On the Use of Sodium Chloride and Calcined Diatomite Sludge as Additives to Improve the Engineering Properties of Bricks Made with a Clay Earth from Jun (Granada, Spain)

Solid bricks manufactured out of clayey earth from a quarry near the city of Granada (Spain) were studied and compared with others to which two additives were added during the kneading of the raw material: Calcined diatomite sludge and sodium chloride. Samples with and without additives were fired at 800 ◦C, 950 ◦C, and 1100 ◦C. New mineral phases were formed in the bricks after firing. These included gehlenite, diopside, and plagioclase, which is gradually enriched in calcium, and microcline, which is transformed into sanidine. Mullite and molysite also appeared in the bricks made with added salt. Porosity increased substantially in the bricks that contained diatomite sludge, while the addition of sodium chloride accelerated the mineralogical transformations and caused sintering at relatively low temperatures of 800 ◦C. The bricks became more compact and less anisotropic as the firing temperature increased. This behaviour was less evident in the bricks with added diatomite sludge, which alters the orientation of the phyllosilicates. The bricks fired at 1100 ◦C had the best physical parameters. However, the bricks with added sodium chloride fired at 800 ◦C reached hygric and compactness values that were at least similar to bricks from the other groups fired at 950 ◦C.


Introduction
Bricks are some of the oldest manmade masonry building materials, which is why they are often found in monuments with great heritage value [1]. For bricks to last, they must have certain physical and mineralogical properties [2][3][4][5][6]. These properties are undoubtedly related to the composition of the raw material and the manufacturing process, such that the chemical and mineralogical composition, the grain size, firing temperature, and the atmosphere (oxidation or reduction) inside the kiln all have a considerable influence on the porosity, mechanical resistance, and colour of the bricks [7][8][9][10][11]. Bricks are usually made out of local earths, which means that the composition varies from one region to the next. This also makes them very economical and easy to produce. The brick production process has changed little over thousands of years. Perhaps the most important change is that today there is tight control over the different stages of the production process to ensure that a more standardized, high quality material is obtained. In spite of these controls, worn or damaged bricks can sometimes be found in new buildings. This may be due to defects in the production process [12]. For this reason, the brick industry is experimenting with new resources and ideas aimed at improving the properties compactness, and colour. This will allow us to identify the best (or worst) characteristics of the fired pieces, thus optimizing their physical properties and reducing production costs due to the need for smaller amounts of raw material (in the case of calcined diatomite sludge) or lower firing temperatures (in the case of sodium chloride).

Raw Materials, Additives, Preparation, and Firing of the Test Samples
The clayey earth selected for making the bricks comes from the village of Jun, near Granada (Spain). Geologically speaking, these are grey clays, silts, and sands from the mid-late Turolian Age. The clays are interspersed with small layers of gypsum, which are increasingly frequent towards the top of the formation, and occasionally reddish carbonates that are rich in gastropods.
As regards the additives, the diatomite sludge with organic remains (above all barley and corn) was supplied by the Spanish beer manufacturer, Cervezas Alhambra S.L. (Granada, Spain), which uses diatomite in the beer filtering process. To reap the maximum environmental benefits, as much sludge as possible should be added to the mix, providing of course that this does not have a detrimental effect on the physical quality of the bricks [19]. In previous research [33], a proportion of 10 wt.% diatomite sludge mixed with clayey earth was found to be the maximum amount that could be added to produce bricks acceptable to the construction sector. The sodium chloride is edible sea salt produced by the company, Salinera Española S.A. (Palma de Mallorca, Spain), which is obtained by natural evaporation of seawater. 30 wt% salt was added to the water used for kneading the clay to analyse possible improvements in the quality of the fired pieces, such as, for example, greater vitrification. The salt was dissolved in running water before being used to knead the clayey earth. The salt concentration used in this work is higher than that used by other authors [37] or than in seawater. The advantage of a higher concentration is that it enhances the changes to be identified at mineralogical and physical levels.
The bricks were made according to craft techniques. The first stage was to remove the largest fragments (centimetre size) composed above all of gypsum and carbonates. Then, approximately 1 kg of clayey earth was mixed with 500 mL of water. This amount was calculated by adding the water little by little until the paste acquired the necessary consistency for moulding the pieces. The paste was then put in a wooden mould measuring 15 × 20 × 4 cm 3 and was compacted by pressing down initially with the palm of the hand and later with a rolling-pin. The pastes with and without additives were taken out of their moulds while still damp and cut with tensor wire to obtain cubic samples with 4 cm sides and elongated, prism-shaped samples measuring 2 × 4 × 8 cm 3 . They were then left outside to dry for approximately one week until they had lost all the kneading water.
The dried raw bricks were then fired in a Herotec CR-35 electric kiln in an oxidation atmosphere at 800, 950, and 1100 • C. 950 • C is one of the most commonly used temperatures in brick production. The other two firing temperatures, one lower than 950 • C and the other higher, were chosen to study the changes that take place in the mineralogy and the physical properties of the bricks over a range of 300 • C. The temperature was kept constant at 100 • C for 1 h to ensure that there was no moisture left in any of the raw, unfired pieces. The samples were then heated up to the selected firing temperature at a heating speed of 2 • C/min. Once this temperature had been reached, it was maintained for 3 h. Finally, the bricks were left to cool inside the kiln until the next day, with the whole firing-cooling process lasting about 24 h. When the bricks were taken out of the kiln, they were submerged in water in order to eliminate possible grains of calcium oxides that could cause them to shatter due to lime blowing [37].
Overall, nine types of bricks were prepared. These were labelled as J, JS, and JD in order to distinguish samples with no additives (J) from those made with added salt (JS) and those made with calcined diatomite sludge (JD). Each label is followed by a number (800, 950, or 1100) specifying the firing temperature in • C.

