 min.
PIONEER BRUKER aXS), flame photometer (FP
A weight equal to 150 g of raw PG was washed
JENWAY 500-731 Model PFP7), and inductively
in a beaker contained 250 ml of distilled water
coupled plasma mass-spectrometry (ICP-MS; HP-
under mechanical stirring for 15 min at ambient
4500 instrument). Infrared spectra were per-
temperature. After decantation, the supernatant
formed by Fourier transform infrared spectros-
was eliminated. After filtration through a
copy (FTIR 8400 s SHIMADZU spectrometer)
Buchner, the washed PG was dried at 40
C and
using KBr pellets in the region of
then the contents of P2O5 and F were analyzed.
4000–450 cm1. Raman spectrum was performed
For the purification of PG, the different sam-
by Timegated Raman Spectrometer. The dehydra-
ples were thoroughly mixed with 10, 20, 30, 50,
tion and carbonation degree of our samples were
and 98% aqueous sulfuric acid solution with the
investigated by thermogravimetric analyses
ratio PG/H2SO4 equal to 1/2, under mechanical
(DTG-60 type SHIMADZU).
Contents of P2O5 and total carbon were ana- stirring (200 tr/min) for 1 or 2 h at 40 and 60
C.
lyzed by spectrophotometer (Rayleigh VIS- The solution was filtered through a Buchner and
7220G/UV-9200) while the fluoride was meas- the purified PG was washed with distilled water
ured by ionometric method (fluoride ion selective for three times. The purified gypsum samples
electrode Tacussel TR 200) after necessary were dried at 40
C. Finally, the contents of P2O5
calibration. and F were analyzed.
The pH of PG was determined by mixing
weight of sample with the same weight of dis- Results and discussion
tilled water. After stirring for a half hour, we
Characterization of PG
have measured the pH and the potential of the
solution with a pH-meter (Jenco Vision Plus The XRD pattern of the PG sample (Figure 1)
pH 6175). shows clearly that PG is composed mainly of
Density of PG sample was measured by water gypsum (JCPDS 33-0311) with a low amount of
pycnometry technique. quartz (JCPDS 87-0703). The previous researches
Solubility was determined by conductivity reveal that the majority of PG samples have this
measurements on a solution of 100 ml of distilled composition (Essabir et al., 2017; El Zrelli et al.,
water to which we have added small masses of 2018; Grabas et al., 2018).
 CHEMICAL ENGINEERING COMMUNICATIONS 3

Figure 2. Infrared spectrum of PG compared with

pure gypsum.

Figure 4. TGA-DTA curves of PG (a) and gypsum (b).

Figure 3. Raman spectrum of PG sample. characteristic bands of SO42 and H2O. A very
weak band which appears at 572 cm1 is assigned
Table 1. Infrared and Raman frequencies of PG. to water torsion
T (H2O). The other bands of
Wave number
(cm1) SO42 appear around 420 (
2), 495 (
2), 620 (
4),
Assignment FTIR Raman 670 (
4), 1012 (
1), and 1140 (
3) cm1
H2O
3600–3400 – (Kloprogge and Frost, 2000).
d 1687; 1624 –

T – 572 Vibration band frequencies and assignments
SO42
3 1132 1140 for PG are listed in Table 1.

1 1004 1012

4 669; 603 620; 670 TGA-DTA curves for PG and gypsum are rep-

2 – 420; 495 resented in Figure 4. The results of TGA indi-
HPO42
837 –
cated that the total mass loss was 20.8% (20.6%
for gypsum). The mass loss starts at 110
C and
Figure 2 shows the infrared spectrum of PG ends at 200
C. It corresponds to the elimination
sample compared to that of pure gypsum. It indi- of the crystallization water. The elimination of
cated the presence of all bands of SO42 and H2O this water was done in two stages: the first weight
characteristic of gypsum, except a band at 837 cm1 loss (13.9 and 13.8% for PG and gypsum, respect-
which corresponds to HPO42 due to the formation ively) corresponds to the weight of 3/2 H2O, the
of CaHPO4.2H2O syncrystallized with gypsum second weight loss (6.9 and 6.8% for PG and
CaSO4.2H2O. This latter band can be observed for gypsum, respectively) represents the weight of
certain PG but generally it was difficult to detect 1/2 H2O.
(Ennaciri et al., 2016; Essabir et al., 2017). DTA curve of PG exhibits three peaks. The
Raman spectrum of PG sample at room tem- first peak endothermic at 144
C (139
C for gyp-
perature (Figure 3) indicates the presence of the sum) is due to the dehydration of dihydrate to
4 Y. ENNACIRI ET AL.

