Preparation and fire behavior of rigid polyurethane foams synthesized from modified urea–melamine–formaldehyde resins

In this study, a series of ethylene glycol modified urea–melamine–formaldehyde resins (EUMFs) were synthesized from urea, melamine, paraformaldehyde and ethylene glycol, and then incorporated into rigid polyurethane foams (RPUFs) as a reactive-type liquid flame retardant. The structure of EUMFs was characterized by Fourier transform infrared spectrometry; the morphology of the foams was characterized by scanning electron microscopy; and the thermal degradation and fire behavior of RPUFs were characterized by limiting oxygen index (LOI), cone calorimetry test and thermogravimetry analysis. The results show that the incorporation of EUMFs results in an increase in thermal stability, smoke suppression and LOI of RPUFs. As the melamine loading in EUMFs increases, the peak heat release rate and the total heat release of RPUFs decrease significantly, but the LOI increases slightly. Compared with the original foam, the cells of RPUFs become less regular with nonuniform diameters. In general, EUMFs show excellent flame retardancy and smoke suppression for RPUFs.


Introduction
Rigid polyurethane foams (RPUFs) have been used in many elds due to their outstanding insulation performance, dimensional stability, and low density. [1][2][3][4] However, RPUFs are ammable because of the weak covalent bond and foam structure, and notoriously, a large amount of smoke and toxic gases are released during the combustion of RPUFs, which can have severe environmental and health consequences. [5][6][7] Obviously, there is a need to improve the ame resistance of RPUFs.
Halogen ame retardants, such as tris(2-chloroethyl) phosphate and tris(1-chloro-2-propyl)phosphate, are mostly used in RPUFs, but their applications are limited due to the release of a large amount of toxic gases during the combustion. 8 Thus, much effort has been made to develop efficient and ecofriendly halogen-free ame retardants for RPUFs, which can be roughly divided into addition-type and reactive-type. The addition-type ame retardants, such as aluminium hydroxide, expandable graphite, and ammonium polyphosphate, are mainly incorporated into the foam by means of physical mixture. 5,[9][10][11][12][13] However, the mechanical properties of RPUFs can be greatly reduced and the ame retardants would separate from the foam, due to the poor compatibility between ame retardants and foam matrix. The reactive-type ame retardants are organic compounds containing ame retardant elements, such as phosphorus, nitrogen and active functional groups, and they have more durable ame retardant performance and better compatibility with the foam matrix compared with additiontype ame retardants. 14 Melamine is oen utilized in polyurethane foams because of its stable triazine ring and high nitrogen content, and it can increase not only the re retardancy of the polyurethane foam, but also the smoke suppression during foam combustion. 15,16 Nevertheless, a high melamine loading can lead to a reduction of the mechanical strength of RPUFs and an increase of the viscosity of the foaming solution. Recently, Liu et al. and Wang et al. developed urea-melamine-formaldehyde (UMF) and melamine-formaldehyde (MF) foams with excellent ame retardancy, respectively. 17,18 However, their applications in RPUFs can be restricted by the high water content.
In this study, a series of ethylene glycol modied ureamelamine-formaldehyde resins (EUMFs) were synthesized from urea, melamine, paraformaldehyde and ethylene glycol, and then incorporated into RPUFs as a reactive-type liquid ame retardant. The effects of EUMFs, as well as the melamine loading in EUMFs, on the cell morphology, compressive strength, ammability, re behavior and thermal stability of RPUFs were investigated.

