CLAY HYDRATION MECHANISMS AND THEIR EFFECT ON DUSTINESS

Abstract Clays are employed in a wide variety of industries such as ceramic industry, manufacture of paper, rubber, etc. In this sense, it is well known that at industrial processes in which clayey materials are used, such as ceramic industry, in order to carry out some specific stages, the wetting of clays is commonly required. Moreover, it is also long established that wetting is an appropriated measure to reduce particulate matter emissions during clays storage and handling. The present study was undertaken to assess the influence of moisture on clay dustiness because, though the complex behaviour of the clay–water system has been known since antiquity, the mechanisms involved in clay hydration and their influence on dustiness are still not well understood. To encompass a wide range of specific surface areas, three clays and a kaolin were studied. Chemical and mineralogical analysis of these four raw materials was performed and their particle size distribution, flowability, true density, plastic limit, and specific surface area were determined. Raw materials dustiness was determined using the continuous drop method. As against what might intuitively be expected, the results showed that the relationship between moisture and dustiness was quite complex and strongly related to the hydration mechanisms. In this regard, to better understand the phenomena involved in the clay hydration process, a specific methodology was developed to estimate the critical points of the clay hydration process (regarding dustiness). This methodology can be readily applied to other clays or even to materials of different nature to predict the optimum moisture and, therefore, it could be employed to propose specific measures which could entail an improvement of outdoor and indoor air quality.


Dust generation associated with clay use
Clays are used in a wide variety of industries, such as the ceramics, paper, rubber, and absorbent industries (Mukherjee and Ghosh, 2013). A vast number of workers are thus potentially exposed to particulate matter emitted from dry mechanical processes such as screening and grinding and clay handling and transfer operations (Mukherjee and Ghosh, 2013).
Some industrial activities, such as ceramics manufacture (floor and wall tiles, sanitary ware, tableware, ceramic refractories, etc.) and cement manufacture (grey and white Portland cement, concrete, etc.), are particularly at issue, as many of these industrial goods are claybased. The raw materials mixture used in manufacturing these products includes considerable amounts of clay, which play a vital part in the different manufacturing process stages. By way of example, figure 1 presents a break-down of the use of national clays in Spain (8.5 million tonnes) in 2012. About 86% of these clays were used in the ceramic industry, 10% in cement manufacture, and 4% in other industries. A parameter of great interest in evaluating, controlling, and minimising the risks associated with the particulate matter emissions generated during clay handling is clay dustiness, i.e. clay of 20 tendency to produce dust when handled. In this regard, as the resulting dustiness depends on the apparatus used, standard EN 15051:2013 proposes two test methods: the rotating drum method and the continuous drop method. The test enables the mass fractions of inhalable dust (w I ), thoracic dust (w T ), and respirable dust (w R ) to be determined. The test results allow the materials to be classified in different categories as a function of their dustiness, according to the criteria laid down in the above standard.

Water as dust suppressor
It is long established that moisture content increases the interparticle binding forces, which leads to less dust generation (WHO/SDE/OEH/99.14). Dustiness was observed to decrease with increasing moisture content in different experimental and observational field studies in which water was added to dusty solids to reduce their dustiness (Alwis et al., 1999;Farrugia et al., 1989;Hjemsted and Schneider, 1996;Plinke et al., 1992;Teschke et al., 1999;Westborg and Cortsen, 1990). Industrial examples of the use of water as a dust suppressor include using water as a wetting agent in the bulk outdoor storage of certain dusty materials, wet processing of minerals, and the use of slurries and wetted materials in the ceramic industry (WHO/SDE/OEH/99.14). Water is the most common dust suppressor because of its availability, low cost, and chemical compatibility.
However, other studies (Cowherd et al., 1989;Plinke et al., 1995) found that the dustiness index of some materials rose when moisture content increased slightly, dustiness subsequently decreasing when moisture content was further raised.
Moisture may therefore be considered to have a rather complex effect, which may vary from one material to another, depending on the extent to which water is absorbed internally or is Page 6 of 20 attached to the particle surface (Leith D. 1991). That is, its effect depends on the material and on the material's surface properties and hygroscopicity (WHO/SDE/OEH/99.14).