Table 2. Major elements in PG sample determined by X-ray fluorescence.

PG sedimentary PG igneous
Major
elements (%) this work Morocco(a) Morocco(b) Tunisia(c) Egypt(d) China(e) Russia(f) Poland(g)
CaO 31.71 30.92 19.3 37.18 28.31 26.77 31–37 29.6–31.4
SO3 43.40 43.93 53.5 37.50 40.45 36.96 44–45 42.11–43.01
H2O 20.80(1) 21.58 – – 19.71 18.89 20–25 19.31–19.80
P2O5 1.20(2) 1.23 0.8 1.11 1.98 1.58 0.91–2.86 0.34–0.6
F 1.10(3) 0.94 1.20 0.52 0.26 0.26 0.14–0.48 –
Na2O 0.27 0.22 0.22 0.05 0.29 – 0.24–0.32 0.079–0.834
K2O 0.05 0.01 – 0.03 0.02 0.15 0.04–0.08 0.041–0.237
SiO2 0.74 0.61 1.5 1.03 8.29 13.94 0.3–0.6 0.36–1.08
Al2O3 0.26 0.18 0.89 0.06 0.17 0.68 – 0.18–0.30
MgO 0.08 0.00 0.2 0.07 0.21 0.69 – –
Fe2O3 0.07 0.04 1.49 0.13 0.31 0.12 – 0.06–0.20
Y2O3 0.02 0.02 – – – – – –
SrO 0.06 0.06 – 0.46 – – – 1.30–1.45
TiO2 0.01 0.02 – – 0.01 – – –
Cl 0.02 – – – – – – –
C 0.25(2) 0.37 – 0.50 – – – 1.14–1.20
(1)
Thermal analysis, (2)spectrophotometric method, and (3)ionometric method.
(a)
Bourgier, 2007; (b)Essabir et al., 2017; (c)El Zrelli et al., 2018; (d)Kandil et al., 2017; (e)
Mi et al., 2018; (f)Kaziliunas et al., 2006; (g)
Grabas et al., 2018.

hemihydrate: CaSO42H2O fi CaSO41/2H2O þ as follows: (i) the origin of Moroccan PG is sedi-

3/2H2O. The second peak endothermic found at mentary rocks; (ii) silicon and alumina are some-
162 (157
C for gypsum), is attributed to the times added to the liquor in the production of
dehydration of hemihydrate to anhydrite III: phosphoric acid to facilitate the crystallization
CaSO41/2H2O fi CaSO4 III þ 1/2H2O. The last process; and (iii) SiO2 can be used to neutralize
exothermic peak at 430
C (353
C for gypsum) fluorine, sodium, or potassium to form the com-
indicates the crystalline transition of the anhyd- pounds H2SiF6, Na2SiF6, and K2SiF6 (Bourgier,
rous form of the gypsum: CaSO4 III fi CaSO4 II. 2007; Manar, 2016).
Generally, the thermal behavior of Moroccan The presence of soluble phosphorus (expressed
PG is similar to pure gypsum. From these results, by P2O5) in PG is due essentially to the remain-
we can confirm that the presence of impurities ing acid H3PO4. In addition, the insoluble phos-
(especially the HPO42 ions) affect on the tem- phorus present in PG can be attributed to the
perature of different thermal transitions residual inattacked phosphate rock or to the sec-
(Bourgier, 2007; Grabas et al., 2018). ondary reactions in the attack environment which
Table 2 shows the chemical composition (% can engender CaHPO4.2H2O. Infrared spectros-
weight) of the main species and major impurities copy indicates previously that phosphorus is in
in PG compared with other gypsums from differ- form of syncristalezed ions HPO42 which
ent origins. replace the ions SO42 in the structure of gypsum
In the light of these analyses, we can notice (Ennaciri et al., 2016).
clearly that the composition of the PG was domi- The analysis of PG confirm also the existence
nated by calcium, sulfur and water, accompanied of Al2O3, Fe2O3, and MgO characteristics of clay
by major content impurities like SiO2, P2O5, and minerals especially with the presence of Na2O
F and also Al2O3, Na2O, K2O, Fe2O3, and MgO and K2O (Ennaciri and Bettach, 2018). PG con-
in lesser amount. Usually, the variation of the tains also SrO, Y2O3, and TiO2 which prove its
percentage of these dominant elements is due to sedimentary origin (Bourgier, 2007). The stron-
the substitution of the calcium or the sulfur by tium co-precipitates with radon and barium to
other chemical elements having the same charge form solid solution (Ba,Sr)SO4 (Grandia et al.,
and similar size during the digestion of phos- 2008). The quantity of yttrium is important
phate rock by sulfuric acid. which makes it detectable by the X-ray fluores-
The content of silicon is low (0.74% SiO2). cence. The content of titanium oxide is similar to
This is already confirmed by X-ray diffraction that reported in other PG (Bourgier, 2007; Kandil
analysis. The presence of silicon can be explained et al., 2017).
 CHEMICAL ENGINEERING COMMUNICATIONS 5

Table 3. Trace elements in PG sample determined by ICP.

