13 Surface Area Measurements Jan Skalny and Nataliya Hearn

1.0

INTRODUCTION

Surface size and surface quality are two of the most important parameters of organic and inorganic solid materials. Surfaces of most solids, in tandem with other basic materials properties and environmental conditions, control to a large degree the rate of the relevant chemical and/ or physicochemical reactions, thus the behavior of the material in realworld applications. The degree of this control relates to the surface area of the reacting material(s), aided by their reactivity—a quality depending on the crystallographic and other basic physical and chemical properties of the reacting solid. Specific surface area is defined as the surface area per unit of mass, typically expressed in square centimeters per gram (cm2/g) or square meters per kilogram (m2/kg). Specific surface area depends on the particle size, particle shape, and any imperfections or flaws present at the surface. Although not linearly related, knowledge of the surface area of a system of particles (and assuming that they are monosized, nonporous spheres) allows estimation of the average particle size of the powder and vice versa. The particle size of a solid body depends on the degree of natural or manmade growth or size-diminution. A larger particle will have a larger mass and volume per unit of surface area than a smaller particle. When a larger 505

506

Analytical Techniques in Concrete Science and Technology

particle is subdivided into smaller particles by cutting, grinding, or other similar process, the total surface area will dramatically increase. For example, division of a 1 m3 cube into one micrometer cubed particles would increase the surface area from 6 m2 to 6 × 106 m2, a millionfold increase. The surface area of a solid also depends on the shape of the particle which, theoretically, can vary from a perfect sphere (minimum surface area-to-volume ratio) to a collection of atoms bonded along a one-dimensional chain (maximum surface area-to-volume ratio). The particle shape of realworld particles lies between these two extremes and so does the related surface area. Exact determination and classification of the three-dimensional shape of a particle is difficult; the best approximation is achieved by two-dimensional evaluation by means of computerized image analysis. In addition to its size and shape, the surface area of a particle depends on the amount, size and shape of the flaws (surface imperfections), varying from visible flaws to flaws at the atomic or crystal lattice level. Porosity of a particle can be then defined as the volume of those surface flaws which have a depth greater than the width.[1] Depending on the size and shape of the flaws or pores, the surface area of a solid can vary widely and the influence of flaws/pores can often overwhelm other factors controlling the surface area. Because of the above phenomena, the technical literature often distinguishes between internal and external surface area and between surfaces (often expressed in terms of porosity) accessible or inaccessible to certain chemicals or measurable by a particular methodology. Considering the above, it is clearly understandable that surface area is much more influential with respect to the properties of a material in highly divided or dispersed solids with higher concentrations of flaws and pores than in a coarse material with lesser concentration of flaws and pores. Thus, for example, the surface of a concrete aggregate is less important with respect to concrete properties than is the surface area, thus reactivity, of the cement. Another example: it is the grinding of the clinker granules to high surface area that enables production of a highly reactive portland cement. It is important to realize that the value of the measured surface area will depend on the experimental technique used to measure it. For example, as discussed in Sec. 2.1, the surface area measured by gas adsorption techniques such as BET method[2] will vary with the quality (crosssectional area and other properties of the molecule) of the gas used as the sorbate. In a similar manner, the surface area of a powder calculated from particle size distribution measurements will differ from the surface area of the same powder estimated on the basis of image analysis techniques. Other examples abound.

Surface Area Measurements

507

Please note that some of the test methods described in this chapter overlap with the measurements of porosity and pore size distribution discussed elsewhere. This overlap is caused by the inherent relationship between porosity, pore size distribution, and surface area. It is impossible to discuss all available techniques in detail. Each section, therefore, contains references to the original publications or a more detailed text for specifics in the derivation and application of each method.

1.1

The Importance of Surface Area in Cement/Concrete Science and Technology

Treval C. Powers and Stephen Brunauer were among the first who recognized the extraordinary importance of surfaces in explaining the complex phenomena leading to the development and degradation of physical properties of cement-based material. The rates of hydration reactions of individual components of portland and other cements are related to the surface area of the particular cement, thus controlling the rate of development and the quality of the desired physical properties. This is true in particular for the calcium silicate components of the portland cement clinker, tricalcium silicate (alite, Ca3SiO5) and dicalcium silicate (belite, beta-Ca2SiO4), both of which hydrate to form high-surface area calcium silicate hydrates (C-S-H, the “heart of concrete”),[3]] the most important component of the portland cement paste.[4] It is the semi-amorphous structure of the C-S-H and its complex porosity and high surface area that are mostly responsible for the time-dependent engineering properties of concrete, e.g., volume stability, strength, and modulus of elasticity. It should be recognized, however, that surface area alone is not an adequate measure of the quality or reactivity of a material. For example, it is known that two cement clinkers made from the same raw material, but in two different kiln systems, may produce clinkers that when ground to the same surface area will have different rates of reaction with water. This is related, among other factors, to the differences in the flaws concentration, thus surface quality.[5][6] Presence of heat-induced crystal lattice defects in the clinker minerals such as C3S and C3A influences the reactivity of the clinker and the resulting cement. In a similar fashion, clinker quality may be modified by introduction of flaws due to rapid rate of cooling or the intensity of grinding.

