Hybrid Structures Made of Polyurethane / Graphene Nanocomposite Foams Embedded within Aluminum Open-Cell Foam

: This paper focuses on the development of hybrid structures containing two di ﬀ erent classes of porous materials, nanocomposite foams made of polyurethane combined with graphene-based materials, and aluminum open-cell foams (Al-OC). Prior to the hybrid structures preparation, the nanocomposite foam formulation was optimized. The optimization consisted of studying the e ﬀ ect of the addition of graphene oxide (GO) and graphene nanoplatelets (GNPs) at di ﬀ erent loadings (1.0, 2.5 and 5.0 wt%) during the polyurethane foam (PUF) formation, and their e ﬀ ect on the ﬁnal nanocomposite properties. Globally, the results showed enhanced mechanical, acoustic and ﬁre-retardant properties of the PUF nanocomposites when compared with pristine PUF. In a later step, the hybrid structure was prepared by embedding the Al-OC foam with the optimized nanocomposite formulation (prepared with 2.5 wt% of GNPs (PUF / GNPs2.5)). The process of ﬁlling the pores of the Al-OC was successfully achieved, with the resulting hybrid structure retaining low thermal conductivity values, around 0.038 W · m − 1 · K − 1 , and presenting an improved sound absorption coe ﬃ cient, especially for mid to high frequencies, with respect to the individual foams. Furthermore, the new hybrid structure also displayed better mechanical properties (the stress corresponding to 10% of deformation was improved in more than 10 and 1.3 times comparatively to PUF / GNPs2.5 and Al-OC, respectively). investigation, S.C.P., I.D., P.A.A.P.M., L.G. and R.V.; writing—original draft preparation, S.C.P. and P.A.A.P.M.; writing—review and editing, S.C.P., I.D., R.V. and P.A.A.P.M.; supervision, I.D.; P.A.A.P. and R.V.


Introduction
In recent years, porous materials have attracted a huge interest from both academia and industry because they may find applications in a variety of fields, such as energy storage [1], catalysis [2], drug release [3], sound and thermal insulation [4], environmental remediation [5] and others. Giving the International Union of Pure and Applied Chemistry (IUPAC) definition, porous materials can be categorized based on their pore sizes: microporous (pore size <2 nm), mesoporous (2-50 nm), and macroporous (>50 nm) [6]. According to these categories, the properties of the porous materials and their subsequent applications will differ. Moreover, they can be found in three-dimensional (3D) and two-dimensional (2D) structures. Common 3D porous materials are sponges, foams, wood, and bone. Two-dimensional porous materials include separation membranes, filter paper, textiles, and so on.

Sample Preparation
Pristine PUF and PUF nanocomposites were prepared by a two-step procedure, as schematized in Figure 1, route A. First, the pre-polymer (polyol), silicone oil (5 wt%), water as blowing agent (5 wt%) and GO or GNPs (1.0, 2.5 and 5.0 wt%) were placed in glass beaker and homogenized for 30 s using a mechanical stirrer at high speed. Next, the proper amount of MDI to obtain a R NCO/OH = 0.80 (ratio between NCO groups of isocyanates and OH groups) was added and the mixture was homogenized again for 10 s. The PUF and PUF nanocomposites were obtained by free expansion in the cup mold at room temperature. The foams are hereafter referred as PUF, PUF/xGNPs and PUF/xGO, where x refers to the carbon nanostructure content.

Sample Preparation
Pristine PUF and PUF nanocomposites were prepared by a two-step procedure, as schematized in Figure 1, route A. First, the pre-polymer (polyol), silicone oil (5 wt%), water as blowing agent ( 5 wt%) and GO or GNPs (1.0, 2.5 and 5.0 wt%) were placed in glass beaker and homogenized for 30 s using a mechanical stirrer at high speed. Next, the proper amount of MDI to obtain a RNCO/OH = 0.80 (ratio between NCO groups of isocyanates and OH groups) was added and the mixture was homogenized again for 10 s. The PUF and PUF nanocomposites were obtained by free expansion in the cup mold at room temperature. The foams are hereafter referred as PUF, PUF/xGNPs and PUF/xGO, where x refers to the carbon nanostructure content.
For the preparation of the Al-OC hybrid (the interconnected porous metallic foam with the optimized PUF nanocomposite) the metallic foam was placed into the cup containing the nanocomposite mixture right after the addition of MDI. The pores of the Al-OC were filled during the nanocomposite foam expansion (Figure 1, route B), hereafter referred as PUF/GNPs2.5-OC. Before characterization, the samples were settled to rest for 24 h at room temperature to ensure complete reaction.