Chemical and Mineralogical Characterization
The clayey material and the calcined diatomite sludge were studied using X-ray fluorescence (XRF). A Philips Magix Pro PW-2440 wavelength dispersive sequential spectrometer with a 4 kV X-ray generator was used to analyse major and trace elements. Prior to the analysis, 5 g per sample was finely milled in an agate mortar. The accuracy of the analytical results was evaluated by comparison with certified values of analysed reference materials [38]. Typical accuracy is higher than 1.5% relative to a concentration of 10%. Loss on ignition was determined gravimetrically as the weight loss was recorded between 110 and 1000 • C. The De Jongh model [39] was followed to convert the XRF intensities into concentrations.
The identification of the mineral phases in the raw material, the calcined diatomite sludge, and the bricks with and without additives fired between 800 and 1100 • C was carried out by X-ray diffraction (XRD) using the disoriented powder method. Samples were milled in an agate mortar to a particle size of less than 0.053 mm. A Philips X'Pert PRO diffractometer was used with the following working conditions: CuKα radiation (λ = 1.5405 Å), 45 kV voltage, 40 mA current, 3-60 • 2θ explored area, and 0.1 2θs −1 goniometer speed. The interpretation of results was performed using the XPowderX computer program [40]. Quantitative analysis of the mineral phases was performed using the non-linear least square method to fit full-profile diffractograms and the results were compared with standard values in the database. For this task, the program implements the following refinements and corrections: a) Theoretical mass absorption coefficient of the samples as a function of the chemistry and density of each phase found in the database sheets; b) pattern intensity ratio (PIR) correction, initially using the experimental RIR (reference intensity ratio) for each mineral phase contained in the sample and finally obtaining a PIR value upon full pattern refinement. The American Mineralogist Crystal Structure Database (AMCSD) was used to identify the mineral phases. An indicative value of the amorphous versus crystalline phases (a/c) ratio was provided. This ratio is based on the mean value of the intensities, standard deviation, and area of the crystal reflection [40]. In order to obtain the mineralogical composition of the grain fraction measuring less than 2 µm, 100 g of the raw material was submerged in a beaker with deionized water treated with acetic acid (CH 3 COOH) with 1N concentration to eliminate carbonates. This treatment was repeated until the reaction ceased. The sample was then washed several times to eliminate the acid in the water, and hydrogen peroxide (H 2 O 2 ) with a concentration of 20 vol.% was added to eliminate any possible organic matter. After that, sodium hexametaphosphate was added (5 g/L) to separate any aggregates that might have formed. The fraction below 2 µm was separated using a Kubota Ks-8000 centrifuge and the oriented aggregate method was used to study it with X-ray diffraction. Air-dried oriented aggregate (OA) was prepared for general study. Other aggregates were also prepared: Dissolved with ethylene-glycol (OA + EG) to detect the presence of smectites and chlorites [41], aggregates dissolved with dimethyl-sulfoxide (OA + DMSO) to confirm the presence of kaolinite [42], and aggregates heated at 550 • C to destroy kaolinite and confirm the presence of chlorite.

