dose. This change dropped the freezing point to 4.0C
What is clear is that both mixes 8 and 9 performed
(24.8F) and the initial set to 6.1 hours.
well at 5C (23F). They gained strength more rap-
It was apparent from mixes 24 that more admix-
idly at that temperature than control mortar cured at
5C (41F). The 5C (41F) represents the lowest tem-
tures were needed to depress the freezing point further.
To do this, mixes 5 and 6 used the Type D, water-re-
perature at which concrete currently can be cured (ACI
ducing and retarding admixture at 6.25 times its maxi-
1988). Concrete (mortar) usually gains 30, 56, and 83
mum dose and the Type C accelerator at both full and
percent of its 28-day room temperature strength when
cured for 7, 14, and 28 days, respectively, at 5C (41F).
twice-normal dose. Though these mixes used admix-
tures at higher than normal, it was hoped that they would
With this as guidance, it can be seen that mixes 8 and 9
performed well down to 10C (14F), even though
provide an indication of how much admixture would
ultimately be needed. As Table 4 shows, the retarding
neither possessed a freezing point that low. It appears
that a mix must have a freezing point of at least 5C
effect of the Type D admixture when used with the Type
(23F) to perform well, as those with 4.0C (24.8F)
C accelerator at full dose caused mix 5 to set at 8.25
hours and to freeze at 3.1C (26.4F). Mix 6 counter-
freezing points did not do well at even 5C (23F).
acted the slow setting of mix 5 by doubling the quan-
Interestingly, both mixes, though they gained little
tity of the Type C admixture. As can be seen, mix 6 set
strength, recovered full strength after being cooled to
at 5.1 hours and froze at 4.0C (24.8F). The accel-
erator seemed to help but, obviously, more was needed.
To attain these higher concentrations while staying
within recommended levels, mix 7 used both the Type
C and E admixtures at 100 percent of their recom-
mended doses and combined them with the Type F
This study showed that admixtures used in today's
water reducer, also at 100 percent of its recommended
concrete could be combined to allow fresh concrete to
dose. This combination nearly produced the sought-
gain appreciable strength at below-freezing tempera-
after admixture as it set at 4.9 hours and froze at 4.9C
tures. Freezing points as low as 9.2ϒC (15.4F) were
(23.2F). In addition, Table 4 shows that the flow of
developed by combining a Type C, a Type E, and a
107 percent was a bit higher than that of the control.
Type F admixture along with calcium chloride, all
This suggested that the w/c ratio of this mix could be
within allowable doses. This combination of admix-
tures produced higher strength in mortar cured at 5ϒC
reduced, which would, in effect, cause the admixture
(23F) for 14 days than did control mortar cured for
concentration in the mix water to increase and the freez-
the same time at 20ϒC (68F). What is more, this ad-
ing point of the mortar to decrease. Mix 8 tried this
approach by reducing the w/c ratio to 0.30 while keep-
mixture combination produced significant strength in
mortar cured at 10ϒC (14F) and allowed mortar to
ing the admixtures the same as those in mix 7. Table 4
shows that this mix achieved a freezing point of 5.7C
recover full strength when thawed after being at 20ϒC
(21.7F). Unfortunately, it was too dry to be practical.
(2C). The problem with this admixture combination
Mix 9 combined the two commercial admixtures
was that it caused the mortar to set too rapidly.
with calcium chloride, each at their maximum allow-
Additional work should be done to find the
able dosage, with the Type F high-range water reducer.
combination(s) of off-the-shelf admixtures that permit
This produced an initially workable mix with a 9.2C
concreting to 10ϒC (14F) without setting too rapidly.
(15.4F) freezing point. However, it set up during the
Though all admixtures conforming to ASTM C 494
casting of strength specimens (within half an hour of
(ASTM 1997c) have individually passed a battery of
water hitting the cement) so that setting times could not
tests to prove that they cause no harm to the concrete,
be measured. Time did not allow additional testing.
little is known on how they react when used in combi-
nation with other admixtures or with admixtures from
other companies. Thus, admixtures in various combi-
The mixes shown in Table 4 were tested for strength
nations and from various sources should be studied for
gain at four temperatures. Table 5 shows those results.
their affects on concrete and embedded metal. For ex-
Although the ones of most interest are mixes 8 and 9
ample, preliminary data suggest that high doses of ad-
(because they had freezing points that were below the
mixtures can sometimes improve the resistance of con-
5C [23F] target set at the beginning of this study),
crete to cycles of freezing and thawing. It is not
all strength results are given. The results from mixes
recommended that the results of this study be applied
27 will not be discussed except to note that mixes 4
directly to field application until additional study and
and 5 achieved very low strengths, even at room tem-
testing are done.
perature. It is not clear why this happened.
Future work should focus not only on portland