Synthesis of EUMFs
EUMFs were synthesized using a neutral-acid-base procedure, as shown in Scheme 1, and their formulations are shown in Table 1. In brief, an appropriate amount of paraformaldehyde, ethylene glycol, urea and melamine were fed into a 1000 mL four-necked round-bottomed ask equipped with a thermometer, a mechanical stirrer and a reux condenser, and the mixture was heated to 90 C and then kept at that temperature for 30 min. The pH of the mixture was adjusted to 4.5-5.0 with NH 4 Cl. The degree of condensation was measured every 5 min by the turbidity point method, and one drop of resin was dispersed in 100 mL of water. 19 When the solution became cloudy, the pH of the mixture was adjusted to 8.5-9.0 with triethanolamine. Aer that, the system was cooled to 60 C and accessional urea was added, and then the resin was allowed to mature at 60 C for 60 min. Finally, the resultant resin, a white opaque liquid, was obtained for future use. It is noted that the NH 4 Cl and triethanolamine present in the EUMF resins may have a negligible effect on the properties of the resultant products due to their small quantity and neutralization. The hydroxyl values of EUMFs were measured according to GB/T 12008.3-2009 (The Light Industry Standard of People's Republic of China), and the average of two samples was recorded.

Preparation of RPUFs
RPUFs were prepared by free-foaming according to the formulations listed in Table 2. Specically, a xed amount of polyols and EUMFs were mixed and stirred at a speed of 200 rpm for at least half an hour, and then the catalyst, surface active agent, and blowing agent were added and stirred for 1 min using a high speed mechanical stirrer. PAPI was added and stirred for 15 s at a speed of 2000 rpm, and then poured into an open mold for free foaming. The foams were placed in an oven for polymerization reaction at 70 C for 24 h, and then samples were cut for characterization. 2.4.5. Thermogravimetric analysis (TGA). TGA was performed on a NETZSCH STA 409 PC instrument at a heating rate of 10 C min À1 , and the foams were heated from 50 C to 700 C at a nitrogen ow of 50 mL min À1 .
2.4.6. Cone calorimeter testing (CCT). CCT was performed using a FTT2000 cone calorimeter instrument according to ISO 5660-1. Each sample of 97 Â 97 Â 20 mm 3 (length Â width Â thickness) was wrapped in aluminum foil and exposed horizontally to an external heat ux of 35 kW m À2 . At least three samples were tested in each experiment.
2.4.7. Scanning electron microscopy (SEM). The morphologies of foams and foam residuals aer CCT were characterized by a Hitachi S-4800 scanning electron microscope at an accelerating voltage of 15 kV. The residual surfaces were coated with a thin gold layer before SEM observation.

FT-IR spectra of EUMFs
The FT-IR spectra of urea, melamine, and four EUMF samples (EUMF-1 to EUMF-4) are shown in Fig. 1. In the FT-IR spectra of urea, the absorption peaks at 3346 and 1681 cm À1 can be assigned to N-H and C]O stretching vibration, respectively. In the FT-IR spectra of melamine, the absorption peaks at 3469, 3419, 3334 and 3129 cm À1 are assigned to -NH 2 stretching vibration; whereas those at 1654, 1550 and 813 cm À1 are assigned to the triazine ring, respectively. 20 In the FT-IR spectra of EUMFs, the peaks at 1135 and 1253 cm À1 are attributed to the absorption of C-O-C and -CH 2 of -CH 2 -O-CH 2groups, respectively. The absorption peak at 3351 cm À1 in the FT-IR spectra of EUMF-1 is assigned to -NH-, and those at 3419 and 3349 cm À1 in the FT-IR spectra of EUMF-2, EUMF-3, and EUMF-4 are assigned to -NH 2 , respectively. 21,22

13 C NMR analysis of EUMF-2
All EUMF samples show similar 13 C NMR spectra. As an example, Fig. 2 shows the 13 C NMR spectrum of EUMF-2, and its  Table 2 The formulations of RPUFs   19,24 All these results indicate the occurrence of amine-aldehyde condensation reaction in the system.