Physical Characterization
Mercury intrusion porosimetry (MIP) was used to study the porous system of the bricks using a Micromeritics Autopore III 9410 apparatus (Norcross, GA, USA). This porosimeter can exert a maximum pressure of 414 MPa and can identify pores with radii comprised between 0.003 and 360 µm. Freshly cut chips of about 1 cm 3 were oven dried for 24 h at 100 • C and then analysed. Pore size measurements were taken on two fragments of each brick type. The specific surface area (SSA), open porosity to Hg (P oMIP ), and apparent and real density values (ρ aMIP and ρ rMIP ) were also calculated.
One of the most important aspects to consider when evaluating the durability of construction materials is their hygric behaviour, i.e., how these materials absorb and evaporate fluids over time. This is because decay processes are often affected by the flow of water inside the porous system [43]. Water absorption at atmospheric pressure (A b ) [44] and water absorption under vacuum (A f ) [45] were measured as follows: where M 0 is the mass of the dried sample, M l is the mass of the sample saturated with water at atmospheric pressure, and M S is the mass of the sample saturated with water under vacuum. The drying index (Di) was determined as follows [46]: where M t represents the decreasing water weight content as a function of time and t 0 and t f are, respectively, the start and end times for the test. A capillary test [47] was carried out to determine the capillarity coefficient (C c ) and the capillarity height (H c ) [45] as follows: where M t is the amount of water absorbed at time t, A is the surface of the sample in contact with the water, and h is the height of the water rise by capillarity at time t. These tests enabled us to ascertain the degree of pore interconnection (A x ) [48], the saturation coefficient (S), apparent (ρ a ) and real densities (ρ r ), and open porosity (P o ) [49]. These parameters were calculated as follows: where M 48h is the mass of the sample after 48 h immersion in water at atmospheric pressure and M H is the mass of the sample saturated with water under vacuum and weighed in water. Cubic samples were used for absorption-drying tests and prism-shaped samples for the capillarity test. Even though the UNE-EN 1925 norm [47] requires the use of cubic samples in the capillary test, elongated samples (twice as long) were used to better measure the capillarity height, given the rapid ascent of water in these materials. Hygric tests were performed under controlled thermo-hygrometric conditions (T = 18 • C and RH = 35%) using deionized water. Three samples from each brick group were analysed.
Of the various available techniques for determining the physical properties of bricks, ultrasounds and spectrophotometry are particularly attractive because they are not destructive. The compactness of the bricks was measured using a Control 58-E4800 ultrasonic pulse velocity tester with a resolution of 0.1 µm, 54 kHz transducers, and a contact surface of 3 cm in diameter under controlled temperature and hygrometric conditions (20 • C and relative humidity of 35%). An ultrasound eco-gel (Transonic Gel, Telic S.A., Barcelona, Spain) was used to obtain a good coupling between the transducers and the brick surfaces. The propagation velocity of the P waves was measured using the transmission method in accordance with the ASTM D2845 standard [50] on dry cubic samples in three orthogonal directions: V P1 is the velocity perpendicular to the compaction plane of the raw material in the wooden box and V P2 and V P3 are the velocities parallel to it. The structural (∆M) and relative anisotropies (∆m) were calculated as follows [51]: Three samples per brick type were tested. The colour of the bricks was determined by spectrophotometry in accordance with the UNE-EN 15886 standard [52] in order to quantify any changes produced by additives or firing temperatures. A Konica-Minolta CM-700d spectrophotometer (Tokyo, Japan) was used for this purpose. The working conditions were: A circular measurement area of 8 mm diameter, D65 illuminant, 10 • vision angle, SCI/SCE mode, and light radiation range between 400 and 700 nm. The CIELab system was used to calculate the lightness (L*, which varies from 0 to 100) and chromatic (a* and b*, which vary from −60 to +60) parameters. The spectrophotometer performs an auto averaging of two values on the measurement area. Six measurements were made for each brick type. In order to verify the possible effects of the additives on the colour, the colours of the samples made with and without additives were compared by measuring the colour difference. The colour difference, ∆E, was calculated as follows: where L * 1 , a * 1 , and b * 1 are the lightness and chromaticity values for the bricks without additives and L * 2 , a * 2 , and b * 2 are those for bricks with additives.

Raw Materials and Additives
The clayey earth from Jun is composed above all of silica and alumina and appreciable amounts of calcium and iron. The diatomite sludge is mainly composed of silica and small amounts of alumina, calcium, iron, and sodium. The loss on ignition (LOI) is 10% in the clayey earth and 15% in the diatomite sludge. The LOI values may be due to the likely presence of carbonates in the clayey earth and of organic matter in the sludge (Table 1). In mineralogical terms, the clayey earth is composed above all of quartz and phyllosilicates and smaller quantities of calcite, dolomite, plagioclases, and K-feldspars ( Figure 1a and Table 2). Gypsum was also identified. This sulphate is quite common in this soil and was formed in the Granada depression during the Turolian Age due to processes of chemical precipitation and evaporation of water from lacustrine environments [53]. The clay fraction is composed of illite, kaolinite, chlorite, smectites, and paragonite ( Figure 1b). After treatment with ethylene glycol, the displacement of the reflection (001) of the smectites shifted from 14 Å to 17 Å, approximately, was observed. This indicates that smectites can expand, leaving the reflection (001) of the chlorite (previously masked by the smectites) more visible. The presence of kaolinite was confirmed using dimethyl sulfoxide, by the displacement of the reflection (001) to 11.2 Å. Finally, heating to 550 • C destroyed both the smectites and the kaolinite, leaving the reflections of illite, chlorite, and paragonite visible (Figure 1b). Paragonite is an unusual clay mineral, which is produced by the transport of metamorphic minerals originally from Sierra Nevada [54].

Mineralogy of Fired Bricks
The mineralogy of the fired samples is similar to that resulting from high temperature metamorphic processes in which the high speed at which the minerals are heated causes extreme reactions, preventing stable phases from developing and favouring the presence of metastable phases [56]. The mineralogy of the diatomite sludge was also studied. It is mainly composed of cristobalite, as well as quartz, calcite, and albite ( Figure 1c). In fact, in chemical terms, this sludge was rich in amorphous silica while also having appreciable amounts of calcium (which led to the formation of calcite) and of aluminium and sodium (which enabled the crystallization of albite). As mentioned earlier, sodium carbonate is normally added during the diatomite calcination process as a melting agent. The curve followed by the diffractogram suggests a substantial amount of the amorphous phase.

