Available online at www.sciencedirect.com
ScienceDirect Procedia Engineering 189 (2017) 158 – 165
Transportation Geotechnics and Geoecology, TGG 2017, 17-19 May 2017, Saint Petersburg, Russia
Engineering properties of carbonated reactive magnesia-stabilized silt under different activity index Liu Song-Yua, Cai Guang-Huaa*, Du Yan-Juna, Zhu Hengb, Wang Pingb a Institute of Geotechnical Engineering, Southeast University, Nanjing, 210096, China. Nanjing Dong Da Geotechnical Engineering Technology Co. LTD, Nanjing, 210018, China.
b
Abstract Engineering properties of soft soils can gain great improvement through the addition of reactive magnesia (MgO) and further carbonation of substantial gaseous CO2 absorbed. The paper studies the influence of MgO activity index on engineering properties of the carbonated silt with different water-MgO ratio. The engineering properties are investigated mainly through unconfined compression tests, and then the strength development are explained by the scanning electron microscopy (SEM). The results demonstrate that the mechanical properties of carbonated MgO-stabilized soils were greatly influenced by MgO activity index and water-MgO ratio, and the unconfined compressive strength of reactive MgO-stabilized soil has increased obviously after CO2 carbonation. With increasing MgO activity index and reducing water-MgO ratio, the unconfined compressive strength increased, and the failure mode of carbonated specimens approximately changes from elastic-plastic to brittleness as well as their failure strain mainly ranges between 0.5% and 2.3%. The deformation modulus of carbonated silt generally increases with increasing unconfined compressive strength, and the ratio of the deformation modulus to unconfined compressive strength is about 35 to 150. A simplified equation with combining MgO activity index and water-MgO ratio is proposed for accurately predicting the unconfined compressive strength of carbonated MgO-stabilized soils. Moreover, the microstructural characteristics explain the strength gain of carbonated MgO-stabilized soils. © 2017 The Authors. Published by Elsevier Ltd. © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license Peer-review under responsibility of the scientific committee of the International conference on Transportation Geotechnics and (http://creativecommons.org/licenses/by-nc-nd/4.0/). Geoecology. Peer-review under responsibility of the scientific committee of the International conference on Transportation Geotechnics and Geoecology Keywords: reactive magnesia; activity index; soft soil; carbonation; engineering properties.
* Corresponding author. Tel.: +86-025-83795086; fax: +86-025-83795086. E-mail address:
[email protected]
1877-7058 © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the scientific committee of the International conference on Transportation Geotechnics and Geoecology
doi:10.1016/j.proeng.2017.05.026
159
Liu Song-Yu et al. / Procedia Engineering 189 (2017) 158 – 165
1. Introduction Foundation reinforcement, in which the native soils are generally mixed with various cementitious materials such as Portland cement (PC) and lime, have been extensively utilized in treating the soft soils existing in the construction of infrastructure projects [1, 2]. However, the strength gain of PC-stabilized soil is relatively slow due to the timedependent formation of hydration products [3]. Moreover, the PC production involves intensive energy consumption and severe environmental impacts (a0.85 to 0.95 t CO2/t PC and a5% to 8% of global anthropogenic CO2 emissions). Considerable efforts have been made to explore alternative low-carbon materials to completely or partially eliminate the use of PC, including supplementary materials such as fly ash and slag [4], geopolymers [5] and calcium carbide residues [6]. In recent years, reactive magnesia (MgO) cements were also put forward to be as one such alternative material owing to their higher hydration rate and greater potential in absorbing CO 2 [7]. Therefore, a prospective technology, in which reactive MgO was mixed with soil and then the mixture was exposed to CO2 for carbonation, could generate the rapid and significant enhancement of soil strength as well as absorb lots of CO2 [8]. Furthermore, reactive MgO is prone to hydrate to form brucite, which could then react with CO 2 and additional water to form a series of hydrated magnesium carbonates (e.g., nesquehonite, dypingite and hydromagnesite), which have been confirmed to possess significant advantages in strength growth and pore filling of untreated soils or reactive MgO-stabilized soils [9, 10]. Previous studies [11, 12] elucidated that MgO reactivity had a significant influence on the hydration kinetics of slags, and higher reactivity would result in a higher hydration rate, producing a larger quantity of hydration products and higher strength as a consequence. Mo et al. [13] studied the influence of the MgO reactivity on the deformation and mechanical properties of the MgO expansive cements. The MgO reactivity is generally measured according to the kinetic analysis method by determining the discoloration time of acidic solution used in the test, which qualitatively reflects the hydration rate of reactive MgO [11]. The Chinese Industry Standard [14] presented a method for measuring the MgO activity content or activity index (cA, %) which can be calculated by using the equation (2).
