- Email: [email protected]
Properties of blended cements with thermally activated kaolin
Available online at www.sciencedirect.com
Construction and Building
MATERIALS
Construction and Building Materials 23 (2009) 62–70
www.elsevier.com/locate/conbuildmat
Properties of blended cements with thermally activated kaolin Metin Arikan a, Konstantin Sobolev b,*, Tomris Ertu¨n c, Asim Yeg˘inobali c, Pelin Turker c b
a Civil Engineering Department, Middle East Technical University (METU), Turkey Department of Civil Engineering and Mechanics, University of Wisconsin-Milwaukee (UWM), USA c Cement and Concrete Research Institute of TCMA, Turkey
Received 5 February 2008; accepted 5 February 2008 Available online 9 April 2008
Abstract Kaolin, one of the materials of major importance for the ceramic and paper industry, is also used in the construction industry as a raw material for the production of white cement clinker and, in the form of metakaolin, as an artificial pozzolanic additive for concrete. Metakaolin is a vital component of high-performance and architectural concrete; however, its application in regular concrete is very limited due to relatively high production costs. This report evaluates the performance of a low-cost metakaolin-based additive called thermally activated kaolin (TAK), in cement. Due to its pozzolanic properties and the densification of cement matrix, the application of TAK provides a 15% improvement of the compressive strength. It was shown that TAK of optimal quality can be manufactured by the thermal treatment of raw kaolin with 74% of kaolinite at 750 °C without the intermediate beneficiation stage. The application of a developed approach can significantly reduce production expenditures and make the application of such an additive feasible even in regular-grade cement and concrete. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Kaolin; Metakaolin; Additive; Thermal activation; Cement; Compressive strength; Microstructure
1. Introduction There is an ongoing interest in the use of selected clay minerals, including kaolinites, in the construction industry. A recent development comprises the application of metakaolin as an artificial pozzolanic additive for concrete [1–7]. The strength and durability of conventional cement-based materials can be significantly improved by using thermally activated kaolin additives [1–8]. Such additives are conventionally manufactured by firing the high-grade purified kaolinite at 650–850 °C according to the following reaction [8,9]: 650850 C
2SiO2 Al2 O3 2H2 O ! 2SiO2 Al2 O3 þ 2H2 O " The main beneficial effect of metakaolin in concrete and cement is related to its high pozzolanic activity (i.e., the ability to react with portlandite, Ca(OH)2 released during *
Corresponding author. E-mail address: [email protected] (K. Sobolev).
0950-0618/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.conbuildmat.2008.02.008
the hydration of portland cement) [1–8]. The application of the super-fine particles of metakaolin results in a microfiller effect and improves the packing of the cement matrix. The micro-bearing’s effect is provided by the flaky particles of metakaolin, resulting in the better sliding of more coarse cement particles and, therefore, facilitating the flow of the system. Furthermore, metakaolin improves the morphology of the interface zone between the cement matrix and the aggregate’s surface [5–8]. Both very light, attractive color shades, and an improved strength and durability of metakaolin-based concrete make such an additive one of the vital components of modern architectural concrete [1,6,9]. Shvarzman et al. demonstrated that useful properties of metakaolin are preserved even at a reduced content (down to 30%) of kaolinite in the raw mix [8]. Based on this study, it was proposed that clays containing more than 35% of kaolinite can be directly processed into pozzolanic additives using thermal activation, eliminating the expensive stage of beneficiation [9]. Realization of such an approach
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
63
3. Experimental program
to a surface area of about 300 m2/kg. The chemical composition of these cements (as determined by the X-ray fluorescence, XRF), is reported in Table 1. It can be observed that these cements had similar composition; however, NPC had somehow lower content of alkalis. The raw kaolin specimens were obtained from Ku¨tahya-Kizilcßukur quarry (KK), Ku¨tahya-Ulasßlar quarry (KU), and S ß ile-DomaliMatel quarry (KS), Turkey [10–12]. The investigated samples were composed mainly of quartz and kaolinite with low to medium crystallinity as defined by the Hinckley Index [10]. The chemical composition of the kaolin based materials used in research was analyzed using the XRF technique and reported in Table 1. According to the XRF analysis, a sample of KK contained a high amount of Al2O3 (31.5%, within the usual range of 15–30%) and a relatively low amount of SiO2 (55%, within the common range of 55–80%). This specimen was predominantly composed of kaolinite and quartz as determined by the XRD (X-ray diffraction, Fig. 1), SEM (scanning electron microscopy) and EDS (energy dispersive spectrometry) [10]. The results of DTA-TG (differential thermal analysis/thermogravimetric analysis) investigation demonstrated the presence of the dehydroxylation endotherm at 580 °C with a characteristic weight loss of 9.2% that corresponds to 73.6% of kaolinite (Fig. 2). Due to the presence of quartz and tridymite minerals, a relatively high content of SiO2 (67.0%) was detected for sample KU (Fig. 1). For KU specimen, the kaolin–metakaolin conversion endotherm was observed at 564 °C with a weight loss of 6.7%, Fig. 2; this provides the estimated kaolinite content of 53.6%. The sample of KS had intermediate contents of Al2O3 and SiO2 with lowest level of kaolinite, 41.6%. The detailed characterization of investigated raw kaolins was reported by Aras et al. [10]. Based on their chemical analysis, all obtained TAK are classified as Class N according to ASTM C349 [13].
3.1. Materials used
3.2. Research program
could significantly reduce the production expenditures related to intermediate wet beneficiation and the subsequent drying of the raw kaolin. Furthermore, the waste streams generated during these stages can be also eliminated. Consequently, thermally activated kaolin (TAK, to differentiate from its metakaolin sibling, which is made from purified kaolin) can be made available at a significantly reduced cost, making feasible the application of such an additive even in a regular-grade cement and concrete. The properties of such products can be tailored to provide the improved strength and durability. To verify this proposal, an extensive research program involving five R&D and educational institutions (METU, LAU, TCMA, MTA, TECHNION) was initiated and supported by the TUBITAK (Scientific and Research Council of Turkey) [9]. The detailed results on the characterization of available raw kaolin sources in Turkey were reported by Aras et al. [10] and Sobolev et al. [11] described the thermal activation and process parameters for the production of TAK. 2. Research significance Metakaolin is a well-known additive for the improvement of concrete strength and durability. The application and performance of metakaolin is well covered in the scientific literature. However, there are limited references to the performance of cement and concrete additives based on low-grade or unprocessed kaolin. Therefore, the investigation of strength, microstructure and other performance parameters of cement containing thermally activated additives based on raw (unprocessed) kaolin was considered important. This research deals with such an evaluation.
