Synergistic effect of phosphorus–nitrogen and silicon-containing chain extenders on the mechanical properties, flame retardancy and thermal degradation behavior of waterborne polyurethane

Abstract

With the aim to keep a balance between the flame retardancy and thermal stability as well as mechanical properties of waterborne polyurethane (WPU), a novel phosphorus–nitrogen–silicon containing flame retardant WPU (FRWPU) was synthesized by conjugating with a cyclic phosphoramidate lateral group bearing diol (named as PNMPD) and silane coupling agent KH-602 in the chain-extension and post-chain extension process, respectively. Significant enhancement in tensile strength (6.1 MPa) is obtained with the combined covalent incorporation of PNMPD and KH-602 rather than using PNMPD alone. A limiting oxygen index (LOI) value of 27.7% and a UL-94 vertical burning V-0 rating is achieved for FRWPU-12.6 with 12 wt% PNMPD and 60% post-chain extension ratio by KH-602, while the peak heat release rate (PHRR), total heat release (THR), peak smoke produce rate (SPR) and total smoke production (TSP) characterized by a cone calorimeter (CC) markedly reduce by 36.0%, 42.9%, 40.1% and 35.4%, respectively, compared to those of pure WPU, which is more efficient in flame retardancy than FRWPU-12.0 merely loaded with 12 wt% PNMPD. Meanwhile, the thermal degradation behaviors and flame-retardant mechanism of FRWPU were consistently confirmed by thermogravimetric analysis (TGA), scanning electron microscopy (SEM), thermogravimetry-Fourier transform infrared (TG-FTIR) spectroscopy, X-ray photoelectron spectroscopy (XPS) and Raman spectroscopy. All results indicate that the interactions between phosphorus and silicon elements in the condensed phase are mainly responsible for the dramatically reduced fire hazards, which inhibits the heat and flammable gas release and facilitates the formation of a more thermally stable graphitized char layer consisting of –P(=O)–O–Si– structures.

---

1. Introduction

Waterborne polyurethane (WPU) is one of the versatile and eco-friendly polymeric materials that has increasingly found numerous industrial applications covering leather/synthetic leather finishing, textile laminating, coatings, packaging, adhesives, sealants, pigment pastes and automotive.^1,2 However, the intrinsic flammability of WPU, which has resulted in enormous fire-related casualties and property losses over the past few years,³ seriously restricts its further application in many fields. Thus, developing flame-retardant WPU has been highly desirable.

Generally, mechanically incorporating halogen-, phosphorus-, nitrogen- and silicon-containing flame retardants into WPU is recognized as the most convenient and efficient method to improve the flame retardancy.^3–5 Nevertheless, such conventional blending technique has the disadvantages of high loading (normally ≥20 wt%), poor compatibility, easy leaching, and deteriorated mechanical property.^5–7 To this end, scientific interest has principally focused on covalent conjugation reactive-type flame retardants into polyurethane backbone or side chains. As recent examples, chain extenders with both phosphorus and nitrogen elements have distinctly demonstrated potential to furnish WPU with permanent fire retardance owing to the synergistic flame retardant effect.^8–10 Despite the aforementioned advantages, the possible threat to the hydrolytic decomposition of organophosphate moieties that conjugated in WPU backbone during storage therein, is particularly problematic. That is, such hydrolization will come at a sacrifice of cleavage of WPU main-chain structure, thus resulting in not merely reduced flame retardancy but also physical properties.¹¹ Very recently, our research group found that bearing cyclic organophosphate and phosphoramidate moieties in WPU lateral chains rather than backbone can efficiently minimize the hydrolysis of phosphorus-based flame retardants.^12,13 Unfortunately, the tensile strength and thermal stability of WPU decrease slightly which is unfavorable. In fact, similar issues also arise in other phosphorus–nitrogen flame retardants.^14–16 Therefore, it remains urgently imperative but challenging to explore inherent flame retardant WPU that combined with favorable hydrolysis-resistance, thermal and mechanical properties.

With regard to the prepolymer dispersion process, the post-chain extension is considered to be a critical green approach to bestowing WPU with better performance since it consumes fewer organic solvents.¹⁰ Although numerous researchers have consistently noted that the thermal stability, aging resistance and mechanical strength of WPU can be significantly enhanced with the amine post-chain extenders,^17,18 little efforts were conducted on improving the flame retardancy of WPU by post-chain extension technique. Luo et al. synthesized a post-chain extension flame retardant WPU by using phosphorus-containing diamine (BPPO), and results showed that the flame resistance was increased with the concentrated BPPO content.¹⁰ However, the phosphoester bond of BPPO, which is sensitive to moisture, was incorporated into WPU main chain.

In this research, we herein demonstrate the preparation of flame-retardant WPU (FRWPU) containing phosphorus–nitrogen–silicon by chemically anchoring a cyclic phosphoramidate lateral group bearing diol (PNMPD) in the chain-extension process, followed by post-chain extension with commercially available silane coupling agent KH-602. The formation of stable reticulated di-substituted Si–O–Si cross-linked structures derived from the condensation reaction of KH-602 is expected to simultaneously overcome the disadvantages of phosphorus-containing PNMPD on the mechanical and thermal property of polyurethane, and exhibit a phosphorus–silicon synergistic flame retardant effect with PNMPD for FRWPU. To verify our assumptions, a complementary study on the mechanical behaviors, fire performances and possible flaming mechanism of this FRWPU was also systematically investigated and discussed.

2. Experimental

2.1 Materials

The cyclic phosphoramidate lateral group bearing diol, namely 2-(5,5-dimethyl-2-oxo-2λ⁵-1,3,2-dioxaphosphinan-2-ylamino)-2-methyl-propane-1,3-diol (PNMPD), was synthesized according to our previous work.¹³ Polypropylene glycol with average functionality of 2.0 and molecular weight of 2000 (PPG2000), neopentyl glycol (NPG), and dimethylolpropionic acid (DMPA) were purchased from Aladdin Reagent Co., Ltd. (Shanghai, China) and dehydrated under vacuum at 120 °C for at least 2 h before use. 4,4′-Diisocyanatodicyclohexylmethane (H₁₂MDI) was obtained from Hersbit Chemical Co., Ltd. (Shanghai, China). Triethylamine (TEA), acetone, stannous octoate (Sn(Oct)₂, catalyst), and 3-(2-aminoethylamino)propyl-dimethoxymethylsilane (KH-602) were supplied by Kelong Reagent Co. (Chengdu, China).