508

Analytical Techniques in Concrete Science and Technology

Conversely, it is the quality of the cement paste, primarily its porosity and permeability, that controls the rate of concrete deterioration. Whether the concrete deteriorates due to physical processes (e.g., freezing-thawing,) or chemical reasons (such as sulfate attack, acid attack, or alkali-silica reaction), the overall porosity and its distribution are crucial in controlling the time dependency of the damage. At a more technical level it is the increase in the liquid-accessible pore surface area as well as the rate of movement of moisture through the concrete rather than the porosity itself that cause a highly porous and permeable concrete to deteriorate at a rate higher than a less porous, less permeable concrete. The importance of surface area as a measure of cement quality is widely recognized by the research and business communities and by ASTM, CSA, CEN, and other standardization organizations; for example, ASTM requires all cement types to pass certain minimum surface area (fineness) requirements, thus helping the cement producer and user to control the product quality.[7] On the other hand, the practical importance of the surface area of the hydration products formed as a consequence of the water-cement reactions is much less appreciated by the typical field engineer and some of the underlying details and principles are still not completely understood. On the following pages a brief review of the most common techniques of surface area determination is presented. The review is not meant to be a comprehensive state-of-the-art compilation of the existing knowledge, just a brief introduction to the most important approaches used. The methodologies discussed include air permeability methods, methods based on particle size distribution, gas adsorption techniques, and scattering techniques.

2.0

TECHNIQUES OF SURFACE AREA MEASUREMENT

In the area of cement and concrete research, two main types of materials need to be characterized: powder materials such as cement, fly ash and silica fume (microsilica), and solid materials such as hardened cement paste and concrete. The surface area of the powder material is of importance due to its effect on the reactivity. That is why the rapid-hardening portland cement (ASTM Type III) has a specific surface area of 450 to 600 m2/kg as compared to 250 to 400 m2/kg for ordinary portland cement (ASTM Type I). Measurement of the surface area of powders is dependent on the particle

Surface Area Measurements

509

size, with coarser powders measured using sedimentation or elutriation (such as Wagner turbidimeter—ASTM C115-93). For cements, air permeability—Blaine method (ASTM C204-94), which is a modified Lea and Nurse method, is used to determine surface area. For powders finer than portland cement (silica fume or fly ash) nitrogen adsorption or mercury intrusion porosimetry are used to determine the surface area. The hydration process increases the surface area as the fragile calcium silicate hydrate gel structure is formed. Figure 1 shows the increase in the surface area with the progress of hydration. The surface area of hydrated cement paste is a thousand times greater than that of the unhydrated cement with the reported values dependent on the measurement technique used and sample conditioning. For instance, the water adsorption results yield surface area of 200,000 m2/kg,[8] while x-ray scattering gives 600,000 m2/kg.[9] Because concrete is by nature an inhomogeneous material, sample preparation often damages the finest portion of the pore structure and many test methods use only very small samples, it is very difficult to obtain statistically significant results. Also, to further complicate the situation, the pore sizes and surfaces to be measured span several orders of magnitude.

Figure 1. Total surface area as a function of hydration time.[10]

To overcome these difficulties, several techniques based on different principles were developed over the years. Some of these techniques measure the resistance of a material to flow of an inert gas/liquid or the adsorption of

510

Analytical Techniques in Concrete Science and Technology

gases and vapors, some measure the porosity by monitoring the pore size filled with a liquid at an applied pressure, and others use the deflection of electron and neutron beams.

2.1

Gas Sorption Techniques

Gas sorption techniques are based on physical adsorption of gases or vapors on surfaces of solids. In contrast to chemical adsorption, also called chemisorption or irreversible adsorption, during physical adsorption the sorbed molecules are not restrained to specific sites on the surface of the measured solid and are free to cover the whole surface. For this reason, determination of surface areas is possible. Physical adsorption is fully reversible and, with the exception of small pores, equilibrium can be easily achieved since no activation energy is involved. Because physical adsorption does not occur at elevated temperatures, sufficiently clean surfaces can be prepared prior to the actual low-temperature surface area measurements. Although numerous methods for surface area measurements based on physical adsorption of gases were developed, the best known method for determination of surface areas of porous solids is the so-called BET method grounded on the work of Brunauer, Emmett, and Teller.[2][3] Since its development in the late 1930s, this method became the most universally used method for surface area determination of such diverse materials as catalysts, carbon blacks, finely divided silica, and hydrated cement pastes and components. The BET methodology gives surface areas two to three times higher than the Lea and Nurse and Blaine methods discussed above. The term sorption is used to describe the interaction of a gas with a solid surface; this interaction may be in the forms of adsorption, absorption, or capillary condensation. When a gas is removed from a surface the process is called desorption. The gas interacting with the surface is usually referred to in the literature as a sorbate. The BET method as well as some previous related work is based on the experimental establishment of the relationship between the pressure of a gas that is in equilibrium with a solid surface and the volume of the gas adsorbed at the particular pressure at the surface. The theoretical basis for this approach is the Langmuir theory based on the kinetic theory of gases.[11][12] Gas sorption techniques are based on the assumption that gas molecules are strongly attracted and adsorbed on a virgin surface. The method assumes also that the sorbent molecules have access to the walls of the pores

