- Email: [email protected]
Experimental and statistical analysis of the alkali–silica reaction of accelerating admixtures in shotcrete
Construction and Building Materials 30 (2012) 330–339
Contents lists available at SciVerse ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Experimental and statistical analysis of the alkali–silica reaction of accelerating admixtures in shotcrete Jong-Pil Won ⇑, Bo-Ra Choi, Jae-Wan Lee Department of Civil and Environmental System Engineering, Konkuk University, Seoul 143-701, Republic of Korea
a r t i c l e
i n f o
a b s t r a c t
Article history: Received 21 July 2011 Received in revised form 22 October 2011 Accepted 24 November 2011 Available online 30 December 2011
This study investigated the alkali–silica reaction of accelerating admixtures for shotcrete. Tests were performed with two types of cement (low- and high-alkali) and three types of accelerating admixture (alkali-free, cement-based mineral and aluminate). An expansion test was performed to determine the alkali–silica reactivity according to ASTM C 1260. The results showed that the expansion increased with the total equivalent alkali content of the specimens. To examine this finding statistically, analyses were conducted at the 95% confidence level. When the low-alkali cement was used, no difference appeared to exist in the expansion after a reasonable period of time regardless of the accelerating admixture. In contrast, when the high-alkali cement was applied, the expansion of the specimens varied with time, although no difference was observed between the expansions of the plain sample and the specimen using the alkali-free accelerating admixture. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: Alkali–silica reaction Accelerating admixture Cementitious composites Expansion Shotcrete Statistical analysis
1. Introduction A high alkali content, a reactive aggregate and moisture are essential for creating the alkali–silica reaction (ASR) [1]. If any one of these factors is absent, then the ASR does not occur. ASR can adversely affect the durability of concrete because the alkali– silica gel which is a reaction product can cause serious cracking and overall expansion. Its process can be expressed as follows [2,3]. (1) Silica tetrahedrons (SiO2) which are one of the main components of reactive aggregate break up due to the attack of hydroxyl ions (OH).
2SiO2 þ OH ! SiO5=2 þ SiO5=2 H (2) If the attack of hydroxyl ions is continued, silicate ions, 2 H2 SiO4 , H3SiO4, and small polymers are formed.
2
SiO5=2 þ OH þ 1=2H2 O ! H2 SiO4
In this process, swelling N(K)–S–H gel is created by combining 2 H2 SiO4 and sodium or potassium ion [4]. N(K)–S–H gel has a low viscosity and can easily move from the aggregates to other parts. The volume of it is increased because it attracts water due to osmosis. At this time, generated local pressure can cause cracking in the concrete. ⇑ Corresponding author. Tel.: +82 2 450 3750; fax: +82 2 2201 0907. E-mail address: [email protected] (J.-P. Won). 0950-0618/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.conbuildmat.2011.11.017
It is hard to predict the degree of expansion by ASR. Because it depends on the viscosity or stiffness of alkali–silica gel and there are many other factors (such as aggregate type and size, alkali content and temperature) that affect the ASR. In the case of alkali content, if it is increased, reactive silica components react more easily with hydroxyl ions so that the ASR is accelerated. The alkali is supplied mainly from cement or the specific environment, and can be supplied directly from additives, such as the aggregate used or deicing chemicals [5,6]. Swamy and Al-Asali [7] showed that the higher expansion occurred when the specimens were stored in a 4% NaCl than in water. Saccani et al. [8] used four types of sodium additives (NaOH, Na2SO4, NaCl, NaHCO3) to examine the effect of alkali content on the expansion due to ASR. As a result, when sodium additives were added, expansion is bigger than those not added. Shehata and Thomas [9] showed that the expansion was influenced by alkali content in cement. An analysis revealed that the equivalent alkali content (Na2Oeq = Na2O + 0.658K2O) of Type I Portland cements produced by eight companies in Korea was 0.81–1.14%. These values exceed 0.6%, below which the equivalent alkali content prevents the ASR from taking place [10]. Therefore, the ASR should occur if reactive aggregate and sufficient moisture are supplied. In addition, if an additive containing high alkali is used, the ASR will be greater [8,9]. Shotcrete is used to construct tunnels or underground structures [11]. A low amount of rebounding, good adhesiveness, good shooting, and quick formation of initial strength are basic requirements of shotcrete [11–14]. Therefore using of accelerating admixture is essential to meet this requirement [11]. The
331
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339 Table 1 Chemical composition of the cements. Alkali content of the cement (%)
SiO2 Al2O3 Fe2O3 CaO MgO SO3 K2O Na2O Na2Oeq
Low
High
20.51 4.86 3.38 62.22 2.57 2.52 1.07 0.11 0.81
20.85 4.74 3.18 61.95 2.73 2.40 1.50 0.15 1.14
Fig. 2. Expansion test results (L: low-alkali cement, H: high-alkali cement, AF: alkali-free accelerating admixture, CM: cement-based mineral accelerating admixture, AN: aluminate accelerating admixture).
Fig. 1. X-ray diffraction patterns for aggregate.
(a) After 14 days
Table 2 Equivalent alkali content of the accelerating admixtures. Alkali content of the accelerating admixtures (%)
K2O Na2O Na2Oeq
Alkali-free
Cement-based mineral
Aluminate
0.01 0.51 0.52
0.18 13.40 13.52
0.02 18.80 18.81
Table 3 Designation of each mixture. Type of accelerating admixture
None Alkali-free Cement-based mineral Aluminate
Alkali content of the cement Low
High
Plain L L-AF L-CM L-AN
Plain H H-AF H-CM H-AN
accelerating admixture for shotcrete is a supportive material that expedites the initial strength development. It is classified as a hardening accelerator, inorganic alkali, and can include salt and silicate or aluminate, alkali-free and cement-based mineral admixtures [15]. The silicate and aluminate accelerating admixtures were once used the most widely, but they have been banned in many countries because they are potentially harmful to human health. Consequently, a new harmless accelerating
(b) After 84 days Fig. 3. Comparison of expansion.
admixture that does not result in a long-term reduction in strength is needed. Therefore, alkali-free and cement-based mineral accelerating admixtures were developed [16]. Nevertheless, the silicate and aluminate accelerating admixtures are still used widely for economic reasons, and the ASR may occur because some of these contain high alkali levels. Although the effects of additive materials on the ASR have been investigated [17], no study has examined the effects of the accelerating admixture used for shotcrete.
332
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
Therefore, this study scrutinised the ASR of accelerating admixtures in cementitious composites when reactive aggregate and sufficient moisture are supplied. Furthermore, the influence of the equivalent alkali contents of the accelerating admixture and cement on expansion was examined.
