strength development and hydration progress of a …

STRENGTH DEVELOPMENT AND HYDRATION PROGRESS OF A

CEMENT-BASED SOLIDIFYING AGENT USED IN CEMENT-

TREATED SOIL CONTAINING VOLCANIC MINERAL

COMPONENTS

Izuru Segawa (1), Yuya Takahashi (1) and Go Igarashi (1)

(1) Department of Civil Engineering, The University of Tokyo, Japan

Abstract

This study investigates the relationship between the strength development and hydration

progress of cement-based solidifying agents (CBSAs) in cement-treated soil with volcanic

mineral components. A CBSA is a cement-based material used as a soil stabilizer. It is known

that the strength development of cement-based materials is inhibited by volcanic mineral

components, e.g., allophane, and CBSAs are specially controlled against hydration inhibitions.

Previous studies showed that allophane might be the governing factor for the inhibition of

cement hydration, and its response to the hydration properties should be investigated.

In this study, cement-paste and cement-treated–soil specimens are made with three types of

binder (ordinary Portland cement, blast-furnace slag cement, and a CBSA) to investigate the

hydration progress and strength development. Uniaxial compression tests were conducted,

and the strength development was measured with the different binder types. It was found that

the strengths of the cement-paste specimens at day 28 were almost the same with the three

binders, while they differed in the mortar specimens that contained allophane. X-ray

diffraction/Rietveld analyses were conducted to measure the hydration properties. From the

results, it was supposed that, owing to the lack of Ca(OH)2, the pozzolanic reaction was

stagnant. Considering the ion concentration of the pore water in the solidified materials,

quantitative estimations of the strength development affected by allophane should be possible

in the future. Keywords: Cement-based solidifying agents, cement-treated soil, XRD Rietveld, ettringite,

allophane

1. INTRODUCTION

A cement-based solidifying agent (CBSA) is a material for soil improvement. Components

that are effective for hardening soil are added to the base cement material. The CBSA

composition is designed to produce the appropriate amount of ettringite during its hydration.

Ettringite should improve the ground through its capacity to absorb large amounts of water

4th International RILEM conference on Microstructure Related Durability of Cementitious Composites (Microdurability2020)

228

and decrease the moisture content in the ground; it also acts like a bridge to firmly connect the

soil particles.

CBSAs are widely used in various places for structure-foundation stabilization, roadbed

improvements, etc. About 8-million tons of CBSAs are used annually in Japan. However,

when soil containing volcanic ash is treated, the strength is often lower than in other soil cases.

Because volcanic-ash soils are distributed all over Japan, understandings of the mechanism

and future prediction models are anticipated in the engineering field.

Allophane, a mineral contained in volcanic-ash soil, is thought to be one of the main

causes of the insufficient strength development in cement treatments. Allophane is the

weathered phase of volcanic glass or pumice, and its chemical composition can be represented

as (SiO2)n･Al2O3･(H2O)m, where n = 1.3–2.0 and m = 2.5–3.0. The amorphous phase of

allophane in the soil inhibits the cement-solidification reaction by adsorbing the calcium ions.

Experiments by the Japan Cement Association showed a correlation between the amount of

allophane and the uniaxial compressive strengths of cement-treated soils, as well as a

correlation between the amount of allophane and a reduction in the calcium-hydroxide

solution concentrations [1]. It was assumed that the strength development was inhibited

because of the insufficient hydration progress. However, no studies have been conducted to

quantify the hydration progress against the solidification inhibition.

Thus, in this study, the degree of solidification inhibition by allophane in cement-treated

soil was investigated by preparing cement-treated soil specimens and quantitatively

measuring the clinker consumption and the hydrate production. The study attempts to clarify

the contribution of each hydrate to the strength development.