Sample Characterization
Attenuated total reflection-Fourier transformed infrared spectroscopy (ATR-FTIR) spectra were collected using a Perkin Elmer FTIR System Spectrum BX Spectrometer (Buckinghamshire, UK) equipped with a single horizontal Golden Gate ATR cell, in the range 4000 to 500 cm −1 and running 64 scans with a resolution of 4 cm −1 . Scanning electron microscopy (SEM) analysis was performed in a TM 4000 Plus (Hitachi, Tokyo, Japan) scanning electron microscope at accelerating voltage of 15.0 kV. Samples (10 × 10 × 10 mm 3 ) were analyzed in a X-ray microcomputed tomography (µCT) equipment from SkyScan 1275 (Bruker µCT, Kontich, Belgium) with penetrative X-rays of 30 kV and 125 µA, in high-resolution mode with a pixel size of 8 µm and 450 ms of exposure time. NRecon and CTVox software (Bruker, Kontich, Belgium) were used for 3D reconstruction and CTan software For the preparation of the Al-OC hybrid (the interconnected porous metallic foam with the optimized PUF nanocomposite) the metallic foam was placed into the cup containing the nanocomposite mixture right after the addition of MDI. The pores of the Al-OC were filled during the nanocomposite foam expansion (Figure 1, route B), hereafter referred as PUF/GNPs2.5-OC. Before characterization, the samples were settled to rest for 24 h at room temperature to ensure complete reaction.

Sample Characterization
Attenuated total reflection-Fourier transformed infrared spectroscopy (ATR-FTIR) spectra were collected using a Perkin Elmer FTIR System Spectrum BX Spectrometer (Buckinghamshire, UK) equipped with a single horizontal Golden Gate ATR cell, in the range 4000 to 500 cm −1 and running 64 scans with a resolution of 4 cm −1 . Scanning electron microscopy (SEM) analysis was performed in a TM 4000 Plus (Hitachi, Tokyo, Japan) scanning electron microscope at accelerating voltage of 15.0 kV. Samples (10 × 10 × 10 mm 3 ) were analyzed in a X-ray microcomputed tomography (µCT) equipment from SkyScan 1275 (Bruker µCT, Kontich, Belgium) with penetrative X-rays of 30 kV and 125 µA, in high-resolution mode with a pixel size of 8 µm and 450 ms of exposure time. NRecon and CTVox software (Bruker, Kontich, Belgium) were used for 3D reconstruction and CTan software (Bruker, Kontich, Belgium) was used in morphometric analysis (total porosity, pore size distribution and cell-wall thickness). The apparent density and porosity were determined geometrically for three specimens of each PUF nanocomposite.
The thermal conductivity properties were evaluated with a Hot Disk TPS 2500 S instrument (Gothenburg, Sweden), at 20 • C; in accordance with the standard ISO 22007-2.2 and ASTM D7984, the specimens were cube shaped with (25 × 25 × 25 mm 3 ). The value of the sound absorption coefficient was estimated from measurements made with an impedance tube according standard ASTM E 1050 [36] for cylindrical shaped specimens with 37 mm of diameter and 22 mm of thickness. The thermal stability of the nanocomposites foams was assessed by thermogravimetric analysis (TGA) using a thermogravimetric analyzer (Netzsch Jupiter, Selb, Germany) at a scanning rate of 10 • C/min, in the temperature range of 30-800 • C, under a synthetic air atmosphere (80% N 2 and 20% O 2 ). The fire retardancy test was based in the direct observation of the response of the specimens when submitted to a flame. The test consisted in applying an ethanol flame, at the specimen's bottom using the set-up in vertical sample position, for 3 s, plus a subsequent application (3 s) if the specimen self-extinguished.
The tests were conducted in half cylindrical shaped specimens, with 30 mm of diameter and 10 mm of thickness.
The mechanical testing machine (Shimadzu MMT, Kyoto, Japan; maximum load 101 N) was used to study the quasi-static compressive response of PUF nanocomposites (10 × 10 × 10 mm 3 ) under a strain rate of 1 mm/min up. The uniaxial compression test of Al-OCF and hybrid PUF structures were performed in a Shimadzu-AGS-X-10kN (Kyoto, Japan) testing machine at a speed of 6 mm/min.