cA
40 (m2 - m1 ) u100% 18 m1
(1)
where m1 and m2 represent the dry mass of sample before and after hydration (g), respectively; 18 and 40 are the molecular weight of H2O and MgO (g/mol), respectively. Previous studies [11, 12] have also indicated that the water-MgO ratio (w0/c) and MgO activity index (cA) significantly controlled the mechanical properties of carbonated reactive MgO-stabilized soils. However, very limited studies have been conducted to investigate the influence of MgO activity index on their engineering characteristics. This study aims at (i) investigating the influence of MgO activity index on the stress-strain characteristics, unconfined compressive strength and deformation modulus of carbonated soils, and (ii) proposing an empirical equation for quantifying the relation of the unconfined compressive strength to MgO activity index and water-MgO ratio. 2. Materials and Methods 2.1. Materials The soil used in this study was sampled from a highway construction site at 2.0 m depth in Suqian City, Jiangsu Province, China (see Fig. 1(a)). The basic physicochemical properties of the soil are shown in Table 1. Two types of reactive MgO powders, MgO-H (denoted as H) and MgO-L (denoted as L), were used in this study (Fig. 1(b, c)). MgO-H was light-burned white MgO powder with higher activity content from Xingtai City, Hebei Province, China; and MgO-L was heavy-burned light pink MgO powder with lower activity content from Haicheng City, Liaoning Province, China. The grain size distribution of materials was determined using a laser particle size analyzer Mastersizer 2000, and the particle size distribution curves were shown in Fig. 2. Moreover, the oxides of
160
Liu Song-Yu et al. / Procedia Engineering 189 (2017) 158 – 165
materials were measured by using an X-ray fluorescence spectrometry, and the chemical compositions are shown in Table 2. a
c
b
Fig. 1. The materials’ pictures (a) silt; (b) MgO-H; (c) MgO-L.
Percentage finer passing (%)
100
5~75Pm
< 5Pm
> 75Pm
80 60 40 MgO-H MgO-L Soil
20 0 0.1
1
10 Particle size /Pm
100
1000
Fig. 2. The particle size distribution curves of the materials Table 1. Physicochemical properties of silt.
Index
Value
Natural water content, wn (%)
Specific gravity, Gs
Liquid limit, wL (%)
Plastic limit, wP (%)
26.1
2.71
33.8
23.9
Grain-size distribution (%) < 2 Pm
2–75 Pm
75–2000 Pm
2.8
67.2
30
pH (water/ soil =1) 8.57
Table 2. Chemical compositions of the materials (w/w, in %). Materials
MgO
Al2O3
CaO
SiO2
Fe2O3
Na2O
K2O
TiO2
SO3
P2O5
MnO2
LOI a
Silt
2.18
11.9
5.57
65.07
3.3
1.28
2.23
0.61
0.04
0.18
0.06
7.58
MgO-H
91.8
1.43
1.26
3.91
0.30
––
0.04
0.13
0.40
0.31
0.02
0.38
MgO-L
84.3
0.42
2.46
5.58
0.22
––
0.01
––
0.1
––
0.01
5.87
Note: ––, not detected; a Value of loss on ignition is referenced to 950°C.