2. EXPERIMENTAL SETUP

Cement-paste and cement-treated–soil specimens were prepared and subjected to

compressive-strength tests, X-ray diffraction tests (XRD), and thermal gravimetric analysis

tests (TGA). Three types of binder were used: ordinary Portland cement (OPC), blast-furnace

slag cement (type B in accordance with Japanese Industrial Standard JIS R 5211 [2]) (BB),

and a cement-based solidification agent (CBSA).

Table 1 shows the physical properties and clinker compositions of the binders. Cement-

paste specimens were prepared at two water-to-binder ratios (W/B): 0.5 and 1.0. For the

cement-treated soil, the particle sizes of commercial sand from Kanuma, Japan (Kanuma-soil)

were adjusted and used in this study.

Table 2 shows the physical properties of the soil. The allophane content in the soil was

estimated by XRD. It should be noted that the estimated allophane content includes other

inorganic amorphous materials. Table 3 shows a series of cement-treated soil specimens. The

cement-treated soil was prepared according to the method in the Japanese Geotechnical

Society's JGS 0821-2009 [3].

Table 1: Binder properties

Binder

type

Ig. Loss

(%)

Chemical composition (mass %) Density

(g/cm3)

Specific surface

area (cm2/g) SiO2 Al2O3 Fe2O3 CaO MgO SO3

OPC 2.47 13.89 3.74 3.47 75.12 0.97 1.94 3.15 3550

BB 1.88 19.63 7.27 1.76 62.99 3.30 3.27 3.02 3800

CBSA 1.38 22.23 7.13 2.01 55.77 2.61 6.87 3.06 3630


229

Three cylindrical specimens (φ5 × 10 cm) were prepared for each cement-paste or cement-

treated–soil case. The top surface of each specimen was covered with plastic wrap until the

test-material age after casting (sealed curing), and the specimens were placed at a constant

temperature of 20 °C.

The specimens were subjected to uniaxial compression tests at their material ages of 1, 3, 7,

and 28 days. After the compression tests, some specimen pieces were collected and subjected

to XRD and TGA tests. The test pieces were pulverized to about 5 mm and immersed in

isopropanol for one week to stop their hydration. Next, one week of vacuum drying was

applied. They were pulverized again and sieved to 100-μm particle sizes for the TGA and

XRD tests.

The TGA measurements were conducted with an STA 2500 Regulus thermal analyzer

(NETZSCH) under nitrogen-flow conditions with a temperature-increase rate of 15 °C/min.

The ignition-loss (Ig. Loss) value is defined as the value of the mass lost from 105 °C to

950 °C divided by the mass at 950 °C. XRD analyses were conducted using the LabX XRD-

6100 X-ray differential system (Shimadzu). The hydrated and dried powder samples were

analyzed under the following conditions: The X-ray source is Cu-Kα, the tube voltage is 40.0

kV, the tube current is 30.0 mA, the scanning range is 2θ = 5–70°, the step width is 0.02°, and

the scanning speed is 2.0°/min.

XRD Rietveld analyses were conducted on the XRD results with SIROQUANT software

ver. 3.0. The following phases were considered in the quantifications as cement minerals or

hydration products: C3S, C2S, C3A, C4AF, MgO, CaO, CaCO3, CaSO4･2H2O (gypsum),

CaSO4 ･ 0.5H2O (basanite), CaSO4 (anhydrite), Ca(OH)2 (CH), C3A ･ 3CaSO4 ･ 32H2O

(ettringite), C3A･CaSO4･12H2O (monosulfate), 3CaO･Al2O3･0.5Ca(OH)2･0.5CaCO3･12H2O

(hemicarbonate), 3CaO･Al2O3･CaCO3･11H2O (monocarbonate), and α-Al2O3. α-Al2O3 of 10

wt% of the samples was mixed as the internal standard.

