Use of Spent Zeolite Sorbents for the Preparation of Lightweight Aggregates Differing in Microstructure

Lightweight aggregates (LWAs) made by sintering beidellitic clay deposits at high temperatures, with and without the addition of spent zeolitic sorbents (clinoptilolitic tuff and Na-P1 made from fly ash) containing diesel oil, were investigated. Mineral composition of the aggregates determined by X-ray diffraction was highly uniformized in respect of the initial composition of the substrates. The microstructure of the LWAs, which were studied with a combination of mercury porosimetry, microtomography, nitrogen adsorption/desorption isotherms and scanning electron microscopy, was markedly modified by the spent zeolites, which diminished bulk densities, increased porosities and pore radii. The addition of zeolites decreased water absorption and the compressive strength of the LWAs. The spent Na-P1 had a greater effect on the LWAs’ structure than the clinoptilolite.


Introduction
Lightweight aggregates (LWAs) are building materials, produced from different minerals (including ordinary soil clay, perlite, vermiculite, and natural and synthetic zeolites) by rapid sintering/heating at high temperatures up to 1300 • C [1].To achieve expanded material appropriately, two conditions are necessary: the presence of substances that release gases at high temperature, and a plastic phase with adequate viscosity, which is able to trap the released gases [2].The expanded clay aggregates are non-flammable and highly resistant to chemical, biological and weather conditions.Their highly porous structure is represented mainly by closed pores surrounded by glassy coatings, which are formed during the thermal transformation of clay minerals.As a consequence, LWAs have relatively low particle and bulk densities, low thermal conductivity and sound dampening characteristics [3][4][5][6][7][8], thereby allowing them to have broad applications in the construction industry, geotechnics, gardening and agriculture [4,5,[9][10][11][12][13][14][15][16].
Much effort has been recently invested to reuse different kinds of waste materials, in order to avoid their disposal in landfills and paying additional environmental taxes, as well as to reduce production costs [17,18].Many waste materials, such as combustion ashes [19], waste glass [15], sewage or industrial sludge [20][21][22][23], incinerator bottom ash [24], mining residues, heavy metal sludge, washing aggregate sludge [4], polishing residue, lignite coal fly ash [25,26], spent adsorbents [27,28] and contaminated mine soil [29], have been used as additives for the production of LWAs.Some of The compressive strength Ca, being the force necessary to pass a piston for a certain depth into a cylinder filled with the studied material, was determined according to UNE-EN 13055-1/AC:2004-10-22 [44].
The freezing resistance F of the aggregates, which express the percentage loss of the mass of the aggregate soaked in water and subjected to 10 cycles of freezing-thawing (−17.5 and 20 • C, respectively) was determined by the UNE-EN 12697-2:1999 standard [45].
The laser diffraction method was applied in order to measure the particle size distribution of the initial materials subjected to 300 W of ultrasonication for 1 min using a Mastersizer 2000 with a Hydro G dispersion unit provided by Malvern UK.When obscuration after adding the sample to the measuring system exceeded 10%-20%, it was lowered by using the procedure that ensures there is no discrimination of any fraction [46].For the solid phase, the refraction index was taken as 1.52 and the absorption index as 0.1; for water, the refraction index was taken as 1.33.

