U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590

Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-HRT-04-150
Date: July 2006

Chapter 6. Voids

Previous | Table of Contents | Next


In hardened HCC, a void is an empty space (other than a crack) in the cement paste that contains nothing but air. The type, size, shape, arrangement, and abundance of the voids are factors controlling many important properties (e.g., compressive strength, resistance to destruction bycycles of freezing and thawing, and resistance to chemical attack on the reinforcing steel and the cement paste). The percentage of air-void volume is generally specified by the design of the mixture. A large number of very small (most are not visible without magnification) air voids are desirable so that the average distance between any point in the paste and a void (spacing factor) is very short and, thus, the paste is protected from freezing and thawing. However, because voids weaken the HCC, spacing factors much less than the maximum ensuring protection from freezing and thawing are counterproductive and should be avoided.

The total air-void content (of voids larger than capillary size) of an unhardened concrete mixture is routinely determined in accordance with ASTM C 231 (pressure method) or ASTM C 173 (volumetric method). The unit weight of fresh concrete (ASTM C 138) may also be used to determine the percentage of air in the mixture. These methods do not ascertain the type of voids present; they merely measure the total void content. These measurements are important. As Bartel (1978) stated:

Tests for air content and unit weight of fresh concrete, carefully made in accordance with the appropriate ASTM test method, will yield an accurate measurement of the amount of air, weight, and volume of concrete being produced. Tests for air content, coupled with intelligently selected specification limits, can ensure the beneficial effects of entrained air in hardened concrete (p. 130).

Specially formed specimens of hardened HCC mixtures may be tested for resistance to the destructive forces of freezing and thawing in accordance with ASTM C 666 (resistance to rapid freezing and thawing). An HCC that is resistant generally indicates that an adequate air-void system is present and the aggregates are sound or that the HCC has not become critically saturated.

It has been variously claimed that the total air- void content increases or decreases as the concrete hardens. It appears that what really happens is that the determination of the total air- void content with field equipment made on the fresh concrete does not agree with the total airvoid content determined by microscopical analysis of the hardened concrete. Except in the case of hydrogen gas being evolved by the corrosion of aluminum fragments (Newlon and Ozol,1969) (see figure 52), no evidence has been offered showing volumetric changes in field placements or test cylinders to corroborate an increase or decrease in total air-void content.

Careful investigation by a combination of controlled sampling and testing with petrographic analyses has shown that air-void content does not change upon hardening. Rather, it may change because of outside influences (e.g., excessive vibration, the further addition of water and retempering, or pumping operations with long vertical drops). The void content in the hardened

Figure 52. Concrete that increased in volume because of corrosion of aluminum fragments.

After it was cast in the cylinder mold, the concrete increased in volume as hydrogen gas evolved from the chemical reaction of aluminum (from an aluminum delivery pipe) with the alkaline fluids of the fresh cement paste. The photo shows two 150 by 300 millimeter heavy steel cylinder molds with concrete rising above the top rim of the molds. To give scale a hand is shown holding a pen on the top rim of one of the molds to show that the concrete has risen about 10 millimeters.

After it was cast in the cylinder mold, the concrete increased in volume as hydrogen gas evolved from the chemical reaction of aluminum (from an aluminum delivery pipe) with the alkaline fluids of the fresh cement paste.

state will usually agree within 1 percent of the void content determined in the fresh, unhardened state. When they do not closely agree, either one of the measurements is in error or the two specimens tested do not represent the same concrete subjected to the same influences (see appendix C and Ozyildirim, 1991).

An air-void content in excess of the amount required for protection against the destructive forces of alternate freezing and thawing that occur in saturated concrete adds no benefit to concrete expected to bear loads and resist abrasion. (For a discussion of the high-air cellular concretes, see Legatski, 1994.) An excessive air-void content will lower the compressive strength of the concrete by about 5 percent for each excess percentage of voids. Because of this dual-edged nature of air entrainment, maintaining control of the air content is one of the most important facets in concrete production. It requires knowledge of the various materials and potential interactions, batching and mixing equipment and operations, and careful testing. Whiting and Nagi (1998) provide excellent information on this critical function.


The overall void content in HCC is composed of four general types of voids (as listed in table 15).