PUFs Nanocomposites
The success of PUF and PUF nanocomposites preparation relies on the appropriate reaction between the precursors, namely the extinction of isocyanate groups through the reaction with hydroxyl groups of the polyol and urethane formation; this was confirmed by FTIR analysis ( Figure S1). The addition of GO or GNP, at the used concentrations, did not prevent the progression of foam formation. However, the GO and GNPs did not disperse in the same way in the polymer matrix, as can be easily observed in the left column of Figure 2. The photographs show a homogeneous black color when GNPs were added, suggesting a good interaction between these nanofillers and the polymer matrix. On the contrary, black spots were observed in the PUF when the GO was used. The GO nanosheets were directly obtained from the chemical exfoliation of graphite and contain several oxygen chemical functionalities, which is what makes the GO highly hydrophilic. The GNPs were also obtained from graphite exfoliation, but without the use of chemical oxidants, thus resulting in a non-oxidized surface [37], GNPs being hydrophobic. This difference in the chemical surface structure of the carbon nanostructures determines their dispersion in the polymeric matrix, which have hydrophobic domains enabling GNPs dispersion. This section will provide a concise and precise description of the experimental results, their interpretation, as well as the experimental conclusions that can be drawn. mainly as co-flame retardants, at loadings from 1.0 to 10 wt% [41]. Also, functionalized graphene designed and prepared from expandable graphite and phosphorus-containing compounds has been described as flame retardants and smoke suppressor of PUF at 6.1 wt% [30]. It is worth mentioning that in our study, the pristine PUF and PUF nanocomposites with 1 wt% of nanofillers also kept their shape after burning. However, due to the better results obtained for 2.5 and 5.0 wt%, the PUF with 1.0 wt% of nanofillers were excluded in the further characterization results. The thermogravimetric (TG) and derivative thermogravimetric (DTG) curves are shown in Figure S3a and Figure S3b, respectively. The pristine PUF and PUF nanocomposites showed similar TGA curves, suggesting that GNPs and GO do not significantly influence the decomposition of PUF. This finding is most likely related to the small amount of nanofillers used. The initial decomposition temperature corresponding to 5% (T5%), and 50% (T50%) of mass degradation, the maximum-rate degradation temperature (Tmax), and mass residue at 750 °C are listed in Table 1. The results show that GNPs had a positive effect on T5% on PUF, while GO was detrimental to the early thermal stability. These results are caused by the earlier degradation of the GO oxygen functionalities at low temperature, thus accelerating the degradation of the PUF matrix. The oxygen functionalities of GO may interact with the PUF precursors during the foam formation thus interfering with the

