Synthesis of a highly efficient phosphorus-containing flame retardant utilizing plant-derived diphenolic acids and its application in polylactic acid

Abstract

Organophosphorus compounds as one of the most preferred flame retardants are traditionally derived from petroleum resources. Of recent interest is the exploration of renewable candidates as the precursors to synthesise these compounds. Here, we report the synthesis of a novel bio-based polyphosphonate (BPPT) through the reaction of plant-derived diphenolic acid (DPA), caged bicyclic phosphorus (PEPA) and phenylphosphonic dichloride (PPDC) in two steps. The BPPT was used as a flame retardant to prepare polylactic acid (PLA) blends. Significant enhancement in the limiting oxygen index of PLA (28.8, 33.7 and 35.4 with 2 wt%, 4 wt%, and 6 wt% of BPPT, respectively, in the PLA blends) was obtained. Additionally, the results in the UL94 flammability tests demonstrates that BPPT is effective to quench the flame, which facilitate PLA blends to achieve a V0 level at a BPPT loading of 4 wt%. The data of the cone tests and thermal gravimetric analysis confirm that a gas phase mechanism is mainly responsible for the highly efficient flame retardancy. Moreover, the tensile test results indicate that the BPPT-induced depression of the mechanical properties of the flame retardant blends is minimal. The flame retardant PLA blends developed herein contain a bio-mass content of more than 95 wt%.

---

1. Introduction

Biodegradable and bio-based polymers have become of immense interest recently due to the growing concerns over sustainable development issues.^1–3 Examples of bio-based materials can be obtained from pectin,⁴ cardbisphenol,⁵ cellulose,⁶ chitosan,⁷ starch,⁸ dextran,⁹ etc. Of note, polylactic acid (PLA), a biodegradable aliphatic thermoplastic polyester derived from starches and sugars, is deemed as one of the most promising polymer materials for wide-range applications due to its multiple benefits, such as high degree of transparency, comparable processing and mechanical properties.^10–12 Additionally, its expected market would be rapidly extended to the transportation, aircraft, electrical and electronic equipment (E&E) fields, in which non-flammability properties are required.¹³ However, similar to its petroleum-derived counterparts, the inflammable deficiency of PLA limits its application. Therefore, the work rendering PLA with flame retardancy is of great significance.

PLA is a type of thermally degradable material that burns at a relatively rapid heat release rate (HRR), negligible chars, and is unable to meet UL-94 flame retardancy standards.¹⁴ It had been reported that flame retardancy can be imparted to PLA in various ways.^15–17 Blending PLA with flame-retardant additives during polymer processing constitutes the majority of flame-retardant PLA formulations, including PLA compounded with flame retardants, such as phosphorus-containing additives,^18,19 carbon fillers,²⁰ silicon and metal-containing flame retardants.^21–23 Given the fact of fossil resource depletion and thirst for the application of environmentally friendly flame retardants, the use of renewable materials to replace conventional petroleum-based materials is becoming popular.²⁴ Acting as renewable flame retardant, the bio-based materials such as cyclodextrin,²⁵ starch,²⁶ cellulose,²⁷ lignin,²⁸ and ramie²⁹ were incorporated into PLA formulations via compounding with other flame retardant additives. However, high loadings (up to 20–30 wt%) of these additives mentioned above are required to achieve a satisfying level of flame retardancy, which may seriously deteriorate other properties of the PLA blends. Thus, developing high efficient green flame retardants to substitute petroleum-derived feedstocks is both challenging and appealing.

Diphenoic acid (DPA) is an aromatic diol with a pendant carboxyl group, which has been deemed as a potential platform chemical derived from biomass.^30–32 The versatile molecule has been heavily studied to prepare various polymer, such as polycarbonates,³³ polybenzoxazines,³⁴ polyurethanes,³⁵ and polyesters.³⁶ The presence of the reactive hydroxyl groups and the carboxyl group allows DPA to be tailored with a variety of functional moieties into its structure. In our previous study,³⁷ DPA was used directly to synthesize a phosphorus-containing flame retardant (FR). However, the FR exhibits poor flame retardancy for ABS resin. The inferior performance may attribute to the low compatibility between the FR and ABS or its insufficient phosphorus content. Moreover, the other deficiencies of DPA, such as large polarity, low thermal stability, and moisture sensitive caused by the carboxyl group have not been noticed in the published references till now.

According to the consideration mentioned above, a novel bio-based polyphosphonate (BPPT) was synthesized in two steps using DPA, caged bicyclic phosphorus (PEPA) and phenylphosphonic dichloride (PPDC) as raw materials in this study. In the first step, the intermediate, DPA-based monomer (DPM) was synthesized through the esterification of DPA and PEPA, which can help to enhance the thermal stability, hydrophobicity and compatibility of the target flame retardant. Secondly, BPPT containing phosphorus in both pendant groups and the main chain was synthesized by the polycondensation between DPM and PPDC. Then BPPT was used a flame retardant to prepare PLA blends. The PLA/BPPT blends developed herein achieve UL-94 V0 rating and LOI value of 33.3 at the BPPT loading content of 4 wt%, which is much lower than those reported in literature. In addition, the mechanical properties of PLA and its flame retardant blends were investigated by tensile test.

