Ageing and Durability of Concrete
Gilles Pijaudier-Cabot, Université de Pau et des Pays de l’Adour – E2S UPPA & Institut Universitaire de France
Ageing of a material can be viewed as the evolution of its properties with time. For most construction materials such as mortar, concrete or plaster at early ages typically, mechanical properties progress with time. The material hardens, it is “getting stronger”. Simultaneously, these materials exhibit shrinkage. It means that during the first days after they have been manufactured, their volume reduces due to the chemical reactions that occur during hardening. In most instances, shrinkage is restrained: a simple illustration is the case of a layer of mortar placed on a brick: the mortar will shrink with time, but not the brick itself. A consequence is that cracking occurs. This time, ageing is detrimental to the mechanical properties of the material.
This example demonstrates that ageing should not be considered with respect to the material properties at stake, but rather with respect to the expected functionalities of the material (resisting to a load, preventing leakage,…). This is the reason why in many technological applications, including construction and infrastructures made of concrete that are the focus of this essay, engineers prefer to speak about durability. Hereby, the underlying issues are “how long the material will last?” and “how long the structural system made of this material will fulfil its duties?” The answer to this question defines the lifetime of the structural system for which ageing is per se detrimental.
Concrete: a complex material used worldwide
Concrete is one of the most widely construction material used on the planet. The volume of concrete cast all over the world per year is above 6 billion of cubic meters (0.85 m3 per inhabitant of the planet!). At the same time and when considering the US solely and year 2016, the cost of maintenance of infrastructures made of concrete was over twice the health expenditures due cancer and cardiovascular diseases altogether1 It is therefore obvious that the economic challenges of ageing of concrete are tremendous.
Concrete is a mix of aggregate, sand, water and cement. It is an “artificial stone” used for its mechanical properties: the compressive strength of classical concrete is in the range of 30-50 Mpa and the tensile strength is one tenth of its compressive strength typically2 . This weakness in tension (compared to compression) explains why concrete is always reinforced: steel bars placed in the moulds before casting are meant to take over the tensile stresses while concrete sustains compressive stresses.
To provide some perspective, the compression strength of concrete is at least ten times less than that of alloys, and the tensile strength is a hundred times less than that of alloys. The reason why this material is so popular is that it is much easier to manufacture a concrete beam than a steel beam. The first one is cast, often on site, and does not require some industrial infrastructure, whereas the second requires some heavy industrial processes such as heating and rolling.
Concrete is a lively material: the binder between the aggregate pieces is the result of hydration of Portland cement; it is mostly made of calcium silicate hydrates (CSH) and calcium hydroxide. Hydration starts upon mixing and casting concrete and may last several months, and possibly years, after casting. The hydrated cement paste has a multiscale porous structure, with pores ranging from the nanometer to the micrometer scale. It is through this connected network of pores that the material interacts with its environment, mostly with water, ages and exhibits time-dependant properties.
Leaving aside the corrosion of steel bars in reinforced concrete for which the reader can find details, e.g. in the book by François et al.3, we are going to illustrate ageing of concrete with two typical situations: in the first one ageing occurs because of applied loads, while in the second, it is simply the result of the long-term interaction of concrete with its environment.
Creep of concrete
Creep is defined as the deformation of a material with time under sustained loads. Creep of concrete has been the subject of a large number of studies, summarised e.g. in Bazant and Jirasek4. It stands among the potential sources of failure of concrete structures, a typical example being the collapse of the Palau bridge in 1996 in the Pacific islands (Fig. 1). This bridge was exhibiting excessive long-term deformations due to creep and was retrofitted. Six months after, it failed suddenly. The bridge was in fact poorly repaired and suffered from too many creep deformations5.

Figure 1: The Palau bridge before and after failure which occurred on September 26th, 1996 (taken from Burgoyne and Scantlebury6, credit: W.E. Perryclear).
