Modification of Lime-Fly Ash-Crushed Stone with Phosphogypsum for Road Base

In order to increase the recycling of phosphogypsum waste, this study explored the feasibility of using phosphogypsum to replace some of the lime and aggregate in the lime-fly ash-crushed stone mixture which is a widely used road base material in China. For this purpose, compaction, compressive strength, composition structures, wetting-drying cycle tests, and shrinkage tests were carried out on the lime-fly ash-phosphogypsum-crushed stone composite to investigate its performance. +e results indicate that lime-fly ash-crushed stone modified with phosphogypsum has the required strength of the road base material and favourable performances in environment (wetting-drying cycle) stability. +e image processing analysis and shrinkage tests demonstrated that phosphogypsum can significantly improve the compactness and shrinkage performance of lime-fly ash-crushed stone mixture. A suitable content of phosphogypsum and a reasonable content of fine aggregate are conducive to improving the roadway engineering properties (i.e., decreasing shrinkage cracks and increasing compressive strength) of lime-fly ash-phosphogypsum-crushed stone composites.


Introduction
Phosphogypsum (PG) a by-product from the phosphoric acid industry is a remarkable solid waste, and about 4.5-5.0 tonnes of dry-based PG is outputted per ton of phosphoric acid (as P 2 O 5 ) recovered [1]. Approximately 20 million tonnes of PG is generated in China every year, and its utilization ratio is not more than 10% [2]. A large amount of it is normally discarded to the environment without any treatment leading to environmental contamination. It has been well documented that PG is an effective modification material for stabilizing special soil and improving the base (or subgrade) behaviors in roadway engineering [2][3][4][5][6]. Studies have been shown that PG alone is not sufficient for road construction. However, it can hasten the pozzuolana reactions between the fly ash (FA) and lime, and the further reaction generates calcium sulfoaluminate hydrate (generally called ettringite) which can bring on a slight expansivity [2,4,7]. erefore, PG is usually used together with FA and lime to increase strength and reduce shrinkage cracks of road base (or subbase) materials.
For lime-FA-PG binder which is usually applied as subbase material binder, the content of lime is generally 6%-12%, and the content of the ratio of PG and FA is usually controlled between 1 : 1-1 : 4 [2]. Meanwhile, the content of PG should not be very high because of the expansivity of ettringite [8][9][10]. However, most of these studies focused on the modification of PG for lime-FA binder or lime-FAstabilized soils. In this paper, the properties of lime-FAcrushed stone (CS) mixture modified with PG are investigated. Lime-FA-CS (or gravel) is a widely used road base material in China [11,12], and it has many advantages such as steady strength development, simple construction, and the utilization of waste FA, but unfortunately the shrinkage cracks often occur, and the early strength of it is very slow. ese disadvantages baffle it from being utilized widely [11,13,14]. It should be known that the mix ratio of binder and CS (or gravel) is also another factor affecting the mixture's strength and anticrack properties. If the content of binder is very low, the strength of mixture will be very low; otherwise, the mixture will easily shrink and crack. e content of aggregate (CS or gravel) is generally 75%-80% for dense skeleton-type mixture, and appropriate amount of fine aggregate (<2.36 mm) is beneficial to improve the roadway engineering properties (i.e., decreasing shrinkage cracks and increasing compressive strength) of mixture [13,15,16].
In this study, PG was used to replace some amount of the lime and aggregate in the typical lime-FA-CS mixture without changing the dense skeleton structure. e properties and performance of lime-FA-PG-CS composite used as road base material were investigated to prepare it for the wider application. e CS used in this study was collected from a roadway construction site in Nanjing, China. e particle-size distribution of CS is decided by dense skeleton structure and satisfies the demand of Chinese highway road base specifications JTG/T F20-2015 [17] (Figure 1).