Surface Area Measurements

511

within the studied solid and that the distance between the walls of a pore is large when compared to the molecular dimensions of the sorbate. As a monolayer of the sorbent gas forms on the surface the repulsion of the previously adsorbed gas molecules makes the formation of a second and subsequent layers less likely. This process is gas pressure dependent, thus with increased relative pressure the degree of gas adsorption increases. Gas molecules may be adsorbed in subsequent layers before the underlying layer is completely covered. The BET Method. To enable generalization of the Langmuir’s approach to multilayer adsorption, the BET theory makes two basic assumptions regarding the heat of adsorption: 1. It is constant throughout the formation of the first layer of adsorbed gas. 2. In the second and higher layers it is equal to the heat of liquefaction. More about the accuracy of these BET assumptions can be found in Refs. 2, 12, and 13. A plot of the amount of gas adsorbed at a certain temperature against the relative pressure is called a sorption isotherm. It is usually presented as the volume of adsorbed gas versus the relative pressure, p/po , see Fig. 2.[12] From such a plot the amount of gas needed to form a monolayer can be determined and, assuming the cross-sectional area of the sorbate molecule, the surface area of the measured solid can be calculated. The relative humidity at which a monolayer completely covers the solid surface depends both on the nature of the used sorptive gas and the nature of the solid. Table 1 gives the approximate cross-sectional areas of some commonly used sorbates. Because the capillary condensation in a set of pores of certain size at adsorption does not occur exactly at the same relative pressure as capillary evaporation from the same pores at desorption, the combined adsorptiondesorption isotherms may, and usually do, show a hysteresis. A schematic example is shown in Fig. 3.[15] Several types of adsorption-desorption isotherms were identified. [16] The BET equation can be given as: v/vm = ckx/(1 - kx) [ 1 + (c - 1) kx] or in its linear form as: kx/v(1 - kx) = (1/vm c) + [(c - 1)/vm c]kx

512 where:

Analytical Techniques in Concrete Science and Technology x is the relative pressure, p/p o k is a number smaller than 1 (see Ref. 11) c is the BET (“adsorption coefficient”) constant vm is the volume of the sorbate at monolayer coverage v is the volume of the adsorbed sorbate

Figure 2. Sorption isotherm for a planar surface (adopted from Ref. 12, Fig. 2).

Table 1. Most Commonly Used Sorbates Sorbate Argon Benzene Nitrogen Oxygen Water vapor Xenon

Area in Å 14.2 40.0 16.2 14.1 10.8 2.5

Surface Area Measurements

513

Figure 3. Adsorption isotherm showing capillary condensation (Fig. 1, Ref. 15).

The actual analysis is usually done by plotting x/[v (1 - x)] against x for the linear (multilayer adsorption) region of the isotherm. This plot, referred to as the BET plot, is fitted with a least-square algorithm. The equations for the slope (S BET) and intercept (I BET ) of the BET plot are given as: SBET = (c - 1)/vmc I BET = 1/vmc The BET method is applicable for solids with pore diameters from above ~4 nm. Below 8 nm the calculations are believed to be inaccurate. It is accepted that BET gives the most accurate results for the mesopores (160 nm to about 10,000 nm). The range of relative humidities (p/p o) at which the BET method applies is from about 0.05 to 0.35. For additional information see, for example, Refs. 12, 17, and 18.

514

Analytical Techniques in Concrete Science and Technology

The BET method has been used in cement science for several decades now. Whereas the technique enabled better understanding of the nature of the hydration products and of the mechanisms of hydration, it also led to controversies regarding the best sorbate to be used in determination of the “true” surface area of the hydrates. The sorbates used most often are water vapor and nitrogen, neither of which can measure accurately all the surface area located in the complex pore system. Water sorption may give erroneous results due to the possible adsorption of the polar water molecules within the vacated interlayer spaces of C-S-H. Nitrogen (N2) and other nonpolar sorbents, on the other hand, may give much lower surface areas; this phenomenon is related, among other reasons, to the failure of these nonpolar sorbents to reach the surface of the pores of certain size or shape (e.g., so-called ink-bottle pores). For more information, see Refs. 12, 15, and 19. The differences can be quite substantial. BET surface areas computed from N2 adsorption isotherms are between 13,000 and 84,000 m2/kg, while calculations made from the water adsorption isotherms are about 200,000 m2/kg. Oxygen, argon, and organic vapors (cyclohexane and isopropanol), give results similar to the N2, and methanol, 50,000 to 114,00 m2/kg.[20] The preparation of specimens for the adsorption methods requires drying. Techniques which do not require predrying of specimens, such as small angle x-ray scattering, small angle neutron scattering, and nuclear magnetic resonance, show that 200,000 m2/kg obtained by water BET is up to three times lower than the surface area of saturated specimens (Table 3). The primary advantage of the BET method is its capability of measuring disconnected microcracks and cracks that are open only at one end. The main disadvantage is its complexity—it requires skilled labor and is time consuming; however, automated BET equipment is now available for routine use. The method does not make any assumptions regarding the pore size or pore size distribution. At relative pressures higher than about p/po above 0.35–0.40, capillary condensation can take place and pore size distribution can be obtained from Kelvin equation linking the vapor pressure of the sorbate with curvature of the meniscus formed by the liquid inside the pores. Several approaches to complete pore structure analysis were developed.[15][21] In most such calculations assumptions must be made about the shape of the pores.