Three raw kaolin specimens with different kaolinite content (41.6–73.6%) and two reference portland cements CEM-I 52.5, NPC and CEM-I 32.5, LPC (ASTM Type III and Type I, respectively) were used in the research program. NPC was a commercial product and LPC was laboratory made cement obtained by intergrinding of 95% clinker and 5% of natural gypsum in a laboratory ball mill
The experimental program involved three main tasks: thermal activation of kaolin at 750 °C for different treatment durations (1–2 h); the investigation of the properties of obtained TAK: mineralogical composition (XRD); fineness, particle size distribution, whiteness, morphology (using SEM);
Table 1 Chemical analysis of materials used in the research Type
SiO2 (%)
Al2O3 (%)
Fe2O3 (%)
TiO2 (%)
CaO (%)
MgO (%)
SO3 (%)
K2O (%)
Na2O (%)
MnO2 (%)
P2O5 (%)
L.O.I. (%)
Kaolinite* (%)
NPC LPC KK KU KS
19.78 19.47 55.0 67.0 57.3
5.25 5.01 31.5 21.5 26.7
3.24 3.07 0.3 0.3 2.7
– – 0.1 0.5 1.2
63.45 63.95 0.1 0.2 0.1
1.21 1.57 <1.0 <1.0 0.7
2.45 2.63 – – –
0.72 0.84 <1.0 0.4 2.5
0.32 0.47 <1.0 <1.0 0.1
– – <1.0 <1.0 <1.0
– – 0.2 0.2 <1.0
1.60 1.24 10.85 9.40 7.90
– – 73.6 53.6 41.6
*
Calculated according to the weight loss at 500–750 °C corresponding to dehydroxylation of kaolinite.
64
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
H
K
QT
KQ A
QA
Q
K
K Q A T H
: Kaolinite : Quartz : Alunite : Tridymite : Halloysite
KK
KU
KS
KK—1/2 hour
KK—1 hour
KK—2 hours
10
20
30
40
50
2Θ
60
Fig. 1. The results of XRD analysis of raw kaolin and selected (KK – based) TAK specimens activated at 750 °C.
Weight Loss
Heat Flow
410 µV*min/g 993 °C -2.93 mV*min/g 580 °C
KK
Δ = -9.2 %
KK
278 µV*min/g 994 °C
KU
-1.93 mV*min/g 564 °C
Δ = -6.7 %
KU Δ = -1.4 %
KS 141 µV*min/g 950 °C -1.37 mV*min/g 553 °C
0
200
400
Δ = -5.2 %
600
800
1000
0
Temperature, °C
200
400
600
800
KS
1000
Temperature, °C
Fig. 2. The results of DTA-TG investigation of kaolin specimens.
the investigation of properties of blended TAK cement: setting times, soundness, heat of hydration, microstructure (using SEM), compressive and flexural strength (as pozzolanic activity).
3.3. Notations used The following notations were used to distinguish the investigated samples:
NPC reference portland cement; KK raw kaolin, from Ku¨tahya-Kizilcßukur; KU raw kaolin, from Ku¨tahya-Ulasßlar; KS raw kaolin, from S ß ile-Domali-Matel.
All TAK specimens were designated with an additional number after the kaolin type notation. This number corresponds to the duration of thermal treatment; for example, KK-1 and KK-2 correspond to TAK based on KK and activated at 750 °C for 1–2 h, respectively.
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
3.4. Mixture proportioning The properties of 14 cements were investigated. These included two reference portland cements (NPC and LPC) and composite cements with 20% of raw kaolin or TAK. The 20% cement substitute level was also selected to assess the pozzolanic activity of mineral additives (as prescribed by ASTM C311) [14]. The mortars were designed using a sand-to-cement ratio (S/C) of 3.0 and W/C of 0.5 according to EN 196 and tested according to ASTM C109, ASTM C311, ASTM C349, and ASTM C618 [13–17]. 3.5. Preparation of specimens Prior to its use in the research program, the raw kaolin was pre-ground in a ball mill for 60 min and the remaining coarse particles were removed by screening on a 90 lm sieve. Raw kaolin was thermally activated in a muffle furnace at 750 °C for 1–2 h respectively, resulting in the TAK samples. Based on preliminary investigation [11], for majority of raw kaolin specimens, the temperature of 750 °C was found to be sufficient for kaolin–metakaolin conversion within 1 h. However, kaolinite-rich specimens usually require longer treatment times (up to 2 h) or higher temperatures [11]; these observations are also supported by the results of XRD and DTA-TG investigation (Figs. 1 and 2). The composite cements with 20% of mineral additive were obtained by dry intermixing in the standard lab mixer, a process that represents the industrial approach applied by
65
many cement companies in Europe. The investigated mortars were produced following EN 196 [13]. The mortars were cast into three-gang (40 40 160 mm) prism molds, and compacted in accordance with EN 196. For SEM investigations cement pastes were prepared following the same mixing and molding method using NPC and 20% of mineral additive based on KK at W/C of 0.3. 3.6. Curing of specimens After the compaction procedure, the molds were placed in a humidity cabinet for 24 h (with a relative humidity of 95% and a temperature of 20 ± 1 °C). Following this period, the specimens were removed from the molds and kept in water until the testing age. The cement pastes prepared for SEM investigations were cured for 7- and 28-day at 20 ± 1 °C. After the curing period, the fractured specimens were dried, treated with acetone and coated with a thin layer of gold. 3.7. Tests performed The particle size distribution of investigated cements was measured by a Masersizer laser diffraction analyzer (Malvern Instruments). The resulting particle size distribution and other characteristics of investigated materials (i.e., cement, kaolin and TAK) are summarized in Fig. 3 and Table. 2. The properties of developed cements (setting times, soundness, heat of hydration, compressive and flexural strength) were obtained using EN 196 [15]. These
Fig. 3. Particle size distribution of investigated materials.
Table 2 Physical properties of selected cements and raw kaolin/TAK Performance characteristics Blaine specific surface area (m2/kg) Median size (lm) Heat of hydration (cal/g) Whiteness (%)
Reference cements
Blended cements based on NPC
NPC
LPC
KK
KK-1
KK-2
Raw kaolin/TAK KK
KK-1
KK-2
356 22 70.5 –
318 27 – –
362 25 78.9 –
392 20 94.7 –
397 20 100.4 –
422 26 – 77.6
474 15 – 81.5
489 14 – 82.4
66
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
results are presented in Table. 3. The mortar samples were tested at the age of 7, 28 and 90 days for flexure and compression. Compressive strength tests were conducted using the portions of three prisms broken in flexure. Therefore, the reported compressive strength results are the average of the six values [15]. SEM observations on the cement paste samples were performed at 7- and 28-day using a LEO scanning electron microscope operated at an accelerating voltage of 15 kV.
much higher than that of raw kaolin (Table 2). The obtained TAK is characterized by a three-modal distribution (Fig. 3). The size reduction associated with such thermal treatment led to an increase in the whiteness of TAK from 77.6% for raw kaolin to 81.5% and 82.4% for samples KK-1 and KK-2, respectively (Table 2). The cements based on TAK had smaller median particle size and higher surface area (Table 2). 4.3. Morphology of TAK
4. Results and discussion 4.1. The effect of thermal activation It was demonstrated that thermal activation at 750 °C is adequate to complete the kaolin–metakaolin transformation (Figs. 1 and 2). Thermal activation for 2 h results in complete disappearance of kaolinite peaks on XRD, which is clear sign of its completed conversion to metakaolin. However, even 1 h treatment is sufficient to complete the transformation with very little kaolinite left in the system. Treatment for ½ h was less effective, leaving significant quantities (about 50%) of unconverted kaolinite; therefore, such activation was not considered for production of TAK.
