4.1. Wide and Small Angle X-ray Scattering (WAXS/SAXS) and Optical Polarized Microscopy (POM)
WAXS spectra of the milled cellulose (C), the milled walnut shell (WS) and of the diatomite (D), used as fillers in the preparation of polyurethane composite foams, are shown in
Figure 1a.
The WAXS profile of the C shows the presence of two very broad halos (which are almost collapsed in only one broad halo) at 2θ values of ~19.57° and 15.28°, suggesting that the used C is characterized by a high degree of amorphous phase and only few and small crystals are present in a mixture of the Form I and Form II [
19]. Two better-resolved diffraction peaks at 2θ values of 20.45° and 15.05° are observed in the WAXS spectrum of powders of the WS (
Figure 1a), revealing the existence of crystals in the Form I and II of the cellulose and a higher crystallinity than the C.
For the diatomite (D), all diffraction peaks detected in the corresponding WAXS profile can be assigned by the attendance of different polymorphic forms of silica [
12,
23].
In order to study the influence of the different natural fillers on the crystallization process of the polyurethane foam, WAXS measurement were carried out on pristine PUG and on composite systems PUG-C, PUG-WS and PUG-D (
Figure 1b).
The diffraction profiles of the pristine PUG and of the composite foams PUG-C and PUG-WS are characterized only by a broad halo centered at 2θ ≈ 19°, indicating that the systems are amorphous. The WAXS profile of the sample PUG-D shows a shoulder at 2θ = 21.6°, which is related to the appearance of a low crystallinity of the polyurethane (PU) [
24]. In fact, since the maximum diffraction peak of the diatomite is located at 2θ = 19.8° (
Figure 1b), the shoulder in the WAXS profile of the PUG-D can be attributed only to the crystallinity of the polymer. This result is also in agreement with the presence of spherulites, with a size of about 20–30 µm, in the POM image of the sample PUG-D (
Figure 1c). Similar spherulitic structures were not found in the POM images of the pristine PUG (
Figure 1d) and of the composite foams PUG-C and PUG-WS (data not shown). Probably, PUG, PUG-C, and PUG-WS systems that appear amorphous from WAXS analysis, are characterized by very small crystalline entities, which are difficult to detect by POM observations. Therefore, WAXS and POM data of the foams suggest that the different fillers participate in different manners on the crystallization process of the polyurethane and in particular, the diatomite induces an increase of the crystallinity acting as a nucleating agent on the crystallization of the polymer.
Furthermore, in the WAXS profiles of the PUG-C, PUG-WS and PUG-D, a shift is observed at slightly higher 2θ values of the amorphous halo compared to the 2θ value of the pristine PUG (from 2θ = 18.76° for the PUG to 19.02°, 19.09° and 19.35° for the PUG-C, PUG-WS and PUG-D, respectively). This effect is ascribed to a modification of the local microstructure of the polyurethane network, which likely arises from the ability of cellulose, walnut shell and diatomite to become involved in the building up of the organic network through crosslinking reactions among -OH groups of the fillers and -NCO groups of the isocyanate [
25].
The polyurethane composite foams were further investigated by SAXS experiments in order to understand how the interaction of the different fillers with the PU network intervenes on the PU morphology, and finally, to correlate the PU morphology to the final properties (such as thermal and mechanical properties) of the systems.
SAXS analysis has been widely used for studying PU morphology [
26,
27,
28,
29,
30] because these materials are usually random segmented copolymers of alternating flexible soft-segments (SS) and more rigid hard-segments (HS) containing urethane and/or urea domains. Hence, PUs typically show a microphases-separated morphology where the different electronic densities among SS and HS is responsible of the broad knee scattering feature present in their SAXS patterns. Thus, the interdomain distance between HS (
L) can be estimated by using the Bragg’s law:
where
qmax is the maximum of the intensity of the scattering vector (
q), determined from the Lorentz corrected plot (
I(
q)·
q2 −
q). Moreover, the electron densities difference can be quantitatively evaluated by the invariant of scattering
Q as follows:
The value of
Q describes the electron density fluctuation of polymer and is a good approximation to estimate the overall degree of phase separation (DPS) in PU [
24,
29].
