Cost-Effective Treatment of Hemihydrate Phosphogypsum and Phosphorous Slag as Cemented Paste Backfill Material for Underground Mine

School of Resources Environment and Safety Engineering, University of South China, Hengyang 421001, China School of Resources and Environmental Engineering, Jiangxi University of Science and Technology, Ganzhou 341000, China School of Resources and Safety Engineering, Central South University, Changsha 410083, China Changsha Design and Research Institute of Nonferrous Metallurgical Co., Ltd, Changsha 410001, China


Introduction
Phosphogypsum (PG) is a main by-product of wet-process phosphoric acid.PG can be classified into dihydrate phosphogypsum (DPG) and hemihydrate phosphogypsum (HPG) according to different processing methods [1,2].DPG, the main phase of which is CaSO 4 • 2H 2 O, cannot form a solidified body without heat treatment.In contrast, HPG is mainly composed of CaSO 4 • 0.5H 2 O, which presents better self-gelling properties during CaSO 4 • 0.5H 2 O conversion to CaSO 4 • 2H 2 O [3].Phosphorus slag (PS) is also a by-product of yellow phosphor production via the electric furnace method and is mainly composed of CaSiO 3 , which contains active SiO 2 and Al 2 O 3 and has weak gelling properties [4,5].Besides, PG and PS contain a small amount of harmful impurities such as phosphorus, fluorine, and heavy metals, which limit its comprehensive utilization [1,6,7].
At present, over 700 million tonnes of PG and PS have been discharged in China and approximately 8000 hectares of land have been occupied [2,[8][9][10].Moreover, the total emissions of phosphorus chemical wastes are still increasing at a rate of 70-90 million tonnes per year [10,11].e recycling rate of PG and PS in China is less than 20%, which is mainly reused as building material [7], cementitious material [12,13], road base material [14], soil improvement material [15], etc. e remaining 80% is directly discharged into an open-air yard without any treatment [16,17].However, the P, F, and heavy metals in the PG and PS are released under the long-term repeated action of natural rainfall and sunlight, causing serious pollution to the surface water and soil [2,18].In addition, a large number of goafs are formed in the long-term mining of phosphate rock, which seriously threatens the surface environment and mining operation safety [19,20].With the rapid development of mineral processing, the problems caused by the accumulation of PG, PS, and goafs are very prominent and seriously restrict the healthy and rapid development of the phosphorus chemical industry in China.
Cemented paste backfill (CPB) is an ideal solution for the treatment of solid waste and goafs based on solid waste solidification/stabilization technology and has been widely used in the global mining industry [21][22][23][24].In order to improve the treatment of goafs and the utilization of PG and PS, Li et al. [25] and Wang et al. [26] developed a cemented backfill technology based on DPG and cement, and this technology was successfully applied to the Kaiyang phosphate mine in Guizhou province, China.However, the consumption of cement was high, and the early uniaxial compressive strength (UCS) of the backfill body was very low.Chen et al. [27], Liu [28], and Dang et al. [29] found that the PS presented better self-gelling characteristics under the excitation of CaO and NaOH and indicated that the PS could replace the cement in CPB-based DPG and cement, which greatly reduced the cost of backfill.Additionally, Li et al. [30] and Chen et al. [31] analyzed the underground environmental impact of soluble P, F, and heavy metals released from the backfill body based on PG through static and dynamic leaching tests and found that CPB could solidify P, F, and heavy metals.ese efforts not only cut down the cost of backfill but also increased the recovery rate of phosphate resources.On the other hand, the utilization proportion of DPG was been greatly improved.
Original HPG is not an ideal backfill material by itself due to its strong acidity, and the large amounts of harmful gases such as H 2 S and SO 2 are generated in the filling process [32].HPG modified by quicklime not only greatly improves the early strength of the backfill body but also eliminates the gases generated during the filling process [32].However, the strength loss coefficient reaches 25% curing for 240 days in a humid environment [3].
e long-term stability of the strength of CPB is the key to effectively control the pressure of goafs, which means that the strength stability of CPBbased HPG must be verified in the subsequent tests.In addition, the recycling PS is also a topic of global concern, and it is necessary to develop a backfill method for the low cost and safe disposal of PS and HPG, with the aim of further expanding the application range of HPG and PS.
is article attempts to develop a new solution for CPB technology-based HPG and PS. e physical and chemical properties of raw materials were first analyzed in detail.e UCS, slump value, bleeding ratio, generated gases, and microstructural features of CPB were then evaluated.en, the underground environmental effects of P, F, and heavy metals released from the backfill body were investigated through a dynamic leachability experiment.Finally, an engineering application of CPB-based HPG-PS is evaluated in detail.