PG sedimentary PG igneous
Trace
elements (ppm) this work Morocco(a) Morocco(b) Tunisia(c) Egypt(d) Florida(e) Togo(f) Poland(g)
(1)
Ba 30.08 64 39.7 10 78 30 26 54
Cd 1.34 1 8.6 17.70 – 0.2 7.2 1.31–1.43
Hg 0.65 – 0.28 0.028 – 27 18 –
Pb 0.73 2 5.01 0.9 – – 9 2.12–3.09
Cr 7.5 12 22 13 34 1.7 1.1 4.06–5.40
Zr 19(2) 9 – 2.6 16 23 16 –
Zn 43(2) 49 30.1 137 25 6.4 31 9.4–66.0
Cu 55(2) 4 2.1 9.6 12 3.3 3.3 17.0–32.6
(1)
Flame photometer and (2)X-ray fluorescence.
(a)
Bourgier, 2007; (b)Gaudry et al., 2007; (c)El Zrelli et al., 2018; (d)
Kandil et al., 2017; (e)
and (f)Rutherford et al., 1995; (g)
Grabas et al., 2018.

complexes such as Na2SiF6, Na3AlF6, K3AlF6 … or

to the ion FPO32 syncrystallized with gypsum
substituting easily the sulfate ions SO42 to form
CaFPO3.2H2O (Bourgier, 2007).
The concentrations (ppm) of some trace ele-
ments (effectively toxic element) in the PG are
shown in Table 3.
The content of the barium (30 ppm) remains
lower than that determined for other gypsums by
Kandil et al. (2017) (78 ppm) and by Bourgier
(2007) (64 ppm). However, it is closer to the val-
ues given by Rutherford et al. (1995) (30 and
Figure 5. Particle size distribution. 26 ppm for two samples from Florida and Togo
PG, respectively). Several factors promise the
Our sample contains chlorine which can be presence of barium in PG such as similar chem-
due to its addition during phosphoric acid pro- ical behavior of barium and calcium, the granulo-
cess when the concentration of cadmium rises. metric fraction less than 20 lm in PG and the
Chlorine can also come from the water used dur- co-precipitation with radium (Ra,Ba)SO4 and
ing the various stages of phosphate rock digestion strontium (Sr,Ba)SO4 (Grandia et al., 2008).
(Ruiz et al., 2002). Our PG sample contains also toxic and heavy
For the carbon total especially the organic metals such as Pb, Cd, Cr, and Hg which are con-
form, PG sample contains a low amount of this sidered as dangerous elements for the environment
element (0.25%) compared to other PG in the and for the human health, especially when their
world: 0.37% (Bourgier, 2007), 0.50% (El Zrelli concentrations exceed the standards. Generally, the
et al., 2018), 0.70% (Moalla et al., 2017), concentrations of these heavy metals are lower
1.14–1.20% (Grabas et al., 2018). Generally, the than those reported by other authors in PG
presence of the carbon in the PG is due to ali- obtained from sedimentary phosphate rocks.
phatic compounds of carbonic acids, amines, Lead is less soluble (2 mg/l) and forms PbSO4 or
ketones, modifiers, or flocculants introduced dur- PbS according to the conditions of oxidation or
ing filtration (Bourgier, 2007). reduction. Generally, Pb, Cu, and Zn can substitute
PG sample contains 1.10% of fluorine, this Ca in PG to generate PbSO42H2O, CuSO42H2O,
value remains closer to that obtained by Bourgier and ZnSO42H2O. The rate of insertion of this
(2007) and Essabir et al. (2017) for Moroccan PG metal in the resulting PG is limited to 30%
which are 0.94 and 1.20%, respectively. The pres- (Lamzougui et al., 2016). However, the adsorption
ence of soluble fluorine in PG results from the of lead by clay and organic matter maintains its
free acid like HF, H2SiF6 (Long et al., 2018). low concentration (Rutherford et al., 1995).
While the insoluble fluorine can be attributed to The presence of cadmium in our sample is
the phosphate rock not attacked, to insoluble often associated with enrichment of this element
6 Y. ENNACIRI ET AL.