Table 15. Types of voids.
  1. Capillary voids: Capillary voids are irregularly shaped and are very small (less than 5 µm on the lapped surface of the slice examined). They represent space originally filled by mixing water, remain after the hydration of the cement gels, and are an integral part of the paste. Although they contain air at the time of examination, they are not considered part of the air-void system.
  2. Entrained air voids: Entrained air voids are defined by VTRC as spherical voids larger than the capillaries, but less than 1 mm on the lapped surface of the slice examined. They are formed by the folding action of the concrete mixer, and their shape, size, and abundance are influenced by the addition of surface-active, air-entraining admixtures to the mixture.
  3. Entrapped air voids: Entrapped air voids are voids that are larger than the entrained voids, but have internal surfaces that indicate that they were formed by air bubbles or pockets. They may be spherical or irregularly shaped.
  4. Water voids: Water voids are irregularly shaped voids whose shape, location, or internal surface indicates that they were formed by water. Usually, they are larger than entrained air voids.

6.2.1 Capillary Voids

The smallest optically visible voids in HCC are the various sizes of capillaries. A few of the larger capillary voids can be seen at the higher magnifications used to determine the parameters of the void system; however, they are generally not that large. Hearn, Hooton, and Mills (1994) provide an excellent discussion of the capillary void system in HCC. Capillary voids are spaces formed by the shape of the hydrated cement gel structures and spaces left between the gel structures as water is used in the self-desiccation of the hydration process. They were occupied by water or gas when the concrete was fresh and are larger and more abundant in concretes with a high w/cm. The extent (volume and continuity) of the capillary system is thus controlled by the w/cm and the degree of maturity of the concrete. The evenness of the distribution of the poresand capillaries is controlled by the distribution of the water. As the concrete hydrates, the water in the pores is used in the hydration of the cement. As the concrete matures, much of the capillary space becomes filled with the products of hydration and the products of any reactions occurring between the chemicals of the paste and the aggregate. Some of the finer capillaries are spaces created by differential crystal growth.

The capillaries are detected only when specialized methods are used. In laboratories thus equipped, the various types of electron microscopes can be used to view the capillary void system. In the VTRC laboratory, the abundance and location of the capillary voids are detected by use of the P/EF microscope in the study of fluorescent thin sections of the specimen concrete (see chapter 13). At the NIST laboratory, the SEM is used (see chapter 14). Rarely, capillary voids can be noted during the determination of the parameters of the void system. In that event,capillary voids are considered to be paste.

6.2.2 Entrained Air Voids

Entrained voids are small spherical voids enfolded by the mixer. Surface-active, air-entraining admixtures are added to the mixture to stabilize a specified percentage of these voids and thus protect the hardened HCC against the destructive forces of freezing and thawing. Thus, the entrained air void is a desirable void. Entrained air voids are generally considered to be larger than the capillaries (at least 5 µm in diameter), but smaller than the entrapped voids (Verbeck,1966, 1978; Hearn, et al., 1994). Entrained air voids have so much surface tension relative to their volume that they are little affected by the shape of nearby particles. Distortion occurs in these small voids only when external forces distort the concrete after the beginning of hardening.

The presence of the proper quantity of well-distributed entrained air voids can prevent deterioration of the concrete by relieving the pressures that can develop during the freezing and thawing process (Helms, 1978; Hover, 1994; Newlon, 1978; Newlon and Mitchell, 1994; Pigeon and Pleau, 1995). Entrained air also facilitates the placement of the concrete because the entrained air voids increase the fluidity of the paste.

Very irregularly shaped small voids (maximum dimension of less than 1 mm) cannot be properly called entrained voids because they lack the characteristic shape showing that they were formed by the surface tension of the air-entraining admixture. It is not known if such voids function to protect the concrete against the deterioration caused by freezing and thawing. Small, irregular voids (particularly if not at an aggregate boundary or a wearing surface) may be evidence of retempering (see appendix C).

Figures 53 and 54 show concretes with different percentages of air voids.

Figure 53. Surface of finely lapped slice of concrete containing 5.6 percent total air voids.

The concrete surface shown has many very fine air voids spaced throughout. It also has a few larger voids, including a 2 millimeter void (larger than an entrained air void), which is shown marked by a white arrow. The void is in the mortar phase, between two coarse aggregate pieces. The void content of this concrete is in the middle of the specification range

The void content of this concrete is in the middle of the specification range. The large void marked with an arrow is about 2 mm across (larger than an entrained air void). Notice the very fine voids throughout the paste.

Figure 54. Surface of finely lapped slice of concrete containing 17 percent total air voids.

