By Matthew D’Ambrosia1 and Nathaniel Mohler2
Early-age cracking can be a significant problem in concrete. Volume
changes in concrete will drive tensile stress development when they
are restrained. Cracks can develop when the tensile stress exceeds
the tensile strength, which is generally only 10% of the compressive
strength. At early ages, this strength is still developing while
stresses are generated by volume changes. Controlling the variables
that affect volume change can minimize high stresses and cracking.
Mechanisms of Early-Age Volume Change
The volume of concrete begins to change shortly after it is cast.
Early volume changes, within 24 hours, can influence tensile stress
and crack formation in hardened concrete.
Chemical shrinkage occurs due to the reduction in absolute volume
of solids and liquids in the hydrating paste. Chemical shrinkage
continues to occur as long as cement hydrates. After initial set,
the paste resists deformation, causing the formation of voids in
Autogenous shrinkage is the dimensional change of cement paste,
mortar, or concrete caused by chemical shrinkage (Figure 1). When
internal relative humidity is reduced below a given threshold (i.e.,
extra water is not available), self-desiccation of the paste occurs,
resulting in a uniform reduction of volume.
|Figure 1 – Chemical shrinkage and autogenous
shrinkage volume changes of fresh concrete. Not to scale.
Creep is the time-dependent deformation of concrete under sustained
load. During early age, concrete creep is generally as much as 3-5
times higher than for mature concrete. Early load application due
to construction forces or prestressing operations can therefore
have a significant impact on total deformation. Furthermore, the
magnitude of creep in tension is greater than in compression, and
early tensile creep can be relied upon as a stress relaxation mechanism.
Creep is influenced by drying or self-dessication at early age,
and this synergy is often referred to as the Pickett Effect, after
Gerald Pickett, a PCA researcher who discovered the phenomena in
the 1940s (Pickett 1947).
Concrete, mortar, and cement paste will sometimes swell when sealed
or in the presence of external water. Swelling is generally caused
by pore pressure, but can be accentuated by the formation of some
expansive hydration products. The swelling is not significant, between
50-100 millionths at early ages; therefore, we will not be discussing
As cement hydrates, the reaction provides a significant amount of
heat. In large elements, this heat is trapped and can induce significant
expansion. When thermal changes are superimposed upon autogenous
shrinkage at early age, cracking can occur. In particular, differential
thermal stress can occur due to rapid cooling of massive concrete
Testing of Early-Age Volume Changes
Chemical shrinkage test
Volume change due to chemical shrinkage can be estimated from the
hydrated cement phases and their crystal densities or it can be
determined by physical test. The physical test places a measured
amount of lime-saturated water in an open pipet over a known amount
of cement paste inside a closed container. The change in water level
within the pipet indicates the change in volume due to chemical
shrinkage (Figure 2).
|Figure 2 – Test for chemical shrinkage
of cement paste showing flask for cement paste and pipet for
absorbed water measurement.
ASTM C157 - Modified for early age shrinkage
The standard drying shrinkage test for concrete can be modified
to capture early age volume change by elimination of the curing
period (usually 7-28 days) and beginning measurements as early as
possible. Prisms may also be sealed after casting to provide an
estimate of autogenous shrinkage. Surfaces should be sealed as quickly
as possible to eliminate loss of moisture. It should be emphasized
that autogenous shrinkage depends on temperature history (maturity)
and will be different in a laboratory prism when compared to a larger
concrete member in service under variable ambient temperatures.
As with drying shrinkage measurements, the test result will not
represent the actual shrinkage in the structure.
ASTM C1581 – Restrained Ring Shrinkage
The restrained ring shrinkage test consists of a concrete ring specimen
150 mm (6 in.) tall, 13 mm (0.5 in.) thick and 330 mm (13 in.) diameter
that is cast surrounding an instrumented steel ring (Figure 3).
The steel ring prevents the concrete from shrinking from the time
that the concrete is first cast. Shrinkage stresses continue to
grow as the concrete passes from initial set to final set and beyond.
Tensile creep relaxation alleviates stress development and is considered
beneficial at early age.The instrumented ring uses strain sensors
to monitor the development of stress. If the shrinkage in the concrete
is significant, the stresses will eventually cause cracking. The
strain sensors provide an indicator of the cracking time, which
is used to compare the cracking tendency between different concrete
|Figure 3 – Restrained ring
shrinkage test setup. (Courtesy of CTLGroup)
ASTM C512 - Compressive Creep
The standard creep test consists of a frame and hydraulic loading
system to apply constant stress to 150X300 mm (6x12 in.) cylindrical
specimens (Figure 4). Deformation is monitored periodically over
time and compard to compansion unloaded specimens to obtain the
creep strain of the concrete, which can then be used to calculate
the creep compliance, or “specific creep” of the material.
Tests are typically started at 7 or 28 days of age, but this test
can be modified for early age by starting the test as early as 24
hours. Sealed tests are used to evaluate “basic” creep
and unsealed tests incorporate the Pickett Effect, or “drying”
|Figure 4 – Standard creep test frames.
(Courtesy of CTLGroup)
Mitigating Early-Age Cracking
Optimization of aggregates to reduce total cementitious
Since volume changes are more a function of the cement paste,
rather than the more volume-stable aggregates, reducing the overall
cementitious content is the best way to mitigate early-age volume
changes. Typical concrete mixtures have gap-graded aggregates
that leave significant void space for cement paste to fill. By
optimizing the aggregate gradation across the entire spectrum,
as opposed to the coarse and fine aggregates individually, the
amount of paste required to surround each aggregate particle and
fill the void space is minimized (Figure 5); thereby minimizing
the effects of early-age volume change of the paste.
|Figure 5 – A comparison of void space
with different aggregate gradations.
Minimum w/cm ratio
Autogenous shrinkage increases with a decrease in water to cementitious
materials ratio (w/cm). Concrete mixtures with a w/cm of 0.30 can
experience autogenous shrinkage upwards of half of the normal drying
shrinkage. Using the highest w/cm that still provides adequate strength
and durability can reduce the impact of autogenous shrinkage.
Internal curing is a method by which water is encapsulated within
a concrete mixture for continued release during the hydration process.
Typical internal curing materials include high absorption lightweight
aggregate particles and super-absorbent polymers. The self-dessication
of the paste draws the water out of these particles to continue
the hydration of the cement particles. This is particularly helpful
in mitigating autogenous shrinkage of concrete mixtures with very
low w/cm (0.30 or less).
Shrinkage-reducing admixtures (SRAs) are typically used as mitigation
of cracking and curling caused by drying shrinkage; however, SRAs
can be utilized to mitigate autogenous shrinkage as well. The SRA,
typically propylene glycol or polyoxyalkylene alkyl ether based,
alters the surface tension of the pore water and reduces the stresses
developed during desiccation, whether self-induced or by evaporation.
Several concreting procedures can be used to minimize early-age
volume changes. When autogenous shrinkage is a concern, the use
of moist curing methods will help mitigate self-desiccation near
the concrete surface. The use of a well-developed thermal control
plan will mitigate the effects of thermal-based volume changes.
Kosmatka, Steven H.; Wilson, Michelle L.;
and Control of Concrete Mixtures, EB001.15, 15th edition,
Portland Cement Association, Skokie, Illinois, USA, 2011.
Pickett, Gerald, The
Effect of Change in Moisture Content on the Creep of Concrete Under
a Sustained Load, Research Department Bulletin RX020, Portland
Cement Association, 1947.
2Concrete Engineer, PCA