Guidelines for Detection, Analysis, and Treatment of Materials-Related Distress in Concrete Pavements

Previous | Table of Contents | Next

Chapter 3.  Development of Advisory Guidelines

This chapter briefly summarizes background information used to develop the guidelines.  The chapter is divided into three sections corresponding to each of the three guidelines.  The complete guidelines are presented in Volume 2 of this Final Report.

3.1 Background for Field Distress survey, Sampling, and Sample Handling Procedures

The identification of MRD generally begins with a field distress survey.  During the field survey, the manifestations of observed distress are noted (such as the prevailing crack pattern, the location and extent of the cracking pattern within pavement slabs, the presence of any staining, and the deposition of any exudate). This information can be used to distinguish MRD from other PCC pavement distresses caused by different mechanisms and to suggest possible MRD types.  However, the similar distress manifestations of many of the MRD types make it extremely difficult to accurately and positively identify the specific MRD type and its associated mechanisms based on visible distress manifestations alone. Furthermore, a particular pavement may be affected by several MRD mechanisms, thereby further complicating field identification of the MRD.  Thus, while field surveys can provide indications of the type(s) of MRD and suggest possible mechanisms, the absolute identification of the MRD can only be accomplished through a detailed laboratory analysis of samples obtained from the pavement.

For the most part, very little information is available regarding the field survey of MRD-affected pavements and the associated field identification of MRD types.  This section reviews several common distress survey methodologies to examine the applicability of their procedures for surveying MRD affected pavement while noting how each method differentiates among the various MRD types.  Based on that information, the framework adopted in the first guideline for conducting a field survey of PCC pavements exhibiting MRD is described.  A summary of additional field testing methods that are potentially available to assist in the field identification of MRD is provided.  A discussion of field sampling and handling procedures is then provided,followed by the approach adopted for use in this study.

Review of Selected Distress Survey Methodologies

Over the years, several attempts have been made at developing pavement condition survey procedures and distress identification manuals [Highway Research Board (HRB) 1957; Hveem 1958; HRB 1970; Smith et al. 1979].  However, the increased interest in pavement management activities in the late 1970’s and the launching of the SHRP initiative in 1987 established the need for more standardized procedures and distress definitions.  For example, many highway agencies have developed procedures for their own conditions and applications, and most of these are intended for network-level condition monitoring [e.g., Washington Department of Transportation (WSDOT) 1983; Pennsylvania DOT (PennDOT) 1996, Florida DOT (FDOT) 1996].  As such, there is little (if any) information on identifying specific manifestations consistent with MRD, although many highway agency distress manuals do include D-cracking and map cracking distress types that are consistent with aggregate freeze-thaw deterioration and ASR, respectively.

Of the various distress survey and identification procedures that have been developed, two stand out as currently being the most widely used and accepted: the SHRP Long-Term Pavement Performance (LTPP) distress identification manual (SHRP 1993) and the Corps of Engineers Pavement Condition Index (PCI) survey procedure (Shahin and Walther 1990).  Each of these was developed for a specific purpose, with the SHRP LTPP manual developed for use in the LTPP program and the PCI procedure for network-level pavement management systems (PMS) (although it can also be used for project-level evaluations).  The field survey methodology and the way the respective procedures account for MRD are described in the following sections, along with a brief description of an ACI publication on making a condition survey of concrete in service (ACI 1992b).

SHRP LTPP Distress Identification Manual

Description and Purpose

The SHRP LTPP distress identification manual was developed to provide a consistent basis for collecting distress data for the LTPP program (SHRP 1993).  Because the pavement sites in the LTPP program are located throughout the country and distress will often be collected by different parties, there was an acute need for a uniform basis in distress data collection to ensure the integrity and the compatibility of the data.

Since it was developed for the LTPP program, the manual is oriented toward research applications. That is, it requires that distress data be collected in considerable detail and at high levels of precision in order for the data to be used in characterizing pavement performance and behavior.  However, even though it is a research-oriented manual, it does little to distinguish cracking patterns or physical manifestations common to the various types of MRD.

Survey Procedure

The SHRP LTPP distress identification manual includes an appendix describing the survey procedures, with detailed instructions on how the survey should be conducted.  Blank survey forms and standard symbols for noting the various types of distress are included as part of the appendix. However, since the focus of the procedure is on distress surveys of previously identified and delimited test sites, no guidance or information is provided on sampling or sectioning; a previously selected 150-m-long section serves as the basic sample unit length for the procedure.

The condition survey is conducted with the survey team walking over the length of the sample unit and recording the observed distresses on the field distress form. This is accomplished by using the distress symbols and numbers shown in figures 1-2 and 1-3 to indicate the occurrence of the various PCC pavement distress types. The type, amount, and severity of the distresses are noted on the field survey sheets and then later summarized on another data sheet.  An example of a completed form is shown in figure 1-4, which represents 30.5 m of the sample unit; additional sheets are needed for the documentation of the entire length of the sample unit.  During the field survey, only distresses in the outer traffic lane are noted.

Figure 1-2.  SHRP LTPP distress symbols for jointed concrete pavements (JCP) (SHRP 1993).

Figure 1-3. SHRP LTPP distress symbols for continuously reinforced concrete pavements (CRCP) (SHRP 1993).

Figure 1-4. Example SHRP LTPP survey form (SHRP 1993)

Distress Types and MRD Identification

Table 1-2 lists the specific distress types recognized by the SHRP LTPP distress identification manual for PCC pavements.  Of these distresses, one is related specifically to a durability problem (D-cracking) and three are possibly characteristic of an MRD: spalling, map cracking, and scaling.  The selection of D-cracking as a distress stands out from the other distresses in Table 1-2 in that identifying the distress type in effect presumes a durability problem.  As discussed previously, many of the various MRD types can display similar manifestations. In the case of D-cracking, the observed manifestation is often associated with aggregate freeze-thaw deterioration, which suggests the possibility of misidentification if an MRD is designated as D-cracking based solely on visual appearance.

Table 1-2. SHRP LTPP distress types for PCC pavements (SHRP 1993).

Spalling and scaling distress manifestations can all be associated with MRD in some manner.  However, in the LTPP survey procedure little additional information is sought on the specific characteristics of these distresses.  For instance, the distress identification manual notes that map cracking is limited to the upper concrete surface, but seeks no other information on the pattern of cracking or on any associated staining or material deposits or exudations. Furthermore, because distress symbols are used in noting the distresses, it is difficult to get an accurate representation of cracking patterns and locations.

As with map cracking and scaling, transverse and longitudinal joint spalling are rated based only on their dimensions, with no information sought on associated crack patterns or staining.  And, in assessing joint spalling, no mention is made of the possible associated presence of D-cracking, which gives rise to the possibility that both distresses are counted in a distress survey, when in fact the joint spalling is caused by D-cracking (the severity of the D-cracking is based on the amount of cracking and the amount of loose or spalled material).

Summary of the SHRP LTPP Procedure

A review of the SHRP LTPP distress identification manual and survey procedure indicates that it is limited in its ability to identify MRD.  Four distress types that are representative of MRD are noted in the manual (D-cracking, map cracking, scaling, and spalling), but there are no attempts made to collect additional information (such as staining or crack patterns) that would be useful in identifying MRD.  Furthermore, the D-cracking distress noted in the manual implies a cause of the distress and not just the actual distress manifestation.  These limitations were addressed in Guideline I by adding methods to describe cracking and spalling patterns and locations and descriptors for staining and exudate.

Corps of Engineers PCI Procedure

Description and Purpose

The PCI procedure was developed by the Army Corps of Engineers as a means of determining the current condition of a pavement network (Shahin and Walther 1990).  Extensive work went into the development of a numerical index value that is used to represent the pavement’s structural integrity and its surface operational condition based on the observed distress.  The resulting index, the PCI, ranges from 0 (failed pavement) to 100 (perfect pavement) and accounts for the types of distress, the severity of the distresses, and the amount or extent of the distresses. The associated effects of these factors are combined into a composite PCI value through established "weighting factors" so that it more accurately reflects the overall performance of the pavement (Shahin and Walther 1990).

The PCI procedure is intended primarily for network-level pavement management purposes, not only in documenting the current condition of the pavement but also in developing prediction models to forecast future pavement condition (Shahin and Walther 1990).  However, the methodology is sufficiently comprehensive and flexible that the PCI procedure can also be used in project-level analyses.

Survey Procedure

The PCI procedure first requires that the pavement section be divided into smaller subsections for inspection purposes.  These smaller subsections are called sample units and, for concrete roads, the recommended sample unit size is 20 + 8 slabs (if the joint spacing of the pavement is greater than 7.6 m, imaginary joints less than 7.6 m apart are assumed) (Shahin and Walther 1990).  The sample unit size (i.e., number of slabs)should be selected as a matter of convenience; for example, if a pavement section consists of 200 slabs, then 8 sample units consisting of 25 slabs provide a convenient means for inspection.  For data management purposes, each of the individual sample units is numbered from 1 up to the total number of sample units in the section; in the previous example, the sample units would be numbered from 1 to 8.

The type of analysis (network- or project-level) will determine the number of sample units to be inspected.  For project-level analysis, a greater number of sample units should be inspected because of the need to accurately assess pavement performance and to prepare accurate work plans and contracts.  The minimum number of sample units (n) that must be surveyed in a project-level analysis in order to obtain a reasonable estimate of the pavement section’s PCI is determined from the following formula (Shahin and Walther 1990):

where:

n = Minimum number of sample units to inspect.
N  = Total number of sample units in the pavement section.
s = Standard deviation of the PCI between sample units (generally assumed to be 15 for PCC pavements).
e = Allowable error in the estimate of the section PCI at 95 percent confidence (generally set to 5).

When the total number of sample units (N) is less than 5, then it is recommended that all of the sample units be surveyed.

Once the minimum number of sample units has been determined, the next step is to determine which specific sample units should be inspected.  This is accomplished by a stratified sampling procedure that first determines the sampling interval (i) as follows:

 where:

i  = Sampling interval.
n = Minimum number of sample units to inspect.
N = Total number of sample units in the pavement section.

The first sample unit to be surveyed is selected at random between sample unit 1 and the sampling interval i (e.g., if i = 3, then the first sample unit to be surveyed is randomly selected between 1 and 3).  After the first sample unit is selected, the next sample units to be surveyed are selected by adding the sampling interval to the first sample unit number; for example, if the first sample unit is 2 and the sampling interval is 3, then the sample units to be inspected would be 2, 5, 8, 11, and so on up to the total number of sample units (N).

As with the LTPP procedure, the inspection is conducted by the survey crew walking over the length of each sample unit and recording the distress found in each slab.  An example form that can be used to conduct the survey is shown in figure 1-5.  This form is for one sample unit only; additional forms are required for each sample unit within the section.  The pavement slabs in the sample unit are drawn in the space on the right portion of the form.  However, according to the procedure, the incidence of distress is not drawn in, but rather is noted to occur in the slab using the number code shown in figure 1-5.  Thus, it is not clear exactly where on the slab the distress is occurring.

Distress Types and MRD Identification

The specific distress types recognized by the PCI procedure are listed in table 1-3.  This table shows that the distress is recorded on a slab-by-slab basis, and that actual distress quantities are not obtained (e.g., linear measures of cracking); rather, only the presence of the distress is noted.  Furthermore, if more than one of a particular distress type occurs within a slab, only the highest severity is recorded so the presence of a less severe distress of that type is ignored.  The data obtained in this fashion are useful in gaining an overall perspective on the performance of the pavement, but the end result is that true distress quantities are not collected.

Although never specifically stated, the PCI procedure is geared to jointed concrete pavements. No mention is made regarding how a CRCP should be surveyed (since the survey is based on a slab-by-slab evaluation procedure). Furthermore, none of the distresses listed in table 1-3 is specifically for CRCP structures; in fact, the punchout distress listed in table 1-3 refers to punchout structural failures occurring between a transverse joint and a nearby crack.

The PCI procedure recognizes one manifestation directly attributable to MRD (D-cracking), clumps three other distress manifestations consistent with MRD together as a single distress type (scaling/map cracking/crazing), and breaks out another distress manifestation common with MRD (spalling) into two varieties (corner and joint).  As with the SHRP LTPP procedure, the designation of the D-cracking distress represents a diagnosis, which could lead to the misidentification of the distress.  And the grouping of the scaling/map cracking/crazing

Figure 1-5. Example PCI sample unit inspection form (Shahin and Walther 1990).

Table 1-3. PCI distress types for PCC roads (Shahin and Walther 1990).

manifestations into a single distress appears to have been done as a matter of convenience and perhaps in recognition of their perceived similar effect on pavement performance; however, in grouping these manifestations together, the PCI procedure makes no attempt to distinguish between the unique causes of each manifestation.  Furthermore, the inclusion of two spalling manifestations appears to complicate the identification of possible distress mechanisms, and no attempt is made to distinguish spalling associated with D-cracking with other joint spalling.  Finally, as with the SHRP LTPP procedure, no attempt is made to capture associated distress manifestations (crack patterns, staining, material depositions) that may be useful in identifying MRD.

The primary goal of the PCI survey is to obtain a numerical index value representing the current condition of the pavement, and this is accomplished through deduct charts used to assess the relative effect of each distress type on the performance of the pavement.  Thus, specific charts for each distress type are needed for this evaluation, meaning that only the distresses listed in table 1-3 are assumed to affect pavement performance.  In order for other distress types to be included, deduct charts would have to be developed and incorporated into the procedure.

Summary of Corps of Engineers PCI Procedure

The PCI procedure is intended more for network-level pavement condition surveys where the presence of the various distress types is important in assessing the overall condition of the pavement section.  The procedure incorporates detailed sampling plans so that representative measures of performance are obtained.  However, the PCI procedure only recognizes three distress types or manifestations consistent with MRD (D-cracking, scaling/map cracking/crazing, and spalling) and makes no attempt to collect additional information that may be useful in identifying mechanisms and causes of distress.  Furthermore, the procedure is geared toward jointed concrete pavements and no guidance is provided on surveying CRCP structures.

ACI Guide for Making a Condition Survey of Concrete in Service (ACI 1992b)

Although not a detailed pavement survey procedure, this ACI document does provide some useful information in evaluating the performance of a concrete pavement.  For example, it provides a comprehensive checklist of information to be considered during a condition survey.  Although the checklist is generic to all concrete structures and not all information given in the document will be needed for each survey, a review of the various items is useful to ensure a thorough survey. A listing of the major heading items found in the checklist are summarized in table 1-4 (the complete checklist contains many sublevel items that are not repeated here).

The ACI document also provides a detailed listing of concrete distress manifestations, including a description of the manifestation and photographs of the distress.  The listing is very comprehensive as it includes distresses for all concrete structures, and introduces many distress types not normally encountered in a pavement distress identification manual.  It also includes specific distresses that fit into the MRD category; these distress items, and their associated definitions, are (ACI 1992b):

• Craze cracks – fine, random cracks or fissures in a surface of concrete.
• D-cracking – a series of cracks in concrete near and roughly parallel to joints, edges, and structural cracks.
• Pattern cracking – fine openings on concrete surfaces in the form of a pattern resulting from a decrease in volume of the material near the surface, an increase in volume of the material below the surface, or both.
• Efflorescence – a deposit of salts, usually white, formed on a concrete surface, the substance having emerged in solution from within the concrete and subsequently precipitated by evaporation.
• Exudation – a liquid or viscous gel-like material discharged through a pore, crack, or opening in the surface of concrete.
• Scaling – local flaking or peeling away of the near-surface portion of hardened concrete.
• Spalling – a concrete fragment, usually in the shape of a flake, detached from a larger mass by a blow, by the action of weather, by pressure, or by expansion within the large mass.
• Discoloration – departure of color from that which is normal or desired.

Table 1-4. Simplified ACI survey checklist (ACI 1992b).

Other than the D-cracking definition (which, like the other procedure reviewed, represents a distress diagnosis), these definitions represent an attempt to record and classify many of the visible distress manifestations associated with MRD.  These are used as the basis for visible distress items included in the field survey guidelines.

Applicability of Available Distress Survey Procedures for the Assessment of MRD Affected Pavements

Although a variety of distress survey procedures and distress identification manuals have been prepared over the years, none fully serve as a comprehensive document for the field survey and evaluation of concrete pavements with MRD.  Two primary distress survey procedures—the SHRP LTPP procedure and the COE PCI methodology—are lacking the ability to identify MRD manifestations.  A guide produced by the ACI provides some useful information on data elements needed for a distress survey and on some MRD-associated distress manifestations, but does not describe survey procedures.  Together, however, these documents provide the foundation for what is required in an MRD-related survey.

Procedures Adopted for a Distress Survey of MRD Affected Concrete Pavement

The goal of an MRD pavement survey is to collect all information needed to assist in the identification of the type of MRD.  Although the MRD survey may provide strong indications as to the causes of the MRD, it is recommended that final determination be made based on the results of extensive laboratory testing and evaluation.  Thus, it is imperative that the field survey collects all information necessary for later testing and evaluation.