Figure 6. SEM of PG sample.

during phosphate rock digestion, adsorption of Table 4. Content of P2O5 and F in raw, washed, and puri-
organic matter, or formation of metallic sulfide fied PG.
precipitates (Rutherford et al., 1995). Balkaya and Content P2O5 F Na2O þ K2O C pH
Raw PG 1,2 1,1 0.33 0.25 2.15
Cesur (2016) have observed a very rapid chemi- Washed PG 0,72 0,88 0.14 0.08 4.01
sorption mechanism of cadmium by PG. Purified PG (20 %, 0.20 0.08 0.06 0.01 5.11
60
C, 2H)
The content of zinc (Table 3) remains import- Plaster (Moalla et al., <0.4 % <0.4 % <30 % <0.15 % >5
ant and almost closer than other Moroccan PG 2017)
Purified PG (50%, 0.06 0.02 0.01 0.00 5.85
analyzed by Gaudry et al. (2007) (49 ppm). This 60
C, 2H)
can be explained by the transfer and the enrich- Cement (Bourgier, <0.05% <0.05% – – >5
2007)
ment of this element during the phosphate rocks
digestion. On the other hand, the content of zir-
The morphological structure of PG sample was
conium is closer to that determined by
observed by SEM technique (Figure 6). We
Rutherford et al. (1995) (16 and 23 ppm).
observe that the crystal form is tabular with a
The laser granulometry curve (Figure 5) allows
a good visualizing of the distribution of the size grading from 40 to 180 mm. This morphology
grains thickness in our sample. It can be noted can be explained by the presence of alkaline and
that there is practically no particle larger than HPO42 syncrystallized ions (Mangin, 1978;
210 mm and that the majority of the weights of Bourgier, 2007).
the sample result from particles of sizes in the The PG has an acidic pH with a mean value of
range 1–200 mm. The granulometric distribution 2.15 (see Table 4). This value is close to those
of the PG is carried out according to two popula- reported in previous work by Perez-L opez et al.
tions. The first population has a significant par- (2010) (pH between 2.42 and 2.68). This PG
ticle number of around 10 mm in diameter. The acidity is probably due to the presence of residual
second population contains grains of about acids, such as H3PO4, H2SO4, and HF.
80 mm in diameter. The density (specific gravity) determined by
The laser granulometry analysis of PG sample water pycnometer of the PG sample and the pure
gives also the average specific surface area value gypsum is 2.24 and 2.31 g/cm3 respectively. The
of 0.521 m2/g. This value is closer to values given difference of these values can be explained by the
by Laganiere (2000) for other Moroccan PG presence of considerable amount of water and
(0.345 m2/g). However, Balkaya and Cesur (2003) impurities within PG.
have determined high specific surface area value Figure 7 show the conductimetric curves of PG
of 11 m2/g for Turkish gypsum. and gypsum dissolution in distilled water (a) and

Figure 7. Conductimetric curves of PG dissolution in distilled
water (a) and in seawater (b).
in sea water (b), respectively. The solubility of
PG in distilled water is 2.89 g/l. It is nearly close
to that of the Tunisian PG (2.7 g/l) (Hammas
et al., 2013). The solubility of gypsum is 2.41 g/l,
which confirms that the PG is more soluble
because of the impurities which it contains. In
addition, the solubility of the same PG in sea
water is higher (4.16 g/l). This value due to the
salinity (the increase of ionic strength) of the sea
water is reasonable. Bourgier (2007) and James
(1992) indicate that the solubility of PG in sea
water is between 3.5 and 4 g/l.
Purification of PG
Figure 8 represents the contents of P2O5 and F in
not sieved PG and sieved PG in different frac-
tions with various particle sizes. The results show
that the fraction of larger particles (above
CHEMICAL ENGINEERING COMMUNICATIONS 7
Figure 8. Contents of P2O5 and F in not sieved raw PG and
for the fractions of particle size under or higher than 200 mm.
200 mm) is rich in P2O5 and F compared to PG
not sieved. For these reasons, we have adopted a
granular cut at 200 mm of PG which allows us to
eliminate the big particles rich in impurities (in
particular, not attacked phosphate and fluorapa-
tite) with a low loss of mass.
In the so-called Cerphos process (Traub, 2003)
a large part of the impurities in PG is removed
by sieving. In an investigation of a Moroccan PG,
it was found that the particle size fractions larger
than 170 mm were heavily enriched with F, Si,
and Na. On the other hand, the particle size frac-
tions smaller than 25 mm were rich in organic
substances and in phosphate combined in the
crystal lattice. Approximately 75% of the gypsum
can be recovered in a clean state by this method.
Generally, the purification method of PG depends
essentially on the nature and the application field
of this raw matter (Singh et al., 1996; Kandil
et al., 2017).
The contents of P2O5 and F for the washed PG
not sieved and for particle sizes bellow 200 mm
are almost similar (Figure 9). The washing of PG
by distilled water makes it possible to eliminate
the maximum of soluble impurities such as the
trace of acid (H3PO4, H2SO4, HF, and H2SiF6).
Whereas phosphorus and fluorine substituted in
the crystal lattice of gypsum are difficult to
remove by washing. According to El Zrelli et al.
(2018), the distribution of chemicals elements Ca,
S, O, and P in sedimentary PG are homogenous,
except for F which was found to be concentrated
with other particles essentially with elements of
clays Si, Al, Fe, Mg, Na, and K. These elements
8 Y. ENNACIRI ET AL.