2.2 Synthesis of flame retardant waterborne polyurethane (FRWPU) dispersions

The FRWPU dispersions are synthesized through a prepolymer mixing process as schematically depicted in Scheme 1. Briefly, a stoichiometric ratio of PPG2000, H₁₂MDI and DMPA were firstly charged into a 500 mL round-bottom, four-necked separable flask fitted with a mechanical Teflon stirrer, thermometer, nitrogen inlet and reflux condenser, and reacted at 80 °C for about 3 h with the presence of catalyst to obtain NCO terminated ionomer segments. After that, NPG and PNMPD dissolved in acetone were injected into the mixture and stirred continuously for another 2–3 h until the NCO content reached a theoretical value, yielding a PNMPD-conjugated prepolymer. Subsequently, the prepolymer was cooled to 50 °C, and TEA (DMPA equiv.) was added for neutralization reaction in a period of 15 min. An aqueous dispersion was accomplished by emulsifying the resulting mixture under vigorous stirring with de-ionized water slowly charged in, followed immediately by post-chain extension reaction with KH-602 at ambient temperature for next 2 h. After removing the acetone, the FRWPU emulsion was obtained, of which the solid content was about 35 wt%. The resultant samples were designated as FRWPU-x.y, and their theoretical compositions are tabulated in Table 1. Here x and y denote the weight percentage of PNMPD and chain extension ratios of KH-602, respectively.

Scheme 1 Schematic illustration for the synthesis of FRWPU.

Table 1 Composition and particle size of WPU and FRWPU dispersions

Sample^a	Theoretical composition (g)								Particle size (nm)
	H12MDI	PPG2000	DMPA	NPG	PNMPD	TEA	H2O	KH-602
a Samples here were abbreviated as FRWPU-x.y, wherein x and y demonstrated the weight percentage of PNMPD and post-chain extension ratios of KH-602, respectively.
WPU	23	30	2.9	3.0	0	2.19	110	0	62.09
FRWPU-6.0	23	30	2.9	1.5	3.65	2.19	114	0	85.23
FRWPU-6.3	23	30	2.9	1.5	3.65	2.19	114	1.38	95.61
FRWPU-6.6	23	30	2.9	1.5	3.65	2.19	114	2.76	106.6
FRWPU-12.0	23	30	2.9	0	7.3	2.19	118	0	105
FRWPU-12.3	23	30	2.9	0	7.3	2.19	118	1.38	117.1
FRWPU-12.6	23	30	2.9	0	7.3	2.19	118	2.76	146.8

---

2.3 Preparation of FRWPU cast films

Polyurethane films were prepared by pouring the above synthesized dispersions into a freshly cleaned polytetrafluoroethylene plate mold (15 cm × 15 cm) at ambient conditions for 5 days, followed by annealed in a vacuum at 80 °C for 24 h. Then the obtained films with typical thickness of 0.3 mm and 3.2 mm were conditioned in a desiccator containing discoloration silicone before performance tests.

2.4 Characterization

The chemical structure of FRWPU samples and the exterior chars obtained after degradation at desired temperature in the tube furnace under nitrogen was determined by a Thermo Scientific Nicolet IS10 Fourier transform infrared (FTIR) spectrometer, and the wavenumber range was set from 400–4000 cm⁻¹.

The average particle size of FRWPU dispersions were measured with a Malvern Model Zeta-sizer ZEN3600 light scattering ultrafine particle analyzer (UK) at 25 °C.

Tensile measurements of FRWPU films were completed on a UTM 6203 universal testing machine in accordance with ASTM D 638 procedure at a crosshead speed of 10 mm min⁻¹ at room temperature. At least five specimens with a gauge size of 50 × 10 mm² (L × W) were tested.

Thermogravimetric analysis (TGA) was evaluated using a Netzsch TG-209F1 thermal analyzer by heating from R.T. to 600 °C at a heating rate of 10 °C min⁻¹ under nitrogen atmosphere.

The limiting oxygen index (LOI) values of FRWPU specimens were determined with an Oxygen Index Instrument (Fire Testing Technology, UK) in terms of ASTM D 2863-2009 standard, and the sample dimensions were 127 × 10 × 3.2 mm³.

The vertical burning (UL-94) test of FRWPU films was assessed by FTT0082 instrument (Fire Testing Technology, UK) according to ASTM Standard D 3801-2010 with bar-shaped sample dimensions of 130 × 12.7 × 3 mm³.

The cone calorimeter tests were evaluated using a FTT cone calorimeter, following the procedures in ISO 5660-1. Squared FRWPU specimens of 100 × 100 × 2 mm³ were horizontally irradiated under an incident flux of 35 kW m⁻².

The scanning electron microscopy (SEM) micrographs (Hitachi Model S-4700, Japan) were carried out to characterize the surface morphology of FRWPU char residues after cone calorimeter tests at an accelerating voltage of 20 kV. And the equipped Energy Dispersive X-ray (EDX) spectroscopy was used for qualitative and quantitative elemental analysis of chars.

The evolved gaseous products from FRWPU films during thermal decomposition were characterized by thermogravimetric analysis-infrared spectrometry (TG-IR), using a Netzsch TG-209F1 thermal analyzer that coupled with Nicolet 6700 FTIR spectrophotometer. About 10.0 mg of the samples were conducted under a nitrogen flow with a heating rate of 10 °C min⁻¹.

X-ray photoelectron spectroscopy (XPS) spectra were recorded on a XSAM80 (Kratos Co, UK) spectrometer, using a monochromated Al Kα excitation radiation (1486.6 eV) at a base pressure of 2 × 10⁻⁹ mbar. And the XPS Peak version 4.1 program was used for data analysis.

Raman spectra of residue were performed with a SPEX-1403 laser Raman spectrometer (SPEX Co., USA) from 600 to 2000 cm⁻¹ at room temperature using a 514.5 nm argon ion laser.

