U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
|This report is an archived publication and may contain dated technical, contact, and link information|
Publication Number: FHWA-RD-01-164
Date: March 2002
This appendix has specific information about the laboratory techniques used to analyze concrete specimens. This information is provided for quick reference and the reader is directed to the literature recommended in this guideline for detailed descriptions of the test methods. The following laboratory methods are discussed:
B.1 Determination of Water-to-Cement Ratio (w/c)
B.2 Staining Concrete to Identify Sulfates and Alkali–Silica Reactivity (ASR)
B.3 Optical Microscopy (OM)
B.4 Scanning Electron Microscopy (SEM)
B.5 X-ray Diffraction (XRD)
Determination of w/c by chemical methods involves independently determining the apparent water and apparent cement content, and then calculating w/c. The free water is determined by a combination of saturation/drying and weighing methods consistent with the methods described in ASTM C 642 Standard Test Method for Specific Gravity, Absorption, and Voids in Hardened Concrete. Determining the combined water requires drying the concrete at 550 °C and recording the weight loss.
These procedures require the following major equipment or equivalent:
Cement content can be determined using one of the methods described in ASTM C 1084 Test Method for Portland Cement Content of Hardened-Cement Concrete. One method accomplishes preferential dissolution of the cement in concrete using a cold-HCl solution. The other method presented, and the one adopted for this guideline, uses methanolic maleic acid (MMA) to digest the cement. This procedure requires less equipment then the cold HCl method. The major required equipment is as follows:
Clearly, accurate determination of the w/c value in a failed concrete section is an important part of understanding why the material failed. This is of course due to the known relationships between w/c, strength, and durability. Methods for determining w/c have been developed based on chemical laboratory methods. These methods involve independently determining the water and cement content and then calculating w/c. The free water is determined by a combination of saturation/drying and weighing methods consistent with the methods described in ASTM C 642 Standard Test Method for Specific Gravity, Absorption, and Voids in Hardened Concrete. Determining the combined water requires drying the concrete at 550 °C and recording the weight loss. Cement content can be determined using methods described in ASTM C 1084 Test Method for Portland Cement Content of Hardened-Cement Concrete. This commonly applied method accomplishes preferential dissolution of the cement in concrete using a cold-HCl solution. However, this solution is known to attack some aggregates resulting in erroneous cement contents (Hime 1993). The method presented in this report is adapted from the method presented by Marusin (Marusin 1981) using MMA to digest the cement. The procedure is presented below.
Step A: Determine Mixing Water (MW) content
Note: Special care and due speed is required for steps 5 and 6 to prevent loss of moisture to evaporation.
MWSSD = (BWSSD + FWSSD) - aggregate absorption
Step B: Determine Cement Content
Step C: Determine Water-to-Cement Ratio
Generally speaking, staining can be carried out in any lab space and may very well be conducted in the petrographic lab if no chemistry lab space is available. The one exception to this is the SHRP uranyl acetate test (Stark 1991). The primary concern with conventional staining is storage of the chemicals. A closed, ventilated, dry, and secure cabinet for storing chemicals is required. Also, a waste stain collection bottle should be used for each type of stain so the unused stain can be reused without re-mixing spent and fresh stain.
It is noted that all of the stains used are hazardous materials and it is therefore necessary that strict policies be developed and followed for handling and disposal. All stains should be clearly labeled and dated.
The required equipment for staining depends on the method being used. Excepting the SHRP uranyl acetate test, standard laboratory beakers, Nalgene storage bottles, nitrile gloves, and safety goggles are all that are required. For staining large sections, a medium depth plastic pan (dishpan) is useful. The SHRP uranyl acetate test requires a special lighting fixture and protective glasses for viewing the stained regions. These lighting devices are available as part of a SHRP testing kit from a variety of vendors.
Staining methods can be conducted in the field but are best carried out in a laboratory where safe and proper storage and disposal of chemicals is accommodated. Four existing staining methods are presented. These include the barium chloride potassium permanganate (BCPP) stain for sulfate minerals in concrete (Poole and Thomas 1975), the recently introduced sodium cobaltinitrite/rhodamine B method for ASR (Guthrie and Carey 1997), the SHRP uranyl acetate method for identifying ASR in concrete (Stark 1991), and phenolphthalein to determine the depth and extent of paste carbonation.
The BCPP stain is very useful for identifying ettringite, gypsum, and anhydrite phases in concrete. The principal staining method 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.
The sodium cobaltinitrite/rhodamine B method is relatively new and as a result, little field experience currently exists regarding its applicability and reliability. Yet some aspects of the technique appear to be very promising, addressing some limitations noted in the uranyl acetate method. The recommended test sequence is as follows:
After drying, the specimens can be visually assessed with a hand lens or using an optical microscope. Regions affected by ASR 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 has been determined that the yellow stained regions were K+Ca+Si±Na reaction products resulting from ASR.
