Production and characterisation of self-blowing lignin-based foams

Industrial kraft lignin powder was supplied by Stora Enso, Helsinki, Finland. Furfuryl alcohol (FA, 98%), glyoxal (40 wt%), Tween 85 and polyvinyl alcohol (PVA) were purchased from Sigma-Aldrich and used as received. Diethyl ether (DE, 98%) and sulphuric acid (H₂SO₄, 98 wt.%) used as blowing agent and catalyst were purchased from CHEMSOLUTE, Th. Geyer GmbH & Co. KG, Renningen Germany.

Self-blown lignin-furanic foams were prepared in various formulations (Table 1). First, known weights of lignin, furfuryl alcohol, glyoxal (40 wt%), and Tween 85 were thoroughly mixed in a beaker under mechanical stirring for 5 min at 500 rpm until a homogenous mixture was obtained. The blowing agent diethyl ether was then added and the mixture was stirred vigorously for 30 s at increased speed. Lastly, diluted sulphuric acid (32%) was added in an amount of 8.1% of the formulation, and the blend was mixed rapidly for 10 s. In order to ascertain the reaction between FA and lignin, a neat PFA matrix was produced using the LF1 formulation without lignin and with 6% surfactant and 3% blowing agent.

For some foam samples, aqueous PVA (in 8.3% solution) was added to the foam mixture before adding the catalyst and blowing agent in an amount equivalent to 5–10% of the dry weight of lignin in the formulation. The resulting mixture was immediately poured into an open silicon mould (10 × 10 × 10 cm) and left under the fume cupboard for the foaming process to occur. The foams obtained were black, viscous and sticky at first but stabilised and hardened at room temperature after about 30 min. The foams were then placed in an oven at 70 °C for 24 h to fully cure. After curing and cooling, the foams were removed from the mould and stored at ambient temperature for several days before characterisation. During this period, the residual blowing agent evaporated.

The apparent density of the foams ranged from 80 to 320 kg m⁻³. It decreased steadily with increasing diethyl ether content (Fig. 2a). To obtain a rigid lignin-furanic foam (Fig. 3) that does not collapse during expansion and curing, the balance of components is crucial. Therefore, several experimental trials were conducted to monitor the effect of the proportions of FA, lignin, and other components on the foaming process. A mixture containing a higher proportion of liquid components than solid components allows for an even distribution of the solid components and an easier quantitative transfer into the foam mould. The influence of FA on the density of the lignin-furanic foams is noticeable as foams with the highest proportion of FA (LF3) gave the highest density range. Lignin foams had homogenous densities (Fig. 2b). FA is a critical component in the formulation as it influences the heat generated during self-condensation. The observations from the foaming process are consistent with a previous study, which showed that FA could act as a co-monomer and a cross-linker in reaction with polyol and an aldehyde such as glyoxal (Delgado-Sánchez et al. 2018a). The acid catalyst initiates the exothermic auto-condensation reaction of FA. Here, the amount of sulphuric acid was adjusted to increase the temperature of the system to a level sufficient for expansion. When concentrated sulphuric acid was used, a violent and rapid reaction which led to the formation of a densed non-expanded resin was observed.

Furthermore, the lignin proportion influenced the blend's viscosity and nucleation period. Nucleation, for example, the formation of isolated bubbles in the mixture, could already be observed during foam preparation. The cell microstructure is formed at this initial stage, and polymerisation is initiated in addition to a series of chemical reactions (Lacoste et al. 2013). This period is crucial for the foaming process. With a higher lignin content than FA, the result was a highly viscous mixture that was difficult to stir. In this viscous mixture, nucleation was restricted, and the foam made was of relatively higher density. The positive aspect of a highly viscous mixture is that the expansion occurs in a slow and controlled manner (Tondi et al. 2016). However, foams produced from a highly viscous mixture tend to possess low porosity, which may increase their thermal conductivity—an essential property for thermal insulating application (Delgado-Sánchez et al. 2018a). If, on the other hand, a higher proportion of FA was used, the result was a mixture with lower viscosity. With lower viscosity, a shorter mixing time was needed to achieve homogeneity. A similar observation in the production of tannin foams was reported (Link et al. 2011). To control the viscosity of the mixture, water was added to the formulation in most studies on tannin foams. Here, adding water led to dilution, slowing down the polymerisation and expansion. Therefore, a component rebalancing was needed to achieve the desired foaming behaviour.

