1. Introduction
Cement-treated aggregates have been widely used in modern construction, such as railway subgrades and road bases. However, when the surrounding environment or the soils contain sulfate, it may cause damage to the cement-treated aggregates, such as expansion, cracking, loosening, and materials peeling off [
1,
2,
3]. Sulfate attack on cement-treated aggregates (SACA) is a well-known problem that causes severe damage to concrete structures in addition to remarkable heave. In northwest China, continuous heave was observed in a high-speed railway subgrade section, and cement-treated graded macadam, ettringite (3CaO·Al
2O
3·3CaSO
4·32H
2O), gypsum (CaSO
4·2H
2O), and thaumasite (CaCO
3·CaSO
4·CaSiO
3·15H
2O) were found in the subgrade bed [
4]. In Tarragona, Spain, the embankments were designed by including soil–cement-treated transition wedges (with gypsum), and they experienced a continuous and severe heave shortly after construction. The ettringite and thaumasite in the embankment were analyzed [
5]. In northern Louisiana, the roads using Winn rock (CaSO
4) gravel as a surface course were rehabilitated via stabilization with Type I Portland cement and an asphaltic overlay, and it heaved within days. The ettringite was analyzed in the expansion position [
6].
Although discussion regarding the sulfate attack mechanism has been ongoing [
7,
8,
9,
10], it is readily admitted that the expansive products (e.g., ettringite, gypsum, and thaumasite) involved in a sulfate attack led to the expansion induced by SACA. The general reaction which causes the attack may be summarized in Equations (1)–(3) [
11,
12]. The formation of crystals in a supersaturated solution may exert pressure on its surroundings and lead to expansion [
13].
The conditions of expansion induced by SACA have been studied by many researchers. The cations (i.e., Na
+, K
+, Mg
2+, Ca
2+) also have a major influence on the reaction’s development [
14]. Gypsum and M-S-H are the major products when Mg
2+ is present in the sulfate solution, whereas ettringite and gypsum are predominantly found in the presence of Na
+ [
15]. Water is the reactant for the formation of ettringite, which is a prerequisite for SACA [
16]. In an alkaline environment (pH > 10.5), aluminate and silicate ions from the clay minerals of soil are released (in potentially large quantities), and then they are made available for the SACA reaction [
17]. Ma et al. [
18] proposed that when mortar attacks a 30 g/L sulfate ion solution, it could induce expansion. Puppala et al. [
19] concluded that the sulfate content that causes soil expansion is between 320 and 43,500 mg/kg. Wang et al. [
6] proposed that the expansion induced by SCAC is proportional to the cement content. Ouyang et al. [
20] found that although aggregates had higher initial strength when the cement content was high than when it was low, it expressed greater expansion after the sulfate attack. Temperature also has a strong effect on the SACA reaction, and 0 to 20 °C is the active temperature range for a sulfate attack [
4]. Tai et al. [
21] noted that porosity could also have a positive effect on durability by generating additional space to accommodate the expansive phase. Through field tests, Tang et al. [
22] proposed that the temperature, sodium sulfate, and cement content may be the main causes of arch expansion on pavement surfaces and cement-stabilized macadam bases. These factors may change the period of sulfate-induced distress initiation from 2 to over 25 years [
14].
Based on the overview of the literature described above, it is worth noting that although the factors that influence the SACA reaction are known, there is no unified standard for the critical sulfate content of expansion induced by SACA. The cement content and the degree of compaction for expansion behavior induced by SACA have seldom been analyzed. Moreover, the quantitative analysis of the influencing factors on expansion induced by SACA has also been rarely analyzed. Therefore, the objective of this paper lies in studying the effects of the sodium sulfate content, cement content, degree of compaction, sulfate types, attack types, aluminum ion supply, concentration for curing the sulfate solution, temperature for the sulfate attack cement-treated aggregate, and the dominant factor of expansion induced by SACA. Through the swelling tests at the constant environment of 10 °C temperature and 70% humidity, the expansion behavior induced by SACA has been monitored up to 60 days. Based on the Sobol sensitivity analysis method, considering the interactions among influencing factors, the sensitivity of each factor affecting the SACA has been quantitatively analyzed, and the dominant factor of expansion induced by SACA has been analyzed. This work reveals the influence factors of SACA and explores the dominant factor of SACA. This work is expected to provide a solid base of information to understand the expansion mechanism of SACA, and it aims to serve as a reference for the design and construction of high-speed railway lines and maintenance of subgrade heave.
2. Materials and Methods
The high-speed railway (HSR) in northwest China was experiencing a continuous and severe heave. The subgrade was approximately 3.2 m high, and the filling of the transition zone was cement-treated graded macadam of different cement content, 5% for the top 0.4 m, and 3% for the remaining part. The surface aggregates of the subgrade bed were treated with 3% and 5% content of ordinary Portland cement. The field heave at different depths from November 2017 to April 2018 was monitored, and it was found that the cement-treated graded macadam filler is the main location of expansion [
4]. The surface aggregates of the subgrade bed were treated with 3% and 5% content of ordinary Portland cement. The mineral composition of the subgrade bed has been analyzed, which is mainly composed of quartz, calcite, and albite. The sulfate content on the top of the subgrade was close to 0.3% after expansion. Some products of SACA (e.g., ettringite, thaumasite, and gypsum) have been analyzed in the surface layer of the subgrade bed.
2.1. Materials
To simulate the field conditions in the laboratory, the fresh stone powders (without salt) with a diameter of less than 2 mm were used in this study. The stone powder contained albite (42%), quartz (19%), calcite (14%), microline (8%), amphibole (7%), muscovite (6%), and clinochlore (4%). The optimum moisture content (11.2%) and maximum dry density (2.36 g/cm3) of the stone powders were obtained.
