Surfactant-induced growth of a calcium hydroxide coating at the concrete surface.
Keywords Concrete, Calcium hydroxide, Surface, Image analysis, Surfactants, Demolding agent
High-performance concrete (HPC) was developed in recent decades to meet an increasing need for high mechanical resistance and demanding structures. Its high compressive strength is mainly achieved through two properties: a low water to cement ratio, made possible by the addition of plasticizers and superplast-icizers, and an optimized granular repartition due to a wide particles size distribution. These major differences from standard concrete also impact the surface properties: HPC exhibits a smooth surface with low open porosity and minor color variations, which makes it an ideal candidate for esthetic applications--outdoor buildings, but also furniture, floors, artwork, etc.
In spite of its low porosity, HPC is very sensitive to external attacks from atmospheric pollutants, microorganisms, or accidental staining (coffee, oil, etc.) which compromise both its esthetic properties and its durability. Methods exist to either prevent or repair the effects of these attacks. On the preventive side, the most effective existing method is to apply a protective coating on the concrete surface, which induces both operational and environmental costs that are no longer acceptable today. (1) As for the curative methods (laser cleaning, chemical cleaning, painting), they have to be frequently repeated and are not always completely efficient or environmentally friendly.
There are several possible ways to limit the impact of external pollution on a concrete surface. One of them is to close the surface porosity to prevent the deposition and incrustation of staining agents or microorganisms, like coatings or sealers do. (2) Another approach is to facilitate the self-cleaning of the surface. This can be achieved with super-hydrophilic surfaces, which can be easily washed by dripping water, or with self-cleaning agents that destroy the polluting agents and limit the biological growth. Photocatalytic titanium dioxide has proved very efficient to confer hydrophilic properties to concrete surfaces and to neutralize a wide range of pollutants, but its action mechanisms are not fully understood yet and this cleaning technology still needs improvements before being widely used on the concrete market--more information on this can be found in the review by Maury and De Belie. (3)
This paper suggests an innovative method to protect HPC from environmental factors, which can be used as a complementary method to Ti[O.sub.2] self-cleaning. This new method generates no costs and allows a gain of time as it necessitates no post-demolding treatment. It consists of provoking the formation of a Ca[(OH).sub.2] layer at the concrete surface. This layer, which is super-hydrophilic, closes the surface porosity and facilitates water cleaning.
Although this article focuses on concrete, it may be of interest to all of the coating community as an example of an innovative approach to coating science. (4)
Ca[(OH).sub.2], also called porllandite, is one of the main Portland cement hydrates (up to 25% in concrete volume). Spontaneous formation of large Ca[(OH).sub.2] crystals can occur at the HPC surface in very specific cases. For example, Martin (5) showed that the use of a smooth formwork made of glass or PMMA provokes an accumulation of saturated water at the concrete/formwork interface and the precipitation of Ca[(OH).sub.2] once the water is reabsorbed during hydration. This study shows that this phenomenon can be triggered and amplified by the introduction of nonionic surfactants at the concrete/formwork interface. Under adequate conditions, it is possible to grow a continuous layer of Ca[(OH).sub.2] which significantly reduces the water uptake of the surface and decreases the static water contact angle.
Materials and methods
Preparation of mortar samples
Since only the surface properties matter, the samples were made of mortar rather than concrete, i.e., no coarse aggregates or fibers were used. The preparation of the mortar samples was based on a white HPC mix-design. It was made of 31% white Portland cement (CEM I 52.5 PMES from Lafarge, Le Teil--France), 9% of limestone filler (from Omya), 7% of silica fumes (from MST), 43.5% of sand with a [D.sub.50] of 307 [micro]m (from Sibelco France) and 1.5% of adjuvant (polyoxyalkyene polycarboxylate). The mortar was mixed with a water to cement ratio (W/C) of 0.26. The samples were prepared by pouring the fresh mixture of mortar into a rectangular formwork (12 cm x 15 cm x 1 cm) made of polyvinyl chloride (PVC). The surfactants were applied on the surface of the formwork (15 g/[m.sub.2]) about 15 min before pouring the mortar mix. All the mortar samples were removed from their formworks after 18 h. They were stored 28 days under ambient conditions (25[degrees]C; 50% relative humidity) in order to complete the hydration.