The literature states that pink staining regions were found to be generally less prevalent and isolated in highly deteriorated areas and that they commonly occurred only when extensive yellow staining also existed in the sample. Interestingly, yellow staining was often present in the reaction rims around aggregate while the pink staining was outside this area along fractured faces. Prior to staining, the entire reaction product appeared white, but after staining, it had clearly defined areas of yellow and pink. The pink staining material was identified as being alkali poor Ca+Si gels associated in some way with advanced ASR (Guthrie and Carey 1997). It is hypothesized that the pink staining gels formed from yellow staining gels either through leaching of alkalis or through a reaction with the cement paste. In practice, it has been observed that the rhodamine-B stained the calcium rich hydrated cement paste pink, therefore limiting the usefulness of this stain. As a result, it might be desirable to use the sodium cobaltinitrite alone as an indicator for ASR.
Guthrie and Carey (1997) compared the results of their staining method to the uranyl acetate method. They found excellent agreement between the two methods with the added benefit that their 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 cobaltinitrite/rhodamin B method is that the chemicals are not as rigorously regulated as uranyl acetate and that a UV light source is not required for illumination. Thus conventional optical microscopes (OM) can be used for viewing stained specimens.
In the SHRP uranyl acetate test, concrete is treated with a uranyl acetate solution, which is then rinsed off. This leaves uranyl (UO2+2) sorbed to the negatively charge ASR gel which fluoresces with yellow-green glow under UV light (Natesaiyer and Hover 1992). It is this glow that is attributed to the presence of ASR gel.
Although this test has been very helpful to the practitioner, it has a number of problems associated with it. One major problem is the key ingredient of the stain is radioactive uranyl acetate. Although the concentration of uranyl acetate is relatively small, it is significant enough to warrant special handling and disposal procedures from a regulatory standpoint. Therefore, obtaining permission to receive this material at some labs is not trivial and documentation is often required. Consult with your radiation safety personnel or State agencies governing radioactive materials as to specific laws in your State. Once received, this material must be stored in a closed, ventilated, dry, and very secure cabinet to minimize any chance of spillage. The radiation given off is minimal and radiation shielding is not required in the storage cabinets. Once exposed to uranyl acetate, the concrete becomes contaminated and must be disposed of using acceptable procedures.
Another disadvantage of the uranyl acetate test is that it requires the use of a UV light source to fluoresce the gel. This is problematic as it makes it almost impossible to simultaneously view the gel and other relevant concrete characteristics such as void structure, aggregate, paste, and fractures. This is also a limiting factor in field analysis where bright sunlight may make it difficult for the analyst’s eyes to adjust for adequate viewing of the concrete under UV light.
The final disadvantage is that the uranyl ion is nonspecific, meaning that it will associate with most cations in concrete (Guthrie and Carey 1997). This nonspecificity means that false positives are not uncommon when using the uranyl acetate test, possibly leading one to believe that there is an ASR problem when it actually might not exist.
The use of phenolphthalein as a stain for determining the depth and extent of carbonation is based upon the pH of the cement paste. Any cement paste within the concrete that has a pH greater than 10 will be stained a magenta color. All other constituents including carbonated paste remain unchanged. Carbonation results in reduction of pH in the paste and thus carbonated paste will not stain. The required equipment is the same as the BCCP stain, but instead a 0.5 percent phenolphthalein in methanol solution is used. The procedure is as follows:
The distinct difference in color allows for directly imaging the depth of carbonation using a stereo OM or, if it is a polished slab, a computer scanner.
The most common types of concrete specimens observed in an optical microscope are broken pieces, polished slabs, picked grains, and thin sections. Broken pieces require no sample preparation if examined by stereo OM. This is often very useful for examining fractures, aggregates, and filled voids. Also, reaction products can sometimes be detected and sampled, or picked, for analysis by petrographic OM or scanning electron microscope (SEM). A polished slab is commonly more useful than a broken piece because the cross section of a whole core can be examined. All constituents of the concrete can be seen and examined in detail. Often, staining is applied to the polished slab to reveal specific types of features. Aggregate type and abundance can be readily determined from a polished slab and cracks are easily seen and traced through the concrete. Reaction rims and products are clearly observed in polished slabs. Picked grains are minute pieces of concrete extracted from a specimen, usually while observing it in the stereo microscope. These grains may be analyzed on the petrographic microscope using refractive index liquids or on the SEM. Thins sections are 20-micron-thick pieces of concrete bonded to a glass slide commonly using fluorescent dyed epoxy. These are useful for transmitted light observation on the petrographic OM and backscattered imaging in the SEM.
