Concrete Cracking Problems: A modern-day phenomenon?

October 13, 2016

Images courtesy PJ Materials Consultants Ltd.

By Paul Jeffs
Anecdotal evidence suggests cracking of concrete structures and components appears to be on the rise, often resulting in conflict between involved parties. Sometimes described as severe and/or excessive, it appears the cracks occurred even when specifications were strictly followed and the concrete designed, supplied, placed, and cured in accordance with formal standards and good practices of the construction industry. Often, cracks occurred soon after the concrete hardened, but then continued to grow over many months or even years.

This author believes the increase in the occurrence of cracks is caused by modern concrete mixtures that lead to a higher magnitude of naturally occurring autogenous shrinkage. The problem is compounded by the fact that stresses developed within the concrete by the resultant volume change affected by autogenous shrinkage are in addition to the known problems associated with drying shrinkage cracking—the latter having been shown to typically follow or combine with the former to aggravate the extent and severity of the cracks.

When cracking problems occur, investigation will typically focus on the practices and factors causing or influencing excessive drying shrinkage, such as excessive workability, rapid drying conditions, and poor curing practices. However, although drying shrinkage certainly should be considered when investigating the potential cause of excessive cracking, the influence of autogenous shrinkage is typically overlooked, but can sometimes be dramatically more important. Many researchers believe autogenous shrinkage strains can 
be at least equal to strains caused by drying shrinkage and, under certain conditions, can be considerably greater.

What is autogenous shrinkage?
Autogenous shrinkage and its adverse effects on hardened concrete have been well-studied by researchers and concrete scientists for decades, but the phenomenon remains a mystery to many who are involved at the grassroots of the concrete industry, and possibly even to a large number of seasoned practitioners. Perhaps one of the reasons for this may be the natural reluctance of the various involved parties to publicize their problems.

In simple terms, the complex phenomenon can be described as a physical, volume-reducing response to chemical reactions generated by the natural hydration of the cementitious binder, but occurring without moisture transfer to the environment. (Another form of shrinkage known as chemical shrinkage—often confused with autogenous shrinkage—also takes place because of cement hydration reactions, but this occurs very early and has no influence on the potential for cracking after hardening. It will therefore not be discussed further within this article.) Autogenous shrinkage can also be described as being the result of internal self-desiccation—a reduction in the amount 
of moisture that could otherwise be available for binder hydration reactions.


The American Concrete Institute (ACI) defines self-desiccation as:

the removal of free water by chemical reaction so as to leave insufficient water to cover the solid surfaces and to cause a decrease in relative humidity of the system.

However, the International Union of Laboratories and Experts in Construction Materials, Systems, and Structures’ (RILEM) book TC 196-ICC, State-of-the-Art Report on Internal Curing of Concrete, goes somewhat further by defining it as:

the reduction in internal relative humidity of a sealed system when empty pores are generated. This occurs when chemical shrinkage takes place at the stage where the paste matrix has developed a self-supportive skeleton, and the chemical shrinkage is larger than the autogenous shrinkage.

Self-desiccation can lead to a lower hydration compared with a moist-cured paste; this is because cement gel can only form within a water-filled space. In addition, according to the textbook Properties of Concrete, by expert Adam Neville, because hydration of a sealed system can proceed only if the amount of water present in the paste is at least twice that of the water already combined, self-desiccation becomes important for concrete mixtures when the water/cement (w/c) ratio is below about 0.50.

Mechanisms of autogenous shrinkage
As water is lost to internal autogenous hydration reactions, tensile stresses are generated in a similar mechanism that occurs when water is lost by evaporation to cause drying shrinkage. Both of these mechanisms can result in the formation of cracks. The most widely accepted theory to explain the volume change that subsequently occurs is capillary tension. This develops within the cement paste pores as moisture is lost and menisci are thereby formed. However, drying shrinkage focuses the stresses at the exposed concrete surfaces, whereas autogenous shrinkage issues occur throughout the mass of concrete.