3. Results and discussion

3.1 Structural characterization of FRWPU

FT-IR spectroscopy was conducted to qualitatively investigate the chemical structure of FRWPU. Fig. 1 shows the FT-IR spectra of WPU and FRWPU. The absorption bands centered at 3353 cm⁻¹, 1713 cm⁻¹, 1536 cm⁻¹ and 1105 cm⁻¹ are the characteristic peaks of WPU, assigning to urethane N–H stretching, stretching vibration of C=O groups, urethane N–H bend vibration, and nonsymmetric stretching vibration of C–O–C, respectively. Concerning FRWPU, two new peaks at 1660 cm⁻¹ and 971 cm⁻¹ appear, which are respectively attributable to N–H bending vibration of phosphamide and P–N stretching vibration from PNMPD. Unfortunately, due to the band overlapping in the 1000–1200 cm⁻¹ region of interest, the P–O–C and Si–O–Si characteristic peaks are visually indistinguishable from C–O–C. However, the intensity of N–H stretching vibration in FRWPU around 3200–3500 cm⁻¹ is obviously increased. Simultaneously, the absorption intensity increases with the increment of PNMPD content and post-chain extension ratio. This is presumably ascribed to the extra hydrogen bonding interactions induced by carbonyl groups with phosphamide in PNMPD segment and urea N–H bonds, wherein the urea groups are yielded by the polyaddition of NCO-terminated prepolymer and amino silane coupling agent KH-602.

Fig. 1 FTIR spectra of WPU and FRWPU.

3.2 Flame retardancy of FRWPU

LOI and UL-94 vertical burning tests are used to evaluate the flammability of WPU and FRWPU films, and the corresponding results are presented in Table 2. WPU is readily ignitable with a low LOI value of 18.4%, and it burns out quickly yielding serious melted polymer droplets after first ignition in UL-94 test, thus receiving no rating. With the conjugation of PNMPD, the LOI values of FRWPU-6.0 and FRWPU-12.0 films increase to 22.3% and 27.2%, respectively. This demonstrates that the reactive PNMPD monomer positively endows polyurethane with flame retardancy, which is also verified by UL-94 results. Specifically, the average flaming time for FRWPU-6.0 after ignition is 19 s and 8 s, respectively, but the flaming drips ignite the cotton placed below specimen. Hence, V-2 rating is achieved for FRWPU-6.0. While for FRWPU-12.0, the flames can self-extinguish with 3 s after removing the burner and no molten droplets are observed. Consequently, FRWPU-12.0 reaches a UL-94 V-0 rating. After post-chain extension by silane coupling agent KH-602, the flame retardancy of FRWPU specimens is further improved. For example, the LOI values of FRWPU-6.6 and FRWPU-12.6 are 22.6% and 27.7%, respectively, increased by 0.3% and 0.5% over that of FRWPU-6.0 and FRWPU-12.0. Comparatively speaking, the gains in LOI values are larger than the discrepancies between FRWPU that chain-extended only by KH-602 and referenced WPU (growth by 0.2%, see Table S1, ESI†). Furthermore, the dripping cannot ignite the underlying cotton for FRWPU-6.6 during UL-94 test, whilst the duration of flaming combustion for FRWPU-12.6 is shortened to 2 s and 0 s after the first and second flame application, respectively, which denotes that synergistic effectiveness on flame retardancy probably exist between phosphorus, nitrogen and silicon.

Table 2 LOI values and UL-94 ratings of WPU and FRWPU specimens

Sample	LOI (%)	UL-94 rating	Drips	AFT^a (s)	Igniting cotton
a Average flaming combustion after the first and second burner flame application.b Specimen burns up to holding clamp after the first ignition.
WPU	18.4 ± 0.2	No rating	Yes	53 ± 4/—^b	Yes
FRWPU-6.0	22.3 ± 0.1	V-2	Yes	19 ± 3/8 ± 3	Yes
FRWPU-6.3	22.4 ± 0.2	V-2	Yes	19 ± 3/8 ± 2	Yes
FRWPU-6.6	22.6 ± 0.2	V-1	Yes	17 ± 2/6 ± 2	No
FRWPU-12.0	27.2 ± 0.2	V-0	No	3 ± 1/2 ± 1	No
FRWPU-12.3	27.4 ± 0.1	V-0	No	3 ± 1/2 ± 1	No
FRWPU-12.6	27.7 ± 0.2	V-0	No	2 ± 1/0	No

---

Cone calorimetry is one of the most effective methods for assessing the dynamic combustion behavior of polymers under a well-defined developing fire scenario.¹⁹ From which, various salient combustion parameters such as time to ignition (TTI), heat release rate (HRR), peak heat release rate (PHRR), total heat release (THR), smoke produce rate (SPR), total smoke production (TSP), and etc. are provided. Fig. 2 displays the HRR, THR, SPR and TSP versus time curves of WPU and FRWPU, and the detailed data are summarized in Table 3. The WPU is inherently combustible with a high PHRR value of 1117 kW m⁻². In comparison with WPU, the conjugation of PNMPD significantly reduces the PHRR of FRWPU-6.0 and FRWPU-12.0 to 904 kW m⁻² and 742 kW m⁻², respectively, clearly signifying great enhancements in the flame retardancy of FRWPU. Moreover, after incorporating with KH-602, the PHRR of FRWPU-12.6 shows a further reduction to 715 kW m⁻². This result offers another supporting evidence to validate the possible synergistic flame-retarded effects of P and Si elements on polyurethane. The THR shows quite similar trend to HRR, and those values of FRWPU-6.0, FRWPU-12.0 and FRWPU-12.6 are remarkably decreased by 26.2%, 37.7% and 42.9%, respectively, compared to that of WPU. On the other hand, TTI and time to PHRR (T_PHRR) are predated by covalent conjugating with PNMPD due to the early decomposition of phosphoester bond. Nevertheless, a slight prolongation in TTI of FRWPU-12.6 is achieved when KH-602 is loaded, which should be ascribed to the formation of thermal stable siloxane linkages between polyurethane chains.

Fig. 2 HRR (a), THR (b), SPR (c) and TSR (d) versus time curves of WPU and FRWPU films.

