Early Age Cracking in Cementitious Systems 47
3.6 DRYING SHRINKAGE CRACKING AT EARLY AGES
Jan Bisschop
Microlab, Faculty of Civil Engineering and Geosciences, Delft University of Technology, The
Netherlands
3.6.1 Introduction
This section deals with the driving forces and microstructural aspects of early ages shrinkage and
cracking due to external drying. First, the nano- and microscale mechanism of drying shrinkage
will be described. Although these nano- and microscale mechanisms are the driving forces for
drying shrinkage cracking, it is the type and degree of restraint that determines the type of drying
shrinkage cracking. The different types of internal and external restraints and the associated types
of cracking will be described in this section. An important aspect is the effect of sealed-curing
duration on drying shrinkage, because it prescribes the required wet-curing duration of concrete in
engineering practice. Finally, drying shrinkage of low w/b, high-strength concrete is compared
with drying shrinkage of ordinary concrete.
3.6.2 Mechanisms of Drying Shrinkage
Drying shrinkage is defined as the time-dependant deformation due to loss of water at constant
temperature and relative humidity [1]. Although, drying shrinkage has been extensively measured
on Ordinary Portland Cement (OPC) pastes and concretes, the basic nanoscale mechanisms of
drying shrinkage have so far not been precisely established. For OPCsystems four prominent
shrinkage mechanisms have been proposed: 1. surface free energy (Gibbs-Bangham shrinkage); 2.
capillary tension; 3. movement of interlayer water; and 4. disjoining pressure [1]. It is generally
believed that more than one mechanism may be involved in drying shrinkage depending on the
relative humidity. Hansen [1] investigated shrinkage mechanisms of OPC with w/c-ratios of 0.4
and 0.6 based on measurements of shrinkage, weight loss, and pore structure using nitrogen
sorption and mercury intrusion porosimetry (MIP). He found that the Gibbs-Bangham effect (i.e.
the surface free energy effect) is a major stress mechanism, which was active in the entire RH
range from 0-95%. Shrinkage due to the capillary tension was found to be active above 25% RH.
A large proportion of first drying shrinkage is irreversible and can therefore not be explained by
the reversible Gibbs-Bangham or capillary tension effects. Helmuth and Turk [2] and Pickett [3]
showed that irreversible shrinkage can be as large as 60% of first drying shrinkage following
long-term drying. Hansen [1] concluded that true Gibbs-Bangham shrinkage could only account
for 33% of the first drying shrinkage. The remaining 67% of first drying shrinkage might have
48 RILEM TC 181-EAS: Final Report - July 2002
been irreversible shrinkage due to a decrease in interlayer spacing caused by Gibbs-Bangham and
capillary induced stresses.
As indicated by Jennings and Xi [4] drying shrinkage of Portland cement paste and concrete is
controlled by a number of mechanisms that operate in various parts of the microstructure and at
various length scales. Besides the aforementioned nanoscale mechanisms of drying shrinkage,
they describe shrinkage mechanism on a micrometer and millimetre level. It is not only the
different hydration (C-S-H) products (i.e. inner and outer products) that have their own shrinkage
characteristics [4], but also differences in shrinkage characteristics in the bulk paste and interfacial
paste can be expected, because of differences in porosity [5]. Therefore, the distribution, sizes, and
amount of cement particles and aggregates are important factors controlling shrinkage in these
length scales.
3.6.3 Drying Shrinkage Cracking due to Internal Restraints
Although drying shrinkage is the driving force for drying shrinkage cracking, it is the type and
degree of restraint against shrinkage that determines the type of drying shrinkage cracking. Thus,
the mechanism of drying shrinkage cracking depends upon the type of restraint that caused the
cracking. A distinction can be made between cracking due to internal restraints and cracking due
to external restraints. Cracking due to external restraints will be dealt with in the next paragraph.
In this paragraph three types of cracking due to internal restraints will be described.