Thermal Stability
Although the dispersion of both nanofillers (GNPs and GO) was not the same, their flame-retardant action was remarkable. A control test with pristine PUF showed flame propagation at some extension with smoke release, even if no dripping was observed, (Figure 2a). On the contrary, PUF nanocomposites prepared with 2.5 wt%, showed fire retardant properties, with the flame extinguished after 1 s. Importantly, the specimens maintained their shape after burning without dripping or smoke release (Figure 2b,c). Similar behavior was observed for the compositions with 5.0 wt% of the GO or GNPs.
For the lower nanofillers amount tested (1 wt%), the results were not as good, since at 3 s there was still flame propagation, mainly for PUF/GNPs1.0, with smoke release in both cases ( Figure S2). Carbon nanostructures are known to play a key role either in slowing down the flame propagation or even in providing self-extinction to thermoplastic or thermosetting polymeric matrices [38]. Their positive effect in the thermal stability of PUF nanocomposites may be attributed to the high specific area and layered structure of the nanofillers, which tend to form a dense and continuous char layer acting like a physical barrier at the PUF surface. This barrier becomes an obstacle to the release of the volatile degradation products, preventing or causing the delay of the degradation of the whole composite [39]. It is remarkable that, although the GO nanosheets are not as good dispersed as the GNPs in the polymer foam (photographs in the left side of Figure 2b,c), their effect on the flame retardancy was as efficient or even more efficient than the GNPs. GO's benefits in the fire retardancy efficacy over graphene have been referred to and are attributed to the GO oxygen functionalities that decompose and dehydrate at quite low temperatures. This causes a cool down of the polymer substrate during the combustion process and simultaneously release gaseous species that dilute the oxygen atmosphere near the ignition zone [30,39,40]. Graphene and GO are described mainly as co-flame retardants, at loadings from 1.0 to 10 wt% [41]. Also, functionalized graphene designed and prepared from expandable graphite and phosphorus-containing compounds has been described as flame retardants and smoke suppressor of PUF at 6.1 wt% [30]. It is worth mentioning that in our study, the pristine PUF and PUF nanocomposites with 1 wt% of nanofillers also kept their shape after burning. However, due to the better results obtained for 2.5 and 5.0 wt%, the PUF with 1.0 wt% of nanofillers were excluded in the further characterization results.
The thermogravimetric (TG) and derivative thermogravimetric (DTG) curves are shown in Figure S3a,b, respectively. The pristine PUF and PUF nanocomposites showed similar TGA curves, suggesting that GNPs and GO do not significantly influence the decomposition of PUF. This finding is most likely related to the small amount of nanofillers used. The initial decomposition temperature corresponding to 5% (T 5% ), and 50% (T 50% ) of mass degradation, the maximum-rate degradation temperature (T max ), and mass residue at 750 • C are listed in Table 1. The results show that GNPs had a positive effect on T 5% on PUF, while GO was detrimental to the early thermal stability. These results are caused by the earlier degradation of the GO oxygen functionalities at low temperature, thus accelerating the degradation of the PUF matrix. The oxygen functionalities of GO may interact with the PUF precursors during the foam formation thus interfering with the crosslinking reaction, as reported by Gama et al. [19]. In fact, at high temperatures there is the complete burning of GO, proved by the lower percentage of residue at 750 • C [42]. The sample PUF/GNPs2.5 presents the higher thermal stability with T 50% of 404 • C. It was reported by Liu et al. [43] that graphene can act as a heat source and accelerate the decomposition of PUF. As so, the 2.5 wt% seems to have a more positive effect on the thermal properties of PUF than 5.0 wt%. show an inhomogeneous open-cell structure composed by quasi-spherical interconnected pores for all specimens. However, depending on the presence or absence of nanofillers, some small differences were noticed. As shown in Figure 3a insets, the GNPs and GO sheets are located, and sometimes wrapped, in the PUF cell walls. It is reported that carbon nanostructures have a nucleating effect during foam formation, thus altering the PUF morphology [44]. Usually, the presence of such fillers decreases the average cell size and increase foam density improving damping properties, flame-retardancy, and mechanical properties [45]. By comparing PUF/xGNPs with PUF/xGO SEM images, GO seems to promote thicker cell walls and joints which can be related with the agglomeration of GO or from the affinity of polyol and MDI with GO. The GO nanosheets are located between adjacent cavities of the foam, and the cell wall sticks together around them, producing thicker cell walls. This effect is accentuated with the increase in the quantity of GO: PUF/GO5.0 presented bigger pore size and thicker cell walls than PUF/GO2.5. On the contrary, the increase of GNP content in the PUF promotes a pore size decrease.