The concrete surface shown can be characterized by large numbers of voids of approximately 1 millimeter diameter as well as smaller voids. The area of darker paste (lower left) has a lower void content. If an H C C contains more than one kind of paste, this generally indicates that the mixture had begun to hydrate before additional water was added. The void content of this concrete is much greater than the upper limit of the specification range. An example 1 millimeter void is shown marked with an arrow.

The void content of this concrete is way above the upper limit of the specification range. The void indicated by the arrow is about 1 mm across. The area of darker paste (lower left) has a lower void content. If an HCC contains more than one kind of paste, this generally indicates that the mixture had begun to hydrate before additional water was added (see section 8.4 and appendix C).

6.2.3 Entrapped Voids and Water Voids

All voids, regardless of shape, that have a maximum dimension (on the surface examined) of more than 1 mm are defined by VTRC as entrapped voids (large spherical) or water voids (large irregular). If voids occur flattened out at the boundary between the aggregate (usually coarse aggregate) and the paste, they are a class of entrapped voids called boundary voids.

Voids larger than entrained voids have no appreciable beneficial effect and weaken the HCC. Such voids are controlled by the efficiency of whatever system of consolidation is in use and the workability of the concrete at the time of consolidation. Particular voids may be caused by too much water in the HCC, a strong affinity of a particular aggregate lithology for water, improper consolidation, and occasionally by the dissolving away of Ca(OH)2. Water- or air-filled pockets not removed by consolidation will leave irregularly shaped voids (see figure 55). Water voids are irregularly shaped voids created in the HCC by bleed water prevented from rising to the surface by an aggregate particle or the hardening of the paste. Water voids contained water when the HCC was fresh and unhardened; however, in the hardened state, these voids are filled with air and thus might be more properly termed water-formed air voids.

Figure 55. Concrete core with about 4 percent large, irregularly shaped voids.

The surface of the concrete cross section shown has irregularly shaped voids. In this instance, the concrete, which had not yet been consolidated, became hard and unworkable while repairs were being made on the paving equipment. The specimen is 100 millimeters across, and some of the large voids exceed 5 millimeters in size.

In this instance, the concrete, which had not yet been consolidated, became hard and unworkable while repairs were being made on the paving equipment. The specimen is 100 mm across.


6.3.1 Overview

In hardened concrete, the air-void system may be characterized by one of two methods described in ASTM C 457—the linear traverse method or the modified point-count method. The parameters calculated include:

The air-void analyzer (AVA) is an apparatus that determines the characteristics of the air-void system from a sample of fresh concrete. Developed in Europe, the device measures the rate of air loss from a sample of mortar mixed with a viscous liquid and then uses Stoke’s law to determine the size of the bubbles rising through a water column. The device reports the air content by volume of concrete contained in the entrained air voids, the specific surface of the entrained airvoid system, and a value for the spacing factor. The AVA is a focus technology of the American Association of State Highway and Transportation Officials (AASHTO) Technology Implementation Group. See http://www.aashtotig.org/?siteid=57&pageid=689 for more information.

6.3.2 Methods and Equipment Overview

New methods and equipment are continually being devised to monitor and determine the air-void parameters of hardened concrete. It is part of the job of the petrographer to assess the value of new methods and equipment and to decide which method is of value in which situation and, therefore, which equipment is worthy of a place in the budget of the organization. If the results of an air-void determination are to be presented in court and the testimony of opposing expert witnesses will be heard, any deviation from the principles of ASTM C 457 that has not been agreed upon by the litigants may invalidate the results of the analysis. Within an organization, particular deviations from the strict interpretation of ASTM C 457 may be acceptable if justified by their experience.

According to ASTM C 457, air-void system analyses can be efficiently performed with several methods and kinds of equipment. Suitable equipment for the determination of air-void parameters in hardened concrete includes, but is not necessarily limited to: (1) linear traverse, (2) modified point-count, and (3) image analysis equipment. Typically, the types of equipment to be used permit or facilitate the movement of the specimen of HCC on the stage of a microscope so that data may be collected over the specified area and from the specified length of traverse Generally, the data collected are:

  1. Total length of the traverse over which the determination is made: In the modified point-count method, the total number of points examined and the distance the equipment moves between the points are required for the calculations.
  2. Portion of the traverse that is across air voids: In the linear traverse method, this portion is the sum of the chord lengths across air voids; in the modified point-count method, this portion is the number of points that occur in air voids multiplied by the distance the equipment moves between points.
  3. Number of voids occurring in the traverse examined: The accuracy of the determination of the specific surface and spacing factor is completely dependent on the accuracy of the count of the number of voids on the line traversed. In the linear traverse method, the number of voids in the traverse is the number of chords collected; in the modified point-count method, it is the number of voids counted along the traverse line.