2. Experimental

2.1. Materials

Ingeo 3052D polylactic acid (PLA) was obtained from Nature Works (USA). Diphenolic acid (DPA) was supplied by J&K Scientific Co., Ltd. (China). Phenylphosphonic dichloride (PPDC) was purchased from Acros Organics. 1-Oxo-2,6,7-trioxa-1-phosphabicyclo[2,2,2]octane-4-methanol (PEPA) was synthesized in our laboratory according to reference literature,³⁸ yield 70%. FTIR spectra (KBr, cm⁻¹): 3390 (–OH), 2960, 2840 (C–H), 1299 (P=O), 1163 (C(CH₂)₄), 1019, 988 (P–O–C), and 870 (skeleton vibration of caged bicycle phosphate). ¹H NMR (500 MHz DMSO-d6, ppm): δ 5.07 (s, 1H), 4.58 (t, 6H), 3.28 (s, 2H). ¹³C NMR (126 MHz DMSO-d6, ppm): δ 77.09, 58.40, 39.38. ¹H NMR and ¹³C NMR spectra are included in the ESI.† Other materials were purchased from Sinopharm Chemical Reagent Co., Ltd (China) and used as received except acetonitrile, which was distilled over calcium hydride prior to use.

2.2. Synthesis of DPA-based monomer (DPM)

DPA-based monomer (DPM) was synthesized via the esterification reaction between DPA and PEPA, as shown in Scheme 1(a). A mixture of DPA (0.1 mol, 28.6 g), PEPA (0.1 mol 18.0 g) and p-toluenesulfonic acid (0.01 mol) in acetonitrile (250 mL) was introduced into a 500 ml three-necked round bottom flask equipped with a nitrogen inlet, a mechanical stirrer and a reflux condenser device. The mixture was heated to reflux with a moderate flow of nitrogen gas for 24 h. Then the solvent of the resultant was removed by rotary evaporation at reduced pressure. The crude product was purified by triple boiling with deionized water 500 mL for 30 min. The precipitate was filtrated and dried in vacuum at 80 °C for 8 h to afford light pink powder (yield: 25.6 g, 55%); mp: 291 °C (by DSC). ¹H NMR (500 MHz DMSO-d₆, ppm): δ 9.21 (s, 2H), 6.96 (d, J = 8.6 Hz, 4H), 6.66 (d, J = 8.6 Hz, 4H), 4.68 (d, J = 6.5 Hz, 6H), 3.90 (s, 2H), 2.26 (m, 2H), 2.08 (m, 2H), 1.47 (s, 3H). ¹³C NMR (126 MHz DMSO-d₆, ppm): δ 172.8, 155.1, 139.1, 127.7, 114.6, 75.5, 59.5, 43.7, 36.9, 35.9, 29.4, 27.2. ³¹P NMR (202 MHz DMSO-d₆, ppm): δ −7.56, ³¹P NMR spectrum is included in the ESI.†

Scheme 1 Synthesis route of BPPT.

2.3. Synthesis of bio-based polyphosphonate (BPPT)

BPPT was synthesized through the polycondensation between DPM and PPDC, as shown in Scheme 1(b). A mixture of DPM (22.4 g, 0.05 mol) and acetonitrile (120 mL) was added into a 250 mL flame-dried round bottom flask equipped with a nitrogen inlet, reflux condenser device and mechanical stirrer. The reactants were heated to 40 °C under the protection of N₂, which was fitted with a vent to a saturated sodium hydroxide solution to trap the hydrochloric acid byproduct. Then PPDC (9.75 g, 0.05 mol) in acetonitrile (20 mL) was slowly added into the flask in one batch. The reactants were heated to 80 °C within 10 min and maintained for 5 h. Giving a brown solid crude product at the bottom of the flask. The crude product was twice washed by hot deionized water (500 mL) and ethylacetate (100 mL), respectively, and then dried in vacuum at 60 °C for 8 h to reveal a purple product, (yield: 27.3 g, 85%); mp: 35 °C (by DSC). FTIR spectra (KBr, cm⁻¹): 1740 (–COO–), 1298 (P=O), 1019, 988 (P–O–C), and 870 (skeleton vibration of caged bicycle phosphate). ¹H NMR (500 MHz, DMSO-d₆, ppm): δ 7.75 (m, 2H), 7.56 (m, 2H), 7.46 (m, 1H), 7.01 (m, 2H), 6.72 (m, 2H), 4.66 (dd, 6H), 3.94 (d, 2H), 2.31 (m, 2H), 2.13 (m, 2H), 1.52 (s, 3H). ¹³C NMR (126 MHz, DMSO-d₆, ppm) δ: 172.68, 155.03, 139.1, 134.8, 133.7, 131.20 (d, J_P–C = 9.9 Hz, aromatic carbon atom ortho to phosphorus), 130.82 (d, J_P–C = 2.8 Hz, aromatic carbon atom para to phosphorus), 130.50 (d, J_P–C = 9.8 Hz, aromatic carbon atom meta to phosphorus), 128.04 (d, J_P–C = 14.1 Hz, aromatic carbon atom attached to phosphorus), 127.7, 131.20, 119.7, 114.86, 76.83, 59.48, 43.75, 37.14, 36.85, 29.45, 27.26, ¹³C NMR spectrum is shown in the ESI;† ³¹P NMR (202 MHz, DMSO-d₆, ppm) δ: 12.76, 11.87, −7.54; GPC (in DMF): the number-average molecular (M_n) = 1874, and the polydispersity (PDI) = 1.14.