Mechanical Properties
The mechanical properties of the foams, the apparent density, the compressive strength at 10% of compression, and the compressive modulus are summarized in Table 2. The results indicate that the foams' densities increase with either GO and GNPs loading. In addition, the compressive strengths and compressive modulus steadily grow with the increase of both nanofillers content. It is reported that the addition of nanofillers to cellular materials can have different effects depending on several aspects, the loading content, the size and shape, the compatibility between the nanofillers and the matrix [33,46]. For the same fillers loading, the GO provides higher compressive modulus and higher resistivity to undergo load without total collapse of pores (higher stress plateau) than the GNPs. From the SEM images it was observed that GO induced thicker cell walls, while PUF/GNPs were similar to PUF. The improvement of compressive modulus and strength by the addition of carbon nanostructures to PUF was also reported by other experimental works, where phosphorus-

Mechanical Properties
The mechanical properties of the foams, the apparent density, the compressive strength at 10% of compression, and the compressive modulus are summarized in Table 2. The results indicate that the foams' densities increase with either GO and GNPs loading. In addition, the compressive strengths and compressive modulus steadily grow with the increase of both nanofillers content. It is reported that the addition of nanofillers to cellular materials can have different effects depending on several aspects, the loading content, the size and shape, the compatibility between the nanofillers and the matrix [33,46]. For the same fillers loading, the GO provides higher compressive modulus and higher resistivity to undergo load without total collapse of pores (higher stress plateau) than the GNPs. From the SEM images it was observed that GO induced thicker cell walls, while PUF/GNPs were similar to PUF. The improvement of compressive modulus and strength by the addition of carbon nanostructures to PUF was also reported by other experimental works, where phosphorus-functionalized GO [30] and expandable graphite [19] were used. Furthermore, MDI has isocyanate terminal groups (NCO) that can react with carboxylic and hydroxyl groups of GO leading to the chemical cross-linking between GO and PU matrix via the formation of amides or carbamate esters, thus improving the mechanical response [47].
From the average stress-strain compressive curves (Figure 4a), the foams exhibit the typical behavior of cellular materials. An elastic region at low strain values (5% of strain), where the stress-strain curve is linear, followed by a near constant stress until 55-60% of strain, is designated the stress plateau. Finally, densification takes place, with the complete collapse of the cells and the formation of a compact material, like a thin film, which is characterized by an abrupt increase of stress.

