Influence of Ca(OH)2 and CaSO4 · 2H2O on hydration reaction of amorphous calcium aluminate

CEMENT and CONCRETERESEARCH.Vol. 20, pp. 824-832, 1990. Printed in the USA. 0008-8846/90. $3.00-~00. Copyright(c) 1990 PergamonPress plc.

I N F L U E N C E OF Ca(OH)2 A N D CaSO4 • 2H20 ON H Y D R A T I O N REACTION OF A M O R P H O U S CALC[UM A L U M I N A T E .

K. Nakagawa, I. Terashima, K. Asaga,* and M. Daimon,* Denki Kagaku Kogyo k.k. Ohmi Factory, Niigata-Prefecture. *Department of Inorganic Materials, Tokyo Institute of Technology, Meguro-Ku, Tokyo, Japan. (Communicatedby J.P. Skalny) (ReceivedMarch22, 1990) ABSTRACT The influences of Ca(OH)z and CaSO4 • 2I-I20 on the hydration of amorphous calcium aluminate with the chemical composition similar to C,zA7 were studied by determining the degree of hydration according to the selective dissolution method. The hydration reaction of amorphous calcium aluminate progresses very rapidly and amorphous AH3 is produced along with C2AH8. The reaction is retarded slightly by the addition of Ca(OH)2. When a small amount of Ca504 • 2H20 is added, the degree of hydration is increased compared with the case of no addition. However, the hydration reaction is greatly retarded, in the presence of large amount of CaSO, • 2HH.~O. When Ca(OH)z and CaSO, • 2H20 are added, the retarding effect is larger than in the case of adding 1 mole of CaSO, • 2H20 alone. ( A b b r e v i a t i o n s : A = A l 2 0 3 , C=CaO, S=S03, H = H 2 0 )

Introduction In a previous paper (1), the applicability of the selective dissolution method using methanol solution of salicylic acid to determine the degree of hydration of amorphous C12A7 (hereinafter Am. C.A.) was reported. The Am.C.A. modified with a suitable amount of gypsum is the main component in quick setting admixtures for shotcreting under tunnel construction and for emergency works. In the present study, the hydration reaction of Am.C.A. was examined by the selective dissolution method in the presence of CH and CSHz in order to obtain basic informations of the abovementioned practical systems.

Experimental Materials and Hydration Procedure The clinker, CH, CNH2, and the hydration procedure were the same as described in the previous paper (1). One mole of CH was added to 1 mole of Am.C.A. while in the case of CSH2 1 mole and 2 moles were added. When CH and CSHz coexisted, 1 mole each were added. 824

The additives The hydration

were dispersed

reaction

The degree

in deionized

was stopped

dryed in the same as described

Determination

825

AMORPHOUSCALCIUMALUMMATE,HYDRATlON.CH,C~H2

Vol.20.No. 5

a certain

in the previous

of Hydration of hydration

after

water,

and mixed curing

time,

quickly with Am.C.A. at 20°C. and the hydrated samples were

paper.

Degree

was calculated

from the ignition

loss of the hydrated

sample

and the

soluble amount in the selective dissolution procedure reported in our previous paper (1). When both CH and CSH, were added, the hydration degree was calculated assuming the all CH added had dissolved completely in methanol solution of salicylic acid, while CSH, would not dissolve

at all,

Identification

and remain

completely

and Quantitative

to form CSH+ when dried at 110°C.

Determination

of Hydrated

Products

Powder X-ray diffraction (XRD) was performed for the identification, and differential scanning calorimetry (DSC) was used for the quantitative determination of CH and CSH, remaining in a hydrated sample, and &AH,, C,A3CSH,, and CsACSH,, which formed during the hydration porcess. The reagents of CH and CSH, were employed as references. and for ettringite and monosulfate. reference samples were synthesized by hydrating GA with CSH,, and drying the products according to the procedure used in the present study. The calorimetric data used for quantitative determination were 430°C and 22.2 kcal/mole for for C3A3CSH,,, 187°C and 45. CH, 14O’C and 27.7 kcal/mole for CSH,, 115°C and 272.5 kcal/mole 2 kcal/mole for GACSH,, respectively. It was also impossible to synthesize a pure C,AH,, and calorimetric data could not be available in the literature. The calorimetric data for &AH, were obtained from DSC curve of C,,A, paste hydrated for 30 minutes at 20°C. The quantity of heat was calculated from the degree of hydration and the endothermic peakarea of C,AH, (75’C), assuming that the hydration of C,,A, expressed by proceeded stoichiometrically, The calculation gave a quantity as shown in the following equation. of heat of 10.3 kcal/mole. &A, + 51H + G&AH, + AH, The specific surface area of hydrated samples were determined by the krypton gas adsorption method (BET), and hydrated products were observed by scanning electron microscope (SEM).