Clay hydration mechanisms
In order to understand the effect of moisture on clays dustiness, it is needed to know the mechanisms involved in the hydration of clays. Clay water retention takes place by two mechanisms: adsorption and capillary condensation, and the following domains may be distinguished (Prost et al., 1998): In this adsorption domain, water is first adsorbed on the material's hydrophilic sites (exchangeable cations, edge cations, and surface OH groups). It can induce the creation of pillars between clay layers, expanding the layers (interlamellar swelling) and eventually, results in a creation of a film of water on surfaces. Capillary condensation at contact points between particles may also occur in this domain. Domain 2: Multilayer adsorption then occurs on "free surfaces", which are either unsaturated pore walls or surfaces of layers or particles that can expand freely. This does not necessarily mean that the entire water structure adsorbed by this process is affected by the surface: only about 2-3 layers of adsorbed water are perturbed by clay surfaces (Prost et al., 1998;Revil and Lu, 2013).
In this regard, Hagamassy (1970) obtained the statistical number of water layers versus relative pressure (P/P 0 ) for a wide variety of non-porous adsorbents. Hagamassy classified the obtained results in different groups based on the heat of adsorption of water on the adsorbents, to this regard each group was characterised by its BET constant (C), which is an index linked with the affinity between adsorbate and adsorbent. Hagamassy results can be summarised as follows: for low relative pressures (P/P0 < 0.5) four groups were established (C=5.2, C=10-14.5, C=23, C=50-200), whilst for high relative pressures (P/P0 > 0.5) two groups were distinguished (C=5.2 and C=10-200) The C values obtained in our study were in the range 6.9-7.2 and, therefore, the results obtained by Hagamassy for the group with C=5.2 have been considered the most appropriated for all the range of relative pressures. Ormerod and Newman (1983) and Moore et al. (2007) postulated that water sorption on the external surfaces of clays was similar to that on reference non-porous oxides (Hagamassy, 1970), so that the relation between relative pressure and the statistical water layers may also be assumed to be valid for porous materials. As the amount of water increases, the added water is attracted by the water already adsorbed at the solid surface by the cohesion forces while tending to minimise the water-air interface area, so that particles start to be held together by liquid bridges at their contact points (pendular state). If more water is added, some pores become fully saturated by liquid, while of 20 air-filled voids still remain (funicular state). Note that the transition between the pendular and the funicular state (5% saturation degree) entails about a threefold increase in granule tensile strength from the pendular state (Oliva, 2015). Water then starts to fill the interparticle voids until all these voids are filled with liquid, but the surface liquid is drawn back into the pores under capillary action (capillary state).

Domain 4:
This domain is related to the multilayer adsorption process on grain or particle "free surfaces" and to a capillary condensation mechanism in saturated pores whose size increases or decreases as swelling or shrinkage occurs. Finally, particles are fully immersed in liquid (slurry state).

Figure 2. Scheme of the hydration process
In view of these mechanisms, these stages may reasonably be assumed to influence interparticle attractive forces and therefore clay particle cohesion and dustiness. To identify these mechanisms and their effect on dustiness, the identification of the critical points was deemed crucial. However, no clear and compressive methodology to experimentally determine and quantify these critical points was found in the reviewed literature.
Page 9 of 20

Objective and scope
Though water is the main dust suppressor in the use of clays in industrial practice and the complex behaviour of the clay-water system has been known since antiquity, the clay hydration mechanisms and their influence on dustiness are still not well understood. This study was therefore undertaken with the following objectives:  To evaluate the effect of moisture on clay dustiness.
 To develop a methodology to establish the critical points in clay hydration.
 To analyse the relationship between dustiness and the clay hydration mechanisms.
As the effect of moisture depends on the extent to which water is adsorbed, the study was conducted on four clays that encompassed a wide range of specific surface areas (12-33 m 2 /g). Many industrial clays may therefore be assumed to behave like the studied raw materials. Moreover, the relationship found between hydration mechanisms and dustiness could help to predict or comprehend the effect of moisture on other clays or even on materials of different nature, and this information could be deemed to be essential in order to propose specific preventive measures, which could entail an improvement of air quality and in occupational health when clays are industrially processed.