Mineralogy of Fired Bricks
The mineralogy of the fired samples is similar to that resulting from high temperature metamorphic processes in which the high speed at which the minerals are heated causes extreme reactions, preventing stable phases from developing and favouring the presence of metastable phases [56].
The samples without additives fired at 800 • C are composed above all of quartz, K-feldspars, and plagioclases (Table 2 and Figure 2). Of the different carbonates identified in the raw material, dolomite is no longer present although there is a small reflection of the calcite at 3.03 Å. The types and concentrations of phyllosilicates also diminish, leaving only the dehydroxylated phases of illite and paragonite (Table 2 and Figure 2). At 950 • C, the calcite and paragonite also disappear and the concentrations of illite are reduced. New mineral phases begin to develop at this temperature as a result of the reaction between carbonates and silicates. These include gehlenite and plagioclase, which is now anorthitic due to its enrichment in calcium (Table 2 and Figure 2).
Anorthite may appear after the reaction of calcite with quartz and potassium feldspar as follows: calcite quartz K-feldspar anorthite At 1100 • C there is a considerable increase in anorthite, while the orthoclase is transformed into the most stable high temperature polymorph, sanidine. The diopside formed by the reaction between quartz and dolomite also appears: As the firing temperature increases, hematite starts developing due to the crystallization of the iron present in the raw material.
The concentration of the quartz diminishes above all at 1100 • C ( Table 2) as the edges of the grains of this mineral react with the carbonates to form calcium and magnesium silicates. The fired bricks normally contain anhydrite due to the presence of gypsum in the raw material.
As the firing temperature increases, hematite starts developing due to the crystallization of the iron present in the raw material.
The concentration of the quartz diminishes above all at 1100 °C (Table 2) as the edges of the grains of this mineral react with the carbonates to form calcium and magnesium silicates. The fired bricks normally contain anhydrite due to the presence of gypsum in the raw material. The samples that contain sodium chloride and diatomite sludge behave similarly to the samples without additives in the evolution of the mineral phases (Table 2 and Figures 3 and 4). The small differences are manifested in an increase in the concentration of K-feldspar at 950 and 1100 °C. As happens with the samples without additives, an increase in the concentration and the calcium enrichment of plagioclases can be observed as the firing temperature increases. The phyllosilicates undergo a similar process, with a progressive decline in the illite/muscovite, until they disappear completely between 950 and 1100 °C, while paragonite can only be detected in traces up to 800 °C. In the presence of sodium chloride, the decline in the amount of quartz begins earlier at 950 °C, when it is consumed in the reaction with the dolomite to form diopside ( Table 2).
As the firing temperature increases, there is a constant increase in the proportion of the amorphous phase compared to the crystalline phase in all the samples (with and without additives) (a/c, Table 2). This suggests that the samples are undergoing vitrification. It is important to bear in mind that the a/c ratio in JD samples is due not only to the vitrification of samples, but also to the amorphous phase (organic matter) present in the diatomite sludge (see background noise of XRD The samples that contain sodium chloride and diatomite sludge behave similarly to the samples without additives in the evolution of the mineral phases (Table 2 and Figures 3 and 4). The small differences are manifested in an increase in the concentration of K-feldspar at 950 and 1100 • C. As happens with the samples without additives, an increase in the concentration and the calcium enrichment of plagioclases can be observed as the firing temperature increases. The phyllosilicates undergo a similar process, with a progressive decline in the illite/muscovite, until they disappear completely between 950 and 1100 • C, while paragonite can only be detected in traces up to 800 • C.
In the presence of sodium chloride, the decline in the amount of quartz begins earlier at 950 • C, when it is consumed in the reaction with the dolomite to form diopside ( Table 2).
As the firing temperature increases, there is a constant increase in the proportion of the amorphous phase compared to the crystalline phase in all the samples (with and without additives) (a/c, Table 2). This suggests that the samples are undergoing vitrification. It is important to bear in mind that the a/c ratio in JD samples is due not only to the vitrification of samples, but also to the amorphous phase (organic matter) present in the diatomite sludge (see background noise of XRD pattern in Figure 1c and LOI value in Table 1). The a/c ratio values indicate that the samples with sodium chloride start vitrifying earlier than the others. Indeed, as NaCl is a low temperature melting agent, Stimmell et al. [35] observed that it reduced the temperature at which sintering and vitrification of the bricks began at 800 • C. The mineralogy also undergoes more profound changes, bringing forward the appearance of diopside to 950 • C and increasing its concentration at 1100 • C. Mullite only appears in this group of bricks (Table 2 and Figure 3). On this question, Cultrone et al. [36] observed greater development of newly formed phases in the presence of sodium chloride even when this salt had been added to the bricks in very low concentrations (0.5%). Mullite is formed at 1100 • C, replacing the illite/muscovite (epitaxial replacement according to Rodríguez Navarro et al. [57]) in line with the following reaction: illite quartz mullite K-feldspar 3KAl 2 (Si 3 Al)O 10 (OH) 2 + 2SiO 2 → Al 6 Si 2 O 13 + 3KAlSi 3 Another mineral phase that was identified exclusively in the samples with sodium chloride was molysite (FeCl 3 ), a chloride that is rarely found in bricks and is more typical of volcanic fumaroles. This seems to be related with the appearance of unusual yellowy-green halos on the surface of the samples. According to Stimmell et al. [35], who had observed similar marks on bricks made with seawater, the sodium chloride interacts with the clayey matrix, forming HCl. This acid then reacts with the hematite in the matrix, forming iron chloride. These halos are most likely to be found in the places with the highest Cl concentrations, i.e., where the halite crystals were present.
Another mineral phase that was identified exclusively in the samples with sodium chloride was molysite (FeCl3), a chloride that is rarely found in bricks and is more typical of volcanic fumaroles. This seems to be related with the appearance of unusual yellowy-green halos on the surface of the samples. According to Stimmell et al. [35], who had observed similar marks on bricks made with seawater, the sodium chloride interacts with the clayey matrix, forming HCl. This acid then reacts with the hematite in the matrix, forming iron chloride. These halos are most likely to be found in the places with the highest Cl concentrations, i.e., where the halite crystals were present. In the samples made with diatomite, traces of cristobalite can be observed at 800 and 950 °C with higher concentrations at 1100 °C (Table 2 and Figure 4). The increase in temperature produced the same silicates (gehlenite, anorthite, diopside) observed previously in other bricks. In the samples made with diatomite, traces of cristobalite can be observed at 800 and 950 • C with higher concentrations at 1100 • C (Table 2 and Figure 4). The increase in temperature produced the same silicates (gehlenite, anorthite, diopside) observed previously in other bricks. In the samples made with diatomite, traces of cristobalite can be observed at 800 and 950 °C with higher concentrations at 1100 °C (Table 2 and Figure 4). The increase in temperature produced the same silicates (gehlenite, anorthite, diopside) observed previously in other bricks.