(a)
(b)
Figure 3.6.1: Drying shrinkage microcracks due to self-restraining in cement paste specimen of
40x40x160 mm3 at 20% drying. (a.) Cell-like microcracking on drying surface. (b.) Four
combined longitudinal cross-sections. The deeper microcracks in (b.) are also the wider
microcracks (~30-40 µm) on the surface. The smaller microcracks have a width < 20 µm, [9].
The first type of cracking due to internal restraint is called self-restraining. It is the result of non-
uniform shrinkage caused by the moisture gradient that develops during drying. Self-restraining
Early Age Cracking in Cementitious Systems 49
leads to tensile stresses near the drying surface and compressive stresses deeper into the specimen
[6,7]. When the tensile stresses near the drying surface reach the tensile strength of the material,
cracks appear. Self-restraining tensile stresses near the drying surface are parallel to the drying
surface, and therefore cracks have an orientation perpendicular to the surface [8,9]. On the drying
surface well-known cell-like crack-patterns can be observed as shown in Figure 3.6.1 [9,10]. The
scale of cracking (crack length and crack width) due to self-restraining probably depends on the
specimen size and the degree of drying [6,11]. Bisschop and Van Mier [9] observed two or three
crack sizes due to self-restraining in a 40x40x160 mm3 prism of plain cement paste at a degree of
drying of 20% using light microscopy (see Figure 3.6.1). These crack types may represent the
primary, secondary and tertiary cracks formed when drying progresses as predicted by shrinkage
cracks models described by Bazant [6] and Granger et al. [11]. As calculated by Bazant [6] the
crack width remains too low in normal sized concrete for low degrees of drying to be visible with
the unaided eye (i.e. 0.2 mm). Colina and Acker [10] used specimen dimensions of >1 m3 to make
self-restraining cracks in micro-concrete visible to the unaided eye and determined the geometrical
characteristics of the crack network on the drying surface. In addition to specimen size
dependence of the crack mechanism, there might also be a specimen shape dependence. In thin
concrete slabs self-restraining results in curling of the slab [12], and possibly this curling may
diminish drying cracking compared to specimens having dimension not allowing curling.
The second type of internal restraint is caused by the presence of stiff aggregates. When the
restraint stresses reach the tensile strength of the material, microcracks in the paste matrix will
form with an orientation perpendicular to the aggregate perimeter [13]. Bond cracks may also be
formed due to shearing of the paste along the aggregate surface. However, bond cracks are in
general closed (and thus invisible), because of the radial compressive stresses caused by this type
of restraint [9,13]. The importance of this type of internal restraint in causing microcracking in
shrinking concrete (for both drying and autogenous shrinkage) remains unclear. It has been shown
by Bisschop and Van Mier [9] that this type of microcracking occurs in a cement based composite
with 35% (by volume) of mono-sized aggregate particles of 6 mm.
The third type of internal restraint is caused by the steel bars in reinforced concrete. Ohno et al.
[14] showed that autogenous shrinkage of the high-strength concrete alone was not large enough
to cause cracking due to D25 restraining bars. On the other hand, at external drying of 60% R.H.
cracking occurred in this concrete for 1-day and 3-day sealed cured specimens after an average of
10 days and 19 days drying, respectively. Only some of the 7-days sealed cured specimens
cracked due to external drying.
3.6.4 Drying Shrinkage Cracking due to External Restraints
In addition to cracking due to the internal restraints described above, external restraints contribute to
cracking in drying concretes. In practice stresses occur if shrinkage is prevented by the surrounding
structure. For example, thin concrete slabs or concrete repair layers may crack due to external restraint
caused by bond with the subgrade [15]. Since internal restraints (especially self-restraining) are always
present in concrete, drying concrete will never show cracking due to external restraints only. Concrete
which dries in a structure which is externally restrained or in a restrained shrinkage test will always
show cracking due to superimposed stresses of both external and internal restraints. This is an
important issue to consider, when using restrained shrinkage test to determine the crack sensitivity of
concrete. In the section ‘Early age cracking tests’ (section 6.6) the different categories of restrained
shrinkage test are described for evaluation of early age cracking. To understand cracking behaviour of
50 RILEM TC 181-EAS: Final Report - July 2002
concretes in these restraint cracking tests it is important to determine the relative importance of both
external restraint (i.e. set-up) effects and self-restraining (i.e. material) effects in inducing cracks.