Table 4 Descriptive statistic of expansion. Name of specimen
After 14 days
After 84 days
Mean
Standard deviation
Mean
Standard deviation
L L-AF L-CM L-AN H H-AF H-CM H-AN
0.101 0.137 0.152 0.205 0.123 0.148 0.178 0.227
0.047 0.005 0.026 0.010 0.031 0.030 0.032 0.020
0.381 0.407 0.512 0.483 0.405 0.430 0.487 0.548
0.076 0.022 0.080 0.008 0.033 0.020 0.010 0.015
2. Alkali–silica reaction test 2.1. Materials 2.1.1. Cement The Type I Portland cements of eight companies in Korea were classified according to the equivalent alkali content. Of these, the cements with the highest (1.14%) and lowest (0.81%) alkali contents were selected. The chemical composition of these cements is shown in Table 1.
Table 5 Analysis of variance for the cements and accelerating admixtures (after 14 days). Source
Sum of squares
df**
Mean square
F***
P****
Cement Accelerating admixture* Cement accelerating admixture Residual
0.002 0.034 0.000
1 3 3
0.002 0.011 5.833E5
3.186 14.752 0.075
0.093 0.000 0.973
0.012
16
0.001
–
–
Total
0.656
24
–
–
–
* ** *** ****
Statistically significant at the 95% level. df: degree of freedom. F: F-distribution. P: P-value, significance probability.
Table 6 Analysis of variance for the cements and accelerating admixtures (after 84 days). df**
Mean square
F***
P****
0.003 0.065 0.006
1 3 3
0.003 0.022 0.002
1.617 11.903 1.130
0.222 0.000 0.367
0.029 5.107
16 24
0.002 –
– –
– –
Source
Sum of squares
Cement Accelerating admixture* Cement Accelerating admixture Residual Total * ** *** ****
2.1.2. Aggregate Crushed andesite was used. This aggregate is classified as a reactive aggregate because it contains a large amount of amorphous silica components [4,18]. According to a study by the Cement and Concrete Association of New Zealand, the expansion behaviour varies with the type of andesite [19]. Fig. 1 shows the X-ray diffraction (XRD) result of aggregate used in this study. In figure, the highest point mean amorphous silica component [20]. In other words, if point is high, there are lots of amorphous silica components that can cause ASR. 2.1.3. Accelerating admixture for shotcrete The alkali-free, cement-based mineral and aluminate accelerating admixtures for shotcrete that are used the most widely were selected. The alkali content of each admixture was measured using inductively coupled plasma atomic emission spectrometry (ICP). As Table 2 shows, the cement-based mineral and aluminate accelerating admixtures had higher equivalent alkali contents. The accelerating admixtures used had lower potassium contents than the cement, but higher sodium contents. 2.2. Experimental procedure An expansion test was performed to determine the alkali–silica reactivity according to ASTM C 1260 using 440 g of cement, 990 g of aggregate and 206.8 g of water. The alkali-free, cement-based mineral and aluminate accelerating admixtures were mixed with 7%, 5% and 5% cement by weight, respectively, based on typical usage. The mixtures are listed in Table 3. Experimental procedure can be summarised as follows: Step 1. Three 25 25 285-mm specimens were made from each mixture. To measure the length change, a stud was installed at both ends of each specimen. Step 2. The specimens were cured for 24 h in air after moulding and first measured after removing them from the mould. Step 3. The specimens were cured in water at 80 ± 2 °C for 24 h. The zero length was measured within 15 s.
Statistically significant at the 95% level. df: degree of freedom. F: F-distribution. P: p-value, significance probability.
Table 7 Multiple comparisons of the accelerating admixtures (using low-alkali cement). Source
Mean difference
Std. error**
After 14 days Plain alkali-free Plain cement-based mineral* Plain aluminate* Alkali-free cement-based mineral Alkali-free aluminate* Cement-based mineral aluminate*
0.03633 0.05167 0.10433 0.01533 0.06800 0.05267
0.022169 0.022169 0.022169 0.022169 0.022169 0.022169
After 84 days Plain alkali-free Plain cement-based mineral Plain aluminate* Alkali-free cement-based mineral Alkali-free aluminate Cement-based mineral aluminate
0.02600 0.05833 0.10167 0.03233 0.07567 0.04333
0.041046 0.041046 0.041046 0.041046 0.041046 0.041046
* ** ***
Statistically significant at the 95% level. Std. error: standard error. P: p-value, significance probability.
P***
95% Confidence interval Lower bound
Upper bound
0.140 0.048 0.002 0.509 0.015 0.045
0.08745 0.10279 0.15545 0.06645 0.19912 0.10379
0.01479 0.00055 0.05321 0.03579 0.01688 0.00155
0.544 0.193 0.038 0.454 0.103 0.322
0.12065 0.15299 0.19632 0.12699 0.17032 0.13799
0.06865 0.03632 0.00701 0.06232 0.01899 0.05132
333
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339 Table 8 Multiple comparisons of the accelerating admixtures (using high-alkali cement). Source
Mean difference
Std. error**
After 14 days Plain alkali-free Plain cement-based mineral* Plain aluminate* Alkali-free cement-based mineral Alkali-free aluminate* Cement-based mineral aluminate
0.02500 0.05467 0.10333 0.02967 0.07833 0.04867
0.023384 0.023384 0.023384 0.023384 0.023384 0.023384
After 84 days Plain alkali-free Plain cement-based mineral* Plain aluminate* Alkali-free cement-based mineral* Alkali-free aluminate* Cement-based mineral aluminate*
0.02467 0.08167 0.14300 0.05700 0.11833 0.06133
0.017314 0.017314 0.017314 0.017314 0.017314 0.017314
P***
95% Confidence interval Lower bound
Upper bound
0.316 0.048 0.002 0.240 0.010 0.071
0.07892 0.10859 0.15726 0.08359 0.13226 0.10259
0.02892 0.00074 0.04941 0.02426 0.82441 0.00526
0.192 0.082 0.000 0.011 0.000 0.088
0.06459 0.12159 0.18293 0.09693 0.15826 0.10126
0.01526 0.84174 0.10307 0.01707 0.07841 0.02141
* Statistically significant at the 95% level. Std. error: standard error. P: p-value, significance probability.
(a) Plain
(b) Alkali-free
(c) Cement-based mineral
(d) Aluminate
Fig. 4. Expansion trends of the mixture using accelerating admixture according to the alkali content of the cement (after 14 days).
Step 4. After that, the specimens were cured in 1 N NaOH at 80 ± 1 °C and the length was measured at 3, 6, 9, 12, 14, 21, 28, 35, 42, 56, 70 and 84 days. The procedure was repeated three times in the same order as above.
2.3. Statistical procedure Analysis of Variance (ANOVA) was carried out based on the data by the expansion test. ANOVA is original method to compare the interaction effects. In other
334
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
(a) Plain – AF
(b) Plain – CM
(c) Plain – AN
(d) AF – CM
(e) AF – AN
(f) CM – AN
Fig. 5. Comparison of the expansion trend among mixtures (after 14 days).
words, it is the method to confirm whether or not the samples have similar distribution. In this study, twoway ANOVA was conducted because there were two variables as cement and accelerating admixture. The procedure of twoway ANOVA can be summarised as follows.