Porous System
The shape of the porometric curves and the height of the main peak are similar to those found in modern brick production [58]. The MIP test showed that the bricks had a unimodal pore size distribution with the maximum peak of pores at around 1 µm ( Figure 5). This peak shifts towards larger pores as the firing temperature increases, although the said shift is almost imperceptible in the case of bricks without additives, as can be seen in the porometric curves (from 0.4 µm at 800 • C to 0.5 µm at 950 and 1100 • C). The shift is more pronounced in samples with added NaCl (from 0.5 µm at 800 • C to 1 µm at 1100 • C) and above all, in those with added diatomite sludge (from 0.5 µm at 800 • C to 1.5 at 1100 • C). This means that both additives cause alterations in the porous system. The greatest alterations take place with the addition of diatomite sludge, and above all at the highest firing temperature. However, as can be seen in Figure 5, the first six curves (J and JS) are similar, which means that the differences described above are probably not very significant.
While the largest pores generally increase in number from 800 • C to 1100 • C, the smallest pores fall in number due to the decline in specific surface area (see Table 3 for SSA and pore range of less than 0.1 µm). This happens constantly and in all types of brick. The reduction of SSA is more evident in JS bricks because, as commented earlier, sodium chloride is a low temperature melting agent and causes alterations in the texture of the bricks at 800 • C. In fact, if one compares the results in Table 3, at 800 • C, the specific surface area of JS bricks is slightly less than 3 m 2 /g, while in the other two groups, SSA is well above 5 m 2 /g. This indicates that the porous system of the bricks undergoes greater changes in the presence of sodium chloride. These differences are less pronounced at 950 • C when the other two brick groups also begin to vitrify. However, even at 1100 • C, the JS bricks still show the lowest SSA values (Table 3). These results are important because a lower specific surface area means that the material is less exposed to pollutants and other decay agents [59]. The highest open porosity values were obtained in JD bricks, mainly because of the combustion and subsequent loss of the organic matter present in the diatomite sludge. These bricks reached open porosity values of 45-46%, substantially higher than those for J and JS bricks (P oMIP , Table 3). Open porosity is a crucial parameter, together with specific surface area and pore size distribution, for assessing the durability of porous materials [58,60]. There is normally an inverse relation between porosity and durability [61]. These results suggest that the addition of 10% in weight of diatomite sludge can have negative effects on the durability of bricks. However, in general, all the bricks are very porous because the method of production used in this research (handmade) imitates traditional methods rather than industrial processes (i.e., pressed or extruded) in which porosity values of around 20% are usual [62]. Real density values (ρ rMIP , Table 3) were similar in all the samples and were characteristic of the mineralogy of a clayey earth rich in quartz and phyllosilicates. The lowest values were measured in samples with diatomite because of the low density of cristobalite (2.32 g/cm 3 ), a mineral phase that is only found in JD bricks ( Table 2). The lowest apparent density value (ρ aMIP , Table 3) was again obtained by JD bricks, which is normal since they were the most porous.

Porous System
The shape of the porometric curves and the height of the main peak are similar to those found in modern brick production [58]. The MIP test showed that the bricks had a unimodal pore size distribution with the maximum peak of pores at around 1 µm ( Figure 5). This peak shifts towards larger pores as the firing temperature increases, although the said shift is almost imperceptible in the case of bricks without additives, as can be seen in the porometric curves (from 0.4 μm at 800 °C to 0.5 μm at 950 and 1100 °C). The shift is more pronounced in samples with added NaCl (from 0.5 μm at 800 °C to 1 μm at 1100 °C) and above all, in those with added diatomite sludge (from 0.5 μm at 800 °C to 1.5 at 1100 °C). This means that both additives cause alterations in the porous system. The greatest alterations take place with the addition of diatomite sludge, and above all at the highest firing temperature. However, as can be seen in Figure 5, the first six curves (J and JS) are similar, which means that the differences described above are probably not very significant. Figure 5. Pore size distribution curves for bricks with and without additives fired at 800, 950, and 1100 °C. Pore radius (in μm) versus log (DV/Dr) (in mL/g). The radius at the maximum peak is indicated.
While the largest pores generally increase in number from 800 °C to 1100 °C, the smallest pores fall in number due to the decline in specific surface area (see Table 3 for SSA and pore range of less than 0.1 µm). This happens constantly and in all types of brick. The reduction of SSA is more evident in JS bricks because, as commented earlier, sodium chloride is a low temperature melting agent and causes alterations in the texture of the bricks at 800 °C. In fact, if one compares the results in Table 3, at 800 °C, the specific surface area of JS bricks is slightly less than 3 m 2 /g, while in the other two groups, SSA is well above 5 m 2 /g. This indicates that the porous system of the bricks undergoes greater changes in the presence of sodium chloride. These differences are less pronounced at 950 °C when the other two brick groups also begin to vitrify. However, even at 1100 °C, the JS bricks still show the lowest SSA values (Table 3). These results are important because a lower specific surface area means that the material is less exposed to pollutants and other decay agents [59]. The highest open porosity values were obtained in JD bricks, mainly because of the combustion and subsequent loss of the organic matter present in the diatomite sludge. These bricks reached open porosity values of 45-46%, substantially higher than those for J and JS bricks (PoMIP, Table 3). Open porosity is a crucial parameter, together with specific surface area and pore size distribution, for assessing the Figure 5. Pore size distribution curves for bricks with and without additives fired at 800, 950, and 1100 • C. Pore radius (in µm) versus log (DV/Dr) (in mL/g). The radius at the maximum peak is indicated. Table 3. Results of MIP test on brick samples with and without additives fired at 800, 950, and 1100 • C. SSA = specific surface area (m 2 /g); ρ aMIP = apparent density (g·cm −3 ); ρ rMIP = real density (g·cm −3 ); P oMIP = open porosity (%); relative porosity distribution (%) according to pore size ranges (<0.1, 0.1-1, 1-10, >10 µm).