Table 3 Detailed experimental results of WPU and FRWPU obtained from cone calorimetry

Sample	WPU	FRWPU-6.0	FRWPU-12.0	FRWPU-12.6
TTI (s)	35	27	23	24
THR (MJ m^−2)	74.9	55.3	46.7	42.8
PHRR (kW m^−2)	1117	904	742	715
TPHRR (s)	240	200	205	200
FIGRA (kW m^−2 s^−1)	4.65	4.52	3.62	3.58
TSR (m^2 m^−2)	441	368	327	285
PSPR (m^2 s^−1)	0.066	0.059	0.043	0.039
Residue (%)	2.3	5.9	15.1	18.6

---

To evaluate the burning propensity and fire hazards of polyurethane more clearly, the fire growth rate index (FIGRA) is calculated according to the proportion of PHRR and T_PHRR.²⁰ Basically, lower FIGRA value implies there is ample time to evacuate people in distress.²¹ For WPU, the FIGRA is 4.65 kW m⁻² s⁻¹. The FIGRA values of FRWPU decrease to 4.52 kW m⁻² s⁻¹ and 3.62 kW m⁻² s⁻¹ when PNMPD content is 6 wt% and 12 wt%, respectively. While the FRWPU-12.6 possesses the lowest FIGRA value of 3.58 kW m⁻² s⁻¹, declaring the interaction between PNMPD and KH-602 imparts better fire safety.

The smoke release property during combustion has long been considered as another critical parameter to estimate the fire hazards of polymers since the poisonous smoke can reduce the visibility or even result in the trapped people to death by suffocation.^19,21 As depicted in Fig. 2(c) and (d), WPU releases large amount of smoke with a high PSPR of 0.066 m² s⁻¹ and TSP of 441 m² m⁻², while the presence of 12 wt% PNMPD reduces the PSPR and TSP to 0.043 m² s⁻¹ and 327 m² m⁻², declined by 34.8% and 25.9%, respectively. With the conjugation of KH-602, the FRWPU-12.6 exerts a better smoke suppression effect. This can be described as follows: during fire, the flame retardant P and Si elements facilitate the carbonization of polyurethane, which plays a main role as char enhancer to impede the emission of smoke.²² The assumption is supported by the dramatically promoted burning residues as shown in Table 3.

3.3 Thermal degradation behaviour

The TGA and derivative thermogravimetric (DTG) curves of FRWPU films in N₂ atmosphere are presented in Fig. 3 with the corresponding data summarized in Table 4. As shown in Fig. 3, the 5% weight loss temperature (T_5%) and the temperature at maximum mass loss rate (T_max) of all FRWPU films are lower than that of WPU, which is attributed primarily to the degradation of minor stable O=P–O and P–O–C bond in PNMPD. This phenomenon was also observed in other phosphorus-containing polyurethane.¹⁵ However, the amount of char residue at 600 °C (Y_c) gradually increases with increasing PNMPD content. Usually more char layers help to achieve better flame retardant. It should be noteworthy that after post chain-extension with KH-602, the T_5% and T_max2 of FRWPU shift to sensible high temperature while the residue yield is increased, which is in accord with the cone results, suggesting that silane coupling agent can somewhat compensate the thermal stability deteriorated by organophosphorus-based flame retardant. This is reasonable as the interpenetrating and crosslinking network formed in KH-602 modified FRWPU restricts the movement of polyurethane molecular chains and increases the melt viscosity.^23,24 Additionally, the bond dissociation energy of Si–O linkage (110 kcal mol⁻¹) is higher than that of C–O bond (85.5 kcal mol⁻¹),²⁵ and thus accounts for the improved thermal resistance.

Fig. 3 TGA (a) and DTG (b) curves of the WPU and FRWPU films under nitrogen atmosphere.

Table 4 TGA data of WPU and FRWPU films under nitrogen atmosphere

Sample	T5% (°C)	Tmax1 (°C)	Tmax2 (°C)	Yc (wt%)
WPU	283.3	348.9	410.2	0.83
FRWPU-6.0	263.0	341.1	385.5	1.35
FRWPU-6.3	266.8	341.2	389.9	2.07
FRWPU-6.6	269.1	340.0	391.5	3.16
FRWPU-12.0	255.6	335.3	386.7	4.01
FRWPU-12.3	262.1	333.9	389.9	4.48
FRWPU-12.6	264.2	335.2	391.5	4.70

---

To better understand the thermal oxidative process, the chemical structures of WPU and FRWPU solid residues obtained after heat treatment at several representative temperatures for 15 min in tubular furnace under nitrogen atmosphere were investigated by FT-IR, as shown in Fig. 4. For WPU in Fig. 4(a), the distinct absorption peaks attributing to N–H (3403 cm⁻¹, 1546 cm⁻¹) and C=O bonds (1708 cm⁻¹) in urethane groups decrease gradually after 283 °C (T_5%) and almost vanish at 349 °C (T_max1), indicating that the degradation of hard segment, or more exactly, the cleavage of urethane linkage occurs in this corresponding temperature. At the meantime, the relative intensity of other characteristic absorption bands approximately at 2969 and 2853 cm⁻¹ (C–H stretching vibration), 1454 and 1375 cm⁻¹ (bending vibration of C–H), and 1108 cm⁻¹ (stretching vibration of C–O–C) hardly dim before T_max1. As the pyrolysis temperature elevates to 410 °C (T_max2), these absorptions dramatically decreases. Upon 450 °C, all bands vanish, which means the complete decomposition of WPU. It is worth noting that an obvious absorption peak appears at 1628 cm⁻¹ in the high-temperature region, unambiguously implying the formation of graphite-like carbonaceous structures during pyrolysis process.²⁶

Fig. 4 FTIR spectra for condensed products of WPU (a), FRWPU-12.0 (b) and FRWPU-12.6 (c) at different pyrolysis temperatures.