Surface Area Measurements

2.2

515

Mercury Intrusion Porosimetry (MIP)

One of the best known techniques used for determination of particle size or pore size distribution that can also be used to indirectly determine the surface area is the Mercury Intrusion Porosimetry. MIP is based on the relationship between pressure and the corresponding volume of pores filled with a non-wetting liquid. Due to the high pressure needed to infiltrate the pores, this method is applied mainly to hydrated cement paste and concrete both of which can withstand the applied pressure without much damage to the measured microstructure. The liquid typically used is mercury. It is a non-wetting liquid having a contact angle with concrete of 130–140°. The applied pressure and the resulting volume of mercury entering the tested material can be related to the pore diameter and the amount of pores of the particular size: 2·γ ·cos θ where:

γ θ P r

=P·r

is the surface tension of mercury is the contact angle is pressure is the radius of pore

The method uses two major assumptions: (i) the surface tension and the wetting angle are constant throughout the tested specimen, and (ii) all pores have the shape of an ice cream cone. The first assumption may not be accurate because the contact angle changes with the changing pore solution in the tested material. The second assumption causes the space behind the neck of the ink-bottle to be treated as a cylinder with radius of the pore opening. The pressures typically applied during testing ranges from 0.5 psi (~3,800 Pa) to about 60,000 psi (~45 MPa). These pressures allow the measurement of pores from about 2 nm to about 20 µm.[22]–[24] The upper limit on the pressure used for testing is set to limit the damage caused to the pore structure damage. MIP is one of the more useful methods because it enables measurement of porosity over several orders of magnitudes; however, the range of pores present in hydrated portland cement or concrete is wider than what can be measured by MIP. Smaller pores are not detected due to the restraint on the applied pressure, whereas larger pores are misrepresented because it is

516

Analytical Techniques in Concrete Science and Technology

difficult to define boundaries between the largely interconnected pores.[25] Furthermore, MIP does not measure the true pore distribution, but rather, indicates the accessibility to mercury of the overall porosity as a function of pore size. This is a form of invasion percolation. The use of MIP requires complete drying of the pore structure. This process damages the microstructure and subsequent measurement of the surface area may yield erroneous results. As with many other techniques used to measure surface area the tested sample is typically very small, on the order of a few grams. One has to, therefore, bear in mind the possible analytical errors, the heterogeneity of concrete and of the other measured solids, and the statistical significance of the obtained results. Rootare and Prenzlow derived another method for surface area measurement using mercury.[26] This method is based on the amount of work (given by the area under the pressure vs. volume curve) needed to cover the surface of the examined solid with mercury. This method requires no assumptions regarding the pore shape distribution; however, it has been thirty years since its development and it has been hardly used.

2.3

Wagner Turbidimeter

The primary use of this technique is in measurement of surfaces of powders. The Wagner turbidimetry is based on measurement of the terminal free fall velocity governed by Stoke’s law. By using a beam of light, the concentration of particles suspended in kerosene is measured by determining the percentage of light transmitted through the suspension to a photocell.[27] The method was adapted by ASTM as Standard Test Method C 115. The technique usually gives consistent results, the main error being the assumption of uniform distribution of particles smaller than 7.5 µm. The use of the average particle size of 3.8 µm below the 7.5 µm is an overestimation resulting in the lower calculated surface area. The assumption that the average particle below 7.5 µm is 3.8 µm is overestimated; thus the calculated surface areas are too low. Because it is these finest pores that govern the specific surface area, their influence on the calculated results is overwhelming; a modification to reduce this error was proposed by Hime and LaBonde.[29] Both MIP and Wagner Turbidimeter results have to be recalculated into specific surface area using the equation:

Surface Area Measurements

S = 6 ⋅ 10 6 ⋅ F ⋅ where:

517

∑f

d⋅ρ

S is the specific surface area in m2 kg-1 F is an empirical constant that takes into account the assumed pore shape distribution specific surface area f the weight fraction of material consisting of grains assumed to have a diameter d, in µm r is the density, given in kg m-3

These techniques of surface area measurement do not always yield satisfactory results because, among other reasons, the specific surface area is highly dependent on the assumed particle or pore shape distribution. More about these techniques is given in Ch. 14.