The major pieces of equipment required are as follows:
Thin sections are often required when examining concrete to allow for identification of components in the concrete microstructure. These specimens require practice to prepare and require more elaborate equipment as compared to polished slabs. The required major equipment is as follows:
Usually, thin sections are epoxy impregnated to fill voids and cracks, improving
the image quality in the OM. In the case of fluorescence microscopy, the
epoxy contains a fluorescent dye that must be mixed in the epoxy prior to
use. Detailed instructions for impregnating specimens and preparing the fluorescent
dye are given in various manuals (Roy et al. 1993a; Walker 1992). The additional
equipment needed to perform epoxy impregnation includes:
The stereo zoom OM is arguably the most useful tool for examining distressed concrete. A typical stereo zoom OM can provide magnifications to 150 to 200x and has a considerable depth of field at any given magnification allowing for observations of cracks, voids, and crystalline grains. It is recommended that all laboratories analyzing concrete obtain a quality stereo zoom OM. When purchasing one, some desirable features are click stops on the magnification control, multiple types of light sources (e.g., gooseneck, ring lights), a sturdy, vibration isolated adjustable boom stand, and a maximum magnification of at least 140x. An automated stage may be useful for manipulating the specimen. Most commercially available stages provide programmability so, if in-house support is available, software can be written to perform air-void system analysis in accordance with ASTM C 457. Digital frame grabbers are also available to capture digital images from the microscope.
The petrographic OM is also very useful but requires considerably more acquired skill as compared to the stereo OM. The petrographic OM can be used to identify various constituents in concrete by passing plane polarized illumination through a prepared thin section. As light passes through materials, it bends, or refracts, with the degree of refraction dependent upon the material. Light traveling through most translucent crystalline materials is broken down into two rays that behave differently, with one being decelerated (extraordinary ray) and one unaffected (ordinary ray). The refraction of these waves within a crystal varies as the specimen is rotated and this refraction results in image contrast that can be used to identify different mineral species. When purchasing a petrographic OM, the key features are:
The choice of lenses is very important and should be thoroughly considered. The lenses are the heart of the microscope and will determine the quality of the image. Some common lens types are achromats, apochromats, and fluorites. The term “plan” is included in the lens type if it has been corrected to remove the curvature of the lens from the resulting image. Generally, buy the best lens you can afford. If fluorescence microscopy is to be used, the highest quality lenses will almost certainly be required. The two common descriptors of a lens are the magnification factor and numerical aperture (NA). Common magnification factors are 3.5x, 5x, 10x, 20x, and 50x. The numerical aperture is a measure of the lens’ resolution capabilities with an increased NA indicating a higher resolution lens.
Refractive index liquids are used to identify materials mounted as grains on a microscope slide. As a beginning set, the following refractive index liquids are recommended:
As with the stereo OM, some form of image recording is required. Petrographic OMs typically support 35-mm cameras or video camera/frame grabber systems.
To perform an ASTM C 457 analysis, some type of automated microscopy system is preferred. A number of small firms offer limited production turnkey systems that perform this test in either a fully automated or semi-automated mode. Any system to be purchased should be compared to the systems described in ASTM C 457 to confirm compliance. In general, a semi-automated approach is almost always desired to allow the operator to make judgments about features in the microscope image. Fully automated systems have been developed to analyze specially prepared samples where the entire matrix has been blackened and the available pores filled with a white powder (e.g., wollastonite dust.) The resulting image is a “binary image” where only two colors exist and the computer can easily make judgments about what is being considered a void. One drawback to this approach is that no information about the degree of void infilling is obtained. If possible, try to purchase a system that will allow for determination of the percentage of filled air voids.
The first step in microscopic evaluation of concrete is sample preparation. Detailed methods for preparing samples are given in a variety of references (Walker 1992; Roy et al. 1993a; ASTM C 457; ASTM C 856). Not all methods or all types of samples need to be analyzed in every case. It is up to the analyst to decide which test will provide insight into the mode of failure. The purpose of this section is to provide an overview of the common types of laboratory procedures used in preparing specimens to investigate MRD in concrete pavements.
Often, for a case of MRD in a highway structure, broken pieces are readily available in addition to cored specimens. Broken pieces can be very useful for identifying secondary deposits in addition to studying the nature of the fractured surfaces. When the surfaces are freshly broken, fresh gel may be seen in cases of advanced ASR. No special sample preparation is required for viewing broken pieces in a stereo OM. To view these pieces in a CSEM, a conductive layer must be deposited as described in section B.4 of this appendix.
Often, material can be picked from the surface of concrete using a metal pick tool. This material can be placed on a glass slide in a drop of liquid of known refractive index. The grains of material can then be viewed on the petrographic microscope to determine the material’s optical properties such as refractive index and birefringence, from which the material can be identified. This process is usually done in conjunction with stereo OM observation. Table 6 in ASTM C 856 lists the important optical properties of common concrete phases.