The procedures detailed in ASTM C 457 are those to be used with nonelectronic types of equipment (see ASTM C 457, figures 2 and 6). When equipment is used that includes automatic devices for moving the specimen, electric or electronic counters or totalizers and calculators, or measuring devices, the equipment must allow adherence to the principles of ASTM C 457 and permit or perform the calculation of the same parameters of the air-void system from the same data. The exact procedures followed for the operation of the equipment must be those described and specified by the fabricator of the equipment.

It is not known which type of equipment produces the most accurate results or how accurate the results need to be. Modified point-count equipment is favored by those who need speed. If the air-void parameters of spacing factor and specific surface are not required, a simple point count can be performed. Those who want a record of the chord-length distribution for research purposes favor linear traverse equipment, whereas those who want speed and the ability to collect a lot of data and manipulate them on a computer in many different ways favor image analysis equipment. Image analysis equipment can strain the equipment budget,but requires less operator time since a human does not actually perform the examination.Research laboratories will usually require either point-count or image analysis equipment for speed in making routine determinations,and linear traverse equipment for its ability to determine chord-length distribution on a surface unaltered by the fillers and dyes required by image analysis.

NOTE: The air-paste ratio method of calculation detailed in ASTM C 457 is to be used ONLY: (1) when proportions of the ingredients in the mixture are known with some certainty; (2) when it can be assumed that no change in mixture proportions has occurred (e.g., retempering has not occurred (i.e., the amount of paste can be closely calculated)); AND (3) when, because of the lack of exposure of a generalized specimen of the HCC or because of the extremely large size of the aggregate, it is impossible to obtain a specimen of the HCC for microscopical analysis with an aggregate distribution that is representative of the placement. The air-paste ratio calculations use the aggregate-paste ratio of the mixture design to transform mathematically the air-paste ratio and other air-void properties determined microscopically to an estimated percentage of air voids, specific surface, and spacing factor for the concrete. In these situations, it is convenient to select a specimen of HCC that is low in aggregate so that the microscopist will not have to spend excess time moving over aggregate. Linear Traverse

Using the linear traverse equipment (see figure 56), the operator tabulates the chord lengths across all phases of interest and records them for later analysis (Walker, 1988). This sort of data permits the straightforward calculation of the void parameters by the summing of the lengths of the chords and the counting of each occurrence of a phase. Because the calculations are extremely sensitive to errors made in the determination of the number of voids traversed, the method of deciding whether a void is or is not touched or transected by the line of traverse must be consistently applied throughout the analysis. If the individual lengths of the chords of the air voids are recorded and particular shape assumptions are made, a graphical representation of the chord lengths will indicate the size distribution of the air voids. The collection of the air-void data requires one pass of the microscope along the traverse line. The data necessary for calculation of the paste content may be collected at the same t me or a separate determination can be made for the paste content. This procedure is further discussed in section 7.1.2. With some types of linear traverse equipment, all the air-void parameters are automatically calculated; with others, the calculations must be performed on a calculator or computer. Point Count

With the point-count equipment, the operator records the type of substance (air void, paste, or aggregate) appearing at the index point of the reticle at a large number of points as provided by the click stop locations on the HCC specimen as the stage and specimen are moved under the microscope. The points may be randomly distributed or regularly distributed on a randomly placed grid or on a traverse line. Data concerning the relative amounts of all the phases and the voids encountered can be collected from one pass along the traverse line. If difficulty in distinguishing the exact paste-aggregate boundary is encountered, the operator may consider a separate pass over a lightly etched surface to collect these data (see section 7.1.2). The air-void parameters may be calculated by the analysis equipment or separately calculated using a calculator or computer.

Figure 56. Partially automated linear traverse equipment for determining air-void parameters.

The photo shows a stereoscopic microscope with a polished sample set on the stage. Either linear traverse or point count software can be used to control the motion of the stage and collect data. A video camera is also mounted on the microscope so that the image of the polished surface can be shown on a screen.