The comparison of surface features and morphology of investigated raw kaolin and TAK (using KK samples) were performed using the SEM technique (Fig. 4). It was observed that raw kaolin particles are of the predominant size of 5–50 lm, with agglomerates clearly seen around the quartz particles (Fig. 4a). Due to thermal treatment at 750 °C and subsequent recrystallization to metakaolin, the particles of kaolin are separated from quartz and significantly reduced in size. The resulting TAK is characterized by the particles of the predominant size of 1–15 lm (Fig. 4b). SEM observations of KK-2 morphology demonstrated the presence of star-like agglomerates of low-density metakaolin (Fig. 4c). 4.4. Effect of TAK on properties of cement
4.2. Particle size distribution of TAK The particle size distribution of investigated materials (observed by means of laser diffraction analysis) is presented in Fig. 3. It can be observed that the particles of raw kaolin are somehow larger compared with portland cement. The particle size distribution of raw kaolin is characterized by a bi-modal distribution. It was found that due to thermal treatment at 750 °C and its recrystallization to metakaolin, the kaolin particles are significantly reduced in size (with the predominant size of 1–15 lm). Due to this transformation, the Blaine specific surface area of TAK is
The replacement of the NPC with TAK (such as KK-1 and KK-2) at a dosage of 20% does not affect the final setting time of blended cements (Table 3) and the initial setting time of such cements was only slightly prolonged. However, the addition of raw kaolin and TAK to LPC cement results in delayed setting of blended cements. The soundness (tested according to le Chaˆtelier method of EN 196 [15]) of all investigated cements was at the same level of 1–2 mm for all the cements, except the reference cement LPC and blended cement based on NPC with 20% of KK-1 additive, these had zero soundness.
Table 3 Performance of the investigated cements Type of cement
Activation (h)
Normal consistency (%)
Setting time (min) Initial
Final
Soundness* (mm)
Compressive/flexural strength (MPa) 7 days
28 days
90 days
NPC NPC–KK NPC–KK-1 NPC–KK-2
– – 1 2
25.4 28.4 29.3 28.0
155 195 170 165
200 245 205 195
1 1 0 1
39.6/5.4 37.4/4.9 45.2/5.7 41.1/5.6
53.4/6.6 51.5/6.2 61.7/6.6 59.7/6.7
58.4/6.8 55.4/6.4 65.8/6.9 62.9/6.8
LPC LPC–KK LPC–KK-1 LPC–KK-2 LPC–KU LPC–KU-1 LPC–KU-2 LPC–KS LPC–KS-1 LPC–KS-2
– – 1 2 – 1 2 – 1 2
25.6 26.8 27.1 27.0 28.0 27.4 27.3 26.9 27.1 26.9
145 175 185 215 195 200 190 200 215 195
190 215 225 250 240 245 230 235 250 230
0 1 1 2 2 1 1 2 2 1
31.2/4.5 23.0/3.9 28.0/4.0 28.0/4.1 23.7/4.0 25.6/3.9 27.2/4.1 22.3/3.8 30.3/4.4 29.6/4.7
39.2/5.6 26.9/4.2 39.0/5.2 41.4/5.3 30.0/4.3 38.0/5.2 39.4/5.5 26.0/3.8 38.7/5.4 38.2/5.2
42.7/6.0 29.9/4.4 41.8/5.8 44.1/6.3 37.6/5.6 44.1/6.3 44.9/6.3 27.4/4.9 40.8/5.6 40.4/5.7
*
Tested according to le Chaˆtelier method of EN 196 [15].
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
67
observed that with hydration of the reference portland cement (NPC), the unhydrated clinker grains are surrounded by radiating fibers of C–S–H resembling the pattern of Type I C–S–H. Randomly oriented CH crystals and the prismatic ettringite crystals are widely dispersed through the paste (Fig. 5a, left). The microstructure of the hydrated paste of NPC at the age of 28 days is presented by the amorphous gel filling the spaces between the hydrated particles. In NPC pastes, the layered accumulations of CH crystals of about 15 lm in width are intermingled through the paste (Fig. 5a, right). It is observed by SEM that the KK and TAK particles are well dispersed within the cement paste, resulting in a dense structure with low porosity (Fig. 5b–d). There is a visible densification around the TAK particles, due to the pozzolanic reaction of TAK, leading to the formation of C–S–H. It can be seen that at the age of 7 days the kaolin and TAK grains are already well covered with a pseudomorphic layer composed of hydration products. The matrix phase is composed mainly of short acicular outgrowths of C–S–H around the clinker grains and needleshaped ettringite crystals (Fig. 5c–d, left). At the age of 28 days the kaolin and TAK grains are well embedded into the matrix and are well connected through the C–S–H gel. It can be distinguished that cement pastes based on TAK, especially cement with KK-1, are of higher density comparing to reference cement and blended KK cement (Fig. 5). 4.6. Fresh properties of mortars In spite of somehow higher normal consistency of TAKcements (Table 3), all investigated mortars were of perfect workability allowing for easy casting of the specimens as per as EN 196. 4.7. Compressive and flexural strength of mortars based on NPC Fig. 4. SEM observations of kaolin morphology. (a) raw kaolin KK, (b) KK-1, (c) KK-2.
Based on the heat of hydration values (as per as EN 196, Table 2) it can be proposed that the addition of TAK results in significant acceleration of cement hydration process. Even 20% replacement of NPC with TAK (KKbased) results in 34% and 42% increase in heat of hydration in case of TAK activated for 1–2 h, respectively. This behavior can be explained by intensive pozzolanic reaction that consumes significant quantities of portlandite (i.e., consumption of CH would shift the hydration reaction towards intensification of hydration) and also by the formation of cement crystallization centers/sites on smaller particles of TAK. 4.5. Effect of TAK on microstructure The results of SEM investigation of the cement pastes hardened for 7- and 28-day are presented in Fig. 5a. It is
The compressive test results of investigated mortars (following EN 196) are presented in Table 3. The 7-day compressive strength of cement with KK-1 is the highest among the investigated cements, 45.2 MPa. The cement with KK-2 had a 7-day compressive strength of 41.1 MPa, which is close to the strength of reference NPC (39.6 MPa). The cement with raw kaolin KK had the lowest 7-day compressive strength of 37.4 MPa (Table 3). Acceleration of the strength development of TAKcement with optimized TAK, KK-1 at 7-day age by 14% can be explained by the pozzolanic reaction and the densification of cement matrix (as it was also observed by the SEM). According to the test results, the best 28-day compressive strength of 61.7 MPa was obtained in cement with TAK activated for 1 h (KK-1); this value is 15% higher than the strength of reference cement. The cement with TAK activated for 2 h (KK-2) reached a 28-day compressive strength of 59.7 MPa, which is only slightly less than that for KK-1.
68
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
Fig. 5. Effect of TAK on microstructure of cement paste at the age of 7 days (left) and 28 days (right).