Lignin-furanic foams produced without glyoxal were brittle and could not be processed for further characterisation. This was similar to formaldehyde-free tannin foams, which had low mechanical properties (Link et al. 2011) and underlines the importance of glyoxal as a cross-linker in foam formulation. Furthermore, the density of foam produced with Tween 85 increased slightly with increasing surfactant content. This trend stabilises as the surfactant content rises further. A recent study on the production of tannin-based foams reported 2–5 times increase in the density of the foam prepared with Tween-80 and sodium dodecyl sulfate (SDS) (Sepperer et al. 2021). Adding PVA to the foam formulation slightly influenced the density of the foam. The densities of LF4 samples prepared with 5%, 10%, and 15% aqueous PVA were 181.8 ± 0 190.0 ± 0, and 186.8 ± 0 kg m⁻³ (mean value ± standard deviation), respectively. Based on density, lignin-furanic foams prepared in this study can be classified as low-foaming foams (Jin et al. 2019; Basso et al. 2013).

The compression curves of some of the lignin-furanic foams (Fig. 4a) showed a very low linear elastic region at low strain up to about 7%, a stress-plateau region representing cell collapse and densification due to the breakage under increased stress (Nielsen 1966). These curves are similar to those reported for furanic-glyoxal foams reinforced with olive powder (Xi et al. 2020). The shape of a compression curve depends on a material's elasticity or brittle nature (Li et al. 2006). Some lignin-furanic foams were brittle and cracked easily under stress to generate fragments. Thus, they did not yield compression curves that would have been suitable for further evaluation. The average compression strength was highest in lignin-furanic foams produced with the highest proportion of FA (0.48 ± 0.02 MPa). The compression strengths of lignin-furanic foams increased in direct proportion to the density (Fig. 4b). Lignin-furanic foams prepared with the addition of 5, 10 and 15% PVA exhibited compression strengths of 0.18 ± 0.01, 0.17 ± 0.07, and 0.21 ± 0.01 MPa, respectively. Without the addition of PVA, the foam had a compression strength of 0.21 ± 0.01 MPa. The compression strengths of lignin-furanic foams were lower than those reported for tannin-based foams (Wu et al. 2020; Tondi et al. 2009b) but fall within the average value reported for lignin-furanic foams produced with magnesium bisulfite liquor (0.25–0.35 MPa) (Tondi et al. 2016). The difference in compression strength in tannin or lignin-based foams may be attributed to the difference in the interaction of the two aromatic polymers with FA and glyoxal because this interaction determines the strength of the internal cross-linking of the foams (Lacoste et al. 2013; Tondi et al. 2009a). In this regard, tannin is more reactive than lignin (Tondi et al. 2016) as it exhibits more aromatic sites for hydroxyalkylation. The trend toward higher compression strength of foams based on magnesium bisulfite lignin compared to kraft lignin (this study) may be attributed to the water solubility of the former lignin, which might allow for a more homogeneous reaction. Enhancement of lignin reactivity might be achieved through modification processes such as amination or nitration (Laurichesse and Averous 2014).

The relative water absorption (%) of lignin-furanic foams decreased with an increase in foam density (Fig. 5). There was a steady increase in water absorption of lignin-furanic foams within the first 5 h of water immersion. After that, the water absorption increased slightly at a very slow rate. Foams with comparable density values (157 and 161 kg m⁻³) showed similar water absorption. The water absorption of lignin-furanic foams corresponds to that of rigid polyurethane foam (Thirumal et al. 2008). Lignin-furanic foams reinforced with PVA have about 15% higher relative water absorption irrespective of density, probably due to the hydrophilic character of PVA.

In the leaching test, the foams lost between 5 and 11% of their dry weight (Fig. 6). The mass loss increased with an increase in the FA content. PFA lost 9% of its dry weight. These loses are attributed to some unreacted furfuryl alcohol, lignin, surfactant and acid recovered (Sepperer et al. 2021). Kraft lignin lost 35% of the dry weight during the leaching.