For the cement-treated–soil specimens, albite, anorthite, and quartz were added to the

phases for the quantifications. The amorphous-substance content was calculated according to

Eq. (1) from the quantified value of the internal standard α-Al2O3 [4]. The amounts of

unhydrated clinkers and hydrated products in the samples were calculated by correcting the

quantitative values obtained from the XRD Rietveld analyses with the Ig. Loss and the

amorphous content, according to Eq. (2).

where A is the amorphous-material content, S is the mixing ratio (%) of α-Al2O3, SR is the

quantified value of α-Al2O3 (%), QR is the quantitative value of each phase (%), and L is Ig.

Loss.

Table 2: Physical soil properties Table 3: Mix proportions of cement-treated soil specimens

Moisture

contentDensity

Amount of

allophane

26.1% 2.7 g/cm3

82%

Gravel Sand Clay& Silt

12% 87% 1%

W/B S/BUnit (kg/m

3)

W B S

1.75 2.50 585 334 835

2.33 3.33 601 257 858

3.50 5.00 617 176 882

W/B S/BUnit (kg/m

3)

W B S

1.13 2.50 475 422 1056

1.50 3.33 491 328 1092

2.25 5.00 509 226 1131

A = {100 × (SR - S)}/{SR × 100 - S) / 100} (1)

Q = QR × (100 - A) / (100 - L), (2)


230

3. EXPERIMENTAL RESULTS OF CEMENT-PASTE SPECIMENS

Figure 1 shows the results for the compression tests of the cement-paste specimens. The

legend in the figure represents the (Binder Type)-(W/B). From the results in Fig. 1, BB and

CBSA showed strengths similar to OPC at day 28, while their initial strength developments

were slower than OPC in both the 0.5 and 1.0 W/B cases.

Table 4 and Figure 2 show the phase compositions of the mass calculated by the XRD

Rietveld analyses. The W/C 0.5 cases of the OPC and CBSA specimens are displayed in Fig.

2. There was no significant difference in the C3S hydration rate between OPC and CBSA,

which is considered to contribute the most to the strength in the early ages. Regarding the

other clinkers, CBSA had a slower hydration rate than OPC. In CBSA, anhydrite was

consumed at almost the same rate as C3S, and an accordingly large amount of ettringite was

produced.

Figure 2: Calculated mass of phase compositions

Using the density of each cement mineral and the hydrate phase from previous research [4,

5] (Table 5), the phase compositions in the volume basis were calculated [5, 6]. Figure 3

shows the calculated volume of the phases of the cement-paste specimens. The volume of

ettringite in the CBSA case is larger than that of the OPC case. In addition, the theoretical

porosities of the cement pastes were estimated from the calculated phase volumes. The

porosities of the cement-paste specimens (Φcp ) were calculated by Eq. (3):

where Vpore is the volume of free-water and blank spaces and Vtotal.ini is the volume of the

cement paste at day 0.

Figure 1: Strength development of paste

Table 4: Mineral composition of binders

C3S C2S C3A C4AF

OPC 68 12.2 10.2 5.3

BB 57.3 13.9 6.6 4.4

CBSA 56.1 10.7 5.4 5.9

BassaniteAnhydriteGypsum Others

OPC 0.4 0.1 0.5 3.3

BB 1.9 0.5 3.7 11.7

CBSA 0.7 14.3 2.1 4.8

OPC-0.5 case CBSA-0.5 case

Φcp = Vpore / Vtotal.ini, (3)


231

Table 5: Density of each phase in kg/m3

phase C3S C2S ettringite monocarbonate portlandite anhydrite gypsum

density 3150 3310 1778 2175 2251 2968 2311

phase C3A C4AF calcite hemicarbonate monosulfate amorphas water

density 3030 3740 2710 1985 2015 2410 1000

Figure 4 shows the estimated porosities of the cement-paste specimens. A correlation

between porosity and strength was shown in previous research [7]. However, in this study, the

CBSA and OPC cases showed almost the same porosity decrease rates, while the strength

developments were different, as shown in Fig. 1.

