- Email: [email protected]
Carbonation of low heat portland cement paste precured in water for different time
J o u d of University of Science and Technology Beijing Volume 14, Number 2, April 2007, Page 178
Carbonation of low heat portland cement paste precured in water for different time Deping Chen”, Etsuo Sakai2’,Masaki Daimon2’,and Yoko Ohba’’ 1) Civil and Environmental Engineering School, University of Science and Technology Beijing, Beijing 100083, China
2) Graduate School of Science and Engineering. Tokyo Institute of Technology, Tokyo 152-8552, Japan (Received 2006-06-01)
Abstract: The carbonation technique was applied to accelerate the hydration of low heat portland cement (LHC). Before carbonation, the demoulded pastes were precured in water for 0,2,7, and 2 1 d, respectively. The results show that precuring time in water strongly influences the carbonation process. The phenolphthalein test indicates that the paste precured in water for a shorter time is more quickly carbonated than that for a longer time. The content of calcium hydroxide increases with increasing the precuring time in water, whereas, the amount of absorbed carbon dioxide changes contrarily. Scanning electron microscope (SEM) observation shows that portlandite always fills up big air bubbles in the paste during precuring in water, and the mercury intrusion porosimetry (MTP) results show that there are less large capillary pores in the paste precured in water for a longer time. It is found that the paste without precuring in water has more carbon dioxide absorption during curing in carbon dioxide atmosphere, and its total pore volume decreases remarkably with an increase in the carbonation time than that precured in water. X-ray diffraction (XRD) and Brunauer-Emmett-Teller (BET) surface area analyses indicate that the carbonate products are vaterite and calcite; C,SH, formed from carbonation has low BET surface area in comparison with that of C-S-H formed from curing in water. Key words: carbonation: carbon dioxide absorption; vaterite; BET surface area; pore size distribution
[Thisstudy was financially supported bv the Ministv of Education, Culture, Sports, Science, and Technology, Japan.]
1. Introduction Low heat portland cement (LHC) is attracting a great deal of interest worldwide for its lower energy consumptions and CO, emissions on manufacture than ordinary portland cement. But its application is comparatively restricted because of the slow hydration and low early strength. These disadvantages can be improved by using a number of techniques, such as, activating of belite or introducing reactive additives [ 1-21. It was reported [3-41 that LHC could accelerate hydration under a C02 curing condition accompanied by highly developed strength. In this process, a significant proportion of C 0 2 was fixed by the carbonation reaction, which could provide an effective way of sequestering C02 as important solid phases to produce construction materials [5]. However, the precuring condition that is most suitable for the carbonation reaction of LHC, and the details of the carbonation process with different precuring time in water are not clear. K. Koibuchi et 111. (41 said short precuring time in water is in favor of absorbing carbon dioxide; E. Sakai er al. [3] investigated the carbonation reaction of hardened low Corresponding author: Deptng Chen. E-mail: [email protected]
heat cement under 2 d precuring in water. In this article, the effect of precuring time in water on the carbonation process of LHC was studied.
2. Experimental 2.1. Materials The LHC was from Taiheiyo Cement Co., Japan. Its composition (wt%) was: SO,, 26.12; Fe203, 2.29; A1,0,, 2.87; CaO, 63.94; MgO, 0.75; SO3, 2.26; Ti02, 0.12; LOI, 0.87. Its mineral composition (wt%) by Bogue is: f-CaO, 0.10; C3S, 32.30; C2S, 50.60; C3A, 3.74; C4AF, 6.96; CaSO,, 3.84, with a density of 3.22 g/cm3and a Blaine surface area of 3280 cm2/g.Moreover, an antifoamer and two compositions of the viscosity improver (viscoA and viscoB) were used to reduce air bubble formation when mixing, to prevent the slurry from segregation.
2.2. Preparation The mix proportions were: water-to-cement mass ratio 0.4, antifoamer to water 5 ~ 1 0 - and ~ , viscoA and viscoB to water 0.04 respectively. Initially, viscoA, antifoamer, and distilled water were mixed together, and Also available online at www.sciencedirect.com
D.P.Chen et al., Carbonation of low heat Portland cement paste precured in water for different time
then poured into the LHC, and stirred for 2.5 min. Next, viscoB was added and again stirred for 2.5 min. The slurry turned to viscous paste. The fresh paste was cast into a mould of 20 mmx20 mmx80 mm. The paste with mould, after vibrating for 1 min, was put into a closed plastic box for 24 h at 20°C. The demoulded samples were first immersed in distilled water for 0,2,7, and 21 d (it was called 'precuring in water' below), and then placed in the carbonation apparatus (SH-221 from ESPEC Corp.) for 0,3,7,21, 42, and 70 d (curing in CO,). Throughout the curing temperature is 20°C; the carbonation condition is relative humid, 60%, with 5~01%COz.
2.3. Testing methods (1) Measuring the depth of carbonation. The carbonation depth was measured in a traditional way, by spraying phenolphthalein alcohol solution onto the surface of a split prism of the carbonated LHC paste. The neutralization area of transect was then calculated. The ratio of the neutralization area to the whole area of transect was defined as the carbonation area ratio. (2) XRD analyses. The paste samples were crushed and manually grounded with an alumina mortar and pestle. Their phase compositions were examined by X-ray diffraction (XRD) with a Toshiba powder diffractometer using Cu K, radiation. (3) Thermogravimetric/differential thermal (TG/ DTA) analysis. The powder samples were submitted to thermal analysis using the TG-DTA 2000s equipment (Mac Science Co.), which permits performing TG and DTA analysis simultaneously. In testing, approximately 33 mg of powder samples were heated from room temperature to 1000°C at a heating rate of 10"C/min in flowing N, (150 mL/min). After testing, TG and DTA curves were obtained. With the help of TGDTA curves, the Ca(OH), content in the paste was determined from the following equation:
where CH(wt%) is the content of Ca(OH),, WL,H(wt%) is the weight loss of Ca(OH), that occurred during the dehydration of portlandite, MW,, and MWH are the molecular weights of Ca(OH), and water, respectively. (4) CO, absorption tests. The inorganic carbon content C(wt%) of the powder samples was tested at 200°C with 5 % H3PO4on alumi-
z 79
na carrier boats using TOC-505OA Total Organic Carbon Analyzer equipped with SSM-5000A Solid Sample Module (Shimadzu Corporation, Japan). The amount of CO, absorption by LHC was calculated by using the following equation: CO,(wt%) =
C(wt%)x MWco, ( 1- LOI%)x M w c
where CO,(wt%) is the amount of absorbed CO, by LHC, C(wt%) is the content of inorganic carbon, LOI% is the mass loss by heating at 1000°C for 30 min, MW,, and MWc are the molecular weights of CO, and carbon, respectively. (5) Brunauer-Emmett-Teller (BET) specific surface area analyses.