SAXS profiles of the polyurethane composite foams and the corresponding Lorentz corrected plots are reported in
Figure 2, while the values of
qmax,
L and
Q are listed in
Table 2.
The samples PUG-C and PUG-WS show values of
L of 5.32 nm and 5.15 nm, respectively, which are higher than the
L value of 4.42 nm of the pristine PUG. Usually, an increase of the interdomain distance between HS is also related to an increase of the size of the domains and to an increase of the
Q value. As reported in
Table 2, the
Q values for the PUG-C and PUG-WS are slightly lower than the
Q value found for the PUG. These results can be reasonable with the formation of some urea and/or urethane hard domains in the PUG-C and PUG-WS, which are not incorporated with the polymer network [
28], causing a disruption of the long range connectivity between the HS. These unattached urea/urethane-HS do not contribute to the phase separation degree in the polyurethane, because they are isolated domains and are more mobile than those within the PU microdomains, justifying the slightly decrease of the
Q values and the increment of the
L values of polyurethane composite foams containing C and WS as fillers.
SAXS analysis of the PUG-D system (
Figure 2 and
Table 2) demonstrates that the diatomite plays a completely different role from those observed for cellulose and walnut shell in the morphology of the PU. In fact, the presence of the D induces an increase of the interdomain distance (from 4.42 nm for the PUG to 4.76 nm for the PUG-D) and also an appreciable increment of the invariant of scattering
Q (from 0.022 for the PUG to 0.102 for the PUG-D) of the polymer. Therefore, the addition of the D as filler in the polyurethane foam induces a growth of the HS that participate to the polymer network, resulting in an improvement in the degree of the microphase separation of the PU.
4.2. Infrared Spectroscopy (FTIR)
The effect of filler addition on chemical structure of polyurethane foams was also evaluated by FTIR-ATR investigation. The main absorption vibration peaks of polyurethane composite foams (PUG-C, PUG-W and PUG-D) and pristine PUG are listed in
Table 3. All the spectra highlight the disappearance of the characteristic NCO vibration peak (around 2270 cm
−1) associated to the isocyanate consumption inherent to the urethanic linkage formation [
10,
11,
30]. The characteristics band features of the mainly functional groups (urethanes and urea linkages), absorbed in the spectral region 1400–1800 cm
−1, are analyzed and discussed and the main deconvoluted peaks are shown in
Figure 3. For instance, as widely investigated [
10,
24,
30,
31] in the 1600–1800 cm
−1 wavenumber region, the carbonyl vibration bands are correlated to the urethane and urea linkages,
H-bonded (
H-C=O) or free C=O (
f-C=O) as summarized in the
Table 3. Typically,
H-C=O linkages are due to the HS domains present in the polyurethane structure formed between NH groups proton accepting (urethane or urea carbonyl), while the
f-C=O linkages can exist dispersed in Soft segment (SS) or in interphase domains (urea phase).
In this region, the pristine PUG highlighted a stretching vibration peak related to
H-C=O urethane linkages at 1716 cm
−1; for the composite foams (PUG-C, PUG-W and PUG-D), the carbonyl region is split into two peaks. The
H-C=O band of urethane disappears and the
f-C=O correlated to urethane (around 1724–1722 cm
−1) and urea (around 1677–1673 cm
−1) linkages display (see the deconvoluted spectra in
Figure 2) [
31]. It was hypothesized that, the addition of fillers resulted in the decrease of the isocyanate (NCO) available to react with the polyol (to form HS) due to the reactions that can be occurred between the OH groups of fillers with NCO creating covalent-like bonds to the PU molecular chains. This leads to a decreasing of urethane density (reduction of
H-C=O in urethane linkages) and to the formation of new urethane chains (
f-C=O) and urea domains (
f-C=O) more flexible [
30,
31,
32].
Furthermore, the urethane domains for rigid polyurethane foams were also identified through the vibrational modes, νas C-N and δ N-H, around 1511 cm−1 for PUG samples, at a lower wavenumber, 1509 cm−1, for the PUG-C and PUG-W. An increase in wavenumber for the PUG-D (1516 cm−1) was observed. This means that the addition of filler C and W induces a decreasing of networking in polyurethane structure, while the addition of D acts as cross-linking agent increasing the hardness of polyurethane chemical structure.