A very useful way to examine concrete cores is to section a slab from the core and polish one or both surfaces for examination. This provides a large area for examination and also allows for examining concrete in longitudinal and transverse directions. The number of slabs required will be determined by the tests being run. For example, within ASTM C 457 there are provisions for the minimal surface area to be analyzed. To facilitate microscopic examination, the polished slab needs to be as flat as possible with opposite sides being parallel. Stains can be applied directly to the slab, as described in Section B.2, to identify phases in the concrete. Aggregate rock type and distribution can be readily determined for both the coarse and fine aggregate using a polished slab. If the core being sectioned is fragile, it may be necessary to stabilize the concrete with epoxy, Carnuba wax, or 1:5 solution of nylon fingernail hardener in acetone or methanol. The following method is used for preparing concrete slabs for microscopic inspection:
Epoxy impregnation is often used to stabilize concrete, greatly facilitating most sample preparation techniques. In general, this should be done as a last resort because the epoxy impregnation is irreversible. Carnuba wax or 1:5 solution of nylon fingernail hardener in acetone or methanol are substances that can be dissolved away from the concrete, once hardened, making them a preferred stabilization method.
Often, impregnation is done for other reasons. It can be used to fill air voids with a material of known index of refraction or a material containing a fluorescent dye for fluorescent microscopy techniques. It is therefore common to use epoxy for this task. The following method is recommended for vacuum impregnating concrete with epoxy:
Thin sections are not difficult to produce but they are difficult to produce consistently well. There are a number of publications that detail the steps (Walker 1992, Roy 1993) and only a brief summary is given here. The art lies in the final polishing required to bring the section down to the desired thickness. Unfortunately, polishing is irreversible, and a specimen polished unevenly or too far is ruined. All of the sectioning and grinding steps are very important as the final thin section quality is greatly controlled by the uniformity and thickness of the slide before final polishing begins.
Thin sections are usually prepared on standard geological thin section slides using standard geological sample preparation methods. An ultimate thickness of 25 to 30 microns is obtained by first cutting and grinding the section to very near the desired thickness. Then, the section is thinned by hand until the desired thickness and taper is achieved. The latter is important as some petrographers purposefully taper their sections to provide a variety of thickness for observation. For example, thin sections near the standard thickness can be used for routine observations, while thinner areas are used for examining features such as unhydrated cement grains, or dolomite grains in ACR reactive aggregates. The following method is recommended for the preparation of thin sections:
The same specimens studied by optical microscopy can be studied using the scanning electron microscope (SEM). As a convention within these guidelines, the acronym SEM will refer to the general instrument including the high-vacuum conventional scanning electron microscope (CSEM), the environmental SEM (ESEM) and the low-vacuum environmental scanning electron microscope (LVSEM). Common types of concrete specimens for electron microscope observation are broken pieces, polished slabs, and thin sections. All specimens require a conductive coating if examined in the CSEM or in “CSEM mode” using an LVSEM or ESEM. Usually no coating is required for examination at low vacuum by the ESEM or LVSEM.
The SEM is a convenient tool for identifying reaction products and other phases in distressed concrete. The SEM is also very useful for examining cracks, aggregates, and deposits in fractures or filled voids. The SEM forms an image by rastering a focused electron beam across a small rectangular area of a specimen. The electron beam and specimen interact resulting in two electron signals of interest and characteristic x-rays, to be discussed later. As the beam strikes the specimen, electrons from the material are released. These electrons are called secondary electrons, have a very low kinetic energy (<500 eV), and are used to form topographical images of a specimen. Electrons that originate in the electron beam have a very high energy (15-25 keV) and these electrons scatter in proportion to the specimen density at the point of impact. These backscattered electrons are used to form images showing specimen composition. As the electron beam is scanning the specimen area, the electron yield from the specimen is monitored using an electron detector. This electron signal is used to modulate the brightness or intensity of another electron beam scanning inside a viewing cathode ray tube (CRT). This modulation/scanning process maps out an image on the viewing CRT that is an “electronic representation” of what the specimen surface looks like. By design, an SEM produces images with extraordinary depth of field allowing for examination of rough surfaces.
When a point on a specimen is excited with the incident electron beam, characteristic x-rays are emitted from the specimen that are unique identifiers of the elements present in the irradiated volume. By identifying the energy or wavelength of the characteristic x-rays produced using a spectrometer, the elements present are identified. For a given incident beam current, the rate of characteristic x-ray production for a given element will be proportional to that element’s concentration. This effect forms the basis of x-ray mapping and quantitative x-ray microanalysis.Conventional Scanning Electron Microscope (CSEM)
In a conventional SEM (CSEM), the electron beam/specimen interaction occurs in a high vacuum (10-6 torr) allowing for transmission of the incident electron beam and emission of the low energy secondary electrons from the sample with a minimum of collisions between electrons and gas molecules in the chamber (i.e., high vacuum, low number of gas molecules in the chamber). This allows for the optimum operating conditions with a well-defined incident electron beam and a maximum yield of electrons from the specimen. However, this high vacuum creates severe dehydration and cracking of concrete, affecting the ability to make direct observations on cracking. In essence, the CSEM produces an image of altered concrete as compared to what was sampled. In many cases, this does not matter. ASR reaction products and secondary deposits can easily be identified based on their cation ratios. Other constituents, such as aggregates, are unaltered by the vacuum.
