*2.4. SEM*/*EDX and Porosimetric Analyses*

An electron microscope FESEM-EDX Carl Zeiss Sigma 300 VP (Carl Zeiss Microscopy GmbH (Jena, Germany)) was used to characterize the morphology and the chemical composition of the samples which were applied onto aluminum stubs and sputtered with gold (Sputter Quorum Q150 Quorum Technologies Ltd (East Sussex, UK)) before the test. In this respect, the composition of the normalized sand was: C (4%), O (52%), Si (35%), Ca (2%), the composition of the sieved sand was: C (10%), O (45%), Ca (45%), polystyrene composition was: C (30%), O (70%), the composition of the cement paste was: C (4.2%), O (40%), Si (7.6%), Ca (44%), Fe (1.5%), Al (2.5%). Ultrapyc 1200e Automatic Gas Pycnometer, Quantachrome Instruments (Boynton Beach, FL, USA) was used for the porosimetric measurements and helium was used to penetrate the material pores.

## **3. Results and Discussion**

Flow data of the non-consolidated samples are reported in Figure 1 and were obtained after measuring the diameters of the mixture before and after the test [51]. The flow of a sample is represented by the percentage increase of the diameter over the base diameter.

**Figure 1.** Flow-test results.

The Sand sample showed a higher flow (+35%) with respect to the Normal sample due to the absence of finer aggregates. The EPS specimens were more fluid than both references and specifically with respect to the normalized mortar (Normal). This behavior may be ascribed to the low surface energy, the low roughness (smooth surface), the hydrophobic features (synthetic organic polymer) and the low density of the EPS particles (10–30 g/dm3 with respect to 1700 g/dm3 of sand) which may induce aggregate segregation in a cement conglomerate. The lower fluidity of EPS3 (+126%) with respect to EPS2 (+174%) and EPS4 (+150%) is likely due to a better compaction of the aggregates in the mixture (better distribution of the granules), while in the case of the Sand/EPS specimen, the presence of the inorganic aggregate contributed to a reduction of fluidity (Figure 1). In Figure 2 and Table 3 the flexural and compressive strengths of the samples are reported as a function of the specific mass. The Sand sample showed slightly higher mechanical strengths than the Normal sample due to the presence of larger size aggregates which contribute to the increase of the specific mass. The addition of EPS determined the formation of voids in the composite with a sensible reduction of the specific mass of the mortars (Table 2) which depends not only on the matrix and polymer characteristics (foaming structure of EPS), but also on the interface properties [53–55]. For this reason, after total replacement of the sand volume, a decrease was observed of the mechanical strengths of the conglomerates, this

effect is ascribable to the low density/high porosity of the EPS beads (inset Figure 2) and to the voids created by the aggregate at the cement/EPS interface during mixing [53,54]. In fact, the porosity of these samples is approximately two times higher than the references (Table 2). For this purpose, the flexural and the compressive resistances of EPS2, EPS3, and EPS4 samples were approximately ~80% lower than the references, with compressive strengths passing from nearly 50 MPa to less than 10 MPa as the specific mass was lowered from 2100 to 900 kg/m3. After substitution of 50% of the sand volume with EPS beads (Sand-EPS), an increase was observed of the mechanical strengths with respect to the EPS specimens. In fact, the flexural strength decrease was approximately 25% with respect to both references, while the compressive strength was 25%–30% lower than the references.

**Figure 2.** Flexural and compressive strengths of the samples (28 days curing). The label EPS (expanded polystyrene) represents EPS 2, EPS3, and EPS4. White squares represent the compressive strengths, while black squares represent flexural strengths. In the inset: inner porosity of an EPS bead (SEM image).

The EPS mortars did not show a flexural brittle behavior which can be observed in the sand specimens (Normal and Sand), but the rupture was more gradual and the mortars containing 100% EPS volume did not show a separation of the two parts [56,57]. The Sand-EPS sample, characterized by 50% of sand and 50% of EPS, showed a semi-brittle behavior. As in the former case, the compressive failure of the EPS2, EPS3, and EPS4 mortars was gradual with high energy absorption because of the load retention after rupture without collapse [56,58,59]. As expected, the reference samples showed a typical brittle failure. It was observed that most of the aggregates of the EPS3 and EPS4 specimens sheared off along the failure plane (Figure 3A,B), on the contrary no damage was observed to most of the aggregates in EPS2 mortar and some of the EPS2 beads were de-bonded from the matrix (Figure 3C).


**Table 3.** Mechanical strengths (28 days curing) of the samples.

**Figure 3.** (**A**) SEM image of the cement paste/EPS interface in the EPS3 sample. (**B**) SEM image of the cement paste/EPS interface in the EPS4 sample. (**C**) SEM image of the cement paste/EPS interface in the EPS2 sample, in the inset an image of the de-bonded EPS bead.