Apart from this, the spectra of PUG-C and PUG-W exhibit additional stretching vibration bands, around 1528–1539 cm−1, related to the NH of isolated urea domains (interphase) located in the polyurethane matrix that induce a reduction of the continuity of urethane structure.
4.4. Thermal Properties (Thermo-Gravimetric Analysis-TGA, Differential Scanning Calorimetric-DSC and Thermal Conductivity)
TGA analysis has been used as a tool to characterize the thermal degradation properties of polyurethane foams [
29], it is important to point out that all the produced foams were manufactured without the addition of flame retardants. TGA results of produced samples are listed in
Table 4. The thermograms of PUG and PUG composite foams show several degradation steps. The first degradation step occurs, for all polyurethane foam samples, in the range 150–220 °C (maximum of the derivative weight loss, Tmax1 around 220 °C) with a weight loss equal to 6wt% attributed to the elimination of bound water molecules. The second step which occurs in the temperature range of 250–350 °C (with Tmax2 equal to 311) is ascribed to the breaking of urethane links: the isocyanate dimerises to form carbodiimide with evolution of volatiles compounds as CO2, alcohols, amines etc. The carbodiimide is stable up to higher temperature when it degrades giving rise to the subsequently degradation steps (Tmax3).
As observed in
Table 4, the maximum decomposition temperatures (Tmax) of the foams increased with the addition of filler. The enhancement of thermal stability for PUG-C and PUG-W can be ascribed to the high thermal stability of all fillers [
12,
34,
35].
Furthermore, it is important to point out that for the PUG-WS and PUG-D a marked increase of the char (around 21 and 22.4 wt%, respectively) at high temperature (800 °C) with respect to the pristine PUG was detected. For the PUG-WS, this result could be ascribed to the presence of lignin in the walnut shell composition. For the PUG-D the further enhancement can be attributed to the presence of diatomite (inorganic) domains, which are able to reduce the amount of more-combustible organic molecules and to produce a siliceous layer that inhibits any mass transfer [
12,
34,
35,
36] producing more char component.
Finally, the higher amount of char, produced during the pyrolysis of composite foam, can potentially inhibit flame propagation through the substrate of polyurethane.
The thermal conductivity values for the produced foams range around 31 mW/mK, while for the sample PUG-D (diatomite) a reduction of the thermal conductivity, 27 mW/mK, with respect to the other foams, was recorded. This result is consistent with the ones reported in the literature for foam composites filled with preformed aerogel particles; in this case, the addition of diatomite reduces the thermal conductivity of the foams thanks to the intrinsic nanoporosity in its morphological structure. This behavior can be explained considering that the diatomite (located within the cell strut of the foam, see the micrographs in
Figure 4b) due to its nanometric porosity (
Figure 4b) includes voids into the solid phase of the cell-walls thus inducing a reduction of the radiative contribution λr of thermal conductivity. Generally, the thermal conductivity of a porous material can be described as a combination of different specific terms:
where:
λ is the apparent thermal conductivity of the insulation material,
λs is the thermal conduction through the solid structure of the pore walls,
λg is the conduction through the gas-filled inside the pores,
λr is the heat radiation through the cell walls,
λc is the convention within the cells, that is negligible.
The diatomite is able to reduce the λr parameter.
DSC thermograms are reported in
Figure 5 and the resulting data are summarized in
Table 5. As observed in
Figure 5, all the foams exhibit a glass transition temperature (T
g) around 26 °C for the pristine PUG and a cold crystallization temperature (T
cc) at 1.34 °C ascribed to the organized microphase (urea “nuclei”) present within the polyurethane foams.
By analyzing the DSC data of the composite foams, a slight decreasing in Tg and an increase of Tcc were recorded. Furthermore, an increase of the enthalpy values correlated to the cold crystallization phenomenon was mainly found for the PUG-C and PUG-W. These outcomes could be explained considering that cellulose, walnut shell (in particular) and diatomite both act as reactive fillers, subtracting isocyanate groups form the polyurethane reaction (because of the reaction of their OH groups with NCO), and thus, causing a decrease in the cross-linking and consequently of Tg, and as nucleating agents leading to the increasing of Tcc and urea formation (ΔH). These confirm the SAXS/WAXS and FTIR analysis.
