The Use of Phosphogypsum as a Source of Raw Materials for Gypsum-Based Materials

Abstract

Gypsum binders and the products based on them are widely in demand in the construction materials market, which is due to their easy production technology, lower energy consumption, and low environmental impact in relation to Portland cement. Not only natural gypsum (NG), but also phosphogypsum (PG), which is a by-product of the synthesis of orthophosphoric acid from phosphorite rock, can be used as a source of raw materials for the production of gypsum materials. PG is produced annually in large quantities throughout the world. In chemical composition, PG mainly consists of calcium sulfate dihydrate CaSO₄·2H₂O, so it is a good potential analogue of natural gypsum, which is used as the main component of gypsum building materials. Thus, the useful recycling of PG as a technogenic resource with valuable properties will expand the raw material base for the production of gypsum materials. This approach to handling technogenic resources fits well with the principles of a circular economy. However, like any technogenic resource, PGs from different enterprises normally differ in their deposits of the original phosphate rock and production technologies. Therefore, PG contains a large number of undesirable impurities, the proportion and composition of which vary over a wide range. This feature does not allow for predicting the properties of PG-based materials without a preliminary detailed study of PG. This research was aimed at carrying out a comprehensive study of the characteristics of PGs from three different industrial plants to evaluate their relationship with the properties of gypsum materials based on them. It was found that PGs have significant differences in their structural and morphological characteristics both in relation to each other and in relation to NG. Also, binders based on PG and NG have significant differences in their physical properties. The average density, compressive strength, and flexural strength for the PG binders with equal workability are lower than those of NG binders. At a water/solid ratio (W/S) < 0.7, all PG binders exhibit comparable compressive strength to NG binders. Thus, PG can act as an alternative to natural gypsum in gypsum binders.

1. Introduction

Gypsum-based materials (plasterboard, blocks, plaster, etc.) are widely in demand in the building materials market due to their easy manufacture, as well as their good physical and mechanical properties (fire resistance, low sound and thermal conductivity, light weight, etc.) [1,2,3]. It should also be noted that the production of gypsum binders, in comparison with Portland cement, is lower energy consumptive and is also not accompanied by the emissions of harmful impurities and CO₂ into the environment [4,5] (Equation (1)).

CaSO₄·2H₂O = β-CaSO₄·0.5H₂O + 1.5H₂O

(1)

Normally, natural gypsum (NG) is used for the production of gypsum binders. However, in countries where deposits of NG are limited, gypsum-bearing wastes from various industrial enterprises, such as phosphogypsum, citrogypsum, FGD-gypsum, etc., can be used as an alternative source for gypsum binder production [6,7,8,9,10]. The most large-tonnage of the above-mentioned wastes is phosphogypsum (PG), the annual production of which in the world is ≈300 million tons [8,9,10,11,12].

PG is a primary by-product formed during the production of phosphoric acid and an intermediate product in the production of phosphate fertilizers from phosphate rock (calcium phosphate or apatite) [9,10,11,12].

Only 15% of all phosphogypsum produced in the world is recycled; the remaining volume is stored or dumped into water reservoirs, which causes significant damage to the environment [12]. In this connection, the search for ways to use PG as a resource for the production of gypsum binders is a promised trend that will solve the environmental problems in regions where there are accumulations of FG, but will also expand their raw material base for the production of gypsum materials. This concept is in agreement with the principles of the circular economy.

Despite the fact that PG is mainly represented by calcium sulfate dihydrate (CaSO₄ 2H₂O) (Reaction (2)) [13], it also contains impurities of phosphorus, fluorine, organic substances, heavy metals, radioactive substances, etc. [11,14], which can have an ambiguous effect on the quality of gypsum materials. Therefore, PG, as a gypsum-bearing resource, requires additional, more detailed study.