Figure 9. Contents of P2O5 and F in washed PG for particle Figure 10. Content of P2O5 and F in purified PG by sulfuric
sizes under 200 mm. acid 10 and 20 % for particle sizes less than 200 mm.

are presented in fine particles less than 20 mm

(Renterıa-Villalobos et al., 2010). In this case, the
similarity of contents of P2O5 and F in different
granular cuts may be explained by uniformity
and homogenous distribution of these insoluble
phosphorus and fluorine. The rate of elimination
of P2O5 and F from raw PG to washed PG with
granular size below 200 mm are almost 40 and
20%, respectively. Singh et al. (1996) improved a
process for the purification of PG by sieving
under 300 mm and washing it with water in the
proportion 1:3 by volume. It reveals that water
reduces considerably the soluble impurities P2O5,
F, organic matter, and alkalies (from 1.28 to Figure 11. Content of P2O5 and F in purified PG by sulfuric
0.42% for P2O5 and from 1.80 to 0.60% for F). In acid 30, 50, and 98 % for particle sizes less than 200 mm.
others works, Khadilkar et al. (2009), Lokshin
allows the formation of water-soluble phosphoric
and Tareeva (2015) have proposed a process to
acid according to the following reactions:
neutralize the soluble impurities by using milk of
lime (Ca(OH)2, CaO). The obtained results prove CaðH2 PO4 Þ2  H2 O þ H2 SO4 ! 2H3 PO4 þ CaSO4 þ H2 O
that the totalities of soluble P2O5 and F are neu- CaHPO4  2H2 O þ H2 SO4 ! H3 PO4 þ CaSO4 þ 2H2 O
Ca3 ðPO4 Þ2 þ 3H2 SO4 ! 2H3 PO4 þ 3CaSO4
tralized and converted into insoluble calcium
phosphate and calcium fluorides. Fluorides can react also with aqueous sulfuric
Figure 10 shows the content of P2O5 and F in acid and form water-soluble sodium sulfate
purified PG by sulfuric acid 10 and 20% for par-
(Na2SO4), hydrofluorosilicic acid (H3SiF6), hydro-
ticle sizes less than 200 mm at 40 and 60
C. It’s
fluoroaluminate (H3AlF6), hydrofluoroferrate
important to note that the phosphates and fluo-
(H3FeF6), and hydrofluoric acid (HF) according
rides impurities are reduced when the concentra-
to the following reactions:
tion of sulfuric acid, the temperature and the
time of contact increase. The maximum removal Na2 SiF6 þ H2 SO4 ! Na2 SO4 þ H2 SiF6
rate of these impurities from washed to purified 2Na3 AlF6 þ 3H2 SO4 ! 3Na2 SO4 þ 2H3 AlF6
2Na3 FeF6 þ 3H2 SO4 ! 3Na2 SO4 þ 2H3 FeF6
PG is 53% and 82% for P2O5 and F respectively.
CaF2 þ H2 SO4 ! CaSO4 þ 2HF
The removal of insoluble phosphates present
in PG can be explained by the reaction of these These water-soluble compounds can be easily
phosphates with aqueous sulfuric acid which removed with distilled water.