Step 1. To identify the interaction between two factors (cement and accelerating admixture). Step 2. If there is no interaction, carry out the multiple comparisons after separation the effect of two factors.
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
(a) Plain
(b) Alkali-free
(c) Cement-based mineral
(d) Aluminate
335
Fig. 6. Expansion trend of the mixture using accelerating admixture according to the alkali content of the cement (after 84 days).
Step 3. If there is interaction, remove it by using variable change and go back to step 2.
3. Results and discussion 3.1. Expansion test The results of the expansion tests according to the type of accelerating admixture are displayed in Fig. 2. In this figure, L and H are plain specimens using low-alkali and high-alkali cement respectively. And AF, CM, and AN are specimens using alkali-free, cement-based mineral, and aluminate accelerating admixture respectively. Therefore, LF means the specimens mixed the low-alkali cement and alkali-free accelerating admixture. This principle is applied to the other specimens. The plain specimens without an accelerating admixture suggest that the ASR occurred as the expansion exceeded 0.1% at 14 days. The expansion of the specimens made using high-alkali cement exceeded 0.2% after 28 days, which could have negative effects. After 35 days, the expansion of the specimens made using low-alkali cement also exceeded 0.2%. The expansion increased until 84 days and was greater using high- rather than low-alkali cement. The expansion of all specimens using the alkali-free accelerating admixture exceeded 0.1% after 14 days, suggesting that the ASR occurred. After 28 days, harmful expansion exceeding 0.2% was seen. As indicated in figure, using the alkali-free accelerating
admixture, the expansion was similar regardless of the equivalent alkali content of the cement. In the case of the specimens using the cement-based mineral accelerating admixture, the expansion was greater as this contained more alkali than the alkali-free accelerating admixture. Again, the expansion increased with time. When the cement-based mineral accelerating admixture was used with high-alkali cement, the expansion was greater than with the low-alkali cement. Using the aluminate accelerating admixture, harmful expansion exceeding 0.2% occurred at 14 days regardless of the equivalent alkali content of the cement. These specimens expanded the quickest. Using the low-alkali cement, 56 days were required for the expansion to exceed 0.4% versus 84 days with the cement-based mineral accelerating admixture. The expansion values after 14 and 84 days are shown in Fig. 3. As you can see, the expansion is greater using high-alkali cement or accelerating admixture. Based on our results, the expansion increased with the equivalent alkali content of the specimen. The equivalent alkali content of the accelerating admixture had more influence on the ASR than the cement, since the difference in expansion depended more on the equivalent alkali content of the accelerating admixture. Sodium ions appear to affect the ASR more than potassium ions, since the accelerating admixture contained more sodium ions than the cement did, as shown in Tables 1 and 2. In addition, the expansion increased with time. Alkali is believed to have been supplied endlessly from the NaOH solution. Therefore, it is expected that the expansion will increase until all
336
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
(a) Plain – AF
(b) Plain – CM
(c) Plain – AN
(d) AF – CM
(e) AF – AN
(f) CM – AN
Fig. 7. Comparison of the expansion trend among mixtures (after 84 days).
amorphous silica components in reactive aggregate are consumed due to the ASR. 3.2. Statistical analysis Statistical analysis of the expansion at 14 and 84 days was used to confirm whether the difference in the equivalent alkali content
of the cement and accelerating admixture caused the actual difference in expansion. At first, analysis of descriptive statistic was conducted. The descriptive statistic is the method which can indicate the characteristics of data and the result is shown in Table 4. Table 5 shows the result of an Analysis of Variance (ANOVA) of the expansion at 14 days. No interaction occurred between the two
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
(a) Plain L
(b) L-AF
(c) L-CM
(b) L-AN
337
Fig. 8. Expansion prediction of the specimens using low-alkali cement (L: low-alkali cement, AF: alkali-free accelerating admixture, CM: cement-based mineral accelerating admixture, AN: aluminate accelerating admixture).
factors (the cement and the accelerating admixture) at the 95% confidence level. Therefore, the effect of each factor on the expansion was examined: cement was not significant at the 95% confidence level, while the accelerating admixture had a significant effect. The results of the ANOVA for the expansion 84 days matched this finding. As indicated in Table 6, no interaction took place between the effects of the cement and accelerating admixture on the expansion at the 95% confidence level. Only the accelerating admixture had a significant effect on the expansion. Then, we analysed whether a difference in expansion occurred depending on the type of accelerating admixture. The results are shown in Tables 7 and 8. Using low-alkali cement, the difference in expansion at 14 days between the plain specimen, and the specimen using the alkali-free or cement-based mineral accelerating admixture, was not significant at the 95% confidence level. In contrast, the expansion in the plain specimen and in the specimens using cement-based mineral, alkali-free and aluminate and cement-based mineral, and aluminate accelerating admixtures were significant at the 95% confidence level. After 84 days, the difference in expansion was not significant at the 95% confidence level, except for the expansion of the plain specimen and of that using the aluminate accelerating admixture. When the high-alkali cement was used (Table 8), the differences in the expansion of the plain specimen and the specimens using the alkali-free accelerating admixture, the specimens using the alkali-free and cement-based mineral, and the cement-based mineral and aluminate accelerating admixture, were not significant at the 95% confidence level. After 84 days, however, the differences using the low-alkali cement were significant in all cases, except in the expansion of the plain specimen and the specimen using the alkali-free accelerating admixture.
In summary, no actual difference was observed in expansion over time depending on the type of accelerating admixture using the low-alkali cement, while a difference was seen using the high-alkali cement. Considering the main sources of the alkali (i.e. the accelerating admixture and cement), cement had less impact on the ASR than the accelerating admixture, although considering only the equivalent alkali content of the cement, it can make a difference. Fig. 4 is the result of the analysis of expansion at 14 days, which showed that the tendency to expand increased with the alkali content of the cement. Fig. 5 compares the expansion at 14 days. The difference depending on the alkali content of the cement was not large between the plain specimen and the specimens using the alkali-free accelerating admixture, and between the specimens using the alkali-free and cement-based mineral accelerating admixtures. The analysis of the results at 84 days indicated that the tendency to expand increased with the alkali content of the cement (Fig. 6). At 84 days, the difference in expansion depending on the alkali content of the cement was large comparing the plain specimen and the specimens using the alkali-free accelerating admixture (Fig. 7). In contrast, the expansion of the specimens made using the aluminate accelerating admixture differed markedly from the other specimens. The difference in expansion between the specimens was greater when high-alkali cement was used compared to low-alkali cement. 3.3. Discussion As the expansion continued to increase (Fig. 2), we tried to predict the expansion after 84 days. It is carried out based on the data of expansion and the results are Figs. 8 and 9. The expansion will
338
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
(a) Plain H
(b) H-AF
(c) H-CM
(b) H-AN
Fig. 9. Expansion prediction of the specimens using high-alkali cement (H: high-alkali cement, AF: alkali-free accelerating admixture, CM: cement-based mineral accelerating admixture, AN: aluminate accelerating admixture).