2.4. Preparation of PLA/BPPT blends

PLA was dried in vacuum at 60 °C for 12 h prior to use. All PLA/BPPT blends were prepared in a melt mixer (ThermoHaake rheomixer, Polylab, Germany) at 170 °C with a rotor speed of 60 rpm for 8 min. PLA blends with 2 wt%, 4 wt% and 6 wt% BPPT are coded as PLA2, PLA4 and PLA6, respectively. After mixing, the blends were transferred into a mold and preheated for 5 min at 170 °C, and then hot-pressed at 20 MPa for 2 min, followed by cold-press at room temperature for 5 min to form specimens for all tests.

2.5. Characterization

¹H NMR, ¹³C NMR and ³¹P NMR spectra were recorded on a Varian Unity Inova spectrometer (500 MHz, Advance2B, Bruker, Germany) using DMSO-d₆ as the solvent. FT-IR spectra were recorded on a Vector-22 FT-IR spectrometer (Bruker, Germany) using KBr pellets. M_n, and PDI of BPPT were determined by gel permeation chromatography (GPC) in a PL 220 GPC instrument (Waters, USA) at 25 °C in DMF. The concentration of the sample solution for GPC measurement was 5 wt‰. Three PL gel 10 μm mixed-B columns were used and polymethylmethacrylate was used as the standard. Thermal gravimetric analysis (TGA) was performed on TGA 209 F1 (Netzsch, Germany) at a heating rate of 20 °C min⁻¹ in N₂ from 50 to 600 °C. Each specimen with about 6 mg was examined in triplicate. Melting points were determined by differential scanning calorimetry (DSC) using a DSC 200 PC (Netzsch, Germany). All samples were analyzed at a heating rate of 5 °C min⁻¹ in N₂. Thermogravimetric analysis coupled to Fourier transform infrared spectroscopy (TGA-FTIR) measurements were performed with a TGA 209 F1 instrument (Netzsch, Germany), coupled with a Thermo Nicolet iS10 FTIR spectroscopy (Thermosher, Germany). The volatiles evolved from TG can be transferred into the gas cell of FTIR through the transfer line by a suitable gas flow. The spectra were obtained with a scan interval of 2.23 s. The resolution of the spectra was 4 cm⁻¹. Limiting oxygen index (LOI) values were determined using an HC-2 Oxygen Index Instrument (Jiangning Analysis Instrument Company, China) according to ASTM D2863, with the dimension of the sample sheets of 150 × 6 × 3 mm³. The vertical burning test was performed on a CZF-III Horizontal and Vertical Burning Tester (Jiangning Analysis Instrument Company, China) according to UL94 test (ASTM D3801 standard), with the dimension of the sample sheets of 127 × 12.7 × 3 mm³. Cone calorimetry was carried out on a FTT Cone Calorimeter with the dimension of the sample sheets of 100 × 100 × 3 mm³ at a heat flux of 35 kW m⁻² according to ISO 5660 standard. The tensile test was performed using an electronic universal testing machine (Model RTW10, Shenzhen Reger Instrument Co. Ltd., China). Dumbbell specimens with a 25 × 4 × 2 mm³ neck were used. Tests were performed at room temperature with a constant crosshead speed of 2 mm min⁻¹. At least eight samples were tested to obtain the average values of tensile properties for all blends.

3. Results and discussion

3.1. Characterization of DPM

Among the organophosphorus flame retardants, caged bicyclic phosphate (PEPA) have been widely studied as excellent flame retardants.^39–41 According to the advantage of this structure, the intermediate DPM with a pendant caged bicyclic phosphate was designed and prepared using DPA and PEPA as raw materials. The esterification between the carboxyl of DPA and the hydroxyl of PEPA catalyzed by p-toluenesulfonic acid can avoid the deficiency such as the poor thermal stability and the hydrophilicity caused by carboxyl and hydroxyls of virgin DPA and PEPA. The FT-IR spectra of DPA and DPM are shown in Fig. 1. In Fig. 1, the absorption peak at 1708 cm⁻¹, which is attributed to the carboxyl group in DPA, disappears in DPM. Moreover, a new peak appeared at 1740 cm⁻¹ corresponding to the ester group in DPM appears, indicating the successful reaction between DPA and PEPA.