Test Method.
e testing mixture of lime-FA-CS and lime-FA-PG-CS were prepared separately, for comparative analysis, and the content of reactive CaO (5.4%-5.7%) in these two materials was approximately the same. e mix composition of lime-FA-CS is as follows: lime 8%, FA 12%, and CS 80%, in which the fine aggregate ranged from 0.06 mm to 2.36 mm in size accounting for 8% of total mass. PG is used to replace part of lime and CS in the lime-FA-CS mixture to prepare the lime-FA-PG-CS mixture containing the following composition: lime 6%, FA 12%, PG 6%, and CS 76%. e effect of fine aggregate was also investigated, and its content was varied from 5%, 8%, to 12%.
After storing in a sealed plastic bag for 4 hours, the mixture was compacted into a cylinder mold with the size of ø 150 × 150 mm 3 to prepare test specimens. e compaction degree of the specimen was 98%. e compaction degree was obtained by dividing the specimen's dry density with the mixture's maximum dry density. e maximum dry density was determined from the modified Proctor compaction test. e modified Proctor compaction tests of lime-FA-CS and lime-FA-PG-CS were carried out according to the test specification JTG E51-2009 of Chinese Ministry of Transport [18], which is quite similar to AASHTO T180/ASTM D1557 [19,20]. e unconfined compressive strength tests were also conducted in accordance with the test specification JTG E51-2009 [18,21]. e specimens of unconfined compressive strength test for lime-FA-CS mixture and lime-FA-PG-CS mixture were sealed in a plastic bag and stored in a curing room with a relative humility above 95% and at a constant ambient temperature of 20 ± 2°C. Considering the influence of the curing condition for the specimens, two different curing methods were performed: standard curing and soaking curing. In standard curing, the specimen was cured in the curing room for the designed curing period (e.g., 7 days or 28 days) and was immersed in water for 24 hours before the strength test [9]. In soaking curing, the specimen was cured for 7 days in the curing room and then immersed in water for 21 days. Meanwhile, the stability of the mixtures was also investigated by dry-wet cycle tests. Before the drywet test, the specimens were cured 7 days in the curing room. en, the specimens were taken and placed in an oven at 60°C for 12 hours, followed by immersion in the water for 1 day at an ambient temperature of 25°C. ese two steps were defined as one dry-wet cycle. After 6 successive dry-wet cycles, the strength test was performed. For each data point, six specimens were tested [22]. e composition structures of the mixture samples were studied with computed tomography (CT) and image processing analysis [23][24][25]. As shown in Figure 2, four sections of the 28-day standard cured strength test specimen were scanned by CT equipment. en, the CT images were adjusted to grayscale mode and converted to binary segmentation by maximum variance, and then the independent crushed stone image and pore image were obtained by contour tracking programs; finally, the pore area ratio was measured and calculated.
In order to reveal the anticracking property of PGmodified mixture as roadway material, shrinkage tests were also carried out. e shrinkage tests were performed conforming to the procedure of a shrinkage test in JTG E51-2009 [18].
e specimens for shrinkage test were prisms of 100 mm × 100 mm × 400 mm. After the specimens were cured in the curing room for 7 days, comparator probes with dial indicator were set on the specimen end surfaces, and then specimens were placed in a testing room with a relative humility of 60 ± 5% and at an ambient temperature of 20 ± 2°C (Figure 3). e shrinkage ratio (ε i ) was measured at each curing age and calculated according to equation (1). ree prism specimens were measured for each data point [26]. After the test, the sample was weighed, and the water loss rate was calculated: where δ i is the shrinkage deformation of the specimen at the ith moment and L is the initial length of the specimen. Table 2, five mix formulas of typical solidified materials were performed. Compaction characteristics (maximum dry density ρ dmax and optimum moisture content w opt ) of the mixtures are also summarized in Table 2.    Advances in Civil Engineering 3 material. Figure 5 shows the strengths of different mixtures under standard curing condition and soaking curing condition, and it is very clear that all the strengths of the testing specimens under the soaking curing condition at 28 days were lower than those under the standard curing condition. is means that the unconfined compressive strength is weakened by the infiltrating water. erefore, the PGmodified mixtures used as road base materials in engineering practice cannot be immersed in water at the initial stage of curing, and if it rains during the construction, some covering measures must be taken [9,10]. e mass loss of PLFS-2 and PLFS-3 is 1.8% and 2.4%, which is as large as LFS (2.3%). Compared to them, the mass loss of PLFS-1 and PLFS-4 almost doubled.

Results and Discussion
is is mainly because the low content of fine aggregate in PLFS-1 prevented the coarse aggregates from being effectively bound together, while the excessive content of blinder in PLFS-4 resulted in the poor integrity. e dry-wet environment stability of the solidified materials was also studied, and the test results are presented in Figure 6; dry-wet environment stability was assessed with the compressive strength in the specimens after 6 successive dry-wet (DW) cycles. From Figure 6, it can be seen that the strength development of lime-FA-PG-CS mixtures after DW cycles was similar to that of the typical lime-FA-CS road base material, and all the strengths of the specimens increased after 6 cycles of wetting-drying, especially the strength of LFS and PLFS-2 increased by 50.4% and 57.8%.
is is   mainly because the higher temperature (60°C) in the drying environment was conducive to the strength development. is behavior indicates that using PG to replace part of lime and CS in typical lime-FA-CS road base material did not reduce the environmental stability.

Composition Structures of Solidified Material.
e composition structures of different solidified materials were analyzed by the image processing method. e second scan section of LFS, PLFS-1, PLFS-2, and PLFS-3 are presented in Figure 7. In the segmented images, the black color, gray color, and white color represent the air voids, binder which contains the fine aggregate, and coarse aggregates, respectively. e area ratios of coarse aggregate, binder, and voids in the four scan sections are calculated and listed in Table 3. It can be seen that all the solidified materials had good performance, and their void area ratio were no more than 10%; however, the porosity of each section for one specimen is very discrete because of the uneven distribution of coarse aggregate. For PG-modified mixtures, the content of fine aggregate has impact on composition structures. PLFS-2 with 8% fine aggregate has the smallest void area ratio (3.8%-4.7%), whereas the void area ratios of PLFS-1 and PLFS-3 are higher. e content of fine aggregate of PLFS-1 is 5% which may not completely fill the voids between larger particles. On the contrary, the content of fine aggregate of PLFS-1 is 12%, and more fine aggregate in the binders is prone to dehydration shrinkage cracks. Hence, a reasonable content of fine aggregate can enhance the conjoint points between larger particles and improve the compactness of mixtures.

Dry Shrinkage.
e anticrack ability of different solidified materials can be evaluated by dry shrinkage tests. e test results of four mixtures LFS, PLFS-1, PLFS-2, and PLFS-3 are presented in Figure 8. e shrinkage ratio of LFS Advances in Civil Engineering 5 mainly occurred within 0-30 hours, and then the shrinkage ratio tended to be stable, while the dry shrinkage ratios of PG-modified materials (PLFS-1, PLFS-2, and PLFS-3) tended to increase at first and then decrease. e dry shrinkage of the various solidified materials is LFS > PLFS-3 > PLFS-1 > PLFS-2. PLFS-2 has the lowest shrinkage ratio and has a slight expansivity after curing 160 hours. It is also evident that PG can obviously decrease the dry shrinkage ratio of mixtures due to its enhancement on the pozzuolana reactions and the reaction product AFt's expansivity.

Conclusions
In this paper, the lime-FA-PG-CS composite is prepared to substitute the typical lime-FA-CS mixture as a road base material. Based on laboratory tests, its mechanical property, composition structures, and volume stability are investigated comparing with the typical lime-FA-CS mixture.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this article.