Addition of TAK based on KK did not affect the flexural strength of investigated mortars in all ages of hardening (Table 3). Therefore, it can be stated that addition of TAK has higher effect on the improvement of compressive
strength. The cement with raw kaolin KK had a 28-day compressive strength of 51.5 MPa, which is somewhat close to the strength of reference NPC (53.4 MPa). The observed trend was also preserved at the age of 90 days with KK-1
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
cement gaining the highest compressive strength of 65.8 MPa (or 13% higher than the reference). 4.8. Compressive and flexural strength of mortars based on LPC Reference cement LPC had a lower strength in all ages of hardening vs. NPC (Table 3). This can be explained by lower Blaine specific surface area of LPC, as well as higher content of alkalis. Addition of TAK led to the reduction of the 7-day compressive strength of mortars based on LPC. For blended cements with TAK based on raw kaolin composed mainly of quartz and kaolinite mix (i.e., KK and KS), the reduction of 7-day strength was minor, about 10% or less. However, the specimen processed from raw kaolin containing tridymite had a significantly reduced strength at the age of 7 days. This trend was overcome at the 28-day age when all investigated cements with TAK had a compressive strength similar to that of reference cement LPC. Such performance level satisfies and exceeds the requirements of ASTM C618 set for Class N pozzolanic additives. After 90 days of hardening the strength of blended cements with TAK based on raw kaolin with more than 50% of kaolinite was higher than the strength of reference cement; an exception to this rule was the behavior of cement with KK-1, which had slightly lower 90-day strength that that of the reference. Such a strength reduction can be explained by the incomplete kaolin–metakaolin conversion (Figs. 1 and 2). Still, cements with 20% of low-kaolinite based TAK (KS-1/ KS-2 based on clay with 41.6% of kaolinite) demonstrated only 5% strength reduction vs. the reference LPC; such additives can also be considered as effective pozzolans. The flexural strength of mortars based on LPC followed fairly similar trends, usually keeping the compressive-toflexural strength ratio within the limits of 6.8–7.3. Strength of LPC – raw kaolin blended cements was low in all ages of hardening (about 20–30% less than the reference); based on the criteria set forth by the ASTM C618, such reduction of strength does not allow classifying the raw kaolins as pozzolanic additives. 5. Conclusions It was shown that effective mineral additives can be manufactured by the thermal activation of raw kaolin with at least 41.6% of kaolinite, corresponding to at least 21.5% of Al2O3 and less than 55% of SiO2. For the investigated kaolin types, the thermal activation at 750 °C for 1 h is sufficient to obtain the product of desired performance. The resulting mineral additive, called thermally activated kaolin (TAK), meets or exceeds the requirements of ASTM C618 set for Class N pozzolanic additives. It can be proposed that addition of TAK results in significant acceleration of cement hydration process due to intensive pozzolanic reaction and also due to the formation of cement crystallization centers on smaller particles of TAK.
69
Based on the conducted investigation, it can be concluded that the application of TAK manufactured from kaolin with a high kaolinite content (73.6%, such as KK) is very effective in blended cements based on high-strength cement (i.e., EN CEM-I 52.5 or ASTM Type III). The 28day compressive strength of such cements is improved by up to 15%. The compressive strength of blended cement with 20% of TAK (KK-1/KK-2) reaches the level of 65 MPa at 90-day age. Based on the microstructural investigation, the improvement of the strength of TAK-cement is attributed to the pozzolanic reaction and the densification of cement matrix due to the addition of TAK. It is proposed that the TAK-based additive provided quite a significant improvement of cement properties, which is similar to the effect of ‘‘purified,” commercial grade metakaolin. This effect can be further improved when the TAK additive is applied with an effective superplasticizer. The application of TAK is less effective when it was blended with laboratory-grade portland cement (i.e., LPC, classified as EN CEM-I 32.5 or ASTM Type I). In spite of the reduction of 7-day compressive strength of mortars based on LPC, the 28- and 90-day compressive strength of such cements was within the ±5% range of the reference cement strength. It was demonstrated that the addition of TAK does not affect the flexural strength of investigated mortars in all ages of hardening. It is demonstrated that the developed approach can significantly reduce the expenditures related to production of TAK due to the elimination of the steps of intermediate wet beneficiation and the subsequent drying of the raw materials that are common for commercial grade metakaolin additives. Furthermore, the proposed process can reduce the waste streams generated during these eliminated stages. Therefore, it is expected that TAK could be available at a significantly reduced cost, making its application feasible even in a regular-grade cement and concrete. Additional investigation may be necessary to explain and quantify the hydration mechanism and the long-term microstructural development of TAK cements containing different amounts of TAK. Further research is also required to investigate the behavior of TAK cements in concrete, as well as to investigate the durability of such materials. Acknowledgements This study was performed under TUBITAK research _ ¸ TAG – 680, in cooperation with MTA research Grant IC team represented by Dr. Mustafa Albayrak and Dr. Aydin Aras. The authors express their gratitude to Ipek S ß ener and Bahadir Erdogan for their help with preparation of the samples. Extremely useful comments of Prof. Konstantin Kovler (TECHNION) are highly appreciated. The financial contribution of TUBITAK, TCMB, LAU, MTA, CONACYT, PROMEP and PAICYT is gratefully acknowledged.
70
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
References [1] Kostuch JA, Walters GV, Jones TR. High performance concrete incorporating metakaolin—a review. Concrete 2000 1993;2:1799–811. [2] Wild S, Khatib J, Jones A. Relative strength, pozzolanic activity and cement hydration in superplasticized metakaolin concrete. Cem Concr Res 1996;26(10):1537–44. [3] Khatib J, Wild S. Sulfate resistance of metakaolin mortar. Cem Concr Res 1998;28(1):38–92. [4] Jones TR, Walters GV, Kostuch JA. Role of metakaolin in suppressing ASR in concrete containing reactive aggregate and exposed to saturated NaCl solution. In: Proceedings of the nineth international conference on alkali-aggregate reaction in concrete, vol. 1; 1992. p. 485–96. [5] Sabir BB, Wild S, Khatib J. On the workability and strength development of metakaolin concrete. In: Dhir RK, Dyer DT, editors. International congress on concrete in the service of mankind, concrete for environment enhancement and protection, theme 6, waste materials and alternative products, University of Dundee Spon, London; 1996. p. 651–62. [6] Caldarone MA, Gruber KA, Burg RG. High-reactivity metakaolin: A new generation mineral admixture. Concr Int 1994;16(11):35–40. [7] Sobolev K. White cement: problems of production and quality. Cem Concr World 2001(July–August):34–42. [8] Shvarzman A, Kovler K, Grader GS, Shter GE. The effect of dehydroxylation/amorphization degree on pozzolanic activity of kaolinite. Cem Concr Res 2003;33(3):405–16. [9] Arikan M, Sobolev K, Ertu¨n T, Yeg˘inobali A, Albayrak M, Aras A, et al. Development of cement and concrete additives based on _ _ ¸ TAG ¨ BITAK thermally activated kaolin. TU project proposal IC 2003;680:55 [in Turkish].
[10] Aras A, Albayrak M, Arikan M, Sobolev K. Evaluation of selected kaolins as a raw material for the Turkish cement and concrete industry. Clay Miner 2007;42(2):233–44. [11] Sobolev K, Arikan M, Ertu¨n T, Yeg˘inobali A, Erdogan B. The development of thermally activated kaolin admixture for highperformance blended cements. J. Am Ceram Soc submitted for publication. _ ¨ zel Ihtisas [12] Besß Yillik Kalkinma Plani: Madencilik O Komisyonu Raporu Endu¨striyel Hammaddeler Alt Komisyonu Toprak Sanayii Hammaddeleri – I (Seramik Killeri-Kaolen-Feldspat-Pirofillit-Wollastonit-Talk) C ¸ alisßma Grubu Raporu. Ankara: DPT. 2001. 204 pp _ 622). ISBN: 975-19-2837-0 [in Turkish].. [13] ASTM C349. Compressive strength of hydraulic cement mortars (using portions of prisms broken in flexure). Annual book of ASTM standards, vol. 04.02. West Conshohocken, Pennsylvania: ASTM; 2005. [14] ASTM C311. Standard methods of sampling and testing fly ash or natural pozzolans for use as a mineral admixture in portland cement concrete. American society for testing and materials. Annual book of ASTM standards, vol. 04.02. West Conshohocken, PA: ASTM; 2005. [15] EN 196. Test method for determining compressive strength of cement mortar. European Standard; 1994. [16] ASTM C109. Compressive strength of hydraulic cement mortars (using 2-in or 50-mm cube specimens). Annual book of ASTM Standards, vol. 04.02. West Conshohocken, PA: ASTM; 2005. [17] ASTM C618. Standard specification for coal fly ash and raw or calcined natural pozzolan for use in concrete. Annual book of ASTM standards, vol. 04.02. West Conshohocken, PA: ASTM; 2005.