The reference sample was cast in a PVC mold with no demolding agent. The study itself was conducted on six surfactants, each of them being used pure or introduced in various proportions (1, 5, 10, and 20 wt%) in demineralized water before being applied to the formwork. All the surfactants were elhoxylated fatty alcohols, with various lengths of hydrophobic chains and various ethoxylation degrees. Table 1 gives the hydrophilic-lipophilic balance (HLB) of each surfactant, i.e., the degree to which they are hydrophilic (the higher the HLB, the more hydrophilic the surfactant). The HLB value was calculated using Griffin's method: HLB = 20 [M.sub.h]/M, where [M.sub.h] is the molecular mass of the hydrophilic part of the molecule and M the molecular mass of the whole molecule. (6)
Table 1: HLB of the six surfactants Surfactant A B C D E F HLB 8 9.2 10.5 13.8 14.5 15.9
The mortar samples are designated in this paper by the notation [X.sub.y], where X is the surfactant (A-F) and y is the proportion of surfactant introduced in the solution. For example, [A.sub.10] designates a mortar sample molded with a solution composed of 90% water and 10% of surfactant A.
Scanning electron microscopy
SEM observations were conducted on both surfaces and cross sections of the mortar samples to determine the CH crystals' microstructure. Small cubes (1 cm x 1 cm x 1 cm) were cut from each mortar sample. For surface observation, the cubes were directly carbon-coated on the adequate face and observed in secondary electron mode. For cross section observations, the cubes were impregnated, polished, and carbon-coated, and then observed in back-scattered electron mode. An SEM FEG Quanta 400 from FEI Company was used at an accelerating voltage of 15 kV and current intensity of 1 mA.
Image analysis has been successfully used in previous studies7 9 to assess the amount of portlandite in mortar as well as the geometrical parameters of the particles (size, shape, etc.). In those cases, image analysis was conducted on SEM micrographs from polished sections. This study focuses on the portlandite that grows at the mortar formwork/interface, which is visible on the mortar surface, so there was no need to work on sections. Furthermore, the crystals are so large that observations under a binocular were sufficient to distinguish them individually. For each mortar sample, five small areas of the surface (1 cm2) have been colored with a black felt-tip pen to increase the contrast between cement paste and portlandite. The black ink is absorbed by the cement paste but not by the portlandite crystals, so they appear very dearly as white areas on a black background.
These colored areas were then photographed under a binocular. The images were treated and analyzed using the open-source software Image J. (10)
The most difficult step in image analysis is to separate the studied particles from the background. In our case, the following procedure was used (Fig. 1):
[FIGURE 1 OMITTED]
- The brightness and contrast of the images were optimized using the automatic function on the software. The optimization was not based on the whole image, but on the histogram analysis of a small area of one of the particles on the image. This step created a highly contrasted image where particles appeared red and the background black.
- The images were then converted to binary and inverted. At this point, the background appeared white and the particles black.
- The noise was removed using the "remove outliers" function of the software, which replaces each pixel by the mean value of its neighbors in a given area.
- Finally, the "watershed" function was used to automatically draw the outlines of the particles. This step might be the main cause of errors in the measurements, as watershed segmentation works best for convex objects that do not overlap too much. In some cases, the software was not able to correctly separate clusters of particles, which were then counted as one big particle.
- The particles were counted and their Feret's diameters measured (longest distance between two points of a particle), as well as the fraction area they covered on the surface.
Mercury intrusion porosimetry was inadequate for assessing the surface porosity because the high pressure involved led to the rupture of the very thin crystals of portlandite and the results were irrelevant. The limitation of this technique was discussed in a review by Diamond. (11) That is why an alternative method was chosen to determine the effects of portlandite on permeability, which consists of measuring the mass of water that is absorbed by capillarity through the mortar surface. This method is a good indicator of what happens at the mortar surface when staining by aqueous solutions occurs or when the mortar is exposed to rainy weather or high humidity. Similar tests could be done with nonaqueous solvents to prevent calcium leaching, but they would not be as representative of the real conditions as this one.