Based on information obtained from the distress survey procedures previously presented, and in anticipation of the needs for a detailed MRD survey, the approach presented in Guideline I contained in Volume 2 of this Final Report was developed.  Since the method is presented in its entirety in Volume 2, it will not be duplicated here.  Instead, only a brief summary of some salient points is provided below.

Records Review

The first part of the MRD survey should begin with a preliminary records review.  This should include all information needed to adequately identify and assess the pavement in the field, such as project location, structural design (slab thickness, base type, joint design, subgrade support, and so on), mix design and materials information (including aggregate type, aggregate source, and cement type), and traffic data (traffic volumes, truck volume composition).  Forms for recording this information are provided in Guideline I.

Shoulder Survey and Sample Unit Selection

Immediately before conducting the distress survey, a shoulder survey of the entire project should be conducted in both directions when both directions represent the same project (structural design, mix design, construction, and so on). The purpose of the shoulder survey is to assess the overall performance of the project and to assess the extent and uniformity of MRD.  Any directional differences in performance or distress manifestations should be noted by location to assist later in the selection of sample units.

After the shoulder survey has been conducted, the project must be laid out so that sample units may be selected for survey.  Sample units are intended to be representative of the entire project and are used so that a complete evaluation of the entire project (which typically is several kilometers long and would require a substantial amount of time and effort to evaluate) is not necessary.

The number and location of the sample units will depend largely on the overall condition of the pavement.  This is because the purpose of the field evaluation is to collect data and material samples necessary to aid in the identification of the type and causes of MRD.  In other words, a problem has already been identified in the pavement and a detailed evaluation is being conducted to determine the causes of the distress.  This is in contrast to a network-level pavement survey in which statistical sampling procedures are employed requiring that typical, representative sample units be evaluated to make statistically valid assessments of overall pavement performance.

The results of the shoulder survey are used in the selection of the location and number of sample units to be surveyed.  Generally, if the shoulder survey indicates that the extent and severity of MRD are fairly uniform throughout the length of the project, then a single sample unit selected at a representative location is believed to be sufficient.  If, however, the shoulder survey indicates a range of MRD severity levels, or areas where there is no distress and other areas where there is severe distress, then surveying sample units in each of these areas may be appropriate to determine why one area is performing differently than another.

Pavement Distress Survey

It is recommended that the distress survey be conducted in general accordance with the procedures and distress definitions found in the SHRP LTPP distress identification manual (SHRP 1993) with several modifications to address the previously described limitations. These include the removal of D-cracking as a distress manifestation and the addition of procedures to provide a detailed description of the location and patterns associated with cracking and spalling.  Additional descriptors were added for staining and exudate.

However, there is no need to draw every crack associated with the MRD acting on a pavement. Such documentation is time-consuming and does not contribute to the identification of the distress.  Instead, the general area of MRD will be noted and the physical manifestations characterized on specific MRD survey forms.

The limits of the sample unit are first identified and laid out; the distress survey then proceeds from the beginning of the project.  All distresses occurring within the sample unit are noted and, where appropriate, drawn approximately to scale on the general distress survey form.

The primary distress survey should generally be conducted in only one lane (e.g., the outer lane of a multi-lane facility or one direction of a two-lane roadway). However, in viewing the adjacent lane from the primary lane, any distresses that are clearly visible should be marked on the distress forms to provide some overall indication of the presence of distress in those lanes. 

Although identification of the MRD will not be made during the field surveys, attempts should be made to characterize to the greatest extent possible the visible signs of MRD, including:

• Cracking pattern (location, orientation, extent, crack size).
• Staining (location, color).
• Exudate (presence, color, extent).
• Scaling (location, area, depth).
• Presence of vibrator trails.

The typical MRD cracking pattern will also be recorded during the field surveys. This will be done for only one typical slab to illustrate a schematic of the prevailing crack pattern.  However, to facilitate the field survey, the MRD cracking is not drawn on the distress survey forms (but the general area of distress is noted).  Detailed quantities of MRD are entered on another form.

As mentioned, several different data collection forms were developed to record observed distress manifestations.  These include:

• A general information form that describes the location and characteristics of the project. Most of this form can be completed prior to visiting the field site.
• A field distress survey form similar to that used in the SHRP LTPP program (SHRP 1993).  All pavement distresses are to be recorded on this form, in addition to the general area of occurrence of the MRD on the pavement.
• Two additional forms were developed to help characterize the MRD.  One of the forms provides space to draw the typical MRD pattern observed on a single slab, including the crack pattern and the presence of staining and any exudation.  The second form is then used to categorize the MRD distress in terms of its cracking pattern (location, extent, size), staining (location, color), material deposits or exudation (presence, color, extent), and scaling.

Photo/Video Documentation

At the conclusion of the distress surveys,a complete photo or video summary of the sample unit is performed.  The purpose of this is to fully and completely document the performance of the pavement, as well as to record the prevailing foundation and drainage characteristics of the roadway.  Because some MRD “disappear” under poor lighting conditions, the photo documentation should be obtained in such a manner that the distresses are clearly visible.

Field Chemical Testing

Several test procedures are available that might be useful in the field evaluation of pavements suspected of suffering ASR.   These tests use chemicals that are applied to a pavement sample in the field to indicate whether ASR is present.  Two tests for the determination of ASR distress are described below; currently, no other chemical tests are available for the preliminary identification of MRD in the field.

Gel Fluorescence (Uranyl Acetate)Test

The gel fluorescence test was first developed as a laboratory procedure at Cornell University and was later modified for field use (Natesaiyer and Hover 1992).  In this test, the gel-like reaction products of the ASR reaction are identified by staining them with uranyl ions; these ions have a characteristic bright greenish-yellow fluorescence when viewed under UV light.  Other normal compounds in concrete are not typically stained by the uranyl ions. The test can be conducted in the field, but proper guidelines must be followed because of the radioactivity of the uranyl acetate compound.  The conduct of this testing is included as an annex to ASTM C 856.

Although preliminary results with this test have been encouraging, a positive identification by this technique does not necessarily mean that ASR is present (Farny and Kosmatka 1997).  For example, some aggregates fluoresce naturally and some carbonated concrete can show a mild fluorescence in the areas of carbonation(Natesaiyer and Hover 1992).  Thus, the test is not intended to be a stand-alone method for diagnosing ASR, but rather an ancillary test to more definitive laboratory testing (Farny and Kosmatka 1997).  Furthermore, the test requires experienced and knowledgeable technicians for correct interpretation of the findings.

Additionally, because of the perceived hazards associated with uranyl acetate and the disposal of the stained concrete, this test is not currently recommended for use in field surveys of pavements with potential MRD problems.

Sodium Cobaltinitrite/Rhodamin B Test Method

In recent years, a test developed at the Los Alamos National Laboratory has become available for the field identification of ASR gel (Guthrie and Carey 1997).  The advantages of the test procedure are that it uses non-radioactive chemicals (sodium cobaltinitrite and rhodamin B) to stain ASR reaction product, the stains are specific for ASR gel, and the staining is visible in normal light.

Due to the recent introduction of this test, little information is currently available regarding its effectiveness to make positive identification of ASR in the field.  It does promise to be simpler and safer to run, but additional laboratory testing is required to confirm the both presence and extent of ASR.  Thus, while consideration should be given to the further evaluation of this technique, it is not recommended as part of an MRD field survey at this time.

Assessing the Likelihood of MRD

Once the visual assessment is completed, the analyst is directed to a flowchart to assess whether the observed distress is likely the result of an MRD.  The flowchart asks questions regarding the nature of the cracking, whether staining and/or exudate are present, evidence of expansion, or whether the distress is progressive.  An affirmative answer suggests that an MRD may be at work, thus a laboratory analysis should be conducted necessitating the acquisition of samples.

Review of Sampling and Handling Methodologies

This section summarizes currently accepted and recommended practices for obtaining and handling samples of existing PCC pavements for the purpose of diagnosing MRD. It also discusses techniques for retrieving, handling, and storing the specimens in a manner that both maintains the physical integrity, and chemical composition and preserves the nature and structure of the microstructure present.

The importance of effective sampling and testing procedures should not be overlooked.  The basic principles and approaches on the topic have been developed and disseminated by the American Society for Testing and Materials (ASTM) Committee E11 on Statistical Methods (ASTM E 105; ASTM E 141) and others (ASTM 1948; Tanner and Deming 1949; Slohim 1960; Montgomery 1985; Tanner 1961; Cochran 1963; Bicking 1964; Bicking 1965; Duncan 1967; Visman 1969; ASTM 1973; ASTM 1975; SHRP 1990). 

Additional information on sampling procedures for concrete and concrete ingredients are found in various ASTM standards (C42, C823 and C 856) and other handbooks and manuals such as those by the Waterways Experimental Station (WES) (1949), ACI (1981), and the United States Bureau of Reclamation (USBR 1975).  Useful references to existing ASTM standards dealing with probability sampling concepts and the sampling of various concrete component materials and hardened PCC pavements are presented in table 1-5.

Table 1-5. ASTM sampling standards for concrete and concrete-making materials.

As few of these practices are not directly applicable to sampling concrete pavements for the investigation of MRD, they will thus not be discussed in detail in this report.  Instead, the information presented in this section will focus on the unique sampling requirements to obtain PCC specimens for subsequent laboratory testing and analyses to provide accurate and representative indications of the presence of MRD.

Approach Adopted for Sampling and Handling MRD Affected Pavements

This section discusses the preparation activities, sampling and handling procedures, and shipping procedures recommended for use when sampling a concrete pavement suspected of being affected by MRD.  The sampling procedures presented are designed only for the purpose of identifying MRD, and are not to be used in the course of construction quality control.

General Approach to Sampling

ASTM C 42, Standard Test Method for Obtaining and Testing Drilled Cores and Sawed Beams of Concrete, governs the retrieval of cores or the sawing of beams from hardened concrete.  However, this standard does not mention how to decide at what point or points such specimens should be secured.  ASTM C 823, Standard Practice for Examination and Sampling of Hardened Concrete in Construction, describes this, as well as many other aspects of sampling hardened concrete, in considerable detail.  Other sources of information include Abdun-Nur (1970) and ACI 214, Evaluation of Results of Tests Used to Determine the Strength of Concrete.  It is noted that standardization of the sampling process is difficult because it involves a great deal of judgment that is often difficult to standardize.

If the purpose of sampling hardened concrete is to assess overall properties and variability, it should be done randomly and objectively.  The selection of samples in such cases should not be biased by procedures that intentionally select either the best or poorest materials.  However, sampling in the course of investigations of MRD are distinctly different from quality control/assurance testing in that samples should be obtained to investigate unusual or extreme conditions or features that will aid in the identification of the causes of distress or failure of the concrete.  It is this type of sampling approach that has been adopted.

In hardened concrete being analyzed for MRD, specimens could be obtained by random sampling from within a panel to determine the variability of the development of MRD within the panel.  Or sampling could be concentrated in areas of known or suspected MRD development in order to characterize the conditions that have led to that development.  The latter approach is recommended due to the time and expense involved in analyzing concrete in the laboratory.  Thus it is extremely important that sampling in the field (including the development of an appropriate sampling plan) is not a simple matter and must not be delegated to the untrained or careless, or relegated to a laborer for the sake of convenience (Abdun-Nur and Poole 1994).

The sampling plan must be selected with consideration of the results of field condition surveys that have been discussed previously in this report.  The visual inspections and review of the construction records will indicate whether or not the concrete is believed to be in similar condition and of similar quality.  The sampling plan must also reflect the full prior knowledge of the laboratory-testing program that will be performed, as is discussed later in this report.  For the purpose of an MRD investigation, evidence of distress will be obtained from the results of the visual assessment conducted in accordance with the guidelines.  Specimens for use in the laboratory investigation should be selected based on “targeted sampling,” since they are being used solely to characterize the type and extent of distress.

Data collected during the field survey will be used in the selection of coring locations.  For example, information detailing the location and characteristics of the project should be reviewed by the person(s) developing the sampling plan as well as by the persons that will be executing the sampling plan prior to their visit to the field.  This information should include a detailed sketch or drawing of typical observed MRD, including the crack pattern and the presence of staining and any exudation.  It may also include categorization of the MRD distress in terms of its cracking pattern (i.e., location, extent, and size), staining (i.e., location and color), and material deposits or exudation (i.e., presence, color, and extent).

The first step in determining the number of coring sites for a given project is to use the results of visual survey to assess the overall continuity of MRD within the project length.  If the distress is uniformly distributed over the entire project length, one or two coring sites are considered sufficient to accurately characterize the type and extent of MRD.  If observed distress severity and/or extent vary along the project length, at least one coring site must be selected for each observed condition level. 

Coring sites should always correspond to the sample units inspected during the visual assessment.  In this way, the detailed crack maps can be used to initially select specific core locations.  An adequate number of cores must be obtained to represent the condition of the pavement, yet the number must be restricted due to practical restraints on resources and the specific needs of the investigation.  In the guidelines developed in the course of this study, each PCC pavement coring site is 150 m long.  It is recommended that a minimum of four cores be obtained per coring site.  Specific details concerning recommended core locations are presented in the guidelines contained in Volume 2 of this Final Report.  The following presents a brief summary of the procedures.

The quantity or number of concrete samples and the dimensions of the pieces retrieved in the sampling operation must conform to the stipulations of the tests to which the samples will be subjected.  In some instances, two or more tests may be performed on a single sample, provided that the preceding tests do not modify the properties of the concrete to be evaluated by subsequent tests as described later in this report.  One example of acceptable successive testing is the determination of air void content by linear traverse or modified point count followed by petrographic examination.  An unacceptable sequence would be to stain a specimen using uranyl acetate and then attempt to analyze it using the scanning electron microscope (SEM).

The recommended coring locations are selected to provide more than enough specimens for analysis while preventing the acquisition of too many specimens, which leads to waste. If the distress in a JCP is concentrated at the joints, a minimum of four core samples (recommended 150-mm diameter) is required if the pavement is doweled.  Two of the cores are positioned directly over the joint, with the center of the 150-mm core barrel offset 50 mm from the middle of the joint.  In this way, the core can be cut across the joint in the direction of traffic to analyze the concrete at the interface to a distance of 125 mm from the joint.  One of these two cores is obtained directly over a dowel bar located between the wheel paths to assess the corrosion of embedded steel.  The second is obtained between dowel bars in an area affected by the MRD.  The third core is obtained at a corner location, located approximately 0.3 m from both the transverse joint and outside lane stripe.  This core will be used to assess the extent of deterioration away from the joint, and thus may or may not be in an area visibly affected by MRD.  The final core is positioned in a center slab location free of observable distress to determine if distress is present at a microstructural level even though it is not yet visible on the surface. 

For a JCP suffering from MRD that is not concentrated at the joints, five cores are required.  The first four are located the same as previously described.  The fifth core will be obtained in the slab interior over visible deterioration to assist in identifying its cause. This deterioration may be a crack (as is common in map cracking), spall, or delamination.  In some cases, this core may be pulled from deteriorated vibrator trails.  It is noted that, as stated previously, the fourth core should be taken from a visually non-distressed area.  In some severe cases of MRD, a non-distressed area might not be visible, in which case the core can be taken from a deteriorated area or not at all.

A similar pattern was developed for CRCP.  One notable difference is that instead of obtaining a core through a dowel bar, it instead should be taken through the reinforcing steel at a transverse crack.  All other cores should be obtained between the steel, if possible.  If corrosion of embedded steel is the specific distress being investigated, additional cores can be obtained through the steel at different locations at the discretion of the investigator.

It is noted that these recommendations present the minimum number of cores needed for the evaluation.  In some instances, additional cores would be useful, particularly if certain unusual features were observed.  Field crews should be instructed to take additional samples if unusual features are observed, carefully noting the location and the feature of interest.

All core samples taken should include the full thickness of the concrete because the development of cracking, deterioration of the cement paste, progress of cement-aggregate reactions, and other features may vary significantly with depth.  The diameter of core specimens for MRD identification is not critical as long as it is sufficient to provide a representative section of the concrete structure once it is prepared for examination.  In general, the use of 150-mm-diameter cores is preferable because it provides more material for evaluation and archiving, although 100-mm-diameter cores may suffice.  Cores that are intended for use in petrographic examination must be unaltered cores that have not been subjected to strength testing or other destructive processes.

Pavement surfaces should be marked at the coring location to indicate the direction of traffic before coring or sawing.  The markings should be clearly visible and easily interpreted after the coring operation is complete.  Procedures that cause mechanical or thermal damage to cores should not be used in the retrieval of specimens and samples must be protected from contamination, damage, and other processes that might change the character of the material being examined or tested.  Possible sources of contamination include other layers at the same site as well as materials carried by coring rigs and other equipment from other sites.  It is desirable that the surface of all cores or beams be rinsed with fresh water to remove coring or cutting slurry prior to wrapping or packing for transport and/or shipping.