Lignin-furanic foams have FTIR spectra comprising bands of the polycondensation complex formed from the acid-catalysed condensation of FA (Fig. 7, Table 2). Kraft lignin is characterised by a broad and intense peak at 3410 cm⁻¹, representing O–H stretching vibrations (Mansouri et al. 2011). The peaks at 2932 and 2837 cm⁻¹ represent C–H stretching vibrations in the methyl and methylene groups. The conjugated C=O stretching of ester, aldehyde and unconjugated ketones appeared at 1728–1704 cm⁻¹. The bands in the 1594–1477 cm⁻¹ region were assigned to lignin aromatic skeletal vibrations. C–O stretch of guaiacyl (G) ring and syringyl (S) ring were noticeable at 1246–1211 cm⁻¹ and 1081 cm⁻¹, respectively. The small peak at 811 cm⁻¹ represents the C–H deformation at the aromatic rings (out-of-plane) (Faix 1991). Furthermore, OH-stretching band is typical for furfuryl alcohol between 3600 and 3000 cm⁻¹ (Burket et al. 2006). The reduction of the OH band in the spectra of the foam (LF1) suggests that the condensation of KL with PFA hydroxyl may have occurred at the free position on the aromatic ring of KL (Choura et al. 1996). The spectrum of LF1 can be further explained by the overlapping of the two spectra of Lignin and PFA, suggesting that eventual new bonding between the 2 constituents is absorbing in the same spectral area as the constituents itself. For comparison, PFA prepared using similar formulation without KL was analysed with FTIR. From the analysis of spectrum, it is clear that the broad band that characterizes OH group in FA is not present in the PFA spectrum. PFA has a complex structural property that can only be accurately elucidated with multiple analytical tools. Nonetheless the peaks in the PFA spectrum represent the products of the intermolecular rearrangements after the acid-catalysed polymerisation of FA. In addition to the affected OH groups, the C–O stretch, guaiacyl (G) ring at 1246–1211 cm⁻¹ was absent in LF1 suggesting that the reactive group of the G unit was occupied by the other groups. The presence of polyaromatic signals in the foam was indicated by several peaks. The peaks between 2921 and 2850 cm⁻¹ are assigned to the furan CH₂-stretching vibration (Xi et al. 2020; Guigo et al. 2007), that at 1013 cm⁻¹ to OH vibration, as well as those at 1508 and 778 cm⁻¹ to C = C stretching vibration from the aromatic furan ring (Burket et al. 2006), while asymmetric CH₂ bending in PFA was assigned at 1421 cm⁻¹ (Tondi et al. 2019). C–O stretching of a lactone ring in α, β-unsaturated γ-lactones was at 1355 and 1310 cm⁻¹ (Wewerka 1971). The distinct signal at 1710 cm⁻¹ is of a Diels–Alder rearrangement product and the peaks between 1149 and 1145 cm⁻¹ represent C–C furan stretching in plane wagging (Tondi et al. 2019).

The thermogravimetric analysis (TGA) curve shows the weight loss as a function of temperature, while the first derivative of the TG curve (DTG) shows the corresponding weight loss rate (Fig. 8). A small weight loss of kraft lignin around 109 °C corresponds to the loss of water through evaporation. DTG thermograms of kraft lignin showed that the thermal behaviour ranged from 200 to 600 °C. The temperature of the highest degradation rate of kraft lignin was around 384 °C. Above 400 °C, several competing reactions are known to occur, including the release of monomeric phenols and decomposition of the aromatic ring (Sahoo et al. 2011). The residual mass of the initial kraft lignin at the end of pyrolysis was 28.6%. The peak at 234 °C in DTG curve of PFA is attributed to the volatiles produced as by-products of the condensation polymerisation. The major peaks between 424 and 510 °C result from the degradation of methylene groups and the furan ring in the PFA chain (Xu et al. 2022).

Lignin-furanic foams showed similar thermal behaviour to kraft lignin, albeit with higher residual mass. The first weight loss region in lignin-furanic foams was noticeable between 54 and 330 °C, which is attributed to the vaporisation of volatile gases, including moisture and blowing agent in the foams (Burket et al. 2006). The lignin-furanic foams' fastest decomposition and major weight loss occurred between 330 and 600 °C. In this temperature range, the weight loss of the foam is about 30–40% depending on the composition of the foam. This weight loss is due to the degradation of large polymer network structures into small fragments (Ng et al. 2022). Due to the formation of carbonaceous residues, low additional weight loss of 5–10% occurs above 680 °C (Burket et al. 2006). The thermal parameters of lignin-furanic foams with similar composition but a different proportion of diethyl ether indicate that the temperature at 5% mass reduction (T₅) increased with an increasing amount of diethyl ether from 149.2 °C (2° DE) to 153 °C (7% DE). The T₁₀, T₃₀ and T₄₀ temperatures did not follow the same trend. The difference in the residual mass was also noticeable. The addition of PVA affected the thermodegradation profile of the foams. For example, foams made with 5% PVA (LF4 – 5%) showed a faster degradation rate at about 209 °C than foams made without PVA (LF3). This is attributed to the presence of water in the formulation. An increase in the PVA content from 10 to 15% had no distinguishing effect on the thermal behaviour. All the studied lignin-furanic foams have a residual mass between 40 and 45%, which was about 25% higher than that of KL. This shows that the PFA polymerised with KL to form a thermally stable complex cross-linked structure, resulting in the heat resistance of the foams.