For example in the ring test (see paragraph 6.6.2) the relative importance (or timing) of self-restraining
and external restraining in causing cracking is determined by the dimensions of the ring test set-up (i.e.
the ri/re ratio). When the ring specimen starts to dry, a moisture and shrinkage gradient develops over
the cross-section of the specimen. Cracking in the specimen is controlled by self-restraining (and thus
invisible to the unaided eye) until the ‘bottom’ concrete connected to the steel inner ring starts to
shrink. From this point onwards, the external restraint of the inner steel ring contributes tothe formation
of tensile stresses over the cross-section. For thicker ring specimens (i.e. lower ri/re ratio) it takes a
longer time before external restraining stresses become effective. The reason for this delay is the longer
time it takes the ‘bottom’ concrete to start to shrink. This might also explain why it took longer for
through-cracking to be observed in thicker ring specimens as was experimentally shown by Weiss and
Shah [15].
3.6.5 Effect of Age on Drying Shrinkage and Cracking
Drying shrinkage may contribute to early-age cracking in engineering practice if wet or sealed
curing is not applied for a sufficient long time. To make predictions about this source of early-age
cracking it is important to understand the effect of age (i.e. the duration of wet-curing or sealing)
on drying shrinkage and cracking.
At early ages there is fast reduction of the free water content due to ongoing hydration. Therefore,
less water can be lost from the concrete at later ages (i.e. with longer sealed curing duration) by
external drying as has been shown in [16-18]. Therrien et al. [18] measured the ultimate moisture
loss as a function of age, as well as the ultimate moisture loss as a function of relative humidity
(see Figure 3.6.2). It can be seen that at low relative humidity (i.e. 53%), the moisture loss
decreased with the initial curing duration. This can be explained by Kelvin’s Law, according to
which the lower the relative humidity, the lower the pore size that is emptied. Hence, materials
cured for 28 days, containing more small pores than those cured for only one day, will retain more
water at low relative humidity. Conversely, at high relative humidity, the moisture content
increases with a diminution in curing duration because there are more large pores.
Early Age Cracking in Cementitious Systems 51
Figure 3.6.2. Specimen
weight loss as a function
of relative humidity
[18].
Figure 3.6.3: Drying shrinkage as a function of relative humidity [18].
The loss of water from the concrete results in shrinkage of the paste matrix. Although there is a
clear trend of the moisture loss as a function of curing time, the resulting shrinkage does not
always show a similar trend. Miyazawa and Tazawa [19] showed that for concretes with 0.4, 0.5,
and 0.6 water-cement ratios, the age of drying did not have a significant effect on drying
shrinkage at later ages. The same results were obtained by Ueda et al. [16] for a 0.6 w/c-ratio
concrete within the first 100 days of drying: no clear effect of age. On the other hand, the high-
strength concretes with w/c-ratio of 0.2. and 0.3 studied by Miyazawa and Tazawa [19] showed a
lower drying shrinkage for longer sealed curing duration. This was also shown by Ohno et al. [14]
for a high-strength concrete with a w/c-ratio of 0.24. It should be mentioned that in these tests of
high-strength concretes, autogenous shrinkage was included in the measured shrinkage. Therrien
et al. [18] found for their 0.4 w/c-ratio mortar that ultimate drying shrinkage decreased for longer
sealed curing duration (Figure 3.6.3). This effect appeared to become more pronounced as the
relative humidity decreased. The ratio between long-term shrinkage values of 1-day and 28-day
52 RILEM TC 181-EAS: Final Report - July 2002
cured specimens at 53% RH is in the order of 1.5. Finally, Neubauer et al. [20] measured
shrinkage on very small cement paste samples (with w/c-ratio of 0.5) in an Environmental SEM
and found that drying shrinkage decreased with the age of the sample. They found that at
intermediate humidities (40-80%) stresses associated with capillary tension, which was the
predominant mechanisms of drying shrinkage at intermediate relative humidities, had a decreasing
effect on drying shrinkage of the older mortars.
