Effect of flame retardants on mechanical and thermal properties of bio-based polyurethane rigid foams

A soy oil-based polyol (HSBP) was synthesized from epoxidized soy oil through a ring-opening reaction with distilled water. A phosphorus-containing flame retardant (DOPO–HSBP) was synthesized through the reaction of 9,10-dihydro-9-oxa-10-phosphaphenanthrene-10-oxide (DOPO) and HSBP. A nitrogen-containing flame retardant (T–D) was prepared by the reaction of diethanolamine with glycol diglycidyl ether. The structures of HSBP, DOPO–HSBP, and T–D were characterized by Fourier transform infrared spectroscopy (FT-IR) and nuclear magnetic resonance (1H NMR). The flame-retardant rigid polyurethane foam (PPUFs and NPUFs) was prepared successfully by mixing HSBP, DOPO–HSBP, and T–D. The effects of DOPO–HSBP content on the mechanical, thermal, and flame-retardant properties of PPUFs and NPUFs were investigated by tensile tests, thermogravimetric analyses (TGA), limiting oxygen index (LOI), and UL-94 vertical burning level. The morphology of PPUFs and NPUFs was studied via scanning electron microscopy (SEM). With the increase in the percentage of DOPO–HSBP added, the flame retardant property of rigid polyurethane foam (RPUF) was greatly improved. When the phosphorus-containing flame retardant DOPO–HSBP was added to 50% of the RPUF with the nitrogen-containing flame retardant T–D, the LOI value of the foam increased from 18.3 to 25.5, and the UL-94 result was classified as “V-0” with almost no effect on the mechanical properties of the RPUF. The results showed that the phosphorus and nitrogen synergistic flame retardants of DOPO–HSBP and T–D can endow excellent flame retardant properties to RPUF without affecting its mechanical properties.


Introduction
Rigid polyurethane foam (RPUF) is an economical and efficient energy-saving material that is widely used in elds of furniture, transportation, construction materials, refrigerator insulation because of its light weight, high strength, and excellent mechanical and thermal insulation properties. [1][2][3][4] However, its porous structure makes it easy to be ignited and burn rapidly aer being exposed to re, releasing large amounts of heat and smoke, leading to serious re accidents. 5,6 This limits the application of rigid polyurethane foam in many elds. Therefore, the study of ame retardant rigid polyurethane foams is of great importance.
In order to improve the ame retardancy of RPUF, adding ame retardants to RPUF would be effective. 6 There are two main approaches to add ame retardants to RPUF: additive-type ame retardants and reactive-type ame retardants. 7,8 Examples of additive ame retardants include expandable graphite (EG), 9,10 ammonium polyphosphate (APP), 11 melamine and its derivatives, 12,13 inorganic phosphorus-containing compounds, 14 etc. However, they are gradually replaced by another way due to their poor compatibility, the large negative impact on the mechanical properties, and the easy leaching of the RPUF. 15 Reactive ame retardants are more stable in RPUF, and the ame retardant element P/N is introduced into the main chain of RPUF by participating in the foaming reaction to achieve the ame retardant effect. [16][17][18][19] DOPO and its derivatives are some of the reactive ame retardants that have attracted much attention in recent years. Compared with other straight-chain small molecule ame retardants, DOPO and its derivatives have better stability due to their aromatic structure. 20,21 For instance, in the work of Wang et al., a bifunctional ame retardant (PDEP) based on DOPO and phosphate was synthesized, and the LOI of RPUF was increased from 18.5 to 22.9 by the addition of this ame retardant. 22 However, a single ame retardant system can only improve the ame retardancy of the foam to a very limited extent. Some studies indicate that the synergistic effect of phosphorus and nitrogen can greatly improve its ame retardant efficiency. 23 In addition to the ame retardancy of rigid polyurethane foams, biodegradability has also been a major concern for environmental protection and sustainable development in recent years. The adoption of bio-based polyols to prepare RPUF would be a good answer for those concerns. 24 Vegetable oil can be used to produce bio-based RPUF due to the presence of many reactive sites on its aliphatic chain. For example, Veronese et al. synthesized RPUFs by using soybean oil or castor oil. 25 Guo et al.
prepared soy-based polyol from epoxidized soy oil ring-opened by methanol and the resulting RPUF exhibited comparable mechanical and insulating properties to other foams from petrochemical feedstocks. 26 Ji et al. synthesized different soybased polyols by reacting epoxidized soy oil with methanol, phenol, and cyclohexanol. At 25 wt% of soy-based polyol, the introduction of phenol can improve the mechanical and thermal properties of the foam. 27 The purpose of this work is to synthesize soybean oil-based polyol as a raw material for RPUF, a polyol DOPO-HSBP as a P-containing ame retardant, and polyol T-D as an Ncontaining ame retardant. The effects of the content of DOPO-HSBP and the phosphorus-nitrogen synergistic effect of T-D addition on the mechanical properties and ame retardant properties of RPUF were analyzed using the mechanical test, scanning electron microscopy (SEM), thermogravimetric analysis (TGA), LOI, and UL-94. Finally, the RPUFs with greatly improved ame retardancy and almost no effect on mechanical properties were successfully synthesized.