Structural Characteristics
X-ray computational microtomography was applied for 3D scanning of the studied LWAs using a Nanotom S device (General Electrics, Frankfurt, Germany).The X-ray source with a molybdenum target, operated at a cathode current of 230 µA and a 60 kV voltage was used for X-ray generation.The scanning process consisted of two stages: an initial pre-scan and a main measurement scan.Prior to the final measurement scan, each sample was subjected to a short 40 min pre-scan in order to heat it up and reach thermal stability, which was maintained further during the main scan lasting 150 min.The scanned specimens were dry, so the only effect of heating by X-rays on the measurement could have been caused by the thermal elongation of the sample holder.The pre-scan eliminated this problem.During the main scan, 2400 2D cross-sectional images were acquired with a spatial resolution (voxel size) of about 0.0063 mm and then used for 3D porous space reconstruction.The resulting 3D 16 bit grey-level images represent the spatial structure of specimens.Image analysis techniques were used for further processing.Initially, the bit depth of images was reduced from 16 to 8 bit.After that, a 3D median filter including a uniform kernel with a diameter equal to 3 px was used for noise reduction.The next step was the thresholding procedure, which utilized the Otsu algorithm.Threshold images had a 1 bit color depth with black areas representing pores.These preprocessing steps were performed using ImageJ software.For further analysis, Avizo software was used.The 3D watershed-based segmentation algorithm and then the labelling algorithm were used to separate the connected pores into individual ones.Geometrical characteristics of the pores including equivalent diameter (a diameter of the sphere with the same volume as a pore), volume, surface and fractal dimension of pores according to the maximal ball method [47] were calculated from three 3D images.
Mercury intrusion porosimetry (MIP) tests were performed for pressures ranging from about 0.1 to 200 MPa (pore radii ranged from about 10.0 to 3.8 × 10 −3 µm).The intrusion volumes were measured at stepwise increasing pressures, which allowed for equilibration at each pressure step.The maximum deviations between the mercury intrusion volumes were no higher than 6.9% and occurred mainly at low pressures (largest pores).The volume of mercury V [m 3 •kg -1 ] intruded at a given pressure P [Pa] gave the pore volume that can be accessed.The intrusion pressure was translated into an equivalent pore radius R [m] following the Washburn equation: where σ m is the mercury surface tension (0.485 N•m −1 ), α m is the mercury/solid contact angle (taken as 141.3 • for all studied materials) and A is a shape factor (equal to 2 for the assumed capillary pores).The total range for the pore radii in the mercury intrusion curve was divided into sections in steps of 0.1 log(R).
Knowing the dependence of V vs. R, a normalized pore size distribution, χ(R), was calculated and expressed in the logarithmic scale [48]: By knowing χ(R), the average pore radius, R av , was calculated from: If a range of pore sizes, wherein the pore volume depends on a power of the pore radius, could be found, this was interpreted in terms of pore surface fractal behavior.In this case, the dependence of log(dV/dR) against logR was plotted and, from the slope of its linear part, the fractal dimension of •pore surface D was derived according to Pachepsky et al. [49]: To define the linear range of fractality, the Yokoya et al. [50] procedure was applied.According to this procedure the measure of the linearity L for the set of the points in a x-y plane is: where σ xx , σ yy and σ xy are the variances of x-coordinates, y-coordinates and the covariance between x and y coordinate sets, respectively.The L value falls between 0 (for uncorrelated and random points) and 1 (for points on a straight line).To separate out the linearity range, the value of L is computed for the first three points, then for the first four, five and so on until the value of L increases.The end of the linearity range is within the points after which the value of L begins to decrease.From the estimated linearity range, the two first and/or two last points were rejected if this caused an increase in the linear regression coefficient between the considered data.
The apparent solid phase skeletal densities of the samples SSD (which are lower than true solid phase densities due to the residence of the finest pores in the solid phase that are not filled by mercury at its highest pressure) and the total surface of MIP available pores S (MIP) were calculated by the porosimetric data analysis program provided by the equipment manufacturer.
Nitrogen adsorption isotherms were measured at the temperature of liquid nitrogen using ASAP 2020MP manufactured by Micromeritics (Norcross, GA, USA).
The scanning electron microscope (SEM) images of the tested materials were taken using an FEI Quanta 250 FEG microscope equipped with the energy dispersion scattering EDS-EDAX system for chemical composition analysis(FEI, Hilsboro, OR, USA).From three SEM images, the sizes of the finest pores were estimated using the Aphelion 4.0.10 image analysis package and the Vogel and Roth procedure [51].

Mineralogical and Physical Properties
Figure 1 illustrates the particle size distribution of the initial materials.The clay material is composed of the finest particles with an average diameter of 58 µm, followed by NaP1 with a similar average diameter of 52 µm.The largest particles occur in natural clinoptilolitic tuff, for which the average diameter is 178 µm.
The main mineral components of the raw clay material were around 51% of beidellite (d hkl 15 2).The main mineral component of the clinoptilolitic tuff was clinoptilolite as recognized by d hkl = 8.95, 7.94, 3.96 and 3.90 Å XRD reflections.The presence of the Na-P1 phase in the product of fly ash conversion was confirmed by d hkl = 7.10, 5.01, 4.10 and 3.18 Å.
The scanning electron microscope (SEM) images of the tested materials were taken using an FEI Quanta 250 FEG microscope equipped with the energy dispersion scattering EDS-EDAX system for chemical composition analysis(FEI, Hilsboro, OR, USA).From three SEM images, the sizes of the finest pores were estimated using the Aphelion 4.0.10 image analysis package and the Vogel and Roth procedure [51].