From these results, the following process can be assumed for the strength development of

CBSA. In the early stage, large quantities of ettringite are produced, owing to the extensive

anhydrite content. Because the cement-paste skeletons are not very stiff yet, the cement paste

itself expands or is destroyed, due to the expansion pressure of the ettringite. Hence, the

strength of CBSA can be lower than the strength of OPC at the early stage. A similar

tendency was observed in previous research [8].

After the paste has stiffened, the ettringite formation can contribute to the strength, as well

as other hydrates. Previous research has shown that the expansion pressure changes,

depending on the shape of the ettringite, and the shape changes depend on the Ca2+

concentration in the pore solution [9]. In this study, neither the expansion nor the void

distribution was measured. Measuring the volume changes or pore-size distributions can help

researchers better understand the CBSA hydration characteristics.


Figure 3: Calculated volume of the phase compositions

Figure 4: Theoretical porosity of the cement-paste specimens


232

4. EXPERIMENTAL RESULTS OF CEMENT-TREATED–SOIL SPECIMENS

Figure 5 shows the compressive strengths of the cement-treated–soil specimens. The

strength-development tendencies in the early stages vary, depending on the W/B. The OPC

and CBSA cases have almost the same strength transitions at higher W/Bs (more than 2.25);

these are an available mix-proportion range in actual soil-improving methods. On the other

hand, for lower W/B specimens, the CBSA specimens initially have a higher strength than the

OPC, while the OPC specimens have a higher strength at day 28. The tendency of the cement-

treated–soil specimens is different from the cement-paste specimens. The BB specimens have

significantly lower strength than the rest of the series.

Figure 5: Strength development of cement-treated–soil specimens

Figure 6 shows the XRD profiles of the cement-treated–soil specimens. Each result is a

profile of a corundum mixture. The data for day 0 show the results of Kanuma-soil that does

not contain cement, and the other charts show the results of the cement-treated–soil specimens

of the W/B 1.75 cases for the OPC and CBSA cases. A gentle broad peak around 20–30°,

which is unique to allophane, is observed; however, the height of this peak decreases in the

cement-treated–soil specimens.

In OPC, peaks for hemicarbonate and monocarbonate are observed around 10.7 ° and

11.7 °, respectively. In CBSA, ettringite peaks are observed at around 9.1 °, for example. In

both the OPC and CBSA cases, no CH peaks can be observed.

Previous studies have reported that ettringite can be produced by adding Ca(OH)2 and

gypsum to allophane [10]. It is supposed that the following reaction occurs between allophane

and the cement system in cement-treated soils:

These reactions are assumed to be pozzolanic reactions, and they might be effective for

long-term strength development. With these reactions, large amounts of CH, which are

generated by clinker hydration, are almost completely consumed.

(SiO2)n･Al2O3･(H2O)m + 3Ca(OH)2 + 3(CaSO4･2H2O) + (23-m)H2O

→ 3(CaO･Al2O3)･3CaSO4･32H2O + nSiO2

(4)

2SiO2 + 3Ca(OH)2 + H2O→3CaO･2SiO2･4H2O. (5)


233

Figure 7 shows the calculated masses of the phases in the cement-treated–soil specimens,

based on the XRD results. Regarding the clinkers' consumption rates, there is no significant

difference between the cement-treated soil specimens (Fig. 7) and the cement-paste specimens

(Fig. 2).

From these results, the following strength-development mechanisms for CBSA in cement-

treated soil containing allophane can be assumed. Large amounts of ettringite are produced in

the early stage and absorb the moisture. Then, the moisture content in the soil decreases and

the connection between the soil particles can be strengthened, which can contribute to an

increase in strength, as opposed to the OPC case.

On the other hand, after the early stage, because a large amount of gypsum is present, a

large amount of CH is consumed to produce ettringite, and the allophane's pozzolanic reaction

does not occur sufficiently. In addition, CBSA produces only about half of the CH that OPC

does in cement paste. As a result, the amount of calcium-silicate-hydrate (C-S-H) in the

CBSA case is less than that in the OPC case. These phenomena could be the cause of the

strength-development slowdown in the CBSA case. This assumption can also explain the

lowest strength of the BB case. The calcium concentrations in the pore water decrease in the

presence of allophane, and the reaction of the blast-furnace slag can stagnate.