Raw materials used
The study was conducted on three clays and a kaolin, selected to encompass a wide range of specific surface areas. The main characteristics of these raw materials were determined: chemical analysis (table 2); mineralogical analysis (table 3 and table 4); particle size distribution by laser diffraction measurements (figure 2); true density and specific surface area (table 5). The procedures used in these characterisations are set out in point 3.2.     To evaluate the influence of moisture on dustiness and to cover a wide moisture range (0.05-14%), twenty-seven test samples were prepared by respective wetting of the four studied raw materials.

Experimental procedure
The methodology followed can be divided into three parts: physic-chemical characterisation of the raw materials, physical characterisation of the bulk samples (dry and wet samples), and identification of the critical points in the clay hydration process.

Physic-chemical characterisation of the raw materials
Raw materials chemical analysis (table 2) was carried out by wavelength dispersive X-ray fluorescence spectrometry (WD-XRF model Axios Panalytical) using reference materials to guarantee measurement traceability. Loss on ignition at 1000ºC was carried out using a LECO model TGA-701 thermogravimetric analyser. (table 3) was performed by XRD quantitative mineral analysis. A Bruker Theta-Theta D8 Advance diffractometer was used for this purpose. The samples were ground to a particle size of less than 40 μm. Readings were taken at 2θ intervals between 5º and 90º. TOPAS software was used to perform Rietveld quantitative analysis.

Mineralogical characterisation
Particle size distribution (PSD) (figure 2) was determined by laser diffraction (ISO 13320 -1), using the wet method, i.e. solid samples were dispersed in water, which acted as a dispersion medium.
Specific surface area (

Characterisation of the bulk samples
In determine these parameters is briefly described below.

Hausner ratio
The flowability of the particulate samples was evaluated by determining the Hausner ratio (HR), i.e. the quotient of the bulk density of the packed particle bed (by vibration or tapping) and the aerated bulk density of the particle bed (Mallol et al., 2008). The Hausner ratio allowed the materials to be classified as a function of their flowability, as detailed in table 6. basically, of a cylindrical pipe through which air circulated in an upward direction at a volume flow rate of 53 l/min. The test material was dropped from the top of this pipe at a rate of 6 to 10 g/min. The dust dropped through an inner pipe, concentric to the pipe through which the air rose. This tube was shorter than the outer pipe, so that the dust was released into a counter-current airflow. A pair of sampling heads for health-relevant particle size fractions were located slightly above the discharge position of the material, enabling the inhalable and of 20 the respirable fractions to be studied simultaneously. The inhalable fraction represented the mass fraction of total airborne particles which is inhaled through the nose and mouth, whereas the respirable fraction represented the mass fraction of inhaled particles penetrating to the unciliated airways (EN 481:1995).

Figure 4. Dustiness tester used and detail of the feeder and sampling heads
In addition, the results obtained allowed the materials to be classified in different categories as a function of their dustiness, according to the criteria laid down in the above standard (table   7).

Methodologies to identify critical stages or points
Page 15 of 20 As mentioned above, no clear and compressive methodology to experimentally determine and quantify the critical points in clay hydration were found in the reviewed literature. An experimental methodology, based on three experimental determinations (moisture adsorption isotherm, saturation degree, and plastic limit), was therefore developed for this study. This methodology and its relationship with the domains described in the introduction are summarised in figure 5.

Moisture adsorption isotherm
The water adsorption isotherms of the clays were determined in a DVS Aquadyne Quantachrome instrument using water vapour as adsorbent. The amount of adsorbed water vapour was measured by a static gravimetric method, obtaining moisture gain values depending on the relative humidity. For each relative humidity, equilibrium was reached. The isotherms were obtained at room temperature.

Saturation degree
Page 16 of 20 The saturation degree is the ratio of the liquid volume (V w ) to the total void volume (V v ) in a porous material.

Air volume
Water volume The saturation degree was obtained from the ratio of free water volume to void volume (equation 1), void volume being determined from the true density and the aerated bulk density of the particle bed (equation 2).