As for FRWPU-12.0, the absorption peak of P–O–C groups around 1102 cm⁻¹ (superimposed by C–O–C) begins to weaken at 256 °C, explicitly demonstrating the conjugation of PNMPD lower the thermo-stability of FRWPU, which corresponds well with the aforesaid TGA results. Actually, the degradation of phosphoester bond can accelerate the thermal decomposition of polyurethane due to the formation of polyphosphoric acids, which act as strong Lewis acids.⁷ This is supported by the dramatic diminution of all absorption intensities over 334 °C. Furthermore, the characteristic peaks representing FRWPU structures completely disappear at about 387 °C. However, other than the graphite structure located at 1629 cm⁻¹, FRWPU-12 exhibits some new absorption bands above 334 °C in comparison to WPU, as depicted in Fig. 4(b). Specifically, a strong peak at 1260 cm⁻¹, which is designated to the stretching vibration of P=O groups derived from phosphate–carbon complexes,²⁷ can be observed. Meanwhile, the noticeable peaks at 1042 cm⁻¹ combined with 930 cm⁻¹ in the finger-print region belong to the symmetrical and asymmetrical stretching vibrations of P–O–P bond, respectively, confirming the presence of thermal stable polyphosphate.²⁸ Hence, it can be interpreted that the PNMPD facilitates the formation of phosphorous-containing residues, which function as an efficient barrier to prevent the underneath polyurethane from further degradation, thus higher char yield is obtained as revealed by TGA and MCC analysis.

The FTIR spectra of FRWPU-12.6 at different degradation temperatures are given in Fig. 4(c). The thermal decomposition process of FRWPU-12.6 is basically similar to that of FRWPU-12.0 below 335 °C. Nevertheless, the serried sharp absorption peaks between 2800–2900 cm⁻¹ and 1400–1500 cm⁻¹, representative of hydrocarbon structures, are still clearly discerned at 391 °C. This implies the chemical tethering of KH-602 can enhance the thermal stability of phosphorus–nitrogen polyurethane. At elevated temperature, just as FRWPU-12.0, the absorption bands for P=O and P–O–P groups appear at 1256 cm⁻¹, 1052 cm⁻¹ and 930 cm⁻¹, respectively. It is worth noting that the new peak at 1073 cm⁻¹ is the signal of Si–O–Si stretching vibration,²⁹ while this is not visible in the spectra of WPU or FRWPU-12.0. That is to say, the silane coupler principally has activity in condensed phase. According to the above analysis, we speculate that some interaction between polyphosphate and siloxane structures should be existed in residual char layer, plausibly through the –P(=O)–O–Si– linkages.^29,30 Additionally, a small absorbance band at 883 cm⁻¹ corresponding to Si–O–P is supplementary evidence of the aforementioned reactions.³⁰

In order to further investigate the thermal degradation mechanism, the volatilized products of polyurethane films were collected and monitored by TG-FTIR analysis. Fig. 5 exhibits the three-dimension TG-FTIR spectra of gaseous volatiles evolved during the whole decomposition processes of WPU, FRWPU-12.0 and FRWPU-12.6 in nitrogen atmosphere. Intuitively, there is no significantly difference in pyrolysis products between WPU and FRWPUs during the thermal degradation process. More detailed information is obtained from the FTIR spectra of WPU, FRWPU-12.0 and FRWPU-12.6 at selected specific temperatures demonstrated in Fig. 6. For WPU, no volatile product is detectable when the pyrolysis temperature below 200 °C, as presented in Fig. 6(a). At 283 °C (T_5%), the distinct peaks between 2331 and 2358 cm⁻¹ are detected, which contributes to carbon dioxide. Meanwhile, some evolved volatile products attributing to H₂O (3500–3800 cm⁻¹), hydrocarbons (2976, 2932 and 2883 cm⁻¹), carbonyl compounds (1735 cm⁻¹), and C–O–C bond from ethers (1108 cm⁻¹),³¹ are initially identified. These unambiguous characteristic FTIR signals are continuously strengthened with increasing temperature, maximizing at around 410 °C (T_max2) and then gradually weakening. Notably, nitrile compounds (HCN and –NCO, 2265 cm⁻¹) are emitted after 283 °C due to the degradation of urethane groups in WPU, which still exists at 349 °C (T_max1). When the pyrolysis temperature is 349 °C and above, another sharp absorption peak with low intensity at 927 cm⁻¹ for NH₃ appears.

Fig. 5 3D TG-FTIR spectra of volatilized products for WPU (a), FRWPU-12.0 (b) and FRWPU-12.6 (c).

Fig. 6 TG-FTIR spectra of pyrolysis products of WPU (a), FRWPU-12.0 (b) and FRWPU-12.6 (c) at different temperatures.

For the FTIR spectra of FRWPU-12.0 given in Fig. 6(b), the gaseous volatiles exhibit characteristic bands of water (3500–3800 cm⁻¹), hydrocarbons (2850–2980 cm⁻¹), CO₂ (2358 and 2310 cm⁻¹), –NCO containing product (2265 cm⁻¹), carbonyl compounds (1735 cm⁻¹), ether components (1108 cm⁻¹), and NH₃ (927 cm⁻¹), which is similar to that of WPU. Although the pyrolysis species of WPU and FRWPU-12.0 are almost the same, the conjugation of PNMPD decreases the absorbance intensity of flammable organic gases, manifesting less fuel to fed back to the flame, and thus unequivocally reducing the heat release and fire hazards during pyrolysis,³² which corroborates with flame resistance measured by cone calorimeter. On the other hand, all the pyrolysis products for FRWPU-12.0 release earlier than WPU, in consequence of the inferior thermal stability of O=P–O in PNMPD. Noteworthily, no phosphorus-containing volatile products emit in the FTIR spectra of FRWPU-12.0 on the whole degradation process. Hence, it is reasonable to interpret that the phosphorus compounds mainly play positive parts in condensed phase flame retardancy mechanism. Additionally, a new absorption band belonging to NH₃ emerges at around 964 cm⁻¹ after 256 °C (T_5%). In fact, the non-flammable ammonia gas has a dilution effect on the oxygen concentration and organic volatiles surrounding the burning polyurethane,³³ slowing down the flame propagation.

Compared with FRWPU-12.0, FRWPU-12.6 has similar FTIR spectra of pyrolysis gaseous products. Still, there are some differences. Specifically, the absorbance intensity of CO₂ at 264 °C (T_5%) is reduced, signifying that the degradation of FRWPU has been delayed in the presence of thermal stable Si–O–Si crosslinking networks. Furthermore, fewer volatile combustible gases release for FRWPU-12.6, which may be ascrible to the synergistic effect of PNMPD and KH-602 on char formation. Another evidence to fortify the above assumption is that no discernable characteristic absorbance for silicon products is detected in the FTIR spectra of FRWPU-12.6. To this end, we surmise that the mechanism of PNMPD and KH-602 based flame retardant polyurethane is primarily condensed-phase through the formation of thermally stable char layers as a synergism of phosphorus, nitrogen and silicon.

