Summary: The curing kinetics of diglycidyl ether of bisphenol A (DGEBA) and 4,4′-diaminodiphenylmethane (DDM) was analyzed using isothermal differential scanning calorimetry (DSC) modes by using a simple mechanistic model which includes two rate constants, k1 and k2, two reaction orders, n1 and n2, and the ratio of initial concentration of hydroxyl group to initial epoxy concentration, c0. Analyses of DSC data indicated that an autocatalytic reaction existed in the curing process. The mechanistic model proposed in this paper fits the experimental data exactly. Rate constants, k1 and k2 have been found to increase with rising curing temperature. The activation energies for the relative reactions were determined to be 66.00 ± 4.21 and 50.74 ± 8.92 kJ/mol, respectively. The complex equivalent constant, K, decreased with increasing temperature. Diffusion control was incorporated to describe the cure in the latter stages.
Comparison of experimental data with the mechanistic model for the curing kinetics of DGEBA with DDM.
DOI: 10.1039/C6RA13168A (Paper) RSC Adv., 2016, 6, 59226-59236
Impact of halogen-free flame retardant with varied phosphorus chemical surrounding on the properties of diglycidyl ether of bisphenol-A type epoxy resin: synthesis, fire behaviour, flame-retardant mechanism and mechanical properties†
Received 20th May 2016 , Accepted 30th May 2016
First published on 1st June 2016
This work aimed to investigate the effect of two types of phosphorus-containing flame retardants (P-FRs) with different chemical surroundings (phenylphosphonate-based (PO–Ph) and phenylphosphoric-based (PO–OPh)) on the flame-retardant efficiency for diglycidyl ester of bisphenol-A type epoxy (EP) resin. The two series of P-FRs which were named as FPx and FPOx (x = 1, 2 and 3), respectively, showed reactivity with epoxy group that was examined by differential scanning calorimetry (DSC) and variable temperature FTIR spectroscopy (VT-FTIR). A comparative study between the FPx and FPOx (x = 1, 2 and 3) containing flame-retardant epoxy was carried out via investigating their flammability, thermal stability and mechanical properties. The most significant difference in flame retardancy between them was that FPx (x = 1, 2 and 3) endowed EP with a V-0 rating in UL 94 test at 5 wt% loading, while FPOx (x = 1, 2 and 3) showed no rating at such loading. Importantly, it is found that there was almost 10 times difference in the flame-retardant efficiency for EP between FPx and FPOx, though they had similar chemically molecular structures. Moreover, TGA-FTIR and TGA-MS coupling techniques (TGA, thermogravimetric analysis; MS, mass spectroscopy) were employed to study the thermal decomposition of FP1 and FPO1; the impacts of FP1 and FPO1 on the thermal decomposition of EP were studied by TGA-FTIR as well. Furthermore, an online temperature detection experiment was designed to collect the temperatures by thermocouples and infrared thermometers in the UL 94 test. Based on the above results, the flame-retardant mechanisms of FP1 and FPO1 in EP are discussed. In addition, the impact of P-FRs on mechanical properties of EP was studied by means of tensile test and dynamic mechanical analysis.
1. IntroductionEpoxy resin, one of the most popular polymeric materials, plays an indispensable role in many application fields, such as coatings, transportation, and electronic and electrical industrials (EE), due to its high performances in mechanical, electrical and chemical resistance.1–3 Bifunctional diglycidyl ester of bisphenol A (DGEBA) based epoxy resins are often utilized for semiconductor encapsulation in EE applications. A common curing agent for DGEBA is 4,4-diaminodiphenylsulfone (DDS). Nevertheless, the intrinsic flammability of DGEBA/DDS is one of the challenges for their products.4–9 In the past two decades, many research groups and industries have devoted themselves to improving the flame retardancy of epoxy, such as using halogenated flame retardants, nano-clays, organophosphorus compounds and polyhedral oligomeric silsesquioxane.5–17 Generally, phosphorus-containing flame retardants (P-FRs) are regarded as one of the best choices to reduce flammability of epoxy resin since lots of halogenated flame retardants are forbidden owing to their serious hazards for the environment and humans.7–10
9,10-Dihydro-9-oxa-10-phosphaphenanthrene 10-oxide (DOPO) and its derivatives are considered as some of the most competitive phosphorus flame retardants for epoxy resin.12–34 The addition of 6.3 wt% of DOPO to DGEBA/DDS led to a high LOI value of 35%; however, the samples did not pass any rating in UL 94 test.31 Recently, Artner et al. synthesized DOPO-based diaminic hardener for DGEBA/DDS carbon fiber-reinforced composites. The LOI values of flame-retardant composites were increased by 17% and the UL 94 test results improved from an HB to a V-1 classification with around 1 wt% of phosphorus content.12 Li et al. found a good synergistic effect between polyhedral oligomeric octadiphenylsulfonylsilsesquioxane (ODPSS) and DOPO on flame-retardant epoxy resins. Epoxy with 2.5 wt% ODPSS and 2.5 wt% DOPO passed V-0 rating in UL 94 test.32 More recently, it was reported that DGEBA/DDS system with 5 wt% of TNTP passed V-0 rating in UL 94 test.33 Overall, most P-FRs reported in the above literature performed well in flame retardant tests by means of large quantity of loading or higher content of phosphorus in epoxy resins.12,18,20,35–37 In 2015, Schartel et al. proposed a two-step approach for condensed phase action to interpret the P-FR flame-retardant mechanism for two epoxy resins of DGEBA/DDS and RTM 6.34 Since P-FRs are regarded as very promising and effective non-halogenated flame retardants for polymers, in particular for epoxy resin, further understanding about the impact of the phosphorus chemical surrounding of P-FRs on the flame-retardant efficiency is very important for the development of high-efficiency flame retardants for epoxy resin.
This work mainly aimed to study the impact of phosphorus chemical surroundings on the flame-retardant efficiency of P-FRs in EP (DGEBA-type epoxy resin/DDS). Two series of reactive P-FRs (phenylphosphonate-based (PO–Ph) and phenylphosphoric-based (PO–OPh)) were synthesized with minor difference in the phosphorus chemical surroundings. The two series of P-FRs were named as FPx and FPOx (x = 1, 2 and 3), respectively. LOI, UL 94 and cone calorimeter tests were used to evaluate the impact of FPx and FPOx (x = 1, 2 and 3) on improving the flame retardancy of EP. In order to simplify the study, FP1 and FPO1 were chosen as one representative pair to study the reactivity and flame-retardant mechanism of synthesized P-FRs in EP. First of all, the reactivity was characterized by means of heat flow curves in DSC test and variable temperature infrared spectra of EPC, EPC/FP1 and EPC/FPO1 from 100 to 220 °C. A relative IR absorbance intensity ratio of R (i(epoxy group)/i(ether group)) was defined in order to described the changes of epoxy group during the heating. Aiming to understand the flame-retardant mechanism of FP1 and FPO1 in EP, comprehensive research was done as follows: TGA-FTIR and TGA-MS coupling techniques were employed to study the evolved gaseous products of FP1 and FPO1 during thermal decomposition; the impact of FP1 and FPO1 on thermal decomposition of EP was studied by TGA-FTIR; an online temperature detection experiment was designed to record the sample temperatures near to the surface and on the surface by thermocouples and infrared thermometers respectively, aiming to describe the heat phenomena in UL 94 test. In addition, tensile and dynamic mechanical analysis (DMA) tests were used to evaluate the impact of P-FRs on tensile strength and viscoelastic properties of EP.
2.1 MaterialsPhenylphosphonic dichloride (PPDCl, 90%), phenyl dichlorophosphate (PDClP, 95%), allylamine (AA, 98%), n-octylamine (OA), 2-phenylethylamine (PA), diethyl ether, potassium bromide (KBr), triethylamine (TEA) and curing agent DDS were purchased from Sigma-Aldrich Corporation. Epoxy resin (brand name, EPC; epoxy equivalent, 0.54) was supplied by Faserverbundwerkstoffe® composite technology.
2.2 General procedure for synthesis of P-FRsThe modified general synthetic procedure of P-FRs was developed from ref. 38. In a 500 mL three-neck flask, appropriate amine compound (AA, OA or PA) (0.21 mol) and TEA (0.2 mol) were dissolved in diethyl ether at 0–5 °C. Then a mixture of appropriate chloride compound (PPDCl or PDClP) (0.1 mol) and diethyl ether (100 mL) was added dropwise into the flask at 0–5 °C. The reaction mixture was stirred at 0–5 °C for 2 h. Then, stirring was continued for 5 h at room temperature (RT). After the reaction finished, the solid–liquid mixture was filtered off to remove triethylamine hydrochloride salt and the filtrate was evaporated under vacuum to obtain the crude product of P-FR. The product was purified by washing with deionized water three times. In this work, PO–Ph FRs and PO–OPh FRs were named as FPx and FPOx (x = 1, 2 and 3), respectively, in order to simplify the naming of synthesized P-FRs. Scheme 1 shows the synthetic routes and structures of FPx and FPOx (x = 1, 2, and 3).
|Scheme 1 The synthetic routes to and structures of P-FRs (FPx and FPOx, x = 1, 2 and 3).|
Synthesis of N,N′-dioctyl-P-phenylphosphonicdiamide (FP2). Preparation of FP2 followed the procedure described in 2.2 by using amine compound OA (0.21 mol) and chloride compound PPDCl (0.1 mol). FP2 was obtained as a white solid. FP2: m.p. 75 °C; yield, 80%; 1H-NMR (400 MHz, DMSO-d6, δ) (ppm): 7.7–7.4 (Ar–H, 5H), 2.7 (–CH2(NH), 4H), 1.3–1.5 (–CH2–, 12H), 0.8 (–CH3); 13C-NMR (100 MHz, DMSO-d6, δ) (ppm): 136.8 (135.2), 131.8, 131.1, 128.4, 44.4, 31.0, 28.2, 23.3, 22.0, 13.8; 31P-NMR (DMSO-d6, δ) (ppm): 20.4.
Synthesis of N,N′-diphenylethyl-P-phenylphosphonicdiamide (FP3). Preparation of FP3 followed the procedure described in 2.2 by using amine compound PA (0.21 mol) and chloride compound PPDCl (0.1 mol). FP3 was obtained as a light yellow solid. FP3: m.p. 85 °C; yield, 85%; 1H-NMR (400 MHz, DMSO-d6, δ) (ppm): 7.7–7.2 (Ar–H, 15H), 2.7–3.1 (–CH2, 8H); 13C-NMR (100 MHz, DMSO-d6, δ) (ppm): 140.4, 136.5 (135.0), 131.8, 128.9, 126.6, 42.5, 38.7; 31P-NMR (DMSO-d6, δ) (ppm): 20.1.
Synthesis of N,N′-diamyl-P-phenylphosphoricdiamide (FPO2). Preparation of FPO2 followed the procedure described in 2.2 by using amine compound OA (0.21 mol) and chloride compound PDClP (0.1 mol). FPO2 was obtained as a colorless liquid. FPO2: m.p. 73 °C; yield, 95%; 1H-NMR (400 MHz, DMSO-d6, δ) (ppm): 7.7–7.4 (Ar–H, 5H), 2.6 (–CH2(NH), 4H), 1.3–1.5 (–CH2–, 12H), 0.9 (–CH3); 13C-NMR (100 MHz, DMSO-d6, δ) (ppm): 152.5, 129.9, 124.1, 121.1, 41.3, 31.9, 29.5, 26.9, 22.7, 14.6; 31P-NMR (DMSO-d6, δ) (ppm): 20.4.
Synthesis of N,N′-diphenylethyl-P-phenylphosphonicdiamide (FPO3). Preparation of FPO3 followed the procedure described in 2.2 by using amine compound PA (0.21 mol) and chloride compound PDClP (0.1 mol). FPO3 was obtained as a yellow solid. FPO3: m.p. 82 °C; yield, 95%; 1H-NMR (400 MHz, DMSO-d6, δ) (ppm): 7.4–7.1 (Ar–H, 15H), 2.7–3.1 (–CH2, 8H); 13C-NMR (100 MHz, DMSO-d6, δ) (ppm): 152.3, 140.2, 129.9, 129.3, 128.9, 124.3, 121.1, 43.1, 38.3; 31P-NMR (DMSO-d6, δ) (ppm): 13.9.
Table 1Results of LOI and UL 94 tests of EP, EFPx and EFPOx (x = 1, 2 and 3)
Nuclear magnetic resonance (NMR). 1H, 13C and 31P NMR spectra were recorded with a Varian Mercury AS400 spectrometer at room temperature using DMSO-d6 as solvent. The chemical shifts were reported in parts per million (d) relative to tetramethylsilane (TMS) as a reference for 1H and 13C NMR (100 MHz). The 31P NMR (162 MHz) spectra were referenced to 85% H3PO4.
Differential scanning calorimetry (DSC). DSC experiment was carried out with a DSC Q200 (TA Instruments) under nitrogen atmosphere. 4–7 mg samples were examined from 25 to 250 °C at a heating rate of 10 °C min−1.
Variable temperature Fourier transform infrared spectroscopy (VT-FTIR). VT-FTIR spectra were collected using a temperature-controlled heating device coupled with a Nicolet iS50 FTIR spectrometer from 25 to 250 °C at a heating rate of 10 °C min−1. The liquid sample was dropped fully into the space between two pressed KBr disks. Then the two disks were put inside the heating sample holder and were subjected to a small force during the test. The scanning range for FTIR spectra was 4000–500 cm−1 with a resolution of 4 cm−1. Each sample was scanned 16 times and the sampling interval was 15.58 s. In this work, VT-FTIR test for each sample was repeated three times. The results were well repeatable.
Flammability tests. LOI values of EP and flame-retardant EP were measured using an oxygen index meter (Fire Testing Technology (FTT), UK) with a precision of ±0.2%. The sheet dimensions of the samples were 130 × 6.5 × 3.2 mm3 according to ASTM D2863-97. Vertical burning tests were carried out with a UL 94 Horizontal/Vertical Flame Chamber (FTT, UK) and sheet dimensions of the samples were 130 × 13 × 3.2 mm3 according to ASTM D3801. Cone calorimeter test was carried out according to the ISO 5660-1 standard with a FTT cone calorimeter. Specimens with sheet dimensions of 100 × 100 × 3.2 mm3 were irradiated at a heat flux of 50 kW m−2. Each sample was tested at least twice.
Scanning electron microscopy (SEM). The morphology of char residues of EFP1-5 and EFPO1-5 after cone calorimeter tests was investigated by SEM (EVO MA15, Zeiss, Germany). The samples were coated with a fine gold layer under 20 kV condition.
Normal TGA test was performed with the same instrument at a heating rate of 10 °C min. 5 ± 1 mg samples were examined under air flow (90 mL min−1) from room temperature to 800 °C.
Thermogravimetry coupled with mass spectroscopy (TGA-MS): a mass spectrometer (Balzers (MS), MSC200) was connected with TGA in order to analyze the evolved gaseous during the thermal decomposition of flame retardants. 10 ± 0.2 mg samples were tested under argon and oxygen with a flow rate of 100 mL min−1. The transfer lines between the TGA and MS instruments were heated to 250 °C in order to prevent the condensation of evolved gases.
Mechanical tests. Tensile tests were conducted with a universal electromechanical testing machine (INSTRON 3384) at a rate of 0.5 mm min−1. The cross-section dimension of a dumbbell-shaped specimen was 4 mm × 4 mm and the length of testing was 35 mm. At least 5 specimens were tested for each sample in this study. Thermomechanical properties were examined with a dynamic mechanical analyzer (Q800, TA Instruments). The dimensions of samples were 35 mm × 10 mm × 2 mm. The samples were tested in a single cantilever clamp with a frequency of 1 Hz at a heating rate of 10 °C min−1 from room temperature to 240 °C.
3. Results and discussion
3.1 ReactivityDSC and VT-FTIR tests were employed to study the reactivity between N–H group in the structures of P-FRs and epoxy group in the structures of EPC. In this work, FP1 and FPO1 were chosen as one representatives pair of FPx and FPOx (x = 1, 2 and 3). The mole ratio of N–H group and epoxy group was a stoichiometric ratio (1:1) for DSC and VT-FTIR tests.
The heat phenomena of EPC and EPC/FP1 (or FPO1) occurring from RT to 220 °C were investigated by DSC (Fig. 1(A) and (B)). Meanwhile, VT-FTIR was used to record changes of IR absorbance intensity of epoxy group of EPC and mixture of EPC/FP1 (or FPO1) during the heating from RT to 220 °C. From VT-FTIR results, a relative IR absorbance intensity ratio of R (i(epoxy group)/i(ether group)) was defined in order to study the reactivity (Fig. 1(C)). The ether group on the chain of epoxy resin was chemically stable during 100 to 220 °C. The epoxy group reacted with N–H group from FP1 and FPO1 if a reaction occurred.
|Fig. 1 (A) DSC thermograms of EP, EPC/FP1, EPC/DDS and FP1. (B) DSC thermograms of EP, EPC/FPO1, EPC/DDS and FPO1. (C) R (i(epoxy group)/i(ether group)) vs. temperature curves of EP, EPC/FP1 and EPC/FPO1 from VT-FTIR tests. (D) Schematic structure diagrams of EFPx and EFPOx (x = 1, 2 and 3).|
In the heat flow curves in Fig. 1(A), EPC and FP1 did not show intensive heat phenomena during RT to 220 °C. However, there was a remarkable exothermic phenomenon at around 110 °C in the DSC curve of EPC/FP1. Compared with the curing DSC curve of EPC/DDS, the exothermic phenomenon was also obvious. These results indicated that reaction occurred between FP1 and EPC. Furthermore, the R ratio of EPC remained at a stable value during the heating, whereas that of EPC/FP1 decreased above 110 °C in VT-FTIR test (Fig. 1(C)). The decreased R ratio was induced by the decrease of epoxy groups. In view of the analysis of DSC and VT-FTIR results, N–H group in structure of FP1 was considered to be reactive with epoxy groups. From Fig. 1(B), EPC/FPO1 showed a remarkable exothermic phenomenon at around 130 °C in the DSC curve. Meanwhile, the R ratio also decreased in VT-FTIR test (Fig. 1(C)). The results showed that N–H group in structure of FPO1 was also reactive with epoxy groups. Overall, the remarkable exothermic phenomena from DSC and the decreased R ratios from VT-FTIR results revealed that N–H group in the structures of synthesized P-FRs was reactive with the epoxy group in the structures of EPC during 100 to 220 °C. FPx and FPOx (x = 1, 2 and 3) were considered to be incorporated into EP by chemical bond linking. Fig. 1(D) shows a schematic diagram of molecular chain structures for EFPx and EFPOx (x = 1, 2 and 3) after curing with DDS. Such structures were supportive for avoiding the leakage problem of flame retardants.
3.2 FlammabilityLOI and UL 94 tests were used to evaluate the impact of FPx and FPOx (x = 1, 2 and 3) on flammability of EP. The results are listed in Table 1 and the videos of UL 94 tests of EP, EFP1-5 and EFPO1-30 are in ESI (videos†).
LOI of EP was 22% and samples burnt out after 1st ignition in UL 94 tests. The results revealed that EP was high flammable. In Table 1, both the addition of FPx and FPOx (x = 1, 2 and 3) increased LOI of EP. LOI of EP was increased after the addition of both of them. However, the increase of LOI caused by the addition of FPx (x = 1, 2 and 3) into EP was more significant than that caused by the addition of FPOx (x = 1, 2 and 3). For instance, LOI of EFPO1-1 was 25% while that of EFP1-1 was 28%. EFP1-5 showed LOI of 31% which was 9% and 4% higher than LOI of EP and EFPO1-5, respectively. LOI of EFP3-7.7 was 37% which was 3% higher than that of EFPO3-7.7 which had the same P content of 0.67 wt% in the flame-retardant epoxy.
Significant different impact induced by the addition of FPx and FPOx (x = 1, 2 and 3) in EP was shown in UL 94 tests. FPOx (x = 1, 2 and 3) showed almost no impact on self-extinguishing ability of EP within 5 wt% loading, whereas FPx (x = 1, 2 and 3) endowed EP with strong self-extinguishing ability in UL 94 tests even with 1 wt% loading. EFP1-1 self-extinguished the flame after 1st and 2nd ignition although the burning time was above 10 s for each ignition. Moreover, the final burnt length of tested sample was only 50 ± 5 mm. In comparison with the total burning of EP, EFP1-1 showed great reduction of fire propagation in UL 94 tests. As P contents were fixed at 0.67 wt% in EP, all EFPOx (x = 1, 2 and 3) showed no rating, whereas all EFPx (x = 1, 2 and 3) passed V-0 rating with short burnt length of 15 ± 5 mm in UL 94 tests. Notably, it was found that until using 30 wt% of FPO1 in epoxy (EFPO1-30) it can only pass V-1 rating, which was the same level as the EFP1-3.
Above all, FPx (x = 1, 2 and 3) showed significantly different impacts on reducing flammability of EP compared with FPOx with small loadings. In the following research, FP1 and FPO1 were chosen as one representative pair of FPx and FPOx (x = 1, 2 and 3) for further detailed research about the flame-retardant mechanism.
3.3 Fire behaviourCone calorimeter test is considered to be one of the most effective bench-scale tests to study the fire behaviours of polymer materials. Fig. 2 presents the characteristic curves of heat release rate (HRR) vs. time (Fig. 2(A)) and mass vs. time (Fig. 2(B)) of EP, EFP1-y and EFPO1-y (y = 1, 3 and 5) from cone calorimeter test. The main performance parameters such as time to ignition (TTI), peak of heat release rate (pHRR), average of HRR (avg. HRR), effective heat of combustion (EHC) and the residue at 500 s are listed in Table 2.
|Fig. 2 (A) Heat release rate (HRR) curves and (B) mass curves of EP, EFP1-y and EFPO1-y (y = 1, 3 and 5) as a function of time. (C) SEM images of inner and outer surfaces of EFP1-5 and (D) SEM images of inner and outer surfaces of EFPO1-5 after cone calorimeter test.|
Table 2Data from cone calorimeter test of EP, EFP1-y and EFPO1-y (y = 1, 3 and 5) at 50 kW m−2
EP underwent intensive burning under the cone heater at 50 kW m−2. After igniting around 55 s, the burning of EP rapidly developed and terminated in a short time. HRR of EP increased rapidly to pHRR of 1079 ± 20 kW m−2 as shown in Fig. 2(A). This intensive combustion terminated within 300 s as shown in Fig. 2(A). During the burning, EHC of EP was 22.0 ± 1.5 MJ kg−1. After cone calorimeter test, the residue of EP was around 9 ± 3 wt% at 500 s, indicating that EP showed low charring ability to form a protective char residue during fire conditions.
From Fig. 2 and Table 2, the addition of FP1 and FPO1 in EP had a remarkable impact on the fire behaviour of EP. First of all, HRR curves for both EFP1-y and EFPO1-y (y = 1, 3 and 5) show two main peaks rather than one sharp peak from 50 to 500 s in Fig. 2(A). All the first peaks included shoulder peaks; the second peaks were broad and lower than 300 kW m−2. In addition, both EFP1-y and EFPO1-y (y = 1, 3 and 5) formed intumescent char residue in cone calorimeter test, as shown in Fig. S1 in ESI.† This type of char residue was beneficial for retarding heat and gaseous compound exchanges between outer and inner parts of samples in cone calorimeter tests. In Fig. 2(B), the weight losses of EFP1-y and EFPO1-y (y = 1, 3 and 5) were depressed compared with that of EP, which was consistent with the changes of HRR curves. The similar impacts of FP1 and FPO1 on fire behaviour of EP were summarized as both EFP1-y and EFPO1-y (y = 1, 3 and 5) showed shorter TTIs, lower pHRRs, lower avg. HRRs and higher yields of residue at 500 s in comparison with EP; the residual amounts of EFP1-y and EFPO1-y (y = 1, 3 and 5) were similar after the test.
Nevertheless interesting differences were specifically found between EFP1-y and EFPO1-y (y = 1, 3 and 5). TTIs of EFP1-y (y = 1, 3 and 5) were longer than those of EFPO1-y (1, 3 and 5) at the same loadings. For instance, TTI of EFPO1-5 was 50 ± 2 s, 5 s lower than that of EP, while TTI of EFP1-5 was 55 ± 3 s, keeping the same value as that of EP. In developed fire zone (52–350 s), EFP1-y and EFPO1-y (y = 1, 3 and 5) presented a shoulder peak in HRR curves in Fig. 2(A). The shoulder peak of EFPO1-y (y = 1, 3 and 5) was sharp and terminated in a short time within 20 s, while the shoulder peak of EFP1-y (y = 1, 3 and 5) was broad and terminated within 45 s. Furthermore, HRR of EFPO1-y (y = 1, 3 and 5) increased rapidly again after the shoulder peaks terminated. pHRR of EFPO1-5 was 702 ± 35 kW m−2. However, HRR of EFP1-y (y = 1, 3 and 5) increased slowly after the shoulder peaks terminated. Until 175 s, pHRR of EFP1-y (y = 1, 3 and 5) appeared with an extremely low value in comparison with that of EP and EFPO1-y (y = 1, 3 and 5). pHRR of EFP1-5 was 419 ± 47 kW m−2 which was 61% and 40% lower than those of EP and EFPO1-5, respectively. In addition, the impacts of FP1 and FPO1 on EHC of EP were different. EFPO1-y (y = 1, 3 and 5) showed EHC of 21.0 ± 0.4, 21.5 ± 0.6 and 20.5 ± 0.5 MJ kg−1, respectively. There was almost no change compared with that of EP. However, EHC of EP greatly decreased from 22.0 ± 1.5 MJ kg−1 to 16.7 ± 0.5 MJ kg−1 with 5 wt% loading of FP1. This indicated that FP1 showed flame inhibition effect in the gas phase during the burning of EFP1-y (y = 1, 3 and 5). Moreover, the mass losses of EFP1-y (y = 1, 3 and 5) were lower than those of EFPO1-y (y = 1, 3 and 5) during the test, although the residue amounts were similar at 500 s.
The shoulder peaks of EFP1-y and EFPO1-y (y = 1, 3 and 5) were considered to be caused by the char residue formed after the sample was ignited. The formed char residue was able to delay the further decomposition of epoxy. In combination with the different mass loss curves with the different pHRRs of EFP1-y and EFPO1-y (y = 1, 3 and 5), this indicated that the addition of FP1 endowed EP with a faster charring process than that of FPO1 in cone calorimeter tests.
In addition, the morphologies of inner and outer surfaces of the char residues of EFP1-5 and EFPO1-5 were characterized by SEM (Fig. 2(C) and (D)). Both the inner and outer surfaces of char residue from EFP1-5 were denser than those from EFPO1-5 after cone calorimeter test. This indicated that the char residue formed during fire behaviour of EFP1-5 was better for isolating effects than that of EFPO1-5.
3.4 Thermal decomposition behaviour and flame retardant mechanismIn Sections 3.2 and 3.3, flame retardants FPx and FPOx (x = 1, 2 and 3) presented multiple impacts on flame retardancy of EP. In common, the additions of both FPx and FPOx (x = 1, 2 and 3) increased LOI of EP, decreased pHRR and accumulated the charring process of EP in cone calorimeter tests. However, FPOx (x = 1, 2 and 3) showed almost no impact on self-extinguishing ability of EP, while FPx (x = 1, 2 and 3) endowed EP with strong self-extinguishing ability within small loadings in UL 94 tests. In addition, FPx (x = 1, 2 and 3) increased LOI and decreased pHRR of EP to a greater extent than did FPOx (x = 1, 2 and 3). All these significant differences urged us to study the flame-retardant mechanism of EFPx and EFPOx (x = 1, 2 and 3). In this part of the study, FP1 and FPO1 were also chosen as one representative pair of FPx and FPOx (x = 1, 2 and 3).
Fig. 3(A) and 4(A) present intensity variation of ion fragments with m/z of 78, 57, 41, 17, 66 and 94 detected by TGA-MS during the thermal decomposition of FP1. Fig. 3(B) presents FTIR spectra of evolved gaseous products at different temperatures detected by TGA-FTIR during the thermal decomposition of FP1. Fig. 4(B) presents absorbance intensity of IR peak at 923 cm−1vs. temperature curve.
|Fig. 3 (A) Intensity variation of ion fragments with m/z = 78, 57, 41 and 17 with temperature curves. (B) FTIR spectra at different temperatures of evolved gaseous products from thermal decomposition of FP1.|
|Fig. 4 (A) Intensity variation of ion fragments with m/z = 66 and 94. (B) Absorbance intensity variation of IR peak at 932 cm−1.|
From Fig. 3(A), allylamine, allyl and ammonium were the main evolved gaseous products from 200 to 350 °C during thermal decomposition of FP1. Ion fragments of benzene were detected with m/z = 78 above 350 °C. FTIR spectra shown in Fig. 3(B) mutually verified this information.
IR peak at 1260 cm−1 assigned to PO bond appeared on FTIR spectra since 350 °C, indicating that phosphorus-containing compounds were released during thermal decomposition of FP1. Moreover, IR peak at 932 cm−1 gradually become noticeable in FTIR spectra at 450, 500 and 600 °C. This peak was assigned to P–O–P bond. Meanwhile, ion fragments with m/z = 94 were detected at around 400 °C in TGA-MS test, as shown in Fig. 4(A). It was not assigned to phenol ion fragment due to the lack of signal of ion fragment with m/z = 66. Combining with the IR peak at 932 cm−1, the fragments with m/z = 94 were assigned to be from P2O2 compound which had been reported by Lopez.39 P2O2 was considered to be formed by radical PO˙ produced during thermal decomposition of FP1.
Overall, P–N and C–N bonds in the structure of FP1 were broken at around 200 to 400 °C. Ammonium, allylamine and allyl compounds were formed and released into the gas phase. Allyl compounds might also bond again with phenylphosphonyl structure in the condensed phase. The further decomposition of reformed structures in the condensed phase produced complex mixture of gaseous products such as allylamine, benzene and phosphorus compounds. P(O)–Ph bond in FP1 structure dissociated at nearly 350 °C which was indicated by the release of benzene structure until 600 °C. Speculative thermal decomposition of FP1 is shown in Fig. 6(A). In the condensed phase, phosphorous acid (H3PO3) was considered to be formed as shown in Fig. 6(A).40
|Fig. 5 (A) Intensity variation of ion fragments with m/z = 94, 66, 57, 41 and 17 with temperature curves of FPO1. (B) FTIR spectra at different temperatures of evolved gaseous products from thermal decomposition of FPO1.|
|Fig. 6 (A) Speculative thermal decomposition route of FP1. (B) Speculative thermal decomposition route of FPO1.|
Fig. 5(A) presents the intensity variation of ion fragments with m/z of 94, 66, 64, 57 and 41 detected by TGA-MS during thermal decomposition of FPO1. Fig. 5(B) shows FTIR spectra of evolved gaseous products at different temperatures detected by TGA-FTIR of FPO1.
The thermal decomposition of FPO1 was different from that of FP1. From 200 to 400 °C, the detected ion fragments were mainly from phenol, allylamine and allyl compounds, as shown in Fig. 5(A). Phenol ion fragments were detected until 550 °C. In Fig. 5(B), FTIR spectra at different temperatures are almost the simply added spectra of phenol and allylamine as shown in Fig. S2 in ESI.† Furthermore, few signals from phosphorus compounds were detected in the evolved gaseous products from both TGA-MS and TGA-FTIR results. The only possible phosphorus compound was considered to be metaphosphorous acid (HPO2) which was assigned to the ion fragment with m/z = 64 shown in Fig. 5(A). The above results indicated that phosphorus was mainly in the condensed phase during thermal decomposition of FPO1. Speculative thermal decomposition path of FPO1 is shown in Fig. 6(B). Phosphoric acid (H3PO4) was considered to be formed in the condensed phase of FPO1 during thermal decomposition.40 In the case of FPO1, there was no significant signal indicating the formation of phosphorus radical at least during the main thermal decomposition of FPO1.
|Fig. 7 (A) TGA curves of EP, EFP1-5 and EFPO1-5. (B) Absorbance intensity of hydrocarbons vs. temperature curves of EFP1-5 and EFPO1-5 from TGA-FTIR tests.|
Compared with the TGA curve of EP, the addition of both FP1 and FPO1 reduced Td5% and Tmax of EP to lower temperatures. Meantime, the char residue of EP was also increased at 600 °C, indicating that both FP1 and FPO1 induced the formation of char residues during thermal decomposition of EP. As seen from Fig. S3 in ESI,† FTIR spectra at Tmax of EP, EFP1-5 and EFPO1-5 were very similar to each other, indicating that the addition of FP1 and FPO1 did not change the main decomposition route of EP.
In detail, specific differences were found between TGA results of EFP1-5 and EFPO1-5. Td5% and Tmax of EFP1-5 were 31 and 5 °C higher, respectively, than those of EFPO1-5. The char residue of EFP1-5 was 27.8%, while that of EFPO1-5 was 33.4% at 650 °C. H3PO3 and H3PO4 formed during thermal decomposition of FP1 and FPO1 were considered to be effective for the formation of char residue during thermal decomposition of EFP1-5 and EFPO1-5, respectively. The different acidities of H3PO3 and H3PO4 were able to induce the formation of different char residues. Hydrocarbons from thermal decomposition of polymer materials were regarded as the main “fuel” to be ignited under heating or fire exposure.8Fig. 7(B) shows that hydrocarbons produced from thermal decomposition of EFPO1-5 started to be detected at 350 °C, while those produced from thermal decomposition of EFP1-5 started to be detected at 400 °C.
In condensed phase, FP1 and FPO1 acted on burning of EP in the same way. The typical intumescent components formed in both EFP1-5 and EFPO1-5 under heating or igniting: the acid compounds (H3PO3 and H3PO4) from thermal decomposition of FP1 and FPO1 acted as acid sources; the EP matrix acted as char former; the gaseous products from thermal decomposition of EFP1-5 and EFPO1-5 acted as blower. Foam-like char residues were observed both in EFP1-5 and EFPO1-5 during cone calorimeter tests. The char residue acted as a physical barrier that slowed down heat and gaseous product transfer between inner and outer parts of the matrix.
However, the release temperature of acid compounds and evolved gaseous products and the charring rate of epoxy matrix affected the formation of foam-like char residues. For instance, char residue of EFP1-5 was denser than that of EFPO1-5, even though their mass and the intumescent shape were similar to each other. As seen from Fig. 2(B), EFP1-y (y = 1, 3 and 5) showed less of a mass loss than EFPO1-y (y = 1, 3 and 5) did before 200 s in cone calorimeter tests. There was a similar occurrence before 400 °C in TGA test, shown in Fig. 7(A). This indicated that the charring rate of EFPO1-y (y = 1, 3 and 5) was lower than that of EFP1-y (y = 1, 3 and 5) at the initial stage of heating.
In the gas phase, FPO1 showed no significant impact, while FP1 showed efficient flame inhibition effect on burning of EP. As to FPO1, firstly, the dilution effect was not evident due to the small loading of FPO1 and the minor impact of FPO1 on thermal decomposition of EP. Secondly, there was no active moiety and radicals that were able to capture H˙ and OH˙ radicals to interrupt the combustion chain during thermal decomposition of FPO1. The gas phase action of FPO1 is speculatively described in Fig. 8(A).
|Fig. 8 Proposed gas phase effect and burning scheme of EFPO1-5 (A) and EFP1-5 (B) in UL 94 test. (C) TTCvs. time curves of EFP1-5 and EFPO1-5 in UL 94 test. (D) TIRvs. time curves of EFP1-5 and EFPO1-5 in UL 94 test.|
As to FP1, hydrocarbons were released around 400 °C for EFP1-5, indicating that H˙ and OH˙ were continuously generated since then. The radical reactions involved with these two kinds of radicals greatly impact the burning velocity of EFP1-5. However, PO˙ radicals was produced during thermal decomposition of FP1 since around 400 °C. The PO˙ radical was able to annihilate H˙ and OH˙ as shown in Fig. 8(B). The concentrations of H˙ and OH˙ decreased owing to the annihilation effect, meaning that the radical flame reaction cycle was depressed.
In recent years, researchers have made a lot of effort to study the specific fire behaviour in different tests, such as LOI, UL 94 and cone calorimeter tests.41,42 In this work, the near-to-surface temperature (NTST) and surface temperature (ST) of UL 94 test samples of EFP1-5 and EFPO1-5 were determined by thermocouple thermometer and infrared thermometer (IR thermometer), respectively, as shown in Fig. 8(C) and (D). During the experiment, the emissivity was always set as 0.950 for IR thermometer. The TTC (NTST) vs. time and TIR (ST) vs. time curves of EFP1-5 and EFPO1-5 are shown in Fig. 8(C) and (D).
In Fig. 8(C), TTC of EFP1-5 and EFPO1-5 increased continuously no matter the 1st or 2nd 10 s ignition in UL 94 test. In the 1st 10 s, TTC of EFPO1-5 was 215 ± 15 °C and that of EFP1-5 was 211 ± 14 °C. Meantime, the slopes of TTC curves were very close to each other. After the 1st 10 s ignition, TTC of EFPO1-5 continued increasing until 35 s when the flame was out, while TTC of EFP1-5 decreased within 5 s after the first 10 s. During the 2nd 10 s ignition, the rising rates of TTC curves were also similar for EFPO1-5 and EFP1-5. After the 2nd 10 s ignition, TTC of EFPO1-5 increased to around 500 °C and kept relatively stable until the burning finished. However, TTC of EFP1-5 decreased within 3 s after the 2nd 10 s ignition. The trends of TTC curves were consistent with the burning behaviour: EFPO1-5 burnt more than 10 s, while EFP1-5 burnt within 5 s no matter after the 1st or 2nd ignition.
In Fig. 8(D), interesting differences of TIR between EFP1-5 and EFPO1-5 are shown. Both EFP1-5 and EFPO1-5 presented up-down trends during 1st and 2nd ignition. However, the increase of TIR of EFP1-5 slowed down when TIR was 300 ± 40 °C, whereas that of EFPO1-5 slowed down when TIR was 400 ± 40 °C during the 1st 10 s ignition. The maximum difference between TIRs of EFP1-5 and EFPO1-5 was around 80 °C during the 1st ignition. During the 2nd 10 s ignition, TIR of EFPO1-5 and EFP1-5 increased in the same way at first. However, TIR of EFP1-5 did not increase obviously when the value was around 350 ± 40 °C. TIR of EFPO1-5 increased to 550 ± 40 °C until the 2nd 10 s ignition. The maximum difference was up to 300 °C during the 2nd ignition. In the following time, TIR of EFPO1-5 increased to around 600 °C until the burning finished, while that of EFP1-5 decreased immediately when the flame was out.
Basically, the heat energy to increase TIR (QST) was described by the following equation: QST = QF + QB − QD − QC. QF represents the heat from UL 94 test flame; QB represents heat from burning of sample; QD represents the decomposition heat from samples; QC represents the conductive heat of samples. Herein, we assumed that QD was the same for the two systems.
From the TTCvs. time curves, it was noted that TTC of EFP1-5 and EFPO1-5 showed almost the same rising rate during 1st and 2nd ignitions. This indicated that QC was considered the same for the two systems as well. Under such a hypothesis, QST only depended on QB, because QF had been also controlled in the same way. According to Fig. 8(D), obviously, QB of EFPO1-5 was much higher than that of EFP1-5. This meant that the burning of EFP1-5 was much weaker than that of EFPO1-5, which was attributed to the flame inhibition effect of PO˙ in the gas phase.
As a whole, the following two factors induced the highly efficient flame retardancy of FP1 on EP: (i) flame inhibition effect in the gas phase; (ii) catalyzing the charring process in the condensed phase. We believed that the flame inhibition effect was the main factor endowing EP with strong self-extinguishing ability. Due to a combination of flame inhibition and catalyzing char formation effects, FP1 induced V-0 rating in UL 94 tests, high LOI value and low HRR for EP at very low loading. It also explains the fire behaviours of other FPx (x = 1, 2 and 3) containing flame-retardant epoxy.
3.5 Mechanical propertiesStrain–stress curves, tensile strength and Young's modulus charts of EP, EFP1-5 and EFPO1-5 are shown in Fig. 9(A) and (B). Storage modulus curves and loss factor (tanδ) curves are displayed in Fig. 9(C) and (D) as obtained using DMA. The addition of FP1 and FPO1 did not have a significant impact on tensile behaviour of EP as shown in Fig. 9(A). The tensile strengths of EFP1-5 and EFPO1-5 were lower within 5% compared with that of EP (Fig. 9(B)). This meant that the impact of FP1 and FPO1 on tensile strength of EP was negligible. Young's modulus of EP was 1911 ± 65 MPa. It was increased to 2103 ± 60 and 2107 ± 75 MPa after the addition of FP1 and FPO1 respectively, indicating that the addition of FP1 and FPO1 slightly increased the rigidity of EP.
|Fig. 9 (A) Strain–stress curves. (B) Charts of tensile strength and Young's modulus of EP, EFP1-5 and EFPO1-5. (C) Storage modulus and (D) tanδ curves of EP, EFP1-5 and EFPO1-5 from DMA tests.|
The results of DMA in Fig. 9(C) and (D) show that the impacts of FP1 and FPO1 on the viscoelastic property of EP were also negligible. Either in glass state or rubbery state, storage modulus of EFP1-5 and EFPO1-5 showed the same magnitude as that of EP. In addition, the peak of tand was used to evaluate the glass transition temperature (Tg) in this work. Tg of EP was 187 ± 3 °C, whereas those of EFP1-5 and EFPO1-5 were also 187 ± 4 °C respectively. Fig. 9(D) shows the selected curves of tand. Overall, the mechanical property of EP was not impacted greatly with 5 wt% loading of P-FRs.
4. ConclusionsIn this work, two series of phenylphosphonate-based (PO–Ph) and phenylphosphoric-based (PO–OPh) flame retardants (FPx and FPOx, x = 1, 2 and 3) were synthesized successfully with high yields. The remarkable exothermic phenomena and the decreased R (i(epoxy group)/i(ether group)) ratios of EPC/FP1 and EPC/FPO1 revealed the N–H group in the structures of synthesized P-FRs was reactive with epoxy group in the structures of EPC during 100 to 220 °C. It indicated that synthesized P-FRs was chemically bonded with EP during the curing procedure. Interesting findings between FPx and FPOx (x = 1, 2 and 3) on improving flame retardancy of EP (DGEBA-type epoxy resin/DDS system) were shown in different fire tests. FPOx (x = 1, 2 and 3) showed almost no effect on self-extinguishing ability of EP, while FPx (x = 1, 2 and 3) endowed EP with strong self-extinguishing ability with small loadings in UL 94 tests. The addition of FPO1 made EP pass V-1 with a high loading of 30 wt%, while EFP1-3 already passed V-1 rating and EFP1-5 passed V-0 rating only with 5 wt% loading (P content lower than 0.7 wt%). In addition, FPx (x = 1, 2 and 3) increased LOI and decreased pHRR of EP to a greater extent than FPOx (x = 1, 2 and 3) did in cone calorimeter tests. Comparing flame-retardant mechanisms, both FP1 and FPO1 acted via inducing char formation in the condensed phase of EP. Intumescent char residues formed owing to the following conditions in cone calorimeter test of EFP1-5 and EFPO1-5: the acid compounds (H3PO3 and H3PO4) from thermal decomposition of FP1 and FPO1 acted as acid sources; the EP matrix acted as char former; the gaseous products from thermal decomposition of EFP1 and EFPO1 acted as blower. However, FP1 showed flame inhibition in the gas phase caused by the PO˙ radical produced during thermal decomposition of FP1, whereas FPO1 did not show any effect in the gas phase. Furthermore, the results of online temperature detection experiment for NTST and ST in UL 94 tests showed that flame inhibition effect was the main factor to endow EP with strong self-extinguishing ability in UL 94 tests. Moreover, the small loading addition of FP1 and FPO1 had negligible impact on tensile strength, storage modulus and Tg of EP.
AcknowledgementsOne of the authors (Xiaomin Zhao) would thank the financial support from China Scholarship Council. In addition, this work is partly funded by Ramón y Cajal grant (RYC-2012-10737) and the European Commission under the 7th Framework Program (Marie Curie Career Integration Grant, GA-321951).
- C. May, Epoxy resins: chemistry and technology, CRC press, 1987 Search PubMed.
- H. Q. Pham and M. J. Marks, Kirk-Othmer Encyclopedia of Chemical Technology, 2004 Search PubMed.
- X. Zhang, Q. He, H. Gu, H. A. Colorado, S. Wei and Z. Guo, ACS Appl. Mater. Interfaces, 2013, 5, 898–910 Search PubMed.
- H. Gu, J. Guo, Q. He, S. Tadakamalla, X. Zhang, X. Yan, Y. Huang, H. A. Colorado, S. Wei and Z. Guo, Ind. Eng. Chem. Res., 2013, 52, 7718–7728 Search PubMed.
- L.-J. Qian, L.-J. Ye, G.-Z. Xu, J. Liu and J.-Q. Guo, Polym. Degrad. Stab., 2011, 96, 1118–1124 Search PubMed.
- T. Mariappan, Y. Zhou, J. Hao and C. A. Wilkie, Eur. Polym. J., 2013, 49, 3171–3180 Search PubMed.
- B. Perret, B. Schartel, K. Stöβ, M. Ciesielski, J. Diederichs, M. Döring, J. Krämer and V. Altstädt, Eur. Polym. J., 2011, 47, 1081–1089 Search PubMed.
- X. Wang, Y.-T. Pan, J.-T. Wan and D.-Y. Wang, RSC Adv., 2014, 4, 46164–46169 Search PubMed.
- M. Rakotomalala, S. Wagner and M. Doering, Materials, 2010, 3, 4300–4327 Search PubMed.
- K. Wu, L. Song, Y. Hu, H. Lu, B. K. Kandola and E. Kandare, Prog. Org. Coat., 2009, 65, 490–497 Search PubMed.
- M. Gao, W. Wu and Y. Yan, J. Therm. Anal. Calorim., 2009, 95
|Sample||TTIa (s)||pHRRb (kW m−2)||Avg. HRRc (kW m−2)||EHC (MJ kg−1)||Residue (500 s, wt%)|
|EP||55 ± 2||1079 ± 50||220 ± 10||22.0 ± 1.5||9 ± 3|
|EFP1-1||54 ± 3||600 ± 45||200 ± 9||20.2 ± 0.6||20 ± 2|
|EFP1-3||55 ± 3||530 ± 53||190 ± 12||18.5 ± 0.8||23 ± 3|
|EFP1-5||55 ± 3||419 ± 47||171 ± 15||16.7 ± 0.5||24 ± 2|
|EFPO1-1||54 ± 2||864 ± 57||208 ± 6||21.0 ± 0.4||22 ± 3|
|EFPO1-3||52 ± 2||720 ± 52||189 ± 11||21.5 ± 0.6||23 ± 3|
|EFPO1-5||50 ± 2||702 ± 35||173 ± 13||20.5 ± 0.5||25 ± 3|