Fig. 1 FT-IR spectra of DPA, DPM and BPPT.

The ¹H NMR and ¹³C NMR spectra of DPM are shown in Fig. 2(a) and (b), respectively. ¹H NMR spectrum reveals that no peaks appeared at 13.93 ppm (carboxyl groups in DPA) and 5.07 ppm (hydroxyl in PEPA), which indicates that the ester bonds between DPA and PEPA were successfully formed in the product of DPM. The peaks at 2.26 ppm (H^f), 2.08 ppm (H^e) and 1.47 ppm (H^d) could be assigned to the chemical shifts of protons on methylene and methyl groups in DPA subunit, respectively. The methylene protons in PEPA subunit appeared at δ 4.68 ppm (H^h) and 3.90 ppm (H^g). It is noticed that proton signals in PEPA subunit were shifted to higher frequencies compared to the corresponding protons in pure PEPA due to deshielding effect caused by the ester group. The peaks at 6.96 ppm (H^c) and 6.66 ppm (H^b) were corresponded to the hydrogen atoms in the phenyl group. The peak at 9.21 ppm (H^a) was due to the hydrogen atoms on the hydroxyl group attached to the phenyl group.

Fig. 2 (a) ¹H NMR and (b) ¹³C NMR spectra of DPM in DMSO-d6.

The ¹³C NMR spectrum of DPM is shown in Fig. 2(b). The peak assigned to the carbon of ester carbonyl was observed at 172.8 ppm (C⁹). The peaks appeared at 155.1 (C¹), 139.1 (C⁴), 127.7 (C³), 114.6 (C²), 43.7 (C⁸), 36.9 (C⁶), 29.4 (C⁷), 27.2 (C⁵) ppm were attributed to the carbon atoms in the DPA subunit. The peaks at 75.5 (C¹²), 59.5 (C¹⁰) and 35.9 (C¹¹) ppm were assigned to the carbon atoms in caged PEPA subunit. The carbon in the caged structure attached to the ester group appeared at 59.5 ppm, which was higher than the corresponding carbon peak appeared at 58.4 ppm in pure PEPA. This is also caused by the deshielding effect of the ester group.

In ³¹P NMR spectrum (shown in the ESI†), the peak at −7.6 ppm was assigned to the phosphorous atom in DPM. The above spectral data was in good agreement with the proposed structure of DPM.

3.2. Characterization of BPPT

Scheme 1(b) shows the synthesis route of BPPT through a direct condensation polymerization of DPM and PPDC. The polycondensation reaction began with the immediate yield of hydrogen chloride, which proceeded for 4 h until no visible gas product evaporated. Then, the brown crude product was twice washed with hot deionized water and ethylacetate, respectively, to remove the residual hydrogen chloride and the unreacted reactants. The product BPPT was soluble in tetrahydrofuran (THF), N,N-dimethyl formamide (DMF) and dimethyl sulfoxide (DMSO). The oligomeric nature of the polyphosphonate made from DPA-based biphenol by condensation polymerization may stems from the low reactivity of the hydroxyl in the large DPM or the side reactions such as chlorination and alkylation of hydroxyl groups or reaction with residual water.⁴²

Fig. 3(a) and (b) shows the ¹H NMR and ³¹P NMR spectra of BPPT, respectively. In Fig. 3(a), the absorption of the aromatic protons of the PPDC subunit (Hⁱ, H^k, H^j) appeared at 7.75, 7.56 and 7.46 ppm, respectively. The peaks appeared at 7.01 and 6.72 ppm are broad multiplet owing to some heteronuclei coupling with phosphorus, which corresponds to the aromatic protons (H^c, H^b) in DPA subunit. Seen from the ³¹P-NMR spectrum of BPPT in Fig. 3(b), the chemical shifts for an aromatic polyphosphonate were observed as expected. The peaks at 11.6 and 12.7 ppm could be assigned to the chemical shifts of phosphorus attached to the benzene ring in the terminal group and the repeated unit of the oligomer. The peaks at −7.52 ppm could be assigned to the chemical shifts of phosphorus in caged structure. According to the data of the ³¹P NMR spectrum, the intensity ratio of P (connected to the benzene) to P (in caged structure) is close to 1 : 1, which is in accordance to the composition of oligomer.

Fig. 3 (a) ¹H NMR and (b) ³¹P NMR spectra of BPPT in DMSO-d₆.