Raw Materials.
e main materials used in this study were HPG, PS, quicklime, and Portland cement, of which HPG and PS were, respectively taken from the waste dam of a phosphate fertilizer plant and a yellow phosphorus plant in Guizhou province, China.
(1) e HPG had a moisture content and pH value of 12% and 3.1, respectively.e main chemical components of HPG are shown in Table 1.e X-ray diffraction (XRD) of HPG is shown in Figure 1(a).
e surface of HPG particles was coarse, and the deposits were impurities such as phosphide and fluoride [34], as shown in Figure 1(b).
e particle-size distribution of HPG is shown in Figure 2(a), for which the particle size of less than 50 μm accounts for 45.34%.
(2) e PS had a moisture content and pH value of 8% and 7.8, respectively.e main chemical components of PS are shown in Table 1.According to the Chinese Standard of Technical Specification for Phosphorous Slag Powder Use in Hydraulic Concrete (DL/T 5387-2007), the basicity coefficient (M o ), the mass coefficient (K), and the activity coefficient (M n ) were 1.18, 1.24, and 0.05, respectively.us, PS was characterized by weak gelling activity.e particle-size distribution of PS after grinding is shown in Figure 2(b), for which the particle size of less than 50 μm accounts for 17.83%.
(3) e CaO content of quicklime powder in this paper was greater than 92%.e cement used in this study was ordinary Portland cement (P.O 32.5).

Mixing Procedures.
To initially determine the range of mass fractions of backfill slurry satisfied for pipeline transportation, which were subjected to the limitations of the test materials, the auxiliary tests were first carried out.e results showed that the slurry was in the form of a paste state when the mass fractions of CPB-based HPG-PS were 66%-69%, as shown in Figures 3(a) and 3(b).Large amounts of pungent odor gases were generated during the test, and the surface of the backfill body specimen was covered with a large number of pores after being demolded, as shown in Figure 3(c).If the paste was directly filled into the goafs, these gases would cause serious pollution to the underground air.erefore, the HPG and PS mixtures should not be directly used as backfill materials before modification.
According to the results of the auxiliary tests, 24 mix proportions were designed, as shown in Table 2.

2
Advances in Materials Science and Engineering Specimens S1-S15 were used to investigate the effects of the HPG/PS ratio and the quicklime content on backfill characteristics of CPB.Specimens S16-S24 were used mainly to study the strength stability of CPB over a long time period.
Based on these experiment schemes, the HPG, PS, quicklime, cement, and water were first weighed separately, and then those materials were thoroughly mixed and stirred for 5 minutes by the electric mixer.Secondly, the uniformly stirred slurry was poured into the standard triunit models with a size of 70.7 × 70.7 × 70.7 mm 3 [35], and then, the slurry was compacted with a stir bar and the mold was vibrated to keep the slurry higher than the top surface 6-8 mm of the mold.While the surface moisture was slightly dried, higher slurry is scraped from the mold to keep the surface smooth.When the slurry was completely consolidated, the mold was removed and the specimens were    Advances in Materials Science and Engineering numbered.Finally, the specimens were placed in a curing box with a temperature of 20 °C and relative humidity of 95%.

