Modified raw cellulose filaments material as polyol substitute in rigid insulating polyurethane foam | Scientific Reports

Scientific Reports volume 15, Article number: 6934 (2025) Cite this article

1652 Accesses

1 Altmetric

Metrics details

Current research in building insulation is primarily focused on enhancing the performance of polyurethane foam or exploring alternatives with biobased materials, with particular attention placed on polyols. This study investigated the use of modified cellulose filaments as a polyol to enhance the environmental performance of polyurethane foam. Two distinct etherification methods were employed to modify cellulose filaments (CFs), yielding accessible and reactive ether functions from the hydroxyl (OH) groups of CFs. Polyols and the resulting polyurethane foam were characterized. Kinetics of foam formation, morphology, density, thermal conductivity, and mechanical properties in compression were studied. Analysis revealed reduced reactivity with modified CFs as a substitute of the petroleum-based polyol, affecting foaming. Impacts on the resulting properties of the foam were observed, such as the size of the cells (from 0.14 ± 0.06 mm2 for petroleum-based polyol foams to 0.03 ± 0.03 mm2 for foams with 70% substitution with biobased polyols 1 and 2), and cells opening (from 92 ± 2% for petroleum-based polyol foams to 8 ± 3% with 70% substitution with biobased polyols). These results lead to non-compliance with the canadian polyurethane foam standard, requiring a closed cell rate of over 90%. A deterioration in mechanical properties through loss of stiffness and a drastic reduction in the maximum strength (yield strength) the material can withstand below the required standard were also measured. However, noteworthy conductivity results were obtained (0.041 ± 0.004 W m−1‧K−1 with 70% of substitution with biobased polyols 1 and 2). Foam properties were partly due to different polyol properties, such as functionality and viscosity.

The chemical and petrochemical industries are the leading consumers of industrial energy, representing 28% of global energy consumption in 2014, and the third largest industry subsector in direct CO2 emissions1,2. In the spirit of the 2 °C trajectory of the COP 21 agreement, the industry must limit its annual increase in process energy consumption, considering an anticipated 40% increase in the demand for primary chemicals from 2017 to 20503. Particular attention should be given to biobased materials to mitigate environmentally harmful emissions4.

In parallel, the construction and building sector contributes for 30–40% of the world’s total energy consumption4,5. The envelope structure significantly influences the building’s carbon footprint, necessitating improvements if the Net Zero Emissions (NZE) threshold is to be reached by 20505,6. The building envelope design plays a crucial role in defining heating and cooling requirements, guaranteeing comfort, indoor air quality, and the safety of occupants. Thus, numerous studies seek to improve building insulation or find alternatives using biobased materials7,8,9,10,11,12. Considering its notable performance characteristics, the scientific community is exploring methods to produce low environmental impact polyurethane foams such as biobased ones13,14,15,16,17.

Poly methylene diphenyl diisocyanate (pMDI) is the predominant choice of isocyanates for preparing polyurethane spray foams thanks to its high reactivity partly due to its aromatic structure18. So far, none of the biobased alternatives presented in the literature have been recognized as truly competitive, especially regarding chemical reactivity19. Hence, research focuses primarily on polyols, the other main compound that forms insulating polyurethane foam.

Several studies have examined the feasibility of easily modifying bioresources to substitute chemicals. Biobased raw materials such as soybean20,21,22,23, chitosan24, vegetable oils25, fatty-acid oil26, palm oil27, castor oil14, and microalgae oil17 have been considered as raw material for bio-based polyol synthesis in polyurethane foams28. The use of soy has given rise to products currently on the market with 22% recycled plastic and renewable soybean oil29,30. It represents a relatively limited amount of biobased content.

The traditional use of pulp and paper is declining, necessitating the exploration of new applications to support the forest industry sector31,32. The use of biobased materials such as cellulose from variable sources has been studied for the development of many materials, including polyurethane foams, due to the abundant availability of these biobased polymers33. Incorporating CFs into the construction sector holds promise for extending CFs durability and mitigating the environmental footprint associated with polyurethane foams32.

Interest has first been directed towards lignocellulosic resources as a filler13,34,35. However, the use of cellulose as a mere filler has demonstrated its limitations. When used in excessive quantities (5%wt of polyol), cellulose deteriorates the material’s properties, allowing only a small proportion of cellulose to be integrated. These results lead to minimal improvements in properties and low substitution of petrochemical components. It shows that using cellulose as a filler is insufficient to reduce the environmental impact of polyurethane foams and that focusing on substituting the petroleum-based products contained in the foams is still necessary.

Cellulose can be easily modified through its high hydroxyl (OH) functional groups content. Cellulose liquefaction with acid in various solvents, such as glycerol and polyethylene glycol has been studied for biobased polyols15,28,36,37. Another strategy consists of grafting easily accessible functions of interest through the OH functions of cellulose16,38,39.

Combining these two methods can result in the creation of cellulose-based polyol. It has been studied through the conversion of cellulose to furfural structures and is a promising avenue for the development of biobased materials. It is crucial to simplify the process of converting cellulose to polyol for future industrialization40. However, while cellulose has shown promise as a polyol in polyurethane foams in the past, the conductivity and closed-cell content, which are essential properties required to comply with both Canadian and European standards, have not been systematically measured15,16,38,41,42,43.

This study explores the potential of CFs modification as a sustainable polyol solution for enhancing the environmental performance of polyurethane foam. For this purpose, CFs have been modified following two different etherification protocols: the first one uses CFs liquefaction in glycerol in the presence of sulfuric acid, and the second grafts glycidol and ethylene carbonate to generate available and reactive ether functions from the CF’s OH groups38,41. Glycerol and glycidol are already available as biobased products, making them valuable for transitioning towards greener materials44. Polyol properties were studied, and their properties in the polyurethane foams were tested at variable percentages through foam formation kinetics, foam morphology, closed cell content, thermal conductivity, and mechanical performances.

The CFs used in this study were produced by Kruger Biomaterials Inc. (Montreal, Canada). Their manufacturing involves longitudinally peeling the wood pulp filaments, specifically Northern bleached softwood kraft (NBSK), to maintain their length, decrease their diameter, and obtain lignin-free CFs. This process achieves a nearly 100% yield without using chemicals or enzymes, eliminating the need for effluent treatment. This product is composed entirely of cellulose I and is both biodegradable and compostable45,46.

As the DP affects the properties of the resulting polyol, the DP of CFs was calculated according to the TAPPI 230 OM 08 method to confirm the effectiveness of the hydrolysis process47. This analysis involves measuring the flow time of a cupriethylene diamine solution (0.5 M) in which cellulose is dissolved. The longer the CFs, the more viscous the solution is, and the longer its flow time is. This analysis was also conducted on the hydrolyzed CFs used for the polyol 2 (P2) synthesis.

CF-based polyol was synthesized by liquefaction of CFs with crude glycerol with a mass ratio solvent to CFs of 10:1 in the spirit of Kosmela et al.41. The reaction was performed at 150 °C for 6 h in the presence of sulfuric acid as a catalyst (at 0.15%wt of CFs and glycerol in the media, from a 5% solution of sulfuric acid). The obtained polyol was neutralized with an aqueous sodium hydroxide solution (NaOH) at 50%, then at 2%, and oven-dried at 103 °C until constant mass.

CFs were first hydrolyzed, resulting in a lower DP and allowing a better solubilization in glycidol. Following the procedure of Szpilyk et al.16, CFs’ hydrolysis was performed in an Erlenmeyer flask in an autoclave at 120 °C for 90 min in the presence of a 20% sulfuric acid solution and 8%wt of CFs. The flask was then placed into an ice and water bath to quench the hydrolysis. The solution was neutralized with NaOH from the most concentrated to the least concentrated state (pellets, followed by NaOH at 6 mol L− 1 and then at 0.5 mol L− 1) until reaching a pH of 7. The neutralized mixture underwent three rounds of water washing through centrifugation, allowing the separation of CFs from the sodium salt that resulted from the neutralization process of the CFs solution. The resulting CFs was dried at 100 °C for 12 h.

The polyol synthesis following the hydrolysis was performed in a 2 L glass reactor equipped with a reflux condenser, mechanical stirrer, and thermometer, following Szpilyk et al. 2021 protocol38. The reaction was monitored by mass loss due to ethylene carbonate (EC) consumption under FTIR analysis16. The resulting polyols were oven-dried at 100 °C until constant mass.

The polyols compositions were studied to ensure that the reaction proceeded as expected and to understand better the properties linked to the chemical structures.

The weight average molecular weight (Mw) and number average molecular weight (Mn) were evaluated by gel permeation chromatography (GPC). GPC analysis were conducted in N, N-dimethylformamide (DMF), with 0.1% lithium bromide (LiBr), allowing better solubilization with 0.02 mL min−1 flow, 300 µL capacity loop, 80 °C temperature, and standard polystyrene reference.

Nuclear magnetic resonance (NMR) 1H spectra were obtained on an Inova model spectrometer from Agilent (Santa Clara, USA). The analysis was performed with a proton resonance frequency of 400 MHz, equipped with a 5 mm ATB probe for 1H NMR study with DMSO as the solvent.

In order to follow the grafting reaction of ethylene carbonate in P2 and analyze the final product of P1 and P2, FTIR spectroscopy was conducted with an INVENIO-R FTIR (Bruker Optics Inc., Billerica, Massachusetts, USA) in the range of 4000–400 cm− 1 in attenuated total reflectance (ATR) mode with 32 scans and a resolution of 4 cm− 1.

Polyols and formulations viscosity measurements were conducted at a temperature of 22 °C using a Bohlin Visco88 viscometer (Malvern Instruments Limited, Worcestershire, United Kingdom) featuring a concentric cylindrical design comprising a rotating inner cylinder of 14 mm diameter and a stationary outer cylinder with a diameter of 15.4 mm. Approximately 10 mL of the polyol sample was introduced into the viscometer cylinder. The viscosity was recorded at 20, 35, 61, 107, 187, 327, 572, and 1000 rpm, employing a torque within the 0.5 to 9.5 mN m range to ensure precise measurements. P2 was found to have a very high viscosity, making it impossible to measure viscosity using the available equipment. As a result, only the viscosity of the polyol blends was measured with the Bohlin Visco88 viscometer. The viscosity of the petroleum-based polyol was measured and labeled as the reference. The viscosity of the petroleum-based polyol with 2.5% of raw CFs was also measured and used as a threshold not to be exceeded by the polyol blends.

Products and quantities used to prepare the polyurethane foam formulations are detailed in Table 1 as well as their suppliers. Table 2 provides the details of the isocyanate used in polyurethane foams, and Table 3 summarizes the information on the commercial polyol reacting in the polyurethane foams. Table 4 details the polyols used in each formulation.

Polyol proportion in the formulations is detailed in Table 4. The amount of each chemical introduced in the formulation is based on the total polyol weight and expressed in parts per hundred of polyol (php) in Table 1. All products were added and homogenized in the formulation using a RZR 1 (Caframo, Wiarton, Ontario, Canada) mechanical stirrer at 2500 rpm at room temperature.