Ca₅(PO₄)3F + 5H₂SO₄ + 10H₂O → 3H₃PO₄ + 5CaSO₄·2H₂O + HF

(2)

There are a large number of studies where the authors have shown the possibility of obtaining PG-based binders by the heat treatment of phosphogypsum. In particular, according to a number of works studying PG from different countries (China, India, Russian Federation) as a raw material for β-CaSO₄ 0.5H₂O, produced by high-temperature treatment in range of 140–220 °C, it has been established that, despite comparable conditions for the production of PG binders, the compressive strength varies from 4 to 13.8 MPa. There are also significant differences in the setting time [15,16,17,18,19,20].

This suggests that the quality of PG binders and the gypsum materials based on them is significantly influenced by the origin characteristics of the PG, such as mineralogical and chemical composition, impurities, grain size, etc. In turn, these characteristics depend on the technological parameters of orthophosphoric acid production, as well as the origin of the phosphate rock [9,21]. According to statistical data, 75% of the world’s sources of phosphorites are of sedimentary origin and 20–25% are of magmatic origin [9,22,23].

Previous research has mainly focused on the possibility of producing gypsum materials to produce a specific PG component. However, the number of works devoted to the study of the relationship between the characteristics of PGs and the properties of the materials based on them is very limited. At the same time, it is believed that the genetic characteristics of PG (mineral and chemical composition, impurities, grain size, etc.) have a critical impact on the quality of gypsum materials and the efficiency of their production.

This study is aimed at comprehensively investigating the characteristics of PG from three different industrial enterprises to evaluate their relationship to the properties of the gypsum materials based on them.

2. Materials and Methods

The objects of study were the PGs, which were the by-products of the three following Russian enterprises:

1. CJSC PhosAgro AG–PG_Bal (Moscow, Russia): Here, the source of raw materials for the production of phosphoric acid is the Khibiny Mine of the alkaline rocks of the Kola Peninsula with sheet- and lens-shaped deposits of complex apatite ores associated with nepheline syenites.

2. Phosphogypsum EuroChem-Belorechenskiye Mineral Fertilizers LLC–PG_Bel (Zug, Switzerland): Here, the source of raw materials for the production of phosphoric acid is the apatite concentrate of the Kovdor Mine of ultra-basic alkaline rocks with carbonatites (igneous and metasomatic) of Lower Paleozoic age, as well as phosphorites of the Karatau phosphorite-bearing basin (Republic of Kazakhstan).

3. Phosphogypsum LLC “PG” Phosphorit”–PG_Kin (Kingisepp, Russia): Here, the source of raw materials for the production of phosphoric acid is the same as that of PG_Bel, but the proportion of raw materials, equipment used, and technology production are different. This may affect the characteristics of the PG and, as a consequence, the gypsum binders based on it.

For this research, the PG sample was used, which was obtained by mixing 20 kg of samples taken from 5 different points of each source.

The reference raw material was NG from the Shedok deposit (Russia).

The chemical and elemental compositions of the studied samples were determined by X-ray fluorescence analysis (XRF), and their mineral composition by X-ray diffraction analysis (XRD). Both methods were realized using an X-ray workstation WorkStation ARL 9900 (Thermo Scientific, Waltham, MA, USA), with co-anode radiation.

Differential thermal analysis (DTA-analysis) was carried out using a thermal analyzer STA 449 F1 JUPITER (Netzsch, Germany).

The study of the microstructure of the PGs and NG particles, as well as the morphology of new formations for synthesized gypsum binders, was carried out with scanning electron (SEM) microscopy using a Mira 3 FesSem scanning electron microscope (Tescan, Czech Republic), operating in high vacuum mode (InBeam) using a Schottky cathode high brightness. The samples were coated with chromium.

The specific surface area (SSA) and the pore size distribution of the PGs and NG particles and the gypsum binders based on them was measured using a multi-functional device PSH-12 (SP) (Khodakov Devices, Moscow, Russia) by Kozeny-Carman gas permeability method. The gas permeability of a powder layer is determined by the duration of air filtration through it at a fixed initial and final vacuum in the working volume of the device.

