Effect of silica fume on the chloride and sulphate attack ...nopr.niscair.res.in/bitstream/123456789/24365/1/IJEMS 8(4) 228-234.pdf · Indian Journal of Engineering & Materi als Sciences

Indian Journal of Engineering & Materi als Sciences Vol. 8, August 2001 , pp. 228-234

Effect of silica fume on the chloride and sulphate attack of sulphate resisting and high alumina cement composite

I M Helmy", H El-Didamony", A H Alib & T M EI-Sokkarl aFaculty of Science, Zagazig University, Zagazig, Egypt

bBuilding Research, Centre, Dokki, Cairo, Egypt.

Received 27 November 2000; accepted I Jun e 2001

T he present investigation aims to study the chloride as well as sulphate attack on the composite of 85% SRC and 15% HAC. T he effect of substitution o f HAC with silica fume on the aggressive attack was also studied . The results indicate that the free lime contents decrease with si lica fume and curing time in sulphate as well as chloride solution. The XRD resu lts are in a good agreement with the compressive strength and chemical analys is. The substitution of HAC with SF improves the attack of chloride and sulphate ions on cement pastes. It can be concluded that superplastecizer must be added to the sil ­ica fume blended cements.

It is we ll known that chlorides constitute damage for reinforcing steel embedded in concrete as they promote corrosion in addition to the reaction with the cement pastes of the concrete. The ch loride ion takes part in chemical reac tions simil ar to those involving in the sulphate ion and yields chloroaluminate hydrate '. However, the corrosive ac tion of the chloroaluminate is significantly different from sulphoaluminate in that the fo rmer does not cause ex pansion but softening2

.

It has been recom mended to use dense concrete to prevent deterioration in chloride sol ution3

. The greater penetration of chloride compared with that of sulphate ions was observed by Stratful4

. It was found that concrete retai ned 15 times more chloride than sulphate. The chloride binding capacity of silica fume in cement pastes has been characterized5

. Increase of silica fume from 0-30% dimini shed the chloride bind ing capaci ty and pH value. Addition of silica fume enhanced the resistance of cement-mortar to chemical attack of acid and sulphates6

.

Pozzolanic cement pastes were prepared from SRC as well as granul ated slag with 8% silica fume and immersed in seawater up to one year7

. Determining the free lime, total chloride and sulphate contents as well as compressive strength fo llowed the durability, of these pastes. The aggress ive attack was discussed and correlated with the type of cement. The silica fume and slag blended cement pastes give the lower values of total chl ori de than S RC pas tes and vice versa of sulphate contents.

Aggressive attack of sulphate ion is one of the factors responsible for damage to Portl and cement concrete8

. Sulphate ions can react with some constituents of cement pastes producing sulphoaluminate hydrates and gypsum, which cause expansion and cracking of concrete. The blast-furnace slag, e .g., supersulphate, Portland blas t-furnace slag or pozzolanic cements increase the chemical res istance to sulphate attack, and possess impermeability and lower heat of hydrat ion9

.

Resistance of silica fume blended cement pastes to sulphate attack was comparable to sulphate resisting cement 10

• This has been attributed to fine pore structure and reduction of lime content of silica fume concrete. The ability of silica fume to ho ld alkali s was found responsible for reducing alkali-silica expansion reaction .

The influence of silica fume up to 20% by weight of cement on the properties of blended cement pastes has been studied 11

• It has been found that the addition of silica fume to Portland cement enhances the hydration kinetics, i.e ., increases the combined water and decreases the liberated Ca(OHh contents. It was also found that 5% silica fume improves the mechanical properties of cement paste in tap water. This mix is the suitable mix to be sulphate-resisting cement (in MgS04 and Na2S04 solution). This work aims to study the chloride and sulphate attack on 85% SRC and 15% HAC composite. Also, the effect of substitution of HAC by silica fume on the attack up to one year was studied.

HELMY et al: EFFECT OF SILICA FUME SUBSTITION IN HAC ON THE CHLORIDE AND SULPHATE ATTACK 229

Experimental Procedure The materials used in this work were SRC,

supplied from Helwan Portland Cement Co., Egypt, HAC, secar 40% Ah03, imported from Lafarge Co., France, and silica fume from Egyptian Company for ferrosilicon alloys Edfo, Egypt. The chemical analysis and surface area of the materials are given in Table I.