As extensively described in section 3.3 ‘Evolution of solid behaviour’, it is not only the magnitude
of shrinkage which determines the risk of early-age shrinkage cracking. Important factors are also
the degree of restraint, the development of the mechanical properties and the relations between
them and the rate of stress relaxation with respect to the rate of stress built-up. Therefore, it is
difficult to conclude that higher drying shrinkage for younger cement-based materials should also
mean a higher crack sensitivity. For example, Ueda et al. [16] found no obvious effect of the age
at the start of drying on the cracking time in a restrained shrinkage test. Also, Bisschop et al. [17]
found for three different types of unreinforced high-strength cement-based material no effect of
sealed curing duration on degree of drying shrinkage microcracking due to internal restraints. The
lower moisture loss, the higher strength and stiffness, and the lower relaxation for the material
sealed cured for 7-days compared to the 1-day sealed cured material did not have a net effect on
the degree of drying shrinkage microcracking.
3.6.6 Drying Shrinkage of High-Strength Concrete
In early-age high-strength concretes self-desiccation and external drying occur simultaneously if
no sealing is applied. Especially in the surface zone of the concrete, this leads to a competition
taking place between water evaporation and water consumption by the chemical reaction (see
section [Link] ‘Combined effect of hydration and external drying on moisture distribution’). The
observed total shrinkage strain is not a simple sum of autogenous and drying shrinkage strains in
this case [21]. In this paragraph drying shrinkage of high-strength concrete is compared to drying
shrinkage of ordinary concrete. High-strength concretes differ from ordinary concretes by having
a lower w/c-ratio and by the addition of SF and superplasticizers. The effects of these three factors
on drying shrinkage are described in this paragraph.
Tazawa et al. [22] showed data of the relationship between the w/c-ratio and three kinds of
shrinkage: drying shrinkage, autogenous shrinkage and total shrinkage (see Figure 3.6.4). Drying
shrinkage was obtained by subtracting autogenous shrinkage from the total shrinkage. Although,
this might be not completely justified due to the aforementioned competition taking place between
water evaporation and hydration, the general statement can probably be made on basis of these
results that drying shrinkage decreases with reduction in w/c ratio. Moreover, Baroghel-Bouny
and Godin [21] showed that also the ‘pure’ drying shrinkage of high-strength concretes (low w/c-
ratio) measured within an uncoupled framework is lower than drying shrinkage of ordinary
concretes.
Early Age Cracking in Cementitious Systems 53
Figure 3.6.4: Relationship between w/c-ratio and shrinkage of concrete at 60% R.H. [22].
The effect of SF on drying shrinkage of concrete is still not clear, because conflicting results have
been presented by researchers. This might have to do with the different curing period and methods
that have been used to study this effect [23]. Al-Sugair [23] studied the effect of the quantity of SF
in concrete on their drying shrinkage behaviour (with unknown contribution of autogenous
shrinkage). It was found that in general drying shrinkage strain increased as more SF was added to
the concrete mixes. Two competing mechanisms were proposed to explain this trend. On the one
hand SF reduced the permeability and the pore size in concrete, and therefore lowers the rate of
drying. On the other hand, this densification effect of SF causes a reduction of the size of capillary
pores, and this should lead to an increase in capillary tension. From the results it was concluded
that is was generally this latter mechanisms that increased drying shrinkage as a function of the SF
contnet. Also Bloom and Bentur [24], and Wiegrink et al. [25] found increased free shrinkage
(autogenous and drying shrinkage) for concretes with SF additions. Haque [26] found on the other
hand that both 5% and 10% addition of condensed SFresulted in a substantial reduction of drying
shrinkage of the concretes. These lower values of drying shrinkage for these concretes was
attributed to the slower drying caused by the densification effect of SF.