continue to increase according to the linear prediction. The curve prediction I shows that the expansion rate of increase will drop during a specific period but after that, it will rise again. However, the curve prediction II shows that the expansion rate of increase will drop or ended after specific day. In this study, it is expected that the expansion curve will be similar with curve II because reactive silica components which can cause the ASR are limited. The expansion curve due to ASR can be classified into two or three major sections; first section where swelling begins (in some cases, this section is not clearly separated), second section where the significant expansion occur, and third section where the expansion reaches a final plateau [21]. The final plateau of expansion curve has already mentioned in previous researches such as Nielsen [22] or Capra et al. [23]. However Carles-Gibergues and Cyr [21] showed the possibility of continuous expansion. Therefore further study is needed to confirm whether or not the expansion reaches a final plateau. 4. Conclusions This study analysed the effects of the accelerating admixture and cementitious composites on the ASR for shotcrete containing comparatively high alkali levels made with a reactive aggregate and sufficient moisture. The following conclusions were made: (1) At 14 days, the expansion test result suggested that the ASR occurred, as the expansion of every specimen exceeded 0.1%. The expansion increased with the equivalent alkali content
(2)
(3)
(4)
(5)
of the cement and accelerating admixture. The accelerating admixture had a greater effect on the ASR. The aluminate accelerating admixture had the greatest effect, as the expansion at 14 days exceeded 0.2%. The expansion was increased with the time. And at 84 days, the expansion was greater than 14 days as the expansion of every specimen exceeded 0.3%. Especially, the expansion of specimens using high-alkali cement and aluminate accelerating admixture exceeded 0.5%. Statistically, no difference existed according to the type of cement at the 95% confidence level, while a significant difference in expansion was observed according to the accelerating admixture. Using low-alkali cement, no difference was detected in expansion over time according to the type of accelerating admixture at the 95% confidence level, while the difference in expansion was greater using the high-alkali cement. It is expected that the expansion rate of increase will decline due to the limited reactive silica components which can cause the ASR. However, there is a possibility to continue the expansion. Therefore long-term observation is needed.
References [1] Broekmans Maarten ATM. Structural properties of quartz and their potential role for ASR. Mater Charact 2004;53:129–40.
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339 [2] Bulteel D, Rafaï N, Degrugilliers P, Garcia-Diaz E. Petrography study on altered flint aggregate by alkali–silica reaction. Mater Charact 2004;53:141–54. [3] Garcia-Diaz E, Riche J, Bulteel D, Vernet C. Mechanism of damage for alkali– silica reaction. Cem Concr Res 2006;36:395–400. [4] Mindess Sidney, Young J Francis, Darwin David. Concrete, 2nd ed. Prentice Hall International Editions; 2003. p. 144–49. [5] Mitchell Lyndon D, Beaudoin James J, Grattan-Bellew Patrick. The effects of lithium hydroxide solution on alkali silica reaction gels created with opal. Cem Concr Res 2004;34:641–9. [6] Bérubé MA, Duchesne J, Dorion JF, Rivest M. Laboratory assessment of alkali contribution by aggregates to concrete and application to concrete structures affected by alkali silica reactivity. Cem Concr Res 2002;32:1215–27. [7] Swamy RN, Al-Asali MM. New test methods for alkali–silica reaction. In: Grattan-Bellew PE, editor. Proceedings of the 7th international conference on alkali-aggregate reactions in concrete, Ottawa, Canada; 1987. p. 324–29. [8] Saccani A, Bonora V, Monari P. Laboratory short-term ASR: a contribution. Cem Concr Res 2001;31:739–42. [9] Shehata Medhat H, Thomas Michael DA. The role of alkali content of Portland cement on the expansion of concrete prisms containing reactive aggregates and supplementary cementing materials. Cem Concr Res 2010;40:569–74. [10] ASTM C 150/C 150M. Standard specification for portland cement; 2009. [11] Park Hea-Geun, Sung Sang-Kyoung, Park Chan-Gi, Won Jong-Pil. Influence of a C12A7 mineral-based accelerator on the strength and durability of shocrete. Cem Concr Res 2008;38:379–85. [12] Paglia C, Wombacher F, Bohni H, The influence of alkali-free. The influence of alkali-free and alkaline shotcrete accelerators within cement systems I. Characterization of the setting behaviour. Cem Concr Res 2001;31:913–8. [13] Paglia C, Wombacheri F, Bohni H. Influence of alkali-free and alkaline shotcrete accelerators within cement systems: hydration, microstructure, and strength development. ACI Mater J 2004;101:353–7.
339
[14] De Belie N, Grosse CU, Kurz J, Reinhardt HW. Ultrasound monitoring of the influence of different accelerating admixtures and cement types for shotcrete on setting and hardening behaviour. Cem Concr Res 2005;35:2087–94. [15] Park Hea-Geun, Lee Myeong-Sub, Won Jong-Pil, Kim Jae-Kwon. Performance of wet-mixed shotcrete with powder types cement mineral accelerator. J Korean Soc Civ Eng 2003;24:65–9. [16] Won Jong-Pil, Kim Hwang-Hee, Park Kyoung-Hoon, Lee Si-Won. Performance and mix proportion of high performance shotcrete accelerator type and silica fume replacement ratio for permanent support. J Korean Soc Civ Eng 2007;27:157–62. [17] Giebson Colin, Seyfarth Katrin, Jochen Stark. Influence of acetate and formatebased deicers on ASR in airfield concrete pavements. Cem Concr Res 2010;40:537–45. [18] TRB. Chemical reactions of aggregate in concrete, special report no. 31, Highway (Transportation) Research Board; 1958. p. 1–12. [19] Cement and Concrete Association of New Zealand. Alkali-aggregate reaction minimizing the risk of damage to concrete. Guidance notes and model specification clauses 63, cement and concrete association report no. TR3; 1991. [20] Schneider José F, Hasparyk Nicole P, Silva Denise A, Monterio Paulo JM. Effect of lithium nitrate on the alkali–silica reaction gel. J Am Ceram Soc 2008;91:3370–4. [21] Carles-Gibergues André, Cyr Martin. Interpretation of expansion curves of concrete subjected to accelerated alkali-aggregate reaction (AAR) tests. Cem Concr Res 2002;32:691–700. [22] Nielsen A. Alkali-aggregate reactions, strengthening or total collapse? The different effects of AAR on concrete structures. In: Proceeding of the 11th international conference on alkali-aggregate reaction in concrete, Quebec City; 2000. p. 1009–18. [23] Capra B, Sellier A, Bourdarot E. Anisotropic modelling of alkali-aggregate reaction in concrete. In: Proceeding of the 11th international conference on alkali-aggregate reaction in concrete, Quebec City; 2000. p. 929–38.