Harmful Gas Detection.
e chemical compositions of HPG and PG showed that a small amount of free ions such as S 2− , SO 2− 3 , CO 2− 3 , PO 3− 3 , HPO 2− 3 , and F − might be present in the slurry, which reacted with the alkaline species in the PS to generate CO 2 , H 3 P, H 2 S, and SO 2 .e gases generated during the pulping process were detected by the method designed in a previous study [32].

Fluidity Tests.
e use of a minicylinder to test the slump value is simple and accurate and can replace the ASTM standard slump test [36,37].It has been proven that backfill slurry with a slump value of 15-23 mm can be pumped to a stope successfully and 29-36 mm can be transported by gravity flowing [38].To achieve better backfill application, the slump value was measured by a minicylinder with a height and diameter of 50 and 60 mm, respectively.e detailed test procedure can be found in a previous study [20].

Bleeding Tests.
e bleeding rate is the main parameter that reflects the shrinkage characteristics of the backfill body [39].In order to ensure completely roof-contacted filling in the stope, the bleeding rate should be less than 5%.In this test, the well-stirred slurry was poured into a 1000 ml cylinder to a position of 600 ml and stirred with a glass rod until the surface was free of bubbles, after which the cylinder was gently shaken and sealed.en, the total mass of the slurry and the change of the sedimentation layer were recorded.When the sediment level ceased to change, the volume of the supernatant was read.Finally, the bleeding rate was calculated by the ratio of the quality of the bleeding to the total mass of the slurry.

UCS Tests.
e UCS was tested for S1-S15 which had only been cured for 7 and 28 days, whereas S16-S24 had been cured for 3, 7, 28, 60, 120, and 180 days [40].e UCS of specimens was tested by the WDW-2000 rigid hydraulic pressure servo machine (Ruite, Guilin, China).In order to ensure the accuracy of the test, three specimens were tested for each group.

Microstructural Analysis.
After the UCS test, small specimen fragments were used for microstructure tests.Microstructural tests were carried out through commissioned tests using a scanning electron microscopy (Quanta 200, FEI, USA) in Central South University Test Center.

Dynamic Leachability Tests.
To investigate the underground environmental impact of P, F, and heavy metals in CPB, the horizontal oscillating method (HJ 557-2010) (HVM) was used to investigate the dynamic leaching  4 Advances in Materials Science and Engineering behavior of original HPG, PS, and specimens cured for 28 days with the recommended parameters [41].en, the results were compared with Categories III and IV water quality standards of the Chinese Groundwater Quality Standard (DZ/T 0290-2015).
e methods of chemical measurement were presented in a previous study [30].

Effect of Quicklime Content on CPB.
e effects of the quicklime content on the bubbles and strength of the backfill body are shown in Figure 4. e backfill slurry was weakly acidic when no quicklime was added.With the increase of the quicklime content, the gas generated in the backfill slurry gradually decreased, and the pH value gradually increased.However, the gases were not generated when quicklime was added to 3%, and the pH value of the slurry was 7.5-7.8.erefore, in order to reduce the cost, 3% of quicklime could be added.e generated gas was confirmed to be a mixture of H 2 S, SO 2 , and CO 2 .Because of the presence of free ions such as SO 2− 3 , CO 2− 3 , and S 2− in the backfill slurry, these ions reacted with H + to form H 2 S, SO 2 , CO 2 , and H 2 O.In addition, the neutralization reaction of free SO 2− 3 , CO 2− 3 , S 2− , and F − with Ca(OH) 2 in the slurry and the formation of insoluble compounds of CaSO 4 , Ca 3 (PO 4 ) 2 , Ca(H 2 PO 4 ) 2 , and CaF 2 resulted in an increase in the pH value of the slurry.e reactions were as follows: (2) As shown in Figure 4, when the quicklime was added from 0% to 1%, the UCS of S4 and S10 cured for 28 days was greatly decreased from 1.63 to 0.8 MPa and from 2.32 to 1.1 MPa, respectively.is indicated that HPG was sensitive to the pH value of the hydration environment, and a weakly acidic environment was beneficial to CaSO 4 • 0.5H 2 O converting to CaSO 4 • 2H 2 O (Equation 4).When the quicklime content increased from 1% to 3%, the UCS of S4 and S10 cured for 28 days increased from 0.8 to 1.10 MPa and from 1.02 to 1.35 MPa, respectively, and continued to increase to 5%, while the UCS decreased slightly.erefore, the addition of 3% quicklime can achieve better strength properties for the CPB.