In an environmental SEM (ESEM), the electron beam/specimen interaction occurs at elevated pressures (0.2 to 20 torr), which greatly reduces or eliminates dehydration (4.6 torr is the minimum pressure that can sustain liquid water). With the pressure in the chamber set greater than the vapor pressure of water, dehydration and the associated cracking are reduced to near zero. Another benefit of the elevated gas pressure is it allows for discharging of any surface charge by ionization of the gas molecules. This reduces the need for conductive coatings. In combination, this greatly improves the applicability of electron microscopy to concrete imaging.
In the ESEM, collisions between incident beam electrons, emitted sample electrons, and gas molecules in the chamber are common. These interactions have both positive and negative effects. For the emitted secondary electrons from the specimen, the interaction with gas molecules results in ionization of the gas and a “gas amplification effect” that greatly enhances the electron signal strength. Special detectors have been developed to take advantage of this effect and this is one of the fundamental differences between an ESEM and an LVSEM. In an LVSEM, the backscattered electron signal is used for both compositional and topographical imaging. On the negative side, the high pressure environment creates a scattering and subsequent defocusing of the incident electron beam, leading to an uncertainty as to where the electron beam is actually striking the specimen. In imaging, this effect is not a problem as the randomly scattered electrons simply form a background to the electron signal that is easily filtered. In the case of x-ray microanalysis, it is a serious problem as x-rays could be coming from any part of the specimen and an element in the analyzed spectrum may in fact not be present at the assumed analysis point. In the end, the ESEM provides an excellent tool for imaging concrete and a limited tool for performing x-ray microanalysis when performed in the ESEM mode. An ESEM is capable of running at high vacuum for performing microanalysis.
The third type of scanning electron microscope is the low vacuum SEM (LVSEM). This instrument is very similar to a CSEM but has been adapted to also operate at elevated pressures (0.2 to 2 torr). In this environment, liquid water cannot be sustained and free water that is present freezes instantly at 4.6 torr and sublimates at lower pressures. For concrete thin sections or polished sections, there will be very little to no free water present if the specimens are not prepared using water. The water of hydration is chemically bound and will dissipate very slowly in the LVSEM. Although desiccation cracking will ultimately occur, it will be very slow process and will not be seen during a typical analysis. The gas environment of the LVSEM will also eliminate surface charging, eliminating the need for coating.
First, in comparing the different instruments there are obvious differences. Any of the three SEM types described can be used for analyzing concrete. A CSEM will provide the highest possible resolution and the best microanalysis data, but there will be cracking and dehydration, and the need to conductively coat specimens to eliminate surface charging. An ESEM is capable of functioning as an LVSEM or as a CSEM. However, design restrictions required for ESEM operation make the high vacuum performance less than optimum. An LVSEM operates well as a CSEM and allows for the examination of concrete with little to no dehydration cracking but liquid water cannot be sustained.
When deciding to buy an SEM, look for instruments with the capabilities for the largest possible sample size and stage travel. Many current SEM designs allow for samples with a diameter of over 200 mm, making possible observation of polished cross sections of cores. If the SEM is to be used for quantitative x-ray microanalysis, examine the geometry between the x-ray detector port and the specimen surface plane. The key parameter is the detector take-off angle, which is the angle of inclination between the specimen surface plane and the x-ray detector. A high take-off angle is preferred with 35 to 40° being the optimum. Other geometries may suffice and it is best to evaluate each instrument analytically with selected mineral standards before making a final decision. Also inspect the take-off angle of all other detectors. Avoid instruments with a negative take-off angle for any detector. That type of detector take-off angle puts the detector “behind” the specimen, which usually means a decrease in electron gathering efficiency.
Finally, when purchasing an SEM, evaluate the vendor and its performance by calling customers on its user list. Most instrument vendors provide this list upon request. Most importantly, evaluate the field service cost and performance. An SEM purchase is a major investment over the equipment’s 10 to 20 years of service. In addition to the purchase price, expenditures for planned maintenance and supplies will be incurred and, thus, financial plans for addressing those costs before they occur need to be developed. As a rule of thumb, the maintenance cost of an SEM varies between 1 and 5 percent of the original purchase price, per year. This does not include the cost of an operator or analyst. For example, a first estimate of the support cost for a $300,000 SEM/EDS system is between $3000 and $15,000. The lower number represents a situation where all maintenance is conducted by the instrument operator and the high end represents a service contract purchased from the vendor.
To perform x-ray mapping or microanalysis with an SEM, an x-ray analysis system must be purchased in addition to the SEM. The most common x-ray analysis system used on an SEM utilizes the energy dispersive (ED) spectrometer. The ED spectrometer, or ED detector, is an electronic device that determines the energy of individual x-ray photons. As previously stated, when a point on a specimen is excited with the incident electron beam, x-rays are emitted from the specimen that are characteristic of the elements present in the irradiated volume. By identifying the energy or wavelength of the characteristic x-rays produced using a spectrometer, the elements present are identified.