To investigate the influence of MgO activity index on the engineering properties of carbonated MgO-stabilized soils, MgO-H and MgO-L were mixed at various ratios in which the MgO-H weight content varied from 0 to 100% (w/w) with respect to the mixture of MgO-H and MgO-L. The symbol of HiLj denotes the mixture with i% MgO-H and j% MgO-L, the corresponding MgO activity index including H100L0, H75L25, H50L50, H25L75 and H0L100 were 85.9%, 82.3%, 77.6%, 69.7% and 66.4%, respectively. The MgO activity index was determined as per the Chinese Industry Standard [14] (refer to Eq. (2)), and the corresponding MgO activity index was depicted in Fig. 3.
161
Liu Song-Yu et al. / Procedia Engineering 189 (2017) 158 – 165
Moreover, the CO2 gas (99.9%) used was supplied by Nanjing Third Bridge Industrial Gases Co, Ltd. 2.2. Specimen preparation and methodology The initial water contents (w0) were set as 20%, 25% and 30%, and the MgO contents were selected as 10%, 15%, 20% and 30%. The weighted MgO and the dry soil were mixed by a mechanical agitator for uniformity, followed by adding predetermined volume of distilled water, and the mixtures were thoroughly mixed to achieve homogeneity. Later on, the homogenous MgO-admixed soil was placed into a steel cylindrical mold (Φ50mm × H100mm) in 3 layers with equal weight, and compacted to achieve the maximum dry density (1.67 g/cm3). Subsequently, the specimen was then immediately extruded from the molds by using a hydraulic jack. After their diameter and height were measured, two identical samples prepared were placed into a fabricated apparatus (Fig. 3) containing CO2 (99.9%), and subjected to the accelerated carbonation under a gas pressure of 200 kPa for 12 h [9, 10]. After the diameter and height of carbonated specimens were measured, all carbonated specimens were immediately subjected to the unconfined compression tests whose strain rate was selected as 1.0 mm/min. For scanning electron microscopy (SEM) test, the tested samples (~1 cm3) were retrieved from the carefully handbroken carbonated specimens after unconfined compression tests, frozen by liquid nitrogen with a boiling point of 195ć and then placed in a freezing unit with a vacuum chamber to be dried at -80ć[3]. Subsequently, the freezedried pieces were used for SEM analysis by using a FEI Inspect F50 field emission scanning electron microscope.
Fig. 3. The test setup used for carbonation.
3. Results and analyses 3.1. Unconfined compressive strength Fig. 4(a, b) illustrates the change laws of unconfined compressive strength (qu, in MPa) with MgO activity index and water-MgO ratio, respectively. It can be found from Fig. 4(a) that the qu slightly increases with MgO activity index, and the smaller the w0/c is, the more remarkable the growth of qu is. According to Fig. 4(b), the qu decreases with an increase in w0/c at a given cA (i.e., cA = 66.4%, 69.7%, 77.6%, 82.3% or 85.9%); and the lower the cA is, the more moderate the variation tendency of qu is. As suggested by a previous study [10], the variation in qu with w0/c could be expressed by the Eq. (2).
qu
a u ( w0 c)b
(2)
162
Liu Song-Yu et al. / Procedia Engineering 189 (2017) 158 – 165
where a and b are dimensionless constants dependent on the cA, and can be expressed by the following relationships (see Eqs. (3) and (4)) and depicted in Fig. 5:
a 0.224 cA 12.8 R 2
(3)
0.92
(4)
14
b
w0/c=2.5 (c=10%) 12
w0/c=2.0 (c=15%)
10
w0/c=1.5 (c=20%) w0/c=1.0 (c=20%)
8
w0/c=0.8 (c=25%)
6 4 2 0 65
70 75 80 MgO activity index, cA(%)
85
Unconfined compressive strength, qu (MPa)
a
Unconfined compressive strength, qu (MPa)
b 0.0326 cA 0.0256 R 2 0.93 14 12
H100L0 H75L25 H50L50 H25L75 H0L100
10 8
Fitting parameters 2 a b R 6.93 -2.91 0.975 5.34 -2.66 0.981 3.99 -2.48 0.820 3.28 -2.33 0.995 1.98 -2.21 0.873
6 qu=a×(w0/c)
4
b
2 0 0.5
1.0 1.5 2.0 2.5 Ratio of initial water content to MgO content, w0/c
Fig. 4. Unconfined compressive strength of carbonated specimens (a) different water-MgO ratio; (b) different MgO activity index. 9
-2.0
8
2
b = -0.0326cA-0.0256 R =0.93
-2.2
6 -2.4
5 4
-2.6
3 2
-2.8
2
a = 0.224cA-12.8 R =0.92
Dimensional constant, b
Dimensional constant, a
7
1 0 60
65
70 75 80 Activity content of MgO, cA(%)
85
-3.0 90
Fig. 5. Relationships between dimensionless constants (a, b) and MgO activity index.