Mineralogical and Physical Properties
Figure 1 illustrates the particle size distribution of the initial materials.
(a) (b)  The clay material is composed of the finest particles with an average diameter of 58 µm, followed by NaP1 with a similar average diameter of 52 µm.The largest particles occur in natural clinoptilolitic tuff, for which the average diameter is 178 µm.
An extremely high degree of uniformization of the mineral composition of the sintered substrates is observed in the XRD spectra of the LWAs produced from different materials.The main mineral components of all LWAs are mullite (d hkl 3.39, 5.41, 3.42 and 2.21 Å) and quartz (3.34, 4.25 and 1.81 Å).The presence of mullite is an effect of the melting of the original clay minerals (beidellite, illite, Minerals 2017, 7, 25 6 of 14 kaolinite) [52].One can observe that iron hydroxides were transformed into well-defined hematite (d hkl 2.70 and 2.51 Å), while the feldspars remained intact.Apart from the defined mineral phases, a significant contribution of an amorphous glassy phase can be distinguished by the rise in the background line within the range 15 • -30 • 2θ, which was the highest for LWAs admixed with Na-P1.

Figure 2. X-ray diffraction (XRD) patterns of the original clay deposit (black), clinoptilolitic tuff (green) and zeolite Na-P1 (red).
The XRD patterns of LWAs obtained from the clay and a mixture of clay and spent zeolitic sorbents are shown in Figure 3.The physical properties of the studied LWAs are shown in Table 1.The solid phase density decreases slightly due to the spent sorbents addition, whereas the particle density decreases markedly, indicating the effect of the spent zeolites on the aggregates' expansion.Despite the decrease in particle density suggesting much larger porosity, the water absorption decreases for all LWAs produced with spent sorbents admixtures.Taking into account the similar mineral composition of the LWAs, resulting most probably in similar surface properties (wettability/hydrophobicity) of the aggregates material, the above differences may be connected to differences in the pore system, particularly in the amount of closed pores unavailable for water.According to Hung and Hwang [53], a particle with isolated pores or a vitrified surface tends to absorb less water than one having connected or open pores.Water absorption of the studied LWAs containing the spent sorbents is markedly lower than for several commercial ones as stated by their manufacturers, such as Lytag (17.55%),Arlita (20%) and Leca (30.3%).Lower water absorption may have technological advantages for building purposes.The frost resistance test showed that all LWAs lost not more than 1% mass after freezing that indicates their high resilience against variations in climate conditions.The aggregate grains did not show any occurrence of cracks after the test, probably because water penetrating the grains has not filled their whole pore space, meaning that it did not cause any visible aggregate damage after freezing.The compressive strength of the studied LWAs significantly decreased after the addition of spent sorbents.However, their mechanical resistance is still higher than that of some commercially available LWAs, such as Lytag (0.43 MPa), Leca (0.09 MPa) [54] or Leca Weber (0.75 MPa) (Saint-Gobain Construction Products Minerals 2017, 7, 25 7 of 14 Poland) and markedly higher than 0.44 MPa, which is the internationally accepted standard for solid waste materials used for land levelling [55].