For simplification, petroleum-based polyol is referred to as R, the CFs/glycerol-based polyol (etherpolyol) as P1, and the CFs/glycidol/ethylene carbonate-based polyol (etherpolyol) as P2 throughout the article.

Formulations were stored at 4 °C overnight to stabilize the HFO blowing agent in the formulation as it tends to evaporate. An excess of HFO was added to ensure its concentration remained superior to 15 php. Then, pMDI was added and mechanically stirred at 2500 rpm until reaching the cream-time (change of the visual aspect before expanding). The foam was then poured into a 946 mL paper cup, and rigid polyurethane foams were obtained within a few seconds. The methods undertaken to calculate the OH number in polyols synthesized from CFs did not enable this parameter to be determined due to the poor solubility of polyols. Foam formation tests revealed similar reaction behavior with the isocyanate between the synthesized polyols and glycerol alone. Therefore, it was approximated that the OH number of the synthesized polyols corresponds approximately to the OH number of glycerol. The pMDI index (moles of NCO groups/moles of OH groups) was therefore adjusted to 1.13 for each foam by approximating biobased polyol parameters with glycerol values. The foams were kept under ambient conditions for one week after preparation to allow the structure to stabilize. The preparation protocol is summarized in Fig. 1.

Summary of the polyurethane foam formation protocol and illustration of temperature recording with thermocouples.

Unlike the foams industrially applied in buildings, the foam samples were prepared in paper cups instead of being sprayed. It can significantly influence material properties and lead to properties different from those of spray-prepared foams, as highlighted by Hawkins et al.34,48. However, similar trends can be expected with substituting petroleum-based polyol (R) with a CF-based polyol in the foam.

All samples were conditioned at 23 ± 2 °C and 43 ± 3% relative humidity (RH) for 24 h before testing. The samples were cut using a bench saw, and the dimensions of all samples were determined using a digital caliper with a precision of ± 0.1 mm. A minimum of three samples per cup were used for each analysis to ensure a representative measurement regardless of the portion taken from the cup.

Samples from three different panels supplied by a sprayed polyurethane foam manufacturer of the same commercial reference product were also analyzed to facilitate a comparison between the foams developed in this project and commercial products and standards. For this purpose, it was decided that the panels’ properties were representative of the products available on the spray polyurethane foam market.

Statistical analyses were conducted to monitor the reaction kinetics, closed cell content, density, thermal conductivity, and mechanical properties in compression tests, varying the CF-based polyol content in the foam with 19 replicates. R software (R Foundation for Statistical Computing, Vienna, Austria) was used for statistical analyses49. The ggsignif package was employed to design boxplots and provide statistical information50.

The test residues did not adhere to a normal distribution. Moreover, the presence of extreme values may indicate the fat-tailed distribution characteristic of the phenomenon under study51. So, the Kruskal–Wallis test, a non-parametric version of ANOVA that does not assume the normality of residuals, was employed52.

The null hypothesis was rejected with a low p-value (p-value < 0.05), stating that the test’s median is the same for all percentage levels. This indicates a significant association between the proportion of the percentage factor and a difference in the median values of test variables. It is worth mentioning that the median is used instead of the mean due to the non-parametric nature of the Kruskal–Wallis test, which, while indicating a statistically significant difference, does not specify which specific groups differ from each other. Dunn Post-hoc tests were then conducted to allow pair comparisons and the identification of significative statistical differences between groups.

To study the proper progress of the chemical reaction, the temperature was monitored at various points during foam formation, as indicated by the red dots in Fig. 1. It was accomplished using four thermocouples linked to a computer. At the outset of foam expansion, the formulation was in contact with the first thermocouple positioned at the base of the paper cup. As the reaction advanced and the foam expanded, it reached the other thermocouples. The highest temperature among the four thermocouples was monitored over time, and temperature recordings started once it surpassed the ambient temperature. A total of nineteen samples were analyzed for each type of foam in this study (R, R + P1 50%, R + P1 70%, R + P1 + P2 50%, R + P1 + P2 70%).

The microstructure of the foam samples, aligned with the direction of expansion, was examined by scanning electron microscopy using a FEI Quanta 250 microscope (FEI Company Inc. Thermo-Fisher Scientific, Hillsboro, OR, USA). Before imaging, the samples were coated with a gold–platinum alloy to enhance their electrical conductivity and stabilize them, allowing them to withstand the conditions of high vacuum and high-energy electron beams. An acceleration voltage of 15 kV was applied for all analyses.

Cell size was determined by using area data acquired through ImageJ software on SEM images captured on the surface perpendicular to the foaming direction of the foam at 100× magnification. A total of two captures were studied on two foams for each CF-based polyol percentage studied, for a minimum of 25 cells observed per image captured by SEM. All cells present in the SEM images were measured, and a minimum of three images per sample type were studied.

Foam density is essential for determining foam performance as it is closely connected to the ratio of open and closed cells, thermal conductivity, and mechanical properties53. Foam density was calculated from the weight and geometric volume of cubic specimens of 25.4 mm side length in accordance with ASTM D1622 (2020)54. A total of nineteen cup samples were analyzed for each foam type in this study (R, R + P1 50%, R + P1 70%, R + P1 + P2 50%, R + P1 + P2 70%), with a minimum of six samples per cup.

The foam’s closed and open cell content significantly impacts the material’s thermal insulation characteristics55. The percentage of open and closed cells was determined using a gas pycnometer (UltraPyc 1200e, Quantachrome, Boynton Beach, FL, USA) in the spirit of ASTM D6226 (2021) following the protocol of Beaufils-Marquet et al. (2023) study34,56. A total of nineteen cup samples were analyzed for each foam type in this study (R, R + P1 50%, R + P1 70%, R + P1 + P2 50%, R + P1 + P2 70%), with a minimum of three samples per cup.

The thermal conductivity of insulation materials was measured using a custom-built small-scale device developed in compliance with ASTM E1225 (2020). This device allows for thermal conductivity measurements using smaller samples than those required by the heat flow meter described by the standard ASTM E1225 (2020). The specimens were cut to 50 × 50 mm2 dimensions, with a thickness of 12.25 mm aligned with the foam’s expansion direction. Before the thermal conductivity was tested, the thickness was measured using a digital caliper. Each sample was placed between thin aluminum sheets and plates with controlled temperatures of 35.5 °C (bottom plate) and 10.5 °C (top plate) for an average temperature of 23 °C and ΔT of 25 °C between the plates. A weight was applied to the system to ensure proper contact between the components of the measuring device and the sample. The temperature was maintained stable by water-cooled Pelletier plates (Model K20, Haake, Vreden, Germany), while the equilibrium heat flux was measured using a PHFS-01 heat flux sensor (Flux Teq LLC, Blacksburg, VA, USA). The k values reported were calculated via Fourier’s law, with k the thermal conductivity (in W m− 1 K− 1), Q the heat flux (in W m− 2 K− 1), L the size of the material (in m), and ΔT the temperature difference (in K):

The thermal conductivity of the foams was determined by analyzing three samples per foam batch and 19 batches for each CF-based polyol content considered.

The compressive characteristics of the foams were measured using a universal testing machine QTest/5 Elite Controller (MTS, Eden Prairie, MN, USA) equipped with a 5 kN load cell and operating at a constant 2.5 mm min− 1 loading rate. The specimens were cut into 50.8 × 50.8 mm2 pieces with a thickness of 25.4 mm running parallel to the direction of foam expansion. All the tests were performed in the spirit of ASTM D1621 (2016).

To accommodate more specimens from each foam sample, a thickness of 25 mm was used instead of the 50 ± 3 mm suggested by the standard57. The compression tests were conducted parallel to the direction of foam expansion, following the specified test procedure. The testing machine recorded the strength applied during the tests on the 19 cup samples for each foam type in this study (R, R + P1 50%, R + P1 70%, R + P1 + P2 50%, R + P1 + P2 70%), with a minimum of three samples per cup.

Young’s modulus was calculated by subjecting the stress-strain curve from the testing machine to data processing. It involved employing the ggpmisc R package to plot the slope, enabling data verification58.

Thermogravimetric analysis (TGA) was performed to allow comparison in terms of the thermal stability of the foam in the presence of biobased polyol. It was performed on a TGA 851e analyzer (Mettler Toledo, Greifense, Switzerland). Experiments were carried out in triplicate under a nitrogen flow (50 mL min− 1) from 25 to 800 °C at a heating rate of 10 °C min− 1. Characteristic data such as the temperature corresponding to 5% of degradation (T5%), the temperatures of the maximum degradation rates (Tmax), and the percentages of residues at 800 °C were determined.

The DP defines the length of a polymer chain. It represents the degree of polymerization and indicates In the case of cellulose ((C6H10O5)n), the number of glucose units in the chain59. The DP influences the properties of CF-based materials, especially mechanical ones60. It depends on crystallinity, cellulose extraction method, or characteristics of the bioresource used61. The DPv of CFs was compared to that of Avicel, a commercially used carboxymethyl cellulose (CMC), serving as a reference.

Various measures of DP exist, including DPn (weighted mean of mole fraction), DPv (measured by viscometry, related to average volume or size of macromolecules in a dilute solution), DPw (measured in Size Exclusion Chromatography, related to weighted mean of weight fraction).

In this study, CFs’ DPv (related to average volume or size of macromolecules in a dilute solution) was found to be 593 ± 9, which is comparable to wood nanofibrils (evaluated between 410 and 1100)62. This value is consistent with the weighted mean of mole fraction DPn = 280 and the DP related to weighted mean of weight fraction, DPw = 1910 of NBSK pulp reported in the literature63. Avicel typically has a DPv of 161 ± 1, and literature reports a DPn of 21864. In comparison, hydrolyzed CFs leading to P2, had a DPv evaluated to 69 ± 1. It can, therefore, be concluded that hydrolysis has indeed reduced the chain length of CFs from more than three times Avicel’s at half Avicel’s length. The resulting DPv is lower than for many cellulose sources65. This variation in DP can impact the properties of the resulting polyol polymeric chain, including mechanical properties60,66. In view of the low DP obtained after CFs hydrolysis for P2 synthesis, it is possible that the remaining CFs are solely or predominantly crystalline CFs. This result was confirmed by FTIR analysis in polyol properties section. It could affect the accessibility of the chemical functions or the mechanical properties of the resulting polyol.

The molecular weight of a polymer strongly influences various properties of the resulting polyurethane foam, such as viscosity, elasticity, or tensile strength, and thermal stability. Thus, a polyol of low molecular weight likely results in hard plastics, while high molecular weight polyols will tend to give flexible elastomers55. It underlines the importance of establishing this characteristic since it determines all the properties of the resulting polyurethane foam. The retention time of the raw materials was measured (Fig. 2). Subsequently, the resulting polyols’ molecular weights were determined and summarized in Fig. 2; Table 5, where Mn characterizes the number average molecular weight and Mw the weight average molecular weight.

Gel permeation chromatography results for P1 and P2 and raw materials.