Preparation of DOPO-HSBP
The DOPO-HSBP was prepared by reacting DOPO with HSBP. During a process, 40 g HSBP, 25.8 g DOPO, 12.5 g TEA, and 40 g dichloromethane were added to a round-bottomed threenecked ask equipped with a mechanical stick. Carbon tetrachloride was added dropwise to the reaction mixture at ice bath over 30 minutes. The temperature of the reaction was maintained at 25 C for 24 h. Aer the reaction completion, the reaction mixture was extracted with ethyl acetate and washed with deionized water. Ethyl acetate and deionized water were removed by rotary evaporator. The preparation route of DOPO-  HSBP was shown in Scheme 1. The properties of the HSBP and DOPO-HSBP were listed in Table 1.

Preparation of T-D
Diethanolamine (54 g) and glycol diglycidyl ether (62 g) were charged into a round-bottomed three-necked ask equipped with a mechanical stick and a thermometer. The temperature of the reaction was maintained at 85 C for 6 h. Aer the reaction completion, the resulting product can be used for the next step without any treatment. The preparation route of T-D was shown in Scheme 2.

Preparation of rigid polyurethane foams (RPUFs)
The RPUFs were prepared using a free-rise method according to the formulation as shown in Table 2. The -NCO/-OH ratio of the systems was set as 1.1. The polyol blends were mixed with surfactant (AK8805), catalyst (dibutyltin dilaurate), and blowing agent (distilled water), in the required proportions under ambient conditions for approximately 120 s. The pMDI was added quickly into the mixture and mixed for another 15 s. Finally, the mixture was immediately poured into an open mold to produce free-rise foam. The obtained RPUFs were completely cured at room temperature for 7 days before analysis. The preparation route and structure of RPUFs are shown in Scheme 3.

Characterization
FT-IR spectra were measured by Spectrum One PerkinElmer Fourier transform infrared spectrometer (PerkinElmer Co., American). The FT-IR spectra were collected using 64 scans in the wavenumber range of 4000-400 cm À1 with a resolution of 4 cm À1 .
1 H NMR spectra were tested by Varian Inova 600 nuclear magnetic resonance spectrometer (American Varian Co., American) at room temperature, using tetramethylsilane (TMS) as a reference, CDCl 3 as solvent for sample.  The density of RPUF samples was measured according to the American Society of Testing Materials (ASTM) designation: D162208 (2008). The average values of ve samples were recorded.
The compressive strength of foams was recorded on a CMT4000 universal testing machine according to (Shengzhen, China) according to ASTM designation: D1621-10. Each sample used for the test was 50 Â 50 Â 50 mm 3 (length Â width Â height). At least ve samples were tested to obtain average values in mechanical tests.
A NETZSCH 209F1 TA Instruments was employed for thermogravimetric analysis (TGA), and the RPUF samples were heated to 750 C at a heating rate of 10 C min À1 under a dynamic nitrogen ow of 50 mL min À1 .
The morphologies of the foams were analyzed with a JSM-6510LV scanning electron microscope (SEM) operating at 20 kV, and the samples were sputter-coated with a thin layer of gold. The scanning electron microscopy was measured along parallel foam rise direction.