According to these results (Figure 10 and
Table 4), the treatment of Moroccan PG by sul-
furic acid 20% at 60
C for 2 h is sufficient to
make it suitable for use in the plaster industry.
Compared to other work realized by Moalla et al.
(2017), the optimum conditions to remove P2O5
and F for the same temperature (60
C) corres-
pond to concentration acid 5%, a ratio PG/
H2SO4 equal 1/3 with optimized pH equal 6.2.
Under these conditions removal fraction of P2O5
and F are 60% and 95%, respectively.
To make the Moroccan PG suitable for use in
the cement, several experiments have been real-
ized to determine the optimum conditions of PG
purification. Figure 11 represents the contents of
P2O5 and F in purified PG after treatment with
sulfuric acid 30, 50, and 98% (ratio PG/H2SO4: 1/
2) at 40 and 60
C. The obtained results show
that the treatment of PG by sulfuric acid 50% at
60
C for 2 h is sufficient to make it suitable for
use in cement manufacture (Table 4). The max-
imum rate of elimination of the insoluble impur-
ities from washed PG to purified PG is 98.5% for
P2O5 and 97% for F for granular sizes lower than
200 mm. Kandil et al. (2017) have treated a sieved
PG at 200 mm by using sulfuric acid and obtained
almost the same results. Concentrations of P2O5
and F in PG decreased to 0.026% and 0.002%,
respectively, under the optimum conditions: sul-
furic acid concentration 50%, ratio PG/H2SO4: 1/4
at 80
C for 30 min contact time. In other hand,
Singh et al. (1993, 2002), have treated PG (sample
DA) by using ammonium hydroxide and citric
acid. The purified PG was found to have lesser
amount of phosphates impurities (0.44 and 0.46%,
respectively), fluorides (0.52 and 0.65%, respect-
ively), and organic matter (0.03 and 0.03%, respect-
ively) than the impure material. The optimum
concentrations of ammonium hydroxide and citric
acid solution were found to be 20 and 4%, respect-
ively, under room temperature. Generally, the puri-
fication of PG by sulfuric acid remains more
efficient than that by ammonium hydroxide, citric
acid, or nitric acid (Aliedeh and Jarrah, 2012).
The chemical components of raw, washed, and
purified PG are regrouped in the Table 4. Based
on these results, the purification process studied
on this work permits to obtain a purified PG
suitable to use in plaster and cement industry.
CHEMICAL ENGINEERING COMMUNICATIONS 9
Conclusion
In this work, we have presented the various tech-
niques of analyses which are complementary and
which allow realizing a complete characterization
of PG sample. We have identified the crystalline
structure, visualized the morphology, determined
the granulometry, and measured the contents of
chemical elements that exist in the form of
impurities. The latter, essentially P2O5 and F are
capable to modify the various physical and
chemical properties of the PG and affect signifi-
cantly the quality of cement and plaster prepared
by this matter.
Generally, the nature and the physical and
chemical properties of the Moroccan PG
studied in this work are similar to those of
other PGs of sedimentary origin in the world.
From an environmental point of view, it is clear
that the impurities contained in PG remain
low, but they present a risk of environmental
impact especially when they are concentrated in
large quantity and moved by leaching and/
or erosion.
Sieving and purifying the PG with sulfuric acid
allowed its transformation into valuable products.
PG grains smaller than 200 mm are less rich in
fluorides and phosphates. The PG recovered after
sieving is treated with concentrated sulfuric acid.
The weight ratio: PG/H2SO4 is 1/2. The treat-
ment temperature is 60
C. By respecting the
standards that limit the impurities of P2O5 and F
in gypsum, the treatment of PG with sulfuric
acid 20% is sufficient for use in the plaster indus-
try. On the other hand, it must be treated with
sulfuric acid 50% for use in the cement industry.
To use this purified PG in plaster and cement
industry, it’s necessary to take into account other
factors such as the microstructure and the micro-
morphology. Also the performance of the puri-
fied PG can be compared to that of
natural gypsum.
ORCID
Yassine Ennaciri http://orcid.org/0000-0002-2442-7397
Ilham Zdah http://orcid.org/0000-0002-4681-7003
Hanan El Alaoui-Belghiti http://orcid.org/0000-0001-
5888-0758
Mohammed Bettach http://orcid.org/0000-0003-1795-1796
10 Y. ENNACIRI ET AL.
References
Aliedeh, M. A., and Jarrah, N. A. (2012). Application of full
factorial design to optimize phosphogypsum beneficiation
process (P2O5 Reduction) by using sulfuric and nitric
acid solutions. Paper presented at the Sixth Jordanian
International Chemical Engineering Conference, Amman,
Jordan (pp. 1–10).
Balkaya, N., and Cesur, H. (2003). Influence of operating
parameters on lead removal from wastewater by phos-
phogypsum, Environ. Technol., 24, 727–733. doi:10.1080/
09593330309385609.
Balkaya, N., and Cesur, H. (2016). A kinetic study on cad-