The Brunauer–Emmett–Teller (BET) and Barrett-Joyner-Halenda (BJH) methods were realized with an automatic device BELSORP miniX (MicrotracBEL Corp., Osaka, Japan). The BET method is a method of mathematical description of physical adsorption, based on the theory of polymolecular (multilayer) adsorption. The BJH method is a method for calculating the pore size distribution in a porous material using adsorption or desorption isotherms. The method is based on the assumption that the pores are cylindrical in shape and that the pore radius is equal to the sum of the Kelvin radius and the thickness of the film adsorbed on the pore wall.

The particle size modulus of PGs was determined using standard methods according to the Russian Standard GOST 8736-2014.

The normal consistency, setting time, compressive, and flexural strength of gypsum binders were determined according to the Russian Standard GOST 23789-2018 and EN 13279-2:2004, NEQ “Gypsum binders. Test methods”.

The strength testing of the gypsum samples was carried out using a laboratory press with a load of 10 tons.

The synthesis of the PG binder was carried out using a laboratory oven with electronic temperature control. For the additional monitoring of air temperature and PG temperature in the oven, thermocouples connected to a Mastech MS6514 control and measuring device (Mastech Digital Inc., Pittsburgh, PA, USA) were used. The temperature of treatment for PG was 175 °C.

3. Results and Discussion

3.1. PGs Characteristics

3.1.1. Chemical and Elemental Composition of PGs

The analysis of the chemical composition (

Table 1. Content of oxides in the studied phosphogypsum and natural gypsum and the standard error (StdErr) of the device.

PGBel			PGBal			PGKin			NG
Oxide	%	StdErr%	Oxide	%	StdErr%	Oxide	%	StdErr %	Oxide	%	StdErr %
CaO	47.84	0.250	SO3	48.84	0.250	CaO	47.03	0.250	SO3	49.83	0.250
SO3	47.18	0.250	CaO	46.44	0.250	SO3	46.85	0.250	CaO	49.62	0.250
P2O5	1.87	0.070	SrO	2.520	0.080	P2O5	1.58	0.060	MgO	0.14	0.006
SiO2	1.18	0.050	P2O5	0.68	0.034	SiO2	1.27	0.060	SrO	0.12	0.006
F	1.04	0.180	CeO2	0.36	0.018	SrO	1.05	0.050	Other	0.30	–
SrO	0.28	0.014	PuO2	0.17	0.010	F	0.98	0.170
MgO	0.15	0.010	MgO	0.15	0.008	Al2O3	0.32	0.016
Al2O3	0.12	0.006	SiO2	0.13	0.009	CeO2	0.22	0.011
Other	0.35	–	Fe2O3	0.12	0.006	MgO	0.21	0.010
			Na2O	0.12	0.018	Na2O	0.10	0.019
			La2O3	0.12	0.009	Other	0.40	–
			Nd2O3	0.11	0.006
			Other	0.23	–

) and the elemental composition (

Table 2. Elemental composition of the studied phosphogypsum and natural gypsum and standard error (StdErr) of the device.

PGBel			PGBal			PGKin			NG
Element	%	StdErr%	Element	%	StdErr%	Element	%	StdErr %	Element	%	StdErr %
Ca	34.21	0.180	Sx	19.56	0.100	CaO	47.03	0.180	SO3	49.83	0.100
Sx	18.89	0.100	Ca	33.21	0.180	SO3	46.85	0.100	CaO	49.62	0.180
Px	0.82	0.030	Sr	2.130	0.070	P2O5	1.58	0.027	MgO	0.14	0.004
Si	0.55	0.025	Px	0.30	0.015	SiO2	1.27	0.026	SrO	0.12	0.005
F	1.04	0.180	Ce	0.29	0.015	SrO	1.05	0.043	Other	0.22	–
Sr	0.23	0.012	Pu	0.15	0.009	F	0.98	0.170
Mg	0.09	0.006	Mg	0.09	0.005	Al2O3	0.32	0.008
Al	0.06	0.003	Si	0.06	0.004	CeO2	0.22	0.009
Other	0.29	–	Fe	0.09	0.004	MgO	0.20	0.006
			Na	0.09	0.014	Na2O	0.10	0.014
			La	0.10	0.007	Other	0.33	–
			Nd	0.01	0.005
			Other	0.17	–