In P1, a low Mn and Mw with low polydispersity was observed, which appears consistent with the retention time of glycerol (8.8 min). It suggests that the cellulose molecules might be too large to solubilize and pass through the filter, resulting in the detection of glycerol alone. Thus, it is suspected that the GPC analysis of the P1 sample primarily detected glycerol. Thus, no conclusions about the molecular weight of P1 can be drawn.

The molecular weight increase impacts the mechanical performance of the resulting product (maximum strength and Young’s modulus), the Tg, and the viscosity55. P2 has the highest viscosity and exhibits a higher molecular weight, higher than the product synthesized by Szpiłyk et al.16. They obtained two molecular weight fractions, Mn of 9850 and 1120 and Mw of 21,080 and 1470, resulting in polydispersities of 2.1 and 1.3, respectively. The second Mn approaches the one observed in our study (1120 g mol− 1 against 1200 g mol− 1 in P2). However, the Mw in this study is considerably higher than that reported in the literature (almost two times higher than the first Mw measured in Szpiłyk et al. (2021) and 26 times higher than the second Mw measured in Szpiłyk et al. (2021)), leading to a high polydispersity.

In Szpiłyk et al. (2021) study, cellulose was hydrolyzed upstream of polyol synthesis with glycidol and ethylene carbonate16. This hydrolysis allows the polyol to pass through the filter compared to P1. Moreover, in our case, less acid was used for hydrolysis compared to Szpiłyk et al. (2021) study16. It is, therefore, consistent that a higher molecular weight is measured.

The P2 GPC peak exhibits a shoulder at a retention time of 8.7 min, which appears to correspond to ethylene carbonate. P2 illustrates high polydispersity, which seems to be further enhanced by detecting ethylene carbonate.

FTIR spectroscopy was performed to monitor the progression of the P2 reaction and to compare the characteristics of the polyols derived from CFs with those of the initial reaction materials.

Figure 3 illustrates FTIR spectra for each polyol synthesized. In CF, glycidol, and glycerol, as well as polyols P1 and P2 spectra, the broad peak over 3000 cm− 1 (3000–3600 cm− 1) is characteristic of the stretching vibration of the O–H bond, while peaks around 2750–3000 cm− 1 can be attributed to the symmetric and asymmetric stretching vibrations of C–H and CH2 in polysaccharides67,68. CFs used in this study were dried at 99%, which explains the absence of the peak around 1630 cm− 1, characteristic of water absorption in cellulose67. On P1 and P2, C–OH peak is observable between 1000 cm− 1 and 1300 cm− 1 (1029 cm− 1 and 1250–1400 cm− 1 characteristic of secondary and tertiary alcohols). The 1170 cm− 1 and 980 cm− 1 peaks can be associated with C–O–C ether groups.

The broad peak at around 1420–1430 cm− 1 is associated with the amount of the crystalline structure of the cellulose, while the peak at 897 cm− 1 is assigned to the amorphous region in cellulose67. Based on previous work, raw CFs are semi-crystalline, which this spectrum confirms34. By studying the absorbance ratio of peaks area assigned to crystalline and amorphous parts of cellulose (A1420/A897), a ratio of 1.68 is obtained. As this result is > 1, this spectrum indicates that CFs contain a larger fraction of crystalline than amorphous domains60,69. The procedure was repeated for hydrolyzed CFs, producing a ratio of 13, indicating a larger crystalline domain than the amorphous domain. It aligns with the theory previously discussed in the DP section, which states that only the crystalline part remains after rinsing hydrolyzed cellulose. As CFs are modified in polyols P1 and P2, extending these results to the FTIR spectra of synthesized biobased polyols is impossible.

Glycerol represents a greater mass fraction than CFs in polyol P1 (10:1 ratio), which can explain the similarity between the FTIR spectra of glycerol and P1. The same observation was made by Kosmela et al.41.

The progression of the P2 reaction was tracked by observing the consumption of ethylene carbonate (EC), specifically by monitoring the disappearance of the peaks in the range of 1560–1933 cm− 1, which is in the region of the spectrum associated with C=O stretching vibrational mode, in this case, associated to lactones70. This peak disappeared, confirming that the reaction had proceeded correctly.

FTIR Spectra obtained for unmodified cellulose filament (CFs), Ethylene carbonate (EC).

Figure 4 illustrates the molecule’s structure through NMR analysis for CFs, P1, and P2 polyols. P2 presented a similar structure than the one obtained by Szpiłyk et al.16. The spectrum shows overlapping proton resonances of methylene and methine groups in both polyols (3.2–3.6 ppm for P1 and 3.2–3.9 ppm for P2). Secondary hydroxyl proton resonance is observable in the 4.4–4.6 ppm range for P1 and 4.4–4.8 ppm for P2.

P1 structure was not represented as the exact structure could not be determined due to the numerous secondary reactions. Acid catalysis, for instance, can lead to cellulose depolymerization, and by-products from glycerol condensation with itself can also be formed36,37.

1H NMR spectra of CFs and polyol, P1 and P2.

Figure 5 presents the impact of modified CF-based polyol on the viscosity as a function of spindle velocity. The viscosity of the petroleum-based polyol with 2.5% CFs was used as a threshold not to be exceeded. Our previous work demonstrated that viscosity increases with the addition of CFs34. Using a 5% CFs concentration resulted in high viscosity, making foam formation challenging and negatively impacting the foam properties. Therefore, staying below the viscosity threshold of the polyol phase containing more than 2.5% CFs was advisable. It was impossible to study the viscosity of R + 2.5% CFs at all the speeds available on the viscometer due to torque being out of the range above 200 rpm. This was also the case for the commercial polyol R above 200 rpm, for substitution with P2 biobased polyol, and for P1 alone above 330 rpm.

The viscosity of P1 was evaluated at substitution ratios of 50%, 70%, and 100% of the petroleum-based polyol. The viscosity of the polyol blends is depicted in Fig. 5. These blends exhibit the lowest viscosity regardless of the spindle velocity. No correlation between the biobased content and the viscosity of the R + P1 blend was observed. Regardless of the spindle velocity, P1 has a higher viscosity than the 50% substitution blend, which has a higher viscosity than the 70% substitution blend. It could be due to the chemical interaction between the two polyols (R + P1). P2 was found to have a very high viscosity, making it impossible to work with alone and measure its viscosity on its own using available equipment. This high viscosity of P2 could, therefore, be predicted. Cellulose and chitosan have structural similarities, allowing for comparisons between their properties. Previous research has investigated the reaction of glycidol and chitosan, which resulted in high-viscosity products24. The P1 + P2 blend used in this study contains 90% of P1 and 10% of P2. Its viscosity was tested by substituting 50% and 70% of the petroleum-based polyol. At low frequencies, polyol mixes containing P2 exhibit higher viscosity than petroleum-based polyol R. Their viscosity decreases with increasing velocity, resulting in a lower viscosity than the petroleum-based polyol R. The viscosity of the petroleum-based polyol decreases slightly and linearly with velocity, so the viscosity values of R and R + 2.5 CFs are close at 200 rpm. R + P1 + P2 polyol mixes depict higher viscosity than P1 (100 rpm at 50% substitution and 60 rpm at 70% substitution). Considering the standard deviation, it is impossible to conclude whether the R + P1 + P2 blend with 70% substitution is more or less viscous than the 50% substitution. However, their viscosity does not exceed that of the reference polyol blend with 2.5% CFs. Therefore, the viscosity criteria were met, and this formulation was used to create foams with P1 + P2 biobased content.

High rheofluidizing properties were observed for all the tested polyols except the petroleum-based polyol, which were also observed in Beaufils-Marquet et al.‘s article (2023)34. It can be explained by the fact that agitation makes the CFs orient themselves in the mixture, reducing viscosity as shear rate increases. This phenomenon is further elaborated in the literature, particularly in the case of cellulose and crystalline structures71,72,73. Limited information is available on petroleum-based polyol to explain the observations. However, it can be assumed that the spatial organization of chains in the formulation is increased with petroleum-based polyol, acting as a lubricant to limit friction and interaction between the complex grafted cellulose polymer chains. Viscosity values support this hypothesis at 50% and 70% of substitution of R by P1 and P1 + P2.

The high viscosity of P2 suggests a high molar mass, results confirmed by molecular weight measurements55.

Profile of formulations viscosity as a function of the spindle velocity (R + CFs 2.5%34).

Foam formation plays a crucial role as it influences the properties of the resultant foams. The foaming process involved temperature monitoring, which is summarized in Fig. 6, while the temperature at a specific time (60 s) after the beginning of the foaming process is presented in Fig. 7. Figure 8 illustrates the stirring time needed before cream-time for each foam.

Boxplots in Fig. 7 highlight the statistically significant differences between foams R (without biobased substitution) and foams with substitution, except 70% R + P1 + P2, whose variability increases with a decreasing temperature trend. No significant differences are observed between batches of the same polyol mix (R + P1 50% and 70%, or R + P1 + P2 50% and 70%). This observation is particularly well respected for R + P1 + P2 in Fig. 6 since the two curves of 50% and 70% overlap. Conversely, the median temperature of R + P1 at 50% is slightly higher than at 70%, but the difference remains small. Furthermore, there are significant differences between 50% R + P1 and batches incorporating P2. The difference in Fig. 7 is not significant between 70% R + P1 and P2 batches, notably due to samples with extreme values, but the median of 70% R + P1 foams is in line with the abovementioned observation. P2 seems to reduce the reaction temperature slightly. It aligns with the theory based on the empirical Guzman-Andrade relationship, according to which viscosity and temperature evolve inversely: the higher the viscosity, the lower the temperature74. In fact, in Fig. 5, shows that at a high shear rate, polyol blends containing biobased polyol tend to be less viscous than R petroleum-based polyol R itself (µ(R) > µ(R + P1 + P2) > µ(R + P1)). While samples without biobased polyol (0% CF-based polyol) exhibit the lowest temperature profile, those containing R + P1 display the highest temperature profiles in Figs. 6 and 7 (T(R) < T(R + P1 + P2) < T(R + P1)). This observation was also made by Kosmela et al. (2018), who explained that the modified polyol system compared to the pure petrochemical polyol is more reactive15. These results support the argument that reactivity and temperature expectancy during foaming are highly dependent on viscosity. In terms of temperature variation in Fig. 6, once cream-time has been reached, a more gradual slope for R is observed, while a steeper slope for R + P1 and an intermediate slope for R + P1 + P2 are observed. This means that once the reaction has started, foams with biobased polyols react faster. In this sense, our results are in accordance with their hypothesis. However, a lower viscosity was observed at 70% compared to 50% substitution (µ(R) > µ(R + PB50%) > µ(R + PB70%)), suggesting that a difference in kinetics should have been observed through lower temperatures as follows: (T(R) < T(R + PB50%) < T(R + PB70%)), which is not the case here.

The temperature profile during the polyurethane foam formation determined the time needed to mix the formulation and react for each foam, illustrated in Fig. 8. Petroleum-based foams exhibit statistically significant differences with other batches except for 50% R + P1, probably due to samples with extreme values. Differences are observable in referencial foams (R_0%) and R + P1 50% while studying the median. Substitution batches at a rate of 70% exhibit statistically significant differences with other batches (50% substitution foams and 0% foams). However, no statistically significant differences are proven between the 70% foam batches, likely due to the variability in their results observed.