Either linear traverse or point-count software can be used to control the motion of the stage and to collect data. Image Analysis

Image analysis systems incorporating a variety of equipment have been used to examine HCC. The typical setup uses a video camera with a magnifying lens (see figure 57) or a standard stereomicroscope with a video camera (see figure 56). Peterson, Swartz, Sutter, and Van Dam (2001) describe the use of a flatbed scanner for imaging. The collected images are then analyzed by computer software to calculate the air-void parameters. To enable the image analysis software to recognize air voids as distinct from other features requires that the specimen be specially prepared so that each of the three major phases of interest (voids, paste, and aggregate) is a distinct tone (e.g., white, black, and medium gray). A typical surface preparation is to ink the surface black, fill the voids with a white powder (e.g., kaolin) and remove the excess from the surface, and finally ink out any filled voids that the system should not classify as air voids (i.e.,voids in aggregate particles). Peterson, et al. (2001) used a thr e-step procedure by collecting

Figure 57. Image analysis equipment.

Photo shows an assembly of video camera with magnifying lens mounted vertically and scanning a concrete slice.  A computer screen in the background shows the progress of the analysis. (Photograph by R.H. Howe, courtesy of the Pennsylvania Department of Transportation.)

Photograph by R.H. Howe, courtesy of PennDOT

The instrument is shown in the process of analyzing the air-void system of a slice of concrete. The screen in the background shows the progress of the analysis.

images of : (1) the untreated surface, (2) the surface treated with phenolphthalein (to differentially stain the paste), and (3) the inked and filled surface. The specimen preparation methods for image analysis can be exacting and can make the surface used useless for ordinary stereomicroscopic examination (as described in chapter 8). If it is desirable to perform normal examinations, these should be performed before the special surface treatments or, alternately, on an opposing lapped surface. Missouri DOT is currently managing a pooled-fund study to develop image analysis software with processing and pattern recognition capabilities to determine the airvoid parameters from an ordinary lapped surface (Missouri DOT, 2000).

Automatic systems that require filling the voids (thus hiding their interior surface) cannot make particular distinctions by a human operator possible. The human operator can often mentally reconstruct what the surface examined would have been if this or that flaw had not occurred. The human operator can judge whether an observed void is an air void, a fly-ash cenosphere, or the hole left where a small round grain of sand has fallen out. These distinctions are generally possible by the study of the reaction products and the luster of the interior of the void. Other Considerations

At VTRC, it has long been recognized that the accuracy of a linear traverse determination of the air-void parameters is as dependent or more dependent on the number of voids encountered and measured along the traverse as it is on the length of the traverse. Once 1000 voids have been measured and counted, the results from the data obtained subsequently change very little. Snyder, Hover, and Natesaiyer (1991) conducted an analytical investigation of the effect of the number of voids and the length of the traverse on the minimum expected error that can be encountered in a linear traverse determination of the void parameters in hardened HCC. Their work supports the belief that little additional accuracy is achieved if the determination includes more than 1000 voids and that almost no additional accuracy is achieved with more than 2000 voids.

When the air-void parameters are determined for intradepartmental purposes , the traverse length can be limited to that needed to count the 1000 or more voids. In such cases ,the traverse length needed should be estimated and the specimen set up so that the traverse length is evenly distributed over the available surface. Under circumstances when ASTM C 457 requires 2540 mm of traverse, it is estimated that only 1270 to 1778 mm are required for the collection of data from 1000 voids in ordinary within-specification concrete. Should a legal controversy arise concerning the subject concrete, any traverse length deficiency can be made up by collecting additional data from traverse lines (also evenly distributed over the surface). If the traverse direction and starting point are randomly chosen, in both cases, the randomness of the data collection will be maintained. It is our view that spreading the data collection area over as large a surface area as possible so that any irregularities of void distributi n (any clumping or areas devoid of voids) become part of the data recorded and examined is more important than the length of the traverse line.

The corollary is that if the number of voids is very small because of low air content or large voids, the length of traverse recommended in ASTM C 457 is probably not sufficient to obtain accurate air-void parameter data (Snyder , et al., 1991). Under such circumstances and with borderline compliance with specifications, it may be wise to use an additional length of traverse to ensure accuracy. In most cases, the small specific surface and large spacing factors caused by the lack of sufficient small voids will decisively indicate the inadequacy of the void system to provide freezing-thawing durability.