Fluidity of Backfill Slurry.
As shown in Figure 5(a), when the quicklime was added from 0% to 5%, the slump value of the backfill slurry did not change significantly.erefore, the quicklime only neutralized the free acid ions in the CPB, improved the pH value of the gel system, and had no substantial effect on the fluidity of the backfill slurry.As shown in Figure 5(b), the slump value of the slurry was between 13 and 38 mm, which decreased from 18.92%-53.57%with the slurry mass fraction increasing from 66% to 69%.When the HPG content increased from 78.13% to 90.79%, the slump value decreased from 24.32%-56.67%.
is indicated that the slump value was very sensitive to mass fraction and HPG content.Due to the increase in the HPG content, water consumption increased during the CaSO 4 • 0.5H 2 O conversion to CaSO 4 • 2H 2 O, so the free water in the slurry was reduced.In order to satisfy the requirements of pumping or gravity flowing for the backfill slurry, the slump value should be controlled between 15 and 36 mm.erefore, HPG/PS ratios were determined to be between 6 and 10.

Bleeding of Backfill Slurry. Figure 6 presents the results
of the bleeding rate tests of specimens with quicklime of 3% and different HPG/PS ratios.e bleeding rate decreased from 41.43% to 56.04% when the mass fraction increased from 66% to 69% and decreased from 35.00% to 51.22% with the HPG content increasing from 78.13% to 90.79%.erefore, the bleeding rate of the slurry in the tests was less than 5% [39], which indicated that the shrinkage of the backfill body was small and could meet the requirements of engineering applications.

Effect of HPG Content on the Strength of Backfill Body.
As shown in Figure 7, when the mass fraction of the backfill slurry with the addition of 3% quicklime increased from 66% to 69%, the UCS of the backfill body cured for 7 and 28 days increased by 33.84%-78.57%and 51.56%-74.51%,respectively.e UCS increased with the increase of the HPG content, and the growth rate continued.e main reason for this was that the increase in the HPG content led to a significant increase in the amount of CaSO 4 • 2H 2 O converted from CaSO 4 • 0.5H 2 O in the gelling system, which indicated that HPG played a major role in the growth of UCS.When HPG/PS ratios were less than 6, the UCS of backfill bodies cured for 7 and 28 days was 0.42-1.10MPa and 1.02-2.25 MPa, respectively, while their UCS cured for 7 and 28 days with HPG/PS ratios of greater than 6 was 0.8-2.65 MPa and 1.35-3.88MPa, respectively.In order to meet the requirements of the upward slicing mining method and maximize the utilization of HPG, the UCS of backfill bodies cured for 7 and 28 days should be higher than 1 MPa.erefore, the mass fraction of CPB should be 69%, and HPG/PS ratios should be more than 6.