Information concerning the location, manner of retrieval, and field condition of all hardened concrete samples should be recorded for future reference. Self-adhesive labels that clearly identify the core (including the coring site, slab number, core location, and date of coring) should be prepared and affixed to each sample. A sample hardened concrete sampling log sheet and sample identification label have been provided in the guidelines contained in Volume 2 of this Final Report.

Handling and Storage of PCC Field Samples

Simply establishing sampling plans, methods, and procedures are not enough to ensure that proper, undamaged samples will reach the laboratory for testing.  For example, Wills (1964) has shown the undesirable effects on concrete properties caused by the use of bags contaminated with sugar, flour, or treated with chemical preservatives.  Although many problems concern fresh concrete specimens, proper protection, packaging, and care of hardened concrete cores in the field and during shipping are still important.

In general, samples must be properly identified and shipped in clean, strong containers.  If the moisture content of the sample is important, the container must be moisture-tight.  Containers should always be clean and free of any potential contaminants.  Care should be taken to prevent damage to concrete specimens during transport and shipping.  Intact cores or large core fragments can generally be transported from the field and shipped over short distances without damage when wrapped in plastic bubble wrap packing material and placed in appropriately sized plastic cylinder molds with lids.  Similarly, sawed beams can also be wrapped in plastic bubble wrap packing material and shipping tape and placed in protective tubs or containers for transport from the field and shipping over short distances.  Specimens that will be shipped by common carrier should be afforded additional protection from damage. One other interesting consideration is that specimens obtained in an area under United States Department of Agriculture quarantine must only be shipped in accordance with the relevant regulations if contaminated soil is present.  Specific recommendations are provided in the guidelines provided in Volume 2 of this final report.

There are generally no limits on the duration of storage for hardened concrete specimens, provided that the desired storage conditions (if any) are maintained with respect to specimen moisture and temperature, exposure to atmospheric conditions, and so on.  However, the condition and composition of the specimen must closely represent the condition of the field concrete at the moment of retrieval. Significant delays in testing may reduce the usefulness of the test results.

Summary of Background for Field Distress Survey, Sampling, and Handling Procedures

This section of the report has reviewed the background information and quickly summarized the approach presented in the first guideline on field distress survey, sampling, and handling procedures.  Several pavement distress identification and survey procedures were reviewed to determine their applicability for use on MRD pavements.  Current procedures provide a basis for evaluating MRD affected pavements, but are lacking in ways of identifying and collecting distress manifestations specific to MRD.  Based on the current procedures and recognizing the specific needs for characterizing MRD, a framework for the components of an MRD survey was developed and presented in this chapter.  A review of field chemical staining tests was also conducted, but it is recommended that no additional testing should be included in the guidelines at this time.

A review of procedures for obtaining and handling field samples of PCC for use in identifying the presence of MRD in PCC pavements was also presented.  The procedures adopted in the guideline were then reviewed. The section describes a systematic approach for determining the numbers of samples required and appropriate locations from which to retrieve those samples, as well as techniques for retrieving, handling, and storing the specimens in a manner that maintains the physical integrity and chemical composition of the material. Compliance with these guidelines will help to ensure that subsequent laboratory tests and analyses provide accurate and representative indications of MRD presence or potential.

3.2 Background for Laboratory and Data Interpretation Procedures

This section of the report provides the background used in the development of the second guideline that includes laboratory methods and data interpretation procedures.  It provides a review of the available procedures followed by a summary of the methodology adopted in this work.

When MRD is suspected of playing a role in the premature deterioration of concrete, laboratory tests are almost always required to help understand the underlying mechanisms at work.  In reviewing the various types of MRD, it is clear that the distress mechanisms involve physical and/or chemical processes that occur between the concrete and its environment.  These processes ultimately lead to changes in the concrete microstructure, which may in turn affect the durability of the concrete.  The relationship between material characteristics and microstructure is not unique to concrete.  The study of material microstructure forms the basis of materials science and engineering.  In the same vein, the typical laboratory methods used to characterize concrete’s microstructure are the same as those used to characterize the microstructure of other materials.

Common References, Standards, and Practices

A number of reference materials exist that provide an excellent presentation of the current body of knowledge related to methods of performing diagnostic tests on concrete.  Walker (1992) developed an excellent summary of the fundamental methods of concrete analysis, including chapters on equipment and materials, sample documentation, preparation of specimens, and the observation of cracks, voids, and paste using the stereo and petrographic optical microscopes.  In addition, other concrete constituents and parameters are discussed including mineral admixtures and w/c determination.  AAR is discussed and numerous photographs and micrographs are included to illustrate the different types of distresses.  Various types of optical microscope methods are presented, including a section on fluorescence microscopy techniques.  Numerous color pictures of various concrete specimens are included and overall, detailed steps on the fundamental tests for concrete analysis are presented. 

Roy et al. (1993b) published a detailed description of the fluorescence microscopy method, including numerous micrographs showing various features as seen in fluorescence microscopy and the effects of poor sample preparation.  Detailed instructions on the method of gathering and the interpretation of data are included. The textbook by St. John et al. (1998) is an excellent reference that discusses concrete petrographic techniques.  It provides information and references along with details of specific techniques.  The text also discusses defects in, and deterioration of, concrete and provides a wide range of micrographs illustrating common microstructural features. Also, details of optical and chemical properties of phases found in concrete are presented.

A large number of ASTM standards address laboratory practices relative to concrete.  In general, three of these standards directly address the examination of hardened concrete and should serve as a basis for any laboratory examination of concrete exhibiting MRD.  The applicable standards are ASTM C 457 Test Method for Microscopical Determination of Parameters of the Air-Void System in Hardened Concrete, ASTM C 823 Practice for Examination and Sampling of Hardened Concrete in Constructions, and ASTM C 856 Practice for Petrographic Examination of Hardened Concrete.  One important impact of these standards is in the case of litigation where the credibility of laboratory results will be based in part on adherence to these accepted standards.  Unfortunately, as detailed procedures, these ASTM standards often fall short, forcing the analyst or engineer to make decisions about methods without sufficient background information.

The scope of ASTM C 823 is as follows: "This practice outlines procedures for visual examination and sampling of hardened concrete in constructions.  Reference is made to the examination and sampling of concrete in prefabricated building units, pre-cast products, and laboratory specimens." This practice presents a procedural plan for examining concrete that includes defining the objective, stating clearly the purpose, and defining the scope of the examination.  The objective is usually the same; try to obtain information that can describe the condition of the concrete or the construction.  However, in forming a clear picture of the objective, also keep in mind other constraints such as time, budget, or the interest of an another party such as a contractor.  In this latter case, sampling should be done judiciously in case disagreements turn into legal action.  The purpose of the investigation can vary widely, including predication of service life, identification of distress, assessment of performance, or support for legal action.  The type of sampling required for each purpose may vary.  For example, MRD may be identified in a broken section removed by a shovel.  Scientific evidence of the same distress for court action would require extensive coring and rigorous adherence to sampling statistics.  The scope of the examination may be limited to one panel or 10 miles of a highway.  In either case, the sampling approach required varies.  It also discusses methods of performing a Preliminary Investigation indicating the need to verify the existence of a problem, describe its nature, and estimate its extent.  The remainder of ASTM C 823 describes considerations and procedures for sampling under various conditions.  Given the wide variety of possible Procedural Plans, ASTM C 823 should be referred to based on the specifics of each examination required.

ASTM C 457 is the standard method describing the procedure for determining the hardened air content of concrete.  It results in calculated values of specific surface, frequency, spacing factor, and paste-air ratio for the air-void system of the concrete.  The test method discusses equipment, procedures, and calculations for making these measurements.  Recommendations of minimal sample size are provided.  Note that these sampling requirements may affect the size or number of core samples required from a pavement.  This method does not discuss treating the case of an air-void system filled with secondary deposits.  This may be an important factor in concrete performance.  It is recommended that anyone performing ASTM C 457 report both the original air content and an estimate of existing air content after infilling.  This is a more tedious analysis that requires multiple analyses of the specimen.  During the first examination, the analyst marks the air void at its original edge and in the second examination, the analyst marks at the edge of the filling rim, skipping completely filled voids.  This is an effective way to quantify secondary deposits in air voids.  Also ASTM C 457 does not produce a measure of the air-void size distribution.  This characteristic is important because two concretes with the same spacing factor may have quite different freeze-thaw resistance, depending on the size and relationship of the air voids to each other (Hover 1993).

ASTM C 856 outlines many of the procedures required for the petrographic examination of concrete.  This Practice should be used as the framework of any laboratory practices developed.  Included in ASTM C 856 are sections on qualifications of petrographers, purposes of examination, required apparatus, sampling, sample preparation procedures, microscopical examination, and suggested diagnostic features to examine in concrete.  Also, a table is included that summarizes the diagnostic optical properties of common phases found in PCC.  This table is very useful for performing petrographic microscope examination of thin sections and grain mounts.  Although ASTM C 856 is very useful and comprehensive, it falls short of relating observed diagnostic features to specific MRD types.  This is to be expected given the general applicability desired in a practice of this type.  It is impossible to address all possible combinations of concrete failures that might occur.  However, the general observations made in ASTM C 856 are sound and applicable to examination of MRD in concrete pavements.  It is up to the analyst to decide how to interpret the results obtained.

To glean useful information from a laboratory analysis of concrete, proper sample preparation is a very important consideration.  Once specimens are selected and prepared, the stereo optical microscope (OM) and the petrographic optical microscope are the most versatile and widely applied tools for diagnosing causes of MRD.  Electron microscopy is becoming more prevalent, especially for chemical identification of reaction products and other secondary phases using energy dispersive spectroscopy (EDS).  The conventional SEM (CSEM) operates with the sample in a high vacuum (10^-6 torr), leading to accurate microanalysis but desiccation cracking of the concrete.  The environmental SEM (ESEM) operates with the sample in a very low vacuum (2-20 torr), leading to poorer microanalysis but little to no desiccation cracking of the concrete.  The low vacuum SEM (LVSEM) provides a reasonable compromise to the problems of the CSEM and ESEM, operating at a vacuum between the extremes of the two.  X-ray diffraction (XRD) has been used in very limited instances for analysis of hardened concrete produced outside the laboratory.  Analytical chemistry is used primarily to determine w/c.  All of these procedures and techniques are described in the following sections.

Sample Preparation

French (1991) has presented a comprehensive review of cement petrography and detailed steps for sample preparation, analysis, and interpretation.  For concrete petrography with an OM or SEM, specimens may be prepared as polished thin sections, polished plates, or freshly broken surfaces.  In a thin section, a section of concrete at least 10 mm thick is cut from a slab and impregnated with low viscosity epoxy in a vacuum.  The vacuum helps force the epoxy into cracks and pores to stabilize the slab and fill voids that would otherwise make microscopic imaging difficult.  Often, the epoxy is colored or fluorescent to enhance the visibility of voids in a microscopic image (French 1991).  The slab is reground, removing approximately 1 mm of material to ensure that no damage from the saw cut will be seen in the microstructure.  Also, at this point, a second epoxy impregnation is often applied before bonding the concrete slab to a glass slide and grinding to the desired thickness.  In all grinding and cutting steps, non-aqueous lubricants and coolants must be used to avoid altering the cement microstructure by water addition (French 1991). 

The final thickness required is determined by the type of microscopy used and the information to be obtained.  For observation by optical microscopy, a section thickness of 40 micrometers is suitable for observation of ASR while a maximum section thickness of 20 micrometers is required to observe the microstructure of the cement paste (French 1991).  If the thin sections are to be analyzed in an SEM, section thickness is not as critical but should not be any less than 10-20 micrometers.  If the specimen is too thin the electron beam, used to form an image in an SEM, penetrates the concrete specimen and creates unwanted radiation from the glass slide (Goldstein et al. 1992).  The surfaces of thin sections are polished on a mechanical polishing wheel with decreasing diamond abrasive size.

French (1991) also described the benefits of polished plates.  Polished plates are prepared for observation by reflected light optical microscopy or stereo microscopy and have the advantage of giving the petrographer a larger specimen area to analyze.  Also, sections may be cut along the core length, providing the petrographer a different insight into the concrete microstructure.  The surface of a polished plate must be polished as flat as possible and to the same degree as a thin section.  In addition, it may be necessary to use a dye to enhance the contrast between phases.  Polished plates can be used to examine the aggregate rock type, aggregate distribution, magnitude of alkali reactive aggregates and amount of reaction products, crack patterns, and the structure and distribution of microcracks.  The air-void system parameters such as air content and spacing factor, along with the volume fraction of concrete ingredients, can be readily measured using the stereological techniques described in ASTM C 457.

In a more recent publication, Kjellsen (1996) discussed the need for flat, polished specimens when high-strength concrete is being examined.  In particular, the author noted that care should be taken to prevent differential polishing of phases in the cement as a result of their different, respective hardnesses.  Also, the author cautioned that high-strength concrete is typically less permeable than normal concrete and, as a result, epoxy will not penetrate as deeply during the impregnation step.

Broken surfaces are useful for examining aggregate and crack surfaces and the contents of voids.  Also, fresh cement paste gel can often be observed on aggregates, indicative of ASR (French 1991).  Due to the relief present on any fracture surface, observation must be performed using a high quality stereo OM or, preferably, an SEM.  The ability of an SEM to examine surfaces at a high magnification and still maintain a high depth of field allows for the examination of fine crystallites or phases that may be recessed in crevices or cracks.  Also, the ability to perform a chemical analysis of a cubic micrometer of the specimen allows for the identification of phases not clearly defined by petrography alone.

In addition to choosing the proper type of sample preparation method to see specific features, the quality of preparation can affect the information obtained.  For example, studies of cement paste pore structure, a key parameter affecting strength and durability, can be performed using an SEM only if epoxy impregnation is performed properly.  Pores and cracks on a polished surface will not produce a uniform electron yield, relative to the bulk of the sample.  In many cases, "edge effects" are seen where edges around holes or cracks appear brighter than adjacent material of the same composition (Goldstein et al. 1992).  Although this approach is more prevalent for images formed by secondary electrons (SE), images formed with backscattered electrons (BE) also exhibit this effect.  In addition, material at the base of a pore may produce a high electron yield and appear as a bright island in the otherwise darker pore.  When filled with epoxy, a continuous, low mass phase is created, providing a continuous dark shade in any electron image.  Epoxy impregnation is equally important for OM petrography where the microscopist, or image analysis system, is looking for phases with a bright or fluorescent color, which was added to the epoxy before impregnation.

The most significant sample preparation problem for cement and concrete is the dehydration of chemically bound water.  A vacuum is used to impregnate specimens with epoxy, affecting specimens for both optical and electron microscopy.  However, the relatively low vacuums achieved by most impregnation systems lead to minimal or no damage to the specimen due to dehydration of chemically bound water.  With the CSEM, the inner chamber is maintained at a pressure of approximately 1x10^-6 torr, to maintain the focus of the electron beam.  Given that the vapor pressure for water is approximately 4.6 torr at the freezing point, excess water will evaporate from the concrete, into the SEM, and chemically bound water will be lost, resulting in alteration of hydrated solids such as ASR reaction products and ettringite.  In addition, it has been reported that solid phases in cement that have water as part of their structure may decompose in situ as a result of losing water to the CSEM vacuum.  Efforts to dry a sample may result in the removal of water that is part of the CSH hydration product microstructure (Bergstrom and Jennings 1992).  These problems can be partially overcome by using an LVSEM while maintaining a 1 to 2 torr vacuum in the chamber or completely overcome by using an ESEM with a chamber pressure in excess of 4.6 torr.

Further, vacuum dehydration of a specimen occurs as part of the process used to conductively coat a specimen for observation in a CSEM.  Special dehydration steps are required beyond vacuum dehydration only when microcracking or hydrated phases are to be studied.  The principal methods for drying are oven drying, freeze-drying, and acetone or alcohol replacement.  Oven drying is the simplest approach.  Specimens are placed in an oven at 95-105°C and dried until a constant weight is obtained.  This approach is widely used (Diamond 1997) and can be readily applied to concrete specimens sampled from in-service concrete.  Replacement by acetone or alcohol is also used to dehydrate concrete where the specimen is saturated in the solvent and, by diffusion, the alcohol replaces the water.  Struble and others have used a process of water replacement by ethanol followed by ethanol replacement by ultra-low viscosity epoxy where the epoxy used was an ethanol miscible epoxy (Struble and Stutzman 1989; Stutzman and Clifton 1999).  The impregnation was performed by immersing the specimens in the ultra-low viscosity epoxy solution and holding the solution at 2°C to impede the polymerization of the epoxy and allow it time to diffuse into the concrete matrix.  This process dries the specimen and impregnates it in a single, two-step procedure.  Freeze drying removes the water by initially freezing the specimen in liquid nitrogen or some other cryogenic medium.