The method of deciding whether a void is or is not touched or transected by the line of traverse must be a simple rule that is firmly adhered to throughout any particular analysis. Pleau, et al. (1990) , using the point-count method, recommended: "A simple way to guarantee a random choice is to systematically choose the constituent located in a given quadrant (of the field viewed),say the upper-left corner of the crosshairs" (p. 5). A similar method can be devised for whatever type reference point, reticle, and counting method are in use. Other researchers (e.g., Mather, 1989) have suggested that points in dispute be collected in a separate register and later distributed to the totals of the constituents in the same proportion as are the data concerning which there is no dispute.

6.3.3 Preparation of Specimens

The importance of the proper preparation of the surface of the slice of concrete cannot be overemphasized (see figures 42 through 44). In most laboratories, skilled, highly trained technicians prepare specimens. A poorly prepared specimen can cause a determination of the percentage of air present in a specimen to deviate from the true value by as much as 2 percentage points (20 percent to 50 percent of the true value). A rough surface makes it impossible to detect small voids. This will have the effect of lowering the detected percentage of air, decidedly lowering the specific surface, and thus raising the spacing factor. Quantitative determination on a surface that is undercut and wherein the edges of the voids have been chipped or worn away can provide data that indicate the presence of more air than really exists.

Chapter 5 has several suggestions for the treatment of weak or immature concrete before lapping. Pleau, et al. (1990) suggests soaking the slices (slabs) in water for a few days to promote further curing and cement hydration and thus produce a strengthened paste for lapping. This is a satisfactory procedure for recently cast concretes when the intent of the examination is merely to determine the parameters of the air-void system and its effectiveness may be improved by saturating the soaking water with lime, since soaking in plain water may leach lime and thus weaken the paste. Concretes that contain pozzolans or slag will benefit from gentle moist-heat curing (i.e., at 38°C). However, if the investigation requires a detailed examination of the paste microstructure, the potential altering effects of plain or lime-saturated water must be considered. More important, these and other effects (e.g., dissolving reaction products or liquefying expansive alkali-silica gels) must be considered before using this method in the examination of mature concretes for condition assessment or fault diagnosis.

Each method of producing a finely lapped specimen surface for microscopical study will probably produce different effects on different types of concrete (different w/cm, different kinds of aggregates, different degrees of maturity and deterioration).

In particular concretes in which the shape of the air voids has become distorted (see appendix C), all manner of overlaps and crushing of voids may occur. The operator should be alert and ready to record the data for each void in a logical and consistent manner.

6.3.4 Technician Considerations

The linear traverse and the modified point-count methods are tedious and hard on the eyes. A single determination of the air-void parameters of a concrete by means of the linear traverse method can take up to 7 h , depending on the size and quantity of the voids. A technician cannot spend more than 4 h per day doing this sort of work on a day-to-day basis. Everyone who has tried has found that the ability of their eyes to focus has deteriorated on the following day. The training and keeping of good microscopical technicians can be a major undertaking, requiring tact, skill, understanding, and a flexible schedule of rest periods.

Image analysis systems do not require that the operator be with the equipment after the initial adjustment :thus , eye fatigue and the need to train technicians to perform microscopical analyses are eliminated. The following points are important considerations in the hiring and training of technicians for the microscopical analysis of air-void systems:


6.4.1 Overview

The determination of the abundance of the various types of voids is very useful in concrete research. It can make data available that can change various practices in the mixing and placing of HCC. As an example, it was once thought that the speed of the screed pulling the vibrators through freshly placed concrete did not affect the degree of consolidation. This did not seem logical to some. The Ballenger Construction Company of North Carolina set up a series of test sections of pavement in which the speed of the screed was carefully controlled. A detailed petrographic laboratory analysis of the abundance of the various sizes of voids in the air-void system of 24 cores that had been removed from these test sections showed that there is a good inverse relationship between the speed of the screed and the degree of consolidation (Walker,1972). As a result, the maximum speed of the screed is now limited in many specifications.

The quantitative determination of the abundance of various types of voids can be an important part of the petrographic analysis of a specimen. In the VTRC petrographic laboratory, air voids are classified as entrained or entrapped during routine point-count analyses (see table 15). With equipment designed for this purpose, this determination may be performed concurrently with the determination of the other parameters of the air-void system.