2.4

Permeability Methods

Permeability measurements are applicable for situations when the flow through a material is caused mainly by pressure gradient. If this assumption is true, permeability is the property of a porous material which characterizes the ease with which a fluid may be made to flow through the material. The coefficient of permeability can thus be related to the surface area of the tested material. The relationship that governs the flow through porous media was developed first by H. Darcy in the late 1800s.[30] Originally, it was meant to define the flow of water through sand. The principal equation is given as: Q = A⋅ where:

k ⋅ ∆P ∆L

Q

is the flow rate

A

is the cross-sectional area

∆P

/∆L

is the pressure gradient.

Lea and Nurse applied Darcy’s law to describe the surface area of materials.[11] They developed an apparatus which measures steady state flow of air through a bed of material with known porosity. The surface is then calculated as:

518

Analytical Techniques in Concrete Science and Technology

Sw = where:

ε 3 ⋅ A ⋅ h11 14 ⋅ K ⋅ L ⋅ h2 ρ ⋅ (1 − ε )

ρ

is the density of cement

ε

is the porosity of cement bed (0.475 in the British Standard test)

A

is the cross-sectional area of the bed (5.066 cm2)

L

is the height of the bed (1 cm)

h1 is the pressure drop across bed h2 is the pressure drop across the flowmeter capillary (between 25 and 55 kerosene) K is the flow meter constant Blaine modified the Lea and Nurse method; instead of measuring pressure drop during the steady state flow, the Blaine approach measures the pressure change as a specified volume of air passes through the tested material.[31] The Lea and Nurse and Blaine methods yield similar results, but because both of them measure only the interconnected pores the surface areas are low compared to some other techniques. By leaving out the disconnected pores, especially the very small pores which are highly desirable in good quality concrete, they underestimate the magnitude of the surface area. The Blaine method is used as the standard method to measure the fineness of cement in North America.[32] It has two main advantages: the method is simple and fast. The main disadvantage, not critical in determination of cement fineness, is its poor accuracy which becomes worse with increasing variability of the particle size, pore tortuously, and surface area. Due to these factors the Blaine method becomes extremely unreliable at surface areas exceeding 500 m2/kg. The use of permeability methods requires complete drying of the pore structure. This process damages the microstructure and subsequent measurement of the surface area yield erroneous results. This is one of the reasons that it is primarily used to measure fineness of cements and other dry powders, but not surfaces of hydrated materials or materials that have to be dried. Table 2 shows the surface area of two types of cements measured using the above described methods.

Surface Area Measurements

519

Table 2. Surface Area of Cements (m2/kg) [Adapted from Ref. 33]

2.5

Cement

Wagner Method

Lea & Nurse Method

I II

179 227

260 415

Nitrogen Adsorption 790 1000

Small Angle X-Ray Scattering (SAXS) and Small Angle Neutron Scattering (SANS)

The basic principle of SAXS and SANS is the scattering of the beam of radiation as it passes through a material. At low angles the scattering is particularly noticeable and can be monitored to describe the inhomogeneities or boundaries in a two phase system (e.g. solid and air, or solid and water). Figure 4 shows schematics of an x-ray scattering arrangement. The detailed explanation of the theory and treatment of SAXS and SANS data can be found in Refs. 34 and 35. In brief, the interpretation of the x-ray or neutron scattering for calculating specific surface area is based on the Porod’s theory. [36] The theory is based on the scattering purely by the interfaces where a sharp boundary between phases exists. In cementitious systems other sources of scatter exist, such as pore fluid in the gel spaces, so that the data has to be corrected by subtracting the non-boundary scatter. Mathematical models for interpretation of the experimental data are needed[37] and work has been done relating SAXS data and fractal geometry.[38] The main advantage of the scattering techniques for measurement of the specific surface area is that they do not require any drying or pretreatment of the samples, thus measuring unaltered microstructure of the cementitious systems.[39] The x-ray and neutrons can readily penetrate small size pores which are inaccessible to nitrogen and mercury with SAXS lower limit of detection of 30 nm and SANS of 0.1 nm. Extensive work by Winslow and Diamond using SAXS to analyze the effect of pretreatment conditioning on the pore structure characterization of hydrated cement pastes has shown the effectiveness of the technique to assess the evolution and destruction of the surface area. Their results are shown in Table 3.

520

Analytical Techniques in Concrete Science and Technology

Table 3. Surface Area of Cement Pastes of Different w/c Ratios and in Saturated and Dry Conditions[9] w/c

Age (days)

Hydration (%)

Condition

Surface Area (m2/g)

0.3 0.3 0.4 0.4 0.6 0.6

513 513 514 514 512 512

78 78 86 86 91 91

Saturated D-dried Saturated D-dried Saturated D-dried

527 159 708 224 782 284

Figure 4. Schematic of scattering arrangement. [39]

2.6

Nuclear Magnetic Resonance (NMR)

One of the new techniques for determination of the total pore volume and surface areas is NMR, specifically the technique referred to as spin-spin relaxation.[10][40] The method is based on the relaxation of longitudinal or transverse magnetization at the interface between a solid and the liquid, in other words, on the dependence of the exact resonance frequency upon the local magnetic field. Because the measured relaxation is surface dominated, the NMR experiments enable determination of the total surface area and of the distribution of the surface-to-volume ratio, in other words, of the pore volume distribution function. In theory, interconnected pore structure can be completely characterized by surface-to-volume ratio. The method may be used for saturated samples, does not require drying of the sample, thus damage to the pore structure is not an issue and the volume of the fine pores can be accurately measured.