Contents lists available at SciVerse ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Experimental and statistical analysis of the alkali–silica reaction of accelerating admixtures in shotcrete Jong-Pil Won ⇑, Bo-Ra Choi, Jae-Wan Lee Department of Civil and Environmental System Engineering, Konkuk University, Seoul 143-701, Republic of Korea
a r t i c l e
i n f o
a b s t r a c t
Article history: Received 21 July 2011 Received in revised form 22 October 2011 Accepted 24 November 2011 Available online 30 December 2011
This study investigated the alkali–silica reaction of accelerating admixtures for shotcrete. Tests were performed with two types of cement (low- and high-alkali) and three types of accelerating admixture (alkali-free, cement-based mineral and aluminate). An expansion test was performed to determine the alkali–silica reactivity according to ASTM C 1260. The results showed that the expansion increased with the total equivalent alkali content of the specimens. To examine this finding statistically, analyses were conducted at the 95% confidence level. When the low-alkali cement was used, no difference appeared to exist in the expansion after a reasonable period of time regardless of the accelerating admixture. In contrast, when the high-alkali cement was applied, the expansion of the specimens varied with time, although no difference was observed between the expansions of the plain sample and the specimen using the alkali-free accelerating admixture. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: Alkali–silica reaction Accelerating admixture Cementitious composites Expansion Shotcrete Statistical analysis
1. Introduction A high alkali content, a reactive aggregate and moisture are essential for creating the alkali–silica reaction (ASR) [1]. If any one of these factors is absent, then the ASR does not occur. ASR can adversely affect the durability of concrete because the alkali– silica gel which is a reaction product can cause serious cracking and overall expansion. Its process can be expressed as follows [2,3]. (1) Silica tetrahedrons (SiO2) which are one of the main components of reactive aggregate break up due to the attack of hydroxyl ions (OH).
2SiO2 þ OH ! SiO5=2 þ SiO5=2 H (2) If the attack of hydroxyl ions is continued, silicate ions, 2 H2 SiO4 , H3SiO4, and small polymers are formed.
2
SiO5=2 þ OH þ 1=2H2 O ! H2 SiO4
In this process, swelling N(K)–S–H gel is created by combining 2 H2 SiO4 and sodium or potassium ion [4]. N(K)–S–H gel has a low viscosity and can easily move from the aggregates to other parts. The volume of it is increased because it attracts water due to osmosis. At this time, generated local pressure can cause cracking in the concrete. ⇑ Corresponding author. Tel.: +82 2 450 3750; fax: +82 2 2201 0907. E-mail address: [email protected] (J.-P. Won). 0950-0618/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.conbuildmat.2011.11.017
It is hard to predict the degree of expansion by ASR. Because it depends on the viscosity or stiffness of alkali–silica gel and there are many other factors (such as aggregate type and size, alkali content and temperature) that affect the ASR. In the case of alkali content, if it is increased, reactive silica components react more easily with hydroxyl ions so that the ASR is accelerated. The alkali is supplied mainly from cement or the specific environment, and can be supplied directly from additives, such as the aggregate used or deicing chemicals [5,6]. Swamy and Al-Asali [7] showed that the higher expansion occurred when the specimens were stored in a 4% NaCl than in water. Saccani et al. [8] used four types of sodium additives (NaOH, Na2SO4, NaCl, NaHCO3) to examine the effect of alkali content on the expansion due to ASR. As a result, when sodium additives were added, expansion is bigger than those not added. Shehata and Thomas [9] showed that the expansion was influenced by alkali content in cement. An analysis revealed that the equivalent alkali content (Na2Oeq = Na2O + 0.658K2O) of Type I Portland cements produced by eight companies in Korea was 0.81–1.14%. These values exceed 0.6%, below which the equivalent alkali content prevents the ASR from taking place [10]. Therefore, the ASR should occur if reactive aggregate and sufficient moisture are supplied. In addition, if an additive containing high alkali is used, the ASR will be greater [8,9]. Shotcrete is used to construct tunnels or underground structures [11]. A low amount of rebounding, good adhesiveness, good shooting, and quick formation of initial strength are basic requirements of shotcrete [11–14]. Therefore using of accelerating admixture is essential to meet this requirement [11]. The
331
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339 Table 1 Chemical composition of the cements. Alkali content of the cement (%)
SiO2 Al2O3 Fe2O3 CaO MgO SO3 K2O Na2O Na2Oeq
Low
High
20.51 4.86 3.38 62.22 2.57 2.52 1.07 0.11 0.81
20.85 4.74 3.18 61.95 2.73 2.40 1.50 0.15 1.14
Fig. 2. Expansion test results (L: low-alkali cement, H: high-alkali cement, AF: alkali-free accelerating admixture, CM: cement-based mineral accelerating admixture, AN: aluminate accelerating admixture).
Fig. 1. X-ray diffraction patterns for aggregate.
(a) After 14 days
Table 2 Equivalent alkali content of the accelerating admixtures. Alkali content of the accelerating admixtures (%)
K2O Na2O Na2Oeq
Alkali-free
Cement-based mineral
Aluminate
0.01 0.51 0.52
0.18 13.40 13.52
0.02 18.80 18.81
Table 3 Designation of each mixture. Type of accelerating admixture
None Alkali-free Cement-based mineral Aluminate
Alkali content of the cement Low
High
Plain L L-AF L-CM L-AN
Plain H H-AF H-CM H-AN
accelerating admixture for shotcrete is a supportive material that expedites the initial strength development. It is classified as a hardening accelerator, inorganic alkali, and can include salt and silicate or aluminate, alkali-free and cement-based mineral admixtures [15]. The silicate and aluminate accelerating admixtures were once used the most widely, but they have been banned in many countries because they are potentially harmful to human health. Consequently, a new harmless accelerating
(b) After 84 days Fig. 3. Comparison of expansion.
admixture that does not result in a long-term reduction in strength is needed. Therefore, alkali-free and cement-based mineral accelerating admixtures were developed [16]. Nevertheless, the silicate and aluminate accelerating admixtures are still used widely for economic reasons, and the ASR may occur because some of these contain high alkali levels. Although the effects of additive materials on the ASR have been investigated [17], no study has examined the effects of the accelerating admixture used for shotcrete.
332
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
Therefore, this study scrutinised the ASR of accelerating admixtures in cementitious composites when reactive aggregate and sufficient moisture are supplied. Furthermore, the influence of the equivalent alkali contents of the accelerating admixture and cement on expansion was examined.