When operating at vacuums common in an LVSEM or ESEM, the need for conductive coatings or special dehydration methods are eliminated.  This is a significant advantage when analyzing concrete using an SEM.

Optical Microscopy – Petrography

Petrography is an analysis method dating back to the mid-1800's when a mineralogist named Highley published the first description of the petrographic optical microscope in 1856 (St John, 1998).  From there, the science of petrography grew and became a staple of mineralogical and geological research.  Petrography, in essence, is the study of light passing through a specially prepared rock or mineral specimen where the material is thin enough to allow light to pass.  As it does, the light interacts with the material differently depending on whether it is crystalline (ordered) or amorphous (disordered).  For crystalline materials, the light interaction differs depending upon the arrangements of atoms in the crystal structure.  The resulting effects can be interpreted by a properly trained petrographer and used for identifying the material. 

Concrete petrography refers to the application of petrography to the analysis of "man-made" rocks (i.e., concrete).  One of the first reports of application of petrographic techniques to concrete was published by N.C. Johnson in 1915 (Erlin 1993) and related observations of the concrete microstructure to performance.  Since then copious publications have been written on the subject, including textbooks (St John 1998; Campbell 1985).  In 1962, Mielenz published a review of concrete petrography that remains a very informative resource as a background on the science.  However, advances in instrumentation and development of new instrumentation have occurred since then and more recent reviews provide additional information.  One such review was published by French (1991) and provides detailed descriptions of various diagnostic features of distressed concrete of all types.  Optical microscopy has the advantages of being relatively inexpensive and also relatively reliable, assuming the microscopist has extensive experience examining cement structures (Campbell-Allen and Roper 1991).  An optical microscope can be used to examine amorphous materials while other techniques, such as XRD, cannot easily quantify these phases.  However, there are limitations to the optical microscope, such as a relatively low image resolution and no ability to perform chemical analysis of the specimen or its micro-constituents (French 1991).  Even with these limitations, optical microscopy is the best choice for studying PCC microstructure and it plays a key role in everyday laboratory petrographic analysis. 

Of special importance in the area of petrography is one relatively new technique that has been developed based on fluorescent microscopy.  First proposed by Christensen (Thaulow et al. 1982), the use of UV-fluorescent dye was proposed for the determination of capillary porosity in hardened cement paste.  In a later paper, Thaulow et al. (1982) used this method to estimate the compressive strength of concrete samples using fluorescence microscopy.  After development, the method has become accepted by some as a measure of w/c in hardened concrete (Mayfield 1990; Jakobsen et al. 2000).  Additionally, the UV-fluorescent dye clearly highlights cracks, microcracks, and air voids.  Cracks or voids filled with reaction products also clearly stand out from the surrounding cement paste.

Scanning Electron Microscopy

Scanning electron microscopy and digital imaging have been widely applied to PCC, cement paste, and mortar characterization.  As stated previously, the cement paste microstructure and the quality of the aggregate interface are the main areas of focus for researchers examining strength and durability of PCC.  For cement paste, and for the interfacial transition zone (ITZ), the key parameters affecting strength are total porosity, pore size distribution, the degree of hydration, and the characteristics of calcium hydroxide phases within the paste.  Many of these attributes have been studied in laboratory prepared specimens of PCC, cement paste, and mortar using an SEM and digital imaging, but little has been presented in the literature on the use of these techniques for routine characterization of concrete that has been in service.  However, the same techniques used to study laboratory prepared specimens of PCC are applicable to the study of concrete that has been in service.

One example of the application of the SEM to characterize concrete microstructure was presented by Tamimi (1994).  He developed a new two-stage mixing procedure that resulted in reduced bleeding, an increased minimum and maximum micro-hardness, and a high rate of strength development.  In this study, XRD was used to identify principal phases resulting from the hydration of portland cement and quantify the abundance of these same phases as a function of hydration time.  Additionally, he was able to characterize the crystallographic orientation of calcium hydroxide occurring at the aggregate-cement interface and show the tendency for calcium hydroxide to grow with a preferred orientation.  In this study, an SEM was used to observe the resulting microstructures and to confirm observations made from XRD results by performing microanalysis of key phases.  In a more recent paper, Tamimi (1996) used the SEM to study the aggregate-cement interface in concrete produced using the same two-stage mixing procedure and monitored the morphology change with time.

As discussed previously, the advent of the ESEM offers significant advantages for cement characterization.  The ESEM allows samples to be viewed at pressures greater than the vapor pressure of water, eliminating the dehydration of concrete specimen.  The ESEM has a unique electron detector that takes advantage of the gas molecules in the ESEM chamber to amplify the electron signal.  The ability to observe specimens "wet" allows for specialized tests such as in situ hydration experiments.  Most published ESEM work falls into this category.  Using an ESEM, Bergstrom and Jennings (1992) demonstrated that C₃S phases younger than 16 hours, when observed under a conventional SEM, have a different microstructure compared to the same phases observed in an ESEM.  In another study, Lange et al. (1991) used the ESEM to observe the early hydration of portland cement.  The results indicated unique insights into the development of cement microstructure at early ages.

Analytical Chemistry

The role of the chemistry laboratory in analyzing an MRD in concrete is not always clearly defined.  This is principally due to the fact that analytical chemistry serves multiple purposes.  First, it can help diagnose routine problems.  Also, the chemistry laboratory can provide quality control where tests in controlled environments are used to verify important measurements that are routinely performed using alternative methods.  Some examples are cement content and determination of w/c, chloride concentration, and sulfate analysis.  In addition, the chemistry laboratory provides a controlled environment where long-term studies can be conducted without interruption.

Analytical chemistry tests for the diagnosis of MRD can be conducted at almost any step in the evaluation process, but are often not performed at all.  It should be noted that many chemistry procedures require relatively large volumes of concrete.  Also, some chemical analyses take many days to complete.  It is for these reasons that analytical chemistry testing should be conducted at an early stage of the analytical process. The most common application of analytical chemistry for diagnosing MRD in PCC is for determining the cement content of the hardened concrete, and indirectly, the w/c.

Recommended Laboratory Procedures for Investigation of MRD in Concrete Pavements

A key step in identifying the cause of MRD in concrete pavement is laboratory analysis of the distressed concrete.  The laboratory analysis of concrete is facilitated by the systematic application of test methods specifically designed to identify MRD by searching for known symptomatic indicators, or diagnostic features.  However, laboratory results are susceptible to broad interpretation, and a rigid adherence to laboratory protocol, along with the judgment of an experienced petrographer, analyst, and/or engineer is often required to avoid incorrectly diagnosing a problem.

This section provides a brief overview of the procedures described in the guideline presented in Volume 2 of this report.  It also provides a brief description of commonly applied procedures and a suggested format for gathering and storing data.  This procedure is intended for use by both engineers and analysts where an engineer is defined as the person presenting the problem to the laboratory and the analyst is the laboratory personnel charged with diagnosing the concrete problem.  In many cases, this may be the same person,but often they are members of a team of people that need to communicate clearly to arrive at a cause, or causes, for the suspected MRD.  In this sense, both engineers and analysts must have a clear understanding of the required tests and the possible caveats associated with each test to properly interpret the results.

Approach to Laboratory Analysis of Concrete

When MRD is suspected of playing a role in the premature deterioration of a concrete pavement, laboratory tests are essential to help understand the underlying mechanisms at work. Optical microscopy using the stereo microscope and the petrographic microscope are the most versatile and widely applied tools for diagnosing causes of MRD.  Electron microscopy is becoming more prevalent, especially for chemical identification of reaction products and other secondary phases using energy dispersive spectroscopy.  Analytical chemistry plays a very important role in determining key parameters of the concrete (e.g., w/c, chloride content).  XRD is applied in some cases but is not widely used in the analysis of deteriorated concrete.

Often,when diagnosing concrete distress, there is no clear answer as to which single distress mechanism caused the failure. This has been referred to as "the straw that broke the camel’s back" theory (Erlin 1993) where multiple distress mechanisms are active and it is the combination of these, in concert, that lead to the failure of the concrete.  Various types of distress mechanisms can occur simultaneously in concrete and each can incrementally contribute to the ultimate failure of the material.  Although in many cases it is difficult to attribute concrete distress to one mechanism, in a majority of cases, the MRD can be easily diagnosed.  This is often accomplished by use of the optical microscope alone. As a result, most laboratory diagnostic procedures have focused heavily on the use of optical microscopic methods. Analytical chemistry methods have also been used, such as staining techniques and determination of parameters such as w/c.  Other techniques, such as SEM and XRD, have also been widely used by researchers and are becoming more common for forensic investigations of concrete failures.

When initiating the study of deteriorated concrete, or any material, an analysis plan of how to approach the problem must be followed.  This plan very often reflects a process of elimination; rather than proving what the problem is, prove instead what it is not. In this way, diagnosis of the MRD responsible is achieved.  The basic flow of a typical laboratory analysis is presented in figure 1-6. The general approach is to start very broadly, inspecting the concrete by eye. As the concrete is examined the analyst should look for diagnostic features, which are essentially a condition or physical property of the concrete that will assist in diagnosis. After evaluating or assessing the core visually, a hand lens or stereo microscope can be used to look more closely at interesting features.  In some cases, as a result of this visual and stereo microscope analysis, the probable or certain cause of distress is identified.  In most cases, a few potential MRD types can be eliminated and further analysis can then focus on those remaining.  At this point, the analyst must decide which examination technique can be performed to confirm a given MRD, or eliminate other MRD types, thereby narrowing the choices to the most probable mechanism.

A summary of the recommended process is as follows.  A sample of concrete exhibiting distress comes into the lab and is first visually inspected.  Upon completion of the visual inspection, specimens are produced from the core sample and observed using the stereo OM for initial optical analysis.  It is common at this stage to employ staining techniques to help identify ASR or sulfate phases. Next, the specimen is viewed in the petrographic microscope and/or SEM, as required.  This process of using the stereo OM, petrographic OM, and SEM is iterative and it is quite common to view the same specimen in all three instruments.   Staining in particular can assist in the optical evaluation although it may interfere with SEM analysis.

Although a trained petrographer can use the petrographic microscope and identify practically all minerals and aggregate reaction products present in a concrete specimen, this often requires thin section preparation or detailed analysis of picked grains with refractive index liquids.  This requires a highly skilled concrete petrographer to analyze and interpret the complex and vast array of information revealed by the petrographic OM.  In contrast, the SEM is an instrument almost any laboratory technician can learn to operate.  Another advantage of the SEM is the simple presentation of the results in a form engineers, technologists, and scientists can all understand.  Both x-ray mapping and x-ray microanalysis are very useful ways of identifying components of the microstructure.  However, the use of the SEM has some disadvantages. 

Figure 1-6. Fundamental process for analyzing a concrete MRD sample.

Cracking problems in the CSEM, microanalysis problems in the ESEM, and a much higher initial cost with a significant on-going maintenance cost associated with its operation.

When analyzing a concrete specimen, the concrete should be viewed as an entity consisting of a system of four principal components: air, hydrated cement paste, coarse aggregate, and fine aggregate.  All available methods to examine the system and its components should be used, looking for all features that will help in the diagnosis.  The ability to establish certain features as being normal greatly helps in deducing the cause of the problem.  For example, no apparent coarse aggregate cracking all but eliminates aggregate freeze-thaw deterioration as a cause of distress.  As another example, the presence of an adequate, un-compromised air-void system helps rule out paste freeze-thaw damage as the primary distress mechanism.  It is emphasized that a systematic examination of all components of the concrete is crucial to determining the cause of failure.

In the end, the proper examination of concrete requires the application of independent, unbiased testing methods in a uniform and controlled approach.  The number of required tests is determined by the complexity of the MRD while the implementation may depend upon various factors including scheduling of laboratory equipment and personnel.

Procedures and Data Collection for the Analysis of Deteriorated Concrete Pavement

The purpose of this section is to provide an overview of recommended analytical procedures used to examine MRD in concrete pavements.  To help facilitate this standard analysis, data sheet templates were designed to follow a core through the laboratory evaluation.  These are provided in the guideline presented in Volume 2 of this Final Report.  The use of these data sheets assists in the laboratory analyses and data interpretation in a number of ways.  First, these data sheets provide a framework for the “questions asked about the material” and are intended to be consistent with the requirements of ASTM C- 856 Practice for Petrographic Examination of Hardened Concrete.  It is noted that the individual questions and data queried in these forms do not constitute the full implementation of ASTM C- 856. Instead they focus on the analysis of deteriorated pavement concrete, which is only a small part of the scope included in this very broadly applied Practice.

Another benefit of using data forms is that they serve as the “lab notebook” for recording and archiving laboratory results.  As an example, careful inspection with a stereo microscope alone can often reveal the reason for concrete failure.  Although a diagnosis of MRD based on stereo microscope observations may be correct, it is often necessary to use other techniques to confirm the initial diagnosis.  The data sheets provide a systematic way of gathering and archiving the results of multiple laboratory analyses that may be conducted by multiple technicians, often in different laboratories.  Finally, personnel outside the laboratory want to know that all possible deterioration mechanisms were considered and that no rush to judgment was made on limited data.  By using standard data collection forms, the diagnostic indicators laboratory personnel were looking for will be clear to others.  This reporting protocol helps provide an understanding of the decisions made by laboratory personnel to arrive at a diagnosis of material failure.  The data sheets provided in the guideline are broken up into the following major categories roughly corresponding to the steps outlined in figure 1-6:

One copy of the completed laboratory data sheets, along with sampling and field data sheets, should accompany each core through the laboratory evaluation. Within each data sheet, tests and diagnostic features are presented in a table format along with possible or common results.  It is up to the analyst to observe and make judgments about the concrete relative to the questions posed on each data sheet.  Laboratory personnel should complete only those data sheets that apply to analyses conducted.  These observations are combined with all other data to arrive at a diagnosis of the MRD observed.  Details regarding the data collection process are provided in the guideline in Volume 2 of this final report.

Summary on Recommended Laboratory Testing Procedures

The diagnosis of MRD in PCC often requires the use of various laboratory procedures for identifying the extent and mechanism of distress.  These laboratory tests do not always lead to absolute characterization of the distress for a variety of reasons.  First of all, most cases of concrete distress occur as the result of multiple distress mechanisms.  As such, it is often difficult to isolate the specific cause of failure or even determine the principal cause of failure.  Second, the laboratory methods commonly used often provide results that may be interpreted differently, depending upon the motivation or objectivity of the analyst.  To minimize the latter case of misinterpretation, it is advised that a thorough, complete examination of the concrete be performed using the data collection forms and approach described.

Clearly, the most useful tools for examining concrete are the stereo OM and the petrographic OM and/or the SEM.  There is a significant body of technical information available discussing OM observations of concrete and concrete distress.  This information is useful as a benchmark when evaluating a specific concrete specimen.  The SEM is quickly becoming an equally valuable tool for evaluating concrete.  It is advised that anyone charged with the examination of MRD in concrete become familiar with this equipment and the capabilities it offers.

Finally, the general approach of "asking the materials questions" must be followed.  As is discussed in the next section, it is often the case that a process of elimination is required to determine what distress is not present, thereby leading to a shorter list of possible distress mechanisms.  Also, it is important to remember that the concrete as observed may have undergone a significant metamorphosis over its service life and the degradation seen may be the final product of years of exposure. It is only through the careful and methodical application of the described laboratory methods that the true cause of distress may be identified.

Background to Data Analysis and Interpretation of MRD in Concrete Pavements

In order to make an accurate diagnosis of MRD, four primary sources of data must be analyzed.  The first source of data is that obtained through a review of existing construction records that details the materials used, mix design, and conditions during construction.  Unfortunately, construction records are not always available and when available are not always of sufficient accuracy to assist in the diagnosis.  Next, site-specific information not directly related to the construction of the project under investigation can also contribute valuable insight.  Specifically, climatic information can be used to assess the environment to which the project was subjected and soil analysis can be used to investigate the possibility of whether some potentially harmful compounds (e.g., soluble sulfates) were naturally present.  A third source of valuable records are the responsible agency’s maintenance records, which will not only reflect the level of maintenance activities applied, but also the amount and type of deicers that may have been used.  The last and most important source of data is from laboratory analyses, which are almost always required to reach a conclusion regarding the cause of the observed distress.

Unfortunately, there are no definitive sources of information on how to diagnose a concrete failure given the wide range of material applications, environments, mix designs, and possible reasons for failure.  To develop a protocol for diagnosing MRD in PCC pavements, it is necessary to assemble known information from numerous sources.  Then, the underlying relationships between various observed diagnostic features must be understood, again requiring reference to many sources of information.  Further, there is currently no available protocol for identifying MRD in PCC pavements.  Without a defined method, agency personnel are stymied in their quest to determine the deterioration mechanisms that are shortening the life of PCC pavements.  In the process of developing the protocol presented in the guidelines, the following summary of diagnostic features for each common MRD was gleaned from the literature.