Cell morphology, apparent density and physicalmechanical properties of RPUFs
The effect of EUMFs on the microstructure of RPUFs was examined by SEM. The cells of RPUF-0 are quadrilateral or hexagon shaped with a uniform diameter (Fig. 3), while those of RPUFs become less regular in shape with a larger and nonuniform diameter. This can be attributed to the branched structure of EUMFs, which can have a signicant effect on the mechanical properties of RPUFs. Given the remarkable effects of foam apparent density on the mechanical properties, ammability, and combustion performance, the apparent density is controlled at about 50 kg m À3 . 25 The foams were compressed to 10% of their original thickness at a crosshead speed of 2.5 mm min À1 , and their physical and mechanical properties were determined, as shown in Table 4. The maximum compressive strength is observed in RPUF-0, indicating that the addition of EUMFs can impair the compressive strength of RPUFs, which is likely due to the destruction of cell structure. However, as the melamine loading in EUMFs increases from 18.9 g to 75.6 g, the compressive strength of RPUF-4 is increased by 24.62% compared with that of RPUF-1, which is mainly due to the high stability of triazine ring that can improve the stiffness of RPUFs. 17

Flammability of RPUFs
The ammability of RPUFs was characterized by LOI. As shown in Table 5, the LOI value of RPUF-0 is 18.1%, and it is increased to about 24% with the incorporation of EUMFs. The improved ame retardancy of RPUFs can be attributed to the release of non-ammable gases and the increase of char yield during the combustion of RPUFs. 17 Nevertheless, increasing the melamine loading in EUMFs has a negligible effect on the LOI values of RPUFs.

Fire behaviors of RPUFs
The re behaviors of RPUFs were characterized by CCT at a heat ux of 35 kW m À2 , as shown in Fig. 4 and 5 and Table 6. In this study, the parameters of interest include the time to ignition (TTI), heat release rate (HRR), peak heat release rate (PHRR) and total heat release (THR). As shown in Table 6, the TTIs of both ame retardant foams and original foams are very short due to the cell structure of RPUFs. HRR is considered as a measure of re intensity. As shown in Fig. 4(a), there are two peaks in the HRR curve of RPUF-0. The rst peak is caused by the formation of a char layer on the surface of RPUF-0 during combustion. The inner polymer is exposed to the ame due to the pyrolysis of the char layer, resulting in the formation of a second HRR peak at about 60 s followed by a sudden reduction. 26 The addition of EUMFs results in a decrease in PHRRs of RPUFs to 173.7, 166.5, 160.2, and 146.4 kW m À2 , respectively, and a delay of the second HRR peak to about 120 s. It can also be observed that melamine plays an important role in reducing the PHRRs of RPUFs. The PHRRs of RPUFs gradually decrease with increasing melamine loading in EUMFs, which is attributed to the formation of melam during the thermal condensation of melamine. 27 The THRs of RPUFs are decreased by 17.6, 19.2, 24.3, and 26.7% compared with that of RPUF-0, respectively, indicating that the incorporation of EUMFs can effectively reduce the combustion intensity of RPUFs. This may be because (1) EUMFs decompose at a lower temperature than polyurethane foam and release non-ammable gases, such as NH 3 , HNCO, and HCN, which can reduce the concentration of ammable fragments; and (2) the incorporation of EUMFs enables the char layer to have  better anti-oxidation properties, so that the HRR is lower than that of original foam.
The smoke emission behaviors of RPUFs were characterized by the smoke production rate (SPR), total smoke production (TSP), rate of smoke release (RSR), total smoke release (TSR), and weight ratio of CO to CO 2 . As shown in Fig. 5, the SPRs and RSRs of RPUFs are quite different from those of RPUF-0. The SPR and RSR curves of RPUF-0 show a sharp and strong peak at about 65 s, while those of RPUFs show two peaks of similar strength. RPUF-3 shows the best smoke suppression performance. Table 6 shows that the SPR and TSP of RPUF-3 are reduced by about 67.6 and 42.7% compared with that of RPUF-0, respectively. This is probably due to the reaction between melamine and aromatic hydrocarbon, which is the primary source of smoke during combustion. 15,28 However, the release of the smoke can be restricted by the compact nitrogen-rich carbon layer formed in the initial stage of combustion. Besides, as shown in Table 6, the PSPR, TSP, PRSR, TSP, and CO/CO 2 weight ratio show no obvious decrease. The TGA results of RPUFs show that the initial degradation temperature of RPUF-4 is lower than that of RPUF-3, indicating that RPUF-4 is degraded earlier than RPUF-3. As a consequence, melamine is released earlier in RPUF-4, resulting in a decrease in the reaction between melamine and aromatic hydrocarbon. However, the difference is not so obvious.     The CO yield and CO/CO 2 weight ratio of RPUFs are important factors for evaluating the anti-re performance. 29 As shown in Table 6, the CO/CO 2 weight ratio of RPUF-3 is reduced by 52.7% compared with that of RPUF-0, which is mainly attributed to the compact char layer. The more the decomposed fragments are reserved in the residue, the less the fragments are burned in re and consequently the less the CO is released.