Surface Area Measurements

521

The NMR theory is quite complex. Its use for surface area determination is based on two principles: • In the region containing evaporable water, the first molecular layer on the hydrated surface has a uniform, well-defined relaxation time • All conditions for fast exchange of water present in the first surface layer and of the remaining evaporable water in the pore structure are met The absorption spectrum of a material will depend on the environment of the protons in the sample, thus the technique allows estimation of the water mobility or its state in the sample. For more information, see Ref. 41.

2.7

Image Analysis

Image analyses of visible light (optical) or backscattered electron images of specially pretreated samples can be, in principle, used for determination of pore size distributions and surface areas. However, as discussed elsewhere in this book (Ch. 19), in addition to some advantages (e.g., no need for drying), the technique suffers from several disadvantages: • Two-dimensional images have to be recalculated into three-dimensional pore distribution data • The visible pore images do not represent reality because the measured cross-sections are smaller than the cross-section at the maximum pore diameter • Inadequate resolution of the instrumentation at high magnifications • Limitation on the number of displayable pixels Because of these and some other reasons, at this stage of development, estimation of surface areas from pore size distributions based on image analysis is not practical and is not used routinely.[42][43]

3.0

APPLICATIONS

Knowledge of the surface area is of utmost importance in many concrete-related applications. Whereas some of the techniques discussed

522

Analytical Techniques in Concrete Science and Technology

above are routinely used in the production and quality control/assurance others are of interest only at the scientific level and in the development of new products. The following paragraphs will give a few, very superficial examples of such applications; the list is certainly not complete and serves only as a reminder of the importance of surfaces in better understanding and control of materials properties. Cement Raw Meal. The quality of a cement clinker and the productivity of its manufacturing depends among others, on the rate at which the components of the raw meal are transformed during its burning into the clinker of desired quality. Under otherwise constant conditions the raw meal-to-clinker transformation depends on the surface area of the kiln feed. Knowing the differences in the reactivity of the raw meal components—for example the relative reactivates of the carbonaceous versus siliceous versus iron-containing components—the fineness of the kiln feed can be adjusted to give a surface area that will guarantee a burning rate acceptable from both economic and technical points of view. For the above reasons the surface area of the kiln feed is closely monitored, primarily by control of its particle size, a measure of the available reactive surface. Concrete Mixture Components. The concrete mixture components for which surface area plays a crucial role are the cement and mineral admixtures, such as fly ash, condensed silica fume (microsilica), and granulated blast furnace slag. The quality and quantity of the fine and coarse aggregate surface are important, however, to a lesser degree than the above components. The reason for this is the fact that whereas the cementitious materials participate in chemical reactions of hydration (reactions between water and the components of cement, blending materials, and chemical admixtures), the aggregate surface can be under most conditions considered to be inert. Based on this knowledge, the concrete manufactures expect to receive from the materials suppliers cement and mineral admixtures that conform to the codes and the standards requirements, and enable them to produce high quality concrete (primary requirement) within the essential time and budget (secondary consideration). The most common techniques of mixture components surface area evaluation are particle size measurements and air permeability techniques. Hydrated Portland Cement and Concrete. As implied above, surface area plays a most important role in chemical reactions that lead to development of concrete properties. Whereas the surface area of the anhydrous cement compounds does affect the rate at which a given cement

Surface Area Measurements

523

contributes to the development of the concrete engineering properties, it is the surface area and other qualities of the newly formed hydration products that are responsible for the microstructural behavior and physical performance of concrete. The cement paste component of most influence on concrete properties is calcium silicate hydrate, C-S-H. It constitutes about 50–60% of the volume of the solids in a fully hydrated cement paste and its surface area, depending on the conditions of its formation (temperature, humidity, water-to-cement ratio, etc.), and the type of measuring technique, is reported to vary between 10 and more than 250 square meters per gram. Most literature data agree today that the surface area of a fully hydrated, D-dried portland cement paste is about 200 m2 g-1.[31] However, values as high as 700 to 800 m2 g-1 per ignited weight were obtained using small angle x-ray scattering.[45] The primary properties of concrete that are influenced by the surface area of the C-S-H are the dimensional changes, creep, strengths, and modulus of elasticity. This is related to the structure and composition of the C-S-H, in particular to the movement of water between and within the C-S-H particles.[4][31] Because of this importance, the issue of the surface area of C-S-H and cement paste is an important topic of research for several decades, and many of the obtained results remain controversial.[25] Different experimental techniques give different results and some of the data interpretations are inconsistent and confusing. This is not only related to the differences in the measurement principles of the individual techniques, but also to the fact that the chemistry and fine internal structure of the C-S-H and thus, its “true” surface area, are unknown. Also unknown are the basic differences between the surface properties of the individual forms of C-S-H that can be observed by electron microscopic techniques (e.g., the so-called “inner” versus “outer” C-S-H product). As mentioned earlier, the most common classical research technique for determination of surface areas of cement paste and hydration products is the BET technique. [2][3] Primarily water vapor and nitrogen, but also other gases are used as sorbents. In recent years, these gas adsorption techniques were supplemented by others, such as SANS, SAXS, and NMR, discussed in more detail above. Novel data are being obtained, but clear and experimentally proven interpretations are lacking. Concrete Durability. In a sense, deterioration of concrete is the reverse process when compared to the hydration reactions. In the former case, discussed above, the chemical reactions lead to formation of a