Construction and Building
MATERIALS
Construction and Building Materials 23 (2009) 62–70
www.elsevier.com/locate/conbuildmat
Properties of blended cements with thermally activated kaolin Metin Arikan a, Konstantin Sobolev b,*, Tomris Ertu¨n c, Asim Yeg˘inobali c, Pelin Turker c b
a Civil Engineering Department, Middle East Technical University (METU), Turkey Department of Civil Engineering and Mechanics, University of Wisconsin-Milwaukee (UWM), USA c Cement and Concrete Research Institute of TCMA, Turkey
Received 5 February 2008; accepted 5 February 2008 Available online 9 April 2008
Abstract Kaolin, one of the materials of major importance for the ceramic and paper industry, is also used in the construction industry as a raw material for the production of white cement clinker and, in the form of metakaolin, as an artificial pozzolanic additive for concrete. Metakaolin is a vital component of high-performance and architectural concrete; however, its application in regular concrete is very limited due to relatively high production costs. This report evaluates the performance of a low-cost metakaolin-based additive called thermally activated kaolin (TAK), in cement. Due to its pozzolanic properties and the densification of cement matrix, the application of TAK provides a 15% improvement of the compressive strength. It was shown that TAK of optimal quality can be manufactured by the thermal treatment of raw kaolin with 74% of kaolinite at 750 °C without the intermediate beneficiation stage. The application of a developed approach can significantly reduce production expenditures and make the application of such an additive feasible even in regular-grade cement and concrete. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Kaolin; Metakaolin; Additive; Thermal activation; Cement; Compressive strength; Microstructure
1. Introduction There is an ongoing interest in the use of selected clay minerals, including kaolinites, in the construction industry. A recent development comprises the application of metakaolin as an artificial pozzolanic additive for concrete [1–7]. The strength and durability of conventional cement-based materials can be significantly improved by using thermally activated kaolin additives [1–8]. Such additives are conventionally manufactured by firing the high-grade purified kaolinite at 650–850 °C according to the following reaction [8,9]: 650850 C
2SiO2 Al2 O3 2H2 O ! 2SiO2 Al2 O3 þ 2H2 O " The main beneficial effect of metakaolin in concrete and cement is related to its high pozzolanic activity (i.e., the ability to react with portlandite, Ca(OH)2 released during *
Corresponding author. E-mail address: [email protected] (K. Sobolev).
0950-0618/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.conbuildmat.2008.02.008
the hydration of portland cement) [1–8]. The application of the super-fine particles of metakaolin results in a microfiller effect and improves the packing of the cement matrix. The micro-bearing’s effect is provided by the flaky particles of metakaolin, resulting in the better sliding of more coarse cement particles and, therefore, facilitating the flow of the system. Furthermore, metakaolin improves the morphology of the interface zone between the cement matrix and the aggregate’s surface [5–8]. Both very light, attractive color shades, and an improved strength and durability of metakaolin-based concrete make such an additive one of the vital components of modern architectural concrete [1,6,9]. Shvarzman et al. demonstrated that useful properties of metakaolin are preserved even at a reduced content (down to 30%) of kaolinite in the raw mix [8]. Based on this study, it was proposed that clays containing more than 35% of kaolinite can be directly processed into pozzolanic additives using thermal activation, eliminating the expensive stage of beneficiation [9]. Realization of such an approach
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
63
3. Experimental program
to a surface area of about 300 m2/kg. The chemical composition of these cements (as determined by the X-ray fluorescence, XRF), is reported in Table 1. It can be observed that these cements had similar composition; however, NPC had somehow lower content of alkalis. The raw kaolin specimens were obtained from Ku¨tahya-Kizilcßukur quarry (KK), Ku¨tahya-Ulasßlar quarry (KU), and S ß ile-DomaliMatel quarry (KS), Turkey [10–12]. The investigated samples were composed mainly of quartz and kaolinite with low to medium crystallinity as defined by the Hinckley Index [10]. The chemical composition of the kaolin based materials used in research was analyzed using the XRF technique and reported in Table 1. According to the XRF analysis, a sample of KK contained a high amount of Al2O3 (31.5%, within the usual range of 15–30%) and a relatively low amount of SiO2 (55%, within the common range of 55–80%). This specimen was predominantly composed of kaolinite and quartz as determined by the XRD (X-ray diffraction, Fig. 1), SEM (scanning electron microscopy) and EDS (energy dispersive spectrometry) [10]. The results of DTA-TG (differential thermal analysis/thermogravimetric analysis) investigation demonstrated the presence of the dehydroxylation endotherm at 580 °C with a characteristic weight loss of 9.2% that corresponds to 73.6% of kaolinite (Fig. 2). Due to the presence of quartz and tridymite minerals, a relatively high content of SiO2 (67.0%) was detected for sample KU (Fig. 1). For KU specimen, the kaolin–metakaolin conversion endotherm was observed at 564 °C with a weight loss of 6.7%, Fig. 2; this provides the estimated kaolinite content of 53.6%. The sample of KS had intermediate contents of Al2O3 and SiO2 with lowest level of kaolinite, 41.6%. The detailed characterization of investigated raw kaolins was reported by Aras et al. [10]. Based on their chemical analysis, all obtained TAK are classified as Class N according to ASTM C349 [13].
3.1. Materials used
3.2. Research program
could significantly reduce the production expenditures related to intermediate wet beneficiation and the subsequent drying of the raw kaolin. Furthermore, the waste streams generated during these stages can be also eliminated. Consequently, thermally activated kaolin (TAK, to differentiate from its metakaolin sibling, which is made from purified kaolin) can be made available at a significantly reduced cost, making feasible the application of such an additive even in a regular-grade cement and concrete. The properties of such products can be tailored to provide the improved strength and durability. To verify this proposal, an extensive research program involving five R&D and educational institutions (METU, LAU, TCMA, MTA, TECHNION) was initiated and supported by the TUBITAK (Scientific and Research Council of Turkey) [9]. The detailed results on the characterization of available raw kaolin sources in Turkey were reported by Aras et al. [10] and Sobolev et al. [11] described the thermal activation and process parameters for the production of TAK. 2. Research significance Metakaolin is a well-known additive for the improvement of concrete strength and durability. The application and performance of metakaolin is well covered in the scientific literature. However, there are limited references to the performance of cement and concrete additives based on low-grade or unprocessed kaolin. Therefore, the investigation of strength, microstructure and other performance parameters of cement containing thermally activated additives based on raw (unprocessed) kaolin was considered important. This research deals with such an evaluation.