3.4 Residual chars analysis

Previous studies have preliminarily validated that a condensed phase action is the main mechanism for the flame retardant FRWPU conjugated with PNMPD and KH-602. And the flame retardancy efficiency depends crucially on the morphology, structure and chemical composition of the char layer during combustion.³⁴ Fig. 7 plots the digital photographs of residual chars after cone calorimeter tests. WPU leaves negligible residual char after burning. There is a small handful of thin-bedded and cracked residue in FRWPU-6.0, whereas the char of FRWPU-12.0 becomes compact and intact. In particular, a more rigid and dense char residue for FRWPU-12.6 is observed, which can explain the improved heat and smoke parameters in cone test.

Fig. 7 Photographs of char residues from WPU (a), FRWPU-6.0 (b), FRWPU-12.0 (c) and FRWPU-12.6 (d) after cone calorimeter tests.

To further elucidate the effect of PNMPD and KH-602 on the char forming capability of FRWPU, the microstructure and morphology of the exterior residues obtained after cone calorimetry tests were investigated by SEM. As shown in Fig. 8(a), WPU exhibits a brittle morphology with many obvious holes and flaws on the surface, formed during the vigorous release of volatile gas. Therefore, it provides limited insulation effect for the underlying polyurethane during burning. Different from WPU, the residue morphology of phosphorus-containing FRWPU is smooth and integral with some rugged wrinkles, the “tortuous effect” of which can play an effective part on heat barrier. For FRWPU-12.6, the surface of charred layer becomes more compact and intumescent. In addition, a plentiful of corrugated folds is found on the surface, which serves as a skeleton to strengthen the residual char,³⁵ thus postponing the transfer of heat and oxygen between flame and unburned polyurethane. Simultaneously, the EDX mapping of FRWPU-12.6 in Fig. 9 demonstrates that phosphorus and silicon atoms are homogeneous distributed in the char surface. These results also indicate that PNMPD and KH-602 have a potent synergistic effect on char formation.

Fig. 8 SEM micrographs of the exterior residual char for WPU (a), FRWPU-12.0 (b) and FRWPU-12.6 (c) after cone test.

Fig. 9 EDX mapping of the exterior char of FRWPU-12.6 after cone calorimetry measurement (SE mode, 20 kV, C: carbon, O: oxygen, P: phosphorus, Si: silicon and N: nitrogen).

The semi-quantitative EDX analysis reveals the presence of C, O, P, Si and N atoms in the exterior residual char. To better understand the element composition and chemical bonds form, the char layer of WPU, FRWPU-12.0 and FRWPU-12.6 obtained from a cone calorimeter were further investigated by XPS analysis, whose results are demonstrated in Fig. 10 and Table 5. As can be seen from Table 5, the relative content of P_2p and Si_2p for FRWPU-12.6 char is 13.23% and 5.88%, respectively, the proportion of which (2.25 : 1) is equal to that in FRWPU-12.6 film (2.38 : 1). This phenomenon provides another robust testimony to support the P–Si synergy in condensed phase. In Fig. 10, the C_1s spectra of all samples can be spilt into three characteristic bands at 284.7 eV, 285.8 eV and 287.5 eV, which are designated to the C–H and C–C in aliphatic and aromatic species, C–O (ether and/or hydroxyl group and/or C–O–P) and C=O, respectively.³⁶ To evaluate the thermal oxidative resistance of the char, the C_ox/C_a ratios are calculated, where C_ox is the integrated intensities of oxidized carbons (C–O, C=O), and C_a represents the intensities of aliphatic and aromatic carbons (C–C, C–H).³⁴ As tabulated in Table 6, the C_ox/C_a values of the char layer for WPU and FRWPU-12.0 are 0.615 and 0.395, respectively, manifesting that phosphorous-containing PNMPD promotes the formation of more polyaromatic species in residual char. The formed polyaromatic cross-linked networks show much denser and compacted morphologies that prevent the internal polyurethane from heat and oxygen,²⁴ thereby enhancing the flame resistance. However, the C_ox/C_a value of FRWPU-12.6 increases to 0.484 after post-chain extension by KH-602, this may be ascribed to that some cross-linking structures are bridged by C–O–P and C–O–Si bonds to form a protective char layer.³⁷ In the O_1s spectra of WPU and FRWPU-12.0, two peaks at around 531.6 eV and 533.1 eV are representative of the double oxygen bonds in C=O or P=O, and simple oxygen bonds in C–O–C or C–O–P or P–O–P, respectively.³⁸ The residue of FRWPU-12.6 exhibits an additional shoulder peak at 534.8 eV which corresponds to Si–O.³⁹ For the P_2p spectrum of FRWPU-12.0, the only binding energy centered at 134.6 eV is attributed to –P(=O)–O–C– and PO₃⁻ groups in the phosphorus-rich crosslinks, while a new band at 135.5 eV for FRWPU-12.6 may be assigned to the formation of stable –P(=O)–O–Si– linkages,²⁹ which coincides with FTIR analysis of solid residue. Moreover, the Si_2p spectra of FRWPU-12.6 char shown in Fig. 10(d₃) is fitted to two peaks at 103.6 eV and 104.3 eV, attributed to Si–O–Si and –P(=O)–O–Si– structures, respectively.²²

Fig. 10 High-resolution XPS spectra of C_1s (a₁–a₃), O_1s (b₁–b₃), P_2p (c₂ and c₃) and Si_2p (d₃) of the residues for WPU (1), FRWPU-12.0 (2) and FRWPU-12.6 (3) after cone test.