mium adsorption from aqueous solutions by pre-condi-
tioned phosphogypsum, Desalin. Water Treat., 57,
2515–2521. doi:10.1080/19443994.2015.1038854.
Bourgier, V. (2007). Influence des ions monohy-
drogenophosphates et fluorophosphates sur les proprietes
des phosphogypses et la reactivite des phosphopl^atres
(Doctoral dissertation, Ecole Nationale Superieure des
Mines de Saint-Etienne).
Ennaciri, Y., and Bettach, M. (2018). Procedure to convert
phosphogypsum waste into valuable products, Mater.
Manuf. Process., 33, 1727–1733. doi:10.1080/10426914.
2018.1476763.
Ennaciri, Y., Bettach, M., Cherrat, A., and Zegzouti, A.
(2016). Conversion of phosphogypsum to sodium sulfate
and calcium carbonate in aqueous solution, J. Mater.
Environ. Sci., 7, 1925–1933.
Ennaciri, Y., Mouahid, F. E., Bendriss, A., and Bettach, M.
(2013). Conversion of phosphogypsum to potassium sul-
fate and calcium carbonate in aqueous solution, MATEC
Web of Conf., 5, 04006. doi:10.1051/matecconf/
20130504006
El Zrelli, R., Rabaoui, L., Daghbouj, N., Abda, H., Castet, S.,
Josse, C., Beek, P., Souhaut, M., Michel, S., Bejaoui, N.,
and Courjault-Rade, P. (2018). Characterization of phos-
phate rock and phosphogypsum from Gabes phosphate
fertilizer factories (SE Tunisia): High mining potential
and implications for environmental protection, Environ.
Sci. Pollut. Res., 25, 14690–14702. doi:10.1007/s11356-
018-1648-4.
Essabir, H., Nekhlaoui, S., Bensalah, M. O., Rodrigue, D.,
Bouhfid, R., and el Kacem Qaiss, A. (2017).
Phosphogypsum waste used as reinforcing fillers in poly-
propylene based composites: Structural, mechanical and
thermal properties, J. Polym. Environ., 25, 658–666. doi:
10.1007/s10924-016-0853-9.
Gaudry, A., Zeroual, S., Gaie-Levrel, F., Moskura, M.,
Boujrhal, F. Z., El Moursli, R. C., Guessous, A., Mouradi,
A., Givernaud, T., and Delmas, R. (2007). Heavy metals
pollution of the Atlantic marine environment by the
Moroccan phosphate industry, as observed through their
bioaccumulation in Ulva lactuca, Water Air Soil Pollut.,
178, 267–285. doi:10.1007/s11270-006-9196-9.
Ghazel, N., Saadaoui, E., Ben Romdhane, C., Abbes, N.,
Grira, M., Abdelkebir, S., Aydi, S., Abdallah, L., and
Mars, M. (2018). Assessment of phosphogypsum use in a
nursery for plant propagation, Int. J. Environ. Stud., 75,
284–293. doi:10.1080/00207233.2017.1356631.
Grabas, K., Pawełczyk, A., StreR k, W., SzełeR g, E., and StreR k,
S. (2018). Study on the properties of waste apatite phos-
phogypsum as a raw material of prospective applications,
Waste Biomass Valorization, 1–13. doi:10.1007/s12649-
018-0316-8.
Grandia, F., Merino, J., and Bruno, J. (2008). Assessment of
the radium-barium co-precipitation and its potential
influence on the solubility of Ra in the near-field (No.
SKB-TR–08-07), Swedish Nuclear Fuel and Waste
Management Co.
Hammas, I., Horchani-Naifer, K., and Ferid, M. (2013).
Solubility study and valorization of phosphogypsum salt
solution, Int. J. Miner. Process., 123, 87–93. doi:10.1016/j.
minpro.2013.05.008.
James, A. N. (1992). Soluble Materials in Civil Engineering,
Michigan University Ellis Horwood Limited, ISBN
0138262640, 9780138262648.
Kandil, A. H. T., Cheira, M. F., Gado, H. S., Soliman,
M. H., and Akl, H. M. (2017). Ammonium sulfate prep-
aration from phosphogypsum waste, J. Radiat. Res. Appl.
Sci., 10, 24–33. doi:10.1016/j.jrras.2016.11.001.
Kaziliunas, A., Leskeviciene, V., Vektaris, B., and Valancius,
Z. (2006). The study of neutralization of the dihydrate
phosphogypsum impurities, Ceramics Silikaty, 50, 178.
Khadilkar, S. A., Karandikar, M. V., Anikode, P. R., Lele,
P. G., Vaity, R. S., and Pathak, A. K. (2009). U.S. Patent
Application No. 12/438,593.
Kloprogge, J. T., and Frost, R. L. (2000). Raman microscopy
at 77 K of natural gypsum CaSO4.2H2O, J. Mater. Sci.
Lett., 19, 229–231. doi:10.1023/A:1006762825557.
Laganiere, D. (2000). Caracterisation, stabilisation et valor-
isation des phosphogypses marocains (French text, the-
sis). https://elibrary.ru/item.asp?id¼9036623.
Lamzougui, G., Nafai, H., Bouhaouss, A., and Bchitou, R.
(2016). Determination of the maximum content of heavy
metals in the phosphogypsum, J. Mater. Environ. Sci., 7,
2161–2169.
Lokshin, E. P., and Tareeva, O. A. (2015). Production of
high-quality gypsum raw materials from PG, Russ. J.
Appl. Chem., 88, 567–573. doi:10.1134/
S1070427215040023.
Long, B., Wang, Z., Zhang, Q., Ke, W., and Ding, Y. (2018).
Improved process to prepare high-purity anhydrous
potassium fluoride from wet process phosphoric acid,
Chem. Eng. Commun., 205, 1342–1350. doi:10.1080/
00986445.2018.1450246.
Manar, S. (2016). Increasing the filtration rate of phosphor-
gypsum by using mineral additives, Procedia Eng., 138,
151–163. doi:10.1016/j.proeng.2016.02.073.
Mangin, S. (1978). Le phosphogypse: utilisation d’un sou-
sproduit industriel en technique routiere, Bulletin de
Liaisons Des Laboratoires Des Ponts et de Chaussees,
Special VII, 7–13, ISSN: 0458-5860, The National