Figure 1: Correlation between autogenous shrinkage strain and water/cement (w/c) ratio (adapted from Leivo and Holt’s Autogenous Volume Change at Early Ages).

Researchers and experts such as Neville have explained the strains developed by the formation of the menisci pull the cement paste closer together and register as a volume reduction. If strains continue to be generated to the extent the tensile stresses exceed the strength resistance, the concrete will crack. Researchers have also indicated in high-strength concrete with a low w/c ratio, the finer pore structure within the cement paste causes the menisci to have a greater radius of curvature. These menisci create a larger compressive stress on the pore walls, thus generating greater autogenous volume change as the paste is pulled inward.

Laboratory studies have confirmed autogenous shrinkage typically occurs at a very early age, but, under certain conditions, it can continue over several days and even weeks. As the rate of creep and relaxation in new hardened concrete can be high, the stresses that occur at early ages can rapidly disappear. However, depending on the extent and severity of the influencing factors, internal microscopic cracks can form—although, at early ages, they may have little adverse effect on the short-term durability of the concrete. Unfortunately, should autogenous shrinkage continue to develop, the microcracks can subsequently widen and propagate to facilitate the ingress of moisture. After several weeks, even months, the influence of drying and carbonation shrinkage means more cracks may occur and the existing ones could further widen and lengthen. According to ACI 446.1R, Fracture Mechanics of Concrete: Concepts, Models, and Determination of Material Properties, cracks may propagate at much lower stresses than are required to cause crack initiation.

The influence of free-mixing water content on the strains generated by the autogenous shrinkage resulting from self-desiccation has been well-researched, with typical conclusions indicating strains can be significant as the w/c ratio becomes less than 0.40 (Figure 1). However, it has also been claimed variations in cementitious binder chemistry and particle fineness can result in autogenous shrinkage becoming significant when w/c ratio limits vary between 0.36 and 0.48.

What factors influence the magnitude of volume change?
In addition to the effect of water content, autogenous shrinkage is very sensitive to cement composition, content, fineness, and temperature. However, the materials used in modern concrete, such as specialty admixtures and supplementary cementing materials (SCMs), can also have an influence on the magnitude of autogenous shrinkage.

From traditional to modern concrete
Less than a century ago, conventional concrete typically consisted of four basic ingredients—cement, sand, graded coarse aggregate, and water. To meet the demands for improved resistance to the actions of freezing and thawing, the first commercially available chemical admixture products were predominantly air-entraining agents. These were followed by water-reducing admixtures (WRAs), which typically provided moderate reductions in mixing water so workability could improve or cement content could be reduced. A variety of products subsequently become available, typically formulated by the manufacturers to modify setting properties (i.e. retard or accelerate), provide waterproofing properties, and improve cohesiveness for pumping.

However, before the commercial availability of admixtures, another important change had taken place. This change was initiated by the cement manufacturers in response to the increasing demands for more rapid setting and strength development properties of concrete for taller, thinner buildings, structures with wider spans, and faster construction times. This led to changes in the chemical composition and fineness of cement, so the desired hardened concrete properties could be achieved by a refinement of the hydration reactions.

Figure 2: This chart shows effect of superplasticizer does on early-age shrinkage (adapted from Autogenous Volume Change at Early Ages).

Unfortunately, after these changes were introduced, cracking problems began to occur. It was determined the causes were linked to inadequate strength of the hardened concrete to resist the development of shrinkage, although it is unlikely the influence of self-desiccation and autogenous shrinkage were specifically evaluated during the investigations. Further, this led to changes being made to codes and specifications regarding maximum w/c ratio, minimum cement contents, and minimum compressive strength requirements. One of the main driving forces for the changes was the American Association of State Highway and Transportation Officials (AASHTO), which modified its specifications to require a maximum 0.445 w/c ratio (down from 0.53), a minimum 362 kg/m3 
(610 lb/cy) cement content, and a minimum compressive strength of 30 MPa (4500 psi) at 28 days. For 40 years prior to the changes, AASHTO had required a compressive strength of 20 MPa (3000 psi) at 28 days.