The mortar samples (12 cm x 15 cm x 1 cm) were cured at 45[degrees]C during 6 days in a humidity-controlled atmosphere to make sure that all the samples were charged with the same mass of water prior to analysis. The lateral faces of the sample were coated with an epoxy resin to seal their porosity. The samples were weighed beforehand and immerged face down 3 mm deep in demineralized water. At given intervals of time (1, 4, 9, 16, 25, 36, 49, 64, 100, and 240 min), the samples were removed from water, wiped with a wet cloth, and weighed again. The rate of absorption (g/[m.sub.2]), defined as the change in mass (g) divided by the cross section area of the sample ([m.sub.2]), was plotted against time (min).
Static water contact angle measurements
Static water contact angles were measured using the sessile drop method with an instrument (Drop Shape Analysis DSA-100 from Kruss--Germany) equipped with a CCD camera and an image analysis processor. This technique is sensitive to the topmost of the surface (0.5 nm) that provides information about wettability. The drop volume was 2.5 [micro]L Milli-Q water and the contact angle was read after 5 s. For each sample, the result was determined by averaging data obtained on at least 10 droplets at different locations on the surface. All the measurements were performed under 20[degrees]C and 50% relative humidity.
Influence of the surfactants on the Ca[(OH).sub.2] crystals
Figure 2a shows the surface of the reference sample (no demolding agent) observed by SEM in secondary electron mode. No specific features arc observed at this magnification. The surface is constituted of a homogenous mass of mortar paste separated by hardly visible opened porosity. Several marks due to the formwork defects are also visible.
Figure 2b shows the surface of sample [B.sub.10] observed in the same conditions as Fig. 2a. The mortar surface after demolding appears to be partially covered with large dark areas. The characteristic shapes of these areas (on some samples, perfect hexagons can be observed, see Fig. 2c) as well as their layered structure identify them as portlandite crystals--no other mortar phase would adopt this morphology. Similar observations were made for all the samples except the reference sample.
On samples B to F, all the observed particles share the same layered and flaky structure with different orientations: in all cases, the crystals are composed of thin leaves (less than 1 [micro]m thick--see Fig. 2f) that grow around the nucleation point. For some crystals, these leaves are strictly parallel to the surface (see Figs. 2c and 2d). They grow as "flower petals" around the central point, reaching sometimes a perfect hexagonal shape. For other crystals, the leaves are strictly perpendicular to the surface (see Figs. 2e, 2f, and 2g). When the surface is observed, only cross sections of the leaves are visible and the crystals exhibit a very regular spherulitic structure. But in most cases, the crystals adopt an intermediate orientation, penetrating the surface with various angles (see Fig. 2h). This is why they exhibit a large range of shapes and sizes when observed from the surface.
The proportion of crystals or clusters of crystals perpendicular to the surface tend to increase with the surfactant to water ratio in the demolding agent, as illustrated in Fig. 3, and with the HLB of the surfactant.
The aspect of CH crystals on samples [A.sub.1] to [A.sub.100] is slightly different. The dark areas are much larger than on the other samples (several [cm.sup.2]) and consist of a superposition of many small crystals that grow in an almost fractal pattern (Fig. 4). It is not possible to distinguish the nucleation point or the boundaries of a particulate crystal.
It should be noted that the surfactants have no visible effect on the structure of the calcium silicate phases.
Some samples could not be properly analyzed: sample A because of an excessive overlapping of the crystals and samples [D.sub.100], [E.sub.100], and [F.sub.100] because of their excessive roughness that induced too much background noise on the images.
The results of the image analyses are summarized in Tables 2, 3, and 4.
The results presented in Table 2 show that all the tested surfactants promote the nucleation of Ca[(OH).sub.2] at the mortar surface to a different extent. When no surfactant is used, no crystal of portlandite is observed--a couple of crystals can marginally be observed in some cases, as stated in the introduction, but never in the amount described here--whereas the samples molded with surfactants exhibit between 40 and almost 1000 crystals per [cm.sub.2].