In an ED system, the spectrum of x-rays emitted by the specimen is established by assigning x-rays of similar energy to classification bins. For example, a 1eV x-ray would be in the 0-10eV bin and classified as 0ev. Likewise, a 11eV x-ray would be in the 10-20eV bin and classified as 10ev, etc. This bin process creates an X-Y histogram with bin energy on the abscissa and frequency of occurrence, or counts, on the ordinate. Characteristic x-rays appear as distinct distributions, or lines, superimposed on a slowly varying background. Usually, the identity of an element is determined by matching the families of lines occurring to preset cursor positions. This process is usually accomplished graphically. Once an element is identified, the integrated intensity of any one of its principal lines can be used as a measure of its concentration. The integrated intensity is extracted by defining a region called a window. This window is usually centered on the peak being integrated and adjacent bins either side of the peak centroid bin are summed, forming the integrated intensity for that element.
X-ray mapping is a way of forming an image that is based on characteristic x-ray information and therefore elemental information. In x-ray mapping, as in electron imaging, the incident electron beam is stepped or scanned over an area of the specimen. However, rather than modulate the viewing CRT electron beam with the specimen electron signal, the characteristic x-ray rate signal from pre-selected elements is used. For each element mapped, a window is defined in the ED spectrum. Each element is assigned a color and a separate map is developed for each element analyzed. These maps can be processed later to indicate common assemblages of elements. The intensity of any pixel in an x-ray map represents the relative intensity for that element, and therefore its relative concentration. The term relative indicates that all of the intensities measured in a digital x-ray map are typically normalized to the most intense pixel in the map. Therefore, direct comparison of intensities between different maps can only be done with prior knowledge of the conditions of analysis or by knowing the minimum and maximum raw intensities measured in a specific map. Another important consideration with mapping is that the specimen should be polished. An x-ray map only shows changes in x-ray intensity. That intensity change could be the result of compositional differences but it can also be caused by topographical changes in the specimen surface. Therefore, the surface texture will dominate x-ray maps when mapping unpolished surfaces, resulting in inaccurate interpretation.
Although x-ray mapping is most commonly applied to concrete analysis, x-ray microanalysis is also very useful. The process of microanalysis requires excitation of a phase within the concrete microstructure by the incident electron beam. The x-rays generated will be characteristic of the elements present in the phase and can be used to identify the elements present. This is referred to as qualitative x-ray microanalysis. The intensity of any given x-ray can be measured and the intensity will be proportional to the concentration of that element. By measuring the intensity of a given element, and comparing it to the intensity of a known standard, the concentration of any given element can be calculated. This is referred to as quantitative x-ray microanalysis. Quantitative analysis allows for the absolute identification of phases within the concrete. This is contrasted with x-ray mapping where common occurrences of elements are noted but no measure of elemental concentration is produced.
Most of the steps necessary to prepare a specimen for the SEM are identical to those presented in section B.3 of this appendix. A recent publication details the methods required for sample preparation (Stutzman and Clifton 1999). Some additional steps may have to be employed beyond the sectioning, grinding, and polishing performed to prepare an OM specimen. Most importantly, for specimens being examined in a CSEM, a conductive layer needs to be deposited on the surface of the specimen being viewed. This conductive coating is used to conduct away beam electrons after the incident electron beam strikes the surface. Without the conductive path provided by this deposited layer, a space charge develops on the specimen surface and the incident beam is displaced erratically, often scrambling the image beyond recognition. The coating process will be discussed later. If an ESEM or LVSEM is being employed, the coating process is usually unnecessary as the residual gas in the specimen chamber dispels the surface charge through ionization.
If staining is used, make sure it is done on a specimen not intended for x-ray microanalysis in the SEM as metal ions in the stain may interfere with the analysis. As an example, when using the SEM/EDS for analysis of uranyl acetate stained gels, the uranium M-series x-ray lines obliterate the potassium K-series x-ray lines, making identification of potassium in the gel impossible. Therefore, qualitative ED microanalysis of ASR gels must be performed on nonstained concrete.
Broken pieces make excellent SEM specimens and, if available, the ESEM/LVSEM is the preferred way of examining these specimens. The excellent depth of field obtained on the SEM allows for easy inspection of secondary deposits in concrete. Qualitative microanalysis on broken surfaces is also very useful and will be discussed in a later section. Before placing the broken surface in the SEM, dust off loose fragments with dry compressed air.
Polished slabs are certainly excellent SEM samples if the SEM being used allows for large specimen entry. Some SEMs are limited to specimens of only 25 to 40 mm in diameter resulting in the need for preparing small plug samples. These are sectioned and polished using methods similar to those used in preparing slabs. Polished slabs can make excellent specimens for performing microanalysis given the large, smooth surface area available for inspection.
Thin sections can also be viewed in the SEM providing no cover slip has been used. Often, thin sections for SEM use are prepared thicker than normal and with a fine polish on the exposed side. These specimens are excellent for quantitative microanalysis. Epoxy impregnation will not affect the SEM image.