The ¹³C NMR spectrum (shown in the in the ESI†) of BPPT exhibited 8 peaks in the range of 156–110 ppm corresponding to aromatic carbon atoms. ¹³C-³¹P splitting is particularly noticeable, especially the carbons in the PPDC subunit. Clear double absorptions appeared at 131.20, 130.82, 130.50, 128.04 ppm, because of the J_C–P couplings in the phenyl structure of the PPDC subunit. The peaks at 155.03, 139.1, 127.7, 114.86 ppm were assigned to the chemical shifts of aromatic carbons of DPM unit. The ¹³C NMR spectra showed several additional, less well-marked peaks in the range for aromatic carbons (i.e., 134.8, 133.7, 129.55 and 119.74 ppm), which may be assigned to the aromatic carbon atoms in the terminal groups.

3.3. UL-94 and LOI tests

Vertical burning test (UL-94) and limiting oxygen index (LOI) are used to evaluate the flammability of the PLA/BPPT blends. The results are shown in Table 1. LOI is widely used to evaluate the flame retardant properties of materials. The higher LOI value is, the more difficult it is for combustion to occur. Pure PLA is inflammable with a LOI value of 20.0. With the incorporation of the flame retardant BPPT into PLA, the LOI value increased remarkably. Seen from Table 1, the PLA/BPPT blends achieved the LOI value of 28.3, 33.7, 35.4 at the BPPT loading of 2 wt%, 4 wt%, 6 wt%, respectively.

Table 1 LOI values and UL94 rating for pure PLA and its blends

Sample	LOI	UL94 rating	Dripping	AFT^a (s)	Igniting cotton
a AFT: average flaming time after the first and the second ignition.
Pure PLA	20.0	Failed	Yes	Burn out	Yes
PLA2	28.8	V2	Yes	3.1/0.3	Yes
PLA4	33.7	V0	Yes	0.2/0.1	No
PLA6	35.4	V0	Yes	0.1/0.1	No

---

UL-94 is the most widely used standard for evaluating the plastics flammable performance. It is used to evaluate the self-extinguishing ability of a material after it is ignited.^43,44 The UL-94 V0 rating represents the highest UL-94 level, which is desperately required in the flame retardant industry. The pure PLA burns completely to the clamp in UL-94 test and receives no rating. After the loading of 4 wt% or 6 wt% BPPT into PLA, flames extinguish instantly after removing of the ignitor and cannot ignite the underlying cotton, both of which are rated as UL94 V0 level. The PLA2 sample could be self-extinguished as well after removing of the ignitor within 4 s, but the droplet can ignite the underlying cotton. Consequently, V2 level is assigned for PLA2 (videos of PLA2, PLA4, PLA6 are shown in the ESI†). The LOI and UL-94 tests indicate that PLA/BPPT blends manifested excellent flame retardant behaviors.

3.4. Cone calorimetric test

Cone calorimetry is used for investigating the dynamic combustion behavior of polymeric materials in a forced scenario, which can provide various parameters such as time to ignition (TTI), peak heat release rate (PHRR) and total heat release (THR).⁴⁵ Fig. 4 presents the curves of heat release rate (HRR) versus time of the PLA and PLA/BPPT blends. Detailed information from the cone calorimetric tests is listed in Table 2. For pure PLA, its PHRR value is 418 kW m⁻², THR is 70 MJ m⁻² and TTI is 68 s. The addition of BPPT into PLA induces a slight reduction in PHRR and THR of PLA. Loading 6 wt% of BPPT, the PHRR and THR values decrease to 388 kW m⁻² and 66 MJ m⁻², respectively.

Fig. 4 Heat release rate (HRR) curves of pure PLA and its blends at 35 kW m⁻².

Table 2 Cone calorimetric test data of PLA and its blends

Sample	TTI (s)	PHRR (kW m^−2)	THR (MJ m^−2)	Residual mass (wt%)	Mean CO (kg kg^−1)	Mean CO2 (kg kg^−1)
PLA	68 ± 1	418 ± 3	70 ± 1	1.6 ± 0.2	0.0073 ± 0.0006	1.69 ± 0.02
PLA2	68 ± 1	394 ± 14	69 ± 1	2.5 ± 0.1	0.0087 ± 0.0005	1.68 ± 0.01
PLA4	78 ± 2	396 ± 6	67 ± 2	3.1 ± 0.1	0.011 ± 0.002	1.66 ± 0.02
PLA6	74 ± 1	388 ± 6	66 ± 1	3.5 ± 0.2	0.016 ± 0.001	1.61 ± 0.01

---

It is no surprise, then, that the rather disappointing inconsistency between the mediocre performance measured by cone calorimeter and the excellent behaviour measured by LOI and UL94 tests for the PLA blends. Since the fire scenarios that these tests performed are different, there is rough correlation between these methods, especially for polymers containing flame retardants.⁴⁵ Nevertheless, the results obtained from the cone calorimetry are able to yield insights into the flame retardant mechanism. As shown in Fig. 5, all the flame retardant PLA blends prepared herein burn to low char yield. Unlike most condensed-phase active flame retardants, which form abundant char on the surface of the burning polymer to inhibit the transfer of heat and combustible gas.