From these results, it can be concluded that the bond between the EPS2 aggregate and the cement paste was weaker than the failure strength of the aggregate (poor EPS adhesion to the cement paste), while the bond between the EPS aggregate and the cement paste in EPS3 and EPS4 samples was stronger (better EPS adhesion to the cement paste) than the failure strength of the polystyrene granules [33,60].This effect was noticed in particular on the EPS3 sample (Figure 3A). The latter result is indicative of better cohesion between the aggregate and the cement paste. Thus, EPS3 exhibited higher compaction which packs the aggregate particles together, so as to increase the specific mass of the mortar, and this also explains the lower percentage flow with respect to the other samples which resulted in more fluid and with a higher tendency to segregation [20] (see Figure 1).

The lower specific mass of the EPS2 sample can be demonstrated by large voids at the ligand/aggregate interface, with length comparable to EPS beads and 20–30 micron width, this effect was ascribed to the mentioned poor adhesion of the beads to the cement paste (Figure 4A,B). This result was also observed in the EPS3 sample, but in the latter case the adhesion of the sheared-off particles to the cement paste was better thus demonstrating the higher specific mass of this type of lightweight mortar. Moreover, from Figure 4C, the perfect adhesion of the sand to the cement paste is evident. In fact, from the map relative to the Si element, which is barely present in the limestone, a negligible separation between the sand and the ligand can be observed ascribed to a favorable adhesion.

**Figure 4.** (**A**,**B**) SEM images of the cement paste/EPS interface in the EPS2 sample. (**C**) SEM image of the normalized mortar and, in the inset, EDX map relative to the Si distribution in the sample.

The time variation of the flexural and compressive strengths of the Normal sample, of the EPS3 and Sand/EPS samples is reported in Figure 5 where an increase of the resistances can be observed with stabilization after 45 days. At 60 days, the values did not sensibly change thus demonstrating stability of the materials on consideration of the specific water curing/conservation conditions of the conglomerates.

**Figure 5.** Flexural (**A**) and compressive (**B**) strengths of the samples over time.

EPS based mortars showed lower thermal conductivities and diffusivities than the sand references (Figure 6). This result can be ascribed to the lower specific mass of the specimens due to the low density of the organic aggregates [61,62] (see inset Figure 2) together with the mentioned voids at the EPS/ligand interface which limit heat transport in the composite. Specifically, the thermal conductivities of the bare EPS specimens were ~80% lower than the references. The best results were obtained in the case of the EPS4 specimen (0.29 W/mK) due to the lowest specific mass. Intermediate values (0.8 W/mK) were obtained in samples with 50% of EPS (Sand/EPS sample). Thermal conductivity and diffusivity data showed an exponential decrease with the decrease of the conglomerates specific mass.

**Figure 6.** (**A**) Thermal conductivity and (**B**) thermal diffusivity of the samples.

The wetting characterization of the side surface (Figure 7) and of the inner surface (Figure 8) of the Normal sample was carried out. Figure 7A,B shows the time evolution of the water contact angle (WCA) and of the drop height for the side surface of the Sand sample. A hydrophilic character (WCA < 90◦) [35] was observed although different behavior on various points of observation was detected. Fast WCA decrease and full penetration occurred in few seconds at point 3, slower but full water absorption occurred at point 2, whereas higher WCA and negligible water absorption were observed in the case of point 1. Figure 7C shows the pictures relative to the drop behavior. The side surface of the reference mortar based on normalized sand (Normal) showed similar features. It is worth highlighting that the possibility of detecting and quantifying the spatially non-homogeneous behavior of a surface/material like these is a specific benefit of the spatially resolved evaluation of wettability and absorption made by this technique (the drop volume is 5 μL), which cannot be achieved with water permeability or capillary absorption measurements.

**Figure 7.** (**A**) Contact angle and (**B**) height variation over time for water drops deposited on representative points of the side surface of the normalized mortar (Sand). (**C**) Optical microscope images (down: point 1 drop, top: point 3 drop).

Figure 8A,B shows the wetting parameters relative to the fracture surface. The inner surface resulting from the mechanical breakage can be considered more representative of the composite features because it is a section of the sample showing every component of the mixture. It shows an open porosity characterized by high roughness and a visible distribution of the aggregates, as opposed to what is observed on the side surface. Specifically, the results obtained on every point of the observation were similar. A fast decrease of the water contact angle and of the drop height was observed at every point (Figure 8C). On the contrary to what was observed on the side surface, WCA was lower, thus the fracture surface can be generally considered super-hydrophilic (WCA ~0–5 [35,63] and fast absorbent. As in the former case, similar results were observed on the inner surface of the Normal sample.

The wetting characterization of the EPS3 mortar, with EPS grains in the 2–4 mm (50%) and 4–6 mm (50%) bead size range, is reported in Figures 9 and 10. As described above, EPS totally replaced the sand volume. Figure 9A,B shows the time evolution of the water contact angle (WCA) and of the drop height on the side surface of the sample. Different trends were observed. A slow but full water absorption occurred at point 1, higher WCA and negligible water absorption were observed in the case of points 2 and 3, the latter with WCA ≥ 90◦. In the present case, the side surface resulted in being more hydrophobic than the references.