In the same way, no statistically significant differences were found between the 50% foam batches. The presence of P2, therefore, does not statistically affect the stirring time required for the reaction. However, it slightly affects the variability in the samples regarding stirring time. In terms of reaction time, the substitution rate of the petroleum-based polyol by the biobased polyols has a greater impact on the reaction kinetics than the choice of the biobased polyol between P1 and P2. A significant difference is noted in the reaction time for petroleum-based foams, which require less than 25 s of agitation to react, compared with the foams with 50% and 70% substitution. Substitution at 50% almost doubles the time required, while at 70%, the reaction time reaches almost 2.5 min, which is six times longer than for the reference. This observation contrasts with the literature since Kosmela et al. (2018) did not observe a noticeable change in processing times by adding biobased polyol in the foam formulation15,41.

According to the product sheet, the catalysts are not recommended for one type of polyol rather than another. Thus, it is not supposed to perform better with a specific type of polyol (ester or ether polyol, in our case). However, Szpilyk et al. (2021a, 2021b) evaluated a cream-time of about 30 s with a mix of biobased and petroleum-based polyol with 1.3%wt and 0.3–0.8%wt of tertiary amine as catalyst related to polyol mass in their two articles. They associated this reaction time with the quantity of catalyst introduced in the formulation, indicating that longer cream and rise times occur with a reduced amount of catalyst16,38. This study used two catalysts to promote the blowing reaction and the reaction leading to urethane formation: a tertiary amine and a catalyst based on metals such as bismuth (Polycat 204 and Dabco MB20, respectively). The catalyst quantity was adjusted based on the amount of polyol utilized. The polyol functionality and OH-equivalent determine the isocyanate: polyol mass ratio. This ratio increases from petroleum-based to partly biobased polyol, reducing the introduced catalyst amount relative to isocyanate mass. In petroleum-based foam, the catalyst is nearly 4% isocyanate mass, while it represents 1.8% in foams with 70% biobased polyol. It is, therefore, possible that the catalyst: isocyanate ratio affects the reactivity of the foams, but the amount of catalyst remains almost twice as high as that used by Szpilyk et al.16,38. It can thus be concluded that the catalyst parameter is probably not the only one affecting kinetic results.

The reactivity also varies inherently among different hydroxyl groups. The petroleum-based polyol used in this study is an ester polyol, while the biobased polyols synthesized are ether polyols. The susceptibility of polyester polyols to chemical attacks is higher than the one of ether polyols due to the presence of carboxyl groups in their polymeric chains75. It could partly explain the difference in reactivity between the biobased polyol and the referential polyol. In addition, the reference polyol used by Kosmela was an ether polyol, which could explain the low or no difference in reactivity between their petroleum-based polyol and the biobased polyol P1. Primary alcohols exhibit a high reactivity at temperatures between 25 and 50 °C, whereas secondary and tertiary alcohols are approximately 0.3 and 0.005 times less reactive than primary alcohols55. Thus, this could affect the kinetic difference between the petroleum-based polyol and the biobased polyol studied in our case. In addition, linear alcohols react faster than branched alcohols, whose reaction speed decreases with the number and size of branches76. The polyols synthesized in this study are obtained by grafting polyol chains onto the cellulose chain, thus providing accessible polyol functions to react with the isocyanate. As a result, biobased polyols are branched polyol, which means they will react more slowly than the presumed relatively linear petroleum-based molecule, which is intrinsically more reactive. All these reasons could thus explain the higher reactivity of commercial polyol compared to CF-based polyol.

In addition, viscosity increases with the reaction process. Since the petroleum-based polyol reacted first, viscosity is higher when biobased polyols react with isocyanate77. From experience, the biobased polyols have a similar functionality to glycerol, which has a higher functionality than the petroleum-based polyol studied in this article. High functionality leads to a branched structure. This structure leads to a faster increase in viscosity, making the reaction more difficult. These factors suggest that the difference in reactivity between the petroleum-based polyol and the studied biobased polyols may be more significant.

Temperature profile throughout the polyurethane foam formation reaction.

Analysis of the temperature at 60 seconds of foaming as a function of CF-based polyol content.

The mixing time required to initiate foam formation (cream time) as a function of the polyol used.

Cell size was evaluated as a function of the percentage of CF-based polyol in the foam, measuring the cell area parameter as illustrated in SEM images in Fig. 9 and reported in Fig. 10.

A wide range of cell sizes was observed for each CF-based polyol percentage studied. In addition, the plot shows a difference between foams containing one or more biobased polyols (R + P1 or R + P1 + P2) and those containing only the petroleum-based polyol, with a decrease in average cell size with the introduction of biobased polyols. A slight decrease in average cell size can also be observed for the same percentage of biobased polyol from R + P1 to R + P1 + P2, coupled with a slight decrease in polydispersity. A reduction of 25% in average cell size was also observed in the case of Kosmela et al. (2018), with a 70% substitution of R with P1. This reduction can be considered advantageous for thermal insulation applications15. Smaller cell sizes generally improve thermal insulation performance by reducing heat transfer through the material due notably to inhibition of convection.

An increase cell size homogeneity was also observed by adding P2 biobased polyol. In the article of Szpilyk et al. (2021), irregularity and large pore size were correlated with a low catalyst: polyol ratio in the foam. In our case, the amount of catalyst was fixed in relation to the polyol (1%), so the irregularity cannot be based on the ratio of catalyst compared to polyol16. However, as mentioned before, the catalyst: isocyanate ratio decreases with biobased polyol content. As such, an increase in irregularity could have been observed with biobased polyol content, which was not the case. The same principle can be applied to surfactants. In our case, the surfactant and catalyst do not seem to support the homogeneity of the foam structure. However, the reactivity of foam formation seems to influence foam morphology, which appears to be correlated with the results observed here. In fact, in order of reactivity from least to most rapid, we observe R + P1 + P2 < R + P1 < R. This order corresponds to the smallest to largest cells observed in Fig. 10.

The blowing agent could also affect the morphology of the foam. During the exothermic reaction process, the released thermal energy induces the transition of the blowing agent from a liquid to a vapor state. This vapor disperses within the formulation and forms nuclei, which remain trapped within the matrix. Trapped nuclei are then transformed into bubbles and spherical cells by merging with surrounding bubbles. An excessively long mixing time may cause the blowing agent to evaporate rather than forming nuclei, resulting in a delayed cream time. If the blowing agent has escaped from the formulation, a decrease in blowing agent content for the same volume occurs, leading to fewer nuclei in this volume. Therefore, the bubbles have greater difficulty coalescing, forming smaller bubbles in polyols that take longer to react with isocyanate. It was difficult to determine the size of the cell walls from the images, so this hypothesis is difficult to validate. However, the order of large and small cells respects the order of agitation times, leading to foams with smaller cells size associated with delayed cream time.

As mentioned before, it has been concluded that the biobased polyols have a higher functionality than the petroleum-based polyols studied in this study. Literature has already correlated the increase of polyol functionality with cell size decrease78. This explanation seems also plausible in our case. Indeed, bubble coalescence becomes more challenging with the increase in cell wall elasticity, attributable to an increased cross-linking78. It is highly likely that the polyol’s functionality also impacts the foam formation process. In addition, the temperature slope could suggest that petroleum-based foams take longer to harden, making it easier for bubbles to coalesce, resulting in larger cells. Whereas bubbles formed in biobased polyols are quickly limited in their growth by the presence of already stable bubbles formed with polyol R, and by the quick increase of biobased polyol crosslinking with isocyanate, which their temperature slope seems to suggest. All this could explain a tendency of smaller cells in the cased of biobased polyol compared to petroleum-based polyol R.

In this study, the petroleum-based foams present the biggest cells and the lowest temperature during foam formation (Figs. 6 and 7). However, 70% substitution illustrated the smallest cells, and they demonstrated slightly lower temperatures than 50% substitution foams. In addition, according to boxplot size, these foams are less homogeneous than R + P1 foams, which are less homogeneous than R + P1 + P2 foams (if H represents the homogeneity: H(R + P1 + P2) > H(R + P1) > H(R) ). So, structure homogeneity parallel to the foam expansion and cell size are not correlated to temperatures during foam formation. This observation disagrees with Kosmela et al.15. The comparison of polyol density led to the conclusion that ρP2 ≤ ρR < ρP1, which leads to ρR ≤ ρR+P1+P2 50 < ρR+P1 50 ≤ ρR+P1+P2 70 < ρR+P1 70. So, parallels cannot be drawn between polyol density and foam morphology, unlike the one suggested by Kosmela et al. (2018) in their article15. Similarly, foam structure does not seem to be correlated with the viscosity of biopolyol, as Kosmela et al. (2018) also suggested15.

Despite the large number of cells observed, cell size measurements were carried out on a limited number of foams. The results are not suitable for statistical analysis. Thus, it cannot be claimed that the sample analyzed by SEM is statistically representative of the behavior of foams in the presence of P1 and P2 in substitution of the petroleum-based polyol R. The conclusions of the correlation between cell size and other properties of the formulation and foams could change when SEM analyzes a larger sample.

SEM images of (1) 0% CFs based polyol (×100), (2) 50% CFs based polyol P1 (×100), (3) 50% CFs based polyol P1 + P2 (×100), (4) 70% CFs based polyol P1 (×100), (5) 70% CFs based polyol P1 + P2(×100).

Foam cells projected area of the foams as a function of the polyols used.

Figure 11 presents the results regarding foam density as a function of CF-based polyol content. Regardless of the biobased content in the formulation, all foams surpassed the mandated minimum density standard of 28 kg m− 357. Referential foams exhibit only statistical significative differences with 50% R + P1 + P2 substitution foams. 50% R + P1 + P2 exhibits statistically significant differences with 0% foams and foams with 70% substitution. In addition, it can be observed that the medians of both R + P1 foams are close to the results observed for the petroleum-based foam. The Dunn post-hoc did not allow to identify any statistically significant differences between foams with petroleum-based polyol R and foams containing R + P1 without P2. This suggests that they are equivalent in terms of impact on density. In 70% P1 + P2 substitution, P2 exhibits statistically significant differences with batches with 50% substitution, which is not valid for 70% R + P1. At 70% substitution of the petroleum-based polyol from P1 to P1 + P2 polyol mix, the density increases slightly to reach the median of 35.3 kg m− 3 and a mean density of approximately 37.2 kg m− 3. Meanwhile, at 50% substitution with P1 and P1 + P2, a decrease from a median of 34.9 to 32.0 kg m− 3 (a decrease of the mean from 35.0 to 32.7 kg m− 3) is observed. A higher dispersity of results is observed for foams containing biobased content. This could be explained by the reaction kinetics of foams with biobased content, which affects the content of the blowing agent. In fact, foam density can be controlled by adjusting the quantity of blowing agents added, either increasing or decreasing the amount75. These results may have been also influenced by kinetics. The physical blowing agent (HCFO) is added to the polyol phase and then refrigerated to stabilize due to its low boiling point (approximately 18 °C).