) of PG_Bal, PG_Bel, PG_Kin, and NG showed that the studied gypsum bearing components, regardless of origin, are represented by CaO and SO₃ oxides, with the highest content of Ca and S elements, respectively. All PGs contain phosphorus oxide from 0.684 to 1.870%, the amount of which is determined by the chemical composition of the original phosphate rock. The main difference between PGs and NG is the presence, in addition to the main CaO and SO₃ oxides (Table 1) and Ca and S elements, (Table 2), of a large number of additional components, which can subsequently affect the quality characteristics of the binders.

3.1.2. Mineral Composition of the PGs and NG

The XRD analysis of the studied PGs and NG demonstrated the presence of reflections responsible for the CaSO₄ · 2H₂O phase (d = 7.63, 4.28, 3.80, 3.07, 2.87, 2.69, 2.22, 2.08, 1.90 Å). The intensity of the peaks in the PG patterns varies slightly (Figure 1) among themselves, as well as in comparison with NG. However, it should be noted that PG_Kin has the lowest intensity of the main peak (d = 7.63).

3.1.3. DTA and TG-Analysis of the PGs and NG

According to DTA-analysis (Figure 2), basic phase transformations, such as the loss of adsorption water (at 100–120 °C); the loss of 0.5H₂O (at 220–240 °C); and the inversion of CaSO₄ (at 380–420 °C) are characteristic of NG.

In the case of PGs, the temperature ranges characteristic of the above phase transformations are differ. This is due to the presence of various impurities that can increase or decrease the temperature of phase transformations.

Phase transformations within approximately equal temperature intervals are typical for all PGs. In the DTA diagrams for all PGs and NG, there is a double endothermic effect. The first endothermic peak is associated with the removal of water crystallization during the dehydration of gypsum dihydrate; the second one is associated with the removal of the remaining moisture. However, it should be noted that in the case of PG_Bel (Figure 2), the removal of 1.5 of H₂O molecule occurs at a temperature of 154.4 °C, which is 5.2 °C lower than for PG_Bal (159.6 °C); PG_Kin occupies an intermediate position. For NG, this phase transformation occurs at 152 °C. The formation of soluble anhydrite for PGs occurs in the temperature range of 168.5–175.6 °C, and at 170.3 °C for NG.

Also, the DTA diagrams of PGs have an exothermic peak in the temperature range of 435.5–436 °C, which indicates a restructuring of the crystal lattice with the formation of insoluble anhydrite. In NG, the insoluble anhydrite is formed at lower temperatures: 353.6 °C.

The TG analysis of PGs and NG demonstrated an equal temperature of loss of the adsorbed water. Thus, it can be assumed that replacing NG with PG will not affect the energy consumption of the gypsum binder production process.

3.1.4. Study of the Surface Morphology of PGs and NG Particles

The surface morphology of the PGs particles largely depends on the production technology, the granulometry of the phosphate rock, the concentration of phosphoric acid in the reactor slurry (higher reaction rate, and, as a result, higher ion mobility) and the component ratio.

According to the data in the work presented in [9], there are five morphological types of PG:

(1)

The rhombic type is characterized by needle-shaped crystals with a uniform distribution of thin needles and elongated plates with sizes from 10 to 120 μm and crystallite size, normally from 400 to 550 nm;

(2)

The aggregate small rhombic type is characterized by small agglomerated needles with a size of 5–30 μm and a crystallite size from 70 to 300 nm;

(3)

The cluster type can be characterized as polycrystalline aggregates arranged in a disorderly manner. The crystals tend to form clumps, which are sometimes referred to as “pink sand”;

(4)

The aggregate short-needle type is characterized by the presence of small grains, which, normally, have a round, stone-like shape. This type is significantly different from other ones;

(5)

The needle type consists of very thin needle-shaped crystals.