Table 2: Amount of Ca[(OH).sub.2] crystals per [cm.sup.2] % Surfactant 1% 5% 10% 20% 100% A n.m. (a) n.m. (a) n.m. (a) n.m. (a) n.m. (a) B 930 590 220 210 137 C 750 400 280 90 n.m. (b) D 0 0 0 100 n.m. (b) E 0 0 0 60 n.m. (b) F 0 0 0 40 n.m. (b) (a) Not measurable because the particles overlap too much (b) Not measurable because the sample is too rough
To analyze the results, it is important to note that the clusters of crystals that grow perpendicular to the surface are not as easy to highlight on the pictures as the flat ones. The black ink used to create a contrast between crystals and cement paste is absorbed by the free space between the crystals" flakes, resulting in small gray spots. Most of these spots are either too small or too pale to be correctly isolated as particles by the software. Therefore, the figures in Table 2 are more representative of the proportion of parallel crystals than of the actual amount of Ca[(OH).sub.2] nucleation points.
This is why the measured amount of crystals decreases with both the surfactant to water ratio and the HLB of the surfactant. As stated in the "SEM observation" section, higher surfactant to water ratio and higher HLB induce mainly perpendicular clusters which are not detected by the analysis software.
For 10% of surfactant, the decreased number of Hat crystals is somehow counterbalanced by a slight increase in size (Table 3), which finally leads to covered areas of the same order of magnitude (Table 4).
Table 3: Mean Feret's diameter of the Ca[(OH).sub.2] crystals ([mu]m) % Surfactant 1% 5% 10% 20% 100% A n.m. (a) n.m (a) n.m (a). n.m. (a) n.m. (a) B 126 282 360 214 202 C 223 257 386 200 n.m. (b) D No crystals No crystals No crystals 247 n.m. (b) E No crystals No crystals No crystals 231 n.m. (b) F No crystals No crystals No crystals 232 n.m. (b) (a) Not measurable because the particles overlap too much (b) Not measurable because the sample is too rough Table 4: Percentage of surface area covered with Ca[(OH).sub.2] % Surfactant 1% 5% 10% 20% 100% A n.m. (a) n.m. n.m. (a) n.m. (a) n.m. (a) B 15.1 36.3 25.8 6.20 2.80 C 29.1 19.7 30.7 2.80 n.m. (b) D 0 0 0 4.00 n.m. (b) E 0 0 0 1.45 n.m. (b) F 0 0 0 1.90 n.m. (b) (a) Not measurable because the particles overlap too much (b) Not measurable because the sample is too rough
The surfactant-induced crystals are considerably larger (Table 3) than the clusters that can be found in mortar volume (Diamond describes them as typically 1-30 [micro]m large (7)), which is not surprising as they have much more space to develop. No significant trend can be found in the size evolution of the crystals at first sight. It is possible though that the diameter of the crystals increases with the amount of surfactant, but as the crystals also change orientation they are no longer detected by the software, which explains why there is a maximum size of 10% surfactant for both B and C. Indeed, the proportion of parallel crystals strongly decreases beyond this value.
[FIGURE 2 OMITTED]
Influence of the portlandite on mortar surface properties
All the mortar samples were colored with a black felt-tip pen prior to image analysis to increase the contrast between portlandite crystals and mortar paste. This is possible because the black ink is strongly absorbed by the porosity of the paste, but not absorbed by the smooth and dense Ca[(OH).sub.2] crystals. Figure 5 illustrates this phenomenon.
Similar observations were made for all the samples presenting large crystals parallel to the surface ([B.sub.1], [B.sub.5], [B.sub.10], [C.sub.1], [C.sub.5], and [C.sub.10]). On the samples where perpendicular orientation predominates, the ink tends to accumulate between the crystals1 leaves, making its removal difficult.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
In the case of surfactants A, the protective effect of the Ca[(OH).sub.2] areas also exists, but is slightly counterbalanced by the roughness of the surface as some of the ink remains trapped at the overlapping crystals' boundaries.
Figure 6 shows the water absorption rate of the reference sample and of sample [A.sub.100]. The curve represents the amount of water absorbed by capillarity by the mortar surface through time, which reaches a maximum after a couple of hours. After 4 h, both the reference sample and the sample Am have reached their maximum absorption capacity, but the sample [A.sub.100], which is partially covered by large continuous areas of portlandite, has absorbed half as much water as the reference sample.
Similar results were obtained for all the A samples, but not for the other series of samples. This is not surprising as only a small fraction of the surface area is actually covered and protected by the portlandite crystals (2-30%, see Table 4). Although the crystals are locally very efficient to close the porosity, their effect is not visible on a macroscopic scale.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
Static water contact angles
The static water contact angle measured on the reference sample was 40[degrees]. For all the other samples, the angle was below 10[degrees], showing a high affinity of the surface to water. The mortar surface becomes super hydrophilic after the growth of Ca[(OH).sub.2].