**Figure 8.** (**A**) Contact angle and (**B**) drop height for representative points of the fracture surface of the normalized mortar (Sand). (**C**) In the optical microscope image: point 2 drop.

**Figure 9.** (**A**) Contact angle and (**B**) drop height for representative points of the side surface of the EPS3 mortar. (**C**) In the optical microscope image: point 2 drop.

Figure 10A,B shows the time evolution of the water contact angle (WCA) and of the drop height on the fracture surface of the EPS3 sample. In this case, the drop was stable for all the observation time. Figure 10 also shows a picture of the drop after deposition onto the specimen surface (point 2) which resulted in being hydrophobic with high WCA (WCA > 90◦) [35]. The latter result was confirmed

after a drop deposition onto an EPS slab or onto bare EPS beads, specifically in the first case WCA was approximately 99◦, while higher in the second (100–102◦) probably due to the beads' curvature. The WCA was higher on bare beads with respect to the EPS in the mixture because of the absence of contamination from the hydrophilic cement paste [64,65]. For this purpose, after deposition onto cement paste regions of the EPS3 sample (points 1 and 3), hydrophilic behavior but negligible water absorption were observed. This latter result is ascribed to the hydrophobic and non- absorbing effect of EPS whose sites decrease the mean surface energy of the sample making the presence of the porous and hydrophilic cement regions ineffective [64,65].

**Figure 10.** (**A**) Contact angle and **(B**) drop height for representative points of the fracture surface of the EPS3 mortar. (**C**) In the optical microscope image: point 2 drop.

The wetting characterization of the fracture surface of the EPS4 mortar, with EPS grains in the 1–2 mm (25%), 2–4 mm (25%), and 4–6 mm (50%) bead size range, is reported in Figure 11A, while the results obtained on the side surface were similar to those of the EPS3 specimen. The fracture surface is hydrophobic in the domain of the polystyrene beads (point 2) and hydrophilic in the domain of the cement paste (point 3) because the drop was deposited onto a hydrophilic and absorbent surface. As a matter of fact, the latter result represents a difference between the fracture surface of this sample and the fracture surface of the former composite (EPS3).

Wetting characterization of the fracture surface of the EPS2 mortar, with EPS grains in the 4–6 mm (100%) bead size range, is reported in Figure 11B and in this case, the results obtained on the side surface of this sample were similar to those observed in the case of the former EPS specimens. In the case of the fracture surface, a hydrophilic character was observed at every point of observation, with very low water contact angle and fast water absorption.

**Figure 11.** Contact angle for representative points of the fracture surface of (**A**) EPS4 and (**B**) EPS2 mortars.

Thus, EPS3 is the specimen with the lowest water drop absorption. This may be due to a more efficient organization of the aggregate particles with open spaces (spheroidal microcavities) between larger particles filled with smaller size EPS beads [49,66], which leads to better behavior of the composite. This specimen indeed shows the highest specific mass and lowest porosity among the EPS specimens, reasonably as a consequence of a better aggregate compaction (evidenced by the lowest flow). This property on the one hand results in a slight reduction of the thermal insulation performances, but on the other hand makes the composite definitely less subject to water ingress. The importance of optimizing the level of compaction, by adjusting the size distribution of the EPS aggregates, is due to the relatively large size of the initial EPS beads, which results in the formation of too large channels of the cement matrix among the aggregates in the hardened artifacts.

Hence, once properly distributed in size, EPS beads can represent suitable aggregates in cementbased artifacts both for lightening/insulating and water proofing purposes. Such a double advantage arises from the peculiar combination of low density and low surface energy of this plastic matter as already shown by using other polymeric aggregates, such as granulated rubber from end-of-life tires [53].

#### **4. Conclusions**

In the present work an investigation into the rheological, thermo-mechanical, microstructural and wetting characteristics of cement mortars containing recycled expanded polystyrene (EPS) was carried out. The samples were prepared after partial/total replacement of the conventional sand aggregate with EPS having different grain size and size distribution. The experimental results may be summarized as follows:


**Author Contributions:** Conceptualization, A.P.; methodology, A.P.; software, R.D.M.; validation, A.P., R.D.M. and M.N.; formal analysis, A.P.; investigation, A.P., R.D.M.; resources, A.P.; data curation, A.P.; writing—original draft preparation, A.P.; writing—review and editing, A.P., R.D.M., M.N.; visualization, M.N.; supervision, M.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** Special thanks to Pietro Stefanizzi and Stefania Liuzzi for thermal analysis. Adriano Boghetich is acknowlwdged for SEM-EDX analysis and also Regione Puglia (Micro X-Ray Lab Project–Reti di Laboratori Pubblici di Ricerca, cod. n. 45 and 56). Acknowledgments to the DICATECh of the Polytechnic of Bari for SEM analyses.

**Conflicts of Interest:** The authors declare no conflict of interest.