The specific surface area of the powder samples was analyzed by using the HORIBA SA-6200 serials BET surface area analyzer (Horiba Ltd., Japan). All the samples were placed in an Aspirator for 7 d to dry, and stored in sealed plastic vials before testing. (6) Pore size distributionand total pore volume tests. A piece of 5-10 mm sized sample was submitted to pore size distribution and total pore volume testing with mercury intrusion porosimetry (MIP) using the Pascal 140 and Pascal 240 porosimeter (CE Instruments, Italy). The samples were under 1 d of D-drying before testing.
(7) SEM observations. SEM observations for the characterization of carbonated LHC pastes were carried out using JSM-5310 scanning electron microscope (JEOL Ltd., Japan). To obtain a conductive surface for analysis, a thin gold layer was deposited over it.
3. Results and discussion 3.1. Carbonation area ratio The carbonation area ratio of LHC pastes with carbonation time was shown in Fig. 1. The neutralization rate of LHC paste without precuring in water is higher than that with precuring in water. The paste without precuring in water was soon fully carbonated at the carbonation time of 21 d, whereas, the others needed more time for it. As the phenolphthalein test is based on the color change at which the pH is about 9, it denotes the presence of Ca(OH), [6]. The content of Ca(OH), in the paste, obtained by TG/DTA analysis, with the carbonation time is shown in Fig. 2. The Ca(OH), contents of the paste without precuring and with 21 d precuring in water before carbonation are 5.92wt% and 12.22wt%, respectively. After 21 d carbonation, the
. I Univ. . Sci. TechnoL Beijing, Vo1.14, N0.2, Apr 2007
180
Ca(OH), content without precuring in water falls to Owt%, whereas, that with 21 d precuring in water still remains at 10.74wt%. This indicates that less precuring time in water enables the paste to neutralize easily.
.-
/ :"'
80-
+a
i?
8 60
-
3
./,\
.+
40-
j
20-
a
,
/
,
0I
/
Precuring time in water -.-Od 2d A 7d - ~ - 2 1 d I
v.
,
0
10
20 30 40 50 60 Carbonation time / d
70
Fig. 1. Carbonation area ratio of LHC pastes.
.
T--
v -
1
-.
.
Precuring time in water
.\.\
A
\
,
0
..-
\.
10
20
30 Carbonation time / d
40
Fig. 2. Contents of Ca(OH), remains in the pastes.
3.2. MIP result and SEM observation Fig. 3 shows the curves of cumulative pore volume with pore diameter in the carbonated pastes. The total pore volume of the paste without precuring in water decreases considerably with carbonation time pro-
w
longed from 3 to 42 d, and the amount of large capillary pores (>0.1 pm) also apparently diminishes; but comparatively those with 21 d precuring in water only slightly change with carbonation time (Fig. 3(a)). The total porosity of the paste without precuring in water decreases from 33.4% at 3 d carbonation to 17.1% at 42 d carbonation time, but that with 21 d precuring in water increases from 22.0% at 0 d carbonation to 24.6% at 42 d carbonation time. The carbonated paste without precuring in water has more large capillary pores (>0.1 pm) than those with precuring in water (Fig. 3(b)). It means that precuring in water is favorable for the decrease of large capillary pores. The longer the precuring time in water it is, the fewer large capillary pores it has. The fracture surface of the carbonated LHC pastes under SEM, as in Fig. 4, indicates the pastes without precuring in water seem looser in microstructure than those with long precuring, for example, 21 d in water. Their big air bubbles always remain empty and some of them are filled with a few Ca(OH)2crystals, and large capillary pores exist everywhere in the paste (Fig. 4(a)), even at 42 d carbonation (Fig. 4(b)). On the contrary, the big air bubbles in the paste with 21 d precuring in water are always filled with many tabular Ca(OH), (Fig. 4(c)), and the existence of large capillary pores are not so popular in the paste as mentioned above. But carbonation microcracks are easily found on the surface and in the inner part of the pastes with 21 d precuring in water (Fig.4(d)), even at 3 d's carbonation time (Fig. 4(c)). The existence of microcracks would result in the increase of total porosity in the paste at the carbonation time of 0 to 42 d, measured with MIP.
-Fg . -g 100 w
\
Precuring time Curing timc
140
E
80
2 n
60
.-9
40
m
3 c) 5
Ib)
\,
120
9
o
Pore diameter / pm
160
20 0
3
0.01
Precuring time Curing time in CO,
\
:"I2, 3
2d 7d
0.1 1 Pore diameter / pm
42 d 42 d 42 d
10
Fig. 3. Cumulative pore volume with the diameter of pores for carbonated LHC paste: (a) the effect of carbonation time on pore volume; (b) the effect of precuring time in water on pore volume.
3.3. Amount of absorbed C 0 2 The amount of absorbed CO, for the paste without precuring in water increases quickly at the early car-
bonation age before 7 d, and then slowly changes with the carbonation time as shown in Fig. 5. It has the highest amount of absorbed C02 when compared to those
D.P. Chen et d,Carbonationof low heat portland cement paste precured in water for different time
with precuring in water, although, in the same carbonation time. The amount of absorbed COz for the paste without precuring in water is nearly 25wt% at 7 d carbonation, approximately six times that of 21 d pre-
181
curing in water. The less the precuring time in water is, the more amount of CO, the paste absorbs. It also reveals that precuring in water is inferior for the pastes to absorb CO,.