The next major event that influenced the magnitude of autogenous volume change was the development and the commercialization of high-performance concrete (HPC), which was originally called high-strength concrete.

High-performance concrete
The evolution that led to today’s concept of HPC began about 45 years ago with the development of superplasticizing admixtures in Germany and Japan. Until that time, high-strength concrete—generally classified as having greater than 40-MPa compressive strength at 28 days—had been produced using high- cement contents, often supplemented with fly ash and traditional water-reducing admixtures. This practice continued well into the late 1970s in North America with compressive strengths limited to about 70 MPa at 28 days—until the introduction of superplasticizing technology.

As superplasticizer’s raw material manufacturing costs were high and the rates of addition considerably more than conventional products, it was not initially perceived that these new admixtures could be commercially feasible for increasing compressive strength, particularly as cement was relatively inexpensive at the time. Thus, they were originally used for producing highly fluid concrete without the need for additional water. However, the oil crisis of the mid-70s led to higher cement costs—while the increased international sales volume through the ’80s contributed to more economical manufacturing costs for the admixtures. Therefore, it became economically feasible to utilize the superior fluidizing properties of superplasticizers to considerably reduce mixing water and thereby significantly increase compressive strength, without compromising placement properties. Unfortunately, much higher dosage rates of superplasticizer were required to achieve the large water reductions to obtain these increases. This also increased the magnitude of early age shrinkage (Figure 2).

The next major advancement in HPC technology occurred during the ’80s, and took advantage of the beneficial properties provided by supplementary cementing materials, such as fly ash, ground granulated blast-furnace slag (GGBS), and silica fume. Although the benefits of fly ash and GGBS were known before this time, it was really the highly reactive properties of the micro-fine particles of silica fume that provided the giant leap forward. Today, it is generally recognized that silica fume can be a prime ingredient in HPC mix design, but there are alternatives and often benefits by substituting all or part of the silica fume content with fly ash or slag. Sometimes, a combination of two or more SCMs is used to gain a synergistic effect. Also, today, the use of the modern materials and technologies means concrete can be routinely produced that develops well in excess of 100-MPa compressive strength.

Unfortunately, studies have confirmed the use of SCMs can intensify self-desiccation and this process begins very early. In particular, the refined cement paste pore structure created by the micro-fine particles of silica fume generates a considerably lower internal relative humidity. Pozzolanic reactions further influence this effect. Silica fume also induces acceleration of hydration reactions of cement at early ages, although experimental data has indicated the resulting self-desiccation can continue for several months.

Figure 3: Originally narrow and shallow, these plastic shrinkage cracks widened, deepened, and lengthened due to volume change.

It is well-known the refined cement paste pore structure produced by the use of silica fume also prevents the natural bleeding that otherwise occurs with conventional concrete. Under rapid drying conditions and without adequate precautions being taken, the result can be the formation of plastic shrinkage cracks. Later, when the effect of volume change caused by self-desiccation occurs, these originally narrow and shallow cracks can widen, deepen, and lengthen (Figure 3). This phenomenon led to the introduction of techniques for use during what is now known as initial curing; these practices attempt to reduce the potential for rapid drying of the immediate concrete surface, prior to the application of conventional curing techniques. (Initial curing techniques typically include the use of evaporation retarders and/or mist-spraying [fog-misting].)

Other factors that can influence shrinkage cracking
A critical aspect of autogenous volume change is—since it occurs within the mass—it can only marginally be influenced by surface-applied water-curing techniques. Liquid-applied curing membranes can have no influence on autogenous self-desiccation. Based on the wealth of available research studies, it is therefore highly unlikely that even the most efficient curing techniques will be able to prevent cracks from forming within concrete susceptible to high levels of internal self-desiccation. If improved durability is a critical requirement, it is therefore important the range of factors that can influence whether a particular concrete could crack are thoroughly evaluated.