Different mixes were made by substituting 5, 10 and 15 % of HAC by SF at 85% SRC. The mixes were denoted as Mo. M 1, M2, and M3• Each dry mix was homogenized for one hour in a porcelain ball mill using two balls to assure complete homogeneity. The mixing of cement pastes was carried out with the water of consistenc/ 2

• The amount of cement was placed on a smooth non-absorbent surface and a crater was formed in the center. The required amount of mixing water was poured into the crater of the cement. The cement was allowed to absorb the water for about one minute. The dry cement was trowled over the remaining mixture. The mixing was completed by continuous mixing for 3 min by gauging trowels. Immediately, after mixing the cement paste was put in cubic moulds 2x2x2 em and pressed until homogeneous specimen was obtained. The moulds were manually vibrated to remove any air bubbles, and then smoothed by spatula. The moulds were cured in a 100% R.H. at± 2°C for 24 h, then demoulded and cured under tap water for 28 days. After 28 days (0 time) curing in tap water, hardened samples were immersed in 4% MgCh solution as well 4% MgS04 solutions for 1, 3, 6, 9 and 12 months. The solution was renewed every month to keep the concentration of cr, or so4-2 nearly constant. The progress of the attack was determined through the measurements of compressive strength, total chloride, sulphate, free lime contents 13

, and total porosity 14. After the

compressive strength determination, a representative sample, of about 10 g was taken. The sample was ground in alumina mortar under the surface of the stopping solution (1:1 v) methanol/acetone, and then filtered through sintered glass funnel G4. Washing the contents of the funnel for two times with the

stopping solution and finally with 50 mL fresh diethyl ether and then the sample was dried at 70°C for one hour then kept in a airtight bottles 15

• The chloride contents and sulphate were determined as described elsewhere 16

. Some selected samples were examined by X-ray diffraction to identify the products of the cement pastes.

Results and Discussion Magnesium chloride solution

The free lime contents of cement pastes immersed under 4% MgCh solution up to one year are graphically represented in Fig. 1. The free lime contents of all cement pastes decrease with curing time. This is mainly due to the consumption of liberated Ca(OH)2 with SF forming CSH and the reaction of liberated Ca(OHh with MgC12 giving CaC12 and Mg(OHh. As the amount of SF increases the free lime contents decrease.

7

6

~ 5 a ., a 8 4

"' .§ .....1

"' 3 ., ~

2

0 3 6 9 12

Curing Time, months

Fig. I -Free lime contents of SRC pastes with various propor­tions of SF and HAC immersed in 4% MgCI2 solution up to one year

Table I -The chemical composition and surface area of starting materials, wt%

Oxides, Si02 Al20 3 Fe20 3 CaO MgO so3 LOI Na20 K20 Cl" Surface Area % cm2/g

SRC 20.88 3.697 4.937 64.57 1.987 1.621 2.27 0.0184 O.Q351 0.0212 3561

HAC 7.92 39.81 1.81 44.88 2.6 1 0.00 2.50 0.25 0.20 0.003 2300

S.F. 94.64 0.97 0.93 0.55 0.35 0.10 2.01 0.20 0.25 0.00 20 m2/g

230 INDIAN J. ENG. MATER. SCI., AUGUST 2001

The XRD patterns of 85% SRC and 15% SF cement pastes immersed in 4% MgCh solution up to 12 months are shown in Fig. 2. Sample hydrated for 28 days in tap water shows the presence of hydrated lime, calcite as well as anhydrous silicates and sulphoaluminate hydrates. At 3 months the hydrated lime is completely disappeared with the decrease of anhydrous calcium silicates and calcite. As the time increases up to one year there is only CSH as semi­crystalline phase. Therefore, SRC with 15% SF is durable in 4% MgCh solution. The calcium chloroaluminate is not detected in all hydrated samples up to 12 months. The calcium chloroaluminate may be amorphous which cannot be detected by XRD technique. This is due to that the SF absorbs the chloride ions.