Thermal Conductivity
To evaluate the possible application of these types of foams as thermal insulation materials, their thermal conductivity was evaluated (Table 3). A good thermal insulation material should present low thermal conductivity. The results obtained for all specimens (0.035-0.037 W•m −1 •K −1 ) are in the range of most widely used commercial insulation materials, 0.030-0.040 W•m −1 •K −1 as reported in the literature [18,52,53]. The thermal diffusivity decreases with the addition of carbon nanostructures, except for the PUF/GO5.0, as listed on Table 3. This could be due to the barrier effect created by the nanofillers that difficult the heat flow.
From the combination of different materials, it was expected to create different multifunctional structures with high strength and reduced weight as a result of the synergetic effect of the individual materials which could be applied in different fields. In this sense, following the characterization of PUF nanocomposites regarding their fire retardancy, mechanic, acoustic, and thermal insulator properties, the formulation PUF/GNPs2.5 was selected to be incorporated in the Al-OC skeleton as a filling material and denoted as PU/GNPs2.5-OC. However, it is worth mentioning that all the developed PUF nanocomposites were suitable to be incorporated inside Al-OC.  Figure 4b shows the sound absorption coefficient between 100 and 4000 Hz for the PUF and PUF with 2.5 and 5.0 wt% of GNPs and GO. Globally, the sound absorption coefficient increases with frequency until 1500 Hz, followed by a small decrease and finally a subsequent stabilization between 2000 and 4000 Hz. The sound absorption coefficient is strongly influenced by the morphology of the foams, namely the density (associated to cell-wall thickness) and pore features (size, quantity, interconnectivity, tortuosity) [48]. It was reported that a smaller interconnected pore structure gives a better sound absorption coefficient due to the high airflow resistivity provided by cell walls [49]. Furthermore, higher porosity, associated with low density offers less resistance to sound-wave dissipation which results in a low sound absorption coefficient. Many factors contribute or influence the sound absorption, and therefore the overall values are a balance of all factors [17,48,50,51]. The incorporation of GNPs and GO increases the compressive modulus (with the later presenting higher values), and thus the cells have greater propension to undergo cell stretching, bending, and buckling without deformation. Also, by SEM and µCT analysis it was observed that the incorporation of GNPs and GO fillers decreases the porosity and pore size. The PUF/GO2.5 has the higher sound absorption curve between 1250 and 1750 Hz, reaching the value of 1 and absorbing more than approximately 50% than PUF. In fact, between 1000 and 1750 Hz, all the PUF nanocomposites have superior sound absorption coefficient comparatively to pristine PUF, which covers the sensitive frequency region of the human ear [34]. One parameter often used to describe sound absorption is the noise reduction coefficient (NRC). This parameter corresponds to the average of the sound absorption coefficients at the octave bands of 250, 500, 1000, and 2000 Hz, and rounding the result to the nearest multiple of 0.05. However, the use of NRC has its limitations since equal NRC values do not necessarily translate the same curve profile. Thus, it is worthwhile to use all the available data, especially for sound absorption at frequencies below 250 Hz or above 2000 Hz. The NRC values are gathered in Figure 4b and results showed that, although PU/GO2.5 and PU/GO5.0 have the same NRC that PU/GNPs5.0, they have distinct profiles of sound absorption vs. frequency, in this case, with PU/GO2.5 having better performance between 1200-2000 Hz. Gama et al. [17] obtained similar NRC values around 0.40-0.45 for PUF with a similar density (around 40 kg/m 3 ) and a higher thickness (approximately 40 mm).

Thermal Conductivity
To evaluate the possible application of these types of foams as thermal insulation materials, their thermal conductivity was evaluated (Table 3). A good thermal insulation material should present low thermal conductivity. The results obtained for all specimens (0.035-0.037 W·m −1 ·K −1 ) are in the range of most widely used commercial insulation materials, 0.030-0.040 W·m −1 ·K −1 as reported in the literature [18,52,53]. The thermal diffusivity decreases with the addition of carbon nanostructures, except for the PUF/GO5.0, as listed on Table 3. This could be due to the barrier effect created by the nanofillers that difficult the heat flow. Table 3. Thermal conductivity, thermal diffusivity, and specific heat (n = 5). From the combination of different materials, it was expected to create different multifunctional structures with high strength and reduced weight as a result of the synergetic effect of the individual materials which could be applied in different fields. In this sense, following the characterization of PUF nanocomposites regarding their fire retardancy, mechanic, acoustic, and thermal insulator properties, the formulation PUF/GNPs2.5 was selected to be incorporated in the Al-OC skeleton as a filling material and denoted as PU/GNPs2.5-OC. However, it is worth mentioning that all the developed PUF nanocomposites were suitable to be incorporated inside Al-OC.