SSA
ρ aMIP ρ rMIP P oMIP <0.1 0. The hygric tests show how some parameters obtained with this technique are influenced by brick composition, while others depend more on the firing temperature. The water absorption curves, for example, show two clearly defined groups (Figure 6), already identified by MIP: One group contains the bricks made with diatomite sludge (JD) and the other contains the bricks made without additives (J) and with sodium chloride (JS). J and JS absorb a similar amount of water, substantially less than JD. One hour after starting the test, J and JS had reached a 20% absorption level, while JD was already between 25% and 30%, depending on the firing temperature of the bricks (Figure 6).

JS800
2.98 1. 45  The hygric tests show how some parameters obtained with this technique are influenced by brick composition, while others depend more on the firing temperature. The water absorption curves, for example, show two clearly defined groups (Figure 6), already identified by MIP: One group contains the bricks made with diatomite sludge (JD) and the other contains the bricks made without additives (J) and with sodium chloride (JS). J and JS absorb a similar amount of water, substantially less than JD. One hour after starting the test, J and JS had reached a 20% absorption level, while JD was already between 25% and 30%, depending on the firing temperature of the bricks (Figure 6).  At the end of the water absorption test at atmospheric pressure, the group of bricks that had absorbed the least water was J (A b is around 26%, Figure 7a), closely followed by JS (A b ranges approximately between 26% and 27.8%), while JD showed the highest A b values (between 34% and 36.7%). The differences between the three brick groups observed after the water absorption test at atmospheric pressure remained after the water absorption test under vacuum (A f , Figure 7b) and clearly indicate that the addition of diatomite alters the hygric behaviour of bricks.
At the end of the water absorption test at atmospheric pressure, the group of bricks that had absorbed the least water was J (Ab is around 26%, Figure 7a), closely followed by JS (Ab ranges approximately between 26% and 27.8%), while JD showed the highest Ab values (between 34% and 36.7%). The differences between the three brick groups observed after the water absorption test at atmospheric pressure remained after the water absorption test under vacuum (Af, Figure 7b) and clearly indicate that the addition of diatomite alters the hygric behaviour of bricks.  The high water absorption capacity of the JD brick group is potentially dangerous for its durability, since the water that flows inside a porous material may contain soluble salts that can crystallize in capillaries, generating fissures or fractures that cause irreparable damage [63,64]. The saturation coefficient (S, Figure 7c) falls as the firing temperature increases and the degree of interconnection between the pores worsens (A x , Figure 7d). The latter parameter is related to the tortuous nature of the porous system. As the difference between the values for water absorption at atmospheric pressure (A b ) and water absorption under vacuum (A f ) increases, so A x increases; in other words, it becomes more difficult for water to circulate inside the material [48]. These results indicate that the increase in temperature helps modify the porous system with the development of the vitreous phase and as a result reduces communication between the pores. This hypothesis confirms the MIP data that showed a shift in the maximum pore size peak towards higher sizes and the reduction of specific surface area values (see Figure 5 and Table 3). On this question, Cultrone et al. [9] demonstrated the existence of a direct relationship between the development of vitrification in bricks and the reduction of pore interconnectivity. The XRD results also highlighted an increase in the amount of the amorphous phase in line with the increase in the firing temperature in all the groups of bricks (a/c, Table 2). The drying index (Di) is another key hygric parameter because porous materials are less durable when water is retained inside them over long periods [65]. Di is quite similar in all the samples with only minimal variations according to the firing temperature (Figure 7e). The only exception is J800, which is the slowest sample to dry (see the slope of the drying curve in Figure 6 and the Di value in Figure 7e). Note that J800 is the sample that retained the most water at the end of the drying test followed by JD800 ( Figure 6). This is probably because these samples had the largest amounts of small pores of less than 0.1 µm (Table 3), the last ones to lose water, which may cause local stresses and cracking [66]. The capillarity test shows an analogous trend to that of water absorption. In fact, the two parameters obtained from this test, the capillary coefficient and the capillary rise height (C c and H c , respectively, Figure 7i,l), clearly indicate that in all the bricks, water rises more quickly as the firing temperature increases. The open porosity (P o , Figure 7f) values again show the influence of brick composition in the porous system, especially in the bricks made with diatomite sludge. It is important to remember that this additive has intrinsic porosity that may have increased the capacity of JD bricks to absorb more water and reach higher porosity values than the other bricks. These values are similar to those obtained in the MIP test. The ρ r value determined by the hygric test is always higher than that calculated using MIP and is closer to the density of the mineral phases identified by XRD (Table 2). In fact, ρ r ranges between 2.5 and 2.6 g/cm 3 , close to the density of quartz (2.66 g/cm 3 ) and K-feldspar (2.56 g/cm 3 ). The lowest ρ r value was again measured in JD samples because of the presence of the low-density mineral, cristobalite, as mentioned earlier in the description of the MIP results. J and JD bricks show a similar trend with lower ρr values at 800 • C compared to those at 950 • C and 1100 • C. This is probably due to the Ca-enrichment of plagioclases (ρ r albite is 2.61 g/cm 3 , while that of anorthite is 2.76 g/cm 3 ). Instead, JS shows a gradual reduction in ρ r from 800 • C to 1100 • C. This can be attributed in part to the reduction in anhydrite (ρ r is 2.95 g/cm 3 ), which disappears at 1100 • C. Moreover, JS registers a gradual decrease in hematite (5.25 g/cm 3 ) and molysite (2.91 g/cm 3 ), while in J and JD, there is an increase in hematite after 800 • C. In general, the small differences observed between the MIP and hygric tests are due to the limitations of each experimental technique, which use different liquids and intrusion pressures. In addition, MIP only considers pores with cylindrical geometry and can erroneously interpret the size of "ink-bottle-shaped" pores [67].