Fig. 4. SEM photographs of the fracture surface for the carbonated LHC pastes: (a) precuring time in water: 0 d, carbonation time: 3 d; (b) precuring time in water: 0 d, carbonation time: 42 d; (c) precuring time in water: 21 d, carbonation time: 3 d; (d) precuring time in water: 21 d, carbonationtime: 42 d.
s
30.0
.
25.0 I
g 20.0
,,’ /
-0
$
s:
15.0
-.’
3 10.0 cr 0
z
/
_,
5.0
3
d
--.-Od A
7d
-tT
2d 21d
0.0
0
10
20 30 40 50 Carbonation time / d
60
70
Fig. 5. Amount of absorbed CO, of the carbonated LHC pastes.
3.4. XRD and TG/DTA results XRD analysis shows the predominant phases of the carbonated LHC paste are vaterite, calcite, portlandite, and belite (p-C,S) (Fig. 6). Vaterite and calcite are two forms of calcium carbonate (CaC03) caused by carbonation. Portlandite Ca(OH), is a hydrate phase of LHC, and belite is an anhydrous cement phase, which remains during the curing process. The peaks of vaterite in the paste without precuring in water are identified easily at 3 d carbonation time, and their intensities get stronger with increasing the carbonation time (Fig. 6(a)). However, with 21 d pre-
curing in water, the peaks of vaterite can still not be found in the paste at 21 d carbonation (Fig. 6(b)). By contrast, the peaks of Ca(OH), could be identified in all the XRD patterns of the pastes, even at 42 d carbonation without precuring in water. It is different from the results of phenolphthalein test and TG/DTA analysis, which show that Ca(OH), does not exist any longer at 42 d carbonation without precuring in water. This paradox could be explained as follows. As shown in Fig. 4(a), a few crystals of Ca(OH), formed inside the air bubbles when the paste was still in mould for one day. The crystals grew less homogeneously everywhere. Some of the crystals grew tabularly, with a perfect structure, whereas, others lamellarly or scatteringly with defects in structure. The Ca(OH), crystals with a perfect tabular structure were difficult to carbonate, and remained in paste even after 42 d carbonation. This could be detected by XRD. However, the remaining Ca(OH), was very little in quantity, existed mainly in the air bubbles and was covered with a thick carbonate layer after a longer carbonation time. It is possible that it could not be identified by a phenolphthalein test in this condition. On the other hand, when the paste was crushed and manually ground to powder, the small Ca(OH), granules could not distribute homogeneously. Moreover, the amount of powder
J. Univ. Sci. TechnoL Beging, voL14,N0.2, Apr 2007
182
sample submitted for thermal analysis was as low as approximately 33 mg. Thus there is a possibility that
V
Ca(OH)* could not be detected by TG/DTA analysis. (b)
Portlandite .Calcite Vaterite A p-C,S
Precuring in water 0 d carbonation time
Portlandite 0 Calcite
8
1
30
Precuring in water: 21 d carbonation time:
I
I
60
50
40
p-C,S
8
I
20
A
V Vaterite
30
20
201(")
40 2e1~)
I
I
50
60
Fig. 6. XRD patterns for carbonated LHC pastes with 0 and 21 d precuring in water: (a) precuring in water for 0 d; (b) precuring in water for 21 d.
The typical TGlDTA curves of the LHC paste (Fig. 7) show that there are three major endothermic reactions along with mass loss, which occur during heating: (1) the release of the evaporable and part of the adsorbed water at 100°C approximately; (2) the dehydration of Ca(OH), at 400-500°C; and ( 3 ) the dissociation of carbonate phases at 620-780°C. Without precuring in water, the total mass losses of pastes as shown in Fig. 7(a), increase quickly from 9.7% at 0 d carbonation to 22.4% and 28.7% at 3 and
21 d carbonation time, respectively. When the carbonation time exceeds 2 1 d, the total mass losses hardly change. The endothermic peak at 400-500°C presenting the dehydration of Ca(OH), disappears at longer carbonation time, such as, 21 and 42 d. By comparison with the precuring time in water of 21 d, the total mass losses of pastes as shown in Fig. 7(b), almost do not change at 0,3 and 21 d carbonation time and are seen as 18.496, 19.496, and 19.796, respectively.
I
Carbonation time
100
100
95
5
s
90
m
m
2
4
85
m
80 75
'Precuring time in water 0 d
42
d21
-0.5
'
5
95 90 85 80
75
recuring time in water 21 d
70'
200
400
600
SbO
lo00 -0.5
Temperature I "C
Temperature / SS
Fig. 7. TG/DTA curves for the carbonated LHC paste: (a) precuring in water for 0 d; (b) precuring in water for 21 d. Precuring time in water
3.5. BET surface area Fig. 8 shows the BET specific surface area of LHC pastes with carbonation time at various precuring times in water. The pastes without precuring in water have a low BET surface area of 9.98-14.46 m2/g despite the carbonation time being changed from 0 to 70 d. But the pastes with 21 d precuring in water have a comparatively large surface area of 30.66-33.08 m2/g at the early carbonation age before 7 d, which decreases sharply to 15.9 m2/g at 21 d carbonation.
90 E v
.
-A-
25
7d
-~--21 d
m
s
t; m
: 0
10
20 30 40 50 Carbonation time / d
60
70
Fig. 8. BET surface area of the carbonated LHC paste.