The mechanisms that govern creep of concrete have been the subject of intense debates over the years, hosted for instance within the International Association for Concrete Creep, Shrinkage and Durability (IA – ConCreep7 ). There is today a consensus that creep of concrete is due (1) to the motion of water in the porous network of the material, mostly at the submicrometric scale, and (2) to the relative motion of the nanoplates of CSH, this time at the nano-scale. The first mechanism implies that creep is sensitive to environmental conditions. Depending on the humidity and water content in the porous structure of hydrated cement phase of concrete, the kinetics of creep deformations will change: the higher the water content, the higher delayed deformations. The second mechanisms sits at a lower scale: water serves as a lubricant and promote relative sliding of the nanoplates of CSH. Arguments at the atomistic scale show that the long-term creep deformation should be a function of the logarithm of time8. Therefore, creep deformations should be unbounded, although it is quite difficult to prove experimentally even with experiments carried out over a decade. Consequently, failure due to accumulated creep deformations is to be expected in any practical application, the problem being to keep those deformations small enough over the lifetime of the structure.
In the field of construction, it is little known that structures have indeed a lifetime. To the user, it looks like that reinforced concrete structures are built to last “forever” but this is not correct; for instance, bridges are supposed to last between 50 and 100 years. The lifetime is obtained from a probabilistic definition of the safety coefficients that are applied both on the loads applied to the structure and to the strength of the materials at stake. Considering the typical lifetime of infrastructures suggests that it is necessary to develop predicting models for creep over one century, beyond the current state-of-the-art, and also beyond experimental data available in the literature today.
Long term ageing of concrete
Another archetypal example where ageing of concrete is important is the storage of nuclear wastes. Such wastes (low to medium radioactive activity) are placed into fiber reinforced concrete containers stored into underground disposals. The role of concrete is (1) to support loads as containers are piled onto each other in the disposal, (2) to ensure the safety upon transport, especially if a container falls, and (3) to possibly prevent leakage of contaminants in the storage through cracks that could appear due to the applied loads on the container. Therefore, their serviceability ought to be assessed, but over a period of time which is 300 years typically.
The ageing mechanism at stake is the dissolution of the hardened cement paste due to a difference in calcium concentration in the liquid and solid phases of the material. This is called calcium leaching. Recall that concrete made with Portland cement – what is common today – has been developed in the late 19th century. A lifetime of 300 years is far beyond the experience accumulated since concrete has been used as a construction material. Obviously, such a duration is also beyond the reach of possible experiments and the ageing mechanism should be accelerated in laboratory tests: pure water is considered (so as to maximise the contrast between a calcium rich phase and a liquid phase without calcium) and an electrical field may be used to help the transport of the calcium ions.
The effect of accelerated calcium leaching can be observed on the SEM pictures in Fig. 2. By comparing the two pictures, needles of calcium hydroxide have disappeared upon leaching. Calcium hydroxide does not participate much to the mechanical properties of the hardened cement paste in the sense that, mechanically speaking, it is rather weak. But dissolution induces an increase of porosity of the hardened cement paste. Upon increasing porosity, the stiffness of the material decreases (up to 40%), the tensile strength decreases too, although to a lesser extend as the material becomes more ductile.9 In addition, the increase of porosity induces an increase of permeability. These phenomena, which are typically neglected for usual structures whose lifetime is up to 50 years, becomes prominent when a lifetime of 300 years is considered.

Figure 2: Cement paste before (left) and after (right) accelerated calcium leaching (after Le Bellégo10)
Test data on the evolution of the strength of structural elements upon leaching are seldom in the literature. Figure 3 shows the results of bending tests performed on beams made of mortar at various stages of leaching. Loading has been carried out until failure under displacement control conditions: a monotonically increasing displacement is prescribed at mid-span and the resulting force needed to reach this displacement is applied by the testing frame. After an elastic regime where the applied load is proportional to the displacement at mid-span, a peak load is reached which corresponds to the onset of crack propagation from the bottom face of the beam at mid-span. Then, the crack propagates and the applied load decreases until the beam breaks completely into two pieces.