Three raw kaolin specimens with different kaolinite content (41.6–73.6%) and two reference portland cements CEM-I 52.5, NPC and CEM-I 32.5, LPC (ASTM Type III and Type I, respectively) were used in the research program. NPC was a commercial product and LPC was laboratory made cement obtained by intergrinding of 95% clinker and 5% of natural gypsum in a laboratory ball mill
The experimental program involved three main tasks: thermal activation of kaolin at 750 °C for different treatment durations (1–2 h); the investigation of the properties of obtained TAK: mineralogical composition (XRD); fineness, particle size distribution, whiteness, morphology (using SEM);
Table 1 Chemical analysis of materials used in the research Type
SiO2 (%)
Al2O3 (%)
Fe2O3 (%)
TiO2 (%)
CaO (%)
MgO (%)
SO3 (%)
K2O (%)
Na2O (%)
MnO2 (%)
P2O5 (%)
L.O.I. (%)
Kaolinite* (%)
NPC LPC KK KU KS
19.78 19.47 55.0 67.0 57.3
5.25 5.01 31.5 21.5 26.7
3.24 3.07 0.3 0.3 2.7
– – 0.1 0.5 1.2
63.45 63.95 0.1 0.2 0.1
1.21 1.57 <1.0 <1.0 0.7
2.45 2.63 – – –
0.72 0.84 <1.0 0.4 2.5
0.32 0.47 <1.0 <1.0 0.1
– – <1.0 <1.0 <1.0
– – 0.2 0.2 <1.0
1.60 1.24 10.85 9.40 7.90
– – 73.6 53.6 41.6
*
Calculated according to the weight loss at 500–750 °C corresponding to dehydroxylation of kaolinite.
64
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
H
K
QT
KQ A
QA
Q
K
K Q A T H
: Kaolinite : Quartz : Alunite : Tridymite : Halloysite
KK
KU
KS
KK—1/2 hour
KK—1 hour
KK—2 hours
10
20
30
40
50
2Θ
60
Fig. 1. The results of XRD analysis of raw kaolin and selected (KK – based) TAK specimens activated at 750 °C.
Weight Loss
Heat Flow
410 µV*min/g 993 °C -2.93 mV*min/g 580 °C
KK
Δ = -9.2 %
KK
278 µV*min/g 994 °C
KU
-1.93 mV*min/g 564 °C
Δ = -6.7 %
KU Δ = -1.4 %
KS 141 µV*min/g 950 °C -1.37 mV*min/g 553 °C
0
200
400
Δ = -5.2 %
600
800
1000
0
Temperature, °C
200
400
600
800
KS
1000
Temperature, °C
Fig. 2. The results of DTA-TG investigation of kaolin specimens.
the investigation of properties of blended TAK cement: setting times, soundness, heat of hydration, microstructure (using SEM), compressive and flexural strength (as pozzolanic activity).
3.3. Notations used The following notations were used to distinguish the investigated samples:
NPC reference portland cement; KK raw kaolin, from Ku¨tahya-Kizilcßukur; KU raw kaolin, from Ku¨tahya-Ulasßlar; KS raw kaolin, from S ß ile-Domali-Matel.
All TAK specimens were designated with an additional number after the kaolin type notation. This number corresponds to the duration of thermal treatment; for example, KK-1 and KK-2 correspond to TAK based on KK and activated at 750 °C for 1–2 h, respectively.
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
3.4. Mixture proportioning The properties of 14 cements were investigated. These included two reference portland cements (NPC and LPC) and composite cements with 20% of raw kaolin or TAK. The 20% cement substitute level was also selected to assess the pozzolanic activity of mineral additives (as prescribed by ASTM C311) [14]. The mortars were designed using a sand-to-cement ratio (S/C) of 3.0 and W/C of 0.5 according to EN 196 and tested according to ASTM C109, ASTM C311, ASTM C349, and ASTM C618 [13–17]. 3.5. Preparation of specimens Prior to its use in the research program, the raw kaolin was pre-ground in a ball mill for 60 min and the remaining coarse particles were removed by screening on a 90 lm sieve. Raw kaolin was thermally activated in a muffle furnace at 750 °C for 1–2 h respectively, resulting in the TAK samples. Based on preliminary investigation [11], for majority of raw kaolin specimens, the temperature of 750 °C was found to be sufficient for kaolin–metakaolin conversion within 1 h. However, kaolinite-rich specimens usually require longer treatment times (up to 2 h) or higher temperatures [11]; these observations are also supported by the results of XRD and DTA-TG investigation (Figs. 1 and 2). The composite cements with 20% of mineral additive were obtained by dry intermixing in the standard lab mixer, a process that represents the industrial approach applied by
65
many cement companies in Europe. The investigated mortars were produced following EN 196 [13]. The mortars were cast into three-gang (40 40 160 mm) prism molds, and compacted in accordance with EN 196. For SEM investigations cement pastes were prepared following the same mixing and molding method using NPC and 20% of mineral additive based on KK at W/C of 0.3. 3.6. Curing of specimens After the compaction procedure, the molds were placed in a humidity cabinet for 24 h (with a relative humidity of 95% and a temperature of 20 ± 1 °C). Following this period, the specimens were removed from the molds and kept in water until the testing age. The cement pastes prepared for SEM investigations were cured for 7- and 28-day at 20 ± 1 °C. After the curing period, the fractured specimens were dried, treated with acetone and coated with a thin layer of gold. 3.7. Tests performed The particle size distribution of investigated cements was measured by a Masersizer laser diffraction analyzer (Malvern Instruments). The resulting particle size distribution and other characteristics of investigated materials (i.e., cement, kaolin and TAK) are summarized in Fig. 3 and Table. 2. The properties of developed cements (setting times, soundness, heat of hydration, compressive and flexural strength) were obtained using EN 196 [15]. These
Fig. 3. Particle size distribution of investigated materials.
Table 2 Physical properties of selected cements and raw kaolin/TAK Performance characteristics Blaine specific surface area (m2/kg) Median size (lm) Heat of hydration (cal/g) Whiteness (%)
Reference cements
Blended cements based on NPC
NPC
LPC
KK
KK-1
KK-2
Raw kaolin/TAK KK
KK-1
KK-2
356 22 70.5 –
318 27 – –
362 25 78.9 –
392 20 94.7 –
397 20 100.4 –
422 26 – 77.6
474 15 – 81.5
489 14 – 82.4
66
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
results are presented in Table. 3. The mortar samples were tested at the age of 7, 28 and 90 days for flexure and compression. Compressive strength tests were conducted using the portions of three prisms broken in flexure. Therefore, the reported compressive strength results are the average of the six values [15]. SEM observations on the cement paste samples were performed at 7- and 28-day using a LEO scanning electron microscope operated at an accelerating voltage of 15 kV.
much higher than that of raw kaolin (Table 2). The obtained TAK is characterized by a three-modal distribution (Fig. 3). The size reduction associated with such thermal treatment led to an increase in the whiteness of TAK from 77.6% for raw kaolin to 81.5% and 82.4% for samples KK-1 and KK-2, respectively (Table 2). The cements based on TAK had smaller median particle size and higher surface area (Table 2). 4.3. Morphology of TAK
4. Results and discussion 4.1. The effect of thermal activation It was demonstrated that thermal activation at 750 °C is adequate to complete the kaolin–metakaolin transformation (Figs. 1 and 2). Thermal activation for 2 h results in complete disappearance of kaolinite peaks on XRD, which is clear sign of its completed conversion to metakaolin. However, even 1 h treatment is sufficient to complete the transformation with very little kaolinite left in the system. Treatment for ½ h was less effective, leaving significant quantities (about 50%) of unconverted kaolinite; therefore, such activation was not considered for production of TAK.