The total chloride contents of cement pastes immersed under 4% MgC12 solution up to one year are plotted in Fig. 3. The total chloride content increases gradually with curing time for all cement pastes. This is due to the chemical reaction between MgCh and Ca(OHh to produce CaCh and Mg(OHh as well as with aluminate ferrite hydrate giving chloroaluminate. Therefore, the total chloride increases with curing time up to one year. On the

A : CA (Cnlcium aluminate) CJ : Calcium chloroaluminate T : Ettringite

40 35 30

M : Monosulphate CS : CSH If : Hydrated lime

25 20

20 (drgrrr)

12 months

0 month

'' 10

Fig. 2- XRD patterns of SRC pastes with 15% SF, immersed in 4% MgCI2 solution up to one year

4.5

4

3.5

* 3 ~

"' "' :9 0 2.5 e

-;;;

~ 2

1.5

(• M .. f-1 ~1 , .._ M1 -+- r\1 .• )

0 .5 4----,------'=;=======;====-1 0 3 6 9 12

Curing Time, months

Fig. 3 -Total chlorides of SRC pastes with various proportions of SF and HAC immersed in 4% MgCI2 solution up to one year

28~----------~;=~==;=~~ (• Mo f-1 M , .._ M1 -+- M3 )

26

24

* 22

-.£ "' 20 0

ct: -;;;

~ 18

16

14

12 0 3 6 9 12

Curing Time, months

Fig. 4- Total porosity of SRC pastes with SF and HAC im­mersed in 4% MgCI2 solution up to one year

HELMY eta!: EFFECT OF SILICA FUME SUBSTITION IN HAC ON THE CHLORIDE AND SULPHATE ATTACK 231

other side, as the amount of SF increases the chloride content decreases, i.e., the cement pastes without SF have higher values than those containing SF. The diffusion process of chloride ions in cement pastes is influenced by the cement composition and total porosity .

The total porosity of different cement pastes immersed under 4% MgCh solution up to one year are plotted in Fig. 4. The total porosity decreases with curing time for all cement pastes. This is due to the filling up of a part of the available pore volume with hydration products. The increase of SF content increases the total porosity, due to the increase of water of consistency. The total porosity = 0.99 We x dpll+Wt where, We is free water, W1 is total water and dP is bulk density. The silica fume has a high surface area ::::: 20 m2/g therefore it needs higher mjxing water. Therefore, the free water increases and the total porosity also increases 17

.

The compressive strength of cement pastes immersed under 4% MgCh solution up to one year are graphically plotted in Fig. 5. The compressive strength of cement pastes decreases with duration of immersion in MgCh solution. The chloride ions attack most of the constituent of cement pastes . The MgCh reacts with the liberated Ca(OHh to give CaC12 and

90.-----------------------------~

80

70

"' ~ 60 .<:: til c: 0

bl 50 0 ;>

·u; Vl 0 c.. 40 s 0 u

30

20

10 0 3 6 9 12

Curing Time , months

Fig. 5- Compressive strength of SRC pastes with SF and HAC immersed in 4% MgC1 2 solution up to one year

deposits Mg(OHh The low solubility of Mg (OHh drives the equilibrium to the right.

Ca(OH)2 + MgCh = CaCh + Mg(OH)2

The liberated CaCh acts as an accelerator at the beginning of hydration and by the increase of its amounts, it acts in the reverse direction. CaC12 reacts with C3A and C4AF giving chloroaluminate hydrate, which makes softening. Therefore, on prolonged immersing, the compressive strength decreases. Also, MgCh can decompose the CSH. These reactions tend to decrease the compressive strength with curing time up to one year. Obviously, with the increase of the amount of SF the compressive strength decreases due to the increase of water of consistency. It can be seen that superplasticizer must be added to the blended cements containing SF to decrease the water of consistency as well as the total porosity and increase the compressive strength.

The effect of different dosages of melamine formaldehyde sulphonate MFS admixture as 0.00, 0.25, 0.50, 0.75 and 1.0 wt% of cement on the water of consistency, apparent porosity and compressive strength was studied. The water of consistency was decreased from 33.2 to 30.0, 25.0, 21.5 and 17.50 at 0.0, 0.25, 0.50, 0.75 and 1.0 adrruxture respectively .

Fig. 6 shows the effect of the different dosages on

30 120 Total porosity Compressive strengt~

110

.+

25 100 (j 0 3

~ 90 -= 0 .., 6

II>

"' "' ·;;; ~· 0 II> I.. 20 80 '!: 0

i:l. .., ] II> .,. = 11Q 0 ..