All specimens require a conductive coating if examined in the high vacuum environment common with the CSEM. Usually no coating is required for imaging in the low vacuum environment of the ESEM or LVSEM. Once a specimen has been coated for viewing in the SEM, it may be unusable for other analysis techniques. Therefore, SEM is usually done last if a CSEM is being used. When a conductive coating is required, there are essentially two methods and every lab should plan on obtaining both.
In carbon or metal evaporation, the specimen is placed in a vacuum chamber along with the coating material and a device capable of heating the coating material to its vaporization temperature. A high vacuum is obtained (10-5 torr) and the coating material is evaporated, coating the entire vacuum chamber and everything in it, including the sample. In a sputtering process, a cathode of the coating material is placed facing a steel or aluminum plate acting as an anode. The sample is placed on the anode plate and a relatively low vacuum is obtained (10-1 to 10-2 torr). Residual gas atoms in the chamber are ionized and this ionization plasma between the anode and cathode causes a transfer of cathode (coating material) atoms to the anode. This deposits a thin layer of coating atoms on the specimen.
When x-ray mapping or microanalysis is to be performed, a conductive layer of carbon is preferred because the coating layer material will not appreciably affect the results of analysis. On the other hand, metal conductive layers do affect the x-ray data. With regards to microanalysis, the metallic coatings interfere in two ways. First, x-rays generated in the conductive coating appear as peaks in the measured spectrum and may obscure other x-ray emissions from the specimen. Second, x-rays generated from phases in the specimen may be severely absorbed by the metallic coating, causing an error in calculated concentrations. However, metal coatings are good for imaging as the heavy metals used typically have a high electron yield.
Relative to a typical sputtering apparatus, evaporation systems are capable of producing a much better coat because of the high vacuum obtained. However, obtaining the high vacuum with a concrete specimen is very difficult and always results in cracking of the specimen. If a high vacuum is not used when attempting to evaporate carbon, a coarse grained or possibly soot type carbon layer is formed. These types of coating are of very low integrity and form poor or no conductive paths.
Often, it is necessary to leave larger concrete specimens in the vacuum evaporation chamber longer to reach an equilibrium vacuum suitable for good coating (10-4 to 10-5 torr). Note that any material placed in a CSEM must not out-gas in the sample chamber and degrade the CSEM vacuum. If this occurs, the CSEM will not function. Therefore, before being placed in the CSEM, every sample must first be capable of being pumped to a vacuum of 10-4 to 10-5 torr in an evaporator. This extensive pumping causes severe cracking in the specimen, making physical characterization of the concrete microstructure very difficult. Using an ESEM or LVSEM for imaging concrete eliminates the need for specimen coating and eliminates this destructive sample preparation requirement.
Broken pieces are best observed using the secondary electron signal. This signal is very sensitive to topography and provides the highest image resolution possible. The backscattered electron signal can also be used but caution should be exercised in trying to interpret image contrast as compositional variance. Although backscattered electrons are not normally topography sensitive, large amounts of surface relief can cause significant variances in the backscattered electron signal. Quantitative x-ray microanalysis can usually not be performed on rough broken surfaces. The quantitative analysis programs assume a zero surface tilt with the incident beam striking the specimen normal to the surface. On irregular or rough surfaces, this is rarely possible. The local surface tilt can be entered in the quantitative analysis program but it is usually indeterminable. Measured x-ray intensities can vary significantly as a function of the local specimen tilt. Qualitative x-ray microanalysis can be conducted on broken surfaces. However, the interpretation of the measured intensities should be done cautiously for the same reasons given above. Also, x-ray mapping of very rough surfaces may result in inaccurate data as the topographical variations will greatly affect the measured x-ray intensity.
These specimens are very good for backscattered imaging, secondary electron imaging, and both methods of microanalysis. Backscattered imaging is most useful as true compositional contrast can be obtained when imaging a polished surface. In addition, the plane geometry of the polished surface satisfies the geometric assumptions made by the quantitative analysis program, allowing for the highest possible accuracy in the analysis. In addition to quantitative and qualitative x-ray microanalysis, x-ray mapping can be readily performed on polished slabs.
Thin sections are also readily examined in the SEM. Backscattered imaging is used almost exclusively as the specimen topography is limited and the topography seen is an artifact of the polishing process. However, secondary electron imaging must be used when examining microcracks because of the relatively high image resolution required. Quantitative and qualitative x-ray microanalysis and x-ray mapping are readily performed on thin sections of standard thickness. Quantitative x-ray microanalysis should not be performed on thin sections less than 10 microns in thickness. This is to satisfy another requirement of the quantitative analysis software. Namely, that the specimen is “infinitely thick” with respect to the incident electron beam. Most analysis software is unable to compensate for intensity variations resulting from the excitation of different volumes of material. The material volume reduction created by an extra thin section will cause a commensurate reduction in measured x-ray intensity.
X-ray diffraction (XRD) can be used to identify the type and abundance of crystalline phases within a concrete microstructure. In general, x-ray diffraction results in identification of the inter-atomic distances between all atoms in a crystalline structure. For any crystalline material, this arrangement of atoms in an ordered unit cell is unique and characteristic of that crystalline phase. X-ray diffraction can be performed using a powder diffractometer method or Debye-Scherrer camera method. Both methods are based on the physical process of diffraction.