Table 5 XPS data of the exterior residues of WPU, FRWPU-12.0 and FRWPU-12.6 after cone calorimeter test

Sample	C1s (%)	O1s (%)	N1s (%)	P2p (%)	Si2p (%)
WPU	78.57	13.68	7.74	—	—
FRWPU-12.0	61.27	24.99	2.53	11.21	—
FRWPU-12.6	53.01	26.02	1.86	13.23	5.88

---

Table 6 Results of C_1s XPS of exterior residual char for WPU, FRWPU-12.0 and FRWPU-12.6

Sample	C–C	C–O	C=O	Cox/Ca^a
a Cox: oxidized carbons (C–O and C=O), Ca: aliphatic and aromatic carbons (C–C).
WPU	0.619	0.229	0.152	0.615
FRWPU-12.0	0.717	0.232	0.051	0.395
FRWPU-12.6	0.674	0.226	0.100	0.484

---

Raman spectroscopy was utilized to evaluate the graphitic structure of char residue after cone calorimeter tests, and the spectra are presented in Fig. 11. All the Raman spectra of WPU, FRWPU-12.0 and FWPU-12.6 exhibit two broad and strongly overlapping characteristic peaks with maximum intensity at approximately 1600 cm⁻¹ (G band) and 1380 cm⁻¹ (D band), which correspond to an E_2g mode of hexagonal graphite and the vibration of disordered graphite or glassy carbons, respectively.^36,40 Generally, the relative integrated intensities of D and G bands (I_D/I_G) reflects the graphitization degree of carbonaceous char, and lower I_D/I_G value implies better char structure.²⁴ As shown in Fig. 11, the I_D/I_G ratio follows the sequence of WPU (3.63) > FRWPU-12.0 (3.41) > FRWPU-12.6 (2.61), illustrating that the conjugation of PNMPD and KH-602 promote the formation of thermally stable graphitic structure during burning. The lowest I_D/I_G ratio of FRWPU-12.6 implies that there is synergistic effect between P and Si atoms on char formation, corroborating with XPS analysis. Furthermore, it has been well-established that higher graphitization degree contributes to the enhanced strength of the char layer as well as a better barrier effect,^6,41 thus improving the fire resistance of FRWPU, which is in good consistence with the LOI and cone results.

Fig. 11 Raman spectra of the burning residue of WPU (a), FRWPU-12.0 (b) and FRWPU-12.6 (c) after cone calorimeter test.

Based on the above considerations, we conclude that the flame retardant action of FRWPU is condensed phase-dominant through charring catalyzed by the interaction between PNMPD and KH-602, and a possible mechanism is proposed and schematically illustrated in Scheme 2. Upon decomposition, the simultaneous presence of phosphorus and silicon elements enhances the acid catalytic charring to produce a compact, intumescent and thermally stable graphitic char layer containing –P(=O)–O–Si– structures, which effectively isolates the heat and mass transmission, provides a shielding effect for the underlying polyurethane from further burning, thereby resulting in the improved flame retardancy and smoke suppression.

Scheme 2 Possible flame-retardant mechanism of FRWPU.

3.5 Mechanical property of FRWPU films

As a non-neglectable performance index for coating material, the mechanical properties directly determine the application areas of FRWPU. Here the typical strain–stress curves of FRWPU films are displayed in Fig. 12. On the whole, the elongation at break of FRWPU tends to increase with concentrated PNMPD content, whereas the tensile strength decreases monotonically. Indeed, the bulky cyclothiophosphoramidate pendant group derived from PNMPD can encumber the crystallization of hard segment and the interactions of molecular, meanwhile, the phosphorus-containing chain extender has internal plasticization effect on polyurethane,⁴² thereby facilitating the mobility of the FRWPU chains. Undoubtedly, the adverse decrease in mechanical strength will significantly limit the application of FRWPU in many demanding fields. Considering the conjugation of KH-602, the ultimate tensile strength substantially improves without sacrificing extensibility. For instance, the tensile strength distinctly increases from 8.92 MPa for FRWPU-12.0 to 15.01 MPa for FRWPU-12.6, whilst the broken elongation slightly decreases from 392.6% to 333.7%. As discussed above, the formation of reticulated –Si–O–Si– cross-linked network between polyurethane chains by polycondensation reaction between silanol groups, which then restrains the motion of macromolecular chains, is primarily responsible for the offset of plasticizing effect caused by PNMPD and the enhancement in mechanical properties. In addition, the “excrescent” amide linkages, forming by isocyanate terminals and amino groups in KH-602, can also reinforce the strength of polyurethane.⁴³

Fig. 12 Stress–strain curves for WPU and FRWPU films.

4. Conclusions

In this work, a phosphorus, nitrogen and silicon containing flame retardant waterborne polyurethane (FRWPU) was prepared by simultaneously conjugating with PNMPD and silane coupling agent KH-602 in the chain-extension and post-chain extension process, respectively. Owing to the formation of reticulated di-substituted Si–O–Si cross-linked networks in polyurethane chains contributed by KH-602, the FRWPU exerts better performance in thermal stability and mechanical property compared to that simply incorporating with PNMPD, which is revealed by the increase of T_5% and tensile strength (enhanced by 6.1 MPa), respectively. Meanwhile, the fire resistance and smoke suppression are remarkably ameliorated. For example, the LOI value of FRWPU-12.6 reaches up to 27.7% and a UL-94 V-0 rating is achieved, while cone calorimeter results exhibit a significant reduction of PHRR, THR, PSPR and TSR by 36.0%, 42.9%, 40.1% and 35.4%, respectively, in comparison with WPU, the decrease amplitude of which is greater than FRWPU-12.0 only loading with 12 wt% PNMPD. Furthermore, thermal degradation behaviour and residue char analysis demonstrate that the excellent flame retardancy of FRWPU is attributed to the synergistic effect of PNMPD and KH-602 in condensed phase, which produces a more compact, intumescent and thermal stable graphitic structured char with –P(=O)–O–Si– linkages, acting as “tortuosity effect” to delay the escape of volatile degradation products and heat transfer. The approach described herein offers a promising strategy for addressing the current bottleneck problem of most existing phosphorous-based flame retardants that enhance the flame retardancy but at the expense of thermal and mechanical properties of waterborne polyurethane.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (Grant No. 21576172, 51273128, and 21206096).