Influence of the surfactants on Ca[(OH).sub.2] nucleation
As stated in the "Introduction/' Martin5 showed that occasional precipitation of Ca[(OH).sub.2] occurs at the mortar surface when the formwork is smooth enough to attract a layer of water by capillarity. When water is reabsorbed by the volume, portlandite precipitates at the mortar surface. Adding a surfactant to the form-work triggers this process on two levels. First, it encourages the formation of the water layer--contrary to an oily demolding agent that would prevent it. Second, it promotes homogeneous nucleation by reducing the nucleus/water interfacial energy. Indeed, the theory of homogeneous nucleation explains that when a spherical nucleus forms spontaneously in a supersaturated solution, it has to reach a critical diameter dc to keep growing and form a crystal. The critical diameter is expressed by the following formula (12):
[d.sub.c] = 2[sigma]]M/[rho] RT ln S
where [sigma] is the nucleus/solution interfacial energy (J/[m.sub.2]), M is the molecular mass of the nucleus (kg/mol), p is the volume mass of the nucleus (kg/[m.sup.3]), R is the gas constant (J/mol K), T is the temperature (K), and S is the supersaturation ratio.
The presence of the surfactant decreases the nucleus/water interfacial energy and stabilizes the growing nuclei.
Influence of the surfactants on Ca[(OH).sub.2] structure
In a nonconstrained environment, Ca[(OH).sub.2] crystallizes in the shape of micrometric hexagonal platelets with an [L.sub.1]/[L.sub.2] ratio of typically 2 to 3 (see Fig. 7). For example, Harutyunyan et al. (13) successfully monitored the growth of Ca[(OH).sub.2] from cement by X-ray transmission microscopy and obtained portlandite crystals with an [L.sub.1]/[L.sub.2] ratio of 2.7.
On the mortar samples studied here, the portlandite crystals are extremely thin ([L.sub.1]/[L.sub.2] [approximately equal to] 500) and exhibit a flaky structure composed of superposed layers. This layered structure of portlandite in mortar is not unusual: it has been described by Stark et al., (14) Knapcn and Van Gemert, (15) and Harutyunyan et al. (16) at the Interfacial Transition Zone.
The fact that many crystals grow strictly parallel to the surface is due to the presence of the formwork which constrains the crystals' development. The relatively large size of the crystals (around 300 [micro]m against a few tens of micrometers in the volume) is not surprising either as the crystals have time and space to develop.
The presence of the surfactant acts on two levels. First, it slightly amplifies the natural trend of portlandite to form thin layers, probably because the hyclrophilic chain of the surfactant absorbs on the hexagonal faces of Ca[(OH).sub.2] and inhibits their growth. As a result, the crystals are even thinner and larger than they would spontaneously be. The same phenomenon was observed by Atahan et al. in air voids. (17) The portlandite that precipitated in air voids generated with surfactants was thinner than the one with no surfactants.
[FIGURE 7 OMITTED]
Second, the surfactant modifies the orientation of the crystals. When the HLB or the surfactant to water ratio increases, the crystals grow from a few micrometers deep under the surface mortar and adopt a very oriented spherulitic structure. Although the crystallo-graphic mechanisms behind this are unclear, they are very likely linked to the fact that these surfactants form a gel when they are mixed with a certain proportion of water (see Fig. 8) and adopt a very structured, probably layered, organization. This explanation is consistent with the fact that the amount of perpendicular crystals increases with both the HLB and the surfactant to water ratio. Indeed, the quantity of water necessary to form a gel decreases when the HLB of the surfactant increases. For example, surfactant C starts to form a gel at a surfactant to water ratio of 0.37, while surfactant D forms a gel at very low ratios (1% surfactant in water is enough).
Surfactant A is the only tested product to induce the formation of a continuous layer of Ca[(OH).sub.2]. It is unclear though whether the size of the portlandite areas is due to a massive nucleation or a massive growth. It should be noted that surfactant A has a specific behavior in the presence of cement and water: it forms microscopic bubbles filled with cement and is stabilized by the surfactant. Furthermore, it is too hydrophobic to mix with water. One possible scenario is that each of these bubbles acts as a nucleus at the interface. Since the surfactant does not mix with the cement paste, the growth then occurs strictly parallel to the formwork, leading to the formation of many small overlapping crystals.