On the one hand, the reaction leading to polyurethane foam formation is exothermic, meaning that as the reaction begins, the temperature increases gradually during agitation after adding pMDI. As illustrated in Fig. 8, the cream-time of petroleum-based polyol fraction occurs in less than 25 s of agitation, so the temperature increases around 25 s of stirring. This observation is still true in biobased polyol formulations containing 50% or 30% of the petroleum-based polyol. Biobased content foams require more agitation time for a successful reaction, and agitation ends when the product starts to foam (cream-time). It could increase density due to the evaporation of the blowing agent when the temperature rises.

On the other hand, if a portion of the petroleum-based fraction has already begun to react, cells containing the blowing agent begin to form. As the agitation continues, they are damaged and can release some of the blowing agent. In this case, the blowing agent can be entirely replaced with air, resulting in reduced density due to the lower air density than the blowing agent. This second explanation could clarify the results in the case of foams with 50% substitution, which leads to a decrease in density. These two hypotheses exhibit interdependence in their process, thus potentially contributing to the variability observed in the results rather than indicating a specific trend in the outcomes. It should also be noted that addition and storage methods of formulations can lead to variability in the amount of blowing agent remaining in the formulation before reaction, so that each foam does not contain precisely the same mass of chemical blowing agent, thus increasing the variability of the whole process and the subsequent properties of the foams.

Another plausible hypothesis is that the higher viscosity of the petroleum-based polyol compared to P1 helps retain the blowing agent in the formulation during mixing, incorporating the foaming agent, and then the reaction. Blowing agent evaporation before the reaction leads to an increase in density. While blowing agent evaporation during the reaction influences badly the strength of the bubble interface. It can lead to substituting the blowing agent with air in the case of viscosity-related brittle interfaces. As a result, foam density and the closed-cell content decrease. It could partly explain the decrease in density at 50% substitution. This conjecture finds support with Al-Moameri et al. (2019), who explain that the evaporation rate of a liquid physical blowing agent depends on the vapor pressure of the blowing agent, serving as the driving force for gas formation, and the mass transfer rate from the resin to the gas cell. The overall mass transfer coefficient of the blowing agent vapor declines with increasing resin viscosity79. It is also noteworthy that when evaluating the NCO/OH = 1.13 ratio, the volume ratio of the polyol phase to the isocyanate phase remains relatively constant for the petroleum-based polyol.

Conversely, for P1 and P2, a reduced volume of polyol is necessitated to achieve this NCO/OH ratio. Thus, as the proportion of biobased polyol increases, there is a decline in the volume of the polyol phase compared to the isocyanate phase. It indicates that maintaining a stable blowing agent content in the biobased polyol is more challenging, as it must stabilize within a reduced volume compared to the petroleum-based polyol. The blowing agent is, therefore, more likely to evaporate in formulations with 70% substitution. This, coupled with the extended agitation time required for these formulations, can lead to a rise in foam density at 70% substitution. This hypothesis is more moderate at 50% substitution.

Previously, it was noted that the foam cell size decreases with the presence of biobased polyol. This principle aligns with the scenario of 70% substitution. The smaller the cells, the greater the specific surface area around the bubbles. So, for an equal volume, if that volume contains large bubbles, it will contain less polyol; thus, a decrease in cell size suggests an tendency towards increased density. This anticipated correlation should also be evident at 50% substitution, which was not observed in this context. This hypothesis can, therefore, be rejected.

Similarly, high polyol functionality has been correlated with higher density in the literature78. As our foam density results emphasize, a little or no density difference is observed between petroleum-based polyol foams and polyol blends containing biobased polyols. So, this explanation cannot be validated in our case.

Szpilyk et al. (2021a, 2021b) produced a polyurethane foam with an apparent density of 72.9 kg m− 3 and 60.5 kg m− 3, which means more than two times denser than the present foams16,38. However, they did not incorporate a physical blowing agent. Since the amount of blowing agent influences the density of the foams, it is conceivable that the difference in density between the foams in their study and ours may be attributed to the quantity and properties of the blowing agent. While Kosmela et al. (2016 and 2018) produced PU foams of about 40 kg m− 3 in the presence of a physical blowing agent, falling within the range of results obtained in our study although slightly less dense in our case15,41.

The hypotheses formulated here explain the increase in density at 70%, but not the decrease in density at 50% substitution, especially between 0% biobased polyol and 50% of P1 + P2.

Density of the foam as a function of CF-based polyol content.

To comply with the Canadian standard for good thermal insulation performance, rigid polyurethane foam must have more than 90% closed cells57. This property has an impact on keeping the blowing agent inside the cells. As it has a low thermal conductivity, it is not desirable for the blowing agent to escape from the structure, nor is it intended to be replaced by an air of higher conductivity, which will be able to circulate freely in the case of open cells. Closed cells also allow better control of mechanical performance since the force is distributed throughout the material.

Table 6; Fig. 12 present the closed cells content of the foams produced as a function of CF-based polyol content. No foam with biobased content meets the 90% or more closed cells requirement. Foams with petroleum-based polyol show statistically significant differences with all the batches containing biobased polyols. Substitutions with 70% P1 + P2 also exhibit statistically significant differences from other batches. More specifically, the foams with P1 + P2 at 70% substitution have a deficient number of closed cells (approximately 8 ± 3% closed cells), which contradicts the required standard. Batches with 50% substitution (P1 and P1 + P2) and 70% P1 substitution exhibit more variability than other batches and do not differ statistically from each other. However, 50% P1 + P2 substitution stands out with a higher median and mean (summarized in Table 6) than samples substituted with P1 alone at 50% and 70%. At 50%, a slight advantage from the presence of P2 can be observed, although it correlates to closed-cell content degradation at 70% substitution. However, the dispersity is too high (in the case of 50% R + P1 + P2), and the difference is too small to be statistically significant. It would have been interesting to test lower substitution percentages to see if P2 benefits the closed-cell rate since this phenomenon was only observed at 50% and not at 70%.

The foaming process is very complex since too fast or too slow development can lead to the formation of unstable cells and, hence, the opening of the cells, which is not desired in our case. As mentioned in the section on foam density, it is likely that reaction kinetics and process impact cell structure and integrity. Petroleum-based polyol reacts quickly, whereas biobased polyol requires more mechanical stirring of the formulation to reach cream-time. In foams with biobased content, this likely leads to the rupture of cells formed with petroleum-based polyol, which are stabilized and cured well before cells formed from the biobased polyols. This phenomenon is particularly well illustrated in the 70% P1 + P2 substitution case. In addition, these foams showed lower temperatures at 60 s than other foams containing biobased content. This formulation also exhibited a higher stirring time than needed before cream time. All of this supports the fact that a change in structure would occur for this formulation.

As mentioned previously, the impact of the foaming process and kinetics might be related to the viscosity and functionality of polyols that might affect cell stability during foaming. It is also plausible that surfactants reach their limits for either too high or too low viscosities, leading to cell openings. Due to the prolonged agitation and the time required for the foaming process, the surfactant must maintain sufficient stability to the bubbles formed in the formulation. This is feasible with petroleum-based polyol, as the agitation time is short. However, with extended agitation, the surfactant’s efficiency in stabilizing the bubbles may be compromised. This represents the limitations of surfactants. It can be hypothesized that the reaction time reaches the surfactant’s limit when biobased polyols are present. Therefore, it would be pertinent to investigate various surfactants and catalysts to identify the most effective compounds on closed cells content property for use with biobased polyols. This could either reduce the reaction time of the biobased polyols, thereby staying within the surfactant’s limit, or alternatively, enhance the surfactant’s bubble stabilization capacity despite the extended agitation time required to reach cream time.

Szpilyk et al. (2021a, 2021b) evaluated the closed and open pores content through water uptake and microscopic pictures, providing an approximate result and preventing us from verifying if the closed-cell rate surpasses 90%16,38. While Kosmela et al. (2016, 2018) did not focus on open and closed cell content15,41. Therefore, no numerical comparison data are available for P1 and P2.

Closed cells proportion in the foam as a function of CF-based polyol content.

Figure 13 presents the thermal conductivity results as a function of CF-based polyol content. Foams with petroleum-based polyol R show statistically significant differences with 50% R + P1 and 70% batches. In fact, 50% R + P1 and 70% batches do not differ in a statistically significant way from each other, meaning that 50% P1 + P2 substitution stands out in other batches in line with the results of reference foams. It can also be noted that 50% R + P1 + P2 exhibits much variability. Foams 50% R + P1 + P2 result in superior insulation properties to other foams with biobased content. In addition, the 50% R + P1 + P2 foams exhibit a higher closed cell content and lower density than other CF-based foams, which can be attributed to an increased blowing agent content. Nevertheless, it is important to note that the correlation between thermal conductivity and closed cells is not linear, as illustrated 70% R + P1 + P2, which stood out unfavorably due to its very low closed cell content. So, these foams should have exhibited higher thermal conductivity.

The thermal conductivity of foams depends on several factors. Firstly, it is influenced by the thermal conductivity of the gas within the foam, whether it is a physical blowing agent for closed or air for open cells. This thermal conductivity is also closely related to the foam’s density, which is partly determined by the content of the physical blowing agent. It also depends on the thermal conductivity of the foam formulation component, such as polyol. Finally, the size of the cells in the foam also plays a crucial role in its thermal conductivity80. Smaller cell size can increase a foam’s thermal insulation property, forming more interfaces and reducing thermal conductivity. Moreover, a decrease in cell size can be observed with the introduction of CF-based polyol. The 50% R + P1 + P2 foams’ cell size does not stand out from other foams with CF-based polyol. However, R + P1 + P2 at 70% can benefit from this since it is the formulation that contains the smallest cells.

Furthermore, it was already proven by Beaufils-Marquet et al. (2023) that raw CFs negatively impact the thermal insulation property of the foam by increasing the thermal conductivity with its quantity present in the material, even without affecting the integrity of the material34. It is, therefore, possible to hypothesize that CF, modified for use as a polyol, increases the thermal conductivity of the overall foam, as thermal conductivity is slightly higher in the presence of CF-based polyol P1 and P2 compared to petroleum-based polyol.

Even if a tendency of thermal conductivity increase is observed with the substitution of petroleum-based polyol by CF-based polyol, the values remain close to those of commercial foam and competitive with the results of other conventional types of insulation such as stone wool (0.033–0.040 W m− 1‧K− 1), or wood fibers (0.038–0.050 W m− 1‧K− 1)9. However, it is known that thermal conductivity increases with time in the case of polyurethane foam since thermal conductivity is generally calculated at five years for polyurethane foams. It would, therefore, be relevant to study these foams on a larger scale of time and dimensions to analyze thermal conductivity per the associated standard.

Thermal conductivity as a function of CF-based polyol content.

Young’s modulus is represented as a function of CF-based polyol content in Fig. 14. A significant statistical difference is observed between the referential batch and the other batches containing CF-based polyol. CF-based polyols exhibit lower Young’s modulus, meaning a decrease in rigidity. Only extreme values of 50% P1 substitution achieve a Young’s modulus similar to that of the petroleum-based reference and commercial foam, although 70% R + P1 almost reaches the Young’s modulus of commercial foam. It is not the case for batches containing P2, as its lower median indicates.