Diagnostic Features for Paste Freeze-Thaw Deterioration

When moist concrete is exposed to alternating cycles of freezing and thawing, internal deterioration can result in the paste, as described previously.  Paste freeze-thaw deterioration is observed on the pavement surface as scaling and/or map cracking, typically appearing first near joints and free edges.  Deterioration may have progressed to the point where the surface is pitted as the paste disintegrates, exposing coarse aggregate particles, which eventually may work free.  Confirmation of the diagnosis requires laboratory examination, primarily through optical microscopy.

Concrete affected by paste freeze-thaw damage can be identified through the presence of microcracking in the paste using the petrographic optical microscope.  If the freeze-thaw cycling has led to scaling, these microcracks will predominantly run parallel to the wearing surface.  The best method to determine if paste freeze-thaw is the primary distress mechanism is a microscopic characterization of the air-void system, following ASTM C 457 procedures to assess the adequacy of the entrained air system (Walker 1992).  Analytical systems to perform ASTM C 457 are not common, but are available.  Often, it is easier for the analyst to purchase the necessary hardware and write the necessary controlling software.  For this research, the latter approach was adopted (Sutter 1998).

On normal strength concrete (relatively high w/c), an entrained air-void system producing a Powers spacing factor of 0.20 mm or less will typically provide good freeze-thaw protection.  This value was empirically established primarily through laboratory freeze-thaw testing using ASTM C 666 (AASHTO T 161).  While the Powers spacing factor is not considered to be a truly definitive measure of freeze-thaw protection, it is still used as the standard method of quantifying the distribution of entrained air in concrete.  Other measures for characterizing entrained air have been proposed (Walker 1980; Philleo 1983; Attiogbe et al. 1993), but none have been adopted for general use.

The minimum requirements for an entrained air system in higher strength concrete (lower w/c) are less clear.  Some laboratory research has indicated that, as the w/c decreases below 0.5, the spacing factor required for freeze-thaw durability increases.  While accepted recommendations for a larger spacing factor for higher strength concrete have not been published, ACI 318 - 95 (Section 4.2.1) permits a 1 percent reduction in the recommended air content for concrete with specified compressive strengths above 34.5 MPa.

The microscopic examination of the air-void system in hardened concrete can take the form of a point count, linear traverse, or an areal traverse.  All three methods rely on measurements obtained from a polished plane surface of concrete.  The general mathematical methods used to extrapolate measurements obtained from a two-dimensional surface to three-dimensional space are known as stereology (Russ, 1986).  The mathematical derivations for volume fraction relationships are based upon measurements in a two-dimensional slice of the volume, as is the case with a polished microscope specimen.  The fundamental relationships leading to volume fraction estimation depend upon equivalence between the volume density of quantities measured on a plane section, including point, line, or area fractions.  The fundamental relationships are shown below:

where:             

P_P =    Fraction of total points counted falling in phase of interest.

L_L =    Fraction of total line length traversed falling in phase of interest.

A_A =    Area fraction of phase of interest.

V_V =    Volume fraction of phase of interest.

Both methods of volume fraction estimation, point counting and linear traverse, have been applied to air void measurement in concrete.

In addition to the total volume of air in hardened concrete, ASTM C 457 details procedures for determining a Powers spacing factor.  Powers (1949) developed two expressions for a spacing factor, both of which require a determination of the total air-void specific surface. It can be demonstrated that the total volume of air voids and their total specific surface can be estimated from the mean air-void intercept or chord length obtained from a linear traverse.  Assuming all air voids to be spherical and using geometric probability concepts, the total specific surface, a, expressed in terms of the average chord length, , was shown to be:

(1-4)

Powers first spacing factor expression was obtained by simply calculating the volume of cement paste per unit area of air-void surface.  This is given as:

where:

=    Spacing factor, in units of length.

a  =   Total specific surface of the air voids, in consistent units of length^-1.

p   =   Paste content, in volume percent of concrete.

A   =    Total volume of air voids, in volume percent of concrete.

The second spacing factor expression is based on a hypothetical system of equal sized spherical voids uniformly distributed throughout the paste phase. The size of each of these hypothetical voids is determined by setting their specific surface (3/R) equal to the total measured specific surface of the true void system and then solving for the resulting sphere radius, R.  By making the total air content of the hypothetical system of voids equal to the measured value of air content, the number of hypothetical voids is then determined.

 The cubic packing of this hypothetical void system can be visualized as a system of equal sized adjoining cubes of cement paste with an internal air void located at the center of each.  The maximum distance from anywhere in such a system to the nearest surface of an air void is the distance along the cross diagonal from a cube corner to the enclosed void surface.  This distance is equal to one-half the length of the diagonal minus one-half the sphere diameter.  The Powers spacing factor thus obtained is:

Powers recognized that neither expression for spacing factor provides a true measure of void spacing.  Assuming that both expressions overestimate the true average void spacing, he recommended using the smaller spacing factor obtained from the two equations.  Equation 1-5 yields a smaller factor for p/A less than 4.33, and Equation 1-6 gives the smaller value when p/A is greater than 4.33.

It has generally been found that if the spacing factor is adequate, then the cement paste should be protected against freeze-thaw damage.  But it is readily acknowledged that paste freeze-thaw damage can occur even in concrete having what appears to be an acceptable air-void system.  This type of damage is commonly associated with scaling of the surface. Microscopic observation of the microstructure may reveal a system of cracks running parallel to the wearing surface, with crack spacing typically widening as the distance from the exposed surface increases (Walker 1992).  This type of manifestation is commonly associated with poor finishing in which either bleed water was trapped or the surface was overworked, compromising the air-void system.  The application of deicers can also contribute to this cracking by accentuating the stresses caused by freezing and thawing at the surface. Similar cracking has been reported at joints, except that the cracks may run parallel to the joint face as well.

Paste freeze-thaw deterioration is a physical process, and therefore does not directly involve deleterious chemical reactions.  In light of this fact, one would expect that the observation of microcracking in the paste, not apparently associated with a chemical process, might thus be due to a physical process such as freezing and thawing. Unfortunately, this is not always the case as such microcracking could be the result of aggregate dilation.  In damaged concrete it is very common to find secondary deposits (e.g., ettringite, calcium hydroxide, etc.) in air voids and/or cracks.  It has been stated by some that these secondary deposits in air voids might compromise the freeze-thaw durability of the paste resulting in damage (Niemann and Lehtonen 1997; Ouyang an Lane 1997).  Others have stated that the infilling of voids and the secondary deposition of compounds such as ettringite are not the cause, but instead the result of damage that has already occurred due to other deterioration mechanisms.  It is common to measure the air-void system parameters to determine if they are adequate and to suspect paste freeze-thaw damage as at least a contributing factor if they are not.

Diagnostic Features for Deicer Scaling/Deterioration

Distress related to deicer use is most often associated with surface scaling and delamination, which is almost unique to this distress type.  The deterioration is observed in the paste, exposing the coarse aggregate particles.  This is in contrast to a distress such as "popouts" that is characterized by fractured aggregate particles surrounded by intact paste.  If the concrete under evaluation is subjected to deicer applications and is showing signs of scaling, deicers are likely contributing to the distress.

Chemical test methods, such as AASHTO T 260, can be used to measure the chloride ion concentration in relation to depth of a concrete specimen.  The presence of a high chloride ion concentration at or near the scaling surface is an indicator that the distress is likely related to deicer applications.

Determination of deicer deterioration at joint and crack locations is a much more difficult task. Rarely is the deicer directly responsible for deterioration. It contributes by increasing the level of saturation, increasing the solubility of some compounds, and potentially increasing the alkalinity of the pore. Thermal shock may also contribute to the occurrence of distress.  Under these circumstances, the concrete becomes more susceptible to other distress mechanisms such as paste and aggregate freeze-thaw deterioration as well as ASR.  The deicer may also provide a source of soluble sulfates, resulting in external sulfate attack.

Although the primary distress mechanism may obscure the contribution of the deicer to the deterioration, there are some diagnostic features that are indicators of whether deicer deterioration is at least a contributor to the observed distress.  Muethel (1997) examined concrete at crack locations using phenolphthalein to identify calcium hydroxide depletion and permeability measures to estimate coarsening of the paste. He was able to identify zones at the concrete/crack interface that have significantly altered microstructure compared to the mass concrete, speculating that these differences in microstructure have resulted in joint/crack deterioration and staining.  He further states that primary leaching agent is carbonic acid, which aggressively attacks concrete in a high salinity environment. The presence of chemical deicers thus accelerates the leaching process.  Marchand et al. (1994) described a similar process.

It is logical to assume that salts would concentrate in joints or cracks due to infiltrating melt water produced from the deicer application.  If the chloride ion concentration were mapped from the joint interface into the concrete mass, the gradient would likely be one in which the concentration would be highest at some distance inward from the joint. The concentration at the interface would decrease as the flow of water in or through the crack/joint would have a tendency to wash the deicers away while wetting and drying cycles would concentrate them just below the interface surface. Thus, chemical methods could be employed to determine chloride ion concentrations at various distances from the joint/crack face to measure this phenomenon.

Polished concrete slabs can be examined by noting to what depth the cracking is present. Commonly, a pavement undergoing deicer scaling will have microcracking isolated in and parallel to the surface. This is very similar to common paste freeze-thaw deterioration (Walker 1992), and in fact the mechanisms are closely related.  In addition, if the air-void system parameters in the lower portion of the specimen are adequate, and the paste near the finished surface is devoid of air, poor finishing or curing is likely the responsible mechanism.

A marginal air-void system throughout the concrete mass can initially manifest itself as deicer scaling, but the distress is likely to be progressive in nature as the problem will not be isolated to a weak surface layer.  This would be best categorized as paste freeze-thaw deterioration.

Petrography can also be employed to assist in the diagnosis of deicer deterioration.  Carefully prepared polished thin sections could be used to identify salt crystals that might be present.  Also, changes in the paste porosity/texture will be identifiable if care is taken in sample preparation.  Some researchers are convinced that common deicers react with cement paste to form chloroaluminates and oxychlorides, compounds that can be observed through optical petrography and electron microscopy.

Deicer distress can appear outwardly similar to ASR and aggregate freeze-thaw deterioration in some manifestations.  Wolter speculates that ettringite, in its "amorphous" form, has been confused with alkali-silica gel.  An SEM with EDS capability can identify the minerals definitively.  In the case of deicer distress, ettringite filling the interstitial pore space can be observed using an SEM (Wolter 1997).  The SEM preferably should be an LVSEM or ESEM to reduce the drying shrinkage cracking caused by a high vacuum in the CSEM. Wolter (1997) gives the following characteristics of deicer distress that can be identified using optical microscopy:

• Microcracking propagating around aggregate particles and though the paste only.
• Aggregates relatively intact with virtually no reaction rims of silica gel observed,
• Deterioration within the paste only.
• Microcracking within the paste occurring predominantly sub-parallel to the deteriorating surface.
• Secondary deposits of ettringite within the air-void system and microcracks.
• Less microcracking and secondary deposits as distance increases from the deteriorating surface.

When diagnosing possible deicer distress, the orientation of the sample is important.  Ideally, the sample should contain the vertical joint surface, so that the conditions with distance from the joint can be noted.  Carbonation, microcracking, and the quantity of ettringite often change dramatically over distances of 25 to 50 mm, and samples with unknown orientation to the joint are virtually useless (Wolter 1997).

Thin section analysis should be used to detect ettringite growth within a functional air void. Wolter warns that using melted carnauba wax prior to polishing may alter ettringite due to the heat involved.  Therefore, as much analysis as possible should be done before the wax is used.

It has been shown that concrete subjected to repeated applications of deicers and wetting and drying cycles develops a coarser pore structure, characterized by larger, interconnected pores.  Concrete permeability thus increases, making it more susceptible to many other distresses. This connection has been described by a number of researchers.  For example, Basheer (1994) has proposed a single deterioration of concrete-permeability interaction model that links concrete permeability to corrosion of steel, frost damage, chloride ingress, carbonation, sulfate attack, salt attack, alkali attack, acid attack, alkali-aggregate reactions, and abrasion.  It therefore may be possible to use the increased permeability associated with deicer deterioration as a diagnostic tool by comparing concrete obtained near joint and crack interfaces to that obtained from a slab interior.

Two widely accepted standard test methods for measuring concrete permeability are the British test method (British Initial Surface Absorption Test) that is not commonly used in the United States and AASHTO T277 (ASTM C 1202) rapid chloride permeability test (RCPT).  Many other test methods exist, but none has gained greater acceptance than the RCPT. SHRP researchers recommended modification to the AASHTO T 277 test to make it more rapid (Zia et al. 1993). It was also advocated by SHRP researchers that a "pulse pressure" method for concrete permeability be adopted to measure water permeability in the laboratory (Roy et al 1993a; 1993b).  A third SHRP test method was proposed for field evaluations in which the rate of airflow through a concrete surface under an applied vacuum is measured to estimate permeability.  Unfortunately, this test is not considered sufficiently quantitative for prediction of actual permeability and thus results should only be used as an indicator of permeability (Whiting et al. 1994).  More recently, a concrete sorptivity test has been under development and is under review by ASTM that may provide an indirect method to assess concrete permeability.

Mobasher and Mitchell (1988) published the results of a large study investigating the applicability of the RCPT.  They found that the "test is valid and can be used with confidence."  They investigated the repeatability of the test, finding the single operator coefficient of variation (COV) of a single test result to be 12.3 percent and the multi-laboratory COV to be 18.0 percent.  Misra et al. (1994) report that the RCPT can be used as an important tool for quality control, inspection, and design, although further research is needed to examine the effect that pore solution chemistry has on the results.  This is a very important finding from the perspective of using this test as a diagnostic tool, as the pore solution in concrete near joints and cracks will be different than that elsewhere due to the ingress of deicing chemicals.

Feldman et al. (1994) reported in their investigation of the RCPT how changes in experimental conditions and specimens affected results.  It was observed that the test affected the pore structure and resistivity of concrete, particularly relatively young concrete.  It was also stated that simple measurement of initial current or resistivity gave the same ranking for the four concrete specimens tested, and therefore this might be able to replace the RCPT.  It is noted that all the concrete produced had a w/c of 0.55, and that their findings need to be verified for lower permeability concrete, and that made with blended cements.

Although the RCPT has been embraced by many SHAs due to its ease of use, it suffers some limitations that make it impractical when evaluating some mixtures.   The three main limitations are: 1) the current passed is related to all ions in the pore solution and not just chloride ions, 2) the measurements are made before a steady-state migration is achieved, and 3) the temperature of the specimen increases due to the applied voltage (Stanish et al. 2000).  The first limitation is most problematic for the assessment of concrete permeability in mixtures containing various admixtures (e.g., accelerators, corrosion inhibitors, etc.) that will affect the ion concentration of the pore solution.

Diagnostic Features for Aggregate Freeze-Thaw Deterioration

Field identification of aggregate freeze-thaw deterioration usually starts with a visual inspection of the pavement surface.  Because of the increased access to moisture, the intersections of transverse and longitudinal joints and pavement edges are locations where the characteristic fine-line cracking pattern first appears.  The surface staining that frequently precedes this cracking can also serve as an indication of aggregate freeze-thaw deterioration below the surface. While the characteristic cracking and staining are strong indicators of a problem, in and of themselves they are not proof positive that aggregate freeze-thaw deterioration is the mechanism at work.  The common practice of associating aggregate freeze-thaw deterioration with observations of "D-cracking" should thus be avoided.

For positive diagnosis, cores must be obtained either directly through or in the vicinity of distressed joints or cracks to confirm the source or the surface deterioration. For concrete that is exhibiting surface cracking associated with aggregate related freeze-thaw, it is common for the bottom of the core to be extensively damaged.  In situations where the surface evidence is staining rather than characteristic cracking, observations of microcracking of some coarse aggregate particles near the bottom of a core can confirm early stages of aggregate freeze-thaw deterioration. Finally, it must be established that the aggregate deterioration is a physical process and not related to aggregate reactivity.

In a comprehensive study of procedures for identifying frost-susceptible aggregate particles, Larson and Cady (1969) reported observing four failure modes.  These included aggregate shattering, bed-plane disintegration, aggregate paste bond failure, and general disintegration of the paste matrix.  They also reported that even though the aggregates used in their study were selectively fractionated by size, specific gravity, and mineralogy into relatively homogeneous samples, a relatively small number (5 to 10 percent in some cases) of individual aggregate particles were typically responsible for the disruptive behavior.

It is best to diagnose aggregate freeze-thaw deterioration through optical microscope analysis of prepared concrete specimens. Analysis can be conducted on both polished slabs and thin sections.  Aggregate-paste bond failure through the interfacial zone and possibly disintegration of the paste matrix resulting from the expulsion of water from the aggregate during freezing are diagnostic features of aggregate freeze-thaw deterioration.  Diagnosis based on these features is less certain as the aggregate particle remains intact and thus it can be easily confused with paste freeze-thaw deterioration.  The analyst needs to examine the air-void system in undamaged paste and the pore structure of the aggregate to determine the most likely distress mechanism.