Brooks [27] investigated the effect of admixtures (i.e. plasticizer and superplasticizers) on long-term
deformation (i.e. creep and free shrinkage) on the basis of data of creep and shrinkage extracted from
literature. He concluded that for the addition of five type of admixtures, there is a general increase in
both, creep and shrinkage, between 3% and 130%, but limited data and variability indicate no
conclusive differences between types of admixture or between types of deformation. Bloom and
Bentur [24] also studied the effect of superplasticizer content on drying shrinkage and found that the
free shrinkage (drying and autogenous shrinkage) was higher in the higher superplasticizer content
concrete. On the other hand Wiegrink et al. [25] found a slightly lower free shrinkage for a concrete
containing superplasticizer compared to the control concrete without plasticizers.
54 RILEM TC 181-EAS: Final Report - July 2002
3.6.7 References
1. Hansen, W., ‘Drying shrinkage mechanisms in Portland cement paste’, J. Am. Ceram. Soc. 70 (5) (1987)
323-328.
2. Helmuth, R.A. and Turk, D.H., ‘The reversible and irreversible drying shrinkage of hardened Portland
cement and tricalcium silicate paste’, J. PCA Res. Dev. Lab. 9 (2) (1967) 8-21.
3. Pickett, G., ‘Effect of aggregate on shrinkage of concrete and a hypothesis concerning shrinkage’, J. Am.
Con. Inst. 52 (5) (1956) 581-590.
4. Jennings, H.M. and Xi, Y., ‘Microstructurally based mechanisms for modelling shrinkage of cement paste
at multiple levels’, In Proceedings of 5th Int. Sym. on creep and shrinkage of concrete, RILEM, Barcelona
(1993) 85-102
5. Koenders, E.A.B., ‘Simulation of volume changes in hardening cement-based materials’, PhD-thesis, Delft
University of Technology (1997).
6. Bazant, Z.P. and Raftshol, J., ‘Effect of cracking in drying and shrinkage specimens’, Cem. Con. Res, 12
(1982) 209-226.
7. Wittmann, F. H. and Roelfstra, P.E., ‘Total deformation of loaded drying concrete’, Cem. Con. Res, 10
(1980) 601-610.
8. Hwang, C.L. and Young, J.F., ‘Drying shrinkage of Portland cement pastes. I. Microscracking during
drying’, Cem. Con. Res., 14 (1984) 585-594.
9. Bisschop, J and Van Mier, J.G.M., ‘The effect of aggregates on drying shrinkage microcracking in cement-
based materials’, UEF conference Advances in Concrete and Cement, Mt-Tremblant, Canada, August,
(2000).
10. Colina, H. and Acker, P., ‘Drying cracks: Kinematics and scale laws’, Mat. Struct. 33 (2000) 101-107.
11. Granger, L., Torrenti, J.M., Acker, P., ‘Thougths about drying shrinkage: Experimental results and
quantification of structural drying creep’, Mat. Struct., 30 (1997) 588-598.
12. Berke, N.S., Malone, J.J., Yang, W., ‘Relationship of drying shrinkage to cracking for concrete slabs with
and without shrinkage reducing admixtures’, in ‘Early Age Cracking in Cementitious Systems’, Proc.
International RILEM Conference, [Link] and [Link], editors, Haifa, March (2001) 159-168.