The morphology of PGs particles must be taken into account when using it as a raw material for the production of gypsum binders and the materials based on them. For the production of binders and drywall, the most suitable material is PG, which belongs to morphological types 1, 2, and 5. In turn, the types 3 and 4 are mainly used in the cement industry (as an analogue of NG), and in agriculture as well [9].

The analysis of the morphology of the studied PGs showed that PG_Bel (Figure 3a) is represented mainly by monoclinic needle-shaped crystals. For PG_Bal (Figure 3b), the agglomerates of needles in the sizes of 100–400 μm are observed compared to smaller needle crystals for PG_Bel. At the same time, for PG_Kin (Figure 3c) there are conglomerates, which are stone-shaped grains with a highly developed surface, in addition to needle-shaped crystals.

Thus, morphological analysis showed that PG_Bel belongs to the rhombic type; PG_Bal is of the small rhombic type; PG_Kin is a combination of aggregate and short-needle types.

Despite the identical deposits of phosphate rocks, PG_Bel and PG_Kin have a significant morphological difference, which is associated with differences in the composition of raw materials, technologies, and equipment used [9].

A comparison of the morphology of NG (Figure 4) showed the presence of a dense, highly developed particle surface (compared to PGs), which is caused by mechanical stress during splitting.

3.1.5. Study of the Grain Composition and Granulometry of PGs

Visually, the PGs are a bulk material, consisting of small particles and their large associations, formed through the agglomeration of particles during storage under high humidity. Normally, under minor external loads, the destruction of conglomerates occurs.

Figure 5 shows the PG conglomerates selected from the fraction larger than 5 mm. The conglomerates have the following differences in morphology: PG_Bel is characterized by a flaky shape (Figure 5a), PG_Bal is characterized by a rounded shape with a rounded smooth surface (Figure 5b), and PG_Kin has a porous structure and an angular grain shape (Figure 5c).

The color of PGs varies from white to gray (Figure 6). This depends on the amount and type of impurities in the feedstock, as well as the quality of washing and storage conditions.

To determine the fineness modulus, sieving of the PGs was carried out on standard sieves with a mesh size of 0.14; 0.315; 0.63; 1.25; 2.5; 5 mm. Before a sieving test, PGs were dispersed in a plastic container on rollers for 5 min; rubber balls acted as the grinding bodies, which made it possible to reduce the amount of conglomerates without grinding.

It should be noted that in PG_Kin, the number of large conglomerates practically did not change, which indicates their high hardness and solidity (Figure 6c).

The grain distribution and fineness modulus of the PGs are given in

Table 3. Grain distribution and fineness modulus of the PGs.

PG ID	Sieve Residue Type	Sieve Residue, by wt.% Sieve Mesh Size, mm							Fineness Modulus
		5	2.5	1.25	0.63	0.315	0.16	>0.16
PGBel	Partial	8.6	3.9	4.7	7.1	3.9	14.5	57.3	1.28
	Total	8.6	12.5	17.2	24.3	28.2	42.7	100
PGBal	Partial	11.2	8.2	6.8	10	7.8	18.9	37.1	1.89
	Total	11.2	19.4	26.2	36.2	44	62.9	100
PGKin	Partial	24.6	13.8	8.5	7.8	8.5	14.7	22.0	2.72
	Total	24.6	38.4	46.9	45.7	63.2	77.9	99.9

. According to the presented data, PG_Kin has the largest fineness modulus (2.72), the smallest one is for PG_Bel (1.28); the fineness modulusfor PG_Bal has an intermediate value (1.89).

SSA measured using the air permeability method for PG_Bel, PG_Bal, and PG_Kin was 203.6 m²/kg; 183.9 m²/kg; and 163.5 m²/kg, respectively.

According to the data from the BET and BJH methods, the maximum SSA is typical for PG_Bel, the minimum SSA is for PG_Bal, and the SSA value for PG_Kin occupies an intermediate position. The analysis of NG showed lower total pore volume values compared to the PGs (

Table 4. SSA and the pore size distribution of PGS and NG.