D.P.Chen et al., Carbonationof low heat portland cement paste precured in water for different time
3.6. Discussion LHC pastes precured in water for different time, before curing in C02, are different in many aspects in terms of phase composition,pore size, pore distribution, etc. The paste without precuring in water has the lowest Ca(OH), content (5.92wt%), highest porosity, plenty of big air voids and a large number of capillary pores, compared to those precured in water for 2,7, and 21 d. They seem to form two kinds of Ca(OH), in the paste. One of them scatters in the paste in small or lamellar crystals with defects in structure, which can be easily carbonated, and the other is inside the air bubbles, always in tabular crystals with perfect structure. When the paste was cured in C02, the scattered Ca(OH),, along with other initial hydrates such as AFt or AFm, were soon carbonated. The carbonation reactions are simply expressed as [3,5] Ca(OH)2+C0,-CaC0,+H,0
(3)
3CaO~Al2O3~3CaSO,~32H,0+3CO2+ 3CaC03+ 3(CaSO4~2H20)+Al2O3~nH20(gel)+(26-n)H2O (4) 3CaO~A1203CaS0,~ 12H20+3C02+ 3CaCO,+ CaS04~2H,0+A1203~nH20(gel)+( 10-n) H20
(5)
At the same time, tabular crystals of Ca(OH), are also carbonated via formula (3). With the progress of carbonation, C3S and C2S participate in reaction with C 0 2 and H,O. The overall stoichiometry of the carbonation reaction of C3S and C2S,accompanied by hydration, follows as [7-81 C3S+yH20+(3-x)C02 C,S+yH,O+( 2-x)C02
C,SH,+( 3-x)CaC03
(6)
C,SH,+( 2-x)CaC03
(7)
+
+
Then, further carbonation, changing the composition of the C-S-H gel, occurs.
C,SH,+zCO,-C,_,,SH,+zCaCO,
(8)
Thus, a large amount of C02 was absorbed, and calcium carbonates formed increasingly with carbonation time. The above carbonation reactions were determined mainly by the diffusivity and reactivity of CO,. Paste without precuring in water has high porosity and a large number of connecting air voids and capillary pores, which are in favor of the diffusion and permeation of CO, through the pores. The carbonation products fill up most of the capillary pores, which decreases the total pore volume considerably. On the contrary, the big air bubbles and large capillary pores in the paste, with the precuring time as long as 21 d, are always filled with hydration products during precuring in water. It hinders C02 from diffusing
I83
into the paste, and results in a comparatively low carbonation effect. Microcracks are formed easily on the surface of the paste on account of the huge volumetric dilation during the carbonation of Ca(OH),. Then they stretch and spread into the paste, which becomes the passage for the permeation of C02. With the progress of carbonation,the microcracks form abundantly inside the paste, along with further carbonation, and this is the reason that the amount of absorbed CO, increases noticeably after 21 d carbonation. The hydration products of the LHC paste during precuring in water are mainly Ca(OH), and C-S-H. The BET surface area is determined predominantly by C-S-H, because of the low value of Ca(OH), when compared with C-S-H [9]. The BET surface area of the paste with 21 d precuring in water before carbonation is 30.66 m2/g,and after 21 d carbonation, it falls to 15.09 m2/g. The major carbonation products are vaterite, calcite, and C,SH,. E. S a k i et al. [3] have given the BET surface area of vaterite synthesized from mixing 1 M Na2C03with 3 M CaCl, solution at 20°C for 10 min as 12.30 m2/g;and that of two kinds of calcite as l .77 m2/gfor calcite (A) prepared from vaterite by being heated at 450°C and 0.83 m2/gfor calcite (B) of special grade of calcium carbonate. Although these carbonate phases show a low surface area, their fraction is also low at 5.55% to the mass of cement obtained by TG/DTA analysis. Therefore, it can be concluded that C,SH, formed from carbonation has a much lower BET surface area than that formed from curing in water. J. Thomas et al. [lo], also reported a progressive decrease in the nitrogen BET surface area of dried hydrated portland cement paste during storage in plastic vials, which was found to be caused by carbonation, a reaction between the C-S-H phase and carbon dioxide in the air. Thus, the carbonated paste without precuring in water, exhibiting low BET surface area can also be explained.
4. Conclusions (1) Before curing in CO,, the LHC pastes become much denser in microstructure with increasing precuring time in water. Their big air voids and large capillary pores are always filled with hydration products during curing in water. (2) A longer precuring time in water means less large capillary pores, which hinders carbon dioxide from diffusing into the paste, and results in low amount of absorbed C 0 2 and high content of Ca(OH), remained in the carbonated pastes. (3) The total pore volume of paste without precuring in water decreases remarkably with an increase in car-
J. Univ. Sci. TechnoL Beging, voLz4,N0.2, Apr 2007
184
bonation time, accompanied with absorption of a large amount of C 0 2 .
(4) The carbonate products are vaterite and calcite; C,SH, formed from carbonation has a low BET surface area when compared to that of the C-S-H formed from curing in water.
References K. Quillin, Performance of belite-sulfoaluminate cements, Cem. Concl: Res., 31(2001), p.1341. F.P. Glasser and L. Zhang, High-performance cement matrices based on calcium sulfoaluminate-belitecompositions, Cem. Concr. Res., 31(2001), p.1881. E. Sakai, M. Morioka, K. Yamamoto, X. Zhang, Y. Ohba, and M. Daimon, Carbonationreaction of hardened low-heat cement, J. Ceram. Soc. Jpn. (in Japanese), 107(1999), p.561. K. Koibuchi, M. Morioka, E. Sakai, and M. Daimon, Strength development of low heat cement mortar by carbonation curing, Concr. Res. Technol. (in Japanese). 10(1999),p.65.
[5] M. Fernfindez Bertos, S.J.R. Simons, C.D. Hills, and P.J. Carey, A review of accelerated carbonation technology in the treatment of cement-based materials and sequestration of CO,, J. Hazard. Mater., B112(2004), p.193. [6] Y.Lo and H.M. Lee, Curing effects on carbonation of concrete using a phenolphthalein indicator and Fouriertransform infrared spectroscopy,Build. Environ., 37(2002), p.507. [7] B. Johannesson and P. Utgenannt, Microstructural changes caused by carbonation of cement mortar, Cem. Concr. Res., 31(2001), p.925. [8] L.C. Lange, C.D. Hills, and A.B. Poole, The influence of mix parameters and binder choice on the carbonation of cement solidified wastes, Waste Manage., 16(1996), p.749. [9] S. Kim, H. Taguchi, Y. Ohba, et al., The carbonation of calcium hydroxide and calcium silicate hydrates, Inorg. Mater. (in Japanese), 2(1995), p.18. [lo] J.J. Thomas, J. Hsieh, and H.M. Jennings, Effect of carbonation on the nitrogen BET surface area of hardened portland cement paste, Adv. Cem. Based Mater., 3(1996), p.76.
Carbonation of low heat portland cement paste precured in water for different time Deping Chen”, Etsuo Sakai2’,Masaki Daimon2’,and Yoko Ohba’’ 1) Civil and Environmental Engineering School, University of Science and Technology Beijing, Beijing 100083, China
2) Graduate School of Science and Engineering. Tokyo Institute of Technology, Tokyo 152-8552, Japan (Received 2006-06-01)
Abstract: The carbonation technique was applied to accelerate the hydration of low heat portland cement (LHC). Before carbonation, the demoulded pastes were precured in water for 0,2,7, and 2 1 d, respectively. The results show that precuring time in water strongly influences the carbonation process. The phenolphthalein test indicates that the paste precured in water for a shorter time is more quickly carbonated than that for a longer time. The content of calcium hydroxide increases with increasing the precuring time in water, whereas, the amount of absorbed carbon dioxide changes contrarily. Scanning electron microscope (SEM) observation shows that portlandite always fills up big air bubbles in the paste during precuring in water, and the mercury intrusion porosimetry (MTP) results show that there are less large capillary pores in the paste precured in water for a longer time. It is found that the paste without precuring in water has more carbon dioxide absorption during curing in carbon dioxide atmosphere, and its total pore volume decreases remarkably with an increase in the carbonation time than that precured in water. X-ray diffraction (XRD) and Brunauer-Emmett-Teller (BET) surface area analyses indicate that the carbonate products are vaterite and calcite; C,SH, formed from carbonation has low BET surface area in comparison with that of C-S-H formed from curing in water. Key words: carbonation: carbon dioxide absorption; vaterite; BET surface area; pore size distribution
[Thisstudy was financially supported bv the Ministv of Education, Culture, Sports, Science, and Technology, Japan.]
1. Introduction Low heat portland cement (LHC) is attracting a great deal of interest worldwide for its lower energy consumptions and CO, emissions on manufacture than ordinary portland cement. But its application is comparatively restricted because of the slow hydration and low early strength. These disadvantages can be improved by using a number of techniques, such as, activating of belite or introducing reactive additives [ 1-21. It was reported [3-41 that LHC could accelerate hydration under a C02 curing condition accompanied by highly developed strength. In this process, a significant proportion of C 0 2 was fixed by the carbonation reaction, which could provide an effective way of sequestering C02 as important solid phases to produce construction materials [5]. However, the precuring condition that is most suitable for the carbonation reaction of LHC, and the details of the carbonation process with different precuring time in water are not clear. K. Koibuchi et 111. (41 said short precuring time in water is in favor of absorbing carbon dioxide; E. Sakai er al. [3] investigated the carbonation reaction of hardened low Corresponding author: Deptng Chen. E-mail: [email protected]
heat cement under 2 d precuring in water. In this article, the effect of precuring time in water on the carbonation process of LHC was studied.
2. Experimental 2.1. Materials The LHC was from Taiheiyo Cement Co., Japan. Its composition (wt%) was: SO,, 26.12; Fe203, 2.29; A1,0,, 2.87; CaO, 63.94; MgO, 0.75; SO3, 2.26; Ti02, 0.12; LOI, 0.87. Its mineral composition (wt%) by Bogue is: f-CaO, 0.10; C3S, 32.30; C2S, 50.60; C3A, 3.74; C4AF, 6.96; CaSO,, 3.84, with a density of 3.22 g/cm3and a Blaine surface area of 3280 cm2/g.Moreover, an antifoamer and two compositions of the viscosity improver (viscoA and viscoB) were used to reduce air bubble formation when mixing, to prevent the slurry from segregation.
2.2. Preparation The mix proportions were: water-to-cement mass ratio 0.4, antifoamer to water 5 ~ 1 0 - and ~ , viscoA and viscoB to water 0.04 respectively. Initially, viscoA, antifoamer, and distilled water were mixed together, and Also available online at www.sciencedirect.com
D.P.Chen et al., Carbonation of low heat Portland cement paste precured in water for different time
then poured into the LHC, and stirred for 2.5 min. Next, viscoB was added and again stirred for 2.5 min. The slurry turned to viscous paste. The fresh paste was cast into a mould of 20 mmx20 mmx80 mm. The paste with mould, after vibrating for 1 min, was put into a closed plastic box for 24 h at 20°C. The demoulded samples were first immersed in distilled water for 0,2,7, and 21 d (it was called 'precuring in water' below), and then placed in the carbonation apparatus (SH-221 from ESPEC Corp.) for 0,3,7,21, 42, and 70 d (curing in CO,). Throughout the curing temperature is 20°C; the carbonation condition is relative humid, 60%, with 5~01%COz.
2.3. Testing methods (1) Measuring the depth of carbonation. The carbonation depth was measured in a traditional way, by spraying phenolphthalein alcohol solution onto the surface of a split prism of the carbonated LHC paste. The neutralization area of transect was then calculated. The ratio of the neutralization area to the whole area of transect was defined as the carbonation area ratio. (2) XRD analyses. The paste samples were crushed and manually grounded with an alumina mortar and pestle. Their phase compositions were examined by X-ray diffraction (XRD) with a Toshiba powder diffractometer using Cu K, radiation. (3) Thermogravimetric/differential thermal (TG/ DTA) analysis. The powder samples were submitted to thermal analysis using the TG-DTA 2000s equipment (Mac Science Co.), which permits performing TG and DTA analysis simultaneously. In testing, approximately 33 mg of powder samples were heated from room temperature to 1000°C at a heating rate of 10"C/min in flowing N, (150 mL/min). After testing, TG and DTA curves were obtained. With the help of TGDTA curves, the Ca(OH), content in the paste was determined from the following equation:
where CH(wt%) is the content of Ca(OH),, WL,H(wt%) is the weight loss of Ca(OH), that occurred during the dehydration of portlandite, MW,, and MWH are the molecular weights of Ca(OH), and water, respectively. (4) CO, absorption tests. The inorganic carbon content C(wt%) of the powder samples was tested at 200°C with 5 % H3PO4on alumi-
z 79
na carrier boats using TOC-505OA Total Organic Carbon Analyzer equipped with SSM-5000A Solid Sample Module (Shimadzu Corporation, Japan). The amount of CO, absorption by LHC was calculated by using the following equation: CO,(wt%) =
C(wt%)x MWco, ( 1- LOI%)x M w c
where CO,(wt%) is the amount of absorbed CO, by LHC, C(wt%) is the content of inorganic carbon, LOI% is the mass loss by heating at 1000°C for 30 min, MW,, and MWc are the molecular weights of CO, and carbon, respectively. (5) Brunauer-Emmett-Teller (BET) specific surface area analyses.
The specific surface area of the powder samples was analyzed by using the HORIBA SA-6200 serials BET surface area analyzer (Horiba Ltd., Japan). All the samples were placed in an Aspirator for 7 d to dry, and stored in sealed plastic vials before testing. (6) Pore size distributionand total pore volume tests. A piece of 5-10 mm sized sample was submitted to pore size distribution and total pore volume testing with mercury intrusion porosimetry (MIP) using the Pascal 140 and Pascal 240 porosimeter (CE Instruments, Italy). The samples were under 1 d of D-drying before testing.
(7) SEM observations. SEM observations for the characterization of carbonated LHC pastes were carried out using JSM-5310 scanning electron microscope (JEOL Ltd., Japan). To obtain a conductive surface for analysis, a thin gold layer was deposited over it.
3. Results and discussion 3.1. Carbonation area ratio The carbonation area ratio of LHC pastes with carbonation time was shown in Fig. 1. The neutralization rate of LHC paste without precuring in water is higher than that with precuring in water. The paste without precuring in water was soon fully carbonated at the carbonation time of 21 d, whereas, the others needed more time for it. As the phenolphthalein test is based on the color change at which the pH is about 9, it denotes the presence of Ca(OH), [6]. The content of Ca(OH), in the paste, obtained by TG/DTA analysis, with the carbonation time is shown in Fig. 2. The Ca(OH), contents of the paste without precuring and with 21 d precuring in water before carbonation are 5.92wt% and 12.22wt%, respectively. After 21 d carbonation, the
. I Univ. . Sci. TechnoL Beijing, Vo1.14, N0.2, Apr 2007
180
Ca(OH), content without precuring in water falls to Owt%, whereas, that with 21 d precuring in water still remains at 10.74wt%. This indicates that less precuring time in water enables the paste to neutralize easily.
.-
/ :"'
80-
+a
i?
8 60
-
3
./,\
.+
40-
j
20-
a
,
/
,
0I
/
Precuring time in water -.-Od 2d A 7d - ~ - 2 1 d I
v.
,
0
10
20 30 40 50 60 Carbonation time / d
70
Fig. 1. Carbonation area ratio of LHC pastes.
.
T--
v -
1
-.
.
Precuring time in water
.\.\
A
\
,
0
..-
\.
10
20
30 Carbonation time / d
40
Fig. 2. Contents of Ca(OH), remains in the pastes.
3.2. MIP result and SEM observation Fig. 3 shows the curves of cumulative pore volume with pore diameter in the carbonated pastes. The total pore volume of the paste without precuring in water decreases considerably with carbonation time pro-
w
longed from 3 to 42 d, and the amount of large capillary pores (>0.1 pm) also apparently diminishes; but comparatively those with 21 d precuring in water only slightly change with carbonation time (Fig. 3(a)). The total porosity of the paste without precuring in water decreases from 33.4% at 3 d carbonation to 17.1% at 42 d carbonation time, but that with 21 d precuring in water increases from 22.0% at 0 d carbonation to 24.6% at 42 d carbonation time. The carbonated paste without precuring in water has more large capillary pores (>0.1 pm) than those with precuring in water (Fig. 3(b)). It means that precuring in water is favorable for the decrease of large capillary pores. The longer the precuring time in water it is, the fewer large capillary pores it has. The fracture surface of the carbonated LHC pastes under SEM, as in Fig. 4, indicates the pastes without precuring in water seem looser in microstructure than those with long precuring, for example, 21 d in water. Their big air bubbles always remain empty and some of them are filled with a few Ca(OH)2crystals, and large capillary pores exist everywhere in the paste (Fig. 4(a)), even at 42 d carbonation (Fig. 4(b)). On the contrary, the big air bubbles in the paste with 21 d precuring in water are always filled with many tabular Ca(OH), (Fig. 4(c)), and the existence of large capillary pores are not so popular in the paste as mentioned above. But carbonation microcracks are easily found on the surface and in the inner part of the pastes with 21 d precuring in water (Fig.4(d)), even at 3 d's carbonation time (Fig. 4(c)). The existence of microcracks would result in the increase of total porosity in the paste at the carbonation time of 0 to 42 d, measured with MIP.
-Fg . -g 100 w
\
Precuring time Curing timc
140
E
80
2 n
60
.-9
40
m
3 c) 5
Ib)
\,
120
9
o
Pore diameter / pm
160
20 0
3
0.01
Precuring time Curing time in CO,
\
:"I2, 3
2d 7d
0.1 1 Pore diameter / pm
42 d 42 d 42 d
10
Fig. 3. Cumulative pore volume with the diameter of pores for carbonated LHC paste: (a) the effect of carbonation time on pore volume; (b) the effect of precuring time in water on pore volume.
3.3. Amount of absorbed C 0 2 The amount of absorbed CO, for the paste without precuring in water increases quickly at the early car-
bonation age before 7 d, and then slowly changes with the carbonation time as shown in Fig. 5. It has the highest amount of absorbed C02 when compared to those
D.P. Chen et d,Carbonationof low heat portland cement paste precured in water for different time
with precuring in water, although, in the same carbonation time. The amount of absorbed COz for the paste without precuring in water is nearly 25wt% at 7 d carbonation, approximately six times that of 21 d pre-
181
curing in water. The less the precuring time in water is, the more amount of CO, the paste absorbs. It also reveals that precuring in water is inferior for the pastes to absorb CO,.
Fig. 4. SEM photographs of the fracture surface for the carbonated LHC pastes: (a) precuring time in water: 0 d, carbonation time: 3 d; (b) precuring time in water: 0 d, carbonation time: 42 d; (c) precuring time in water: 21 d, carbonation time: 3 d; (d) precuring time in water: 21 d, carbonationtime: 42 d.
s
30.0
.
25.0 I
g 20.0
,,’ /
-0
$
s:
15.0
-.’
3 10.0 cr 0
z
/
_,
5.0
3
d
--.-Od A
7d
-tT
2d 21d
0.0
0
10
20 30 40 50 Carbonation time / d
60
70
Fig. 5. Amount of absorbed CO, of the carbonated LHC pastes.
3.4. XRD and TG/DTA results XRD analysis shows the predominant phases of the carbonated LHC paste are vaterite, calcite, portlandite, and belite (p-C,S) (Fig. 6). Vaterite and calcite are two forms of calcium carbonate (CaC03) caused by carbonation. Portlandite Ca(OH), is a hydrate phase of LHC, and belite is an anhydrous cement phase, which remains during the curing process. The peaks of vaterite in the paste without precuring in water are identified easily at 3 d carbonation time, and their intensities get stronger with increasing the carbonation time (Fig. 6(a)). However, with 21 d pre-
curing in water, the peaks of vaterite can still not be found in the paste at 21 d carbonation (Fig. 6(b)). By contrast, the peaks of Ca(OH), could be identified in all the XRD patterns of the pastes, even at 42 d carbonation without precuring in water. It is different from the results of phenolphthalein test and TG/DTA analysis, which show that Ca(OH), does not exist any longer at 42 d carbonation without precuring in water. This paradox could be explained as follows. As shown in Fig. 4(a), a few crystals of Ca(OH), formed inside the air bubbles when the paste was still in mould for one day. The crystals grew less homogeneously everywhere. Some of the crystals grew tabularly, with a perfect structure, whereas, others lamellarly or scatteringly with defects in structure. The Ca(OH), crystals with a perfect tabular structure were difficult to carbonate, and remained in paste even after 42 d carbonation. This could be detected by XRD. However, the remaining Ca(OH), was very little in quantity, existed mainly in the air bubbles and was covered with a thick carbonate layer after a longer carbonation time. It is possible that it could not be identified by a phenolphthalein test in this condition. On the other hand, when the paste was crushed and manually ground to powder, the small Ca(OH), granules could not distribute homogeneously. Moreover, the amount of powder
J. Univ. Sci. TechnoL Beging, voL14,N0.2, Apr 2007
182
sample submitted for thermal analysis was as low as approximately 33 mg. Thus there is a possibility that
V
Ca(OH)* could not be detected by TG/DTA analysis. (b)
Portlandite .Calcite Vaterite A p-C,S
Precuring in water 0 d carbonation time
Portlandite 0 Calcite
8
1
30
Precuring in water: 21 d carbonation time:
I
I
60
50
40
p-C,S
8
I
20
A
V Vaterite
30
20
201(")
40 2e1~)
I
I
50
60
Fig. 6. XRD patterns for carbonated LHC pastes with 0 and 21 d precuring in water: (a) precuring in water for 0 d; (b) precuring in water for 21 d.
The typical TGlDTA curves of the LHC paste (Fig. 7) show that there are three major endothermic reactions along with mass loss, which occur during heating: (1) the release of the evaporable and part of the adsorbed water at 100°C approximately; (2) the dehydration of Ca(OH), at 400-500°C; and ( 3 ) the dissociation of carbonate phases at 620-780°C. Without precuring in water, the total mass losses of pastes as shown in Fig. 7(a), increase quickly from 9.7% at 0 d carbonation to 22.4% and 28.7% at 3 and
21 d carbonation time, respectively. When the carbonation time exceeds 2 1 d, the total mass losses hardly change. The endothermic peak at 400-500°C presenting the dehydration of Ca(OH), disappears at longer carbonation time, such as, 21 and 42 d. By comparison with the precuring time in water of 21 d, the total mass losses of pastes as shown in Fig. 7(b), almost do not change at 0,3 and 21 d carbonation time and are seen as 18.496, 19.496, and 19.796, respectively.
I
Carbonation time
100
100
95
5
s
90
m
m
2
4
85
m
80 75
'Precuring time in water 0 d
42
d21
-0.5
'
5
95 90 85 80
75
recuring time in water 21 d
70'
200
400
600
SbO
lo00 -0.5
Temperature I "C
Temperature / SS
Fig. 7. TG/DTA curves for the carbonated LHC paste: (a) precuring in water for 0 d; (b) precuring in water for 21 d. Precuring time in water
3.5. BET surface area Fig. 8 shows the BET specific surface area of LHC pastes with carbonation time at various precuring times in water. The pastes without precuring in water have a low BET surface area of 9.98-14.46 m2/g despite the carbonation time being changed from 0 to 70 d. But the pastes with 21 d precuring in water have a comparatively large surface area of 30.66-33.08 m2/g at the early carbonation age before 7 d, which decreases sharply to 15.9 m2/g at 21 d carbonation.
90 E v
.
-A-
25
7d
-~--21 d
m
s
t; m
: 0
10
20 30 40 50 Carbonation time / d
60
70
Fig. 8. BET surface area of the carbonated LHC paste.