4.5. Mechanical Properties
In order to evaluate the mechanical properties of PUG, PUG-C, PUG-WS and PUG-D polyurethane foams, compression tests were performed as well. The characteristic stress-strain curves of the analyzed materials are reported in
Figure 6a. Each foam presents an elastomeric behavior, in which it is possible to appreciate three different regimes: linear elasticity (LE), elastic buckling (plateau, EB) and densification (DE) [
37,
38,
39]. Depending on the adopted filler, those ranges present different limits. In the cyan area, the empty symbols on the curve represent the LE-EB transition point experienced by each material. On the other hand, in the light-green area the empty symbols are representative of the EB-DE transition points. The mechanical properties and the mechanical behavior of the polyurethane foam matrices are affected by the presence of cellulose, walnut shell and diatomite fillers, each in different ways, as detailed hereafter.
The LE part of the curve is mostly correlated to the intrinsic bending properties of the bulk material that constitutes the cell walls, which is initially deformed when a compression force is applied [
37,
38,
39]. In the inset of the
Figure 6a, it is possible to better appreciate the difference between the slopes of the initial linear part of each curve.
In the case of cellulose and walnut shell, the presence of the filler has a detrimental effect, in terms of elastic modulus and extent of the LE regime with respect to the pristine PUG foams. This can be ascribed to the chemical interaction of the aforementioned fillers with the PUG matrix, whose network results re-constructed, as also stated by SAXS-WAXS analysis. This kind of interaction has also induced the presence of cross-linked network with different crystalline domains, due to the presence of un-linked urea moieties (the urea is not bonded to the urethane polymer structure). Instead, when diatomite is added, the LE regime has the shortest length, while the elastic modulus results in the longest length. This is strictly related to this kind of filler, which is ceramic and rigid. Thanks to the presence of some OH groups (due to the few amount of Si-(OH) groups present in the chemical structure of diatomite), which could interact with the polyurethane structure (through isocyanate group) and alter the structure of the material (see the chemical characterization section), the D filler enhances the stiffness of the material constituting the cell walls, which is then well dispersed within the polymer matrix. This effect endows the bulk material with a completely new crystalline structure, involving the traditional polyurethanes cross-linking network and the different silica allotropic crystals of the diatomite, as stated by SAXS analysis.
The second part of the curve (EB) is characterized by the presence of a plateau. This behavior appears when the cell walls and edges of the foam, after being bent, start to distort, the foam cells appear to be twisted and the material loses its stiffness [
37,
38,
39]. With respect to the PUG material, the others exhibit a transition point from LE regime to the EB one at lower stress and deformation values. From the EB portion of the stress-strain curve of the polyurethanes it is possible to evaluate the so-called Plateau Modulus, which gives information on the chemical structure of the material constituting the cell walls, and not on the foam density [
37,
38,
39].
From the curves, it is possible to notice that the presence of the two strongly chemically interacting fillers (cellulose and walnut shells) produces materials with lower Plateau Modulus with respect to PUG (
Table 6), with PUG-C having a lower slope than PUG-WS. In case of PUG-D, the extent of EB regime is quite similar to the neat material, further proving that the diatomite filler does not alter the chemical structure of the PUG, since it is not seizing much reactive sites of the isocyanate, as cellulose and walnut shells do instead. In this area, the stress at 10% deformation (σ10) is evaluated. As expected, PUG-C and PUG-WS has lower values, while PUG and PUG-D has practically the same value. These results are also in agreement with X-ray analyses, where different cross-linked networks were detected.
After the EB regime, the transition to DE occurs. In this last area, the slope of the curve starts to increase, and the stress escalates. Here, the opposing foam cells walls touch, being almost completely collapsed. If further strain compresses the material, the elastic modulus of the solid itself could be approached [
37]. In this study, the strain stops at 50% of deformation, so that the DE regime is just at its beginning.