Fig. 5 Digital photos of char residues after cone test for PLA and its blends.

Seen from Table 2, the TTI value prolonged with the increased adding amount of BPPT in the PLA blends. The TTIs of PLA4 and PLA6 are 10 s and 8 s longer than that of the pure PLA. Generally, ignition of materials happens until the concentration of fuel from material degradation reaches a critical value. The prolonged ignition time is related to the reduction of fuel gas decomposed from PLA blends during combustion, which is an indicative of the gas-phase mechanism.^46,47 Besides, the mean CO yield of PLA2, PLA4, PLA6 in the cone calorimetric tests are increased by about 20%, 50% and 120% relative to the pure PLA. However, the difference in the mean CO₂ yield is hardly visible for PLA and flame retardant PLA blends. As it is difficult to directly observe the gaseous product evolved during the cone test, such increases in the mean CO yields indicate that the BPPT facilitate the incomplete combustion, which is in contrary to condense-phase mechanism that tends to favor complete combustion. Thus, it is most likely that BPPT endows PLA with flame retardancy through a gas-phase mechanism.

3.5. Thermal decomposition analysis

Given the fact that BPPT endowed PLA with strong self-extinguishing ability in UL 94 test and significantly increased LOI values, all these excellent performance urged us to study the flame retardant mechanism. First of all, the thermal decomposition behaviors for BPPT, pure PLA and PLA blends in N₂ were studied as shown in Fig. 6. The temperature at 5 wt% weight loss is denoted as T_onset, and the maximum decomposition temperature is assigned as T_max. Detailed TGA data are summarized in Table 3.

Fig. 6 TG and DTG curves of BPPT, pure PLA, and their blends in N₂.

Table 3 The TGA data of BPPT, pure PLA, and their blends in N₂

Sample ID	Tonset (°C)	Tmax (°C)	Residue weight at 600 °C (wt%)
			Experimental	Theoretical
BPPT	236.8	251.7, 387.7	41.40	—
PLA	343.5	379.1	1.72	—
PLA2	343.9	380.7	1.70	2.51
PLA4	338.7	379.6	2.38	3.31
PLA6	331.9	379.9	3.01	4.10

---

As for BPPT, it started to decompose at 236.8 °C and had the residue weight of 41.40 wt% at 600 °C. There are two T_maxs for BPPT appeared at 251.7 °C and 387.7 °C, respectively. The T_onset and T_max of the pure PLA appear at 343.5 °C and 379.1 °C, respectively. The residue weight is 1.72 wt% at 600 °C. Compared to the pure PLA, T_onsets of the PLA/BPPT blends decreased with the increased loading of BPPT due to the early degradation of the flame retardant. However, no visible change for T_maxs were observed, which means the addition of BPPT altered the thermal decomposition behavior of PLA in a minor way under nitrogen atmosphere.¹⁵ It is interesting that the experimental residue weight of PLA2, PLA4 and PLA6 are 1.70 wt%, 2.38 wt% and 3.01 wt%, respectively, which are lower than those of the corresponding theoretical values. As the pure BPPT begins to decompose early than neat PLA, the gap between the experimental and the theoretical values of the residue weight indicates that BPPT is a gas-phase active flame retardant.⁴⁸

Thus, TG-FTIR was employed to study the evolved gaseous product of BPPT during thermal degradation in N₂. FTIR spectra of evolved gaseous products of BPPT at different temperature were shown in Fig. 7. At 237 °C (T_onset), the C=O (1775 cm⁻¹) was observed, which indicates that the ester group in pendent structure was broken. Meanwhile, the peak assigned to the absorption of P=O (ref. 49) (1257 cm⁻¹) is visible as well. At 252 °C (T_max1), H₂O (3660 cm⁻¹), C_sp³–H (2970 cm⁻¹), C=C (1605 cm⁻¹), –C–O (1101 cm⁻¹) and P=O are observed in gaseous product of BPPT and exhibit strong absorptions, which corresponds to the main degradation step of BPPT. At 300 °C, the P–O–C (ref. 49) (1101 cm⁻¹) peak was observed. Then at 388 °C, absorptions of C_sp²–H (3050 cm⁻¹, 671 cm⁻¹) vibration is observed, which may correspond to the products evolved during the carbonation process. Therefore, the phosphorus-containing compounds which are deemed as free radical scavenger are presented in the evolved gaseous product during thermal degradation of BPPT.

Fig. 7 FTIR spectra of evolved gaseous products from thermal decomposition of BPPT at different temperatures.

