Expansive cements - Special Inorganic Cements - page 304

• The expansion gets under way only after a critical amount of ettringite has been
produced. This amount increases with increasing initial water/cement ratio of the mix.
It has been postulated that the precipitation of ettringite at the surface of C 4 A 3 grains
causes an increase of the effective volume of these particles; an external expansion of
the paste gets under way when the expanding particles come into contact with and
exert pressure on each other (Bentur and Ish-Shalom, 1974, 1975; Ogawa and Roy,
1981, 1982).
• An uptake of water from the environment is not essential for the expansion to take
place, but it does increase the extent of expansion (Mather, 1973; Odler and Gasser,
1988).
It has also been suggested that swelling of colloidal ettringite formed at high Ca(OH) 2
concentrations, rather than the formation of ettringite itself, is the main cause of
expansion in expansive cements based on ettringite formation (Mehta 1973a, 1982; Mehta
and Hu, 1978). In support of this hypothesis Mehta and Wang (1982) demonstrated a
linear expansion of up to 4% in compacts of “colloidal” ettringite and C 3 A-free Portland
cement. However, from their measurement of expansion in compacts of ettringite, Aluno-
Rosetti et al. (1982) concluded that the swelling of ettringite alone could not account for
all the volume increase seen in expansive cements, and that a simultaneous development
of crystallization pressure due to an anisotropic growth of ettringite must also be
involved.
In conclusion, based on the available experimental evidence, it appears that a
topochemical formation of ettringite and its oriented growth are essential for the
expansion process to take place (Lafuma, 1929; Mather, 1973; Bentur and Ish-Shalom,
1974, 1975; Moldovan and Butucescu, 1980; Ogawa and Roy, 1981, 1982; Aluno-Rosetti
et al., 1982; Wang et al, 1985; Odler and Gasser, 1988; Herrick et al., 1992; Deng and
Tang, 1994). A subsequent swelling of the formed ettringite may also be involved, and
may enhance the extent of the overall expansion, but it does not seem to be essential for
the expansion to take place.
To produce expansion in expansive cements a variety of aluminous phases may be
considered, but calcium aluminate or sulfoaluminate phases are most commonly
employed. Owing to differences in their reactivities (Mehta, 1973b; Odler and Colán-
Subauste, 1999), not all of them perform equally well, to produce expansion without
cracking. Tetracalcium trialuminate sulfate (C 4 A 3 ) appears to be the most suitable, as
most of this phase hydrates in the desirable time range of several hours to several days:
that is, in a period in which the cement paste has already set but is not yet excessively
rigid.
Monocalcium aluminate (CA) hydrates rather rapidly initially, but still a significant
fraction of it reacts shortly after setting: that is, at a time that is favorable for the
expansive process. The reaction ceases within a few days.
U nlike these two phases, tricalcium aluminate (C 3 A) yields significant amounts of
ettringite in the early stages of hydration: that is, at a time when the cement paste is still
plastic, and when this reaction does not cause expansion. After setting, the formation of
ettringite continues at a slow rate for an extended period of time, and is not terminated
even after the concrete has lost its ability to expand without cracking.