Test Results 1 ) The results of ignition loss are shown in Fig. 1. A loss of 14 % was indicated at 15 seconds after with subsequent gradual increase up to the addition of water for hydrated sample with no additive, 29 % after 30 minutes. In case of adding CH, approximately the same values were indicated as When with the case of no additive until 1 minute, but subsequently, values became slightly lower. 1 mole of CSH, was added, it was 8 % at 15 seconds, but there was a rapid increase until1 30 seconds. Afterward, there was a gradual increase in the ignition loss and after 30 minutes 35 % was indicated. For a sample with the addition of 2 moles of CSH*, 9 % was indicated at 15 seconds When 1 mole each of CM and only a slight increase was observed to reach 17 % after 30 minutes. and CSH, were added, the ralue at 15 seconds was the lowest, but there was a rapid increase by 30 seconds to show a trend similar to the case of adding 1 mole of CSH,. However, subsequent values were lower. 2 ) The influence of CH and CSH, on the hydration degree of Am.C.A. determined by the selective dissolution method are shown in Fig. 2. The general trend coincided with that of ignition loss in Fig. 1. In the case of no additive, hydration degree of 24 % was indicated at 15 seconds after the

826

K. N a k a g a w a , et al.

!00 50

O No add. n OH I mle z~ CSH., 1mole

~40

CSH., 2moles

•

o=

20, No. 5

0 No add

_N~

n CH 1mole

"-'80 E .0 "~ 60

n

o•30 :.,,3 "E

V~q

C.SH: 1mole • C-SH~ 2moles • CH C S H 1 mole

/ ~ ~ J "

.1:: 4 0 o

20

~ 20

-~0

a

o 0

0

i

i

15

i

30

1

i

i

5

15

(sec)

i

15

30

i

1

I

5

(sec)

30

i

15

30

(min)

(rain)

Time (log scale)

Fig. 1 Variation of ignition loss with hydration time.

Fig. 2

Variation of the degree of hydration with hydration time.

addition of water, with subsequent gradual increase up to 60 % at 30 minutes. With the addition of 1 mole of CH, the rate of hydration reaction became slightly slower. The addition of 1 mole of CSH2 retarded the hydration reaction till 15 seconds, but after that, it was accelerated sharply and the degree of hydration became 88 % after 30 minutes. When 2 moles of CSH~ were added the reaction was much retarded and the degree of hydration at 30 minutes was only 15 %. Under the coexistence of 1 mole each of CH and CSH2, the hydration reaction was retarded up to 15 seconds even more than the case of adding CStt~ alone. However, similarly to the case of CSI-t~ only. a sudden acceleration occurred, after 30 seconds, and the degree of hydration became greater compared with no additive. Still, the degree of hydration was smaller than that for the case of adding 1 mole of CgH2. ul .'-',100

-- No add.

1O0

-- OH 1 mole

0 C., AH~

so

6o

L

~ c t e d

amorphous p h a s e / ' ' ~ q

3~

/: /

.,-,'~

4O ~

Unreacted amorphous phase

e--

°- la 8 0

• CH

.E 60

:~AH~

~}~40

S.

E

~

20

g

--~'' 0

15

30

E 1

5

(sec)

1-5

~O

(min)

Time (log scale) Fig. 3 Amounts of hydrated and unhydrated phases vs. hydration time. (with no additive and with CH 1 mole.)