Figure 3: Bending tests on notched beams at various stages of leaching. Loading system (left) and force-displacement at mid-span responses (right). The maximum carrying capacity of the material decreases by 60% upon 85% of leaching (after le Bellégo et al.11)
When the material is 85% leached, the maximum load that the beam can support is decreased by 60% approximately. These experimental data are useful for the calibration of chemo-mechanical models that couple leaching and the mechanical behaviour of the material in a computer code for structural design. In view of the lifetime considered here, numerical modelling is the sole possibility for design, keeping in mind that its calibration relies on accelerated experiments, whose representativity remains to be evaluated.
Concluding remarks
Creep and calcium leaching of cementitious materials such as mortar or concrete have in common one feature: the ageing mechanisms occur at the very small scale within the material and induce large consequences at the structural scale. These characteristics are not exceptions, other ageing mechanisms exist with similar features.
Freeze-thaw cycles induce cracking in concrete pavements due to the expansion of water in the pores of concrete upon freezing. This phenomenon is very commonly observed in cold countries where de-icing salts are massively used. Crystallization of salt has similar consequences on concrete (and stones). Upon crystallization, large pressures are generated in the pores of the hardened cement paste and induce cracking. This type of damage, along with corrosion of the reinforcement of concrete which follows, once concrete or mortar no longer protect the steel bars, is observed very often in marine environments. Finally, chemical reactions may develop between the cement and the aggregate (alkali-silica reaction). The chemical product, a gel, is expansive and therefore induces cracking of concrete due to large hydraulic pressures.
As a final remark, let us stress from these examples that ageing of concrete is a multidisciplinary topic, combining physics, chemistry and mechanics, with an unexpected wide spectrum of size and time scales. One cannot rely on past experience, numerical simulation is required, based on robust descriptions of the phenomena at stake. It is from the combination of these disciplines and scales that complexity arises. Concrete looks simple as it is widely used over the world, but it is a complex material, with exciting scientific challenges ahead.
- Angst, U., Elsener,. B. (2016). Chloride Threshold Values in Concrete – A Look Back and Ahead. ACI Special Publication SP-308 “Chloride Thresholds and Limits for New Construction”. American Concrete Institute. [↩]
- Torrenti, J.M., Pijaudier-Cabot, G., Reynouard, J.M., (2010). Mechanical Behaviour of Concrete, ISTE-Wiley Pubs, London UK. [↩]
- François, R., Laurens, S., Deby, F. (2018). Corrosion and its Consequences for Reinforced Concrete Structures, ISTE-Elsevier Pubs, London UK. [↩]
- Bazant, Z.P., Jirasek, M. (2018), Creep and Hygrothermal Effects in Concrete Structures, Springer, 921p. [↩]
- Bazant, Z.P., Yu, Q., Li, G.H., (2012). Excessive Long-Time Deflections of Prestressed Box Girders. I: Record-Span Bridge in Palau and other Paradigms, J. Struct. Engineering ASCE, Vol. 138, pp. 676 – 696 [↩]
- Burgoyne, C., Scantlebury, R., (2006). Why Did Palau Bridge Collapse, The structural Engineer, June, pp. 30-37. [↩]
- See e.g. Creep, Shrinkage and Durability of Concrete and Concrete Structures (2005), proceedings of ConCreep 7, G. Pijaudier-Cabot, B. Gérard and P. Acker Eds, ISTE Pubs. [↩]
- Vandamme, M., Ulm, F.J., (2009). Nanogranular Origin of Concrete Creep, Proc. Nat. Acad. Sci., Vol. 106, pp. 10552 – 10557. [↩]
- Le Bellégo, C. (2001). Couplages chimie – mécanique dans les structures en béton attaquées par l’eau : étude expérimentale et analyse numérique, Thèse de doctorat de l’Ecole Normale Supérieure de Cachan. [↩]
- Le Bellégo, C. (2001). Couplages chimie – mécanique dans les structures en béton attaquées par l’eau : étude expérimentale et analyse numérique, Thèse de doctorat de l’Ecole Normale Supérieure de Cachan. [↩]
- Le Bellégo, C., Gérard, B., Pijaudier-Cabot, G., (2000). Chemo-Mechanical Effects in Mortar Beams Subjected
to Water Hydrolysis, J. Engrg. Mech. ASCE, Vol. 126, pp. 266 – 272. [↩]