524

Analytical Techniques in Concrete Science and Technology

microstructure that controls or guarantees certain mechanical properties of the concrete. In the latter case the chemical reactions lead to destruction of the previously formed microstructure by transformation or decomposition of the matrix components from those possessing cementitious properties, primarily of C-S-H, to new compounds that do not possess them. Because, as discussed earlier, the accessibility to liquids of the surface area in cement-based materials is closely dependent on the porosity and its structure (pore structure, spatial distribution, connectivity), the rate of concrete deterioration by most if not all of the possible mechanisms of chemical deterioration is porosity dependent (e.g., alkali-silica reaction, alkali-carbonate reaction, sulfate attack, acid attack). Thus, stated in a somewhat oversimplified form, the durability is controlled by the water-tocement ratio. The absolute size of the surface area exposed to chemical reactions of deterioration and its time dependency are difficult to measure or are not needed, thus the most common modes of evaluation of the deterioration rate are the changes in porosity, pore structure, water or certain ions permeability, and vapor transmission. These methods constitute an indirect measure of the effects of the reacting surfaces.

REFERENCES 1.

Lowell, S., Introduction to Powder Surface Area, John Wiley & Sons, New York (1979)

2.

Bruanuer, S., Emmett, P. H, and Teller, E., Adsorption of Gases in MultiMolecular Layers, J. Am. Chem. Soc., 60:309 (1938) Brunauer, S., The Adsorption of Gases and Vapors, Oxford University Press, London (1943) Skalny, J., and Young, J. F., Mechanisms of Portland Cement Hydration, Vol. I, Proc.7th. Int. Symp. Chem. Cem., Paris, II-1/3 (1986) Sakurai, T., Sato, T., and Yoshinago, A., The Effect of Minor Components on the Early Hydraulic Activity of Portland Cement Clinker, Proc. 5th. Int. Symp. Chem. Cem., p. 300, Tokyo (1968) Maycock, J. N., Skalny, J., and Kalyoncu, R., Crystal Defects and Hydration, I. Influence of Lattice Defects, Cem. Concr. Res., 4:835 (1974), see also ibid 3:701 (1973)

3. 4. 5.

6.

7.

ASTM, see for example the Standard Test Methods C 115-94, C 184-94, C 204-94a, C 430-94, and C 786-94.

Surface Area Measurements 8. 9.

10.

11. 12.

13. 14. 15.

16.

17.

18.

19. 20.

21. 22. 23.

525

Powers, T. C., Structure and Physical Properties of Hardened Portland Cement Paste, J. Amer. Ceramic Soc., 41:1–6 (Jan., 1958) Winslow, D., and Diamond, S., Specific Surface of Hardened Portland Cement Paste as Determined by SAXS, J. Amer. Ceramic Soc., 57:193–197 (May, 1974) Halperin, W. P, .Jehng, Y-Y., Song, Y-Q, Application of Spin–Spin Relaxation to Measure of Surface Area and Pore Size Distribution in a Hydrating Cement Paste, Magnetic Resonance Imaging, 12(2):169–173 (1994) Langmuir, I., The Adsorption of Gases on Plain Surfaces of Glass, Mica and Platinum, J. Am. Chem. Soc., 40:1361–1403 (1918) Rarick, R. L., Bhatty, J. I., and Jennings, H. M., Surface Area Measurement Using Gas Sorption: Application to Cement Paste, Materials Science of Conc, (J. Skalny, and S. Mindess, eds.), 4:1–40, The American Ceramic Society, Westerville, OH (1995) Brunauer, S., Skalny, J., and Bodor, E. E., Adsorption on Nonporous Solids, J. Coll. Interface Sci., 30:546–552 (1969) Gregg, S. J., and Sing, K. S. W., Adsorption, Surface Area, and Porosity (2nd Ed.), Academic Press, New York (1982) Brunauer, S., Skalny, J., and Odler, I., Complete Pore Structure Analysis, Pore Structure and Properties of Materials, Proc. RILEM-IUPAC Int. Symp., Prague (1973) Brunauer, S., Demming, L. S., Demming, W. S., and Teller, E., On a Theory of the Van der Waals Adsorption of Gases, J. Am. Chem. Soc., 62:309–319 (1938) DeBoer, J. H., The BET Method International Symposium on the Surface Area Determination,(D. H. Everett and R. H. Ottewill, eds.), Butterworths, London (1969) Milburn, D. R., and Davis, B. H., Comparison of Surface Areas Calculated from Nitrogen Adsorption and Mercury Porosimetry, Ceram. Eng. Proc., 14:130–134 (1993) Taylor, H. F. W., Cement Chemistry, Academic Press, London (1990) Mikhail, R. S., and Selim, S. A., Adsorption of Organic Vapors in Relation to the Pore Structure of Hardened Portland Cement Paste, Highway Res. Board, Spec. Rep. No. 90, p. 123 (1966) Wheeler, A., Advan.Catalysis, 3:250 (1951) Winslow, D., Some Experimental Possibilities with Mercury Intrusion Porosimetry, Mat. Res. Soc. Symp. Proc., 137:93–103 (1989) Milburn, D. R., and Davis, G. H., Comparison of Surface Areas Calculated from Nitrogen Adsorption and Mercury Porosimetry, Ceram. Eng. Proc., 14:130–134 (1993)

526 24.

25.

26. 27.

28. 29.

30. 31. 32. 33.

34.

35.

36. 37.

38.

39.

Analytical Techniques in Concrete Science and Technology Diamond, S., Rapid Particle Size Analysis of Fly Ash with Commercial Laser Diffraction Instrumentation, Mat. Res. Soc. Symp. Proc., 137:83 (1989) Lange, D. A., Jennings, H. M., and Shah, S. P., The Influence of Pore Structure on the Properties of Cement Paste, Mat. Res. Symp., Proc., 147:47–54 (1989) Rootare, H. M., and Prenzlow, C. F., J. Phys. Chem., 71:2733 (1967) Wagner, L. A., A Rapid Method for Determination of Specific Surface Area of Portland Cement, Proceedings, ASTM, ASTEA, Vol. 33, Part II, p. 553, (1933) Standard Test Method for Fineness of Portland Cement by the Turbidimeter, ASTM C 115-94, (1994) Hime, W. G., LaBonde, E. G., Particle size Distribution of Portland Cement From Wagner Turbidimeter Data, J. Protl. Cem., Assoc. Research and Development Laboratories, 7:66–75 (1965) Darcy, H., Determination of the Laws of the Flow of Water Through Sand, Victor Dalmont Publisher, Paris (1856) Blaine, R., Measurement of Specific Surface by Air Permeability, ASTM Bull. 108, pp. 17–20 (1943) Standard Method for Fineness Of Hydraulic Cement by Air Permeability Apparatus, ASTM C 204-94a, (1994) Sereda, P. J., and Ramachandran, V. S., Predictability Gap Between Science and Technology of Cements: 1. Characterization of Hydraulic Cements, J. Am. Ceramic Soc., 58, 3–4, pp. 94–99 (1975) Neutron, X-Ray and Scanning: Introduction to an Investigation Tool for Colloidal and Polymeric Systems, (P. Linder, and Th. Zemb, eds.), North Holland, Amsterdam (1991) Kratky, O., and Laggner, P., X-Ray Small Angle Scattering, Encyclopedia of Physical Science and Technology, Academic Press, London, 17:727–781 (1992) Porod, G., Die Rontgenkleinwinkelstreuung von Dichtgepackten Kolloiden Systemen, Kolloid-Z., I. Teil, 124, pp. 83–113 (1951) Haussler, F., Eichor,. F., Rohling, S., and Baumbach, H., Monitoring of the Hydration Process of Cement Pastes by Small Angle Neutron Scattering, Cement and Concrete Research, 20:644–654 (1990) Kriechbaum, M., Degovics, G., Laggner, P., and Tritthart, J., Investigation on Cement Pastes by Small-Angle X-Ray Scattering and BET: The Relevance of Fractal Geometry, Advances in Cement Research, 6(23): 93–100 (1994) Winslow, D. N., Bukowski, J. M., and Young, J. F., The Early Evolution of the Surface of Hydrating Cement, Cement and Concrete Research, 24:1025–1032 (1994)

Surface Area Measurements 40.

41. 42.

43.

44.

527

Bhattachrja, S., Moukwa, M., D’Orazio, F., and Jehng, J. Y., Microstructure Determination of Cement Pastes by MNR and Conventional Techniques, Advn. Cem. Bas. Mat., 1:67–76 (1993) Feldman, R. F., The Porosity and Pore Structure of Hydrated Portland Cement Paste, Mat. Res. Soc. Symp. Proc., 137:59–73 (1989) Eberly, D., Lancaster, J., and Alyassin, A., On Gray Scale Image Measurements, Graphical Models and Image Processing, 53:550–562 (1991) Barnett, M. I., Optical Microscopy for Particle Size Analysis, International Symposium on the Surface Area Determination, (D. H. Everett and R. H. Ottewill, eds.), Butterworths, London (1969) Winslow, D. N., and Diamond S., Specific Surface of Hydrated Portland Cement Paste as Measured by X-Ray Scattering, Coll. Interface Sci., 45:425 (1973)