Table 4 Descriptive statistic of expansion. Name of specimen
After 14 days
After 84 days
Mean
Standard deviation
Mean
Standard deviation
L L-AF L-CM L-AN H H-AF H-CM H-AN
0.101 0.137 0.152 0.205 0.123 0.148 0.178 0.227
0.047 0.005 0.026 0.010 0.031 0.030 0.032 0.020
0.381 0.407 0.512 0.483 0.405 0.430 0.487 0.548
0.076 0.022 0.080 0.008 0.033 0.020 0.010 0.015
2. Alkali–silica reaction test 2.1. Materials 2.1.1. Cement The Type I Portland cements of eight companies in Korea were classified according to the equivalent alkali content. Of these, the cements with the highest (1.14%) and lowest (0.81%) alkali contents were selected. The chemical composition of these cements is shown in Table 1.
Table 5 Analysis of variance for the cements and accelerating admixtures (after 14 days). Source
Sum of squares
df**
Mean square
F***
P****
Cement Accelerating admixture* Cement accelerating admixture Residual
0.002 0.034 0.000
1 3 3
0.002 0.011 5.833E5
3.186 14.752 0.075
0.093 0.000 0.973
0.012
16
0.001
–
–
Total
0.656
24
–
–
–
* ** *** ****
Statistically significant at the 95% level. df: degree of freedom. F: F-distribution. P: P-value, significance probability.
Table 6 Analysis of variance for the cements and accelerating admixtures (after 84 days). df**
Mean square
F***
P****
0.003 0.065 0.006
1 3 3
0.003 0.022 0.002
1.617 11.903 1.130
0.222 0.000 0.367
0.029 5.107
16 24
0.002 –
– –
– –
Source
Sum of squares
Cement Accelerating admixture* Cement Accelerating admixture Residual Total * ** *** ****
2.1.2. Aggregate Crushed andesite was used. This aggregate is classified as a reactive aggregate because it contains a large amount of amorphous silica components [4,18]. According to a study by the Cement and Concrete Association of New Zealand, the expansion behaviour varies with the type of andesite [19]. Fig. 1 shows the X-ray diffraction (XRD) result of aggregate used in this study. In figure, the highest point mean amorphous silica component [20]. In other words, if point is high, there are lots of amorphous silica components that can cause ASR. 2.1.3. Accelerating admixture for shotcrete The alkali-free, cement-based mineral and aluminate accelerating admixtures for shotcrete that are used the most widely were selected. The alkali content of each admixture was measured using inductively coupled plasma atomic emission spectrometry (ICP). As Table 2 shows, the cement-based mineral and aluminate accelerating admixtures had higher equivalent alkali contents. The accelerating admixtures used had lower potassium contents than the cement, but higher sodium contents. 2.2. Experimental procedure An expansion test was performed to determine the alkali–silica reactivity according to ASTM C 1260 using 440 g of cement, 990 g of aggregate and 206.8 g of water. The alkali-free, cement-based mineral and aluminate accelerating admixtures were mixed with 7%, 5% and 5% cement by weight, respectively, based on typical usage. The mixtures are listed in Table 3. Experimental procedure can be summarised as follows: Step 1. Three 25 25 285-mm specimens were made from each mixture. To measure the length change, a stud was installed at both ends of each specimen. Step 2. The specimens were cured for 24 h in air after moulding and first measured after removing them from the mould. Step 3. The specimens were cured in water at 80 ± 2 °C for 24 h. The zero length was measured within 15 s.
Statistically significant at the 95% level. df: degree of freedom. F: F-distribution. P: p-value, significance probability.
Table 7 Multiple comparisons of the accelerating admixtures (using low-alkali cement). Source
Mean difference
Std. error**
After 14 days Plain alkali-free Plain cement-based mineral* Plain aluminate* Alkali-free cement-based mineral Alkali-free aluminate* Cement-based mineral aluminate*
0.03633 0.05167 0.10433 0.01533 0.06800 0.05267
0.022169 0.022169 0.022169 0.022169 0.022169 0.022169
After 84 days Plain alkali-free Plain cement-based mineral Plain aluminate* Alkali-free cement-based mineral Alkali-free aluminate Cement-based mineral aluminate
0.02600 0.05833 0.10167 0.03233 0.07567 0.04333
0.041046 0.041046 0.041046 0.041046 0.041046 0.041046
* ** ***
Statistically significant at the 95% level. Std. error: standard error. P: p-value, significance probability.
P***
95% Confidence interval Lower bound
Upper bound
0.140 0.048 0.002 0.509 0.015 0.045
0.08745 0.10279 0.15545 0.06645 0.19912 0.10379
0.01479 0.00055 0.05321 0.03579 0.01688 0.00155
0.544 0.193 0.038 0.454 0.103 0.322
0.12065 0.15299 0.19632 0.12699 0.17032 0.13799
0.06865 0.03632 0.00701 0.06232 0.01899 0.05132
333
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339 Table 8 Multiple comparisons of the accelerating admixtures (using high-alkali cement). Source
Mean difference
Std. error**
After 14 days Plain alkali-free Plain cement-based mineral* Plain aluminate* Alkali-free cement-based mineral Alkali-free aluminate* Cement-based mineral aluminate
0.02500 0.05467 0.10333 0.02967 0.07833 0.04867
0.023384 0.023384 0.023384 0.023384 0.023384 0.023384
After 84 days Plain alkali-free Plain cement-based mineral* Plain aluminate* Alkali-free cement-based mineral* Alkali-free aluminate* Cement-based mineral aluminate*
0.02467 0.08167 0.14300 0.05700 0.11833 0.06133
0.017314 0.017314 0.017314 0.017314 0.017314 0.017314
P***
95% Confidence interval Lower bound
Upper bound
0.316 0.048 0.002 0.240 0.010 0.071
0.07892 0.10859 0.15726 0.08359 0.13226 0.10259
0.02892 0.00074 0.04941 0.02426 0.82441 0.00526
0.192 0.082 0.000 0.011 0.000 0.088
0.06459 0.12159 0.18293 0.09693 0.15826 0.10126
0.01526 0.84174 0.10307 0.01707 0.07841 0.02141
* Statistically significant at the 95% level. Std. error: standard error. P: p-value, significance probability.
(a) Plain
(b) Alkali-free
(c) Cement-based mineral
(d) Aluminate
Fig. 4. Expansion trends of the mixture using accelerating admixture according to the alkali content of the cement (after 14 days).
Step 4. After that, the specimens were cured in 1 N NaOH at 80 ± 1 °C and the length was measured at 3, 6, 9, 12, 14, 21, 28, 35, 42, 56, 70 and 84 days. The procedure was repeated three times in the same order as above.
2.3. Statistical procedure Analysis of Variance (ANOVA) was carried out based on the data by the expansion test. ANOVA is original method to compare the interaction effects. In other
334
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
(a) Plain – AF
(b) Plain – CM
(c) Plain – AN
(d) AF – CM
(e) AF – AN
(f) CM – AN
Fig. 5. Comparison of the expansion trend among mixtures (after 14 days).
words, it is the method to confirm whether or not the samples have similar distribution. In this study, twoway ANOVA was conducted because there were two variables as cement and accelerating admixture. The procedure of twoway ANOVA can be summarised as follows.