Sulfate is the main reactant in the SACA reaction and participates in the formation of ettringite and thaumisate. Considering that the sulfate content was close to 0.3% after the reaction in the field, thus, the initial sulfate content should be much greater than 0.3%. Sodium sulfate is the only source of SACA in this paper; the sodium sulfate content was mainly set to 3% to ensure that the sulfate ions were sufficient for participating in the actual reaction in this study. In the study of sodium sulfate-influencing factors, to explore the lowest sodium sulfate content causing the expansion induced by SACA, the sodium sulfate content was set at 0.1% to 5%. To study the effect of sulfate solubility, a group of specimens of gypsum (with content of 1% to 3.5%) as the medium soluble sulfate as the sulfate source were researched. To study the effect of sodium sulfate supply, the sodium sulfate solution (0.5% to 5% content) curing environment was set. To simulate the external attack, four specimens without sulfate were cured in the sulfate solution as well. The sodium sulfate and gypsum used in this study were reagent grade with a purity of at least 99%.
Ordinary Portland cement 42.5, obtained from Zhucheng Yangchun Cement Co., Ltd., Zhucheng, China, was used in the tests. Considering the cement contents were 3% and 5% of the surface bed and transition zone in the field, the cement content was mainly set to 3% to remain consistent with the condition of the surface bed. In the study of the cement influenced, the cement content was set at 1%, 3%, 5%, and 7%.
To study the influence of aluminum ions, a set of specimens with 4% content kaolin was tested. The degree of compaction affects the swelling potential of the aggregates; thus, a set of specimens with the degree of compaction (0.85 to 0.97) was studied in this paper.
Considering 0 to 20 °C is the active temperature range for sulfate attack reactions [
4], the curing temperature of 10 °C was chosen in this paper. Considering that water, sulfate, and temperature play important roles in sulfate attack, thus, the deionized water and temperature (20 °C) curing environment was set to study the influence of water and temperature.
2.2. Experimental Procedures
The amount of stone powder, cement, sodium sulfate, and deionized water was calculated based on the maximum dry density, the optimum moisture content of fresh basalt powders, and the degree of compaction (
Table 1,
Table 2,
Table 3,
Table 4,
Table 5 and
Table 6). The process of the materials mixing refers to the Code for Soil Test of Railway Engineering [
23]. The stone powders were mixed with sodium sulfate. Then, the deionized water was added to the mixture and stirred until homogeneous. Next, the cement was added to the mixture and homogeneously stirred. Then, the material above was compacted into a metal ring mold with 61.8 mm diameter and 40 mm height. After compaction, porous stones were placed on two sides of the specimen and transferred to an oedometer. The electronic dial gauge was fixed by the oedometer to measure the vertical swelling (
Figure 1).
At last, the specimens with electronic dial gauges were maintained in a test chamber with a certain temperature and humidity. The initial value of the electronic dial gauge was set to zero, and the expansion behavior of SACA was monitored up to 60 or 240 days. In this study, the strain of the specimen was calculated by the value of vertical swelling (
) and the initial height of the specimen (
) (Equation (4)).
4. Conclusions
In this study, the effects of the sodium sulfate content, cement content, degree of compaction, sulfate types, attack types, aluminum ions supply, concentration of curing sulfate solution, and temperature on the expansion behavior of SACA have been researched. Through swelling tests, the conditions and laws of expansion of SACA have been analyzed. Based on the Sobol sensitivity analysis method, the sensitivity of each factor influencing the SACA has been quantitatively analyzed, and the dominant factor of expansion induced by SACA has also been proposed. The main conclusions can be summarized as follows.
- (1)
The 0.5% sodium sulfate content is the minimum sulfate content causing the expansion of SACA. Under the sulfate content of less than 1%, the expansion of SACA is small on the whole. Under the sulfate content between 1% and 3%, the expansion behavior expresses four stages: rapid strain increases, a short stagnation period, the strain increasing significantly, and a stable stage with a constant strain. Under the sulfate content larger than 5%, there are two stages of the expansion behavior: rapid strain increase and a constant strain.
- (2)
The cement content does not express a consistent tendency for expansion behavior induced by SACA. At a certain sulfate content, it takes time for sulfate ions to migrate to aggregates under external attack, which results in the expansion induced by SACA under external attack being much less than that under internal attack. The expansion of SACA under the deionized water curing environment is smaller than that under the 70% humidity curing environment. This is because the concentration of sodium sulfate in the specimen is diluted by deionized water and reduces the rate of the attack reaction. The expansion of SACA at day 60 still expresses a continues increasing under the sodium sulfate solution curing environment.
- (3)
The expansion induced by SACA under 10 °C curing temperature is obviously larger than that cured in 20 °C curing temperature (the strains increasing about 1.4 to 10 times). The expansion increases with the increasing of the degree of compaction. Under the sulfate content of more than 1%, the addition of kaolin promotes the expansion of SACA. Compared with sodium sulfate attack, when using gypsum as the medium soluble sulfate, the rate of gypsum attack is slower and expansion is smaller.
- (4)
Based on the Sobol sensitivity analysis method, considering the interactions among influencing factors, it is found that sulfate content is the domain factor of expansion induced by SACA, which also presents a logarithmic function relationship with strain. In addition, the temperature and degree of compaction also have greater influences on expansion.
- (5)
For cement-treated aggregates in the high-speed railway subgrade, there are no relevant specifications for limiting the sulfate content of sulfate attack. It is recommended to analyze the sulfate ions content in the subgrade filler and the ground soil before subgrade construction. If the sulfate content reaches 0.5%, it is necessary to conduct the swelling testing in the laboratory to simulate the field conditions, judge the swelling potential of the filler and soil, and determine the availability of the filler and soil.