Notes and references

1. M. S. Kathalewar, P. B. Joshi, A. S. Sabnis and V. C. Malshe, RSC Adv., 2013, 3, 4110–4129.
2. A. Lopez, E. Degrandi-Contraires, E. Canetta, C. Creton, J. L. Keddie and J. M. Asua, Langmuir, 2011, 27, 3878–3888.
3. H. Zhu, Z. Peng, Y. Chen, G. Li, L. Wang, Y. Tang, R. Pang, Z. U. H. Khan and P. Wan, RSC Adv., 2014, 4, 55271–55279.
4. M. Neisius, S. Liang, H. Mispreuve and S. Gaan, Ind. Eng. Chem. Res., 2013, 52, 9752–9762.
5. H. Li, N. Ning, L. Zhang, Y. Wang, W. Liang, M. Tian and T. W. Chan, Polymer, 2015, 67, 1–11.
6. Q. Tai, Y. Hu, R. K. Yuen, L. Song and H. Lu, J. Mater. Chem., 2011, 21, 6621–6627.
7. M.-J. Chen, Z.-B. Shao, X.-L. Wang, L. Chen and Y.-Z. Wang, Ind. Eng. Chem. Res., 2012, 51, 9769–9776.
8. P. Zhang, Y. He, S. Tian, H. Fan, Y. Chen and J. Yan, Polym. Compos., 2015 DOI:10.1002/pc.23603.
9. L. Gu and Y. Luo, Ind. Eng. Chem. Res., 2015, 54, 2431–2438.
10. G. Wu, J. Li and Y. Luo, Polym. Degrad. Stab., 2016, 123, 36–46.
11. M. Sato, S. Endo, Y. Araki, G. Matsuoka, S. Gyobu and H. Takeuchi, J. Appl. Polym. Sci., 2000, 78, 1134–1138.
12. P. Zhang, S. Tian, H. Fan, Y. Chen and J. Yan, Prog. Org. Coat., 2015, 89, 170–180.
13. P. Zhang, Z. Zhang, H. Fan, S. Tian, Y. Chen and J. Yan, RSC Adv., 2016, 6, 56610–56622.
14. W. K. Huang, J. T. Yeh, K. J. Chen and K. N. Chen, J. Appl. Polym. Sci., 2001, 79, 662–673.
15. M. M. Velencoso, M. J. Ramos, R. Klein, A. De Lucas and J. F. Rodriguez, Polym. Degrad. Stab., 2014, 101, 40–51.
16. X. Zhao, F. R. Guerrero, J. Llorca and D.-Y. Wang, ACS Sustainable Chem. Eng., 2015, 4, 202–209.
17. Y.-S. Kwak, S.-W. Park and H.-D. Kim, Colloid Polym. Sci., 2003, 281, 957–963.
18. L. Lei, L. Zhong, X. Lin, Y. Li and Z. Xia, Chem. Eng. J., 2014, 253, 518–525.
19. C. Li, J. Wan, E. N. Kalali, H. Fan and D.-Y. Wang, J. Mater. Chem. A, 2015, 3, 3471–3479.
20. H. Breulet and T. Steenhuizen, Polym. Degrad. Stab., 2005, 88, 150–158.
21. Y. Tan, Z.-B. Shao, X.-F. Chen, J.-W. Long, L. Chen and Y.-Z. Wang, ACS Appl. Mater. Interfaces, 2015, 7, 17919–17928.
22. W. Zhang, X. Li and R. Yang, Polym. Degrad. Stab., 2014, 99, 298–303.
23. D. Sun, X. Miao, K. Zhang, H. Kim and Y. Yuan, J. Colloid Interface Sci., 2011, 361, 483–490.
24. H.-B. Zhao, B.-W. Liu, X.-L. Wang, L. Chen, X.-L. Wang and Y.-Z. Wang, Polymer, 2014, 55, 2394–2403.
25. D. Chattopadhyay and D. C. Webster, Prog. Polym. Sci., 2009, 34, 1068–1133.
26. T. Chang, Y. Chiu, H. Chen and S. Ho, Polym. Degrad. Stab., 1995, 47, 375–381.
27. M. Le Bras, S. Bourbigot and B. Revel, J. Mater. Sci., 1999, 34, 5777–5782.
28. M. Bugajny, S. Bourbigot, M. Le Bras and R. Delobel, Polym. Int., 1999, 48, 264–270.
29. W. Zhang, X. Li, H. Fan and R. Yang, Polym. Degrad. Stab., 2012, 97, 2241–2248.
30. E. Wawrzyn, B. Schartel, H. Seefeldt, A. Karrasch and C. Jäger, Ind. Eng. Chem. Res., 2012, 51, 1244–1255.
31. X. Chen, L. Huo, C. Jiao and S. Li, J. Anal. Appl. Pyrolysis, 2013, 100, 186–191.
32. W. Wang, Y. Pan, H. Pan, W. Yang, K. Liew, L. Song and Y. Hu, Compos. Sci. Technol., 2016, 123, 212–221.
33. M.-J. Chen, Y.-J. Xu, W.-H. Rao, J.-Q. Huang, X.-L. Wang, L. Chen and Y.-Z. Wang, Ind. Eng. Chem. Res., 2014, 53, 8773–8783.
34. X. Wang, L. Song, H. Yang, W. Xing, B. Kandola and Y. Hu, J. Mater. Chem., 2012, 22, 22037–22043.
35. R.-M. Li, C. Deng, C.-L. Deng, L.-P. Dong, H.-W. Di and Y.-Z. Wang, RSC Adv., 2015, 5, 16328–16339.
36. X. Mu, B. Yuan, W. Hu, S. Qiu, L. Song and Y. Hu, RSC Adv., 2015, 5, 76068–76078.
37. W. Xu, S. Wang, A. Li and X. Wang, RSC Adv., 2016, 6, 48189–48198.
38. W. Zhang, X. Li and R. Yang, Polym. Degrad. Stab., 2011, 96, 1821–1832.
39. L. Li, L. Li, B. Wang, L. Liu, Y. Wu, T. Van Ree and K. Thavhiwa, Electrochim. Acta, 2011, 56, 4858–4864.
40. A. Fina, H. C. Abbenhuis, D. Tabuani and G. Camino, Polym. Degrad. Stab., 2006, 91, 2275–2281.
41. R. E. Lyon, Polym. Degrad. Stab., 1998, 61, 201–210.
42. H. Ma, J. Wang and Z. Fang, Polym. Degrad. Stab., 2012, 97, 1596–1605.
43. B. C. Chun, T. K. Cho and Y.-C. Chung, Eur. Polym. J., 2006, 42, 3367–3373.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra15869b

---