15

30

(sec)

1

5

15

30

(min)

Time (log scale) Fig. 4 Amounts of hydrated and unhydrated phases vs. hydration time. (with CSH2 ] mole.)

Vol. 20. No. 5

827

AMORPHOUS CALCIUM ALUMINATE, HYDRATION, CH. C?H2

3 ) The amounts of hydration phase by the selective dissolution

products determined by DSC, and that of unreacted amorphous method are shown in Fig. 3, 4, 5. and 6. As shown in Fig. 3. the

only hydration product recognized by XRD was &AH, in the case of The amount of &AH, produced in 15 seconds after the addition smaller than that expected from the results of ignition loss and degree found that that amount of &AH, after 5 minutes were approximately

no additive. of water nas considerabl!. But, it \vas of hydration. the same as the calculated

from the degree of hydration. When CH was coexisted, the main product recognized by XRD was &AH, as shojvn in Fgi. 3. According to the results of DSC. CIl while the formation of C,AH, was also seen after 5 minutes. decreased gradually, and disappeared completely after 15 minutes. The amount of C,AI-I. produced was fairly smaller than the case of no additive. \vas the onI\4 ) The result in the case of adding 1 mole of C$?HZ is shown in Fig. 4. Ettringite hydration product at 15 seconds. At 30 seconds, the amount of ettringite produced \vas decreased and monosulfate was produced. Almost no CSH? could be found after 1 minute and the production of &AH, was observed, while the amounts of monosulfate and ettringite were 7 to 8 90 and 2 to :i % respectively. Thereafter, the amounts of ettringite and monosulufate were kept almost constant, while C, AH, increased markedly. It is necessary to note, however, that the amount of C,AII, measured \vith the first peak of by DSC might be larger than the real value, because of partial overlapping monosulfate at 90°C. 5 1 Ettringite was the only crystalline hydration product, when 2 moles of CSH, were added and and a colisiderable amount of CSH, remained until 30 minutes, as shown in Fig. 5. Monosulfate &AH, were not produced. Therefore, it may be considered that the reaction of this system progresses very slowly, producing ettringite. 6 1 Fig. 6 shows the result in case of adding CH and CSH, indicating that the initial hydration \vas retarded more strongly than the case of adding CSH, only, and that no hydration product could be recognized at 15 seconds. At 30 seconds, 2 % of ettringite was produced but it was not detected after 1 minute. Monosulfate was recognized at 30 seconds and increased gradually. C,AH, was m

00

loo-

0 C,AH8 Unreacted amorphous phase

80

a Ett. A MS. 0 CSH,

60

(min)

(set)

Time (log Fig. 5 Amounts

of hydrated

and unhydrated

Time (log Fig. 6 phases

Variation

CH

(min)

(s-2)

scale)

vs. curing time. (with CSH, 2 moles.)

\

n

of hydrated

scale)

and unhydrated

vs. hydration time. (with CSH, 1 mole, and CH 1 mole.)

phases

828

K. Nakagawa, et al.

Vol

20, N o 5

=o 40

v 30

2o m

1 m o l eCgH2 /

G) Q.

ar I

m

o

15

I

30

JI

1

CgH2

I

5

(sec)

2moles

I I

I

15

30

(rain)

Time (log scale) Fig. 7 V a r i a t i o n of specific surface area of h y d r a t e d samples vs. h y d r a t i o n time.

observed at 1 minute after, and began to increase at 5 minutes when CSH~ disappeared. Calcium hydroxide decreased gradually, but there was 2 % remaining even after 30 minutes. 7) Fig. 7 illustrates changes in the specific surface area of hydrated samples containing different additives. The specific surface area of the additive-free sample was about 20 m u g at 15 seconds, and increased up to about 3(I m2/g at 1 minute, but almost no increase is observed subsequently. For the sample containing 1 mole of CSH~, the specific surface area was 6 m2/g at 15 seconds, followed by a rapid increasd up to 19 m2/g at 30 seconds. Subsequently the specific surface area gradually increases with time, reaching 33