Results and discussion
FT-IR spectra of HSBP, DOPO-HSBP, DEA, TMA, and T-D The FTIR spectra of HSBP and DOPO-HSBP are shown in Fig. 1. The aromatic C-H absorption around 3066 cm À1 . 32 The absorption around 755 cm À1 and 1596 cm À1 corresponds to vibration with Ph-P and C]C stretching in aromatic functional groups respectively. 33,34 The absent peak at 2385 cm À1 (P-H) indicated that the reaction of HSBP and DOPO has done. 35 Fig. 2 shows the FTIR of TMA, DEA, and T-D. It can be observed from the spectrum of TMA that the peak at 912 cm À1 , 852 cm À1 , 1254 cm À1 , and 757 cm À1 corresponds to epoxy bond characterized absorbing. In the FTIR spectra of T-D, the disappearance of the epoxy absorption peak. The stretching vibration peak at 1100 cm À1 and the deformation vibration peak at 1455 cm À1 in TMA corresponds to C-O-C and -CH 2 respectively still exist. The characterized absorbing peak at 1125 cm À1 and    Fig. 4 compares the 1 H NMR spectrum of T-D with that of TMA. We can see that a new peak is observed at 4.89 ppm (peak g). Furthermore, the peak related to CH-O-C of TMA at 3.17 ppm (peak b) disappears in the spectra of T-D. All the aforementioned results suggested the successful synthesis of T-D.

Apparent density
The apparent density of PPUFs and NPUFs prepared from blends with different ratios of DOPO-HSBP is shown in Fig. 5. The relevant literature shows that the apparent density of rigid polyurethane foam is closely related to the compressive strength. 38 It can be observed from Fig. 5 that the apparent density of both PPUFs and NPUFs tends to decrease with the increase of DOPO-HSBP content. When the content of DOPO-HSBP increased from 0 to 50%, the apparent densities of PPUFs and NPUFs decreased by 18% and 30%, respectively. These changes could be attributed to the decrease in crosslink density of the foam due to the addition of DOPO-HSBP. However, it is not difficult to nd that the apparent density of NPUF-50 is still higher than that of neat RPUF. This indicates that the moderate addition of T-D and DOPO-HSBP has almost no effect on the physical properties of the foams.

Microphotographs
The morphology changes of PPUFs and NPUFs are shown in Fig. 6. A typical RPUF structure was observed in Fig. 6, with open and closed pores of the sphere or polygonal shape. By adding phosphorus-containing ame retardant DOPO-HSBP, it can be seen that there is a signicant increase in the cell size compared  with the PPUF-0, but with the further increase of DOPO-HSBP content from 10% to 50%, the cell size of the foam is reduced by 14%. This may be due to the decrease in the hydroxyl value of DOPO-HSBP compared to HSBP, which leads to a decrease in the cross-link density of the foam, a decrease in the number of pores, and an increase in the cell sizes. When the content of DOPO-HSBP increases further, the cell growth in the foaming process is weakened due to its higher viscosity, which eventually leads to smaller pores. On the other hand, it can be clearly observed that when the N-containing ame retardant T-D is added, there is a slight increase in the foam pore size compared to the foam without the ame retardant added. Comparing PPUF and NPUF, we can see that when T-D is added, the pore size still increases and then decreases with the increase of DOPO-HSBP content, but the cell size of NPUF-50 has been reduced to a size close to RPUF-0. This may be due to the high hydroxyl value and low viscosity of T-D. This means that NPUF-50 has little effect on the foam cell size compared with the RPUF-0, and the foam can still retain its original excellent properties such as thermal insulation.  Table 3. At low strains, the foam exhibited linear-elastic deformation, followed by a continuous deformation platform, which may be the result of brittle fracture of the pore structure of the polyurethane foam, and at high strains, the densication of the foam caused the polymer to harden, making stress continues to increase. 39 It was reported in the literature that the mechanical properties of RPUFs are closely related to the hydroxyl value of the polyol, the size of the foam pores, and the density of the foam. 40 As shown in Table 3 , which leads to a lower cross-link density of the foams, thus resulting in larger foam pores and ultimately a reduction in the compressive strength of the foams. The addition of T-D greatly compensates for this shortcoming, as its higher hydroxyl value greatly increases the crosslink density of the foams and acts as a hard segment during synthesis, increasing the composition of the hard segment of the foams. This leads to an overall increase in the compressive strength of the foams. According to GB T 21558-2008, it can be observed that the compressive strength of all the foams meets the standard of compressive strength ($180 kPa) in rigid polyurethane foams for building insulation. Therefore, the potential application of soy oil-based rigid polyurethane foams in the eld of building insulation can be illustrated.

Dimensional stability
In addition to mechanical strength, dimensional stability is another important characteristic of rigid polyurethane foam used in roong, insulating, or any other constructing materials. Standard specications for dimensional stability had been reported to be less than 3% of linear change at 70 C for 24 h. 41 From Table 3, we can see that the dimensional changes of all the foams are less than 3% and the dimensional changes of the foams aer T-D addition are less than 1%, so the dimensional changes of all the foams meet the standard specications for dimensional stability.