The comparison of surface features and morphology of investigated raw kaolin and TAK (using KK samples) were performed using the SEM technique (Fig. 4). It was observed that raw kaolin particles are of the predominant size of 5–50 lm, with agglomerates clearly seen around the quartz particles (Fig. 4a). Due to thermal treatment at 750 °C and subsequent recrystallization to metakaolin, the particles of kaolin are separated from quartz and significantly reduced in size. The resulting TAK is characterized by the particles of the predominant size of 1–15 lm (Fig. 4b). SEM observations of KK-2 morphology demonstrated the presence of star-like agglomerates of low-density metakaolin (Fig. 4c). 4.4. Effect of TAK on properties of cement
4.2. Particle size distribution of TAK The particle size distribution of investigated materials (observed by means of laser diffraction analysis) is presented in Fig. 3. It can be observed that the particles of raw kaolin are somehow larger compared with portland cement. The particle size distribution of raw kaolin is characterized by a bi-modal distribution. It was found that due to thermal treatment at 750 °C and its recrystallization to metakaolin, the kaolin particles are significantly reduced in size (with the predominant size of 1–15 lm). Due to this transformation, the Blaine specific surface area of TAK is
The replacement of the NPC with TAK (such as KK-1 and KK-2) at a dosage of 20% does not affect the final setting time of blended cements (Table 3) and the initial setting time of such cements was only slightly prolonged. However, the addition of raw kaolin and TAK to LPC cement results in delayed setting of blended cements. The soundness (tested according to le Chaˆtelier method of EN 196 [15]) of all investigated cements was at the same level of 1–2 mm for all the cements, except the reference cement LPC and blended cement based on NPC with 20% of KK-1 additive, these had zero soundness.
Table 3 Performance of the investigated cements Type of cement
Activation (h)
Normal consistency (%)
Setting time (min) Initial
Final
Soundness* (mm)
Compressive/flexural strength (MPa) 7 days
28 days
90 days
NPC NPC–KK NPC–KK-1 NPC–KK-2
– – 1 2
25.4 28.4 29.3 28.0
155 195 170 165
200 245 205 195
1 1 0 1
39.6/5.4 37.4/4.9 45.2/5.7 41.1/5.6
53.4/6.6 51.5/6.2 61.7/6.6 59.7/6.7
58.4/6.8 55.4/6.4 65.8/6.9 62.9/6.8
LPC LPC–KK LPC–KK-1 LPC–KK-2 LPC–KU LPC–KU-1 LPC–KU-2 LPC–KS LPC–KS-1 LPC–KS-2
– – 1 2 – 1 2 – 1 2
25.6 26.8 27.1 27.0 28.0 27.4 27.3 26.9 27.1 26.9
145 175 185 215 195 200 190 200 215 195
190 215 225 250 240 245 230 235 250 230
0 1 1 2 2 1 1 2 2 1
31.2/4.5 23.0/3.9 28.0/4.0 28.0/4.1 23.7/4.0 25.6/3.9 27.2/4.1 22.3/3.8 30.3/4.4 29.6/4.7
39.2/5.6 26.9/4.2 39.0/5.2 41.4/5.3 30.0/4.3 38.0/5.2 39.4/5.5 26.0/3.8 38.7/5.4 38.2/5.2
42.7/6.0 29.9/4.4 41.8/5.8 44.1/6.3 37.6/5.6 44.1/6.3 44.9/6.3 27.4/4.9 40.8/5.6 40.4/5.7
*
Tested according to le Chaˆtelier method of EN 196 [15].
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
67
observed that with hydration of the reference portland cement (NPC), the unhydrated clinker grains are surrounded by radiating fibers of C–S–H resembling the pattern of Type I C–S–H. Randomly oriented CH crystals and the prismatic ettringite crystals are widely dispersed through the paste (Fig. 5a, left). The microstructure of the hydrated paste of NPC at the age of 28 days is presented by the amorphous gel filling the spaces between the hydrated particles. In NPC pastes, the layered accumulations of CH crystals of about 15 lm in width are intermingled through the paste (Fig. 5a, right). It is observed by SEM that the KK and TAK particles are well dispersed within the cement paste, resulting in a dense structure with low porosity (Fig. 5b–d). There is a visible densification around the TAK particles, due to the pozzolanic reaction of TAK, leading to the formation of C–S–H. It can be seen that at the age of 7 days the kaolin and TAK grains are already well covered with a pseudomorphic layer composed of hydration products. The matrix phase is composed mainly of short acicular outgrowths of C–S–H around the clinker grains and needleshaped ettringite crystals (Fig. 5c–d, left). At the age of 28 days the kaolin and TAK grains are well embedded into the matrix and are well connected through the C–S–H gel. It can be distinguished that cement pastes based on TAK, especially cement with KK-1, are of higher density comparing to reference cement and blended KK cement (Fig. 5). 4.6. Fresh properties of mortars In spite of somehow higher normal consistency of TAKcements (Table 3), all investigated mortars were of perfect workability allowing for easy casting of the specimens as per as EN 196. 4.7. Compressive and flexural strength of mortars based on NPC Fig. 4. SEM observations of kaolin morphology. (a) raw kaolin KK, (b) KK-1, (c) KK-2.
Based on the heat of hydration values (as per as EN 196, Table 2) it can be proposed that the addition of TAK results in significant acceleration of cement hydration process. Even 20% replacement of NPC with TAK (KKbased) results in 34% and 42% increase in heat of hydration in case of TAK activated for 1–2 h, respectively. This behavior can be explained by intensive pozzolanic reaction that consumes significant quantities of portlandite (i.e., consumption of CH would shift the hydration reaction towards intensification of hydration) and also by the formation of cement crystallization centers/sites on smaller particles of TAK. 4.5. Effect of TAK on microstructure The results of SEM investigation of the cement pastes hardened for 7- and 28-day are presented in Fig. 5a. It is
The compressive test results of investigated mortars (following EN 196) are presented in Table 3. The 7-day compressive strength of cement with KK-1 is the highest among the investigated cements, 45.2 MPa. The cement with KK-2 had a 7-day compressive strength of 41.1 MPa, which is close to the strength of reference NPC (39.6 MPa). The cement with raw kaolin KK had the lowest 7-day compressive strength of 37.4 MPa (Table 3). Acceleration of the strength development of TAKcement with optimized TAK, KK-1 at 7-day age by 14% can be explained by the pozzolanic reaction and the densification of cement matrix (as it was also observed by the SEM). According to the test results, the best 28-day compressive strength of 61.7 MPa was obtained in cement with TAK activated for 1 h (KK-1); this value is 15% higher than the strength of reference cement. The cement with TAK activated for 2 h (KK-2) reached a 28-day compressive strength of 59.7 MPa, which is only slightly less than that for KK-1.
68
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
Fig. 5. Effect of TAK on microstructure of cement paste at the age of 7 days (left) and 28 days (right).