What is more, the evolved gaseous products were identified using TG-FTIR to understand the thermal decomposition behavior of pure PLA and PLA6 around 380 °C (T_max) in N₂. FTIR spectra are shown in Fig. 8. According to the reported results,⁵ H₂O (3580 cm⁻¹), hydrocarbons (–CH₃ and –CH₂– groups: 3002–2850 and 1400–1200 cm⁻¹), CO₂ (2341 cm⁻¹), CO (2182 cm⁻¹), lactide or cyclic oligomers (carbonyl compounds, 1761 cm⁻¹), aldehyde (2737 cm⁻¹) are the main gaseous products of the pure PLA. Seen from Fig. 8, the main FTIR signals of PLA6 are similar to those of the pure PLA, which indicated that the main decomposition products of PLA were not altered too much by the low loading of BPPT. However, the absorption at 1258 cm⁻¹ may attribute to P=O is observed, indicating the presence of phosphorus-containing fragments in the gas pyrolysis products of PLA6.^21,49

Fig. 8 The comparison of FTIR spectra of evolved gaseous products of PLA and PLA6 around 380 °C.

3.6. Tensile properties

The mechanical properties of PLA and its flame retardant blends were studied by tensile test. Fig. 9 presents the stress–stain curves of PLA and its blends. The pure PLA is a brittle material and shows no yielding behavior during the test. When BPPT was added into PLA, the flame retardant blends exhibits brittle performance as well during the tensile test. With regards to tensile strength and strain at break, the presence of BPPT has a deleterious effect on both properties of the resulting blends, which is presumably due to the impact of BPPT on the crystallization behaviors of PLA. However, the BPPT-induced depression on the mechanical properties of the flame retardant blends is minimal enough, especially at low BPPT loadings (PLA2, PLA4) to warrant their use.

Fig. 9 Stress–strain curves of pure PLA and its flame retardant blends.

4. Conclusions

Taken advantage of the bio-based chemical compound DPA, a novel phosphorous-containing flame retardant BPPT was designed and synthesized in this study. When the BPPT loadings in PLA is only 4 wt%, the LOI value and UL94 rating can reach 33.7 and V0 level, which indicate that BPPT is a highly efficient flame retardant for PLA. Through the analysis of flammability data of the PLA/BPPT blends, it is suggested that the flame retardant mechanism of BPPT was mainly related to gas-phase mechanism. The work demonstrates a new class of highly efficient flame retardants containing sustainable resources PLA.

Acknowledgements

The authors acknowledge the financial support from the Ningbo Science and Technology Innovation Team (No. 2015B11005).