Figure 15 depicts the maximum strength supported by the foams at their elastic limit (yield point). In the same way as Young’s modulus study in Fig. 14, a significant statistical difference is observed between the referential batch and the other batches containing CF-based polyol. Only extreme values of 50% P1 and 70% P1 almost reach the maximum strength of the referential foams. All foams with biobased content approach the necessary values to comply with the standard but fall short. As mentioned before, a decrease in cell size was observed with the introduction of CF-based polyol. A decrease in cell size is generally correlated with forming more interfaces.

However, having more interfaces can lead to interface thinning and opening, which can reduce mechanical performance. Also, as mentioned previously, closed cells allow better control of mechanical performance since the force is distributed throughout the material. However, 70% R + P1 + P2 performs surprisingly better than expected. As mentioned previously, thermal conductivity, closed cell content, and mechanical performances do not evolve linearly. Results can also be linked to CFs studied in our previous work, which illustrated an increase in foam rigidity with the addition of CF. A decrease in maximum supported strength was also observed with the addition of CF34. The literature also supports the assertion that the higher the foam density, the better the mechanical strength of the foams81. However, the results of this study do not support this assertion.

The flexibility of a molecule at the molecular level relies on the freedom of rotation around the single bonds in its main chain43. When a –CH2– in an aliphatic main chain is substituted with oxygen (–CH2–O–CH2–), the rotation around the C–O bond avoids conflicts between H atoms, facilitating rotation and enhancing molecular flexibility. This phenomenon is observed in polyethers, polyesters, and even polyurethane linkages. Conversely, introducing the aromatic ring into the PU backbones provides higher rigidity and thermal stability75. We could, therefore, assume that switching from a petroleum-based ester polyol to a biobased ether polyol would increase the flexibility of the chains. However, polymer rigidity relies especially on the number of functional groups, the degree of branching, and the cross-linking in the hyperbranched polyol. For example, as the number of functionalities in the hyperbranched polyol increases, more efficient cross-linking occurs within the polyurethane network82. So, if a rigid foam is required, a highly cross-linked polymer structure is preferred; conversely, flexible foams require less cross-linking55. Terol 649, the petroleum-based referential polyol, has a functionality of 3, as does glycerol. So glycerol should not make the difference, but rather cellulose, since cellulose is made up of glucose units with a functionality of 5. Although the reactivity of the 5 OH on cellulose is not equal, this is the source of the formation of glycerol chain grafts, which will give rise to a significant number of polyol functions likely to react with the isocyanate. Therefore, an increase in rigidity should be observed, but this was not the case here. It could be due to the hydrolyzation of CFs due to the presence of sulfuric acid. In addition, ester polyols are used to provide high strength to polyurethane elastomers. It could explain the best results for petroleum-based foams55.

Szpilyk et al. (2021a, 2021b) reached 212 and 234 kPa strength in compression, meaning that foams respected the standard16,38. In Kosmela et al. (2016), compression strength between 256 and 367 kPa was attained41. However, density and mechanical performances are related15,41. Kosmela et al. (2016 and 2018) produced PU foams of about 40 kg m− 315,41. It was expected that Kosmela et al. (2016, 2018) foams would have higher mechanical performance.

Young’s modulus as a function of CF-based polyol content.

Maximum strength supported at yield point as a function of CF-based polyol content.

Table 7 summarizes the statistical analysis results to assess reaction kinetics, foam density, closed cell content, thermal conductivity, and mechanical properties evaluated and discussed in the previous sections for every polyurethane foam types (Reference (R), R + P1 50%, R + P1 70%, R + P1 + P2 50%, R + P1 + P2 70%). Since the comparison between groups was conducted in pairs, the null hypothesis (that the medians are equal) is rejected for pairs of groups where the adjusted p-value is less than α/2, with α = 0.05 as the significance threshold. Pairs marked with asterisks signify significant differences83,84. The p-value is also available in the graphs associated with each analysis.

The thermal stability of polyurethane foam was investigated using thermogravimetric analysis under inert conditions. Figure 16 depicts the mass loss curves and their derivatives. Table 8 provides an overview of the key parameters: the temperature at which 5% of the sample mass is lost (T5%), the temperatures at which maximum degradation rates occur, and the percentage of residues remaining at 800 °C.

Thermogravimetric analysis performed under nitrogen and their derivatives polyurethane foam with various biobased polyol contents.

Commercially available polyurethane polymers contain a variety of groups that can degrade during TGA. Biuret, allophanate, urethane, urea, isocyanurate, ether, and ester are the most common85. As an example, four degradation steps were observed for the polyurethane reference foam, as presented in our previous work34. The presence of water was observed in the studied foams34. In the present article, the temperature of the first degradation peak seems too high to be associated with the presence of water in the material86. Hence, no peaks corresponding to water appear on the TGA curves. This is because isocyanate is highly reactive with moisture, forming urea and CO2 functions55. Therefore, the first degradation peak in the mass loss derivative (T1 MAX in Table 9) seems more likely to be associated with allophanate and biuret, among the least thermally stable compounds mentioned. These compounds typically undergo degradation within the temperature range of 110 °C to 170 °C, a range that correlates with the values observed for T1 MAX in Table 8. This peak could also be associated with isocyanate monomers, which, having not reacted in the foam, are known to degrade before 200 °C. Given the absence of instrumentation for fume analysis during TGA assessments, definitive conclusions are challenging to ascertain. This degradation peak is not observed in the case of R + P1 with 50% and 70% substitution, or at least has not been identified on the mass loss derivative. Explaining this difference without further information, such as the smoke analysis, is difficult. The next less thermally stable compounds are urethane and then urea. The degradation of polyether chains is theoretically observed between 250 and 320 °C. This aligns with T2 MAX (Table 8) with a thermal degradation of around 266–270 °C. Thus, this peak results from the degradation of the urethanes associated with the ether compounds through a depolymerization mechanism85. The variation in thermal stability between polyester and polyether is likely due to how molecules move and interact, i.e., molecular rigidity, which affects chemical reactions and degradation55.

Isocyanurates result from the arrangement of several isocyanate compounds through trimerization. These compounds are capable of maintaining their thermal stability at high temperatures. This is why the last degradation peaks could be associated with isocyanurates. Ester is also a particularly stable compound. It is expected that a difference in degradation peaks will be observed between the ester polyol, which is the petroleum-based polyol in our case, and ether polyols, which are P1 and P2. This difference in degradation is effectively observed in this article. According to the literature, T3 MAX, could be associated with ester degradation and urethane degradation associated with the ester compounds through a depolymerization mechanism. T4 MAX probably illustrates the degradation of compounds resulting from the degradation of urethane functions formed with the ester and ether functions. Observing T1 MAX and T5 MAX, it seems likely that different compounds have formed in R + P1 and R + P1 + P2, yet further conclusions are challenging to draw.

Using degradation peaks and mass monitoring, tracking the amount of material degraded at each stage can be interesting. Mass loss equivalent to 5% of the product’s mass occurs at a lower temperature for foams containing the biobased ether polyol (approximately 230 °C) than for the reference ether polyol (264 °C). Similarly, the most significant steps in the reference foam end at 358 °C, resulting in 11.7% mass loss before 358 °C and 61.7% above. With biobased polyols, the two steps in the same range represent respectively the value summed up in Table 9. A mass loss of about 37% occurred below 358 °C and 40% above 358 °C. A greater mass loss associated with temperature range 1 is observed in foams containing P1 and P2 biobased polyols (about 63% of mass loss). It supports the fact that polyester is more thermally stable than polyether.

In addition, the quantity of residual material can be linked to the NCO/OH ratio, resulting in isocyanate or polyol excess. Indeed, an excess of isocyanate leads to improved thermal stability87. In the present case, the mass ratio between the quantity of isocyanate and biobased polyol introduced into the formulation is based on the assumption that the properties of polyol in terms of functionality are equal to those of glycerol. There was, therefore, a risk that the quantity of isocyanate and polyol would not respect the NCO/OH = 1.13 ratio. In contrast, the amount of petroleum-based polyol in the reference foam was calculated using the properties of the polyol itself. Therefore, in this case, it is reasonable to observe a surplus of isocyanate in the thermogravimetric analysis due to the intentional introduction of an excess of isocyanate relative to the polyol manifested as residues remaining at the end of the analysis. In the case of foams containing biobased polyol, it cannot be stated that the effective amount of isocyanate in relation to the polyol has complied with the 1.13 = NCO/OH ratio, as indicated by the residue remaining at the end of the analysis. However, as mentioned previously, ester polyols have better fire retardancy performances than ether polyols. It means that at a similar NCO/OH ratio, it is possible that ester polyol exhibits a higher residue amount than ether polyols.

This study aimed to investigate the viability of modifying cellulose filaments as a sustainable polyol solution to improve the environmental performance of polyurethane foam. To achieve this goal, cellulose filaments (CFs) were modified through two distinct etherification protocols. The two resulting polyols were introduced in polyurethane foams. The properties of polyols and foams were evaluated to evaluate the potential of such products to substitute the petroleum-based reactive in polyurethane foams.

Analyses showed that the introduction of modified CFs as a substitute for the petroleum-based polyol led to a deterioration in the material’s overall properties: cell degradation leading to non-compliance with the standard requiring a closed-cell rate of over 90% (with 8 ± 3% closed cells in 70% R + P1 + P2, 64 ± 13% in 50% P1, 72 ± 14% in 50%P1 + P2 and 61 ± 16% in 70%P1), degradation of mechanical properties through loss of stiffness and a drastic reduction in the maximum strength that the material can withstand under the standard of maximum strength > 170 kPa and loss of rigidity (with 2.5 MPa Young’s modulus against 7.5 MPa for the petrosourced reference). These results are partly due to the different reaction kinetics of the products and polyol properties, such as functionality and viscosity. It was also shown that one of the polyol syntheses from CFs led to a highly viscous product that was difficult to handle as a polyol in polyurethane foams. Therefore, It was shown that the CFs used in this study do not meet the expectations of a product that can compete with existing petroleum-based products.

This article highlighted that not all celluloses are equal for polyol synthesis and that in this particular case of CF, it was impossible to achieve the performance measured elsewhere in the literature. It would be relevant to further research lower percentages of biobased polyols to validate or invalidate the hypotheses linking the percentage substitution of petroleum-based polyol to biobased polyol and the degradation of properties. Although cellulose appears to be an interesting material to valorize as a biobased chemical, using a different type of cellulose would likely allow for better control of the resulting polyol’s viscosity and functionality. It would certainly have a lesser impact on foam properties such as reaction kinetics, cell openness, or mechanical properties, as observed in the present case.

Data available on request from the corresponding author.

IEA & Chemicals IEA. https://www.iea.org/energy-system/industry/chemicals (2023).

IEA & Tracking Clean Energy Progress—Analysis. IEA (2017). https://www.iea.org/reports/tracking-clean-energy-progress-2017 (2017).