As discussed previously, it is well established that the aggregate pore system is the most important factor in determining its freeze-thaw durability.  In an effort to classify aggregates according to their potential freeze-thaw durability, Verbeck and Langren (1960) developed three general aggregate groupings based on their pore systems. In the first group are aggregates with a low enough porosity and/or permeability and a high enough strength so that they can accommodate the internal changes due to freezing by elastic action.  Aggregates in this first group are considered freeze-thaw durable.  In the second group are aggregates that have an internal pore system small enough to compete successfully with the cement paste for moisture and therefore tend to become critically saturated.  At some critical maximum size these aggregates will generate excessive hydraulic pressures that cannot be accommodated by elastic action and therefore tend to not be freeze-thaw durable.  The third group includes aggregates with a high enough permeability along with adequate mechanical properties so that hydraulic pressures from freeze-thaw events are accommodated by elastic action. While aggregates in this third group remain sound during repeated freezing and thawing, they may cause distress to the surrounding paste system or interfacial zone because of the volume of water expelled.

While useful as a means of describing different behavior, the non-uniform nature of many aggregate sources makes this grouping impractical for classifying aggregate sources. But through classification of the aggregate mineralogy and pore structure, a skilled petrographer can attempt to diagnose whether a given aggregate is to blame for an observed physical deterioration of a concrete pavement.  This in combination with the observed cracking and/or deterioration should be sufficient to distinguish the physical mechanism at work.

Diagnostic Features for Alkali–Silica Reactivity (ASR)

ACI (1998), Farny and Kosmatka (1997), ACPA (1995), and Walker (1992) have compiled lists of diagnostic features characteristic of concrete pavements affected by ASR. Visual signs of ASR include map cracking on pavement surfaces, cracking that is more open on the surface, dark or light reaction rims around the coarse aggregates, and visible cracks radiating from within aggregates.  A whitish powder, which is calcium carbonate leaching from the deteriorated concrete, in the specimen or visible on the pavement surface, is another strong indicator that ASR is occurring.

The Cement and Concrete Association of Australia (CCAA)(1996) prepared guidelines to properly determine if an existing concrete structure is in fact undergoing alkali–silica reactivity.  Visual clues include cracking and expansion of the structure such as closing of joints and displacement of fixed members.  Other signs that must be noted include damp patches and efflorescence. Walker (1992) considers the following five items for diagnosis of ASR:

• Crack pattern.
• Structural evidence of expansion.
• Verification of the presence of reactive silica in the aggregate.
• Exudations, coatings, and pore fillings.
• Laboratory investigation of sampled concrete.

Walker (1992) presented a method for distinguishing between plastic and drying shrinkage, stating that the crack pattern commonly associated with ASR affected concrete is similar to that generated when mud dries and shrinks, or plastic shrinkage cracking.  This generates an irregular honeycomb pattern in unrestrained concrete as each portion of the surface pulls away from adjacent portions. In reality, it is not a surface shrinkage phenomenon, but instead an expansion of the underlying concrete that produces the crack pattern.  The differential expansion is the result of the higher moisture content of the substrate concrete that allows the gel to swell.

The actual pattern of the cracking depends on the constraint to which the concrete is subjected.  If the structure is unrestrained, it will be in a honeycomb pattern.  But reinforcing steel in concrete structures, which are considerably longer than they are wide (such as pavements), will not allow the concrete to expand equally in all directions.  In these cases, initial manifestations of surface cracking will occur perpendicular to the direction of expansion.  In an affected pavement, this cracking will thus run longitudinally parallel to the pavement edge.  At the same time, restraint will slow down the deleterious nature of ASR in reinforced structures (Walker 1992).

In most advanced cases of ASR, the silica gel will be visible on the concrete surface (Walker 1992). The gel is most commonly dehydrated and carbonated, being an opaque white or efflorescent, and under traffic may polish to a translucent glaze.  Non-dehydrated gel is translucent and is moist or sticky.

Stark (1991) advocated a staining method for determining whether ASR gel is present based on the reaction of the ASR gel and uranyl acetate.  ASR gel treated with a uranyl acetate solution fluoresces under UV light, making it possible to detect ASR gel before it is visible to the naked eye. There are some difficulties in effectively using this test, however.  One problem is that, although it is generally accepted that treatment with uranyl acetate will cause the ASR gel to fluoresce, it has also been demonstrated that other non-deleterious compounds can also fluoresce.  This leads to false positives, resulting in the necessity for additional analysis if a positive result is obtained.  A second concern surrounds the use of uranyl acetate, which is a radioactive compound.  The fact that uranyl acetate is a hazardous material means that special precautions regarding handling the compound and specimens treated with it are required. Thus, there is a reluctance to employ this method.

More recently, a new test procedure has been developed for the detection of ASR reaction products (Civil Engineering News 1997; Engineering News Record 1997; Guthrie and Carey 1997).  This method uses sodium cobaltinitrite and rhodamin B solutions to stain the ASR reaction product. After treatment with the solutions, the specimens can be visually assessed with a hand lens or using an optical microscope.  Regions where ASR gel is present stain either yellow or pink.  It is observed that yellow staining is associated with a massive precipitate having a distinctive gel-like morphology and granular precipitate that appears to consist of crystals that have grown from the gel (Guthrie and Carey 1997).  It was determined that the yellow staining regions are potassium, calcium, silica, and sodium gels resulting from ASR. Because this test is relatively new, little field experience currently exists regarding its applicability and reliability.

Guthrie and Carey (1997) compared the results of this new staining method to the uranyl acetate method, finding excellent agreement between the two methods with the added benefit that the dual staining technique provided additional information through the appearance of both yellow and pink staining (the uranyl acetate test tagged both).  The major benefit of the cobaltinitrite/rhodamin B method is that the chemicals are not radioactive and UV light is not required for illumination, thus conventional optical microscopes can be used in the analysis.

Farny and Kosmatka (1997) identify several optical methods for use in determining ASR.  It is recommended that a close visual examination should be conducted to identify signs of distress, such as microcracking and separation of aggregates from the paste.  Some type of magnification, whether through a strong magnifying glass or microscope, is needed for this examination.  To positively attribute the distress to ASR requires confirmation of the presence of an ASR reaction product.  A petrographic analysis or analysis using a scanning electron microscope can positively identify whether an alkali-silica gel is present (Thaulow et al. 1996b).

Positive identification of ASR also requires that some of the aggregate within the paste be recognized as reactive or potentially reactive and that these aggregates must show signs of reaction.  Aggregate exhibiting internal fracturing that extends to the concrete matrix is a common diagnostic feature in concrete affected by ASR.  If only the fine aggregate is reacting, cracking in the matrix may be found that does not affect the coarse aggregates.  A network of internal cracks connecting reacted aggregate particles is a strong indication that ASR has occurred.  It has also been observed that silica gel may appear as dark areas or rims around the edge of reacting aggregate.

Walker (1992) provides a detailed description of microscopic diagnostic features observed in ASR affected concrete.  Aggregate will be altered by the reaction, as evidenced by one or more of the following features:

• Rims on the aggregate that seem to make the aggregate exterior less permeable.
• Gel-filled cracks extending into the paste.
• Hollowed-out centers of sedimentary rocks resulting from the consumption of the original cement between sand grains by the reaction.
• Aggregate particles that have been cracked by gel.

Exuding liquid gels will saturate the paste, fill cracks and voids, and will ooze onto the pavement surface.  These gel deposits become more common with depth.  The interior of the concrete will have large cracks that may not be filled with gel because they are the result of expansion elsewhere.

Thaulow and Jakobsen (1997) list the following four main microscopic features that can be used to diagnose ASR:

• The presence of alkali–silica reactive aggregate.
• A characteristic microcracking pattern.
• The presence of alkali–silica gel in cracks and voids.
• Calcium hydroxide (CH) depleted paste.

If the number of reactive aggregate particles is sufficiently high, a continuous crack pattern will develop in the concrete.  The cracks will follow a path of least resistance, running through the cement paste and reacting aggregates and, in rare instances, even through sound aggregate particles (Thaulow and Jakobsen 1997).

Some unique features of the alkali–silica gel discussed by Thaulow and Jakobsen include the fact that it is normally a colorless, isotropic material having a low refractive index of 1.46 to 1.53, but on occasion it may be partly crystalline, having an orange interference color in crossed polarized light.  The gel may also be observed replacing the outer portion of the aggregate particle.  Ettringite commonly is present in ASR affected concrete, and is easily distinguished by the slight birefringence of the ettringite formation.  Thaulow and Jakobsen (1997) suggest using fluorescent epoxy impregnation of thin sections to assist in making the distinction between the ASR gel and ettringite.  They also mention that in intensive ASR, the CH can be dissolved, "leaving a black and opaline shining paste" when observed in cross-polarized light.  This dark paste area is commonly confined to narrow zones around reactive aggregates and along gel containing cracks.

In another paper, Thaulow et al (1996b) discuss using SEM equipped with an EDS to investigate the composition of alkali–silica gel in deteriorated concrete.  In the study described, both optical and electron microscopy were used to view the same areas of prepared, epoxy impregnated thin sections.  The SEM/EDS analysis confirmed the findings of the optical microscope, positively identifying ASR gel in all locations identified by the optical microscope.  Furthermore, the SEM/EDS was able to locate ASR gel in some cracks in which the optical microscope lacked sufficient magnification to make a positive identification.  The SEM/EDS was also able to determine the composition of the observed gel, something that the optical microscope could not do.  An interesting finding of the compositional analysis is that, in most cases, the gel in air voids was richer in calcium compared to gel in cracks in the aggregate, and that the calcium content is directly related to the proximity of the gel to the cement paste. 

It is noted in ACI 221.1R (ACI 1998) that when observed in thin section, disseminated calcium hydroxide is depleted in the cement paste in the vicinity of reactive-aggregate particles. This observation is often noted before gel formation and cracking and is therefore useful in detecting early signs of ASR.

Diagnostic Features for Alkali–Carbonate Reactivity

The visual identification of ACR as the mechanism causing damage within concrete can be difficult because the manifestation of ACR mimics those of various other MRDs. The visual signs common to all AAR are cracking, closing of joints, movement, displacement of members, and pop-outs.  In unreinforced concrete, map cracking is the most common form of cracking. Reinforced concrete structures exhibit cracks that run parallel to the reinforcement (CCAA 1996).  Because these visual signs are common to both ASR and ACR, it is very difficult to distinguish one from the other through unaided visual assessment.  Ozol (1994) determined that in ACR affected concrete, there is no distinct feature or geometry of cracking that precisely identifies the cause to be ACR.  However, there is one major difference: an ASR product gel always accompanies ASR whereas a gel never accompanies ACR (ACI 1998).

It is recommended by Farny and Kosmatka (1997) that both the concrete and the carbonate rocks be examined petrographically in accordance with ASTM procedures C 856 and C 295, respectively, if ACR is suspected as the deleterious mechanism.  As discussed in Appendix A, ACR susceptible aggregate has unique features that can be used in the process of diagnosis.  Swamy (1994) identified three aggregates with which carbonate reactions commonly occur.  These aggregates are calcitic limestones, dolomitic limestones, and fine-grained dolomitic limestone aggregates containing interstitial calcite and clay.  Farny and Kosmatka (1997) outline aggregates that have potential for ACR in the following manner:

• Aggregates with a clay content, or insoluble residue content, in the range of 5 to 25 percent.
• Aggregates with a calcite to dolomite ratio of 1:1.
• Aggregates with increasing dolomite volume up to a point at which interlocking texture becomes a restraining factor.
• Aggregates with discrete dolomite crystals (rhombs) suspended in a clay matrix.

Gress (1997) and Ozol (1994) also report that the observation of large dark gray dolomitic limestone coarse aggregates having cracks radiating into the concrete paste is one indicator of ACR.  Walker (1992) also made this observation.  Another characteristic of ACR mentioned by Gress (1997) was the distinct reaction rims formed around the reactive aggregates, although Ozol (1994) pointed out that rim zones may or may not be produced on reactive particles, and therefore are not in and of themselves definitive in diagnosing ACR. 

Walker (1992) states that the "signs indicating alkali-carbonate reactivity can be very subtle because the reaction does not cause the growth or exudation of any characteristic reaction products, such as the gel associated with alkali-silica reactions."  She argues that the best diagnostic approach is through petrographic examination of the aggregate to identify microstructure and degradation.  Key aggregate microstructural features are the "small rhombic crystals of dolomite suspended in a calcite, micrite matrix that contains the fine, particulate, insoluble constituents."  Sections of extreme thinness must be used, as the dolomite crystals are often less than 10 mm in size. Walker also mentions that thin sections of reacted material will reveal changes in the euhedral dolomite crystals, either through alteration to calcite or the crystal may be partially or completely removed.

Others have suggested that although no characteristic gel is produced, the reaction product formed through ACR, the presence of brucite (Mg(OH)₂) or more complex magnesium silicates can be used to assist in diagnosis (ACI 1998).  Brucite, also known as magnesium hydroxide, can be identified petrographically (Farny and Kosmatka 1997) or through the use of an SEM, although it is difficult to observe, and therefore its absence should not be taken as evidence that ACR is not present (ACI 1998).

Diagnostic Features for External Sulfate Attack

It is not possible to positively identify sulfate attack through visual means alone, although visual assessment of a pavement can contribute to the diagnosis.  Map cracking, particularly in the vicinity of joints, the observation of a white precipitate, evidence of concrete expansion, and concrete paste disintegration are all signs of external sulfate attack.  In the case of external sulfate attack from impurities in chemical deicers, a polygonal crack pattern throughout the pavement becoming more concentrated as it approaches the joint is common (Wolter 1997).  Unfortunately, these descriptions are similar to visual cues for other types of MRD as well.  It is therefore necessary that concrete specimens be obtained for laboratory evaluation to positively identify the distress mechanism.

Site investigation can reveal conditions that may contribute to sulfate attack.  Factors influencing sulfate attack are the amount and nature of the sulfate found in the soil, the level of the water table and its seasonal variations, the sulfate content of the groundwater, the flow of groundwater, the soil porosity, and the type of construction (Mehta and Montiero 1993). Also, if impurities in the deicer salts are suspected, the source of the deicer should be tested.

Barium chloride potassium permanganate (BCPP) stain can be used to detect the presence of sulfate minerals, such as ettringite, gypsum, and anhydrite phases, in concrete (Poole and Thomas 1975).   The principal staining mechanism is a two-step process.  The concrete is initially immersed in the stain solution, causing sulfate ions released from the concrete to precipitate as barium sulfate. As this occurs, the potassium permanganate co-precipitates with the barium sulfate imparting the characteristic "purple" color to the resulting crystal.  This process results in a permanent alteration of the sulfate phase surface at the sulfate/water phase boundary.  The concrete is then rinsed in a saturated solution of oxalic acid to remove any surface coloration from precipitation of excess potassium permanganate.  The remaining purple colored crystals identify where ettringite, gypsum, or anhydrite phases existed. This method is particularly useful in identifying the presence of sulfate minerals that may have filled air voids, since small filled air voids are at times difficult to distinguish from hydrated cement paste using the stereo OM.

Chemical analysis can be conducted on distressed concrete suspected of suffering external sulfate attack to determine if more acid soluble sulfate is present than was contained in the original cement.  If external sulfate attack is responsible, the sulfate content will be many times that derived from the normal calcium sulfate content of the cement.  But in some cases, the excess sulfate may have been leached from the concrete, and therefore this test alone cannot be used to conclusively reject external sulfate attack as the primary distress mechanism. Testing for acid-soluble magnesium determines if concrete has suffered magnesium sulfate attack (Mielenz 1962).

Petrographic analysis can be used to detect symptoms of sulfate attack.  As described previously, ASTM C 856 provides some diagnostic features for ettringite that can be used to identify it presence within the concrete.  Crystalline ettringite occurs as fine, white fibers or needles or spherulitic growths in voids, in the cement paste or lining fractures (ASTM C 856). Marusin (1998) asserts that patches of "gel like" ettringite are often agglomerations of fine ettringite needles. External sulfate attack may manifest itself as an expansion of the concrete cement paste matrix.  In the intermediate stages, the reaction produces great expansion in the cement paste, resulting in progressive distention, warping, and fracturing of the concrete.  The cement paste tends to expand away from the aggregate particles, and the resulting gaps can become filled with secondary ettringite deposits.  Advanced sulfate attack results in soft, friable cement paste, and some of the aggregate rock types may also be altered.  The outer hull of highly carbonated mortar remains as a shell until finally broken by expansion of the underlying concrete (Mielenz 1962).

Given the absence of any abnormal internal sources of sulfate, the presence of ettringite alone is not evidence of external sulfate attack (Thaulow and Jakobsen 1997). Ettringite is a normal hydration product of portland cement.  Although ettringite crystals found in air voids or fractures may be responsible for the fracture, it is just as likely that the ettringite may have precipitated out in the space afforded by the air voids or fractures.  There has been a lack of correlation shown between the amount of observed ettringite filling voids and cracks and the amount of expansion, and there is also believed to be more than one type of ettringite, and not all are expansive (DePuy 1994).