Influence of the Ca[(OH).sub.2] crystals on mortar surface properties
As staled in the "Introduction," there are two ways to improve the mortar surface protection: make the surface super-hydrophilic, so that the dust particles or biological germs can be easily washed by dripping water, and reduce the surface porosity to limit pollutants or staining agents' penetration and incrustation.
[FIGURE 8 OMITTED]
Regarding the hydrophilic properties, the mere presence of Ca[(OH).sub.2], which has strong interactions with water, combined with residues of surfactant is enough to considerably reduce the water contact angle.
Regarding the porosity, the results are very promising. Surfactant A induces the formation of large areas of Ca[(OH).sub.2] (several [cm.sup.2]) that reduce the capillary absorption of aqueous liquids very efficiently. Surfactants B and C in small quantities (10% maximum) induce the formation of smooth flat crystals that close the porosity and locally prevent the incrustation of staining agents. The surface area actually covered with these crystals is less than 30%, which is not enough to efficiently protect the whole surface yet. As for surfactants D to F, they systematically orientate the crystals perpendicularly to the surface, which annihilates their protective effect.
Although some improvements still have to be made, surfactants A to C are excellent candidates and their use is very promising.
It should be noted that all the effects observed here are not altered by the carbonation effects that take place at the surface. Through time, Ca[(OH).sub.2] in contact with atmospheric C[O.sub.2] slowly turns to CaC[O.sub.3], which has the same hydrophilic properties as Ca[(OH).sub.2] (18) and is likely to close the porosity the same way.
This study proposes a new method to protect the surface of HPC from staining and dust deposition, that consists of closing the surface porosity by growing a mineral coating made of large hydrophilic crystals of Ca[(OH).sub.2]. This method presents numerous advantages: the coating grows before demolding, and does not necessitate any costly post-treatment or massive use of additives.
The nucleation and growth of the Ca[(OH).sub.2] crystals are triggered by nonionic surfactants. With a hydrophobic surfactant (HLB = 8), it is possible to grow a continuous layer of Ca(OH)2 that significantly reduces the capillary absorption of aqueous liquids but also induces a roughness in which staining agents tend to accumulate. With more hydrophilic surfactants (HLB = 9.2 and 10.5), it is possible to grow smooth flat crystals that prevent the surface from staining but only cover 30% of the surface.
The most hydrophilic surfactants tested (13.8-15.9) are not good candidates as they favor perpendicular growth of the portlandite crystals. For the same reason, the use of more than 20% surfactant in the demolding solution is not recommended.
In every case, the water contact angle of the HPC surface is reduced from 40[degrees] to less than 10[degrees], improving the self-cleaning properties of the mortar by facilitating the removal of dirt and germs by dripping water.
This study is still at an explorative stage, and further research is necessary to fully explore what seems to be a very promising path.
Improvements can be made on several aspects:
- Increase either the amount or the size of the crystals to cover the totality of the surface.
- Fully exploit the potential of surfactant A, which is the most promising product amongst those tested.
- Investigate other surfactants in the same range of HLB that could have similar effects. In particular, the influence of the carbon chain has not been investigated here.
- Evaluate the interest of combining the methods presented here with other self-cleaning methods such as the use of photocatalytic Ti[O.sub.2]
On a less applicative level, other perspectives could include detailed crystallographic investigations of the Ca[(OH).sub.2] structures, in particular of the spherulitic organization which is very unusual.
Acknowledgments Eleonore Gueit is supported by a Ph.D. fellowship from Lafarge. The authors would like to thank Melanie Dykman for her assistance during the preparation of the samples and the water absorption measurements, Catherine Bouillon for the SEM observation, and Jeffrey Chen for his advice regarding this work.