Hybrid Structures
The effect of filling the voids of Al-OC foams with bulky polymers has been reported [10][11][12]54], showing enhanced mechanical properties due to the pore filling which improves the compressive strength and energy absorption capacity of the hybrid foams. However, the use of bulky polymers to fill the voids of the Al-OC foams can be disadvantageous when lightweight structures are required. Here, we wanted to explore the properties of the hybrid structure resulting from filling the Al-OC foam with a porous one, thus not compromising the lightness of the final structure. An easy and simple process to prepare this hybrid was followed (Figure 1, route B).
3.2.1. Structure   Figure 5a,b show the Al-OC and the PU/GNPs2.5-OC photographs, respectively, and in Figure 5c a 3D reconstruction of the hybrid structure obtained by µCT analysis is presented. It is worth mentioning that the weight ratio of Al-OC and PUF/GNPs 2.5 in the hybrid structure was 48.2% and 51.8%, respectively. It was observed that, during the PUF nanocomposite expansion, when in contact with Al-OC structs, some of the foam cells started to collapse. Because of that, the PUF/GNPs2.5 structure inside the Al-OC foam presented a different porosity and morphology from the optimized one. Although the pores were not uniform and the control of the porosity during the foaming was not possible, it ensured that the procedure were performed under the same operating conditions. In addition, the hybrid structures were reproducible as a whole, taking into account not only the filling material but the also the Al-OC. Though no chemical bonding between the metal surface and polymer was observed or even expected, an excellent form-fitting connection of the PUF/GNPs2.5 to the rough metal surface was visible by direct observation. The PUF/GNPs2.5 density was increased in the vicinity of the metal surface, visible in the color scheme (green color, Figure 5c). fill the voids of the Al-OC foams can be disadvantageous when lightweight structures are required. Here, we wanted to explore the properties of the hybrid structure resulting from filling the Al-OC foam with a porous one, thus not compromising the lightness of the final structure. An easy and simple process to prepare this hybrid was followed (Figure 1, route B).
3.2.1. Structure   Figure 5a,b show the Al-OC and the PU/GNPs2.5-OC photographs, respectively, and in Figure  5c a 3D reconstruction of the hybrid structure obtained by µCT analysis is presented. It is worth mentioning that the weight ratio of Al-OC and PUF/GNPs 2.5 in the hybrid structure was 48.2% and 51.8%, respectively. It was observed that, during the PUF nanocomposite expansion, when in contact with Al-OC structs, some of the foam cells started to collapse. Because of that, the PUF/GNPs2.5 structure inside the Al-OC foam presented a different porosity and morphology from the optimized one. Although the pores were not uniform and the control of the porosity during the foaming was not possible, it ensured that the procedure were performed under the same operating conditions. In addition, the hybrid structures were reproducible as a whole, taking into account not only the filling material but the also the Al-OC. Though no chemical bonding between the metal surface and polymer was observed or even expected, an excellent form-fitting connection of the PUF/GNPs2.5 to the rough metal surface was visible by direct observation. The PUF/GNPs2.5 density was increased in the vicinity of the metal surface, visible in the color scheme (green color, Figure 5c).

Mechanical Properties
The stress-strain curve is shown in Figure 6a and the energy absorption (EAD) and specific energy absorption (SEA) are illustrated in Figure S4. As described earlier for pristine PUF and PUF nanocomposites, these hybrid structures also present the typical behavior of foams, with three distinct regions: elastic, stress plateau, and densification [10]. Comparing the PU/GNPs2.5-OC compressive response with Al-OC, a similar behavior can be observed; however, with higher stress peak and stress plateau values and the densification occurring earlier. The EAD is higher for hybrid structures, with improvements of 27% comparative to the Al-OC foam and 13 times lower than the PUF/GNPs2.5. However, the SEA is lower, suggesting that the increase in EAD does not compensate the increase in weight, as found by Reinfried [14], that combined a steel open-cell foam with an expanded polystyrene foam. The values of apparent density, stress peak, EAD and SEA, and thermal conductivity are gathered in Table 4.

Mechanical Properties
The stress-strain curve is shown in Figure 6a and the energy absorption (EAD) and specific energy absorption (SEA) are illustrated in Figure S4. As described earlier for pristine PUF and PUF nanocomposites, these hybrid structures also present the typical behavior of foams, with three distinct regions: elastic, stress plateau, and densification [10]. Comparing the PU/GNPs2.5-OC compressive response with Al-OC, a similar behavior can be observed; however, with higher stress peak and stress plateau values and the densification occurring earlier. The EAD is higher for hybrid structures, with improvements of 27% comparative to the Al-OC foam and 13 times lower than the PUF/GNPs2.5. However, the SEA is lower, suggesting that the increase in EAD does not compensate the increase in weight, as found by Reinfried [14], that combined a steel open-cell foam with an expanded polystyrene foam. The values of apparent density, stress peak, EAD and SEA, and thermal conductivity are gathered in Table 4.
Contrary to the hybrid structures composed by the Al-OC impregnated with dense polymers, epoxy [12] and polydimethylsiloxane [11], whose compressive mechanical behavior is governed by the dense materials, in this case it is the metal skeleton (Al-OC) that controls the compressive response under loading, as reported in a similar study describing the Al-OC filled with bacterial cellulose nanocomposite foam [15]. However, it should be noted that the weight of these structures is much lighter, and a good compromise between weight and strength can be achieved.   Contrary to the hybrid structures composed by the Al-OC impregnated with dense polymers, epoxy [12] and polydimethylsiloxane [11], whose compressive mechanical behavior is governed by the dense materials, in this case it is the metal skeleton (Al-OC) that controls the compressive response under loading, as reported in a similar study describing the Al-OC filled with bacterial cellulose nanocomposite foam [15]. However, it should be noted that the weight of these structures is much lighter, and a good compromise between weight and strength can be achieved.