Thaulow and Jakobsen (1997) list the following three microscopic features that can be used in the diagnosis of external sulfate attack:

• Surface parallel cracks in the cement paste and along aggegate-paste interfaces.
• The presence of gypsum and/or excessive amounts of ettringite in voids, cracks, and paste.
• The dissolution and decalcification of cement paste.

Surface parallel cracks run parallel to and near the concrete surface, and are commonly filled or partly filled with gypsum. The cracks typically traverse through the cement paste, following aggregate surfaces.  According to Thaulow and Jakobsen (1997), gypsum can be distinguished by its texture and birefringence when observed in crossed polarized light.   If the gypsum has precipitated in the paste, an SEM with an EDS will be useful in identification.  The presence of surface parallel cracking must be present if external sulfate attack is the mechanism at work, as it is an indicator of near-surface paste expansion.  Thaulow and Jakobsen also emphasize that an external source of sulfate must be present.

Bickley et al. (1994) describe other techniques that can be used to identify the type of deterioration, its cause, and the type of cement.  Chemical analysis by Inductively Coupled Plasma (ICP) spectrophotometry, LECO combustion, ion chromatography, microstructural/microchemical examination by scanning electron microscope with energy dispersive spectrometry (SEM-EDS), mineralogical/phase analysis by x-ray powder diffractometry (XRD) and thermogravimetric analysis/differential thermogravimetry (TGA/DTG) can be used on the concrete samples.  Characteristic XRD spacings for thaumasite occur at 0.956 nm and 0.552 NM, while those for ettringite are at 0.973 NM and 0.561 NM

Diagnostic Features for Internal Sulfate Attack

Internal sulfate attack is even more difficult to diagnose than external sulfate attack.  Map cracking, particularly in the vicinity of joints, the observation of a white precipitate, evidence of concrete expansion, and concrete paste disintegration are all signs of potential sulfate attack. Unfortunately, these same manifestations are consistent with other distress mechanisms as well.

It is useful to determine the total acid soluble sulfate content of the concrete to determine if it is significantly greater than that present in the original concrete.  If excess sulfate concentrations are measured, it indicates that the source of sulfates is likely external. On the other hand, if the sulfate content is similar to the expected total sulfate content of the concrete, the sulfate attack may be from internal sources of sulfate.  But in some cases, the excess sulfate may have been leached from the concrete and thus this testing cannot be used to conclusively establish that internal sulfate attack is at work. 

Internal sulfate attack due to delayed ettringite formation can be identified as an expansion of the concrete cement paste matrix.  This is characterized by peripheral cracks or gaps around the coarse aggregate particles (Skalny 1996).  The cracks around the aggregates are proportional in width to the size of the aggregate, consistent with a linear expansion of the cement paste.  In the intermediate stages, the reaction produces great expansion in the cement paste, resulting in progressive distention, warping, and fracturing of the concrete.  The cement paste tends to expand away from the aggregate particles, and the resulting opening can become filled with secondary ettringite deposits.    The identification of early DEF related cracks in the cement paste is difficult given that the cracks are often small, requiring an SEM for identification.  The question then arises as to the source of the crack (i.e., is the cracking the result of paste expansion or a result of the vacuum desiccation of the specimen in the SEM?).  The use of a LVSEM or ESEM would be helpful in this regard.

Gaps between the paste and aggregate are thought to be one of the most distiguishing microscopic features of concrete affected by DEF, a form of internal sulfate attack (Thaulow and Jakobsen 1997).  The gaps may be empty, partially filled with crystalline ettringite, or with a massive, poorly crystalline ettringite.  Although some have speculated that the growth of ettringite created the gaps, Thaulow and Jakobsen (1997) believe that the gaps first formed due to expansion of the paste and secondary ettringite formation later filled them.  It is noted that gaps can also form around aggregates due to paste expansion resulting from freeze-thaw of critically saturated paste.

Another distinguishing feature of internal sulfate attack is that, in contrast to external sulfate attack, gypsum is not normally formed during DEF (Thaulow and Jakobsen 1997).

Diagnostic Features for Corrosion of Embedded Steel

There are several methods available to determine the extent to which a concrete structure has undergone corrosion of reinforcement and the secondary deterioration that has been caused by corrosion.  The simplest method is to conduct a survey using a trained professional to examine the structure and document the location of cracks, spalls, delaminations, and other features of deterioration (Perenchio 1994; Erlin and Hime 1987). Cracks on the surface over reinforcing bars, spalls with steel reinforcement at the bottom, delamination, and rust colored staining are all signs of corrosion distress.

Chapter 4 of ACI 222R provides an excellent summary of diagnostic procedures available for identifying corrosive environments and active corrosion in concrete (ACI 1989). In addition to visual surveys, one recommended tool is a pachometer that can be used to locate steel and determine the thickness of the concrete cover.  The thickness of the cover is important as it has a direct influence on whether corrosion is likely to develop in a corrosive environment.

A delamination survey is recommended as part of the survey, in which delamination detectors are used to sound out the concrete surface.  This survey consists of striking the concrete surface with a metallic object such as a rod, hammer, or chain.  More sophisticated equipment such as a Delamtect can be employed to complete a thorough survey (ACI 1989).  Concrete that is of lesser quality or damaged will make a hollow sound when hit with the object (Perenchio 1994).  Visual and delamination surveys are used together to determine the extent of potential corrosion distress, identifying areas that are candidates for an in-depth study.

There are two wet chemical analyses that are commonly performed on concrete to determine the chloride ion concentration.  One is used to determine water-soluble chloride (ASTM C 1218) and the second to determine acid-soluble chloride content (AASHTO T 260).  In each case, a concrete specimen is obtained and ground to powder, typically using a rotary hammer to avoid the use of cooling fluids.  The most common test is the acid solubility test, in which the pulverized concrete is dissolved in nitric acid.  The water-soluble method has seen limited use because it is sensitive to several factors, including time of leaching, temperature, and degree of grinding.  For the same concrete, the average chloride content measured by the water-soluble method is typically 75 percent to 80 percent of its acid-soluble chloride content (ACI 1989).

Chloride content is most often expressed as the chloride ion (CL) present as a percentage of the mass of cement.  It is recognized that the cement content of in-service structures is not always known so, alternatively, the chloride content can be expressed in terms of percent or weight of chloride ions per volume of concrete.  It is very important that the chloride content being reported is defined according to the method of testing.

Concrete cores may also prove useful in determining the extent to which a structure has undergone corrosion of reinforcement.  Cores can be used to verify whether delamination or debonding of the concrete and steel has occurred.  Concrete cores can also be used to determine the depth of the steel to evaluate whether adequate cover was present.  If dowel lockup is suspected after measuring joint movement, coring through a dowel can verify whether it can still move freely.  A petrographic evaluation of concrete cores can also be used to reveal if the void structure is adequate and whether other deleterious mechanisms are at work in addition to corrosion.

Carbonation can be a contributing factor in corrosion of embedded steel and therefore it is useful to determine the depth of carbonation.  Determination of the depth of carbonation can be accomplished using phenolphthalein stain or other pH indicator, since carbonated cement paste will have a pH significantly lower than that of unaltered cement paste.

Approach to Data Analysis and Interpretation

The interpretation and diagnosis of MRD relies primarily on information collected during laboratory investigation, supplemented with information collected during the review of the records and visual assessment of the pavement surface.  When diagnosing a concrete distress, often there is no clear answer as to which distress mechanism caused the failure as multiple mechanisms are observed.  This makes it difficult to determine which mechanism(s) might be responsible for the initial deterioration versus those that occurred after the fact as opportunistic distress mechanisms.  Various types of distress mechanisms can occur simultaneously in concrete and each can incrementally contribute to the ultimate failure of the material.  This fact must be taken into account when evaluating MRD in concrete pavements.

In approaching laboratory diagnosis of MRD, the analyst must put aside preconceived notions as to what the MRD might be.  Instead, diagnosis should be approached through systematic data collection, linked to a process of elimination.  A general philosophy of "asking the material questions" must be adopted where the analyst determines which diagnostic features are identifiable within the concrete.  For example; “Are there microstructural features indicating AAR?” or “Is the air-void system adequate for the concrete service conditions?”  After examining the concrete and noting all available information, the analyst can only make an educated judgment as to why the material failed.  In some cases, there will be indications of a clear cause while in other cases, there may be multiple mechanisms at work making it difficult to determine precisely which factor is primarily responsible. 

The approach taken, as presented in the guideline in Volume 2 of this Final Report, is to use a series of four flowcharts to systematically guide the analyst through the evaluation.  The analyst inspects the concrete using the methods described in the previous sections, guided by the hierarchy of questions presented in the flowcharts.  The responses to the questions presented in the flowcharts determine what analytical procedures will be performed.  As the analyst moves through the flowcharts, there is the potential for more than one MRD being identified. 

In the procedures described in the guideline, the analyst first performs a visual inspection of the concrete to assess its general properties. In most cases, visual inspection helps build a hypothesis of failure, but does not itself directly lead to the cause of distress.  Most of the observations will be confirmed at a later stage using optical microscopy or an SEM examination.  However, there are some general conditions, with visual diagnostic features, known to contribute to, or directly cause, distress. 

After completing the visual inspection, the concrete is examined with the stereo OM and/or staining techniques to assess the paste and air-void system.  Quantitative measurements of the air system parameters should be obtained using methods consistent with ASTM C 457.  The process may continue, analyzing the paste fraction using the petrographic OM, the SEM, chemical analysis, and/or the XRD.

Having completed the stereo OM and staining evaluations, the analyst can begin the process of assessing the condition of the aggregate structure.  Aggregate quality is fundamental to concrete performance and any degradation of the aggregates should be closely scrutinized.  To complete this analysis, it may be necessary to use the petrographic OM, SEM, chemical analysis, and/or XRD. 

The final flowchart presents the procedure for identifying in-filling material.  The identification is accomplished by using staining techniques, petrography, and/or x-ray microanalysis using the scanning electron microscope.  The identification is straightforward using staining techniques, yielding a yes/no answer whereas x-ray microanalysis yields a full chemical analysis that can be used to identify the specific material.  Optical petrography can also be used, but requires more skill and training, as the optical properties are more difficult to discern.  ASTM C 856 includes a complete table of common minerals found in PCC and their optical properties.  Regardless of the composition of the in-filling material, the analyst must consider the following two things to confirm the MRD type based upon identification of the in-filling material:

• The presence of secondary deposits alone does not mean the associated MRD is the cause of the observed distress.  The presence or absence of other diagnostic features to positively identify the associated MRD must be verified.



• A cause and effect relationship must be established between each possible MRD and the observed distress.  For example, the diagnostic feature of ettringite deposits in air voids, by itself, is not evidence of sulfate attack.  There must be other evidence such as cracking, paste expansion, or gypsum corrosion.  Also, secondary deposits of ettringite are commonly thought not to be expansive.  In the end, more evidence other than ettringite in the air voids is required before identifying sulfate attack.  The same caution must be exercised for all MRD types.

As the analyst works through the flowcharts to complete the laboratory analyses, he/she notes all possible MRDs identified and then consults tables that summarize the common diagnostic features associated with each MRD. These tables summarize the principal, common, diagnostic features for each MRD, as characterized by laboratory methods and as seen during field evaluation.  Included in these tables are the diagnostic features, methods of characterization, and specific comments relative to either the observed characteristic or the test method.  The tables are intended to serve as a quick reference during the diagnosis of MRD by assisting the analyst in identifying other diagnostic features to help confirm the presence of a particular MRD.  The tables also serve to tie together the results of visual inspection performed in the field, with laboratory inspection and diagnosis.

In the end, it is not always possible to identify a single MRD as the cause of the observed distress.  This conclusion should not be viewed negatively, but instead as a recognition that on many occasions more than one MRD may be active in a distressed concrete pavement, making absolute identification of the primary distress mechanism difficult or impossible.  In such cases, the analyst should list all possible MRD mechanisms and assign a relative rating as to the likelihood of each being responsible for the observed distress.  The rating scheme should be simple and subjective, possibly along the lines of a scale ranging from highly unlikely, unlikely, possible, probable, to highly probable.  In this way, the analyst can present to the engineers and other interested personnel what he/she thinks is the most likely cause(s) of distress while still presenting all possibilities.  In the end, this will help focus the repair/rehabilitation efforts and preventative strategies for future construction without turning a blind eye to other possible causes.

Summary on Data Analysis and Interpretation

The procedures presented in the guideline contained in Volume 2 of this final report are intended to lead the analyst through identification of common MRD types based upon typical diagnostic features.  Although one MRD may be present, the flowcharts have the analyst examine diagnostic features for all MRD types, thereby minimizing the probability of prematurely "zeroing in" on a single MRD without considering all possibilities.  The flowcharts do not address every possible combination of MRD nor do they address every set of possible diagnostic features that may be seen in distressed concrete.  It should be understood that the mechanisms responsible for MRD are complex and may manifest themselves differently under different conditions and the exact nature of a given distress may vary.

It is very common to observe multiple distress mechanisms at work when examining distressed concrete.  When multiple distresses are present, it is clearly more difficult to diagnose an absolute cause of failure and the analyst should be cautious when trying to do so.  Such an attempt will almost invariably require some assumptions of the order, magnitude, or cause of the earliest failures.  Likewise, analysts and engineers are cautioned from trying to diagnose MRD based on laboratory results alone.  However, a complete laboratory analysis, when combined with field evaluation data and construction and service records, will help accurately describe the condition, environment, and performance of the concrete pavement in question.  From this type of broad-based evaluation of a MRD problem, conclusions about the source of distress can often be reached.

3.3 Background for Treatment and Prevention of MRD in Concrete Pavements

To this point, this Final Report has focused on the identification of MRD in deteriorating concrete pavements.  In an ideal world, material testing and other preventive measures would eliminate the occurrence of MRD.  However, as evidenced by the many miles of pavements exhibiting durability problems, it is difficult to correctly identify and address all potential durability problems.  Hence, means of addressing such problems need to be established and understood.  And due to the limited effectiveness of treatment and rehabilitation strategies, it is imperative that strategies are implemented to prevent MRD in newly constructed pavements.

Background on Treatment and Rehabilitation Methods

Materials-related problems are becoming increasingly identified as the cause of concrete pavement deterioration.  In many cases, pavements affected by MRD would likely be allowed to deteriorate significantly before the MRD is identified, leaving replacement of the deteriorated concrete as the only option to restore serviceability.  However, a few methods that allow the treatment of in-place materials are now available.  These treatments have evolved from a better understanding of the mechanisms of MRD and from improved techniques that permit early identification of MRD.  As diagnostic technologies improve, better and more timely treatment methods are sure to follow.

Methods for addressing MRD can be broadly categorized as either treatment methods or rehabilitation methods.  Treatment methods are defined as those methods designed to prevent further development of the distress or to reduce its rate of progression.  Examples of treatment methods range from the application of chemicals to alter adverse reactions to the use of joint and crack sealants to prevent excess water infiltration.  Rehabilitation methods, on the other hand, are defined as those methods designed to remove deteriorated areas and to maintain serviceability of the pavement.  Such methods include patching of small, deteriorated areas or more comprehensive methods such as overlays or reconstruction.

This section presents the available treatment and rehabilitation methods for pavements with MRD and the recommended approach for treatment developed in this study.  The discussion within this section includes applicability of the available methods and materials, current practices, laboratory and field performance results, and specific recommendations.

Available Treatment Methods

Treatment methods include chemical treatments, joint and crack sealing, crack filling, surface sealing, retrofitted drainage, drying, and restraint.  The purpose of the various treatments is either to arrest the mechanisms that cause the distress or reduce or inhibit the progression of the distress mechanism.  The different treatment methods work in a variety of ways.  For example, chemical treatments affect the reactions that cause the distress.  Retrofitted drainage, on the other hand, provides its effectiveness by limiting the amount of available water in the pavement system that would otherwise exacerbate the distress.  To be effective, treatment measures need to be performed during the early stages of the deterioration process.  If freeze-thaw deterioration has progressed to the moderate-severity level, treatment methods will no longer be effective, and rehabilitation methods must be used instead.

Chemical Treatments

Chemical treatments are often used in fresh concrete to prevent reactions and, in some cases, can also be effective as treatments on existing concrete pavements.  The main drawback of applying such treatments to existing pavements is the difficulty in achieving penetration of the chemical treatment to a sufficient depth in the slab to where it is effective.  With fresh concrete, the chemical treatment can simply be mixed with the other constituents and spread throughout the concrete.  Since the treatment can only be applied to the surface of existing pavements, to be effective the chemical must be able to penetrate the surface.  While multiple applications can increase the depth of penetration, such treatments will never be able to penetrate the entire depth of the slab.