Step 1. To identify the interaction between two factors (cement and accelerating admixture). Step 2. If there is no interaction, carry out the multiple comparisons after separation the effect of two factors.
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
(a) Plain
(b) Alkali-free
(c) Cement-based mineral
(d) Aluminate
335
Fig. 6. Expansion trend of the mixture using accelerating admixture according to the alkali content of the cement (after 84 days).
Step 3. If there is interaction, remove it by using variable change and go back to step 2.
3. Results and discussion 3.1. Expansion test The results of the expansion tests according to the type of accelerating admixture are displayed in Fig. 2. In this figure, L and H are plain specimens using low-alkali and high-alkali cement respectively. And AF, CM, and AN are specimens using alkali-free, cement-based mineral, and aluminate accelerating admixture respectively. Therefore, LF means the specimens mixed the low-alkali cement and alkali-free accelerating admixture. This principle is applied to the other specimens. The plain specimens without an accelerating admixture suggest that the ASR occurred as the expansion exceeded 0.1% at 14 days. The expansion of the specimens made using high-alkali cement exceeded 0.2% after 28 days, which could have negative effects. After 35 days, the expansion of the specimens made using low-alkali cement also exceeded 0.2%. The expansion increased until 84 days and was greater using high- rather than low-alkali cement. The expansion of all specimens using the alkali-free accelerating admixture exceeded 0.1% after 14 days, suggesting that the ASR occurred. After 28 days, harmful expansion exceeding 0.2% was seen. As indicated in figure, using the alkali-free accelerating
admixture, the expansion was similar regardless of the equivalent alkali content of the cement. In the case of the specimens using the cement-based mineral accelerating admixture, the expansion was greater as this contained more alkali than the alkali-free accelerating admixture. Again, the expansion increased with time. When the cement-based mineral accelerating admixture was used with high-alkali cement, the expansion was greater than with the low-alkali cement. Using the aluminate accelerating admixture, harmful expansion exceeding 0.2% occurred at 14 days regardless of the equivalent alkali content of the cement. These specimens expanded the quickest. Using the low-alkali cement, 56 days were required for the expansion to exceed 0.4% versus 84 days with the cement-based mineral accelerating admixture. The expansion values after 14 and 84 days are shown in Fig. 3. As you can see, the expansion is greater using high-alkali cement or accelerating admixture. Based on our results, the expansion increased with the equivalent alkali content of the specimen. The equivalent alkali content of the accelerating admixture had more influence on the ASR than the cement, since the difference in expansion depended more on the equivalent alkali content of the accelerating admixture. Sodium ions appear to affect the ASR more than potassium ions, since the accelerating admixture contained more sodium ions than the cement did, as shown in Tables 1 and 2. In addition, the expansion increased with time. Alkali is believed to have been supplied endlessly from the NaOH solution. Therefore, it is expected that the expansion will increase until all
336
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
(a) Plain – AF
(b) Plain – CM
(c) Plain – AN
(d) AF – CM
(e) AF – AN
(f) CM – AN
Fig. 7. Comparison of the expansion trend among mixtures (after 84 days).
amorphous silica components in reactive aggregate are consumed due to the ASR. 3.2. Statistical analysis Statistical analysis of the expansion at 14 and 84 days was used to confirm whether the difference in the equivalent alkali content
of the cement and accelerating admixture caused the actual difference in expansion. At first, analysis of descriptive statistic was conducted. The descriptive statistic is the method which can indicate the characteristics of data and the result is shown in Table 4. Table 5 shows the result of an Analysis of Variance (ANOVA) of the expansion at 14 days. No interaction occurred between the two
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
(a) Plain L
(b) L-AF
(c) L-CM
(b) L-AN
337
Fig. 8. Expansion prediction of the specimens using low-alkali cement (L: low-alkali cement, AF: alkali-free accelerating admixture, CM: cement-based mineral accelerating admixture, AN: aluminate accelerating admixture).
factors (the cement and the accelerating admixture) at the 95% confidence level. Therefore, the effect of each factor on the expansion was examined: cement was not significant at the 95% confidence level, while the accelerating admixture had a significant effect. The results of the ANOVA for the expansion 84 days matched this finding. As indicated in Table 6, no interaction took place between the effects of the cement and accelerating admixture on the expansion at the 95% confidence level. Only the accelerating admixture had a significant effect on the expansion. Then, we analysed whether a difference in expansion occurred depending on the type of accelerating admixture. The results are shown in Tables 7 and 8. Using low-alkali cement, the difference in expansion at 14 days between the plain specimen, and the specimen using the alkali-free or cement-based mineral accelerating admixture, was not significant at the 95% confidence level. In contrast, the expansion in the plain specimen and in the specimens using cement-based mineral, alkali-free and aluminate and cement-based mineral, and aluminate accelerating admixtures were significant at the 95% confidence level. After 84 days, the difference in expansion was not significant at the 95% confidence level, except for the expansion of the plain specimen and of that using the aluminate accelerating admixture. When the high-alkali cement was used (Table 8), the differences in the expansion of the plain specimen and the specimens using the alkali-free accelerating admixture, the specimens using the alkali-free and cement-based mineral, and the cement-based mineral and aluminate accelerating admixture, were not significant at the 95% confidence level. After 84 days, however, the differences using the low-alkali cement were significant in all cases, except in the expansion of the plain specimen and the specimen using the alkali-free accelerating admixture.
In summary, no actual difference was observed in expansion over time depending on the type of accelerating admixture using the low-alkali cement, while a difference was seen using the high-alkali cement. Considering the main sources of the alkali (i.e. the accelerating admixture and cement), cement had less impact on the ASR than the accelerating admixture, although considering only the equivalent alkali content of the cement, it can make a difference. Fig. 4 is the result of the analysis of expansion at 14 days, which showed that the tendency to expand increased with the alkali content of the cement. Fig. 5 compares the expansion at 14 days. The difference depending on the alkali content of the cement was not large between the plain specimen and the specimens using the alkali-free accelerating admixture, and between the specimens using the alkali-free and cement-based mineral accelerating admixtures. The analysis of the results at 84 days indicated that the tendency to expand increased with the alkali content of the cement (Fig. 6). At 84 days, the difference in expansion depending on the alkali content of the cement was large comparing the plain specimen and the specimens using the alkali-free accelerating admixture (Fig. 7). In contrast, the expansion of the specimens made using the aluminate accelerating admixture differed markedly from the other specimens. The difference in expansion between the specimens was greater when high-alkali cement was used compared to low-alkali cement. 3.3. Discussion As the expansion continued to increase (Fig. 2), we tried to predict the expansion after 84 days. It is carried out based on the data of expansion and the results are Figs. 8 and 9. The expansion will
338
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339
(a) Plain H
(b) H-AF
(c) H-CM
(b) H-AN
Fig. 9. Expansion prediction of the specimens using high-alkali cement (H: high-alkali cement, AF: alkali-free accelerating admixture, CM: cement-based mineral accelerating admixture, AN: aluminate accelerating admixture).