References

1. A. Mohanty, M. Misra and L. Drzal, J. Polym. Environ., 2002, 10, 19–26.
2. R. A. Gross and B. Kalra, Science, 2002, 297, 803–807.
3. K. Sudesh and T. Iwata, Clean: Soil, Air, Water, 2008, 36, 433–442.
4. H.-B. Chen, B.-S. Chiou, Y.-Z. Wang and D. A. Schiraldi, ACS Appl. Mater. Interfaces, 2012, 5, 1715–1721.
5. S. Li and S. Yan, RSC Adv., 2015, 5, 61808–61814.
6. N. Illy, M. Fache, R. Ménard, C. Negrell, S. Caillol and G. David, Polym. Chem., 2015, 6, 6257–6291.
7. G. Laufer, C. Kirkland, A. A. Cain and J. C. Grunlan, ACS Appl. Mater. Interfaces, 2012, 4, 1643–1649.
8. Y.-H. Guan, J.-Q. Huang, J.-C. Yang, Z.-B. Shao and Y.-Z. Wang, Ind. Eng. Chem. Res., 2015, 54, 3524–3531.
9. M. Barrow, A. Taylor, D. J. Nieves, L. K. Bogart, P. Mandal, C. M. Collins, L. R. Moore, J. J. Chalmers, R. Lévy and S. R. Williams, Biomater. Sci., 2015, 3, 608–616.
10. T. Tábi, P. Tamás and J. G. Kovács, eXPRESS Polym. Lett., 2013, 7, 107–119.
11. B. Gupta, N. Revagade and J. Hilborn, Prog. Polym. Sci., 2007, 32, 455–482.
12. E. T. Vink, K. R. Rabago, D. A. Glassner, B. Springs, R. P. O'Connor, J. Kolstad and P. R. Gruber, Macromol. Biosci., 2004, 4, 551–564.
13. V. K. K. V. Y. Horikoshi, Fujitsu Sci. Tech. J., 2005, 41, 173–180.
14. S. Bourbigot and G. Fontaine, Polym. Chem., 2010, 1, 1413.
15. X. Zhao, F. R. Guerrero, J. Llorca and D.-Y. Wang, ACS Sustainable Chem. Eng., 2015, 4, 202–209.
16. J. Gong, R. Niu, N. Tian, X. Chen, X. Wen, J. Liu, Z. Sun, E. Mijowska and T. Tang, Polymer, 2014, 55, 2998–3007.
17. M. Murariu, L. Bonnaud, P. Yoann, G. Fontaine, S. Bourbigot and P. Dubois, Polym. Degrad. Stab., 2010, 95, 374–381.
18. J. Zhan, L. Song, S. Nie and Y. Hu, Polym. Degrad. Stab., 2009, 94, 291–296.
19. C. Reti, M. Casetta, S. Duquesne, S. Bourbigot and R. Delobel, Polym. Adv. Technol., 2008, 19, 628–635.
20. S. Bourbigot, G. Fontaine, A. Gallos and S. Bellayer, Polym. Adv. Technol., 2011, 22, 30–37.
21. L. Song, S. Xuan, X. Wang and Y. Hu, Thermochim. Acta, 2012, 527, 1–7.
22. D. M. Fox, J. Lee, C. J. Citro and M. Novy, Polym. Degrad. Stab., 2013, 98, 590–596.
23. X. Wen, J. Gong, H. Yu, Z. Liu, D. Wan, J. Liu, Z. Jiang and T. Tang, J. Mater. Chem., 2012, 22, 19974.
24. B. Wang, X. Qian, Y. Shi, B. Yu, N. Hong, L. Song and Y. Hu, Composites, Part B, 2015, 69, 22–30.
25. J. X. Feng, S. P. Su and J. Zhu, Polym. Adv. Technol., 2011, 22, 1115–1122.
26. X. Wang, Y. Hu, L. Song, S. Xuan, W. Xing, Z. Bai and H. Lu, Ind. Eng. Chem. Res., 2010, 50, 713–720.
27. D. M. Fox, M. Novy, K. Brown, M. Zammarano, R. H. Harris, M. Murariu, E. D. McCarthy, J. E. Seppala and J. W. Gilman, Polym. Degrad. Stab., 2014, 106, 54–62.
28. R. Zhang, X. Xiao, Q. Tai, H. Huang, J. Yang and Y. Hu, High Perform. Polym., 2012, 24, 8738–8746.
29. L. Shumao, R. Jie, Y. Hua, Y. Tao and Y. Weizhong, Polym. Int., 2010, 59, 242–248.
30. Y. Liu, Y. Zhang and Z. Fang, BioResources, 2012, 7, 4914–4925.
31. C. Wang and S. Nakamura, J. Polym. Sci., Part A: Polym. Chem., 1995, 33, 2157–2163.
32. P. Zhang, L. Wu, Z. Bu and B. G. Li, J. Appl. Polym. Sci., 2008, 108, 3586–3592.
33. C. Peng, J. Wu, J. Liu, L. Qiu, X. Liu, S. Bo and Z. Zhen, Polym. Chem., 2013, 4, 2703.
34. C. Zúñiga, G. Lligadas, J. C. Ronda, M. Galià and V. Cádiz, Polymer, 2012, 53, 1617–1623.
35. V. Acharya, C. R. Prabha and C. Narayanamurthy, Biomaterials, 2004, 25, 4555–4562.
36. P. Zhang, L. B. Wu and B. G. Li, Polym. Degrad. Stab., 2009, 94, 1261–1266.
37. Y. Zhang, X. Chen and Z. Fang, J. Appl. Polym. Sci., 2013, 128, 2424–2432.
38. Y. Haloern and R. H. Niswander, US4454064A, 1984.
39. S. H. Jiang, H. Y. Yang, X. D. Qian, Y. Q. Shi, K. Q. Zhou, H. Y. Xu, X. Y. Shan, S. M. Lo, Y. Hu and Z. Gui, Ind. Eng. Chem. Res., 2014, 53, 3880–3887.
40. X. Li, Y. Ou and Y. Shi, Polym. Degrad. Stab., 2002, 77, 383–390.
41. H.-Q. Peng, Q. Zhou, D.-Y. Wang, L. Chen and Y.-Z. Wang, J. Ind. Eng. Chem., 2008, 14, 589–595.
42. W. Vogt and S. Balasubramanian, Die. Makromol. Chem., 1973, 163, 111–134.
43. S. Y. Ran, Z. H. Guo, C. Chen, L. P. Zhao and Z. P. Fang, J. Mate. Chem. A, 2014, 2, 2999–3007.
44. S. Chen, X. Li, Y. Li and J. Sun, ACS Nano, 2015, 9, 4070–4076.
45. B. Schartel and T. R. Hull, Fire Mater., 2007, 31, 327–354.
46. X. Zhao, F. R. Guerrero, J. Llorca and D.-Y. Wang, ACS Sustainable Chem. Eng., 2016, 4, 202–209.
47. D.-Y. Wang, Y.-P. Song, L. Lin, X.-L. Wang and Y.-Z. Wang, Polymer, 2011, 52, 233–238.
48. T. C. Mauldin, M. Zammarano, J. W. Gilman, J. R. Shields and D. J. Boday, Polym. Chem., 2014, 5, 5139.
49. X. Mu, B. Yuan, W. Hu, S. Qiu, L. Song and Y. Hu, RSC Adv., 2015, 5, 76068–76078.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra06742e

---