Parameter	PGBel	PGBal	PGKin	NG
Air permeability method
SSA, m2/kg	203.6	183.9	163.5	–
BET method
SSA, m2/kg	18,155	15,773	17,458	–
Total pore volume (p/p0 = 0.9900), cm3/g	0.071	0.067	0.061	0.053
Average pore diameter, nm	15.64	16.98	13.92	14.76
BJH method
SSA, m2/kg	18,065	15,717	17172	–
Total pore volume, cm3/g	0.086	0.084	0.082	0.072
Average pore diameter, nm	19.03	21.40	19.17	19.29

).

Figure 7 demonstrates the curves of the PGs, representing that the pore volume for a specific diameter are multimodal, and the curve of NG is unimodal. The curve of PG_Kin is characterized by four peaks, the two most intense peaks correspond to pore sizes of 3.76 nm (the predominant pore size, see Table 5) and 2.22 nm. The two less intense peaks correspond to particle sizes of 1.93 nm and 2.48 nm. This is most likely due to the fact that PG_Kin consists of particles belonging to two different morphological types (Figure 8).

The PG_Bel curve is characterized by three peaks: the most intense one corresponds to a pore size of 3.78 nm (Figure 7, Table 5) and two less intensive ones, corresponding to the pore sizes of 2.36 nm and 3.36 nm. PG_Bal is characterized by two peaks, corresponding to the pore sizes of 3.76 nm and 2.50 nm. For the curve of NG there is only one peak, corresponding to a pore size of 2.20 nm.

It is assumed that the granulometry of the gypsum-bearing components, as well as the developed surface and porosity of the particles, will influence the amount of mixing water required to ensure the standard consistency of the gypsum binder and, as a result, the strength of the gypsum paste.

At this stage of research, it can be assumed that PG_Kin, which is characterized by a smaller pore volume and a larger particle size, is likely to have a lower water requirement to provide a normal gypsum binder consistency compared to PG_Bel and PG_Bal.

3.2. Study of Binder Characteristics

3.2.1. Study of the Morphology of PG Binders and NG Binder

PG binders were synthesized by the thermal treatment of PGs in a laboratory oven at 175 °C. The NG binder was used as the first reference.

The analysis of the morphology of PG binders showed that thermal treatment does not affect the shape of the particles. However, their surface became more loose and developed due to the detachment of 1.5 of H₂O molecules (Figure 8). It is difficult to judge the effect of thermal treatment on the morphology of particles of the NG binder due to the fact that they are formed by the destruction of the parent rock. At the same time, the surface of the particles of the NG binder has become looser and more developed, just like for PG binders.

Figure 8. Morphology of the binder based on: (a)—PG_Bel; (b)—PG_Bal; (c)—PG_Kin; (d)—NG.

Figure 8. Morphology of the binder based on: (a)—PG_Bel; (b)—PG_Bal; (c)—PG_Kin; (d)—NG.

3.2.2. SSA and Pore Size Distribution of the PG Binders and the NG Binder

The SSA, measured by the air permeability method for PG_Bel, PG_Bal, and PG_Kin binders, was 301.0 m²/kg, 241.8 m²/kg; and 192.3 m²/kg, respectively. Thus, the dehydration of PGs during heat treatment contributes to a slight increase in SSA.

The results of the BET and BJH methods demonstrated that the heat treatment of gypsum-bearing components leads to a decrease in SSA and total pore volume, an increase in the average pore diameter, and an increase in the predominant pore size of the binders (Table 5). At the same time, the most noticeable changes in the above parameters were observed for the PG_Kin binder.

Table 5. SSA and the pore size distribution of PG binders and the NG binder.