Thermogravimetric analysis
To evaluate the thermal stability of the prepared PPUFs and NPUFs, TGA is conducted under the ow of nitrogen. The TG and DTG curves of PPUF and NPUF samples are illustrated in Fig. 8, and the representative parameters are summarized in Table 4. The degradation of foams can be divided into two stages: the degradation in the rst stage of weight loss at 300 C to 350 C is due to the decomposition of carbamate bonds in the hard segment, and the degradation in the second stage of weight loss at 450 C to 500 C is attributed to the thermal degradation in the so segment. 42 Conventionally, the thermal stability of RPUFs is described by the temperatures of 5% weight loss (T 5% ) considered as the temperature for the onset of degradation. 43 It can be seen that as the content of ame retardant DOPO-HSBP increased to 50%, the T 5% of the foam decreased from 277 C to 269 C, and the T 5% of the foam further decreased to 243 C aer the addition of ame retardant T-D. The results showed that the initial decomposition temperature of RPUF showed a decrease aer the addition of DOPO-HSBP and T-D. This may be caused by the lower decomposition temperature of DOPO-HSBP and T-D. In addition, it can be seen from T max1 and T max2 that the temperature of all PPUFs and NPUFs was increased at T max , which indicates that the ame retardant will improve the heat endurance of RPUF.
According to the TGA results of the prepared foams, the char residues of the foams prepared with different ame retardants were enhanced compared to the pure RPUF. Noticeably, the char residue of NPUF-40 has the largest increase, reaching 13.93%, which indicates that the addition of DOPO-HSBP and T-D can signicantly increase the char residue of RPUF. 44

Flame retardancy and combustion behaviors
To evaluate the ammability of PPUFs and NPUFs, the results of LOI and UL-94 tests for PPUFs and NPUFs with different mass ratios of DOPO-HSBP are presented in Table 5, and digital photos of the vertical burn test (UL-94) are shown in Fig. 9. With the addition of ame retardant DOPO-HSBP, the LOI value of the foam increased from 18.3 to 25.2, and the UL-94 result was classied as "V-0 rating", which signicantly improved the ame retardancy. When ame retardant T-D was added, the possible synergistic effect of phosphorus (DOPO-HSBP) and nitrogen (T-D) could improve the ame retardant activity in the gas-phase and condensed-phase, which increased the LOI value to 25.5 and further improved the ame retardant effect. 45 Combined with the UL-94 data in the table and the digital photo gure, it can be seen that the NPUF-50 foam can be selfextinguished and no dripping appears in a short time aer burning, which effectively prevents the spread of re. The formation of a char can be observed on the surface aer combustion, which can effectively prevent heat and oxygen from diffusing into the internal matrix of the polymer and improve the ame retardancy of the foam. 46 All of the above revealed the addition of DOPO-HSBP and T-D is benecial to improve the ame retardant properties of the foam.

Conclusions
In this study, a phosphorus-containing bio-based polyol DOPO-HSBP and a nitrogen-containing polyol T-D were synthesized and added as ame retardants to an environmentally friendly soy oil-based rigid polyurethane foam prepared from soybean oil. It was observed that the synergistic effect of phosphorus and nitrogen of ame retardants DOPO-HSBP and T-D greatly improved the ame retardancy of RPUF by LOI and UL-94, and the LOI value increased from 18.3 to 25.5, and the UL-94 grade was improved to "V0", this may be because DOPO-HSBP and T-D are decomposed from the foam at a lower temperature, thus preventing further burning of the foam earlier and more efficiently. The compressive strength and SEM of the foam showed that both the compressive strength and the size of the pores of the NPUF-50 sample were almost unchanged compared to the polyurethane rigid foam without ame retardant. Therefore, with the addition of ame retardants DOPO-HSBP and T-D in appropriate amounts, it can not only greatly improve the ame retardant properties of the foam but also has little effect on the mechanical properties of the foam. Moreover, the compressive strength and dimensional stability of rigid polyurethane foam prepared from bio-based polyols meet the standards of construction and other elds. From the TG and DTG data of the foams, it can be observed that the addition of DOPO-HSBP and T-D greatly improved the T max and carbon residue of the foams, which also conrms the improvement of the ame retardant properties of the foams by DOPO-HSBP and T-D. Therefore, it can be seen that the phosphorus and nitrogen synergistic ame retardant system for foam ame retardant performance is much higher than the phosphorus ame retardant system alone. This research would help us to further research and develop novel bio-based RPUF materials with excellent ame-retardant effects.

Conflicts of interest
There are no conicts to declare.