In a finely lapped slab or thin section, the size and shape of the voids can be used as indicators of the void origin and type. The luster and texture of the interior of the voids may sometimes be used in the recognition of voids caused by accumulation of water and passageways for water.The properties on which distinctions may be made between the various types of voids are arbitrary and may vary from one laboratory to the next. Because these distinctions are made on the appearance of a void on the surface of a slice (where the third dimension of the void cannot be seen), many large voids will be classified as entrained voids when they are really entrapped voids. As indicated in figures 58 and 59, a small section through a large void can be indistinguishable in two dimensions from a large section through a small void. A cross section that is larger than the defined maximum for entrained voids must be a section of an entrapped air void or a water-formed void. A significant number of large cross sections ndicate a great number of large voids.

Figure 58. Illustration of various sizes of sections that may be expressed on a randomly placed plane.

Drawing illustrates that the area of a plane intersecting a sphere will vary in size depending on how close it may be to the center of the sphere. The top of the drawing shows similar size spheres intersecting a plane. The spheres are located at different distances from the plane, resulting in a smaller projected area on the plane as the sphere center is gradually moved away from the plane. The bottom of the drawing shows how different size spheres can project the same area on the plane if cut by the plane at the correct distance from the sphere center, as long as the sphere is equal to or larger in diameter than the diameter of the intersected circle.

Figure 59. Two equally spaced arrays of voids.

A drawing shows two separate arrays of spheres. The top set has small spheres; the bottom set has large spheres. Each set is arranged in seven rows across and six rows down. Each array of voids is crossed by a randomly oriented plane. There is the same number of voids in a unit area in each array. The plane touches more voids when the voids are larger than when the voids are small. Only one of the small spheres is intersected by a random plane, but four of the large spheres are intersected.

Each array of voids is crossed by a randomly oriented plane. There is the same number of voids in a unit area in each array. Note that the plane touches more voids (see arrows) when the voids are large than when the voids are small.

6.4.2 Distinguishing Entrapped Air Voids From Water Voids

The petrographer can often distinguish between entrapped voids caused by water pockets and entrapped voids caused by air pockets. The estimate will necessitate careful observation and some extrapolation. In general, the interior surface of an air void will appear smoother, sometimes even shiny. A water void will usually have a dull interior that appears to have had small particles and precipitates deposited onto it. In the case of water-formed voids, the shapes of the bounding aggregate particles are often visible in the interior of the void. Water voids may have an interior showing water movement patterns, may be interconnected bleed-water voids, or may show by the nature of the internal deposits and asperities and by their position that they are water pockets trapped by aggregate particles.

6.4.3 Distinguishing Entrained Voids From Entrapped Voids

The determination of the size breakpoint between entrained and entrapped voids varies from laboratory to laboratory and must be interpreted in light of the method of measurement. For example, if the voids whose maximum chord on the surface examined is less than 1 mm are defined as entrained voids, then some voids whose true diameter is larger but not observable because the diameter is not in the plane of observation will be classified as entrained voids. The petrographer should maintain a clear idea of the meaningof the methods of determination in the size sorting of the voids.

A random line of traverse through HCC has a greater chance of traversing a large void than a small one. The probability ratio is the ratio of their volumes (see figure 60). The calculations detailed in ASTM C 457 use the sums of the chord lengths and the count of the voids regardless of the desirability or relative amounts of the various sizes. If the large voids are not counted and measured as part of the overall determination of the void system (suggested by Sommer, 1979), the control against large voids provided by the determination of the specific surface and spacing factor will have been blocked and the apparent precision of the method spuriously improved(Walker, 1980).

Unless the method and the criteria used to obtain data concerning void size are rooted in statistics, the data are only rules of thumb and valid only when compared with data obtained by the same methods.Calculations can be made on the distribution of the void sizes from chord data if particular assumptions are made concerning shape, heterogeneity, and isometric distribution of air voids.

At VTRC, the diameter of the section of the void as seen on the finely lapped surface examined must be equal to or less than 1 mm for the void to be considered an entrained void. In other laboratories, the length of the chord on the traverse line across the void is the parameter measured. The latter method makes it possible to set up an automatic electronic classification and counting system for entrained versus entrapped voids. In some European laboratories, the chord must be 0.3 mm or less for the void to be considered an entrained void (Wilk, et al., 1974). A void viewed in a lapped surface may be transected by the surface either above or below its true diameter, and there is no known way to measure an actual internal diameter. Efforts have been made to peer into a void to try to get an estimate of the true diameter; however, these efforts may serve only to confuse the issue. In many laboratories, decisions on individual void size are made on the lapped surface as viewed. A large void, more than 1 mm diameter, may be so oriented that the surface examined truncates only a small portion of the void (the extreme top or bottom) when considered from the finely lapped surface. Thus, there will always exist a larger proportion of large voids than can be recognized on the surface examined (see figure 58).