Academies of Sciences, Engineering, and Medicine 500
Fifth Street, NW, Washington, DC 20001.
Mi, Y., Chen, D., He, Y., and Wang, S. (2018).
Morphology-controlled preparation of a-calcium sulfate
hemihydrate from phosphogypsum by semi-liquid
method, Cryst. Res. Technol., 53, 1700162. doi:10.1002/
crat.201700162.
Moalla, R., Gargouri, M., Khmiri, F., Kamoun, L., and Zairi,
M. (2017). Phosphogypsum purification for plaster pro-
duction: A process optimization using full factorial
design, Environ. Eng. Res., 23, 36–45. doi:10.4491/eer.
2017.055.
Perez-Lopez, R., Nieto, J. M., L
opez-Coto, I., Aguado, J. L.,
Bolıvar, J. P., and Santisteban, M. (2010). Dynamics of
contaminants in phosphogypsum of the fertilizer industry
of Huelva (SW Spain): From phosphate rock ore to the
environment, Appl. Geochem., 25, 705–715. doi:10.1016/j.
apgeochem.2010.02.003.
Renterıa-Villalobos, M., Vioque, I., Mantero, J., and
Manj on, G. (2010). Radiological, chemical and morpho-
logical characterizations of phosphate rock and phospho-
gypsum from phosphoric acid factories in SW Spain, J.
Hazard. Mater., 181, 193–203. doi:10.1016/j.jhazmat.2010.
04.116.
Ruiz, C., Arribas, A., and Arribas, A. (2002). Mineralogy
and geochemistry of the Masa Valverde blind massive
CHEMICAL ENGINEERING COMMUNICATIONS 11
sulphide deposit, Iberian Pyrite Belt (Spain), Ore Geol.
Rev., 19, 1–22. doi:10.1016/S0169-1368(01)00037-3.
Rutherford, P. M., Dudas, M. J., and Arocena, J. M. (1995).
Radioactivity and elemental composition of phosphogyp-
sum produced from three phosphate rock sources, Waste
Manage. Res., 13, 407–423. doi:10.1177/
0734242X9501300502.
Saadaoui, E., Ghazel, N., Ben Romdhane, C., and Massoudi,
N. (2017). Phosphogypsum: Potential uses and prob-
lems–a review, Int. J. Environ. Stud., 74, 558–567. doi:10.
1080/00207233.2017.1356631.
Singh, M. (2002). Treating waste phosphogypsum for
cement and plaster manufacture, Cem. Concr. Res., 32,
1033–1038. doi:10.1016/S0008-8846(02)00723-8.
Singh, M., Garg, M., and Rehsi, S. S. (1993). Purifying phos-
phogypsum for cement manufacture, Constr. Build.
Mater., 7, 3–7. doi:10.1016/0950-0618(93)90018-8.
Singh, M., Garg, M., Verma, C. L., Handa, S. K., and
Kumar, R. (1996). An improved process for the purifica-
tion of phosphogypsum, Constr. Build. Mater., 10,
597–600. doi:10.1016/S0950-0618(96)00019-0.
Traub, H. (2003). Drytech, Personal Communication.
Tewo, R. K., Maree, J. P., Ruto, S., Rutto, H. L., and Koech,
L. K. (2017). The gypsum reduction process and its valid-
ation using the Mintek Pyrosim model, Chem. Eng.
Commun., 204, 1412–1419. doi:10.1080/00986445.2017.
1365064.