Parameter	Gypsum-Bearing Component
	PGBel	PGBal	PGKin	NG
Air permeability method
SSA, m2/kg	301.0	241.8	192.3	457.0
BET method
SSA, m2/kg	5374.9	3633.4	4019.1	4571.8
Total pore volume (p/p0 = 0.9900), cm3/g	0.040	0.040	0.040	0.038
Average pore diameter, nm	29.66	44.53	40.00	33.11
BJH method
SSA, m2/kg	3732.4	2984.5	3506.1	4157.9
Total pore volume, cm3/g	0.056	0.067	0.068	0.055
Average pore diameter, nm	44.00	90.17	77.93	53.14

Table 5. SSA and the pore size distribution of PG binders and the NG binder.

Parameter	Gypsum-Bearing Component
	PGBel	PGBal	PGKin	NG
Air permeability method
SSA, m2/kg	301.0	241.8	192.3	457.0
BET method
SSA, m2/kg	5374.9	3633.4	4019.1	4571.8
Total pore volume (p/p0 = 0.9900), cm3/g	0.040	0.040	0.040	0.038
Average pore diameter, nm	29.66	44.53	40.00	33.11
BJH method
SSA, m2/kg	3732.4	2984.5	3506.1	4157.9
Total pore volume, cm3/g	0.056	0.067	0.068	0.055
Average pore diameter, nm	44.00	90.17	77.93	53.14

Thus, thermal treatment significantly affects the distribution of pores, which is confirmed by a change in the curves reflecting the pore volume distribution (Figure 9), as well as a shift of the peaks towards pores of a larger diameter.

3.2.3. Setting Times and Physical Characteristics of the PG Binders and the NG Binder

Setting times, compressive and flexural strength, and average density were determined for the gypsum samples with dimensions of 40 × 40 × 160 mm at the age of 2 h after molding, as well as after the complete drying of the samples. To prepare the gypsum samples, binders with a normal consistency (W/S ratio) were used (

Table 6. Physical characteristics of the experimental binders.

Gypsum Component for Binder	W/S ratio	Setting Time, min		Average Density, kg/m3		Flexural Strength, MPa		Compressive Strength, MPa
		Initial	Final	after 2 h of Molding	after Complete Drying	after 2 h of Molding	after Complete Drying	after 2 h of Molding	after Complete Drying
PGBel	0.92	17	22	1461	908	1,.85	2.47	1.72	4.67
PGBal	0.88	11	19	1457	935	1.90	2.55	2.23	5.29
PGKin	0.70	15	25	1494	1056	1.91	3.35	3.38	8.14
NG	0.70	18	25	1570	1120	3.11	5.74	4.76	11.96

).

The analysis of the results in Table 6 showed that to achieve a normal consistency of gypsum binder, the largest amount of water is required for PG_Bel (W/S = 0.92); the least amount of water (W/S = 0.7) is required for the NG binder and the PGB_Kin binder, as well. The higher W/S ratio for the PG_Bel binder results in a longer initial setting time (17 min), compared to the PG_Kin binder (15 min). However, the PG_Bal binder with a high W/S ratio has the shortest initial setting time (11 min). This is most likely due to the different quality of PG washing and, in particular, the higher content of orthophosphoric acid residues in PG_Bal, which, according to the literature sources [24], accelerates the initial setting time for the gypsum binder. Interestingly, the studied PG binders are characterized by a faster initial setting time in relation to the NG binder. This is probably also due to the presence of impurities in the PGs that affect the rate of dissolution and precipitation of hydration products.

The average density and strength for the studied PG binders is significantly lower than NG binder both 2 h after molding and after the samples have completely dried. At the same time, the gypsum samples made from the PG_Bel binder have the minimum values of average density and strength, which is due to the high W/S ratio.

When testing these samples at 2 h after molding under a compressive load, moisture was observed on the surface. This suggests that the morphology of the PG_Bel binder and its high SSA initiate its high water demand, which is not used for hydration, but to ensure the normal consistency of gypsum binder. This phenomenon, in turn, negatively affects both the average density and the strength of the gypsum samples.

Despite the equal W/S ratios for the PG_Kin binder and NG binder, the PG_Kin binder has a lower average density and strength, which is likely due to the presence of impurities in the PG, which can affect not only the shape and size of crystals, but also the adhesion between crystals, and increase the total porosity of the gypsum samples which aligns with the data presented in [24].