Addition of TAK based on KK did not affect the flexural strength of investigated mortars in all ages of hardening (Table 3). Therefore, it can be stated that addition of TAK has higher effect on the improvement of compressive
strength. The cement with raw kaolin KK had a 28-day compressive strength of 51.5 MPa, which is somewhat close to the strength of reference NPC (53.4 MPa). The observed trend was also preserved at the age of 90 days with KK-1
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
cement gaining the highest compressive strength of 65.8 MPa (or 13% higher than the reference). 4.8. Compressive and flexural strength of mortars based on LPC Reference cement LPC had a lower strength in all ages of hardening vs. NPC (Table 3). This can be explained by lower Blaine specific surface area of LPC, as well as higher content of alkalis. Addition of TAK led to the reduction of the 7-day compressive strength of mortars based on LPC. For blended cements with TAK based on raw kaolin composed mainly of quartz and kaolinite mix (i.e., KK and KS), the reduction of 7-day strength was minor, about 10% or less. However, the specimen processed from raw kaolin containing tridymite had a significantly reduced strength at the age of 7 days. This trend was overcome at the 28-day age when all investigated cements with TAK had a compressive strength similar to that of reference cement LPC. Such performance level satisfies and exceeds the requirements of ASTM C618 set for Class N pozzolanic additives. After 90 days of hardening the strength of blended cements with TAK based on raw kaolin with more than 50% of kaolinite was higher than the strength of reference cement; an exception to this rule was the behavior of cement with KK-1, which had slightly lower 90-day strength that that of the reference. Such a strength reduction can be explained by the incomplete kaolin–metakaolin conversion (Figs. 1 and 2). Still, cements with 20% of low-kaolinite based TAK (KS-1/ KS-2 based on clay with 41.6% of kaolinite) demonstrated only 5% strength reduction vs. the reference LPC; such additives can also be considered as effective pozzolans. The flexural strength of mortars based on LPC followed fairly similar trends, usually keeping the compressive-toflexural strength ratio within the limits of 6.8–7.3. Strength of LPC – raw kaolin blended cements was low in all ages of hardening (about 20–30% less than the reference); based on the criteria set forth by the ASTM C618, such reduction of strength does not allow classifying the raw kaolins as pozzolanic additives. 5. Conclusions It was shown that effective mineral additives can be manufactured by the thermal activation of raw kaolin with at least 41.6% of kaolinite, corresponding to at least 21.5% of Al2O3 and less than 55% of SiO2. For the investigated kaolin types, the thermal activation at 750 °C for 1 h is sufficient to obtain the product of desired performance. The resulting mineral additive, called thermally activated kaolin (TAK), meets or exceeds the requirements of ASTM C618 set for Class N pozzolanic additives. It can be proposed that addition of TAK results in significant acceleration of cement hydration process due to intensive pozzolanic reaction and also due to the formation of cement crystallization centers on smaller particles of TAK.
69
Based on the conducted investigation, it can be concluded that the application of TAK manufactured from kaolin with a high kaolinite content (73.6%, such as KK) is very effective in blended cements based on high-strength cement (i.e., EN CEM-I 52.5 or ASTM Type III). The 28day compressive strength of such cements is improved by up to 15%. The compressive strength of blended cement with 20% of TAK (KK-1/KK-2) reaches the level of 65 MPa at 90-day age. Based on the microstructural investigation, the improvement of the strength of TAK-cement is attributed to the pozzolanic reaction and the densification of cement matrix due to the addition of TAK. It is proposed that the TAK-based additive provided quite a significant improvement of cement properties, which is similar to the effect of ‘‘purified,” commercial grade metakaolin. This effect can be further improved when the TAK additive is applied with an effective superplasticizer. The application of TAK is less effective when it was blended with laboratory-grade portland cement (i.e., LPC, classified as EN CEM-I 32.5 or ASTM Type I). In spite of the reduction of 7-day compressive strength of mortars based on LPC, the 28- and 90-day compressive strength of such cements was within the ±5% range of the reference cement strength. It was demonstrated that the addition of TAK does not affect the flexural strength of investigated mortars in all ages of hardening. It is demonstrated that the developed approach can significantly reduce the expenditures related to production of TAK due to the elimination of the steps of intermediate wet beneficiation and the subsequent drying of the raw materials that are common for commercial grade metakaolin additives. Furthermore, the proposed process can reduce the waste streams generated during these eliminated stages. Therefore, it is expected that TAK could be available at a significantly reduced cost, making its application feasible even in a regular-grade cement and concrete. Additional investigation may be necessary to explain and quantify the hydration mechanism and the long-term microstructural development of TAK cements containing different amounts of TAK. Further research is also required to investigate the behavior of TAK cements in concrete, as well as to investigate the durability of such materials. Acknowledgements This study was performed under TUBITAK research _ ¸ TAG – 680, in cooperation with MTA research Grant IC team represented by Dr. Mustafa Albayrak and Dr. Aydin Aras. The authors express their gratitude to Ipek S ß ener and Bahadir Erdogan for their help with preparation of the samples. Extremely useful comments of Prof. Konstantin Kovler (TECHNION) are highly appreciated. The financial contribution of TUBITAK, TCMB, LAU, MTA, CONACYT, PROMEP and PAICYT is gratefully acknowledged.
70
M. Arikan et al. / Construction and Building Materials 23 (2009) 62–70
References [1] Kostuch JA, Walters GV, Jones TR. High performance concrete incorporating metakaolin—a review. Concrete 2000 1993;2:1799–811. [2] Wild S, Khatib J, Jones A. Relative strength, pozzolanic activity and cement hydration in superplasticized metakaolin concrete. Cem Concr Res 1996;26(10):1537–44. [3] Khatib J, Wild S. Sulfate resistance of metakaolin mortar. Cem Concr Res 1998;28(1):38–92. [4] Jones TR, Walters GV, Kostuch JA. Role of metakaolin in suppressing ASR in concrete containing reactive aggregate and exposed to saturated NaCl solution. In: Proceedings of the nineth international conference on alkali-aggregate reaction in concrete, vol. 1; 1992. p. 485–96. [5] Sabir BB, Wild S, Khatib J. On the workability and strength development of metakaolin concrete. In: Dhir RK, Dyer DT, editors. International congress on concrete in the service of mankind, concrete for environment enhancement and protection, theme 6, waste materials and alternative products, University of Dundee Spon, London; 1996. p. 651–62. [6] Caldarone MA, Gruber KA, Burg RG. High-reactivity metakaolin: A new generation mineral admixture. Concr Int 1994;16(11):35–40. [7] Sobolev K. White cement: problems of production and quality. Cem Concr World 2001(July–August):34–42. [8] Shvarzman A, Kovler K, Grader GS, Shter GE. The effect of dehydroxylation/amorphization degree on pozzolanic activity of kaolinite. Cem Concr Res 2003;33(3):405–16. [9] Arikan M, Sobolev K, Ertu¨n T, Yeg˘inobali A, Albayrak M, Aras A, et al. Development of cement and concrete additives based on _ _ ¸ TAG ¨ BITAK thermally activated kaolin. TU project proposal IC 2003;680:55 [in Turkish].
[10] Aras A, Albayrak M, Arikan M, Sobolev K. Evaluation of selected kaolins as a raw material for the Turkish cement and concrete industry. Clay Miner 2007;42(2):233–44. [11] Sobolev K, Arikan M, Ertu¨n T, Yeg˘inobali A, Erdogan B. The development of thermally activated kaolin admixture for highperformance blended cements. J. Am Ceram Soc submitted for publication. _ ¨ zel Ihtisas [12] Besß Yillik Kalkinma Plani: Madencilik O Komisyonu Raporu Endu¨striyel Hammaddeler Alt Komisyonu Toprak Sanayii Hammaddeleri – I (Seramik Killeri-Kaolen-Feldspat-Pirofillit-Wollastonit-Talk) C ¸ alisßma Grubu Raporu. Ankara: DPT. 2001. 204 pp _ 622). ISBN: 975-19-2837-0 [in Turkish].