m~/g after 30 minutes. In contrast to these samples, the reaction was retarded for the sample with 2 moles of CSH2 and the specific surface area remained around 4 m2/g from 15 seconds until 30 minutes, followed by a gradual increase up to 10 m2/g after 90 minutes. 8 ) Fig. 8 shows SEM photographs of hydrated samples. Fig. 8 a) is a sample with no additive at 15 seconds hydration. The card-house structure of hexagonal plates was observed at the surface of unhydrated grains. Fig. 8 b) is a sample with 1 mole of CH at 30 seconds hydration, and shapes and sizes of hydration products were similar to those in the case of no additive. Fig. 8 c) is a sample hydrted for 15 seconds with 1 mole of CSH~. Needle-like ettringite had been produced at the surface of particles. There were many cardhouse-like hydration products on grain surfaces at 1 minute, as shown in Fig. 8 d). This product is a stable form compared with Fig. 8a) and8b). Howver, in the case of adding 2 moles of CSH2, a small amount of cardhouse-like hydration products was observed at 30 minutes after addition of water, as shown in Fig. 8 e). Fig. 8 f) shows a sample hydrated for 15 seconds under the coexistance of 1 mole each of CH and CSH2. Distinct hydration products were not observed at the surface of Am.C.A. particles, because the hydration reaction was retarded as shown in Fig. 1, 2. The forms of subseqently produced hydration products were the same to the case of adding 1 mole of CSH2.

Discussions It is known that C2AH8 and AH3 are produced at the initial stage of hydration reaction of C~2A7 without additive at 20°C, according to the equation of C~2AT+51H--'6C2AHs+AH3. Although the production of crystalline C~AHs was ascertained in this study, crystalline AHa was not recognized. There was a good coincidence between the amount of hydration products calculated from values of ignition loss based on the above-mentioned equation and the amount of dissolution in the selective dissolution method.

Vol. 20, No. 5

AMORPHOUSCALCIUMALUMINATE,HYDRATION,CH, CSH2

829

It can be concluded that the hydration reaction of Am.C.A. progresses similarly to the above equation and amorphous AHa is produced. The degree of hydration is as high as 24 % at 15 seconds and the specific surface area was also increased rapidly. As a consequence, it may be considered that the flash setting property of Am.C.A. is caused by the C2AH8 and colloidal AH3 engulfing with large amounts of water.

a

b

c

d

e

f

Fig. 8

SEM microphotographs of hydrated samples.

830

K. Nakagawa, et al.

Vol. 20, No. 5

It is considered that amorphous calcuim aluminate hydrates and AHa gel have been produced besides crystalline C~AHs, in the hydration reaction at an extremely early stage within 1 minute. because of the facts that the amount of C,AH8 produced is smaller than the value expected from the degree of hydration, and tha specific surface area is conversely large. Specific surface area was not increased at hydration peried longer than 1 minute, in spite of the degree of hydration being increased, and the amount of C2AH8 produced matches the degree of hydration. It is seggested that Am.C.A. hydrates produced at the initial stage crystallize into C2AH~. R. Kondo and S. Ueda (2) expressed their reaction rate equation as ( 1 - I4T7-~~) n = kt, where ot was degree of hydration, and t was hydration time. When n equals 1, reaction rate is constant regardless of the thickness of the reaction layer, and where n equals 2, reaction rate must be controlled by diffusion through reaction layer. If n is a value higher than 2, resistance to diffusion is supposed to be increased due to the reaction layer becoming denser. Fig. 9 shows the log-log plotting of ( 1 - x / 1 - ~ ) and time (t). The variation in value of n obtained from the figure are summarized in T a b l e 1. In case of no additive and addition of CH, n was constant at 4 from 15 seconds. This was due to rapid increase of diffusion resistance, because particle surfaces were covered by amorphous hydrates and AHa gel. Further, it m a y be noted that reaction was slightly retarded when CH was added. According to the results of DSC, the amount of C2AH8 produced was smaller than that in the case of no additive, suggesting that the amount of amorphous hydrates was larger. It is considered that hydration reaction was retarded by the formation of this amorphous hydration products. In the case of C~,A, reported in the previous paper (1), n was 1.5 from 15 seconds untill 5 minutes after addition of water. The value of n was 3 from 5 to 30 minutes after water addition. It is conseivable that reaction rate is controlled by surface chemical reaction at first, and that diffusion

] 0

" 0 •

N o add. Amorphous N o add. Crystal

[] C H

1mole

1mole

A CSH2

~

• CS Hz ~molas

0.01

f 15

1 30

~

I 1

I

.