Ultrasounds
In all the samples, the lowest velocity is always V P1 due to the position of the phyllosilicates perpendicular to the direction of propagation of the ultrasound waves ( Table 4). The phyllosilicates acquired this orientation during the preparation of the test samples due to the pressure exerted when the clayey earth is placed inside the wooden mould. The difference between V P1 and the other two velocities falls as the firing temperature increases, above all at 1100 • C, when the samples are more vitrified and the lamina shape of the crystals is partially lost, so reducing the anisotropy of the pieces (∆M, Table 4). The comparison of the P-wave velocities in the three orthogonal directions therefore allowed us to verify and confirm the vitrification of bricks in line with the increase in the firing temperature, as had already been deduced by the XRD, MIP, and hygric tests. V P2 and V P3 obtained similar values, a fact confirmed by their low relative anisotropy values (∆m, Table 4). Exceptions are the samples made with diatomite sludge, which suggests that this additive helps to alter the texture of the bricks (i.e., it modifies the porous system and the orientation of the phyllosilicates). In fact, the total anisotropy of JD samples shows no clear trend, indicating that this additive is randomly oriented inside the bricks and the firing temperature does not help to homogenize the texture (∆M, Table 4). For their part, the results for the J and JS groups show that there is an inverse relationship between the anisotropy and the vitrification of the samples, a fact that could be indirectly deduced by the a/c ratio (XRD, Table 2). In fact, JS has a lower ∆M value than J when fired at 800 • C because of the high vitrification of bricks with added sodium chloride. JS maintains lower anisotropy at 950 • C (and higher vitrification), but at 1100 • C, the higher a/c value measured in J samples (Table 2) favours a lower degree of anisotropy than JS. The bricks without additives showed the highest velocity values, reaching 2400 m/s at 1100 • C, while in the bricks made with added salt or diatomite sludge, the maximum velocity was 2200 m/s ( Table 4). This velocity is reached in bricks made with salt at 800 • C, another demonstration of the effect of salt as a low-temperature melting agent. When one looks at the mean velocity of P waves (V p , Table 4), an increase in velocity or, in other words, in compactness can be observed as the firing temperature rises. This increase takes place more gradually when no additives are used. JS and JD follow this trend particularly between 950 and 1100 • C when the samples are vitrifying (and NaCl does not have the same influence as at 800 • C). The V P values are in line with those measured in other bricks fired at similar temperatures [68], mostly when no additives are present. Brick compactness (and mechanical resistance) can therefore vary considerably depending on the type of additive and the percentage added to the clayey material [25]. No comparative data were found in the literature for bricks made with sodium chloride, and although Galán Arboledas et al. [33] had researched the use of diatomite sludge, a comparison with their findings was complicated by the fact that they measured the bending strength instead of the ultrasound velocities and because they followed a different brick manufacturing procedure (extrusion instead of manual). Table 4. Speed of ultrasonic wave propagation (in m/s), mean velocity (V p , in m/s), structural (∆M), and relative (∆m) anisotropies (in %) of bricks with and without additives fired at 800, 950, and 1100 • C. V P1 , V P2 , and V P3 are the ultrasound velocities measured in three orthogonal directions.

Colour
Colour is a very important parameter, given that a change in the colour of the bricks can have a negative effect on the image of the building being constructed or restored. Slight variations in colour are not easily perceived by the human eye, which means that the colour measurement techniques are important for ensuring reliable results for evaluation purposes [69]. There are no important colour differences between the unfired samples without additives and those made with added sodium chloride and only very slight differences with those made with diatomite sludge ( Table 5). As the salt is dissolved in the kneading water, it does not alter the colour of the pieces when they dry. The diatomite, however, is added in the form of solid particles in a concentration of 10% in weight compared to the raw material. The result is that the unfired samples made with this additive are slightly less luminous and have a slightly higher chromaticity. After firing, a reduction in the chromaticity (a* and b*) and lightness (L*) parameters was observed as the firing temperature increased (Table 5). This means that the bricks become darker and the colour becomes less chromatic (C* falls). The reddish colour acquired by the bricks is due to the formation of hematite, a phase identified by XRD (Table 2), which, although it is present in low concentrations, is responsible for the red colour of the bricks (Figure 8). The most evident changes took place in the samples made with sodium chloride, which turned lighter (higher L* values and lower a* and b* values compared to the other bricks), with yellow-green halos developing on the surface due to the crystallization of molysite. This fluoride traps the iron in the matrix so preventing the formation of hematite and making the bricks lighter in colour [35]. The halos are of varying size, no larger than 2-3 mm, and have irregularly pigmented the surface of the bricks (Figure 8). The samples fired at 1100 • C show lower chromaticity values compared to the other bricks and are darker (lower L* value), increasing the contrast with the stains. The difference in colour (∆E, Table 5) compared to the samples without additives is higher in the bricks with added salt, in which ∆E values of over 5 were obtained in the samples fired at 950 • C and 1100 • C, a figure above which the changes in colour start to be appreciable by the human eye [70]. Table 5. Lightness (L*), chromatic coordinates (a* and b*), chroma (C*), hue angle (h( • )), and colour difference (∆E) values for unfired and fired ceramic samples with and without additives.