13. Goltermann, P., ‘Mechanical predictions of concrete deterioration - Part 2: classification of crack patterns’,
ACI Mat. J. 92 (1) (1995) 58-63.
14. Ohno, Y., Hayashida, T., Nakagawa, T., ‘Restraint stress due to autogenous shrinkage and drying
shrinkage in high strength concrete’, in International Workshop on Control of Cracking in Early Age
Concrete’, Tohoku University, Sendai, August, Japan (2000) 191-198.
15. Weiss, W.J. and Shah, S.P., ‘Restrained shrinkage cracking: the role of shrinkage reducing admixtures and
specimen geometry’ in ‘Early Age Cracking in Cementitious Systems’, Proc. International RILEM
Conference, [Link] and [Link], editors, Haifa, March (2001) 145-158.
16. Ueda, K., Sato, Y., Kiyohara, S., Nagamatsu, H., ‘Drying shrinkage cracking of young concrete’,
International Workshop on Control of Cracking in Early Age Concrete’, Tohoku University, Sendai,
August, Japan (2000) 245-256.
17. Bisschop, J. and Van Mier, J.G.M., ‘Shrinkage microcracking in cement-based materials with low-water-
cement ratio’ in ‘Early Age Cracking in Cementitious Systems’, Proc. International RILEM
Conference, [Link] and [Link], editors, Haifa, March (2001) 79-88.
18. Therrien J., Bissonnette, B., Cloutier, A., ‘Early-age evolution on the mass transfer properties in mortar
and its influence upon ultimate shrinkage’ in proceedings of Int. RILEM workshop on ‘Shrinkage of
Concrete - Shrinkage 2000’, [Link]-Bouny and [Link], editors, Paris, France, (RILEM 2000) 247-
270.
19. Miyazawa, S. and Tazawa, E., ‘Influence of specimen size and relative humidity on shrinkage of high-
strength concrete’, in proceedings of Int. RILEM workshop on ‘Shrinkage of Concrete - Shrinkage 2000’,
[Link]-Bouny and [Link], editors, Paris, France, (RILEM 2000) 233-246.
20. Neubauer, C.M. et al., ‘Drying shrinkage of cement paste as measured in environmental scanning electron
microscope and comparison with microstructural models’, J. Mat. Sc. 32 (1997) 6415-6427.
Early Age Cracking in Cementitious Systems 55
21. Baroghel-Bouny, V. and Godin, J., ‘Experimental study on drying shrinkage of ordinary and high-
performance cementitious materials’ in proceedings of Int. RILEM workshop on ‘Shrinkage of Concrete -
Shrinkage 2000’, [Link]-Bouny and [Link], editors, Paris, France, (RILEM 2000) 215-232.
22. Tazawa, E., Sato, R., Sakai, E. and Miyazawa, S., ‘Work of JCI committee on autogenous shrinkage’ in
proceedings of Int. RILEM workshop on ‘Shrinkage of Concrete - Shrinkage 2000’, [Link]-Bouny
and [Link], editors, Paris, France, (RILEM 2000) 221-42
23. Al-sugair, F.H., ‘Analysis of time-dependent volume reduction of concrete containing silica fume’, Mag.
Con Res. 47 (170) (1995) 77-81.
24. Bloom, R. and Bentur, A., ‘Free and restrained shrinkage of normal high-strength concretes’, ACI Mat. J.,
92 (2) (1995) 211-217.
25. Wiegrink, K., Marikunte, S. Shah, S.P., ‘Shrinkage cracking of high-strength concrete’, ACI Mat. J., 93 (5)
(1996) 409-415.
26. Haque, M.N., ‘Strength development and drying shrinkage of high-strength concretes’, Cem. Con. Comp.,
18 (1996) 333-342.
27. Brooks, J.J., ‘Influence of mix proportions, plasticizers and superplasticizers on creep and drying shrinkage
of concrete’, Mag. Con. Res., 41 (148) (1989) 145-153.