To study the influence of the W/S ratio on the physical characteristics of the gypsum samples, the cubes with a size of 30 × 30 × 30 mm were molded at the following W/S ratios: 0.5, 0.6, 0.7, 0.8, and 0.9 (

Table 7. Dependence of the average density of gypsum paste on the W/S ratio and binder type.

Gypsum Component for the Binder	Average Density for Completely Dry Binders at Different W/S Ratios, kg/m3
	0.5	0.6	0.7	0.8	0.9
PGBel	1332	1246	1105	1014	944
PGBal	1342	1233	1157	1053	927
PGKin	1321	1202	1095	1062	960
NG	1384	1255	1170	1054	961

and

Table 8. Dependence of compressive strength of gypsum paste on W/S ratio and binder type.

Gypsum Component for the Binder	Compressive Strength for Completely Dry Binders at Different W/S Ratios, MPa
	0.5	0.6	0.7	0.8	0.9
PGBel	16.89	13.96	10.42	8.92	7.57
PGBal	18.24	15.42	12.21	8.84	7.00
PGKin	16.94	13.56	10.90	8.22	7.11
NG	16.86	13.96	11.90	9.34	7.50

).

The analysis of the results (Table 7 and Table 8) showed that the average density and compressive strength of completely dry samples gradually decreased with the increasing W/S ratio. With an equal W/S ratio, the values of the average densities do not differ significantly (Table 7), and the values of compressive strength differ significantly depending on the W/S ratio (Table 8).

The compressive strength of the NG samples is higher than the compressive strength of the studied PG binders only at W/S = 0.8. At W/S ratios of 0.5, 0.6, and 0.7, the compressive strength of PG_Bel and PG_Kin is comparable, but lower than for PG_Bal. When the W/S is <0.7, the compressive strength is higher or comparable to the strength of the PG-based samples. At equal W/G ratios, the strength of samples synthesized from PG_Bel and PG_Kin binders is practically the same. The highest compressive strength at W/G 0.7, 0.6, 0.5 is typical for the PG_Bal binder. It was previously established that PG_Bal at high W/S ratios (Table 6) has a shorter period of initial setting time than other binders. This indicates an earlier structure formation, which at lower W/S ratios can provide higher strengths than other binders.

It is likely that by using various chemical and mechanical methods to reduce the W/S ratio in PG binders, their physical characteristics may be comparable to or higher than NG binders.

3.2.4. Morphology of New Formations in Gypsum Binders

The SEM analysis of the binders with W/S = 0.6 showed differences in their morphology and the density of new formations. The PG_Bal binder (Figure 10b) and NG binder (Figure 10d) are characterized by needle-shaped elongated crystals. New formations in PG_Bel and PG_Kin binders (Figure 10c) have a prismatic reinforced morphology. The differences in the morphology and size of new formations for the PG binders (Figure 10a–c) can be associated with different contents and types of impurities in PG, which, according to the work presented in [24], have a significant effect on the rate of dissolution and the precipitation of hydration products. In turn, the amount and type of impurities depend on the original phosphate rock, the production technology of orthophosphoric acid, the quality of washing, and the storage conditions of the PG.

4. Conclusions

A comprehensive study of PG from three different industrial enterprises and its comparison with NG allowed us to obtain the following conclusions:

-

all studied PGs can be used as an alternative resource of raw materials to NG when the gypsum binders are synthesized;

-

despite the use of the same raw materials for the production of PG_Bel and PG_Kin, differences in the technological characteristics significantly affect the physical characteristics of PGs and PG materials;

-

SSA and pore size distribution for the PG binders are the main factors ensuring a high water demand, and as a result, less satisfactory physical characteristics;

-

to increase the efficiency of PGs use as a raw material for the production of PG-based materials, it is necessary to design measures aimed at reducing its water demand, for example, the use of grinding processes and the introduction of superplasticizers.