~

'

~

J 5

(sec)

'

~j ~

-

~

i 15

I :30

(rain) Time (log scale)

Fig. 9

Relationship of log (1--a lx/'fz'-[-~) and log t of the hydrated samples.

Vol. 20, No. 5

AMORPHOUSCALCIUMALUMINATE,HYDRATION,CH, CSH2

831

Table 1 Variation of n value on hydration degree vs. hydration time. Curing time Sample

(sec) 15

30

(min) 1

5

15

35

No add.

4

)

CH 1 mole add.

4

)

CSH2 1 mole add.

--0.5-):

2.5

-- 0.3-)

4

)

-

4 -)

CSH2 2 moles add. CSH2+CH add. crystalline C~2A7

1.5

) ')

3--) ( 1 - lx/~-a)"=Kt

resistance at particle surfaces increased later. It has become apparent in this way that the hydration kinetics of Am.C.A. is different from crystalline C12A7. When 1 mole of CSH~ was added with or without 1 mole of CH. hydration reaction was strongly retarded at initial stage, and there followed violent reaction with n value smaller than 1. It was considered that particle surfaces were thinly covered by a reaction product containing SO3 produced at the extremely early stage (3). According to the traditional theory (4), covering membrane is broken with reduction in diffustion resistance, when it is crystallized into AFro. After 30 second. C~AH~ and AH3 gel were formed and the reaction progressed similarly to the case of no additive. When CStt2 was added with CH, hydration reaction was slightly slower compared to the case of only CSH~ was added (5). In the case of adding 2 moles of CStf~. n was 5 from 15 second untill 30 minutes, and it may be considered that a fairly dense membrane had been formed. As described previously, Am.C.A. shows a very rapid reaction, and except when 2 moles of CSH~ have been added, the value of n becomes roughly the same at 5 minutes after the addition of water. This means that the violent reaction occured immediately after contact with water can be retarded, and that subsequent reactions are controlled by diffusion through hydrated layers at particle surfaces formed during violent reaction. Conclusions 1. The hydration reaction of Am.C.A. at 20°C proceeds stoichiometrically described by the equation below. A part of calcium aluminate hydrate must be amorphous at initial stage. C,2A7 + 51H --, 6C2AH~ + AH3 2. This reaction is extremely rapid and specific surface of 20 m2/g was indicated at 15 seconds after the addition of water. The flash setting property of Am.C.A. must be caused be the formation of C~AH8 and AH3.

832

K. Nakagawa.et al.

Vol. 20, No. 5

3. The rate of hydration reaction is slightly retarded when 1 mole of CH is added. 4. Hydration is retarded immediately after the addition of water for several tens of seconds, when 1 mole of CSH2 is added with and without 1 mole of CH, but the subsquent hydration is acceierated. The retardation may be caused by the formation of protective layers of amorphous product containing SO3. The coexistence of CgHa and CH makes the retarding effect slightly greater. 5. In the case of adding 2 moles of CSH2, hydration is retarded immediately after the addition of water, while subsequent hydration is also enormously retardedd.

References 1. K. N a k a g a w a , I. Terashima, K. Asaga and M. Daimon, Cem. Concr. Res. (to be publishied). 2. R. Kondo and S. Ueda, Proc. 5th International Symposium on the Chemistry of Cement, Tokyo, Vol. 2, p. 232-248 (1968). 3. J. S k a l n y and J.F. Young, 7th International Congress on the Chemistry of Cement, Paris, Vol. II-1, p. 18-36 (1980). 4. H.E. Schwiete, V. Ludwig and P. Jager, Symposium on the Structure of Portland Cement Paste and Concrete, Special Report 90, Highway Research Board, Washington, P. 353-367 (1966). 5. N. Tsuyuki, N. Hirota, T. Miyakawa and J. Kasai, J. Cerm. Soc. Japan, 92, 554-561 (1984).