Thermal stability of RPUFs
TGA is widely used to evaluate the thermo-oxidative degradation behaviors of different materials. 26 TGA and derivative thermogravimetric analysis (DTG) curves of all foams under nitrogen atmosphere are shown in Fig. 6, and the data is shown in Table 7. As shown in Fig. 6, all RPUFs are decomposed in three stages. The rst decomposition stage of RPUF-0 begins at about 243 C due to the scission of C-O bonds in the urethane group, resulting in the formation of isocyanates and polyols. As the pyrolyzation proceeds, imidodicarbonic diamide is released in the self-reaction of partial isocyanates, which is accompanied by the volatilization of carbon dioxides, alcohols, amines, aldehydes and carbon monoxides. 30 The second decomposition stage is observed at 282-430 C, which corresponds to the degradation of imidodicarbonic diamide and substituted urea generated in the reaction between isocyanates and polyols or water. 31 The third decomposition stage starts at about 430 C due to further decomposition of char residue.
The initial decomposition temperature of RPUFs is slightly lower than that of RPUF-0, because EUMFs can be decomposed more easily. For RPUFs, the EUMFs can decompose into NH 3 , Fig. 4 The HRR curves (a) and THR curve (b) of RPUFs. Fig. 5 The SPR curves (a) and RSR curves (b) of RPUFs. HNCO, HCN, melamine and some stable intermediates at about 170 C. Melem is a direct condensation product of melamine at high temperatures, which can promote the formation of a dense char layer on the surface of the burning material. 27 As a consequence, the residue char of RPUFs is signicantly increased. For instance, the char residue of RPUF-4 is increased to 29.8%. The dense char layer plays a crucial role in isolating the polymer matrix from heat. As shown in Fig. 6(b), the maximum-rate degradation temperature of the second and third decomposition stages of RPUFs is similar to that of RPUF-0, and the addition of EUMFs results in a decreases of decomposition rate. Compared with RPUF-0, the char residues of RPUFs become dense, which contributes signicantly to the good ame retardancy of RPUFs as it can isolate the inner polymer from heat and oxygen. 26

Conclusions
In this study, a series of EUMFs were synthesized by aminealdehyde condensation and aldolization, and then incorporated into RPUFs to prepare reactive-type ame retardants. The incorporation of EUMFs can adversely affect cell morphology and compressive strength of RPUFs, but result in an increase in the LOI value of RPUFs to about 24%. Nevertheless, the melamine loading in EUMFs has a negligible effect on the LOI value of RPUFs. The CCT results reveal that RPUFs exhibit good re resistance and smoke suppression. In particular, the TSP, TSR and CO/CO 2 weight ratio of RPUFs decrease signicantly with the incorporation of EUMFs, and that of RPUF-3 are decreased by about 42.7%, 42.7% and 52.7% in comparison with that of RPUF-0, respectively. The dense char layer contributes signicantly to the good re resistance and smoke suppression of RPUFs.

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