Long-Term Strength Characteristics of Backfill Body.
As shown in Figure 8, before 28 days, the UCS of S16-S19 rapidly increased, whereas the UCS remained constant with curing from 28 to 60 days.After 60 days, the UCS began to decrease and decreased by 18.08%-35.78%until curing for 180 days, and the range decreased with the increase of the HPG content. is phenomenon basically agreed with the previous studies [3,42], which was very unfavorable for controlling the underground pressure of the stope.In addition, when 5% Portland cement was added to the HPG-PS, the UCS of S21-S24 did not change significantly before Advances in Materials Science and Engineering 28 days.After 28 days, the UCS slowly increased and maintained long-term stability, which indicated that the addition of 5% Portland cement could effectively prevent the softening of the backfill body during the later curing stage.erefore, the recommended specimens were S21-S23, and their UCS cured for 7 and 28 days was 1.15-2.10MPa and 2.35-3.32MPa, respectively.3 shows the concentrations of P, F, and heavy metals in the leachate from HPG, PS, S22, and S23. e concentrations of P, F, and heavy metals in the leachate of S22 and S23 cured for 28 days were significantly lower than those of HPG and PS, and the concentration of P decreased from 3528 mg/L in the HPG to 0.38-0.62mg/L in the backfill bodies, which indicated that the gelling system-based HPG-PS had a consolidation effect on P, F, and heavy metals [41,43].In addition, the concentrations of P and F in HPG leachate exceeded Category IV of DZ/T 0290-2015, and the concentrations of Fe, Mn, and Pb exceeded Category III.Moreover, the concentration of P in the leachate of S22 was between Categories III and IV of DZ/T 0290-2015, and the concentrations of the other components could meet the requirements of Category III.However, due to the large   Advances in Materials Science and Engineering inrush of underground water in mines and the highly developed groundwater runoff, P, F, and heavy metals were gradually diluted and transformed during the migration process, and the concentrations of the components in the practical engineering applications would be significantly reduced on the basis of the test.erefore, the CPB-based HPG-PS hardly affected the underground environment.Advances in Materials Science and Engineering e above results indicated that the gelation essence of HPG-PS resulted from the hydration of HG in the HPG and active SiO 2 and Al 2 O 3 in the PS, and the hydration of HG was significantly faster than that of PS. e hydration of HPG was dominated at the early curing stage, whereas the hydration of active SiO 2 and Al 2 O 3 in PS mainly occurred in the late curing stages.

Effect of Soluble P and F on the Hydration of HPG-PS.
e soluble P and F had significant effects on the hydration of HPG and PS [44].e soluble P and F in the HPG and PS dissolved quickly in water and reacted with the Ca(OH) 2 , as shown in equations ( 2) and (3); then, insoluble substances such as Ca(PO 4 ) 2 and CaF 2 were formed.ese substances, adhering to the surfaces of the HPG and PS particles, increased the water impermeability of the formed alkaline membrane on the surface of the particles and retarded the hydration of HPG and PS.In addition, partial Ca 3 (PO 4 ) 2 and CaF 2 gradually covered the surface of the particles and formed Ca 10 (PO 4 ) 6 (OH) 2 (hydroxyapatite), which further reduced the permeability of the semipermeable membrane [45].
erefore, the continuous release process of active substances in PS, HPG, and cement was suppressed.When the soluble P and F were completely converted, the hydration process was gradually restored; so, it did not have a significant effect on the hydration in the later curing stage.

Application Evaluation
A phosphate mine in Guizhou province, China, adopted an upward horizontal slicing filling method with a production capacity of 700,000 t/a and a filling capacity of 100 m 3 /h.HPG and PS are the main by-products of phosphoric acid and yellow phosphorus productions around this mine, with a wide range of sources, no payment, and short transport distances.Based on the recommended parameters, the amount of cement is greatly reduced in the CPB.e total cost of the backfill materials is 3.32 USD/m 3 , which is only 29%-41% of CPB-based DPG-cement [26,27] and 56% of CPB-based total tailings [35].If an effective filling rate of 90% is taken into account as a rough calculation, then there will be a consumption of 300,000 tonnes of HPG and 50,000 tonnes of PS for CPB per year.e engineering application proves that CPB-based HPG-PS is feasible and can be used in mines with similar situations. is technology has the advantages of simple preparation process, low cost, safety, and environment-friendly and can greatly increase the recovery of phosphate resources and the recycling of HPG and PS.