The powder x-ray diffractometer is a major piece of equipment consisting of a high voltage power supply, x-ray tube and shielding, goniometer, signal processing electronics, and computer. The high voltage power supply generates approximately 3000 watts of power at 60 kV. This is used to power the x-ray tube and generate the required x-ray flux. The goniometer is used to move the specimen through a range of orientations while being exposed to the x-ray flux. Simultaneously, the goniometer is moving an x-ray detector through a range of angles to measure the diffracted x-ray intensity from the specimen. The processing electronics converts the detected x-ray pulses to a digital signal for computer processing. Finally, the computer system is used to process and analyze the data
For a given diffraction scan over a specified angular range, a diffraction pattern is produced that is the combination of unique patterns generated from each constituent of the material. The pattern plots the diffracted intensity measured at each angle over the scan range and consists of sharp diffraction peaks, or lines, superimposed on a slowly varying background. For a given crystalline material, a family of lines occurs because diffraction occurs in multiple angular directions. The angular occurrence and relative heights of the diffraction lines forming a family is unique to that crystalline material and serves as a “fingerprint” of the material. These peaks occur at predictable angles and intensities and therefore most XRD computer systems have a set of cursors that can be used to identify phases in multiphase materials such as concrete.
In general, the more abundant phases will have peaks that are very intense and easily identifiable, as the intensity of diffraction is directly proportional to the amount of material diffracting. For the same reason, minor phases will have peaks that are very small, lost in the background, or in some cases convoluted with diffraction lines from other phases. Therefore, identifying secondary deposits by powder diffraction XRD is not always feasible. Unless the material of interest is present in quantities greater than 1 percent by volume, it will be very difficult to detect.Phases occurring in quantities of 1 to 5 percent by volume may or may not be detectable, depending upon the other phases present. Also, note that XRD is used primarily to characterize crystalline materials. Therefore, amorphous gels can not be identified by XRD but they can be confirmed as amorphous.
One way to improve the detection of minor phases in a material is to extract the material from the matrix and analyze it separately. For powder x-ray diffraction, this is still a problem because a significant amount (e.g., 1 to 2 grams) is required to produce a usable pattern.However, the Debye-Scherrer camera method can utilize small mounts of material and produce a diffractograph that is also characteristic of the pattern. In a Debye-Scherrer camera, or powder camera, the specimen is placed stationary on a spindle in the center of a right cylinder. A piece of film is placed around the circumference of the cylinder and the impinging x-ray flux is collimated on to the specimen. Diffraction occurs and the individual diffraction peaks appear as lines on the film. This is actually the source of the term “lines” used for referring to diffraction or spectrographic peaks. After the film is exposed, developed, and laid flat on a light table, the peak position is determined by measuring from the entrance point of the incident beam.
This technique requires a high voltage power supply, x-ray tube, shielding, and a Debye-Scherrer camera. It is easily maintained and considerably lower in cost than a powder diffractometer. However, it is best suited for analyzing single phase powders. When multiple phases are present, the overlapping lines are very difficult to interpret. Also, the exposure, developing, and interpretation can be very time consuming. This method would be most applicable if a set of standard patterns were available for concrete phases.Methods for X-ray Diffraction
Specimens for XRD usually are in the form of a packed powder. If a phase can be physically extracted from the concrete, it can be analyzed separate from the matrix phases. Often this not possible. A sample of concrete can be ground to pass a 100-mesh sieve in any standard laboratory grinder. This is fine enough for packing into the respective specimen holders. However, peaks from minor phases will most probably not be seen when analyzed with concrete as a whole. Extracting minor phases is preferred when they are of interest.
The automated powder diffractometer has the advantage of fully automatic data collection, diffraction pattern analysis software, and peak identification cursors. It has the disadvantage of not being sensitive to phases having low concentration because the minor peaks are obliterated by the peaks from major phases. To perform an analysis, it is usually necessary to specify a scan range, step size, and dwell time. The scan range controls the angular range of Bragg angles examined from x-ray diffraction. A range of 5 ° to 75° for the theta rotation is adequate for most needs. The step size and dwell can be any value within the ranges allowed by the manufacturer. The step size is the change in angle between two successive steps of the drive motor on the XRD goniometer. A smaller step size provides better peak resolution but longer analysis times. Dwell time is the time spent measuring the diffracted x-ray intensity at each step of the goniometer. A longer dwell time will improve the sensitivity for phases present in low concentration but will increase analysis times. Be very skeptical of any software for automatically identifying peaks in an XRD pattern. This type of software usually misidentifies most phases in a material when more than two or three phases are present. Identify all phases by manual indexing of the pattern.
The powder camera is useful for analyzing small amounts of sample. However, the pattern resulting from multiple phase mixtures is often not discernible. If this method is to be used, try to establish a set of comparison patterns from materials of known composition.