D.P.Chen et al., Carbonationof low heat portland cement paste precured in water for different time
3.6. Discussion LHC pastes precured in water for different time, before curing in C02, are different in many aspects in terms of phase composition,pore size, pore distribution, etc. The paste without precuring in water has the lowest Ca(OH), content (5.92wt%), highest porosity, plenty of big air voids and a large number of capillary pores, compared to those precured in water for 2,7, and 21 d. They seem to form two kinds of Ca(OH), in the paste. One of them scatters in the paste in small or lamellar crystals with defects in structure, which can be easily carbonated, and the other is inside the air bubbles, always in tabular crystals with perfect structure. When the paste was cured in C02, the scattered Ca(OH),, along with other initial hydrates such as AFt or AFm, were soon carbonated. The carbonation reactions are simply expressed as [3,5] Ca(OH)2+C0,-CaC0,+H,0
(3)
3CaO~Al2O3~3CaSO,~32H,0+3CO2+ 3CaC03+ 3(CaSO4~2H20)+Al2O3~nH20(gel)+(26-n)H2O (4) 3CaO~A1203CaS0,~ 12H20+3C02+ 3CaCO,+ CaS04~2H,0+A1203~nH20(gel)+( 10-n) H20
(5)
At the same time, tabular crystals of Ca(OH), are also carbonated via formula (3). With the progress of carbonation, C3S and C2S participate in reaction with C 0 2 and H,O. The overall stoichiometry of the carbonation reaction of C3S and C2S,accompanied by hydration, follows as [7-81 C3S+yH20+(3-x)C02 C,S+yH,O+( 2-x)C02
C,SH,+( 3-x)CaC03
(6)
C,SH,+( 2-x)CaC03
(7)
+
+
Then, further carbonation, changing the composition of the C-S-H gel, occurs.
C,SH,+zCO,-C,_,,SH,+zCaCO,
(8)
Thus, a large amount of C02 was absorbed, and calcium carbonates formed increasingly with carbonation time. The above carbonation reactions were determined mainly by the diffusivity and reactivity of CO,. Paste without precuring in water has high porosity and a large number of connecting air voids and capillary pores, which are in favor of the diffusion and permeation of CO, through the pores. The carbonation products fill up most of the capillary pores, which decreases the total pore volume considerably. On the contrary, the big air bubbles and large capillary pores in the paste, with the precuring time as long as 21 d, are always filled with hydration products during precuring in water. It hinders C02 from diffusing
I83
into the paste, and results in a comparatively low carbonation effect. Microcracks are formed easily on the surface of the paste on account of the huge volumetric dilation during the carbonation of Ca(OH),. Then they stretch and spread into the paste, which becomes the passage for the permeation of C02. With the progress of carbonation,the microcracks form abundantly inside the paste, along with further carbonation, and this is the reason that the amount of absorbed CO, increases noticeably after 21 d carbonation. The hydration products of the LHC paste during precuring in water are mainly Ca(OH), and C-S-H. The BET surface area is determined predominantly by C-S-H, because of the low value of Ca(OH), when compared with C-S-H [9]. The BET surface area of the paste with 21 d precuring in water before carbonation is 30.66 m2/g,and after 21 d carbonation, it falls to 15.09 m2/g. The major carbonation products are vaterite, calcite, and C,SH,. E. S a k i et al. [3] have given the BET surface area of vaterite synthesized from mixing 1 M Na2C03with 3 M CaCl, solution at 20°C for 10 min as 12.30 m2/g;and that of two kinds of calcite as l .77 m2/gfor calcite (A) prepared from vaterite by being heated at 450°C and 0.83 m2/gfor calcite (B) of special grade of calcium carbonate. Although these carbonate phases show a low surface area, their fraction is also low at 5.55% to the mass of cement obtained by TG/DTA analysis. Therefore, it can be concluded that C,SH, formed from carbonation has a much lower BET surface area than that formed from curing in water. J. Thomas et al. [lo], also reported a progressive decrease in the nitrogen BET surface area of dried hydrated portland cement paste during storage in plastic vials, which was found to be caused by carbonation, a reaction between the C-S-H phase and carbon dioxide in the air. Thus, the carbonated paste without precuring in water, exhibiting low BET surface area can also be explained.
4. Conclusions (1) Before curing in CO,, the LHC pastes become much denser in microstructure with increasing precuring time in water. Their big air voids and large capillary pores are always filled with hydration products during curing in water. (2) A longer precuring time in water means less large capillary pores, which hinders carbon dioxide from diffusing into the paste, and results in low amount of absorbed C 0 2 and high content of Ca(OH), remained in the carbonated pastes. (3) The total pore volume of paste without precuring in water decreases remarkably with an increase in car-
J. Univ. Sci. TechnoL Beging, voLz4,N0.2, Apr 2007
184
bonation time, accompanied with absorption of a large amount of C 0 2 .
(4) The carbonate products are vaterite and calcite; C,SH, formed from carbonation has a low BET surface area when compared to that of the C-S-H formed from curing in water.
References K. Quillin, Performance of belite-sulfoaluminate cements, Cem. Concl: Res., 31(2001), p.1341. F.P. Glasser and L. Zhang, High-performance cement matrices based on calcium sulfoaluminate-belitecompositions, Cem. Concr. Res., 31(2001), p.1881. E. Sakai, M. Morioka, K. Yamamoto, X. Zhang, Y. Ohba, and M. Daimon, Carbonationreaction of hardened low-heat cement, J. Ceram. Soc. Jpn. (in Japanese), 107(1999), p.561. K. Koibuchi, M. Morioka, E. Sakai, and M. Daimon, Strength development of low heat cement mortar by carbonation curing, Concr. Res. Technol. (in Japanese). 10(1999),p.65.
[5] M. Fernfindez Bertos, S.J.R. Simons, C.D. Hills, and P.J. Carey, A review of accelerated carbonation technology in the treatment of cement-based materials and sequestration of CO,, J. Hazard. Mater., B112(2004), p.193. [6] Y.Lo and H.M. Lee, Curing effects on carbonation of concrete using a phenolphthalein indicator and Fouriertransform infrared spectroscopy,Build. Environ., 37(2002), p.507. [7] B. Johannesson and P. Utgenannt, Microstructural changes caused by carbonation of cement mortar, Cem. Concr. Res., 31(2001), p.925. [8] L.C. Lange, C.D. Hills, and A.B. Poole, The influence of mix parameters and binder choice on the carbonation of cement solidified wastes, Waste Manage., 16(1996), p.749. [9] S. Kim, H. Taguchi, Y. Ohba, et al., The carbonation of calcium hydroxide and calcium silicate hydrates, Inorg. Mater. (in Japanese), 2(1995), p.18. [lo] J.J. Thomas, J. Hsieh, and H.M. Jennings, Effect of carbonation on the nitrogen BET surface area of hardened portland cement paste, Adv. Cem. Based Mater., 3(1996), p.76.