Sound Absorption
The sound absorption ability of the hybrid structure (evaluated by the sound absorption coefficient) present higher sound absorption values (peak at 1200 Hz with a value of 0.8 and followed by a nearly constant value of 0.7 for higher frequencies) (Figure 6b) when compared with the filling material. The increase of stiffness provided by the Al-OC structure can improve the absorption at low frequencies. Increased sound absorption can be obtained for thick specimens [55,56]. The NRC value for PUF/GNPs2.5-OC was 0.45, three times higher than for pristine Al-OC. These values are comparable with other cellular materials reported in the literature with similar thickness (around 22 mm) and density (120 kg/m 3 ) [57].

Thermal Conductivity
The values of thermal conductivity obtained for the hybrid PUF/GNPs2.5-OC are similar to those obtained for PUF and its nanocomposites, around 0.038 (W·m −1 ·K −1 ), suggesting that Al-OC contribution is negligible in the overall thermal conductivity value. This may be due to the small (about 1 mm) layer of filler foam (PUF and PUF/GNPs2.5) surrounding the metal foam that blocks the heat transfer. Another aspect is related to the highly porous structure of Al-OC, so the amount of solid material (Al-SiO3) that would effectively increase the conductivity value is very small. The same trend was observed in our previous works [11,12,15], in which the contribution of Al-OC was small regarding the filling materials. The polymer filler (e.g., polydimethylsiloxane [11], epoxy [12] and cellulose nanocomposites [15]) acts as an insulator, preventing effective heat transfer. On the other hand, some authors [58,59] have demonstrated that the open-cell metal foams can be used to enhance the low thermal conductivity of the pure phase change materials (PCM), composite PCMs and paraffin for application to many situations (e.g., latent heat thermal energy storage system, heat sink and heat exchanger). Furthermore, Fiedler et al. reported [60] that the polymeric adhesives exhibit a low thermal conductivity, and thus form thermal barriers between advanced pore morphology (APM) foam elements (sphere-like closed-cell aluminum foam), resulting a distinct decrease of the thermal conductivity of adhesively bonded APMs. Globally, hybrid PUF/GNPs2.5-OC can be used as a thermal insulation material [18,56].

Conclusions
In the present study, hybrid structures were prepared by impregnating an Al-OC foam with other cellular material, a PUF nanocomposite. Prior to the incorporation into the open structure of Al-OC foams, the standalone PUF nanocomposites were prepared with different amounts of GNPs and GO and fully characterized. The presence of the carbon nanostructures in the PUF nanocomposites provided PUF with excellent fire retardancy, better mechanical strength, and thermal and acoustic insulating properties. The formulation with 2.5 wt% of GNPs was considered the more promising one. The process of filling the Al-OC open structure with the optimized PUF nanocomposite was successfully achieved, and the resulting hybrid structure maintain a low thermal conductivity value (0.038 W·m −1 ·K −1 ), high sound absorption coefficient specially at mid to high frequencies (NRC 0.45) and better mechanical behavior (the stress corresponding to 10% of deformation was improved in more than ten times). Therefore, these hybrid structures are thoroughly multifunctional materials with potential applications in the construction, automotive and, aeronautic sectors.