70 ...... E-o + F'

• ~ ...., 15 60

:.:

.o • -e- D 0.25

-+-D 0.50 50 • -t- D 0.75

~n 1.00 o·

10 40

10 100

Curing Time, days

Fig. 6- Compressive strength and total porosity of SRC pastes with SF and HAC in presence of different dosages of MFS ad-mixture

232 INDIAN J. ENG. MATER. SCI., AUGUST 2001

the apparent porosity and compressive strength up to 90 days. It is clear that as the dosage of admixture in­creases due to the reduction of mixing water (water of consistency). Also, the compressive strength increases with the amount of admixture. This is also due to the decrease of mixing water. As the hydration proceeds, the total porosity decreases and the compressive strength increases due to the accumulation of hydration products in the open pores of the cement paste.

Magnesium sulphate solution The free lime contents of cement pastes immersed

under 4% MgS04 solution up to one year represented in Fig. 7. The free lime contents of cement pastes with SF increase at one month and then decrease up to one year. The initial increase of free lime up to 1M is mainly due to the continuous liberation of lime during the hydration, which is higher than its consumption . It can be said that the SF activates the hydration of belite at later ages. It acts as a nucleating agent for the hydration of cement paste. After one month the free lime content decreases due to the consumption with sulphates. On the other side, the free lime of cement pastes without SF decreases gradually with curing time up to one year. The rate of consumption of free lime in silica fume blended cements is higher than that of only HAC.

"1. 5

6.5

5.5

~ i 4.5 0 c: 0 u 0 8

;::1 3.5 .., " u::

2.5

1.5

'""'1\1, 0.5

0 3 6 9 12

C uring Time, months

Fig. 7- Free lime contents of SRC pastes with various propor­tions of SF and HAC immersed in 4% MgS04 solution up to one year

The XRD patterns of hydrated cement pastes of 85% SRC and 15% HAC substituted with 5, 10 and 15 % SF hydrated for 12 months in 4% MgS04

solution are seen in Fig. 8. The cement pastes without SF is completely corroded with the formation of gypsum and ettringite. The addition of 5% SF decreases the formation of gypsum with the increase of Ca(OH)2. As the SF increases at 10 and 15% the free lime decreases with complete consumption at I 5% SF. On the other side, the CSH increases with SF content due to its reaction with hydrated lime.

The XRD patterns of cement pastes of 85% SRC and 15% SF immersed in 4% MgS04 solution up to one year are seen in Fig. 9. Sample hydrated for 28 days in tap water illustrates the presence of Ca(OHh with CaC03 as well as CSH in addition to unhydrated 13-C2S and C3S. Sample immersed for 3 months shows the decrease of hydrated lime and unhydrated silicates up to 6 months. Sample immersed for 12 months illustrates the presence of gypsum, ettringite and CSH with calcite. The gypsum and ettringite increase with time. Therefore, this blended cement has no resistance against 4% MgS04 solution.

The total sulphate contents of cement pastes immersed under 4% MgS04 solution up to one year

·f l

45

G: Gypsum D : Belite N : Mg(OHh H : Hydrated lime CS : CSH L : C;p,Jcite T : Ettrin ite M : Monru:ulphate

40 JS

L cs •

L cs G

JO 25

29 (dr~-:n•t•)

!\'I a

20 1o 10

Fig. 8- XRD patterns of SRC pastes with SF and HAC im­mersed in 4% MgS04 solution up to one year

HELMY et al : EFFECT OF SILICA FUME SUBSTITION IN HAC ON THE CHLORIDE AND SULPHATE ATTACK 233

G : Gypsum H : Hydrated lime T: Ellringite

45 40 35

G cs

L

JO 25

12 month!

0 month

20 15 10

Fig. 9- XRD patterns of SRC pastes with 15% SF, immersed in 4% MgS04 solution up to one year

2.5 ,.-------- --------,

2

~ 1.5 --::

"' .., «i ..c c.. :; <ll

5 ~

0.5

0 3 6 9 12

Curi ng T ime, months

Fig. 10 - Total sulphates of SRC pastes with SF and HAC im­mersed in 4% MgS04 solution up to one year

33

30

27

~ 24 --::

·~ 0 21 0

(]..