Any material or construction parameters that cause a delay of setting will prolong the period when early-age autogenous shrinkage occurs. Delays in setting can be caused by the use of retarders, some superplasticizers, fly ash, and GGBS. Placing concrete during cold weather can also provide delays in setting and the time of hardening. Some researchers have reported increasing levels of GGBS as a partial cement replacement greatly increases autogenous shrinkage at the same w/c ratio.

Figure 4: Unreinforced concrete, toppings, and overlays are more likely to crack for concrete susceptible to high internal self-desiccation.

In addition to the well-known concerns when placing any type of concrete during hot weather, high-ambient temperatures can also adversely influence autogenous shrinkage cracking. This is because the heat of hydration reactions—and thereby internal self-desiccation—occur more rapidly and at a time when tensile strength development may not be adequate to resist the resulting autogenous volume change. Since the reactions are adversely influenced by high ambient temperature, there is also a possibility that differential autogenous shrinkage can result, concentrating strain development within the surfaces exposed to the higher temperature. The problems can have even greater consequences for mass concrete containing high cement contents.

According to ACI 224.1R, Causes, Evaluation, and Repair of Cracks in Concrete Structures, shrinkage cracking can be controlled by using contraction joints and proper detailing of the reinforcement. Well-defined reinforcement allows the development of stress well beyond that corresponding to maximum stress, permitting concrete to typically withstand greater volume change. Therefore, unreinforced concrete, toppings, overlays, etc., are much more likely to crack for concrete susceptible to high levels of internal self-desiccation (Figure 4). The degree of restraint or relaxation can have a considerable influence on whether volume change can cause cracking. Relaxation provided by the properties of modulus of elasticity and creep is often sufficient to prevent shrinkage cracking at early ages for conventional concrete. However, both of these properties are well-known to potentially have less influence when HPC or high-strength concrete is selected.

When does shrinkage stop?
It has already been stated that autogenous shrinkage takes place very early in the hydration process and can continue for several weeks, if not months. It has also been inferred that drying shrinkage can add to the strains generated by autogenous shrinkage and subsequently cause cracks to occur. Carbonation shrinkage, which typically can take years to occur to any degree, can also impose a cumulative effect on strain development, has also been discussed. So when does total shrinkage stop?

In Properties of Concrete, Neville states  shrinkage takes place over long periods: “some movement has been observed even after 28 years, although a part of the long-term shrinkage is likely to be due to carbonation[.]” From the data he provides, it can be observed for the average of the concrete that was studied, about 40 per cent of the total shrinkage occurred after 28 days and about 80 after one year. The rate of shrinkage then decreased rapidly with time (Figure 5). However, the study Neville includes in his book was published in 1958, which is well before the introduction of the materials and technology used today. It is therefore logical that an even greater effect can be expected when modern concrete materials and mix designs are used.

Figure 5: Range of shrinkage-time curves for different concretes stored at relative humidities of 50 and 70 per cent (source: Neville’s Properties of Concrete).

Mitigating autogenous shrinkage cracking
Established good concrete protection practices are still essential if cracking of conventional concrete is to be prevented or mitigated. However, as already discussed, even efficient wet-curing techniques may not be sufficient to prevent cracking of concrete that is susceptible to severe autogenous shrinkage. Perhaps the best way in which this type of cracking can be addressed is by selecting a concrete mix design that avoids high cement content, as well as low w/c ratios (below about 0.45), and high levels of supplementary 
cementing materials. High rates of superplasticizer addition should also be avoided.