Equation 1
In this study, in order to estimate bound water, the water volume associated with the three adsorbed layers perturbed by the clay surface was determined (equations 2 and 3) from the specific surface area and the thickness of three water molecules (0.9 nm).

Equation 2 Equation 3
Page 17 of 20 According to Oliva (2015), the transition between the pendular and the funicular state (critical point E) lies at 5% saturation degree. The moisture at which this saturation degree was reached was therefore determined.

Plastic limit
In the proposed methodology, the dustiness values (inhalable and respirable fractions) in the plastic limit were assumed to be zero (critical point F). The plastic limit was determined according to the procedure described in UNE 103104:1993. Table 8 details the particle size distribution, specific surface area, moisture content, bulk density, Hausner ratio, and dustiness data (inhalable and respirable mass fractions) obtained for each studied raw material and the corresponding test samples.  K1-7 29 --0 0 A) di: diameter below which lies I% by volume of total particles B) S: specific surface area C) Hausner ratio D) wI: inhalable mass fraction; wR: respirable mass fraction E) Plastic limit

Determination of the critical points by applying the developed methodology
The influence of the hydration mechanisms on dustiness was evaluated by using the methodology described above to identify the critical points or stages in the hydration process.
By way of example, Figure 9 shows the moisture adsorption isotherms for C1 and C2. Table   10 summarises the critical points obtained for the four studied materials.

Discussion of dustiness results based on the identified critical points
The critical points and variation of dustiness and Hausner ratio with moisture for the four studied raw materials are shown in figures 10 to 13. The critical points always appear to be the same for the studied materials, suggesting that the critical points are related to the different effect of water, depending on the mechanism involved. In the first stage, the added water (A→B) increased cohesion (raising the Hausner ratio and lowering dustiness) owing to the generation of liquid bridges, etc. However, as more water molecular layers formed (B→C), interparticle distance increased and interparticle attraction decreased (lowering the Hausner ratio and raising dustiness).
When more water was added (C→D), these water layers, which were not perturbed by the clay surface, presumably tended to stick particulate groups together, thus again increasing cohesion (raising the Hausner ratio and lowering dustiness).
Dustiness maximised in this stage, as a function of the presence of free particles, particle size, and agglomerate size. Finally, when more water was added (E→F), though flowability improved, granulate strength increased about threefold (Oliva, 2015), constraining dust release during handling (funicular state) and thus reducing dustiness. Finally, from the capillary state on, cohesion increased and dustiness decreased until the latter became negligible (presumably at the plastic limit).
It may be noted that these results are consistent with previous studies in which the dustiness index of some materials rose when the moisture content increased slightly, the dustiness index then decreasing when moisture content increased further (Cowherd et al., 1989;Plinke et al., 1995). of 20 Finally, though water adsorption isotherms were not examined in depth in the present study, the clay hydration mechanisms involved were observed to significantly influence the effect of moisture on dustiness. The results indicate that the effect of moisture was quite complex and depended not only on the material but also on water molecule location, which raised or lowered cohesion and therefore increased or decreased dustiness. This behaviour cannot be extrapolated directly to other materials because the effect of moisture depends on the material and on the material's surface properties and hygroscopicity (WHO/SDE/OEH/99.14). However, the methodology developed for the identification of the critical points from the water adsorption isotherm can help predict or better understand the effect of moisture on the dustiness of other clays or even on that of materials of different nature. The methodology may thus be considered general enough to be extrapolable to other types of materials.

Conclusions
In this study of the variation of dustiness of four clay materials with moisture content, as against what might intuitively be expected, the resulting behaviour was quite complex. The effect of moisture is known to depend on the material (WHO/SDE/OEH/99.14), but the results indicate that this also depends on whether the water acts as a binder or as a lubricant, i.e. it depends on water location.
To better understand the phenomena involved in the clay hydration process, a specific methodology was developed to estimate the critical points of the clay hydration process based on the water adsorption isotherm, saturation degree, and plastic limit.
The methodology used and the results obtained allowed the critical clay hydration points and their influence on the effect of moisture on dustiness to be identified. Finally, considering that the developed methodology is based on the mechanisms that take place during the hydration of powders, it should be highlighted that it can be readily applied to other types of clays or, more generally to different types of hygroscopic materials.