IEA. The Future of Petrochemicals—Analysis. IEA. https://www.iea.org/reports/the-future-of-petrochemicals (2018).

IEA. Key World Energy Statistics 2020—Analysis—IEA. IEA. https://www.iea.org/reports/key-world-energy-statistics-2020 (2020).

IEA & Building envelopes. IEA. https://www.iea.org/energy-system/buildings/building-envelopes (2023).

Al-Homoud, D. M. S. Performance characteristics and practical applications of common building thermal insulation materials. Build. Environ. 40, 353–366 (2005).

Article MATH Google Scholar

Palumbo, M., Lacasta, A. M., Navarro, A., Giraldo, M. P. & Lesar, B. Improvement of fire reaction and mould growth resistance of a new bio-based thermal insulation material. Constr. Build. Mater. 139, 531–539 (2017).

Article CAS MATH Google Scholar

Lafond, C. & Blanchet, P. Technical performance overview of bio-based insulation materials compared to expanded polystyrene. Buildings 10, 81 (2020).

Article MATH Google Scholar

Schiavoni, S., D׳Alessandro, F., Bianchi, F. & Asdrubali, F. Insulation materials for the building sector: a review and comparative analysis. Renew. Sustain. Energy Rev. 62, 988–1011 (2016).

Article MATH Google Scholar

Asdrubali, F., D’Alessandro, F. & Schiavoni, S. A review of unconventional sustainable building insulation materials. Sustain. Mater. Technol. 4, 1–17 (2015).

CAS MATH Google Scholar

Abbott, A. et al. Metrics: do metrics matter? Nature 465, 860–862 (2010).

Article CAS PubMed MATH Google Scholar

Manohar, K. Experimental investigation of building thermal insulation from agricultural by-products. Curr. J. Appl. Sci. Technol. 227–239. https://doi.org/10.9734/BJAST/2012/1528 (2012).

Uram, K. et al. Polyurethane composite foams synthesized using bio-polyols and cellulose filler. Materials 14, 3474 (2021).

Article ADS CAS PubMed PubMed Central MATH Google Scholar

Hejna, A. et al. The influence of crude glycerol and castor oil-based polyol on the structure and performance of rigid polyurethane-polyisocyanurate foams. Ind. Crops Prod. 95, 113–125 (2017).

Article CAS MATH Google Scholar

Kosmela, P., Hejna, A., Formela, K., Haponiuk, J. & Piszczyk, Ł. The study on application of Biopolyols obtained by Cellulose Biomass Liquefaction Performed with crude glycerol for the synthesis of rigid polyurethane foams. J. Polym. Environ. 26, 2546–2554 (2018).

Article CAS Google Scholar

Szpiłyk, M., Lubczak, R., Walczak, M. & Lubczak, J. Polyol and polyurethane foam from cellulose hydrolysate. J. Chem. Technol. Biotechnol.. 96, 881–889 (2021).

Article Google Scholar

Peyrton, J., Chambaretaud, C., Sarbu, A. & Avérous, L. Biobased polyurethane foams based on new polyol architectures from Microalgae Oil. ACS Sustain. Chem. Eng. 8, 12187–12196 (2020).

Article CAS Google Scholar

Peyrton, J. & Avérous, L. Structure-properties relationships of cellular materials from biobased polyurethane foams. Mater. Sci. Eng. R: Rep. 145, 100608 (2021).

Article MATH Google Scholar

El Khezraji, S. et al. Recent progress of non-isocyanate polyurethane foam and their challenges. Polymers 15, 254 (2023).

Article PubMed PubMed Central Google Scholar

Gu, R., Khazabi, M. & Sain, M. Fiber Reinforced soy-based polyurethane spray Foam Insulation. Part 2: Thermal and Mechanical properties. BioResources 6, 3775–3790 (2011).

CAS MATH Google Scholar

Kakroodi, A. R., Khazabi, M., Maynard, K., Sain, M. & Kwon, O. S. Soy-based polyurethane spray foam insulations for light weight wall panels and their performances under monotonic and static cyclic shear forces. Ind. Crops Prod. 74, 1–8 (2015).

Article CAS Google Scholar

Khazabi, M., Gu, R. & Sain, M. Fiber reinforced soy-based polyurethane spray foam insulation. Part 1: cell morphologies. BioResources 6, 3757–3774 (2011).

Article CAS Google Scholar

Yu, Z. et al. Bagasse as functional fillers to improve and control biodegradability of soy oil-based rigid polyurethane foams. Korean J. Chem. Eng. 36, 1740–1745 (2019).

Article CAS MATH Google Scholar

Strzałka, A., Lubczak, R. & Lubczak, J. Chitosan Oligomer as a raw material for obtaining polyurethane foams. Polymers 15, 3084 (2023).

Article PubMed PubMed Central MATH Google Scholar

Maisonneuve, L., Chollet, G., Grau, E. & Cramail, H. Vegetable oils: a source of polyols for polyurethane materials. OCL. 23, D508 (2016).

Andersons, J., Kirpluks, M., Cabulis, P., Kalnins, K. & Cabulis, U. Bio-based rigid high-density polyurethane foams as a structural thermal break material. Constr. Build. Mater. 260, 120471 (2020).

Article CAS Google Scholar

Riyapan, D., Saetung, A., Saetung, N. A. Novel rigid PU foam based on modified used palm oil as sound absorbing material. J. Polym. Environ. 27, 1693–1708 (2019).

Article CAS Google Scholar

Ma, Y. et al. Biomass based polyols and biomass based polyurethane materials as a route towards sustainability. React. Funct. Polym. 175, 105285 (2022).

Article CAS MATH Google Scholar

Huntsman building solutions. Heatlok Soya HFO | Huntsman Building Solutions. https://huntsmanbuildingsolutions.com/en-CA/products/closed-cell-insulation/heatlok-soya-hfo (2023).

Polaczek, K., Kurańska, M., Malewska, E., Czerwicka-Pach, M. & Prociak, A. From bioresources to thermal insulation materials: synthesis and properties of two-component open-cell spray polyurethane foams based on bio-polyols from used cooking oil. Materials 16, 6139 (2023).

Article ADS CAS PubMed PubMed Central Google Scholar

Hurmekoski, E. et al. Diversification of the forest industries: role of new wood-based products. Can. J. Res. 48, 1417–1432 (2018).

Article MATH Google Scholar

Hurmekoski, E., Hetemäki, L. & Jänis, J. Outlook for the forest-based bioeconomy. In Forest Bioeconomy and Climate Change (eds Hetemäki, L., Kangas, J. & Peltola, H.) 55–89. https://doi.org/10.1007/978-3-030-99206-4_4 (Springer International Publishing, 2022).

Chapter MATH Google Scholar

Heinze, T. & Liebert, T. 10.05-celluloses and Polyoses/Hemicelluloses. In Polymer Science: A Comprehensive Reference (eds Matyjaszewski, K. & Möller, M.) 83–152. https://doi.org/10.1016/B978-0-444-53349-4.00255-7 (Elsevier, 2012).

Chapter MATH Google Scholar

Beaufils-Marquet, M., Blanchet, P., Hussain, A. & Landry, V. Investigation of cellulose filaments as filler in rigid insulating polyurethane foam. 18, 6086–6117 (2023).

Kuranska, M., Prociak, A., Michalowski, S., Cabulis, U. & Kirpluks, M. Microcellulose as a natural filler in polyurethane foams based on the biopolyol from rapeseed oil. Polimery 61, 625–632 (2016).

Article Google Scholar

Ionescu, M. & Petrović, Z. High functionality polyether polyols based on Polyglycerol. J. Cell. Plast.. 46, 223–237 (2010).

Article CAS MATH Google Scholar

Grilc, M., Likozar, B. & Levec, J. Biofuel from lignocellulosic biomass liquefaction in waste glycerol and its catalytic upgrade. 137–144. https://doi.org/10.1007/978-3-658-04355-1_17 (2015).

Szpiłyk, M., Lubczak, R. & Lubczak, J. The biodegradable cellulose-derived polyol and polyurethane foam. Polym. Test. 100, 107250 (2021).

Article MATH Google Scholar

Liyanage, S., Acharya, S., Parajuli, P., Shamshina, J. L. & Abidi, N. Production and surface modification of cellulose bioproducts. Polymers 13, 3433 (2021).

Article CAS PubMed PubMed Central Google Scholar

Romeo, I. et al. Totally green cellulose conversion into bio-oil and cellulose citrate using molten citric acid in an open system: synthesis, characterization and computational investigation of reaction mechanisms. RSC Adv. 10, 34738–34751 (2020).

Article ADS CAS PubMed PubMed Central Google Scholar

Kosmela, P., Hejna, A., Formela, K., Haponiuk, J. T. & Piszczyk, Ł. Biopolyols obtained via crude glycerol-based liquefaction of cellulose: their structural, rheological and thermal characterization. Cellulose 23, 2929–2942 (2016).

Article CAS Google Scholar

European Standards & DIN EN 12667. https://www.en-standard.eu/din-en-12667-thermal-performance-of-building-materials-and-products-determination-of-thermal-resistance-by-means-of-guarded-hot-plate-and-heat-flow-meter-methods-products-of-high-and-medium-thermal-resistance-english-version-of-din-en-12667/ (2001).

European Standards & UNE EN 14315-. 1:2013 Thermal insulating products for buildings. In-situ formed sprayed rigid polyurethane foam (PUR) products. Part 1: Specification for the rigid foam spray system before installation. https://www.en-standard.eu (2013).

Prete, P. et al. Glycidol syntheses and valorizations: boosting the glycerol biorefinery. Curr. Opin. Green. Sustain. Chem. 35, 100624 (2022).

Article CAS Google Scholar

Kruger Inc. Avantages de Filocell-Biomateriaux-Kruger Inc. Biomaterials. https://biomaterials.kruger.com/fr/products/avantages-de-filocell/ (2017).

Hua, X., Laleg, M. & Owston, T. Cellulose nanofilaments and method to produce same. (2011).

TAPPI. T230 OM-Viscosity of pulp (capillary viscometer method). (2008).

Hawkins, M. C., O’Toole, B. & Jackovich, D. Cell morphology and mechanical properties of rigid polyurethane foam. J. Cell. Plast. 41, 267–285 (2005).

Article CAS MATH Google Scholar

R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2013).

Chen, S. Y., Feng, Z. & Yi, X. A general introduction to adjustment for multiple comparisons. J. Thorac. Dis. 9, 1725–1729 (2017).

Article PubMed PubMed Central MATH Google Scholar

Kemp, M. H. D. Extreme Events: Robust Portfolio Construction in the Presence of Fat Tails (Wiley, 2011).

Kraska-Miller, M., Marie. Nonparametric Statistics for Social and Behavioral Sciences (CRC, 2014).

Leng, W. & Pan, B. Thermal insulating and mechanical properties of cellulose nanofibrils modified polyurethane foam composite as structural insulated material. Forests 10, 200 (2019).

Article MATH Google Scholar

ASTM D1622. Test method for apparent density of rigid Cellular Plastics. https://doi.org/10.1520/D1622-20 (2020).