Structure Characteristics
Exemplary microtomography cross sections of the studied LWAs are presented in Figure 4, wherein quite different porous structures of the studied materials are seen.On the external surfaces of both aggregates containing the diesel oil, a well-developed vitrified layer is seen.However, Gonzáles-Corrochano et al. [5] did not observe the formation of such a layer in LWAs manufactured with used motor oil.The visual analysis of the scans reveals that the LWAs have thick, dense areas, which extend throughout the whole CLAY aggregate, while being limited to the external layer for CLIN and NAP1, for which it is the thinnest.Most probably the thickness of this layer depends on the oil content.More oil evolves more gases during sintering and the resulted more porous structure reduces the number and increases the distance of connections between the molten solid thus its condensation is limited to smaller external space.It is also possible that more time is needed to decompose more oil what provides less time for solid condensation.
Calculated from 3D scans, the pore volume vs. pore radius dependencies and pore size distribution functions are presented in Figure 5.As it is seen in Figure 5a,b, the LWA with used Na-P1 develops the largest pores and the largest pore volumes, particularly in the range of large pores.The volume of small pores is similar for NAP1 and CLAY, whereas CLIN possesses the largest volume of these pores.Pore size distribution functions (Figure 5b) show that CLIN aggregates contain the highest amount of pores lower than 0.1 On the external surfaces of both aggregates containing the diesel oil, a well-developed vitrified layer is seen.However, Gonzáles-Corrochano et al. [5] did not observe the formation of such a layer in LWAs manufactured with used motor oil.The visual analysis of the scans reveals that the LWAs have thick, dense areas, which extend throughout the whole CLAY aggregate, while being limited to the external layer for CLIN and NAP1, for which it is the thinnest.Most probably the thickness of this layer depends on the oil content.More oil evolves more gases during sintering and the resulted more porous structure reduces the number and increases the distance of connections between the molten solid thus its condensation is limited to smaller external space.It is also possible that more time is needed to decompose more oil what provides less time for solid condensation.
Calculated from 3D scans, the pore volume vs. pore radius dependencies and pore size distribution functions are presented in Figure 5.  On the external surfaces of both aggregates containing the diesel oil, a well-developed vitrified layer is seen.However, Gonzáles-Corrochano et al. [5] did not observe the formation of such a layer in LWAs manufactured with used motor oil.The visual analysis of the scans reveals that the LWAs have thick, dense areas, which extend throughout the whole CLAY aggregate, while being limited to the external layer for CLIN and NAP1, for which it is the thinnest.Most probably the thickness of this layer depends on the oil content.More oil evolves more gases during sintering and the resulted more porous structure reduces the number and increases the distance of connections between the molten solid thus its condensation is limited to smaller external space.It is also possible that more time is needed to decompose more oil what provides less time for solid condensation.
Calculated from 3D scans, the pore volume vs. pore radius dependencies and pore size distribution functions are presented in Figure 5.As it is seen in Figure 5a,b, the LWA with used Na-P1 develops the largest pores and the largest pore volumes, particularly in the range of large pores.The volume of small pores is similar for NAP1 and CLAY, whereas CLIN possesses the largest volume of these pores.Pore size distribution functions (Figure 5b) show that CLIN aggregates contain the highest amount of pores lower than 0.1 mm, whereas the lowest amount of these pores is found in NAP1 aggregates.As it is seen in Figure 5a,b, the LWA with used Na-P1 develops the largest pores and the largest pore volumes, particularly in the range of large pores.The volume of small pores is similar for NAP1 and CLAY, whereas CLIN possesses the largest volume of these pores.Pore size distribution functions (Figure 5b) show that CLIN aggregates contain the highest amount of pores lower than 0.1 mm, whereas the lowest amount of these pores is found in NAP1 aggregates.
MIP curves relating the intruded mercury (pore) volume to the logarithm of the pore radius and the normalized pore size distribution functions for the studied materials are presented in Figure 6.It is worth noting that the mercury extrusion branches (data not shown) were, in all cases, practically parallel to the log(R)-axis, indicating that practically all the mercury is accumulated in the pore voids and that the amount of the necks (channels) connecting these voids is negligible.
Minerals 2017, 7, 25 8 of 14 is worth noting that the mercury extrusion branches (data not shown) were, in all cases, practically parallel to the log(R)-axis, indicating that practically all the mercury is accumulated in the pore voids and that the amount of the necks (channels) connecting these voids is negligible.The volume of intruded mercury (Figure 6a) is the lowest for the LWA containing only the original clay deposit, the intermediate for that enriched with the spent clinoptilolite, and the highest for the material containing spent NaP1.The pore size distributions for CLAY and NaP1 are unimodal (Figure 6b).One broad peak is noted for CLAY with the maximum located at around 0.32 µm (logR ~ 0.5), while one narrow peak is found for NaP1 with maximum at R ~ 0.16 µm.Three peaks on the pore size distributions (PSD) function of CLIN are present: two narrow peaks at 32 µm and 2.5 µm, and one broad peak at around 0.16 µm.In contrast, Korat et al. [10] observed only bimodal MIP pore size distributions (peaks with maximum values between 0.1 and 1 µm, above 10 µm, and up to 100 µm) for LWAs prepared from fly ash obtained from coal combustion and silica sludge.Comparing the pore size distribution functions derived from MIP and microtomography, one can see that MIP measurements allocate the sizes of almost the entire volume of the pores towards an underestimation of the large pores and an overestimation of the small pores.This phenomenon, as summarized by Korat et al. [10], appears to be rather intrinsic than accidental, which derives from the lack of direct accessibility for most of the pore volume (including air voids) to the mercury surrounding the specimen.Furthermore, in the case of highly porous structures, errors can also be made due to the breaking of the inner pore's walls, which then give distorted results.
Fractal plots for the studied materials are illustrated in Figure 7.As a rule, the fractal behaviour of the porosity of natural objects occurs in a limited range of pore dimensions (called upper and lower cut-offs) [49].The volume of intruded mercury (Figure 6a) is the lowest for the LWA containing only the original clay deposit, the intermediate for that enriched with the spent clinoptilolite, and the highest for the material containing spent NaP1.The pore size distributions for CLAY and NaP1 are unimodal (Figure 6b).One broad peak is noted for CLAY with the maximum located at around 0.32 µm (logR ~0.5), while one narrow peak is found for NaP1 with maximum at R ~0.16 µm.Three peaks on the pore size distributions (PSD) function of CLIN are present: two narrow peaks at 32 and 2.5 µm, and one broad peak at around 0.16 µm.In contrast, Korat et al. [10] observed only bimodal MIP pore size distributions (peaks with maximum values between 0.1 and 1 µm, above 10 µm, and up to 100 µm) for LWAs prepared from fly ash obtained from coal combustion and silica sludge.
Comparing the pore size distribution functions derived from MIP and microtomography, one can see that MIP measurements allocate the sizes of almost the entire volume of the pores towards an underestimation of the large pores and an overestimation of the small pores.This phenomenon, as summarized by Korat et al. [10], appears to be rather intrinsic than accidental, which derives from the lack of direct accessibility for most of the pore volume (including air voids) to the mercury surrounding the specimen.Furthermore, in the case of highly porous structures, errors can also be made due to the breaking of the inner pore's walls, which then give distorted results.
Fractal plots for the studied materials are illustrated in Figure 7.As a rule, the fractal behaviour of the porosity of natural objects occurs in a limited range of pore dimensions (called upper and lower cut-offs) [49].The geometrical irregularities and roughness of the pore surface have an essential influence on the value of the fractal dimensions, which, for porous solids, may vary from 2 to 3. The lower limiting value of 2 corresponds to a perfectly regular pore surface, whereas the upper limiting value of 3 relates to the maximum allowed pore surface complexity [49].The linearity ranges of log-log plots of dV/dR vs. R can be found for the studied aggregates.One linearity range is found for CLAY in the range of large pores.NAP1 and CLIN exhibit two ranges of linearity: one for large pores and the second for narrow pores.
However, the slopes of the linear log-log plots are very high in all cases, such that the calculated fractal dimensions of the pore surfaces are larger than 3 in all cases except for narrow pores of NAP1 (see Table 2 below).This may result from the specific structure of the aggregates.The large pore voids are accessible through markedly narrower entrances, therefore the volume of mercury forced into a large pore is attributed to the radius of the entrance and not to the radius of the void, falsifying the location of pore volume.In fractal dimension calculations a cylindrical pore model was applied assuming that the pore is a long capillary having the radius of the entrance.It is far from reality and leads to rapid increase of pore volume with pore (entrance) radius.Having such high increase in pore volume V vs. radius R dependence the cylindrical pore model calculates high dV/dR values that gives fractal dimension D values higher than 3.
Extremely low nitrogen adsorption and the calculated surface areas of the produced LWAs, which are less than 1 m 2 /g (see Table 2), indicate that either the vitrified layer produced during heating has an extremely flat surface or the closed intra-aggregate pores are not available for nitrogen molecules.
SEM microphotographs of the obtained LWAs presented in Figure 8 show differences in the finest pores' structure of the aggregates.LWA prepared from clay is characterized by a compact texture, with the smallest pores being oval and frequently elongated.Aggregates with the admixtures of spent zeolites have pores of larger sizes, being the largest ones for NAP1.The placement of the pores is rather irregular.The geometrical irregularities and roughness of the pore surface have an essential influence on the value of the fractal dimensions, which, for porous solids, may vary from 2 to 3. The lower limiting value of 2 corresponds to a perfectly regular pore surface, whereas the upper limiting value of 3 relates to the maximum allowed pore surface complexity [49].The linearity ranges of log-log plots of dV/dR vs. R can be found for the studied aggregates.One linearity range is found for CLAY in the range of large pores.NAP1 and CLIN exhibit two ranges of linearity: one for large pores and the second for narrow pores.
However, the slopes of the linear log-log plots are very high in all cases, such that the calculated fractal dimensions of the pore surfaces are larger than 3 in all cases except for narrow pores of NAP1 (see Table 2 below).This may result from the specific structure of the aggregates.The large pore voids are accessible through markedly narrower entrances, therefore the volume of mercury forced into a large pore is attributed to the radius of the entrance and not to the radius of the void, falsifying the location of pore volume.In fractal dimension calculations a cylindrical pore model was applied assuming that the pore is a long capillary having the radius of the entrance.It is far from reality and leads to rapid increase of pore volume with pore (entrance) radius.Having such high increase in pore volume V vs. radius R dependence the cylindrical pore model calculates high dV/dR values that gives fractal dimension D values higher than 3.
Extremely low nitrogen adsorption and the calculated surface areas of the produced LWAs, which are less than 1 m 2 /g (see Table 2), indicate that either the vitrified layer produced during heating has an extremely flat surface or the closed intra-aggregate pores are not available for nitrogen molecules.
SEM microphotographs of the obtained LWAs presented in Figure 8 show differences in the finest pores' structure of the aggregates.LWA prepared from clay is characterized by a compact texture, with the smallest pores being oval and frequently elongated.Aggregates with the admixtures of spent zeolites have pores of larger sizes, being the largest ones for NAP1.The placement of the pores is rather irregular.The pore parameters of the pore system derived from microtomography, MIP and SEM experiments are summarized in Table 2.The pore parameters of the pore system derived from microtomography, MIP and SEM experiments are summarized in Table 2.
All methods applied give the highest pore volumes and porosities for NAP1 and the lowest for CLAY aggregates.As a rule, MIP measures significantly higher pore volumes and porosities than microtomography.The measuring range of microtomography starts from ~6 µm upwards, while it runs from ~4 nm to ~14 µm for MIP; at the first glance, it does not seem possible that MIP registers larger porosities.However, mercury can invade the whole aggregate interior through narrow entrances to the large pores, thereby filling all large pores inside.Therefore one can state that the total porosity values measured by microtomography are more reliable than these derived from MIP. LWAs made from the clay deposit have the smallest porosity and the highest particle density, whereas LWAs containing spent clinoptilolite and NaP-1 zeolites have larger porosity and smallest particle density that may be due to the presence of the oil in the spent zeolites.
Organic substances produce additional gases during the sintering process, which contribute to the formation of pore beads and the creation of more porous structure of the aggregate [56].However, similarly low densities (0.7-0.9) were achieved by Volland and Brötz [3] for sand sludge LWAs admixed with 20-40% of heulanditic zeolite rock.Such low bulk densities (from 0.95 to 0.7) were also achieved by Mun [57] for LWAs admixed with different doses of a sewage sludge.Kourti and Cheeseman [26] found that sintering 60:40 lignite coal fly ash with waste glass mixes produced LWAs with a mean density of 1.35 g•cm −3 , thereby suggesting that heat treatment of organic material containing substrates gives smaller bulk densities and higher porosities of the resulting LWAs than using organic-derived fly ashes coming from similar organic material.It is worth noting that practically the same bulk densities of the aggregates are measured by mercury intrusion and from the LWA volume and mass (Table 2), indicating that the amount of very fine pores being unavailable for mercury is very small in all LWAs studied.The solid phase (skeletal) density is the highest for CLAY and the smallest for CLIN aggregates.It could be possible that the presence of residual carbon formed from no oxidized oil additions diminished the solid phase density; however, no carbon could be detected in the LWAs studied.The finest close pores are possibly responsible for the above effect.The fractal dimensions calculated from microtomography data are rather high, indicating the complex pore buildup, of which the least diversified is found in the CLAY aggregate.All microtomography fractal dimensions fall within the range between 2 and 3, therefore it is likely that microtomography provides a more realistic picture of the LWAs' fractal pore structure than MIP.This may be due either to the application of the spherical pore model for microtomography data elaboration (instead of cylindrical pore spaces model in MIP) or more probably to a failure of the MIP application in describing LWAs' pore size distribution.

Conclusions
Although the addition of spent zeolite sorbents increased the amount of the amorphous glassy phase in the LWAs, their mineral composition stayed intact, as evidenced by the XRD results.The addition of spent zeolites has fostered a decrease in the particle density, which in turn has involved a decrease in the mechanical resistance.A decrease in water absorption also occurred.The pore structure of LWAs prepared from a clay deposit was strongly modified by the addition of spent zeolites, depending on the composition of the starting mixture.All the methods applied measured the same tendencies of changes in pore volumes and porosities of LWAs due to the addition of spent zeolites.The porosity of the LWAs prepared from a clay deposit was the lowest and the addition of spent NaP1 resulted in the highest porosity of the obtained LWAs.An increase in porosity may also be connected with the amount of the oil present within the added zeolites: with more oil addition the more porous structure is formed.Changes in the average pore radius measured by microtomography and MIP did not run parallel with the pore volume changes.Only the dominant pore radius measured by SEM increased to a similar degree as the porosity.
The reuse (addition) of the spent zeolitic sorbents containing petroleum waste to produce LWAs is a novel method dedicated to this kind of waste utilization.Furthermore, it leads to very advantageous properties of the resulting LWAs (high porosity, low water sorption, enough mechanical resistance, high freezing resistance), indicating their applicability in geotechnics, building construction and agriculture.

Figure 1 .
Figure 1.Particle size (diameter) distributions for the initial materials (a) and cumulative volume (scaled to 100%) versus particle diameter plot (b).

Figure 1 .
Figure 1.Particle size (diameter) distributions for the initial materials (a) and cumulative volume (scaled to 100%) versus particle diameter plot (b).

Figure 4 .
Figure 4. Exemplary 2D cross-sectional images derived from microtomography for the studied materials.Black areas: pores; white areas: solid.

Figure 5 .
Figure 5. Pore volume vs. pore radius dependencies (a) and normalized pore size distribution functions; (b) derived from microtomography scans.The points show average results, while the error bars show differences between the average and the experimental replicates.

Figure 4 .
Figure 4. Exemplary 2D cross-sectional images derived from microtomography for the studied materials.Black areas: pores; white areas: solid.

Figure 4 .
Figure 4. Exemplary 2D cross-sectional images derived from microtomography for the studied materials.Black areas: pores; white areas: solid.

Figure 5 .
Figure 5. Pore volume vs. pore radius dependencies (a) and normalized pore size distribution functions; (b) derived from microtomography scans.The points show average results, while the error bars show differences between the average and the experimental replicates.

Figure 5 .
Figure 5. Pore volume vs. pore radius dependencies (a) and normalized pore size distribution functions; (b) derived from microtomography scans.The points show average results, while the error bars show differences between the average and the experimental replicates.

Figure 6 .
Figure 6.MIP curves (a) and normalized pore size distribution functions; (b) for the studied aggregates.The points show average results, while the error bars show maximum differences between the average and the experimental replicates.

Figure 6 .
Figure 6.MIP curves (a) and normalized pore size distribution functions; (b) for the studied aggregates.The points show average results, while the error bars show maximum differences between the average and the experimental replicates.

Figure 7 .
Figure 7. Fractal plots for the studied materials.The points show average results, while the error bars show maximum differences between the average and the experimental replicates.

Figure 7 .
Figure 7. Fractal plots for the studied materials.The points show average results, while the error bars show maximum differences between the average and the experimental replicates.

Table 1 .
Physical parameters of the LWAs.

Table 2 .
Pore parameters of the studied LWAs.
Figure 8. Representative scanning electron microscope microphotographs (SEM) of the studied aggregate sections.

Table 2 .
Pore parameters of the studied LWAs.