-;;;

~ 18

15

12

9 0 6 9 12

Curing Time, months

Fig. II -Total porosity of SRC pastes with different percentage of SF and HAC immersed in 4% MgS04 solution up to one year

are plotted in Fig. I 0. The total sulphate increases with curing tim~ up to one year due to the reaction of sulphate ions with aluminate ferrite or Ca(OHh forming calcium sulphate and then ettringite phase 18

•

On the other hand, the total sulphate of silica fume blended cement pastes show lower values than those of free SF cement pastes. As SF increases, the total values decreases. The free Ca(OHh reacts with SF therefore, the total sulphate contents are smaller than those in free SF cement pastes. The increase of sulphate content in SRC pastes is mainly due to the increase of alite that liberates more Ca(OH h and reacts with MgS04 given gypsum.

The total porosity of cement pastes immersed under 4% MgS04 solution up to one year are plotted in Fig. II. The total porosity decreases with curing time for all cement pastes due to filling up to a part of pore volume with hydration products. The increase of SF the total porosity increases due to the increase of water of consistency. It is clear that the mai n factor, which affects the total porosity, is the mixing water and the type of hydrates.

The compressive strength of cement pastes immersed under 4% MgS04 solution up to one year are represented in Fig. 12. The compressive strength of all cement pastes decreases with curing time up one year. The sulphate ions attack most of the constituents of the cement pastes; the MgS04 reacts with the liberated Ca(OH)2 and deposits Mg(OHh. Also,

234 INDIAN J. ENG. MATER. SCI. , AUGUST 2001

95

85

75

~ 65 ..c: e:o c:: ~

"' 55 0

·~ Vl 0

0.. 45 s 0 u

35

25

15 0 3 6 9 12

Curing Time , months

Fig. 12- Compressive strength of SRC pastes with SF and HAC immersed in 4% MgS04 solu tion up to one year

MgS04 can decompose the CSH. These two reactions tend to decrease the compressive strength with curing time. Also, as the curing time proceeds the ettringite tends to grow and expand which causes cracking and deterioration of cement pastes. Therefore, the compressive stre.ngth decreases with curing time. Obviously, with the increase of SF the compressive strength decreases due to the mixing water.

It can be concluded that superplasticizer must be added to the cements containing SF to decrease the water of consistency as well as the total porosity and increase the compressive strength.

Conclusion The substitution of HAC with SF increases the

apparent porosity and decreases the compressive strength of cement pastes in MgCI2 and MgS04

solutions up to one year. This mainly due to the high mixing water of cement pastes containing silica fume. On the other side as the amount of silica fume increases, . the chloride content decreases, due to the high resistivity of silica fume towards the chloride ions. Also, the total sulphate decreases with the silica fume content.

References 1 Bensted J, World Cement Techno/ , 8 ( 1977) 171 . 2 Biezok I, Concrete Corrosion , Concrete Protections 8'h, Ed,

(Akadimal Kiado Budapest), 1972). 3 Monhan D, & Mehta P K, Cement Caner Res, 3 ( 198 1) 63. 4 Stratful R F, Mater Port, 3 ( 1964) 74. 5 Page C L & Vennsland D, Mater Constr, 16 (91) ( 1983) 19. 6 Alticne P C, Bedrad C, Plumate M, & Holded D, Proc Mater

Res Soc Symp, 42 (1985) 20!. 7 Ali A H, El-Didamony H & Mostafa K A. Sil lnd, 62 (1 1-12)

( 1997) 149. 8 Amer AA, Helmy I M, El-Hemaly S A S & El-Didamony H,

Indian J Techno/, 29 ( 1991) 379. 9 Gutt W H, Concrete (1977) 35.

10 Sorenson E V, A mer Caner Int. SP-79, 2 Vols ( 1983) 709. II Fu Y, He G, & Ding J, Gu Zhang-Zhao-Guisuanyan Tongba,

6 (4) (1987) 59. 12 ASTM standards, Standard Test Method for Normal Consis­

tency of Hydraulic Cement ASTM Designation; C 187-83 (1983) 195.

13 Kondo R, Abo-El-Enein S A, & Diamon M, Bul Chem Soc Jpn., 48 ( 1975) 222.

14 Copeland L E, JAm Caner lnst, (1952) 836. 15 El-Didamony H, Thermochim Acta, 35 (1980) 201. 16 Bri tish Standards Institution, Test for Chloride, Sulphate and

Sulphoaluminate, 1881 Part 6 ( 1971) 25. 17 El-Didamony H, Haggag M Y & Abo-El-Enein S A, Cement

Caner Res, 8 ( 1978) 351. 18 Tesreanu I & Georgscu M, Rev Roum Chim, 18 (1973) 623.