Even though the foam cell structure can be appreciated from the SEM analysis, the direct correlation between the cell walls or edges structure and mechanical properties is better elucidated by use of the Gibson-Ashby model [
37,
38]. For this model, relative young’s modulus and relative density must be considered. The first is given by the ratio between the elastic modulus of the foamed material (E) and the elastic modulus of the material constituting the cell walls (Es), while the second is the ratio between the foam density (ρ) and the density of the cell wall material (ρs), which was evaluated empirically, following the formula [
40]:
In Equation (4),
Wm and
ρm are the weight fraction and the density of the matrix, while
Wf and
ρf are the weight fraction and the density of the filler, respectively. In our case, the fillers’ densities (from data sheets) are
ρC = 1500 kg/m
3,
ρD = 2300 kg/m
3 and
ρWS = 640 kg/m
3, while the density of non-foamed PU is equal to 1200 kg/m
3. The relationship between relative young’s modulus and relative density follows an exponential trend, as for the equation in
Figure 6b, where the exponential n is a factor depending on the cell structure [
39].
In
Figure 6b, the Relative Young’s Modulus and Relative Density of the selected materials (empty symbols) are plotted. From these values, it is possible to obtain the n value at which the Gibson-Ashby model can be applied, and thus acquire information about the cell walls structure. For n = 2 this model is representative of the ideal behavior of the so-called open-cell polymer foam, in which only cell edges, and no cell walls, are present in the foam arrangement. Despite that most polyurethanes foams exhibit closed cells, their behavior follows the open-cell foams one, as in the cases studied here. This is related to the fact that, during manufacturing, the surface tension, which drives the formation of the cells foam, tends to let the material gather in the cell edges, leaving the cell walls thinned. Then, depending on the used filler, the behavior of each set of samples can be described by the Gibson-Ashby model, using different values of the exponent.
Therefore, in case of PUG, the exponent n results to be equal to 2.23, meaning that even if the material appears to have closed cells (
Figure 4a), the thickness of the cell walls is lower than the one of the edges. In case of PUG-C and PUG-WS, the exponent n is slightly higher than that observed for PUG. This means that the material structure results more similar to the closed cells one, attributable to the distribution of the urea domains not only on the cell edges, but also on the cell walls. Despite that the Young’s Modulus of PUG-WS results were higher than the PUG one, the materials’ LE-EB transition point was actually lowered, as also pointed out by the σ-ε curves, since the urea domains should be located on the cell edges to contribute to the stiffness and to the strength of the material. However, in each case, the cell walls are so weak that they could be considered empty. For PUG-D, where the exponential n has the lowest value, the diatomite acts as a filler, whose high stiffness is exploited in the inclusion in the polymer matrix constituting the cell edges, which will then result as the real supporting column of the foam structure.
4.6. Comparison of Some Key Parameters of Selected Rigid Polyurethane Foams
In
Figure 7, a simple comparison of the key parameters that influence the overall impacts in the construction field (thermal conductivity, density and compressive strength evaluated at 10% of deformation) of composite PU foams produced in this study with those of selected conventional PUs (data from ANPE:
https://www.poliuretano.it/EN/index.html, and technical data sheet from Huntsman:
https://www.huntsman.com/polyurethanes/a/Products, PPUF [
41] all foams were produced by using cyclopentane as blowing agent) and bio-based PUs (data from the recent literature: PU and B-RPF60 foams produced by replacing 40 wt% and 100 wt% of the synthetic polyol with bio-based polyol [
42,
43] respectively; NIPU_1 and NIPU_2 non-isocyanate foams [
44]) is reported.
As observed in
Figure 7, the replacement of a petrochemical precursors with renewable feedstock (mainly when the polyol is totally replacement, sample B-RPF60) in the synthesis of rigid PUR foams affect negatively the functional (increasing of λ for B-RPF60) and mechanical properties (the samples became brittle [
41,
42,
43,
44]) of produced foams with respect to the conventional PUs. Based on the literature, these outcomes could be correlated to the different cellular structure, of the porous material, induced by the complex chemical structure of the bio-based polyols.
Instead, in our case, the thermal conductivity of the PUG-D complies with that of commercial PUs, while the mechanical performance is superior, probably due to the higher density. For this reason they can, potentially, be used in the construction sector as thermal insulators.