Szycher, M. Structure–property relations in polyurethanes. In Szycher’s Handbook of Polyurethanes (2012).

ASTM D6226. Test Method for Open Cell content of rigid Cellular Plastics. https://doi.org/10.1520/D6226-21 (2021).

Standards Council of Canada-Conseil canadien des normes. CAN/ULC-S705-1-Norme sur l’isolant thermique en mousse de polyuréthane rigide pulvérisée de densité moyenne-spécifications relatives aux matériaux. https://nrc.canada.ca/fr/certifications-evaluations-normes/centre-canadien-materiaux-construction/publications-ccmc/document.html?type=techcrit&id=CAN_ULC-S705.1-15 (2015).

Aphalo, P. J., Slowikowski, K. & Mouksassi, S. ggpmisc: Miscellaneous Extensions to ‘ggplot2’. Leanpub (2021).

Blanco, A. et al. Chapter 5-Nanocellulose for Industrial Use: Cellulose Nanofibers (CNF), Cellulose Nanocrystals (CNC), and Bacterial Cellulose (BC). In Handbook of Nanomaterials for Industrial Applications (ed. Mustansar Hussain, C.) 74–126. https://doi.org/10.1016/B978-0-12-813351-4.00005-5 (Elsevier, 2018).

De Silva, R. & Byrne, N. Utilization of cotton waste for regenerated cellulose fibres: influence of degree of polymerization on mechanical properties. Carbohydr. Polym. 174, 89–94 (2017).

Article PubMed MATH Google Scholar

Krause, C., Dreier, L., Fehlmann, A. & Cross, J. The degree of polymerization of cellulosic insulation: Review of measuring technologies and its significance on equipment. In IEEE Electrical Insulation Conference (EIC) 267–271. https://doi.org/10.1109/EIC.2014.6869389 (IEEE, 2014).

Henriksson, M., Berglund, L. A., Isaksson, P., Lindström, T. & Nishino, T. Cellulose Nanopaper Struct. High. Toughness Biomacromolecules 9, 1579–1585 (2008).

CAS Google Scholar

Bouchard, J., Méthot, M., Fraschini, C. & Beck, S. Effect of oligosaccharide deposition on the surface of cellulose nanocrystals as a function of acid hydrolysis temperature. Cellulose 23, (2016).

Matsuoka, S., Kawamoto, H. & Saka, S. What is active cellulose in pyrolysis? An approach based on reactivity of cellulose reducing end. J. Anal. Appl. Pyrol. 106, 138–146 (2014).

Article CAS MATH Google Scholar

Lee, M. W., Park, J. S., Park, D. H. & Seo, Y. B. Comparison of cellulose DP measurements using the CED (cupriethylene diamine) and NMMO(N-methylmorpholine-N-oxide). J. Korea Tech. Assoc. Pulp Paper Ind. 42, 62–66 (2010).

CAS MATH Google Scholar

Ma, Y., You, X., Nieminen, K., Sawada, D. & Sixta, H. Influence of DP and MMD of the pulps used in the Ioncell® process on processability and fiber properties. RSC Sustain.1, 1497–1510 (2023).

Article CAS Google Scholar

Hospodarova, V., Singovszka, E. & Stevulova, N. Characterization of cellulosic fibers by FTIR Spectroscopy for their further implementation to building materials. AJAC 09, 303–310 (2018).

Article CAS Google Scholar

Hong, T., Yin, J. Y., Nie, S. P. & Xie, M. Y. Applications of infrared spectroscopy in polysaccharide structural analysis: Progress, challenge and perspective. Food Chem. X. 12, 100168 (2021).

Article CAS PubMed PubMed Central MATH Google Scholar

Nelson, M. L. & O’Connor, R. T. Relation of certain infrared bands to cellulose crystallinity and crystal latticed type. Part I. Spectra of lattice types I, II, III and of amorphous cellulose. J. Appl. Polym. Sci. 8, 1311–1324 (1964).

Article CAS MATH Google Scholar

Jones, R. N., Angell, C. L., Ito, T. & Smith, R. J. D. The carbonyl stretching bands in the infrared spectra of unsaturated lactones. Can. J. Chem. 37, 2007–2022 (1959).

Article CAS Google Scholar

Iotti, M., Gregersen, Ø. W., Moe, S. & Lenes, M. Rheological studies of microfibrillar cellulose water dispersions. J. Polym. Environ. 19, 137–145 (2011).

Article CAS Google Scholar

Karppinen, A. et al. Flocculation of microfibrillated cellulose in shear flow. Cellulose 19, 1807–1819 (2012).

Article CAS MATH Google Scholar

Mitiakoudis, A. & Gandini, A. Synthesis and characterization of furanic polyamides. Macromolecules 24, 830–835 (1991).

Article ADS CAS MATH Google Scholar

Van Velzen, D., Cardozo, R. L. & Langenkamp, H. A. Liquid viscosity-temperature-chemical constitution relation for organic compounds. Ind. Eng. Chem. Fund. 11, 20–25 (1972).

Article Google Scholar

De Souza, F. M., Sulaiman, M. R. & Gupta, R. K. Materials and chemistry of polyurethanes. in ACS Symposium Series (ed. Gupta, R. K.) vol. 1399 1–36 (American Chemical Society, 2021).

Rios, L. A., Weckes, P. P., Schuster, H. & Hoelderich, W. F. Resin catalyzed alcoholysis of epoxidized fatty esters: Effect of the alcohol and the resin structures. Appl. Catal. A. 284, 155–161 (2005).

Article CAS Google Scholar

Reboredo, M. M., Rojas, A. J. & Williams, R. J. J. Kinetic and viscosity relations for thermosetting polyurethanes. Polym. J. 15, 9–14 (1983).

Article CAS Google Scholar

Lim, H., Kim, S. H. & Kim, B. K. Effects of the functionality of polyol in rigid polyurethane foams. J. Appl. Polym. Sci. 110, 49–54 (2008).

Article CAS MATH Google Scholar

Al-Moameri, H. H., Hassan, G. & Jaber, B. Simulation physical and chemical blowing agents for polyurethane foam production. IOP Conf. Ser. : Mater. Sci. Eng. 518, 062001 (2019).

Article CAS MATH Google Scholar

Obi, B. E. Overview of applications of polymeric foams. In Polymeric Foams Structure-Property-Performance 3–14. https://doi.org/10.1016/B978-1-4557-7755-6.00001-X (Elsevier, 2018).

Thirumal, M., Khastgir, D., Singha, N. K., Manjunath, B. S. & Naik, Y. P. Effect of foam density on the properties of water blown rigid polyurethane foam. J. Appl. Polym. Sci. 108, 1810–1817 (2008).

Article CAS Google Scholar

Maji, P. K. & Bhowmick, A. K. Influence of number of functional groups of hyperbranched polyol on cure kinetics and physical properties of polyurethanes. J. Polym. Sci. Part A Polym. Chem. 47, 731–745 (2009).

Article ADS CAS MATH Google Scholar

Sahai, H., Ageel, M. I. & Ageel, M. I. The Analysis of Variance: Fixed, Random and Mixed Models (Springer Science & Business Media, 2012).

Klockars, A. J., Sax, G. & Sax, G. Multiple Comparisons (Sage, 1986).

Ravey, M. & Pearce, E. M. Flexible polyurethane foam. I. Thermal decomposition of a polyether-based, water-blown commercial type of flexible polyurethane foam. J. Appl. Polym. Sci. 63, 47–74 (1997).

3.0.CO;2-S" data-track-item_id="10.1002/(SICI)1097-4628(19970103)63:13.0.CO;2-S" data-track-value="article reference" data-track-action="article reference" href="https://doi.org/10.1002%2F%28SICI%291097-4628%2819970103%2963%3A1%3C47%3A%3AAID-APP7%3E3.0.CO%3B2-S" aria-label="Article reference 85" data-doi="10.1002/(SICI)1097-4628(19970103)63:13.0.CO;2-S">Article CAS MATH Google Scholar

Kumar, M. N. S. & Siddaramaiah Thermo gravimetric analysis and morphological behavior of castor oil based polyurethane–polyester nonwoven fabric composites. J. Appl. Polym. Sci. 106, 3521–3528 (2007).

Article CAS MATH Google Scholar

Reinerte, S., Kirpluks, M. & Cabulis, U. Thermal degradation of highly crosslinked rigid PU-PIR foams based on high functionality tall oil polyol. Polym. Degrad. Stab. 167, 50–57 (2019).

Article CAS Google Scholar

Download references

The authors are grateful to the Natural Sciences and Engineering Research Council of Canada (NSERC) for its financial support through its IRC and CRD grant programs (IRCPJ 461745-18 and RDCPJ 524504-18) as well as the industrial partners of the NSERC Industrial Research Chair on Eco-responsible Wood Construction (CIRCERB).The authors would also like to acknowledge collaborators who provided technical support: Yves Bédard, Daniel Bourgault, Luc Germain, and Jean Ouellet from the Renewable Materials Research Center (CRMR—Université Laval); Professor Denis Rodrigue from the Center for Research on Advanced Materials (CERMA—Université Laval) for access to the gas pycnometer and the thermal conductivity measuring devices; Nicholas Larouche from CERMA for SEM imaging; and employees from SEREX who have been involved in any way in the analysis of samples at Karl Fischer.

Department of Wood and Forest Sciences, Faculty of Forestry, Geography, and Geomatics, Université Laval, 2405 rue de la Terrasse, Québec City, QC, G1V 0A6, Canada

Manon Beaufils-Marquet, Pierre Blanchet, Loïse Cao, Jérémy Winninger, Simon Pépin & Véronic Landry

NSERC Industrial Research Chair on Eco-responsible Wood Construction (CIRCERB), Université Laval, 2425 rue de l’Université, Québec City, QC, G1V 0A6, Canada

Manon Beaufils-Marquet, Pierre Blanchet, Loïse Cao, Simon Pépin & Véronic Landry

You can also search for this author inPubMed Google Scholar

You can also search for this author inPubMed Google Scholar

You can also search for this author inPubMed Google Scholar

You can also search for this author inPubMed Google Scholar

You can also search for this author inPubMed Google Scholar

You can also search for this author inPubMed Google Scholar

M.B.M. carried out the experimental part of this work and wrote the manuscript. L.C. took part of the experimental part of this work. J.W. contributed to the data exploitation and creation of the graphics. S.P. advised the experimental part of the project and supported the writing of the manuscript. V.L. supported the manuscript writing and supervised the project. P.B. assisted in drafting and editing the manuscript, offered financial support, and supervised both the project and its work plan.

Correspondence to Véronic Landry.

The authors declare no competing interests.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

Beaufils-Marquet, M., Blanchet, P., Cao, L. et al. Modified raw cellulose filaments material as polyol substitute in rigid insulating polyurethane foam. Sci Rep 15, 6934 (2025). https://doi.org/10.1038/s41598-025-89807-2

Download citation

Received: 10 September 2024

Accepted: 07 February 2025

Published: 26 February 2025

DOI: https://doi.org/10.1038/s41598-025-89807-2

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative