Cross-Linking Modification of Ammonium Polyphosphate via Ionic Exchange and Self-Assembly for Enhancing the Fire Safety Properties of Polypropylene

Modified ammonium polyphosphate (MAPP) was prepared as a novel mono-component intumescent flame retardant (IFR) via the ionic exchange between ammonium polyphosphate (APP) and piperazine sulfonate, which is synthesized by self-assembly using 1-(2-aminoethyl) piperazine (AEP) and p-aminobenzene sulfonic acid (ASC) as raw materials. This all-in-one IFR integrating three functional elements (carbon, acid, and gas source) showed more efficient flame retardancy and excellent smoke suppression as well as better mechanical properties than the conventional APP. The incorporation of 22.5 wt.% MAPP into polypropylene (PP) eliminated the melt dripping phenomenon and passed the UL-94 V-0 rating. The results of the cone calorimetry test (CCT) revealed that the release of heat, smoke, and CO is significantly decreased, demonstrating that this novel IFR endows PP with excellent fire safety more effectively. For PP/MAPP composites, a possible IFR mechanism was proposed based on the analysis of the pyrolysis gas and char residues.

illustrated the modification process of ammonium polyphosphate involving two stages. Firstly, the piperazine sulfonate (intermediate) was prepared via self-assembly as shown in stage I. Under stirring, ASC (17.3 g, 0.1 mol) was dissolved in deionized water at 90 • C. Hereafter, AEP (13.1 mL, 0.1 mol) was dropped into the ASC solution at a constant titration rate (2 s / drop). The reaction system was kept 95 • C for 3 h. After vacuum distillation and suction filtration, the straw yellow powder (AEP-ASC) was obtained and then dried at 80 • C in the vacuum oven for 24 h.
Secondly, the modified APP (MAPP) was synthesized via the ionic exchange between intermediate and APP as shown in stage II. To a three-neck round-bottom flask equipped with a stirrer, a thermometer and a reflux condenser were added APP (10 g), 50 mL ethanol, and 20 mL water. The mixture was refluxed under stirring for 12 h. Afterward, the intermediate (10 g) was introduced into the mixture, and the reaction continued for 2 h. Then, the light yellow and viscous solid was obtained from vacuum filtration and washed by deionized water (removing unreacted intermediate) three times. After drying under a high vacuum at 80 • C overnight, the final product (light yellow powder) was obtained and defined as modified APP (MAPP).
Polymers 2020, 12, x 3 of 17 Figure 1 illustrated the modification process of ammonium polyphosphate involving two stages. Firstly, the piperazine sulfonate (intermediate) was prepared via self-assembly as shown in stage I. Under stirring, ASC (17.3 g, 0.1 mol) was dissolved in deionized water at 90 °C. Hereafter, AEP (13.1 mL, 0.1 mol) was dropped into the ASC solution at a constant titration rate (2 s / drop). The reaction system was kept 95°C for 3 h. After vacuum distillation and suction filtration, the straw yellow powder (AEP-ASC) was obtained and then dried at 80 °C in the vacuum oven for 24 h.

Modification of APP
Secondly, the modified APP (MAPP) was synthesized via the ionic exchange between intermediate and APP as shown in stage II. To a three-neck round-bottom flask equipped with a stirrer, a thermometer and a reflux condenser were added APP (10 g), 50 mL ethanol, and 20 mL water. The mixture was refluxed under stirring for 12 h. Afterward, the intermediate (10 g) was introduced into the mixture, and the reaction continued for 2 h. Then, the light yellow and viscous solid was obtained from vacuum filtration and washed by deionized water (removing unreacted intermediate) three times. After drying under a high vacuum at 80 °C overnight, the final product (light yellow powder) was obtained and defined as modified APP (MAPP).

Sample Preparation
The IFR/PP blends were obtained by mixing the APP or MAPP with different mass ratios. The specific formulas were shown in Table 1. Then the IFR/PP blends were mixed at 190 °C for 5 min in an internal mixer (US-70C, Changzhou Suyan Technology Co., Ltd., Changzhou, China). The samples for fire behavior characterization were pressed to the sheet at 195 °C for 5 min under 10 MPa of the pressure under a plate vulcanizer (ZHY-W, Chengde Testing Machine Factory, Hebei, China). The samples for mechanical measurements were injection-molded in a miniature injection molding instrument (WZS-10D, Xinshuo Precision Machinery Co., Ltd., Shanghai, China). The melting temperature was 200 °C, the mold temperature was 40 °C, the holding pressure was 0.6 MPa, and the holding time was 10 s.

Sample Preparation
The IFR/PP blends were obtained by mixing the APP or MAPP with different mass ratios. The specific formulas were shown in Table 1. Then the IFR/PP blends were mixed at 190 • C for 5 min in an internal mixer (US-70C, Changzhou Suyan Technology Co., Ltd., Changzhou, China). The samples for fire behavior characterization were pressed to the sheet at 195 • C for 5 min under 10 MPa of the pressure under a plate vulcanizer (ZHY-W, Chengde Testing Machine Factory, Hebei, China). The samples for mechanical measurements were injection-molded in a miniature injection molding instrument (WZS-10D, Xinshuo Precision Machinery Co., Ltd., Shanghai, China). The melting temperature was 200 • C, the mold temperature was 40 • C, the holding pressure was 0.6 MPa, and the holding time was 10 s.

Characterization
The Fourier transform infrared (FTIR) spectra were scanned by a Perkin Elmer instrument (Waltham, MA, USA) at room temperature. The samples were mixed with KBr pellets including 0.5 mg sample and 50 mg KBr and scanned 32 times over a spectral range of 4000-450 cm −1 with a resolution of 4 cm −1 .
X-ray diffraction (XRD) test was carried out by a power D/MAX2500 diffraction (Rigaku Corporation, Tokyo, Japan) using Cu Ka radiation under a scanning rate of 3 • /s from 5 • to 60 • (2θ).
The microstructures of flame retardants, char residues, and sections were measured using an SEM instrument (Zeiss SUPRA55, Jena, Germany). The specimens were sputter-coated with a conductive gold layer before observation.
The UL-94 flammability classification was measured by a CZF-5 instrument (Shine Ray Instrument Co. Ltd., Nanjing, China) according to ASTM D3801. The dimensions of the samples are 130 mm × 13 mm × 3.2 mm.
Limiting oxygen index (LOI) was measured by an LOI analyzer (JF-3, Jiang Ning Co. Ltd., Nanjing, China). The dimensions of the samples are 130 mm × 6.5 mm × 3.2 mm according to GB/T 2406-93 standard.
Thermogravimetric analysis (TGA) was tested by a Perkin-Elmer TGA 4000 with a heating rate of 10 • C/min under nitrogen and oxygen atmosphere at temperatures ranging from 30 • C to 850 • C. TG-FTIR analysis was carried out under a nitrogen atmosphere at a heating rate of 10 • C/min from 30 • C to 850 • C. The flame retardant samples used in the tests were all fine solid powders with a mass of 8 ± 0.5 mg; the samples of pure PP and its blends used were block solids with a mass of 8 ± 0.5 mg.
The flammability of pure PP and PP blends were tested by a cone calorimeter device (Fire Testing Technology, East Grinstead, UK) according to ISO 5660-1. The dimension of the square sample was 100 mm × 100 mm × 3 mm, and the irradiation power was 35 kW/m 2 .
Raman spectroscopy measurement was carried out with a DXR laser Raman spectrometer (Thermo Scientific,) using a 532 nm helium-neon laser line at room temperature.  Figure 2a illustrates the FTIR spectra of AEP, ASC, and AEP-ASC. The characteristic peaks at 2939 and 2813 cm −1 are attributed to the symmetrical and antisymmetric -CH 2 -stretching vibration of AEP, and the peaks at 1315 and 1120 cm −1 represent the -S=O stretching vibration of ASC [23]. Noticeably, the peak at 2647 cm −1 (S-OH) [24] in the spectrum of ASC disappears and two new peaks at 2446 and 3265 cm −1 (NH 2 + and NH 3 + bending vibration) are observed in the spectra of AEP-ASC, which indicates the formation of -NH 2 + -O-and -NH 3 + -O- [1,25]. Furthermore, the spectrum of AEP-ASC shows sharp peaks at 3469, 3347 and 3232 cm −1 owing to amino groups [11]. These results demonstrate that intermediate (AEP-ASC) is obtained by self-assembly. The chemical structures of the APP and MAPP are also explored by FTIR spectra (Figure 2b). Both APP and MAPP display the typical peaks at 1263 cm −1 (P=O), 1087 cm −1 (P-O symmetric stretching), and 882 cm −1 (P-O asymmetric vibration) [26]. As compared with the spectra of APP, however, that of the MAPP displays some new peaks which are ascribed to the absorption of the benzene ring (1600 and 689 cm −1 , etc.), the stretching vibration of O=S=O (1320 and 1180 cm −1 ) and especially the NH 2 + and NH 3 + bending vibration (2481 and 3248 cm −1 ). The results initially demonstrate that the APP was successfully cross-link modified with AEP-ASC by ionic exchange.

FTIR Analysis
Polymers 2020, 12, x 5 of 17 of AEP, and the peaks at 1315 and 1120 cm −1 represent the -S=O stretching vibration of ASC [23]. Noticeably, the peak at 2647 cm −1 (S-OH) [24] in the spectrum of ASC disappears and two new peaks at 2446 and 3265 cm −1 (NH2 + and NH3 + bending vibration) are observed in the spectra of AEP-ASC, which indicates the formation of -NH2 +-O-and -NH3 +-O- [1,25]. Furthermore, the spectrum of AEP-ASC shows sharp peaks at 3469, 3347 and 3232 cm −1 owing to amino groups [11]. These results demonstrate that intermediate (AEP-ASC) is obtained by self-assembly. The chemical structures of the APP and MAPP are also explored by FTIR spectra (Figure 2b). Both APP and MAPP display the typical peaks at 1263 cm −1 (P=O), 1087 cm −1 (P-O symmetric stretching), and 882 cm −1 (P-O asymmetric vibration) [26]. As compared with the spectra of APP, however, that of the MAPP displays some new peaks which are ascribed to the absorption of the benzene ring (1600 and 689 cm −1 , etc.), the stretching vibration of O=S=O (1320 and 1180 cm −1 ) and especially the NH2 + and NH3 + bending vibration (2481 and 3248 cm −1 ). The results initially demonstrate that the APP was successfully cross-link modified with AEP-ASC by ionic exchange.

XRD Analysis
XRD patterns of APP and MAPP are presented in Figure 2c. Obviously, the position of the diffraction peaks of APP and MAPP are almost the same, indicating that the crystalline structure of APP is not influenced by cross-linking modification.

XPS Analysis
The XPS spectra of APP and MAPP are shown in Figure 2d. As can be seen, MAPP shows much higher carbon content (44%), lower contents of oxygen (32%), and phosphorus (8%) than APP. Furthermore, a new peak appears at 168.1 eV for MAPP, which is ascribed to the S2p from the intermediate (AEP-ASC). The fitted N1S spectra of APP and MAPP are presented in Figure 2e,f. For APP, the binding energy at around 401.2 eV is assigned to NH4 + , and the peak at 399.3 eV might

XRD Analysis
XRD patterns of APP and MAPP are presented in Figure 2c. Obviously, the position of the diffraction peaks of APP and MAPP are almost the same, indicating that the crystalline structure of APP is not influenced by cross-linking modification.

XPS Analysis
The XPS spectra of APP and MAPP are shown in Figure 2d. As can be seen, MAPP shows much higher carbon content (44%), lower contents of oxygen (32%), and phosphorus (8%) than APP. Furthermore, a new peak appears at 168.1 eV for MAPP, which is ascribed to the S 2p from the intermediate (AEP-ASC). The fitted N 1S spectra of APP and MAPP are presented in Figure 2e,f. For APP, the binding energy at around 401.2 eV is assigned to NH 4 + , and the peak at 399.3 eV might correspond to the N (-P-NH-P-). This result of APP spectra agrees with Wang's report [17]. Although the NH 4 + peak still appears for MAPP, three new peaks at 399.7 eV, 400.3 eV, and 400.7 eV are observed [16,17]. It indicates that NH 2 + and NH 3 + are formed and take place of NH 4 + partially. This result further confirms the successful ionic interactions between intermediate and APP.

Surface Morphology and EDS test
SEM is used to investigate the micro-morphology of the APP and MAPP, and the corresponding images are shown in Figure 3. Apparent changes in the surface morphology between APP and MAPP are easily observed. The pure APP particle shows a smooth surface, and its particle size is about 10 µm (Figure 3a,b). However, the MAPP particles look like a cross-linked aggregate of APP particles and display a rough surface (Figure 3c,d). This could be interpreted as due to H-bonding and ion-dipole forces that contribute to particle aggregation and fusion [16]. The changes in the surface morphology of MAPP are consistent with the proposed model as shown in Figure 1. Moreover, the elemental composition of the APP and MAPP is evaluated from the EDS test ( Figure 3e). By comparison with APP, the MAPP presents higher content of C (24.27%) and lower content of P (11.53%) as well as the appearance of S element. Both the changes in the surface morphology and element composition between APP and MAPP demonstrate the successful cross-linking modification of APP evidently.

Surface Morphology and EDS test
SEM is used to investigate the micro-morphology of the APP and MAPP, and the corresponding images are shown in Figure 3. Apparent changes in the surface morphology between APP and MAPP are easily observed. The pure APP particle shows a smooth surface, and its particle size is about 10 μm (Figure 3a,b). However, the MAPP particles look like a cross-linked aggregate of APP particles and display a rough surface (Figure 3c,d). This could be interpreted as due to H-bonding and iondipole forces that contribute to particle aggregation and fusion [16]. The changes in the surface morphology of MAPP are consistent with the proposed model as shown in Figure 1. Moreover, the elemental composition of the APP and MAPP is evaluated from the EDS test ( Figure 3e). By comparison with APP, the MAPP presents higher content of C (24.27%) and lower content of P (11.53%) as well as the appearance of S element. Both the changes in the surface morphology and element composition between APP and MAPP demonstrate the successful cross-linking modification of APP evidently.

Reaction to Small Flame (UL 94 Vertical Burning and LOI)
The flame retardant data of the neat PP and IFR-PP blends are summarized in Table 1. The neat PP shows serious dripping with no rating level during the UL-94 test, and its value of LOI is merely 17%, indicating its extreme flammability. In the case of the incorporation of APP alone, the PP/APP blend fails to pass the UL-94 V-0 test and gives of LOI value of 19% even at its loading level up to 25 wt.%, revealing that APP is not very practical to improve the flame retardancy of PP. However, the LOI values and flame retardant rating of PP/MAPP blends are increased significantly with the incorporation of MAPP. For example, the IFR-PP blend containing 22.5 wt.% MAPP gives 30% of LOI value and UL-94 V-0 rating. It demonstrates that the flame retardant efficiency of MAPP is much higher than that of APP for IFR-PP blends.

Reaction to Small Flame (UL 94 Vertical Burning and LOI)
The flame retardant data of the neat PP and IFR-PP blends are summarized in Table 1. The neat PP shows serious dripping with no rating level during the UL-94 test, and its value of LOI is merely 17%, indicating its extreme flammability. In the case of the incorporation of APP alone, the PP/APP blend fails to pass the UL-94 V-0 test and gives of LOI value of 19% even at its loading level up to 25 wt.%, revealing that APP is not very practical to improve the flame retardancy of PP. However, the LOI values and flame retardant rating of PP/MAPP blends are increased significantly with the incorporation of MAPP. For example, the IFR-PP blend containing 22.5 wt.% MAPP gives 30% of LOI value and UL-94 V-0 rating. It demonstrates that the flame retardant efficiency of MAPP is much higher than that of APP for IFR-PP blends.

Combustion Behavior Under Forced-Flaming Scenario (Cone Calorimetry Test)
The cone calorimetry test (CCT) is currently the ideal test to assess the burning behavior of polymeric materials under an ongoing fire or forced combustion conditions. The characteristic curves including the heat release rate (HRR), total heat release (THR), smoke production rate (SPR), total smoke production (TSP), CO, and CO 2 production (COP, CO 2 P) are illustrated in Figure 4, and some important parameters such as time to ignition (TTI), the peak of heat release rate (PHRR), THR, and average mass loss rates (AMLR) are listed in Table 2.
HRR is generally the most critical performance parameter to characterize the fire intensity, and PHRR indicates the maximum degree of heat release during combustion. As presented in Table 2, the pure PP gives 718.3 kW/m 2 of PHRR value and 57.3 MJ/m 2 of THR, which are reduced to 354.7 kW/m 2 and 47.9 MJ/m 2 with the incorporation of 25 wt.% APP, respectively. The corresponding values are further reduced significantly with the substitution of APP with MAPP. For example, the IFR-PP containing 25 wt.% MAPP presents the lowest PHRR (155.9 kW/m 2 ) and THR values (44.9 MJ/m 2 ), which are reduced by 78.3% and 21.6%, respectively, as compared to those of the neat PP. The sharp decline of the PHRR and THR values indicates that MAPP has a more significant role in fire safety than APP.
Fire growth index (FGI) and fire performance index (FPI) are used to assess the fire hazard of polymers, which can be calculated according to the following equation [27] based on the HRR curves and TTI values: The obtained FGI and FPI values of the neat PP and IFR-PP blends are also listed in Table 2. Generally, the lower FGI and higher FPI values imply higher fire safety properties [28]. As can be seen, the IFR-PP blends display much lower FGI values whilst greater FPI values than that of the pure PP. Noticeably, the FPI value (0.22 m 2 s/kW) of PP/25%MAPP is increased by 57.1% whilst the FGI value (0.72 kW/m 2 s) is decreased by 61.5% as compared with these corresponding values of PP/25%APP (0.14 m 2 s/kW of FPI and 1.87 kW/m 2 s of FGI, respectively). Furthermore, taking the much lower average mass loss rate (AMLR, 0.027 g/s) of PP/25%MAPP than that (0.043 g/s) of PP/25%APP into consideration, it can be concluded that MAPP is superior to APP in fire safety.
It is well known that most victims suffocate to death by inhalation of smoke and CO in fire accidents. As such, smoke suppression is very significant for flame retardant materials. As can be seen in Table 2, the total smoke production (TSP) is reduced from 48.8 m 2 for pure PP to 6.4 m 2 for PP/APP while their peaks of the smoke production rate (PSPR) are almost the same. Fascinatingly, both the PSPR and TSP values of PP/25%MAPP are reduced drastically as compared with that of the pure PP or PP/25%MAPP, indicating the superior smoke suppression of MAPP. Moreover, the CO emissions during CCT are ascribed to the insufficient combustion of decomposed volatiles. It is apparent that CO and CO 2 relaease of IFR/PP composites are weaker than those of pure PP from Figure 4e,f, which was contributed to the formed char at the early stage and barrier performance of the cross-linking structure of MAPP. It is worth noting that the mean CO yeild (0.15 kg/kg) of PP/25%MAPP specimen is half of that of pure PP (0.32 kg/kg) due to the excellent intumescent char residue on the sample surface, implying the possibility of suffocation can be significantly inhibited during evacuation.

Thermal Stability Analysis
Figures 5 and 6 present the TG and DTG curves of the flame retardants and IFR-PP samples under nitrogen and air atmosphere, respectively. Corresponding data, such as the initial thermal decomposition temperature (T5wt%), the temperature at maximal degradation (Tmax), Rate of Tmax, and char residue at 800 °C are listed in Tables 3 and 4. It can be observed that the MAPP shows high T5wt% values, no matter under the nitrogen (303.1 °C) or air (270.9 °C) atmosphere, which fully meets the melt blending temperature of PP (190 °C). While APP presents two primary mass loss processes, ascribing to thermal removal of H2O and NH3, and crosslinking reactions between phosphoruscontaining acid. It is noticeable that thermal decomposition process of MAPP is little different from APP, due to the introduction of AEP-ASC. The deamination, dehydration and decomposition of ASC at around 225-380 °C make the first DTG peak become wider. The second stage after 500 °C corresponds to the pyrolysis of APP chains as well. Furthermore, the pure PP shows a single degradation process with the T5wt% and Tmax at 373.5 °C and 455.4 °C under nitrogen atmosphere. Moreover, there is no char residue left at 800 °C no matter whether under nitrogen or air atmosphere, indicating pure PP fails to form char alone. On the contrary, two-stage degradation is observed for all IFR-PP blends. The first-stage at low temperature is mainly ascribed to the degradation of the IFR and PP, and the second-stage at high temperature is attributed to the decomposition products (polyphosphoric acids, pyrophosphate, and metaphosphate acid) of APP or MAPP [1]. Furthermore, all samples display a lower degradation temperature under the air atmosphere than that under the nitrogen atmosphere owing to thermo-oxidative degradation [24]. As compared with the pure PP, all PP/APP samples display greater T5wt% values, which is ascribed to the protection of the thick

Thermal Stability Analysis
Figures 5 and 6 present the TG and DTG curves of the flame retardants and IFR-PP samples under nitrogen and air atmosphere, respectively. Corresponding data, such as the initial thermal decomposition temperature (T 5wt% ), the temperature at maximal degradation (T max ), Rate of T max , and char residue at 800 • C are listed in Tables 3 and 4. It can be observed that the MAPP shows high T 5wt% values, no matter under the nitrogen (303.1 • C) or air (270.9 • C) atmosphere, which fully meets the melt blending temperature of PP (190 • C). While APP presents two primary mass loss processes, ascribing to thermal removal of H 2 O and NH 3 , and crosslinking reactions between phosphorus-containing acid. It is noticeable that thermal decomposition process of MAPP is little different from APP, due to the introduction of AEP-ASC. The deamination, dehydration and decomposition of ASC at around 225-380 • C make the first DTG peak become wider. The second stage after 500 • C corresponds to the pyrolysis of APP chains as well. Furthermore, the pure PP shows a single degradation process with the T 5wt% and T max at 373.5 • C and 455.4 • C under nitrogen atmosphere. Moreover, there is no char residue left at 800 • C no matter whether under nitrogen or air atmosphere, indicating pure PP fails to form char alone. On the contrary, two-stage degradation is observed for all IFR-PP blends. The first-stage at low temperature is mainly ascribed to the degradation of the IFR and PP, and the second-stage at high temperature is attributed to the decomposition products (polyphosphoric acids, pyrophosphate, and metaphosphate acid) of APP or MAPP [1]. Furthermore, all samples display a lower degradation temperature under the air atmosphere than that under the nitrogen atmosphere owing to thermo-oxidative degradation [24]. As compared with the pure PP, all PP/APP samples display greater T 5wt% values, which is ascribed to the protection of the thick polyphosphoric acid fluid produced by the decomposition of APP [25]. In the case of substitution of APP with MAPP, the T 5wt% values of PP/MAPP samples move to lower temperatures since the T 5wt% of MAPP is lower than that of APP. It is noteworthy that the char residue of PP/25%MAPP is up to 5.4% at 800 • C under the air atmosphere, which is higher than that of pure PP and PP/APP samples. Such an observation demonstrates that the phosphoric acid released by APP can catalyze the carbon formation of the piperazine ring from AEP and the physical carbon formation of the benzene ring in ASC make MAPP itself can form a stable carbon layer, which acts as a barrier between oxygen and PP matrix [11].

Gas Phase Analysis
A TG-FTIR test was performed to analyze the volatile gases released from IFR-PP samples, which helps to explore the flame-retardant mechanism of the gas phase. Figure 7 illustrates the FTIR curves and 3D spectra of the evolved gaseous products during the pyrolysis of pure PP, PP/25%APP, and PP/25%MAPP. All the samples display the following characteristic peaks: alkanes (2960, 2920 cm −1 ), alkenes (1647, 1465, and 1371 cm −1 ) and a diene (890 cm −1 ) [10]. However, the intensity of the abovementioned peaks in the spectra of PP/25%MAPP is weaker than that of the pure PP and PP/25%APP, suggesting that volatile components are effectively suppressed during the degradation process of PP/25%MAPP specimens [21]. As compared with the pure PP, both PP/25%APP and PP/25%MAPP specimens show characteristic peaks ascribed to P=O (1258 cm −1 ), P-O (1087 cm −1 ), and NH 3 (966 and 937 cm −1 ), which contribute to the thermal degradation of APP and scission of polyphosphoric acids [29]. Furthermore, it is noticeable that a new peak at 1499 cm −1 (SO 2 ) is found only in the spectrum of the PP/25%MAPP specimen [22,30]. The appearance of these non-combustible gases (NH 3 and SO 2 ) can dilute the released combustible gases (hydrocarbons, etc.). process of PP/25%MAPP specimens [21]. As compared with the pure PP, both PP/25%APP and PP/25%MAPP specimens show characteristic peaks ascribed to P=O (1258 cm −1 ), P-O (1087 cm −1 ), and NH3 (966 and 937 cm −1 ), which contribute to the thermal degradation of APP and scission of polyphosphoric acids [29]. Furthermore, it is noticeable that a new peak at 1499 cm −1 (SO2) is found only in the spectrum of the PP/25%MAPP specimen [22,30]. The appearance of these non-combustible gases (NH3 and SO2) can dilute the released combustible gases (hydrocarbons, etc.).

Analysis of Char Residue
Digital photographs of the residues after CCT are shown in Figure 8. Obviously, there is no residue left for pure PP and a minute quantity of residues for PP/25%APP. This indicates that the incorporation of the unmodified APP alone cannot promote the formation of an expanded char layer. However, more intumescent and expanded char residues with expansion heights of 1.

Analysis of Char Residue
Digital photographs of the residues after CCT are shown in Figure 8. Obviously, there is no residue left for pure PP and a minute quantity of residues for PP/25%APP. This indicates that the incorporation of the unmodified APP alone cannot promote the formation of an expanded char layer. However, more intumescent and expanded char residues with expansion heights of 1.7 cm (Figure 8c') and 2.5 cm (Figure 8d') are observed for PP/22.5%MAPP and PP/25%MAPP, respectively. The expansion height is usually adopted to assess the swelling degree or quality of the char layer. Accordingly, it is concluded that MAPP has a greater effect on the formation of the intumescent char layer than the unmodified APP, which can obstruct the heat and mass transfer between the gas and condensed phase. expansion height is usually adopted to assess the swelling degree or quality of the char layer. Accordingly, it is concluded that MAPP has a greater effect on the formation of the intumescent char layer than the unmodified APP, which can obstruct the heat and mass transfer between the gas and condensed phase. The microstructure of the residues after CCT was observed further via SEM, and the resultant images are presented in Figure 9. Clearly, numerous large holes are observed both in the outer and inner char residues of PP/25%APP whilst compact and continuous char layer is observed on the surface of PP/MAPP blends. Especially, the appearance of some char bubbles in the PP/25%MAPP surface residues indicates that the incorporation of MAPP enhances the compactness of char residues and thus forms a much excellent IFR system [5]. However, the honeycomb structure still exists in the inner char residue of PP/MAPP samples, which is ascribed to the diffusion of pyrolysis gas. The microstructure of the residues after CCT was observed further via SEM, and the resultant images are presented in Figure 9. Clearly, numerous large holes are observed both in the outer and inner char residues of PP/25%APP whilst compact and continuous char layer is observed on the surface of PP/MAPP blends. Especially, the appearance of some char bubbles in the PP/25%MAPP surface residues indicates that the incorporation of MAPP enhances the compactness of char residues and thus forms a much excellent IFR system [5]. However, the honeycomb structure still exists in the inner char residue of PP/MAPP samples, which is ascribed to the diffusion of pyrolysis gas.
The microstructure of the residues after CCT was observed further via SEM, and the resultant images are presented in Figure 9. Clearly, numerous large holes are observed both in the outer and inner char residues of PP/25%APP whilst compact and continuous char layer is observed on the surface of PP/MAPP blends. Especially, the appearance of some char bubbles in the PP/25%MAPP surface residues indicates that the incorporation of MAPP enhances the compactness of char residues and thus forms a much excellent IFR system [5]. However, the honeycomb structure still exists in the inner char residue of PP/MAPP samples, which is ascribed to the diffusion of pyrolysis gas. The elemental compositions of the char residue of the PP/25%MAPP sample after CCT were investigated by the XPS analysis ( Figure 10). As can be seen, the main elements in the residue are C, O, P, and N. In N1S spectra (Figure 10b), the peaks at 400.5 eV, 401.2 eV, and 402.3 eV are attributed to P-N-P, NH4 +, and P-C-N, respectively [21,23]. The O1S peaks (Figure 10c  The elemental compositions of the char residue of the PP/25%MAPP sample after CCT were investigated by the XPS analysis ( Figure 10). As can be seen, the main elements in the residue are C, O, P, and N. In N 1S spectra (Figure 10b), the peaks at 400.5 eV, 401.2 eV, and 402.3 eV are attributed to P-N-P, NH 4 +, and P-C-N, respectively [21,23]. In the case of P 2P spectra (Figure 10d), the peaks around 133.9 eV and 134.5 eV correspond to P=O and P-O-C structure, respectively [22,31]. The results above demonstrate that the existence of P-C-N and P-O-C structure plays a vital role in forming a dense and continuous char layer.
Polymers 2020, 12, x 12 of 17 correspond to P=O and P-O-C structure, respectively [22,31]. The results above demonstrate that the existence of P-C-N and P-O-C structure plays a vital role in forming a dense and continuous char layer. Raman spectroscopy is a useful tool to characterize the degree of graphitization of char residues. The Raman spectra are mainly divided into G bands (about 1590 cm −1 , showing the graphitic structure) and D bands (about 1365 cm −1 , representing lattice defects of carbon atom). Furthermore, the degree of graphitization of the char residue can be assessed by the area ratio of two peaks (AD/AG). Generally, the lower ratio of AD/AG means the higher graphitization degree and better thermaloxidative stability of the char [32,33]. As shown in Figure 11, the AD/AG values are calculated to be Raman spectroscopy is a useful tool to characterize the degree of graphitization of char residues. The Raman spectra are mainly divided into G bands (about 1590 cm −1 , showing the graphitic structure) and D bands (about 1365 cm −1 , representing lattice defects of carbon atom). Furthermore, the degree of graphitization of the char residue can be assessed by the area ratio of two peaks (A D /A G ). Generally, the lower ratio of A D /A G means the higher graphitization degree and better thermal-oxidative stability of the char [32,33]. As shown in Figure 11, the A D /A G values are calculated to be 2.57 for PP/25%APP, 2.24 for PP/22.5%MAPP, and 2.12 for PP/25%MAPP, respectively. As such, it is concluded that the incorporation of MAPP facilitates the graphitization to form much dense and continuous surface char layers, which is in agreement with the SEM observation results. Raman spectroscopy is a useful tool to characterize the degree of graphitization of char residues. The Raman spectra are mainly divided into G bands (about 1590 cm −1 , showing the graphitic structure) and D bands (about 1365 cm −1 , representing lattice defects of carbon atom). Furthermore, the degree of graphitization of the char residue can be assessed by the area ratio of two peaks (AD/AG). Generally, the lower ratio of AD/AG means the higher graphitization degree and better thermaloxidative stability of the char [32,33]. As shown in Figure 11, the AD/AG values are calculated to be 2.57 for PP/25%APP, 2.24 for PP/22.5%MAPP, and 2.12 for PP/25%MAPP, respectively. As such, it is concluded that the incorporation of MAPP facilitates the graphitization to form much dense and continuous surface char layers, which is in agreement with the SEM observation results.

Possible Flame-Retardant Mechanism
Based on the analysis of the volatile gas and char residues, a possible flame-retardant mechanism is proposed and illustrated in Figure 12. In the initial stage of the combustion, the dehydration of MAPP and the cleavage of ionic bonds result to release of some incombustible gases (such as NH 3 , SO 2 , and water vapor) which can fully dilute the oxygen and heat. With rising temperature, MAPP is thermally decomposed into polyphosphoric acid, pyrophosphoric acid, or metaphosphoric, all of which can esterify and crosslink with the char precursor (piperazine ring) [10]. Moreover, the phosphate ester radicals and pyrophosphate radicals released from the decomposition of MAPP readily quench O·, H·, and OH· radicals [34]. Additionally, the benzene ring in the intermediate acts as both the cross-linking agent and crystal core [23]. As such, the char layer containing P-O-C and P-C-N structure is formed and its quality is improved owing to the crosslinking networks. It is noteworthy that although the initially formed char layer did not swell, its continuous and compact structure hinders the gas overflowing and the accumulation of the incombustible gases keeps the char layer expanding. As a result, an intumescent and dense char layer is formed, which protects the underlying substrate from further thermal decomposition. The physical (expansion and morphology) and chemical (thermal stability) properties of the char layer were enhanced after introducing MAPP and the possible IFR mechanism is shown in Figure 12 [21]. Figure 13 illustrates the typical stress-strain curves of pure PP and IFR-PP blends, and the mechanical properties data are provided in Table 5. The pure PP gives 39.1 MPa of tensile strength, 147.1% of elongation at break, and 4.6 kJ × m −2 of impact strength, respectively. Although the mechanical properties of IFR-PP blends are decreased to a certain extent by comparison with the pure PP, the PP/25%MAPP sample shows better mechanical properties than the PP/25%APP. Namely, the tensile strength, elongation at break, and impact strength of PP/25%MAPP are increased by 12.3%, 110.5%, and 37.9% as compared with PP/25%APP, respectively. This difference might be related to interface interaction between the dispersion phase (MAPP or APP) and the PP matrix. Figure 14 presents the SEM images of the impact fractured surface of PP, PP/25%APP, and PP/25%MAPP. APP particles fail to adhere to the PP matrix, and the interface between the APP particles and PP matrix is clearly visible (Figure 14b). Conversely, the MAPP particles are uniformly embedded in the PP matrix (Figure 14c), which leads to an improvement of the elongation at break and impact strength. This observation is probably ascribed to that the modification of APP improved the interfacial adhesion, and the addition of MAPP had little effect on the mechanical properties of IFR-PP samples.

Mechanical Properties
SO2, and water vapor) which can fully dilute the oxygen and heat. With rising temperature, MAPP is thermally decomposed into polyphosphoric acid, pyrophosphoric acid, or metaphosphoric, all of which can esterify and crosslink with the char precursor (piperazine ring) [10]. Moreover, the phosphate ester radicals and pyrophosphate radicals released from the decomposition of MAPP readily quench O . , H . , and OH . radicals [34]. Additionally, the benzene ring in the intermediate acts as both the cross-linking agent and crystal core [23]. As such, the char layer containing P-O-C and P-C-N structure is formed and its quality is improved owing to the crosslinking networks. It is noteworthy that although the initially formed char layer did not swell, its continuous and compact structure hinders the gas overflowing and the accumulation of the incombustible gases keeps the char layer expanding. As a result, an intumescent and dense char layer is formed, which protects the underlying substrate from further thermal decomposition. The physical (expansion and morphology) and chemical (thermal stability) properties of the char layer were enhanced after introducing MAPP and the possible IFR mechanism is shown in Figure 12 [21].  Figure 13 illustrates the typical stress-strain curves of pure PP and IFR-PP blends, and the mechanical properties data are provided in Table 5. The pure PP gives 39.1 MPa of tensile strength, 147.1% of elongation at break, and 4.6 kJ×m −2 of impact strength, respectively. Although the mechanical properties of IFR-PP blends are decreased to a certain extent by comparison with the pure PP, the PP/25%MAPP sample shows better mechanical properties than the PP/25%APP. Namely, the tensile strength, elongation at break, and impact strength of PP/25%MAPP are increased by 12.3%, 110.5%, and 37.9% as compared with PP/25%APP, respectively. This difference might be related to interface interaction between the dispersion phase (MAPP or APP) and the PP matrix. Figure 14  presents the SEM images of the impact fractured surface of PP, PP/25%APP, and PP/25%MAPP. APP particles fail to adhere to the PP matrix, and the interface between the APP particles and PP matrix is clearly visible (Figure 14b). Conversely, the MAPP particles are uniformly embedded in the PP matrix (Figure 14c), which leads to an improvement of the elongation at break and impact strength. This observation is probably ascribed to that the modification of APP improved the interfacial adhesion, and the addition of MAPP had little effect on the mechanical properties of IFR-PP samples.     Figure 13. Typical stress-strain curves of pure PP and its flame-retardant blends.

Conclusions
A novel mono-component flame retardant (MAPP) was prepared by the modification of conventional APP with piperazine sulfonate. The MAPP showed more effective flame retardancy and smoke suppression than the traditional APP. The PP/MAPP composite passed the UL-94 V-0 rating and achieved 30% of LOI value at 22.5 wt% loading level of MAPP. The HRR, PHRR, and TSP values of PP/22.5%MAPP were reduced significantly as compared with those of the pure PP or PP/25%MAPP. The results of TGA-FTIR and char residue analysis revealed that some incombustible gases were released and intumescent, dense and continuous char residues were produced with the incorporation of MAPP. The PP/25%MAPP sample also displayed better mechanical properties than the PP/25%APP.

Conclusions
A novel mono-component flame retardant (MAPP) was prepared by the modification of conventional APP with piperazine sulfonate. The MAPP showed more effective flame retardancy and smoke suppression than the traditional APP. The PP/MAPP composite passed the UL-94 V-0 rating and achieved 30% of LOI value at 22.5 wt% loading level of MAPP. The HRR, PHRR, and TSP values of PP/22.5%MAPP were reduced significantly as compared with those of the pure PP or PP/25%MAPP. The results of TGA-FTIR and char residue analysis revealed that some incombustible gases were released and intumescent, dense and continuous char residues were produced with the incorporation of MAPP. The PP/25%MAPP sample also displayed better mechanical properties than the PP/25%APP.