Figure 60. Void system produced by early types of high-range water reducers.

The concrete surface shown can be characterized by a high number of large voids (in the 0.5 to 1.0 millimeter range). Such large voids do not add to the resistance of the concrete to cycles of freezing and thawing but do lower the compressive strength.

Scale is in millimeters.

Such large voids do not add to the resistance of the concrete to cycles of freezing and thawing, but do lower the compressive strength.

6.4.4 Procedures

The procedures given here are for the linear traverse method when the chord lengths are collected by an operator depressing an electronic recording device (Walker, 1988). The point-count method does not survey every void on the traverse during the percentage portion of the examination and, therefore, does not allow a classification of every void. Image analysis procedures are not discussed in detail.

This procedure is a three-step process:

  1. Examine each void when the void& leading edge comes to the index point (usually the center of the crosshairs) and determine which type of void is present: With the wide-angle lenses and a magnification of 100X or less, voids of less than 1 mm in diameter will be completely visible in the field of view. Most voids can be classified at a glance as either entrained, entrapped, or water-formed. When borderline cases occur, use a finely marked metric ruler on the slice, in the field of the microscope, to determine void size (at low magnification, an eyepiece micrometer may be used).
  2. Record the presence of the void and the length of the chord across it in the usual manner for a linear traverse determination (by pushing down a button and holding it down until the trailing edge of the void is at the index point): If three buttons are available for the three types of voids, each with its own tabulating device that separately measures, totals, and counts the voids, depress the appropriate button for the type of void determined in step 1. If the chord lengths are recorded by using only one button and are individually printed on paper and the void encountered is not an entrained void, stop the motion of the traverse stage and mark the paper tape at the chord measurement with a symbol to indicate the classification of the void measured. Continue with the analysis.Repeat from step 1 for each void.
  3. When the analysis has been completed, add the lengths of the chords for each type of void (if not added by the linear traverse device employed) and report the percentage by volume and the count (individual voids per specified inches of traverse) of each in the total concrete.


The major parameters of the air-void system are interdependent. Most concrete specifiers require only that the air-void content be within particular limits. Recently, the trend toward end-result or performance-related specifications has introduced the requirement that the spacing factor be below a particular limit or the specific surface be above a particular limit (Holley, et al., 1999; Kaderbek, Tracy, and Krauss, 2002; Schell and Konecny, 2003). Although one-sided limits on the spacing factor or specific surface do not preclude air-void contents that are too high,compressive strength requirements serve as a limit on excessive air.

Although the parameters described in ASTM C 457 remain those commonly used to describe and evaluate air-void systems, there have been attempts to develop a more meaningful descriptor.Philleo (1983) began the process using void-size distribution to determine the volume of the paste within a particular distance (0.20 mm) of an air void and thus unprotected from cycles of freezing and thawing. Furthermore, the distance at which only about 10 percent of the paste remained unprotected was coined the Philleo factor. Similar concepts are discussed by Attiogbe (1993); Pleau and Pigeon, 1996; Pleau, et al., 1996; and Snyder, 1997. Snyder, et al. (2001) provides a detailed examination of the practical and theoretical basis for the measurement of airvoid systems.

The following are some of the issues the air-void parameters will determine:

Visually apparent "vibrator trails" have been reported from some paving operations, with longitudinal cracks developing in the trail. The distinctive "trails" result from segregation of the concrete as mortar moves into the vibrator's zones of influence. Comparison of airvoid characteristics suggested that the vibrator trails had higher spacing factors than the areas of the slab less affected by vibration (Gress, 1997).

Although the quantity of large voids and specific surface are generally not addressed in current specifications, it is important to consider these parameters and to be able to explain their meaning. In general, it is much easier to talk about a significant quantity of large voids than to explain the mathematical derivation of the specific surface. A percentage of large voids exceeding 1.5 percent is considered to be high. More than 2 percent is considered to be excessive, and explanations for the prevalence of the large voids are sought. Does the concrete appear to have been retempered? Improperly mixed or consolidated? What is the reason?


Previous | Table of Contents | Next

Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101