(1.) Jones, FN, "Oulook for Zero-VOC Resins." 7. Coat. Technoi, 73 (916) 63-71 (2001)
(2.) Scheerder, J, Visscher, N, Nabuurs, T, Overbeek, A, 4iNovel Water-Based Fluorinated Polymers With Excellent Antig-raffiti Properties." J. Coat. Technol Res., 2 (8) 617-625 (2005)
(3.) Maury, A, De Belie, N, "State of the Art of Ti[O.sub.2] Containing Cementitious Materials: Self-Cleaning Properties." Mater. Construct., 60 (298) 33-50 (2010)
(4.) Pilcher, GR, "Meeting the Challenge of Radical Change: Coatings R&D as We Enter the 21st Century." J. Coat. Technol, 73 (921) 135-143 (2001)
(5.) Martin, M, "Etude de la texture de la surface coffree des parements verticaux en beton." PhD Thesis, Universite de Cergy-Pontoise (France) et Universite de Lavai (Quebec), 2000
(6.) Griffin, WC, "Classification of Surface-Active Agents by 'HLBV J. Soc. Cosmet. Chem., 1 311-326 (1949)
(7.) Diamond, S, "Calcium Hydroxide in Cement Paste and Concrete--A Microstructural Appraisal." In: Skalny, J, Gebauer, J, Odler, I (eds.) Calcium Hydroxide in Concrete, pp. 37-58. The American Ceramic Society, Westerville (2001)
(8.) Glasser, FP, "The Role of Ca[(OH).sub.2] in Portland Cement Concretes." In: Skalny, J, Gebauer, J, Odler, I (eds.) Calcium Hydroxide in Concrete, pp. 11-35. The American Ceramic Society, Westerville (2001)
(9.) Gallucci, E, Scrivener, K, "Crystallisation of calcium hydroxide in early model and ordinary cementitious systems." Cem. Concr. Res., 37 (4) 492-501 (2007)
(10.) NIH, http://rsbweh.nih.gov/ij (2010)
(11.) Diamond, S, "Review: Mercury Porosimetry. An inappropriate method for the measurement of pore size distributions in cement-based materials." Cem. Concr Res., 30 (10) 1517--1525 (2000)
(12.) Mullin. JW. Crystallization, 1st ed. Butterworths, London (1961)
(13.) Harutyunyan, VS. Kirchheim. AP. Monteiro, PJM, Aivazyan, AP. Fischer, P, "Investigation of Early Growth of Calcium Hydroxide Crystals in Cement Solution by Soft X-ray Transmission Microscopy." J. Mater Sci, 44 (4) 962-969 (2009)
(14.) Stark, J, Frohburg, U, Moser. B, "Investigations by Environmental SEM: Ca[(OH).sub.2] During Hydration and During Carbonation." In: Skalny, J, Gebauer, J, Odler, I (eds.) Calcium Hydroxide in Concrete, pp. 167-189. The American Ceramic Society. Weslerville (2001)
(15.) Knapen, E. Van Gemert, D. "Cement Hydration and Microstructure Formation in the Presence of Water-Soluble Polymers." Cem Concr. Res., 39 (1) 6-13 (2009)
(16.) Harutyunyan, VS, Abovyan, ES, Monteiro, PJM, Mkrlchyan, VP, Balyan, MK, Aivazyan, AP, 4iX-ray Diffraction Investigation of Microstructure of Calcium Hydroxide Crystallites in the Interfacial Transition Zone of Concrete." J. Am Ceram. Soc, 86 (12) 2162-2166 (2003)
(17.) Atahari, HN, Carlos, C, Chae, S, Monteiro. PJM. Baslacky. J. "The Morphology of Entrained Air Voids in Hardened Cement Paste Generated with Different Anionic Surfactants." Cem. Concr. Compos., 30 (7) 566-575 (2008)
(18.) Legens. C, Palermo, T, Toulhoat, H, Fafet, A, Koutsoukos, P, "Carbonate Rock Wettability Changes Induced by Organic Compound Adsorption." J. Pet. Sci. Technol., 20 (3-4) 277-282 (1998)
E. Gueit (*), E. Darque-Ceretti
MINES ParisTech, CEMEF - Centre de Mise en Forme des Materiaux, CNRS UMR 7635, BP 207 1 rue Claude Daunesse, 06904 Sophia Antipolis Cedex, France e-mail: email@example.com
P. Tintillier. M. Horgnies
LCR - Lafarge Centre de Recherche, 95 rue du Montmurier
BP15, 38291 St-Quentin-Fallavier, France