Chemical treatments are designed to mitigate the destructive reaction that causes the distress.  On existing concrete pavements, chemical treatments have been used to:

• Arrest the existing reaction.
• Alter the existing reaction.
• Activate another reaction to offset the existing reaction.

The common component to all of these treatments is a chemical reaction.  As a result, chemical treatments are only effective on distress types that are caused by reactions, such as ASR and sulfate attack.  Other distress mechanisms, such as freeze-thaw deterioration and salt scaling, do not involve a reaction and thus cannot be addressed through chemical treatments.

An example of one of the more promising chemical treatments is lithium salts, which can be used as a treatment for addressing ASR.  Lithium salts were first found to be an effective additive to fresh concrete to prevent abnormal expansion due to ASR (Stark et al. 1993).  They have since been found to be effective in laboratory testing of concrete samples and are beginning to see limited use in the field, although they have yet to be proven effective for widespread use.  There are currently several on-going experimental projects being conducted to evaluate the effectiveness of lithium salts, and early results have been favorable (Stark et al. 1993; Johnston 1997).  As previously noted, the major limitation for field applications is achieving penetration of the lithium solution through the full depth of the concrete slab.

One study tested a series of specimens, which included variations in the amount of expansion allowed before treatment, the type of treatment solution, and patterns of soaking and drying (Stark et al. 1993).  The addition of lithium solutions into hardened mortar exhibiting large expansion due to ASR was found to reduce further expansion, whereas the control specimens continued to expand (see figure 1-7).  Of the three types of treatment solutions, LiOH solutions were more effective in controlling expansion than either Li₂CO₃ or LiF solutions.  As shown in figure 1-8, the expansion nearly leveled off after the sample was soaked in the LiOH solution.

However, it should be noted that the long-term effects of lithium salts have not been studied; the laboratory tests were only 25 months long.  Preliminary results from another study also show signs that lithium salts are effective in field applications (Johnston 1997).  That study avoided the use of LiOH due to safety concerns and the ability of OH ions to exacerbate the ASR.  More recent successes in field applications have been reported at a workshop sponsored by the ACI and the FHWA entitled, Concrete Durability: ASR and Other Deterioration Mechanisms.  At the pilot workshop held in Baltimore on September 11 & 12, 2000, David Stokes of the FMC Corporation reported on successful field applications of lithium nitrate on existing pavements in an attempt to arrest expansion due to ASR.  Long-term monitoring of the cited test sites have not yet been done, so the effectiveness of the treatment could not be evaluated.

The addition of chloride ions offers a potential treatment method for sulfate attack regardless of the source of sulfate ions.  Ettringite has been found to dissolve in the presence of chloride ions, particularly NaCl (Attiogbe et al. 1990; Marks and Dubberke 1995).  Laboratory testing of concrete cores containing ettringite confirmed that treatment with NaCl can dissolve ettringite.  However, this process initially involves further expansion of ettringite before it dissolves.  Of particular concern is the potential harmful effect that chloride ions can have on the pavement, as NaCl can result in scaling of the concrete surface as well as corrosion of embedded steel.  Further, as described previously under deicer attack, salts can also result in the dissolution of calcium hydroxide, which may increase paste porosity and thus permeability.

Chemical treatments often require special mixing, curing, and application techniques.  The manufacturer’s guidelines for each product should be followed closely.  When using a chemical treatment, an important consideration is achieving penetration of the chemical into the concrete.  Achieving penetration through the full depth of the concrete is impossible, but careful application techniques can help increase the effectiveness.  Thorough application of the chemical around and within joints and cracks, where MRDs are often concentrated, can be beneficial.  Repeated application of the chemical treatment can also help increase the effectiveness of the treatment.  Another important consideration when using chemical treatments is safety.  Products that can be potentially harmful should contain special safety considerations; the manufacturer should be contacted if safety guidelines are not clearly identified.

Joint and Crack Sealing

Moisture is key to the development of many durability-related distress types.  As a result, these distress types are often observed to be much more severe at joints and cracks, where moisture can penetrate the pavement surface.  Sealing joints and cracks can help reduce the intrusion of moisture into the pavement.  However, it should be recognized that it is impossible to completely eliminate the intrusion of moisture into joints and cracks, especially for an extended period of time.  For example, a Minnesota study found that the amount of infiltrated water (measured using tipping buckets at drainage outlets) returned to the same levels within 2 weeks after resealing (Hagen and Cochran 1995).

Figure 1-7. Effect of treatment with LiOH solution (Stark et al. 1993).

 Figure 1-8.  Effect of various lithium treatments (Stark et al. 1993).

Consequently, many designers would argue about the effectiveness of any method proposed to eliminate moisture.  Although the amount of water from surface infiltration can be significantly reduced, it can never be completely eliminated.  In addition, the underlying subgrade remains continually moist, even when the most elaborate drainage system is used.  Furthermore, water can also enter the pavement from other sources, such as laterally from ditches and upward from the groundwater table.  Nonetheless, it is still believed that there is some value in taking measures to reduce the amount of available moisture.  However, such methods should be used with caution and are recommended only under certain conditions.

Sealing of joints and cracks is most effective at reducing distresses that initiate at these discontinuities.  Freeze-thaw deterioration of aggregate (D-cracking) is a prime example.  Freeze-thaw deterioration of aggregate typically initiates at joints and cracks where water is allowed to infiltrate, so it only makes sense that measures taken to prevent the intrusion of water can be effective.  Sealing joints and cracks will limit the amount of infiltrated water.  However, the effectiveness of sealing depends on several issues:

• How much water will be prevented from infiltrating by sealing the joint or crack?
• What other sources are contributing moisture and what is the volume of water from other sources?
• How long will the sealant be effective at reducing the infiltration of water?

Sealing joints and cracks will reduce the amount of infiltrated water, although it will not completely eliminate infiltration.  The effect (if any) that this reduction in moisture infiltration will have on reducing freeze-thaw deterioration is debatable.  Studies have shown that even one freeze-thaw cycle can be as damaging as multiple cycles (Janssen 1985).  To be totally effective, sealing must totally eliminate the infiltration of moisture so that even one saturated freeze-thaw cycle is prevented.

Sealing joints and cracks can also help reduce the penetration of sulfate ions that are responsible for sulfate attack.  This method is only viable if the sulfate ions are from an external source applied to the surface; it does not provide any benefit if the source of the sulfate ions is from within the portland cement or from the underlying subgrade.  As a result, methods to limit the amount of available water in the pavement system are not as effective as other treatment methods.

Crack Filling

Crack filling must first be differentiated from crack sealing.  Crack filling refers to the filling of surface cracks (e.g., map cracking) with a material that penetrates into cracks, and not the sealing of individual full-depth transverse and longitudinal cracks.  The purpose of crack filling is not to prevent intrusion of moisture and incompressibles but rather to strengthen the concrete pavement. Crack fillers penetrate into cracks and effectively "glue" the concrete pieces together.

An example of a crack filling material is high molecular weight methacrylate (HMWM).  HMWM strengthens the concrete by filling the cracks and bonding the cracked concrete pieces together.  Such treatments have been most effective when applied to cracks that are wide enough for the material to penetrate (Engstrom 1994). Freeze-thaw deterioration of aggregate and ASR are examples of distress types that might be effectively treated with crack fillers.

Although the cracking associated with freeze-thaw deterioration is not confined to the surface, crack fillers can still provide some benefit.  Crack fillers should only be applied at joints and cracks where freeze-thaw deterioration occurs; there are no benefits to applying them to the entire pavement surface. Field experiments found that HMWM were effective for up to 18 months, which implies that reapplication at such intervals may be required (Engstrom 1994). Several products were tested in the field experiments; the results revealed that Transport T70X and 3M 4R Concrete Restorer performed better and were less costly than the Sika product (Engstrom 1994).  Cores taken at treated areas indicate that HMWM penetrated cracks up to 75 mm deep.

HMWM has also been used successfully on pavements affected by ASR.  The map-like cracking pattern produced by ASR typically extends only 50 to 75 mm below the surface.  In one study, cores taken on State Route 58 near Boron, California found the HMWM penetrated cracks up to 50 mm deep, the maximum depth of the surface cracks (Stark et al. 1993).  On an experimental project on I-80 near Winnemucca, Nevada, significantly lower midslab and joint deflections (11 and 44 percent, respectively) were measured just one day after HMWM application (Stark et al. 1993).

Crack filling with an HMWM requires some special considerations (Engstrom 1994).  First, all bituminous patches should be removed because HMWM deteriorates the asphalt.  In addition, all cracks should be thoroughly cleaned by airblasting in order to help achieve penetration into cracks.  Finally, the treatment should be covered with sand within 20 minutes after application in order to ensure good skid resistance.

Surface Sealing

Surface sealing or coating helps prevent the ingress of moisture into the pavement, which can prevent initiation or limit the extent of moisture-induced distress.  This treatment method forms a penetration barrier on the pavement surface, which expels moisture much like wax on a car. Sealers can also reduce or prevent the ingress of oxygen, carbon dioxide, chloride ions, sulfate ions, and other constituents that contribute to damaging reactions.

Concrete sealers can be divided into two categories—coatings and penetrants.  Coatings form a film on the pavement surface, whereas penetrants are designed either to fill the pores or line them with a water-repellent substance (Campbell-Allen and Roper 1991).  Examples of surface sealers include silane sealants, penetrating oils, and two-part resins.

Sealers have proven to be effective in laboratory testing where concrete samples can be sealed on all sides.  However, as with chemical treatments, sealers can only be applied to the surface of existing pavements. Moisture and other constituents can still penetrate the concrete vertically through the bottom and laterally through the sides of the slab.  Sealers are thus most effective at limiting constituents that infiltrate from the pavement surface, such as chloride ions from deicing salts.  Surface sealers are more effective when used in conjunction with other methods designed to address moisture-related distress.

As with other methods designed to eliminate the intrusion of moisture and deleterious constituents, concrete sealers can never be totally effective.  Although they will significantly reduce the amount of water that penetrates the surface, water can still enter the pavement from other sources.  The usefulness of surface sealers is more questionable because some MRD, such as freeze-thaw deterioration of aggregate, initiate at the bottom of the slab and propagate upward.

Nonetheless, surface sealers have been used to combat the effects of freeze-thaw deterioration.   They have been more effective in laboratory tests (where the sealer completely coats the sample) than in field experiments.  Laboratory testing (later confirmed through field testing) found that water-based and solvent-based silane sealers slowed the rate of deterioration, whereas penetrating oils and two-part resins were not as effective (Janssen and Snyder 1994). Another field experiment using silane sealers indicated mixed results, although they were found to be more effective on pavements with less deterioration (Engstrom 1994).

Surface sealers have proven to be more effective for addressing freeze-thaw deterioration of the cement paste.  Unlike freeze-thaw deterioration of aggregate, where deterioration is worse at the bottom of the slab, freeze-thaw deterioration of cement paste typically results in more damage at the pavement surface. 

Sealing the pavement can also help prevent the penetration of sulfate ions from external sources. Practically, however, only the pavement surface can be completely sealed, and the external sources of sulfate ions (groundwater, soils) make contamination of the bottom of the slab more likely. For this reason, the benefits of surface sealing are questionable for external sulfate attack.

The use of surface sealers for controlling ASR is even more questionable because studies have shown that even water in the vapor phase (relative humidity greater than 80 percent) is sufficient to cause swelling of the gel product (Stark et al. 1993).  In one study, the use of a silane surface sealer was found to have little to no meaningful effect (Stark et al. 1993).  Although the surface sealer did prevent moisture transfer in the liquid phase, it did not prevent moisture transfer in the vapor phase.

For salt scaling, the only feasible treatment method is the application of a surface sealer.  Elimination of the destructive component, namely the chloride ions within deicing chemicals, is not a possibility because they are required for safety (although a nonchloride deicer could be used instead).  Surface sealers must prevent the ingress of moisture and chloride ions into the pavement without disrupting the safety characteristics of the pavement.  This method should only be used for pavements exhibiting limited amounts of low-severity scaling unless previous steps are taken to address the higher severity scaling.

As with chemical treatments, the composition of surface sealers is continually changing as further advancements are made (Campbell-Allen and Roper 1991).  The manufacturer should be consulted to ensure that the appropriate products are used and that proper application techniques are followed.  A clean, dry surface is required to ensure good bonding and thorough penetration of the sealer.  Airblasting or sandblasting may be used for this purpose.  Diamond grinding is a good technique to provide these surface properties, as well as to remove surface irregularities.  Traffic should not be allowed on the pavement until the sealer has fully penetrated the pavement and evaporated.  For silane sealers, a delay of 20 to 45 minutes is typically required (Engstrom 1994).

Retrofitted Drainage

The addition of retrofitted drains will, in theory, remove excess moisture from under the slab and at joints and cracks, which would assist in slowing or delaying MRD.  In reality, however, water cannot readily move through a dense-graded base (typically found in many older, deteriorating pavements) to the retrofitted drains at the edge of the pavement. Therefore, the effectiveness of retrofitted drains is reduced, and other means of moving water from the underlying layers to the drains must be provided.

Moisture-induced distresses generally initiate and progress more quickly at the bottom of the slab, which is exposed to moisture for prolonged periods.  Because the bottom of the slab is not exposed to certain environmental effects, even a light rain can saturate the underlying layers, which will then remain saturated for prolonged periods.  Providing a means to remove moisture at the slab-base interface will help shorten the time the pavement is exposed to moisture.  An important consideration is the permeability of the layers beneath the slab.  Studies have shown that if water cannot move from within the underlying layers to the retrofitted drains, the drainage system will be only marginally effective at removing water and even less effective at reducing moisture-related distress (Smith et al. 1996).  Consequently, retrofitted drainage should only be considered where water can move to the drain, and even then its effectiveness is questionable.

In order to be effective, retrofitted drainage must be applied during the early stages of deterioration.  By the time moisture-related distresses are apparent on the surface, deterioration at the bottom of the slab has often progressed to the point where retrofitted drainage will no longer be effective.

Drying

As the term implies, drying refers to completely removing moisture from the concrete pavement.  Moisture plays a key role in many durability problems, so drying of the concrete will delay the progression of such distresses.  Unfortunately, it is impossible to keep water out of the pavement, and the pavement will eventually become resaturated.  Nonetheless, complete drying can offer prolonged benefits by altering the reaction or changing the reaction products.

Drying has been found to produce some beneficial and irreversible effects for some durability problems.  A prime example is ASR, in which a substantial portion of the alkali hydroxide in the pore solution becomes fixed upon drying and does not return to solution even after resaturation (Stark et al. 1993).  The beneficial effects of drying can also be long lasting.  Laboratory studies have shown that the rate of ASR is substantially diminished by drying even with prolonged soaking and will never be completely reversed (Stark et al. 1993).  This phenomenon, which was discovered by accident but later confirmed through laboratory testing, is believed to be the result of carbonation.  The major shortcoming of this technique is that there is no practical way to sufficiently dry pavements in the field.

Restraint

Restraint can be used to combat the expansive forces developed as a product of some durability problems.  Both ASR and sulfate attack produce reaction products that are larger in volume than the original products, thus causing expansion and cracking of the concrete pavement.  Restraint induces internal compressive stresses in the concrete that prevent the development of cracking, in much the same manner as a prestressed concrete pavement.

Physical restraint has shown some success in controlling the expansive forces produced by ASR.  Three-dimensional restraint has proven effective for controlling expansion in concrete columns but is not feasible for highway pavements.  Laboratory testing on miniature pavement sections revealed low levels of one-dimensional restraint helped control microcracking but higher levels actually accelerated expansion and internal cracking (Stark et al. 1993).  The study also concluded that one-dimensional restraint has a significant positive effect on transverse and vertical strain; finite element modeling confirmed that one-dimensional restraint has a positive influence of restraint in three dimensions (Stark et al. 1993).

The major shortcoming with the use of restraint in actual pavements is that it is extremely expensive and that (in order to be effective)the restraint must be applied early in the life of the pavement before the extent of potential damage has been observed. For pavements, three-dimensional restraint is not feasible; uniaxial restraint offers the only practical approach, yet it is still an expensive proposition.  The degree of restraint is also an important consideration, as too little restraint will not prevent cracking perpendicular to the direction of restraint and too much restraint can cause cracking parallel to the direction of restraint.

Available Rehabilitation Methods

Rehabilitation methods, as opposed to treatment methods, follow a totally different approach to addressing MRD.  Rather than altering the development or progression of MRD, their purpose is merely to repair deteriorated areas to maintain serviceability and possibly extend the life of the pavement.  Rehabilitation methods include partial- and full-depth repairs, slab replacement, diamond grinding, overlays, reconstruction, and recycling.  Some pavement engineers believe that rehabilitation methods are not effective means to repair MRD because they do not address the cause of the distress.  However, when used under the right conditions (such as to maintain the serviceability of the pavement), repair methods can prove cost effective.

Previous | Table of Contents | Next