Conclusions
In this study, we investigated the application characteristics of recycling HPG and PS as cemented backfill materials in CPB under the modification of quicklime and cement by fluidity tests, gas detection, UCS tests, and scanning electron microscopy (SEM).On the basis of the results and discussion, the following conclusions were summarized: (1) e addition of quicklime with 3% of HPG can eliminate CO 2 , H 2 S, and SO 2 generated in the pulping of CPB. e slump value decreases with the increase of HPG content and mass fraction.e slump values range between 15 mm and 35 mm when the HPG/PS ratios reach 6-10, which can meet the requirements for pipeline transportation of the paste.
(2) e UCS increases as HPG content and mass fraction increase.e losses in strength of CPB cured for 60-180 days are up to 18.08%-35.78%with addition of 3% quicklime.e strength of CPB can maintain stability for a long time period while adding 3% quicklime and 5% cement.e UCS of recommended specimens cured for 7 and 28 days is 1.15-2.10MPa and 2.35-3.32MPa, respectively, and the combined materials' cost is only 3.32 USD/m 3 .(3) e strength of CPB is the resultant of the combined hydration of HPG and PS. e soluble P and F delay the hydration of HPG and cement at the early curing stage, but having insignificant effect on the hydration of PS at the later curing stage.(4) e pollutant indexes of the leachates meet Category IV of DZ/T 0290-2015.e HPG-PS materials have a certain consolidation action on P, F, and heavy metals, whereas the curing mechanism of the HPG-PS gelling system on P, F, and heavy metals needs to be further studied in the next work.

Figure 2 :
Figure 2: Curve of particle-size distribution of HPG (a) and PS (b).

FIGURE 4 :
FIGURE 4: Effects of quicklime content on the bubbles and UCS of CPB.(a) HPG/PS ratio of 4 and mass fraction of 66%; (b) HPG/PS ratio of 6 and mass fraction of 66%.

Figure 5 :
Figure 5: Effects of quicklime content and HPG content on the slump value of backfill slurry.(a) Mass fraction of 66%; (b) cement content of 3%.

Figure 6 :
Figure 6: Effect of HPG content on the bleeding rate of backfill slurry.Cement content is 3%.

Figure 7 :
Figure 7: Effect of HPG content on the UCS of backfill body with the addition of 3% quicklime.(a) Mass fraction of 66%; (b) mass fraction of 69%.

)
After 28 days, the particles were completely covered by hydration products such as CaSO 4• 2H 2 O whiskers, Ca(OH) 2 crystals, C-S-H gel, and AFt crystals.e CaSO 4 • 2H 2 O whiskers were continuously thickened and smoothed, and the needle-like AFt gradually increased, as shown in Figures9(b) and 9(d).In addition, the large pores were filled with CaSO 4 • 2H 2 O whiskers, C-S-H gel, AFt, and Ca(OH) 2 crystals.e bond between the hydration products and the particles became dense, but a large number of pores still existed compared with Figures9(a) and 9(c).Consequently, this indicated that the hydration of HG in HPG had been basically completed, but the active SiO 2 and Al 2 O 3 in PS were still slowly released and the hydration continued.

Figure 8 :
Figure 8: Long-term strength characteristics of the backfill bodies.

Table 1 :
Main chemical compositions of HPG and PS (wt.%).Al 2 O 3 SO 3 SiO 2 P 2 O 5total P 2 O 5water MgO K 2 O Na 2 O F total F water Others
*e quicklime and cement were added according to the HPG quality.
Al 2 O 3 • 3CaSO 4 • 32H 2 O) are formed at the edges and between the particles.erefore,thisshowed that HG in the HPG had undergone a large number of hydration, as shown in Equation (4), and partially active SiO 2 and Al 2 O 3 in the PS had also undergone hydration, as shown in the following equations: 3.7.Cementitious Mechanism 3.7.1.Microstructure Analysis.After 7 days, particulate substances such as Ca 3 (PO 4 ) 2 , Ca(H 2 PO 4 ) 2 , and CaF 2

Table 3 :
Leaching elements in the HPG, PS, and backfill bodies by HVM.