To provide insights into the flammability of lignin-furanic foams, representative specimens were characterised using the cone calorimetry technique. A cone calorimeter provides similar real-world fire conditions. It generates a quantitative evaluation of material’s flammability in terms of parameters such as time to ignition (TTI), total heat released (THR), heat release rate (HRR), residual mass, and smoke. Thus, it is widely used in this field. For this study, TTI, THR, HRR and residual mass were used to evaluate the flammability of lignin-furanic foams.

All tested lignin-furanic foam samples showed similar combustion characteristics (Fig. 9a–c). The foam samples caught fire and cracked after ignition but self-extinguish in air within about 45 s after the test (Fig. 10). The TTI decreased as the FA content and the density of the foams increased (Table 3). LF1 was most resistant to ignition. Previous studies have shown that properties such as density and thermal conductivity can influence the TTI and the time taken to self-extinguish (Ira et al. 2020; Patel et al. 2011). The TTI of the lignin-furanic foams is higher than the TTI of rigid polyurethane (Checchin et al. 1999; Günther et al. 2018; Usta 2012), wood-fiber insulation boards (Lee et al. 2019), pure epoxy foam (Auad et al. 2007) and expanded polystyrene foam (Chen et al. 2012). This suggests that lignin-furanic foams are not easy to ignite compared to these materials. Furanic-glyoxal foams produced without lignin showed similar attributes (Xi et al. 2020). Nonetheless, the ability of lignin-furanic foams prepared without additives to self-extinguish after exposure to fire is similar to the fire behaviour of tannin-based rigid foam (Tondi et al. 2009b) and can be a result of the chemical structure of the cured resin (Guigo et al. 2007) and the ability of lignin to delay the burning process (Podkościelna et al. 2020). However, tannin-based foams have also been reported to have a TTI value around 100 s and thus a higher fire resistance (Delgado-Sánchez et al. 2018b).

The heat release rate (HRR) describes the heat released by the burning foam as a function of the fire produced and the flame spread during the burning process. The HRR curves of the investigated samples have a peak between 170 and 197 kW m⁻² between 210 s (LF1) and 325 s(LF3) and then attenuate (Fig. 8a). These peaks remained constant for about 50–60 s before decreasing to level up at about 45 kW m⁻². An earlier peak (125 kW m⁻²) observed at about 60 s can be attributed to the formation of a protective char (Usta 2012), while the highest peak heat release rate (PHRR) was due to the damage of the surface char residue leading to the further combustion of combustibles. The PHRR of lignin-furanic foams declines as the FA content increases. Lignin-furanic foam with the highest FA content also took the longest time (325 s) to reach the PHRR. Compared to rigid polyurethane foams (RPUF) (Günther et al. 2018), expanded polystyrene foam (Chen et al. 2012) and pure epoxy foam (Auad et al. 2007), lignin-furanic foams have lower PHRR. However, tannin-based foams (Li et al. 2012b) and wood–fiber insulation boards (Amaral-Labat et al. 2013) have lower PHHR than lignin-furanic foams. One may suggest that the weaker reaction between lignin, glyoxal, and furfuryl alcohol compared to the reaction with tannin may be responsible for the lower fire retardancy of lignin-furanic foams (Kolbitsch et al. 2012). It is anticipated that additives such as aluminium hypophosphate, ammonium polyphosphate and related compounds will reduce the HRR and improve the fire-retardancy of lignin-furanic foams (Liu et al. 2022). In addition, further variation in the foam composition and resulting properties may positively impact the fire retardance.

The total heat release (THR) represents the amount of heat generated from a specimen during combustion. The THR of lignin-furanic foams averaged 40–50 MJ m⁻². These values are greater than the THR of tannin-based foams (Delgado-Sánchez et al. 2018b). The chemical component of foams can potentially promote char formation on the material, thereby preventing heat transfer and flame spread (Usta 2012). The Fire Performance Index (FPI) was calculated as the ratio of TTI to PHRR. It indicates the safety rank of the materials. The higher the FPI, the higher the safety rank of the materials. The FPI of lignin-furanic foams decreased with decreasing TTI. LF1 containing the lowest FA amount showed the highest FPI, which confirms the fire resistance of the foam. Furthermore, the residue left after the cone calorimeter tests (% RM) was between 43% (LF3) and 33% (LF1). The % RM values fall within the range obtained from the thermogravimetric analysis.