continue to increase according to the linear prediction. The curve prediction I shows that the expansion rate of increase will drop during a specific period but after that, it will rise again. However, the curve prediction II shows that the expansion rate of increase will drop or ended after specific day. In this study, it is expected that the expansion curve will be similar with curve II because reactive silica components which can cause the ASR are limited. The expansion curve due to ASR can be classified into two or three major sections; first section where swelling begins (in some cases, this section is not clearly separated), second section where the significant expansion occur, and third section where the expansion reaches a final plateau [21]. The final plateau of expansion curve has already mentioned in previous researches such as Nielsen [22] or Capra et al. [23]. However Carles-Gibergues and Cyr [21] showed the possibility of continuous expansion. Therefore further study is needed to confirm whether or not the expansion reaches a final plateau. 4. Conclusions This study analysed the effects of the accelerating admixture and cementitious composites on the ASR for shotcrete containing comparatively high alkali levels made with a reactive aggregate and sufficient moisture. The following conclusions were made: (1) At 14 days, the expansion test result suggested that the ASR occurred, as the expansion of every specimen exceeded 0.1%. The expansion increased with the equivalent alkali content
(2)
(3)
(4)
(5)
of the cement and accelerating admixture. The accelerating admixture had a greater effect on the ASR. The aluminate accelerating admixture had the greatest effect, as the expansion at 14 days exceeded 0.2%. The expansion was increased with the time. And at 84 days, the expansion was greater than 14 days as the expansion of every specimen exceeded 0.3%. Especially, the expansion of specimens using high-alkali cement and aluminate accelerating admixture exceeded 0.5%. Statistically, no difference existed according to the type of cement at the 95% confidence level, while a significant difference in expansion was observed according to the accelerating admixture. Using low-alkali cement, no difference was detected in expansion over time according to the type of accelerating admixture at the 95% confidence level, while the difference in expansion was greater using the high-alkali cement. It is expected that the expansion rate of increase will decline due to the limited reactive silica components which can cause the ASR. However, there is a possibility to continue the expansion. Therefore long-term observation is needed.
References [1] Broekmans Maarten ATM. Structural properties of quartz and their potential role for ASR. Mater Charact 2004;53:129–40.
J.-P. Won et al. / Construction and Building Materials 30 (2012) 330–339 [2] Bulteel D, Rafaï N, Degrugilliers P, Garcia-Diaz E. Petrography study on altered flint aggregate by alkali–silica reaction. Mater Charact 2004;53:141–54. [3] Garcia-Diaz E, Riche J, Bulteel D, Vernet C. Mechanism of damage for alkali– silica reaction. Cem Concr Res 2006;36:395–400. [4] Mindess Sidney, Young J Francis, Darwin David. Concrete, 2nd ed. Prentice Hall International Editions; 2003. p. 144–49. [5] Mitchell Lyndon D, Beaudoin James J, Grattan-Bellew Patrick. The effects of lithium hydroxide solution on alkali silica reaction gels created with opal. Cem Concr Res 2004;34:641–9. [6] Bérubé MA, Duchesne J, Dorion JF, Rivest M. Laboratory assessment of alkali contribution by aggregates to concrete and application to concrete structures affected by alkali silica reactivity. Cem Concr Res 2002;32:1215–27. [7] Swamy RN, Al-Asali MM. New test methods for alkali–silica reaction. In: Grattan-Bellew PE, editor. Proceedings of the 7th international conference on alkali-aggregate reactions in concrete, Ottawa, Canada; 1987. p. 324–29. [8] Saccani A, Bonora V, Monari P. Laboratory short-term ASR: a contribution. Cem Concr Res 2001;31:739–42. [9] Shehata Medhat H, Thomas Michael DA. The role of alkali content of Portland cement on the expansion of concrete prisms containing reactive aggregates and supplementary cementing materials. Cem Concr Res 2010;40:569–74. [10] ASTM C 150/C 150M. Standard specification for portland cement; 2009. [11] Park Hea-Geun, Sung Sang-Kyoung, Park Chan-Gi, Won Jong-Pil. Influence of a C12A7 mineral-based accelerator on the strength and durability of shocrete. Cem Concr Res 2008;38:379–85. [12] Paglia C, Wombacher F, Bohni H, The influence of alkali-free. The influence of alkali-free and alkaline shotcrete accelerators within cement systems I. Characterization of the setting behaviour. Cem Concr Res 2001;31:913–8. [13] Paglia C, Wombacheri F, Bohni H. Influence of alkali-free and alkaline shotcrete accelerators within cement systems: hydration, microstructure, and strength development. ACI Mater J 2004;101:353–7.
339
[14] De Belie N, Grosse CU, Kurz J, Reinhardt HW. Ultrasound monitoring of the influence of different accelerating admixtures and cement types for shotcrete on setting and hardening behaviour. Cem Concr Res 2005;35:2087–94. [15] Park Hea-Geun, Lee Myeong-Sub, Won Jong-Pil, Kim Jae-Kwon. Performance of wet-mixed shotcrete with powder types cement mineral accelerator. J Korean Soc Civ Eng 2003;24:65–9. [16] Won Jong-Pil, Kim Hwang-Hee, Park Kyoung-Hoon, Lee Si-Won. Performance and mix proportion of high performance shotcrete accelerator type and silica fume replacement ratio for permanent support. J Korean Soc Civ Eng 2007;27:157–62. [17] Giebson Colin, Seyfarth Katrin, Jochen Stark. Influence of acetate and formatebased deicers on ASR in airfield concrete pavements. Cem Concr Res 2010;40:537–45. [18] TRB. Chemical reactions of aggregate in concrete, special report no. 31, Highway (Transportation) Research Board; 1958. p. 1–12. [19] Cement and Concrete Association of New Zealand. Alkali-aggregate reaction minimizing the risk of damage to concrete. Guidance notes and model specification clauses 63, cement and concrete association report no. TR3; 1991. [20] Schneider José F, Hasparyk Nicole P, Silva Denise A, Monterio Paulo JM. Effect of lithium nitrate on the alkali–silica reaction gel. J Am Ceram Soc 2008;91:3370–4. [21] Carles-Gibergues André, Cyr Martin. Interpretation of expansion curves of concrete subjected to accelerated alkali-aggregate reaction (AAR) tests. Cem Concr Res 2002;32:691–700. [22] Nielsen A. Alkali-aggregate reactions, strengthening or total collapse? The different effects of AAR on concrete structures. In: Proceeding of the 11th international conference on alkali-aggregate reaction in concrete, Quebec City; 2000. p. 1009–18. [23] Capra B, Sellier A, Bourdarot E. Anisotropic modelling of alkali-aggregate reaction in concrete. In: Proceeding of the 11th international conference on alkali-aggregate reaction in concrete, Quebec City; 2000. p. 929–38.