Conclusions
In this paper, bricks made with a raw material from Jun (Granada, Spain) and fired at 800 °C, 950 °C, and 1100 °C were characterized in mineralogical and physical terms and compared to others to which sodium chloride and calcined diatomite sludge were added, so as to find out whether these additives improved the engineering properties of the bricks. The raw material underwent mineralogical changes during the brick firing process. This was due to the decomposition of the carbonates and phyllosilicates and their interaction with quartz and feldspars to form new silicates (gehlenite, anorthite, and diopside). The inclusion of additives led to further mineralogical changes, such as the appearance of molysite and mullite when sodium chloride was added, and of cristobalite when calcined diatomite sludge was used.
All the brick samples vitrified as the firing temperature increased (the amount of amorphous phase increased compared to the crystalline matter). This gave rise to changes in the porous system, such as an increase in the size of the pores, at the same time as reducing the degree of interconnection between them. This is presumably due to the development of a "cellular structure" [71] in which the pores coalesce, increasing in size due to the fusion of the clayey particles. The use of diatomite sludge gave rise to bricks, which were more porous and had a higher water absorption capacity than the other brick groups studied in this paper. This was also due to the diatom particles, which have intrinsic porosity.
Greater vitrification also led to an increase in the compactness of the bricks and was more accentuated when no additives were used. It must be emphasized that the bricks with added sodium chloride were already highly compact at 800 °C, as salt allows melting to take place at lower

Conclusions
In this paper, bricks made with a raw material from Jun (Granada, Spain) and fired at 800 • C, 950 • C, and 1100 • C were characterized in mineralogical and physical terms and compared to others to which sodium chloride and calcined diatomite sludge were added, so as to find out whether these additives improved the engineering properties of the bricks. The raw material underwent mineralogical changes during the brick firing process. This was due to the decomposition of the carbonates and phyllosilicates and their interaction with quartz and feldspars to form new silicates (gehlenite, anorthite, and diopside). The inclusion of additives led to further mineralogical changes, such as the appearance of molysite and mullite when sodium chloride was added, and of cristobalite when calcined diatomite sludge was used.
All the brick samples vitrified as the firing temperature increased (the amount of amorphous phase increased compared to the crystalline matter). This gave rise to changes in the porous system, such as an increase in the size of the pores, at the same time as reducing the degree of interconnection between them. This is presumably due to the development of a "cellular structure" [71] in which the pores coalesce, increasing in size due to the fusion of the clayey particles. The use of diatomite sludge gave rise to bricks, which were more porous and had a higher water absorption capacity than the other brick groups studied in this paper. This was also due to the diatom particles, which have intrinsic porosity.
Greater vitrification also led to an increase in the compactness of the bricks and was more accentuated when no additives were used. It must be emphasized that the bricks with added sodium chloride were already highly compact at 800 • C, as salt allows melting to take place at lower temperatures. In fact, these were the bricks with the largest amounts of the amorphous phase at this temperature.
In terms of their aesthetic appearance, the bricks got gradually darker as the firing temperature increased, above all at 1100 • C and even more so when the diatomite sludge was added. The addition of sodium chloride produced yellower bricks because part of the iron responsible for the red pigmentation of the bricks is consumed in the formation of molysite and cannot crystallize as hematite. The yellow colour was not uniformly distributed on the surface of the bricks and appeared in the form of spots or stains. These stains may indicate a higher concentration of molysite crystals where the presence of halite was highest.
At a general level, the best physical characteristics were found in the samples fired at the highest temperature. However, if bricks can be fired at lower temperatures, this reduces production costs considerably as less energy is consumed. In this case, the bricks made with added sodium chloride fired at 800 • C could be a valid alternative to those made without additives and fired at 950 and 1100 • C. This is because partially vitrified and, therefore, more compact bricks can be obtained at 800 • C thanks to the fact that salt acts as a melting agent. Moreover, at 800 • C, the stains on the surface of the bricks are practically invisible compared to those fired at 950 and 1100 • C, which makes them aesthetically acceptable in situations in which a uniform appearance is required. There is some potential in the use of diatomite in brick-making, as bricks made with this additive have a similar degree of compactness to other bricks. The most obvious difference is their higher porosity. Its possible use as an additive in bricks for thermal insulation should be investigated. The durability of these bricks could be analysed in future research in order to obtain a more complete picture of their technical quality. This study has provided new data on the mineralogical and physical changes that take place in bricks due to the addition of sodium chloride and calcined diatomite sludge. These data will be useful for the brick industry both in new constructions and in the conservation of historic buildings.