In this research, it was not possible to separate the allophane-derived amorphous phase and

the cement-derived C-S-H amorphous phase in the XRD results. By studying the clear-phase

composition and the theoretical porosity, a strength prediction should be achieved in the

future. In addition, it is thought that the mechanism's hypothesis can be confirmed by

measuring the ion concentrations of the pore solutions to discuss the phase equilibrium of the

pore water and hydrates.


Figure 6: XRD profiles of cement-treated–soil specimens


Figure 7: Calculated mass of phase compositions of cement–treated–soil specimens

Day 0

Day 3

Day 7

Day 28


234

5. CONCLUSIONS

To investigate the hydration characteristics of CBSA, cement-paste specimens and cement-

treated soil specimens were prepared, and uniaxial compression, TGA, and XRD tests were

conducted. The relationships between the compressive strengths and the quantitative values of

the hydrate phases were investigated.

The early-stage strengths of the CBSA cement pastes were smaller than those of OPC. A

large amount of anhydrous gypsum was present and a large amount of ettringite was produced

in CBSA. After the stiffness development of the cement-paste systems, it is supposed that the

ettringite contributes to the strength, as other hydrates do.

The early-stage strengths of the CBSA cement-treated–soil specimens were almost equal to

or slightly higher than those of OPC. On the other hand, the subsequent strength of CBSA did

not increase much. In the early-stage strength development, a large amount of ettringite can

contribute to changing the moisture state of the soil, which can increase the initial strength.

The subsequent slow strength developments can be explained by the stagnation of the

pozzolanic reaction, due to the lack of Ca(OH)2.

ACKNOWLEDGMENTS

This study was financially supported by JSPS KAKENHI grants numbers 17H01284 and

18KK0120.

REFERENCES

[1] Japan Cement Association. ‘A study of strength development of volcanic cohesive soil using

geocement: influence of allophane.’ Cement and Concrete, 780 (2012) 3–8. (in Japanese)

[2] JIS R 5211: 2009. https://kikakurui.com/r5/R5211-2009-01.html.

[3] JGS 0821-2009. ‘Practice for making and curing stabilized soil specimens without

compaction.’ The Japanese Geotechnical Society: JAPANESE GEOTECHNICAL SOCIETY

STANDARDS. Laboratory Testing Standards of Geomaterials (vol. 2), 2009, 426-435.

[4] Bish, D.L. and Howard, S.A. ‘Quantitative phase analysis using the Rietveld method.’ Journal

of Applied Crystallography 21(2) (1988) 86–91.

[5] Balonis, M. and Glasser, F.P. ‘The density of cement phases.’ Cement and Concrete

Research 39(9) (2009) 733–739.

[6] Maruyama, I. and Igarashi, G. ‘Cement reaction and resultant physical properties of cement

paste.’ Journal of Advanced Concrete Technology 12(6) (2014) 200–213.

[7] Chen, X., Wu, S., and Zhou, J. ‘Influence of porosity on compressive and tensile strength of

cement mortar.’ Construction and Building Materials 40 (2013) 869–874.

[8] Halaweh, M. ‘Effect of alkalis and sulfates on Portland cement systems.’ (2006).

[9] Mehta, P.K. ‘Mechanism of sulfate attack on Portland cement concrete—Another

look.’ Cement and Concrete Research 13(3) (1983) 401–406.

[10] So, E. and Ying, C. ‘Influence of Allophane Content on Lime-Gypsum Stabilized Volcanic

Cohesive Soils.’ Soils and Foundations 34 (1994) 97–107.


235

https://kikakurui.com/r5/R5211-2009-01.html

https://kikakurui.com/r5/R5211-2009-01.html

strength development and hydration progress of a …

Documents