Thus, the predicting qu can be rewritten into the following Eq. (5) by introducing Eqs. (3) and (4) into Eq. (2): qu
0.224 cA 12.8 u (w0
c)0.0326cA 0.0256
R2
0.97
(5)
Fig. 6 shows the comparison between the predicted and measured values of qu, denoted as qu(pred.) and qu(exp.), respectively. As can been observed, the two values also match very well (R2=0.97), which validates the usefulness and efficiency of Eq. (5) for predicting qu when the cA of MgO and w0/c are given. Moreover, the determination of the optimal cA and w0/c should take into consideration both the cost of MgO with different cA and the minimum qu of carbonated soils in engineering practices.
163
Predicted unconfined compressive strength, qu (pred.) (MPa)
Liu Song-Yu et al. / Procedia Engineering 189 (2017) 158 – 165 14
2
cA=85.9% R =0.97
12
cA=82.3% cA=77.6%
10
1
cA=69.7%
1
cA=66.4%
8 6 4 2 0
0 2 4 6 8 10 12 14 Experimental unconfined compressive strength, qu (exp.) (MPa)
Fig. 6. Comparison between predicted unconfined compressive strength and measured unconfined compressive strength.
3.2. Stress-strain properties Figure 7 depicts the typical stress-strain curves of carbonated specimens with (a) different cA and (b) different w0/c. The stress-strain curves of noncarbonated MgO-stabilized silt or parent silt is not given because the strength of those soil specimens are extremely low. Generally, the stress-strain process can be divided into three stages: (1) the initial loading stage where the stress level is very low and increases linearly with increasing the strain; (2) the nonlinear growth stage of stress where the stress increases gradually with increasing the strain; (3) the descent stage of stress (i.e., failure stage) where the stress decreases suddenly with increasing the strain. Based on the characteristics of the failure stages, it could be seen from Fig. 7 that when cA is larger than 77.6% or the w0/c is lower than 1.5, the stress-strain curves tend towards a brittle failure in terms of rapid drop in the post-peak stress with strain increasing, which is akin to that of the structured natural clays or cement-stabilized clays. Nevertheless, as the cA decreases or the w0/c increases, the stress-strain curves gradually exhibit the plastic-ductile characteristic showing a gradual drop in the post-peak stress with strain, which is similar to that of remoulded natural clays. The analysis results indicate that the cA and w0/c should be effectively controlled according to the engineering requirement in the practical application. 12 10 8
b
H100L0 w0/c=0.8 H75L25 H50L50 H25L75 H0L100
Axial stress, V (MPa)
Axial stress, V (MPa)
a
6 4 2 0 0.0
0.5
1.0 1.5 2.0 Axial strain, H (%)
2.5
3.0
12 10 8 6
w0/c=2.5 cA=85.9% w0/c=2.0 w0/c=1.5 w0/c=1.0 w0/c=0.8
4 2 0 0.0
0.5
1.0 1.5 2.0 Axial strain, H (%)
2.5
3.0
Fig. 7. Stress-strain curves of carbonated specimens (a) different MgO activity index; (b) different water-MgO ratio.
3.3. Deformation modulus Based on the results of unconfined compression tests, the parameter E50 representing the secant modulus of elasticity could be expressed as the ratio of stress to strain at half of strain at failure [3, 9]. The variation of the secant modulus with the qu of carbonated specimens is exhibited in Fig. 8. It is evident from the figure that the
164
Liu Song-Yu et al. / Procedia Engineering 189 (2017) 158 – 165
secant modulus generally increases with an increase in qu, and the E50-qu ratios vary from 35 to 150, which are consistent with those (30 to 200) demonstrated in the literature [9]. 600
0q
u
Number of samples=50
=15 50
E
Secant modulus, E50 (MPa)
500 400
=3 E 50
300
H100L0 H75L25 H50L50 H25L75 H0L100
200 100 0
5q u
0
2 4 6 8 10 Unconfined compressive strength, qu (MPa)
12
Fig. 8. Relationship between secant modulus and unconfined compressive strength.
3.4. Scanning electron microscope analysis Fig. 9 illustrates the scanning electronic micrograph (SEM) images of carbonated soils. It could be observed from Fig. 9 that the carbonated samples with lower cA display looser structures, and a few porous flocculation blobs of brucite exist in each soil sample tested, which has a limited bonding ability [8, 10, 16]. Abundant elongatedprismatic or rod-like crystals and the needle- or rosette -flaky reticulate crystals are observed, and the two carbonation products have been confirmed as nesquehonite and hydromagnesite or dypingite respectively by previous studies [8, 10, 13, 17]. The nesquehonite crystals and hydromagnesite or dypingite are shown in the carbonated samples with lower cA (cA = 66.4% and 77.6%). Besides, a few grains of MgO are observed due to its incomplete hydration at lower cA. A small quantity of polyhedral particles of magnesite are found in the soil when the cA is 85.9%. MgO-H is prone to consume more water to hydrate as compared to MgO-L and the corresponding hydrated product of brucite would continue consuming a certain amount of water as well as abundant CO2 during carbonation. The carbonation of brucite yields a series of hydrated magnesium carbonates including nesquehonite, dypingite or hydromagnesite) whose formations are associated with expansion of carbonated soils [8]. It is noteworthy that these carbonation products are responsible for the pore-filling effect [8, 10], which is substantiated by the SEM analyses. Previous studies have demonstrated that crystalline nesquehonite plays an important role in the strength gain of carbonated soils [8, 10]. As previous researches shown, the dypingite or hydromagnesite and nesquehonite are the main contributors to the effective filling of the available macro pores [8, 10]. Hence, these carbonates would further facilitate the strength gain of the carbonated soils. In summary, it can be opined that the variation in strength characteristics of carbonated soils with cA and w0/c could be established by resorting to SEM analyses. b
a
c Brucite
Magnesite
Nesquehonite
Nesquehonite Brucite
Dypingite/ Hydromagnesite
Nesquehonite Brucite Dypingite/ Hydromagnesite
Nesquehonite
Fig. 9. Scanning electronic micrograph images of carbonated soils (a) cA=85.9%; (b) cA=77.6%; (c) cA=66.4%.
Liu Song-Yu et al. / Procedia Engineering 189 (2017) 158 – 165
4. Conclusions This paper studied the influence of cA on the engineering characteristics of MgO-stabilized soils with varying w0/c after a complete carbonation for 12 h. Based on the obtained results, the following conclusions could be drawn: (1) The mechanical properties of carbonated soils were greatly influenced by cA and w0/c, and the unconfined compressive strength of reactive MgO-stabilized soil has increased obviously after CO2 carbonation. (2) With increasing cA and reducing w0/c, the unconfined compressive strength increased, and the failure mode of carbonated specimens approximately changes from elastic-plastic to brittleness as well as their failure strain mainly ranges between 0.5% and 2.3%. (3) The deformation modulus of carbonated silt generally increases with increasing unconfined compressive strength, and the ratio of the deformation modulus to unconfined compressive strength is about 35 to 150. (4) A simplified equation with combining cA and w0/c is proposed for accurately predicting the unconfined compressive strength of carbonated MgO-stabilized soils. Nonetheless, the present investigation is limited to the low plasticity-unsaturated soils and designated carbonation time (i.e., 12 h). Therefore, it is necessary to extend the utilization of carbonation to other soils, to explore the minimum time for achieving complete carbonation in further studies. Acknowledgements This research is financially supported by the National Natural Science Foundation of China (No. 51279032 and 41330641), “Twelfth Five-Year” National Technology Support Program of China (No. 2012BAJ01B02-01), and Project (KYLX_0147) supported by Graduate Student Scientific Research Innovation Projects of Jiangsu Province. References [1] M. Terashi, H. Tanaka, Ground improved by deep mixing method. In Proceedings of the 10th International Conference on Soil Mechanics and Foundation Engineering, Stockholm, 1981, pp. 777-780. [2] S.Y. Liu, Y.J. Du, Y.L. Yi, A.J. Puppala, Field investigations on performance of T-Shaped deep mixed soil cement column-supported embankments over soft ground, J. Geotech. Geoenviron. Eng. 138 (2012) 718-727. [3] Y.J. Du, N.J. Jiang, S.Y. Liu, F. Jin, D.N. Singh, J.P. Anand, Engineering properties and microstructural characteristics of cement-stabilized zinc-contaminated kaolin, Can. Geotech. J. 51 (2014) 289-302. [4] M. Disfani, A. Arulrajah, H. Haghighi, A. Mohammadinia, S. Horpibulsuk, Flexural beam fatigue strength evaluation of crushed brick as a supplementary material in cement stabilized recycled concrete aggregates, Constr. Build. Mater. 68 (2014) 667-676. [5] P. Sukmak, S. Horpibulsuk, S.L. Shen, P. Chindaprasirt, C. Suksiripattanapong, Factors influencing strength development in clay-fly ash geopolymer, Constr. Build. Mater. 47 (2013) 1125-1136. [6] S. Horpibulsuk, C. Phetchuay, A. Chinkulkijniwat, A. Cholaphatsorn, Strength development in silty clay stabilized with calcium carbide residue and fly ash, Soils. Found. 53 (2013) 477-486. [7] A.J.W. Harrison, Reactive magnesium oxide cements, United States Patent 2008; 7347896. US 11/016,722. [8] Y.L. Yi, M. Liska, C. Unluer, A. Al-Tabbaa, Carbonating magnesia for soil stabilization, Can. Geotech. J. 50 (2013) 899-905. [9] G.H. Cai, S.Y. Liu, Y.J. Du, D.W. Zhang, X. Zheng, Strength and deformation characteristics of carbonated reactive magnesia treated silt soil, J. Cent. South. Univ. 22 (2015) 1859-1868. [10] G.H. Cai, Y.J. Du, S.Y. Liu, D.N. Singh, Physical properties, electrical resistivity and strength characteristics of carbonated silty soil admixed with reactive magnesia, Can. Geotech. J. 52 (2015) 1-15. [11] F. Jin, A. Al-Tabbaa, Characterisation of different commercial reactive magnesia, Adv. Cem. Res. 26 (2013) 101-113. [12] F. Jin, K. Gu, A. Al-Tabbaa, Strength and drying shrinkage of reactive MgO modified alkali-activated slag paste, Constr. Build. Mater. 51 (2014) 395-404. [13] L.W. Mo, M. Liu, A. Al-Tabbaa, M. Deng, Deformation and mechanical properties of the expansive cements produced by inter-grinding cement clinker and MgOs with various reactivities, Constr. Build. Mater. 80 (2015) 1-8. [14] Chinese Industry Standard. Test methods for chemical activity of caustic burned magnesia. Black metallurgical industry standard of the People's Republic of China 2006; YB/T 4019-2006. (in Chinese) [15] C. Unluer, A. Al-Tabbaa, Impact of hydrated magnesium carbonate additives on the carbonation of reactive MgO cements, Cem. Concr. Res. 54 (2013) 87-97. [16] C. Unluer, A. Al-Tabbaa, Enhancing the carbonation of MgO cement porous blocks through improved curing conditions, Cem. Concr. Res. 59 (2014) 55-65. [17] S.Y. Liu, C. Li, Influence of MgO activity on the stabilization efficiency of carbonated mixing method. Chin. J. Geotech. Eng. 37 (2015) 148-155. (in Chinese)
165