If this is unavoidable, then the use of some non-conventional materials or techniques could prove beneficial, although trials are recommended in advance of use so that any side effects can be fully understood. Some of the methods have been developed to provide the benefit of what has become known as internal curing, and include the use of fine, saturated lightweight aggregates and saturated super-absorbent polymers. Both of these provide an internal source of water to replenish moisture lost during self-desiccation reactions. Saturated, fleece-lined, controlled permeability formwork has also been used although the availability of moisture is limited to the immediate interfacial zone of concrete.

Shrinkage-reducing admixtures, typically utilized to counter drying shrinkage, can also be used to mitigate autogenous shrinkage cracking; these materials are designed to alter the surface tension of the pore water, thereby reducing the capillary stresses developed during desiccation—whether self-induced or by evaporation. Expansive cement or shrinkage-compensating cement technology has also been used to counter the volume reduction due to autogenous shrinkage by the development of initial expansion. Finally, the addition of fibres may also have a beneficial influence on the early-age cracking of concrete; although fibres do not typically reduce the shrinkage of concrete, they can increase the resistance to crack propagation and reduce crack width and frequency.

It should come as no surprise to learn that since the commercial availability of high-performance concrete, the problems associated with autogenous shrinkage have become more widespread and profound. Certainly, modern concrete, has all the ‘ingredients’ for dramatic increases in potential autogenous shrinkage and is more susceptible to excessive early and later-age cracking when compared to more traditional concrete mixtures. Hopefully, once the combination of all the factors that exacerbate adverse reactions associated with autogenous shrinkage are better understood—by parties involved with the design and construction of concrete structures—then perhaps problems and conflict can be avoided and the potential for cracking can be reduced to a nominal, more 
acceptable degree.

Terms to describe types of cracks vary from region to region, but cracks resulting from hydration reactions or moisture transfer are most commonly classified by their cause or by their appearance.

Plastic shrinkage cracks
Plastic shrinkage cracks occur within the surface of fresh concrete while it is still plastic, and are characterized by their random pattern. They are usually discontinuous and rarely extend to a free edge. The cracks are caused by the rapid drying of the surface, particularly during warm and wind-drying conditions. The potential for cracking increases with workability—particularly when inadequate curing measures are taken.

Drying shrinkage cracks
Drying shrinkage cracks occur when concrete is subjected to a lower environmental relative humidity (RH) than its internal relative humidity. The mechanism is generally accepted to be due to the development of capillary tension as water is lost from the pores and menisci are formed. Shrinkage of the cement paste of a concrete will be larger the higher the w/c ratio because the latter determines the amount of water in the cement paste and the rate at which water can move toward the surface.

Carbonation shrinkage cracks
Carbonation shrinkage cracks can occur because of atmospheric carbon dioxide diffusing through the cement matrix pore structure to react, in the presence of moisture, with calcium hydroxide. Carbonic acid is formed during the initial reactions that subsequently result in the formation of calcium carbonate. Cracks form as the result of dissolution of the calcium hydroxide crystals while the crystals are under pressure, and deposition of calcium carbonate in places where the carbonate is not under pressure.

Surface crazing and map cracksinadequate curing practices are carried out and rapid drying conditions prevail. The result is that the top surface cracks in a random manner, sometimes referred to as alligator cracking or mud-cracking, with 
the cracks usually only extending a few millimetres below the surface.

Thermal stress cracks
Thermal stress cracks occur as a result of the evolution of heat during the cement hydration process. The potential will increase with increased thickness of section and cement content. Unlike drying shrinkage or carbonation shrinkage, contraction during cooling of mass concrete is greater within the interior.


paul-jeffsPaul Jeffs has more than 45 years of experience in the construction industry around the world. He is principal of PJ Materials Consultants Ltd., a Guelph-based company that provides consulting and sub-consulting services across Canada for the investigation, construction, and restoration of masonry and concrete structures. He can be reached at[4].

  1. [Image]:
  2. [Image]:
  3. [Image]:

Source URL: