Effective Biobased Phosphorus Flame Retardants from Starch-Derived bis-2,5-(Hydroxymethyl)Furan

A series of biobased phosphorus flame retardants has been prepared by converting starch-derived bis-2,5-(hydroxymethyl)furan to the corresponding diacrylate followed by Michael addition of phosphite to generate derivatives with phosphorus moieties attached via P–C bonds. All compounds behave as effective flame retardants in DGEBA epoxy resin. The most effective is the DOPO derivative, 2,5-di[(3-dopyl-propanoyl)methyl]furan. When incorporated into a DGEBA blend at a level to provide 2% phosphorus, a material displaying a LOI of 30, an UL 94 rating of V0 and a 40% reduction in combustion peak heat release rate compared to that for resin containing no additive is obtained. The analogous compounds generated from bisphenol A and tetrabromobisphenol A exhibit similar flame-retarding properties.


Introduction
Polymeric materials have made a huge contribution to the development of modern society [1,2]. Apart from naturally-occurring polymers and modifications, these have generally been derived from petrochemicals. More recently, the generation of biobased polymers has gained interest [3]. For most applications polymeric materials must be flame retarded. Despite the popularity and effectiveness of traditional organohalogen flame retardants, in particular brominated aromatics, there has been a strong move away from the use of these materials. These compounds may be converted to volatile and toxic dioxins at high temperature [4]. However, the much larger concern arises from the migration of these compounds from a polymer matrix into which they have been incorporated. This is particularly a problem for waste items discarded in a landfill [5]. The flame retardants escape into the environment, are not biodegradable, bioaccumulate and pose risks to human health. Widespread human exposure to these materials has occurred and has been associated with a variety of disease states, most arising from endocrine disruption [6][7][8][9]. The negative impacts of human exposure to these compounds has promoted public pressure for the restriction of their use. This has occurred by both regulation and voluntary removal from the market [10][11][12]. The best candidates to replace organohalogen flame retardants are organophosphorus compounds [13,14]. In general, the toxicity of organophosphorus compounds is much lower than that observed for organohalogen counterparts [15][16][17]. Further, the effectiveness of organophosphorus flame retardants may be enhanced by the presence of compounds containing other elements, most notably sulfur [18][19][20][21], nitrogen [22][23][24][25][26][27], boron [28,29], or silicon [30][31][32]. Often the effectiveness of the presence of two flame-retarding compounds is touted as being synergistic. However, this is almost never the case. Usually, the impact of the two compounds on the suppression of flammability is not even additive.
The development of polymer additives from naturally-occurring renewable materials is increasingly important [33]. The precursors to these materials are readily available, renewable annually, environmentally-benign, usually biodegradable, and generally nontoxic. They may be obtained from a variety of sources [34]. Further, the utilization of these materials is independent of the cost or availability of petrochemicals. Most prominently, the development of new flame retardants from biosources has been a focus [35][36][37][38][39][40][41][42][43][44][45]. Plant sources have long been the origin of many commercial materials [46]. More recently there has been an emphasis on starch from seed grains. A major source of starch, particularly in the United States, has been corn [47,48]. Starch may be hydrolyzed to afford glucose which can converted to many useful materials mainly through 5hydroxymethylfurfural (HMF) [49][50][51][52]. Glucose may be reduced and then dehydrated to provide isosorbide, a diether diol, suitable for the generation of a range of phosphorus flame retardants [38][39][40]. Glucose, and other carbohydrates, may also serve as a source of bis-2,5-(hydroxymethyl)furan (BHMF). This difunctional alcohol has been used for the production of poly(ester)s [53][54][55][56] and plasticizers [57]. In this case, it was converted to the corresponding diacrylate. Michael addition of phosphite to the acrylate was then used to generate a series of bioderived phosphorus flame retardants containing P-C bonds [58][59][60].

Results and Discussion
The development of effective flame retardants from renewable biosources offers several advantages over traditional processes based on petrochemicals and is receiving increasing attention. The impetus for this development is a heightened concern for environmental quality and an increased awareness of the toxicity of traditional flame retardants. In this case, bioderived bis-2,5-(hydroxymethyl)furan has been utilized as the base for the generation of a series of phosphorus flame retardants. The diol was first converted to the diacrylate. Michael addition of phosphite to the diacrylate afforded the corresponding phosphorus derivatives [58]. This is illustrated in Scheme 1 for the preparation the DOPO derivative (DDMF). The corresponding diethyl-(DEMF) and diphenylphosphite (DPMF) adducts were prepared in an analogous manner. The structures for all compounds were rigorously established using spectroscopic and thermal methods (see Experimental section); spectra may be found in Supplemental Material). Pertinent infrared absorptions for these compounds may be found in Table 1, proton NMR chemical shifts in Table 2 and the corresponding carbon-13 chemical shifts in Table 3. The corresponding diethyl-(DEMF) and diphenylphosphite (DPMF) adducts were prepared in an analogous manner. The structures for all compounds were rigorously established using spectroscopic and thermal methods (see Experimental section); spectra may be found in Supplemental Material). Pertinent infrared absorptions for these compounds may be found in Table  1, proton NMR chemical shifts in Table 2 and the corresponding carbon-13 chemical shifts in Table 3. The corresponding diethyl-(DEMF) and diphenylphosphite (DPMF) adducts were prepared in an analogous manner. The structures for all compounds were rigorously established using spectroscopic and thermal methods (see Experimental section); spectra may be found in Supplemental Material). Pertinent infrared absorptions for these compounds may be found in Table  1, proton NMR chemical shifts in Table 2 and the corresponding carbon-13 chemical shifts in Table 3. The corresponding diethyl-(DEMF) and diphenylphosphite (DPMF) adducts were prepared in an analogous manner. The structures for all compounds were rigorously established using spectroscopic and thermal methods (see Experimental section); spectra may be found in Supplemental Material). Pertinent infrared absorptions for these compounds may be found in Table  1, proton NMR chemical shifts in Table 2 and the corresponding carbon-13 chemical shifts in Table 3. The corresponding diethyl-(DEMF) and diphenylphosphite (DPMF) adducts were prepared in an analogous manner. The structures for all compounds were rigorously established using spectroscopic and thermal methods (see Experimental section); spectra may be found in Supplemental Material). Pertinent infrared absorptions for these compounds may be found in Table  1, proton NMR chemical shifts in Table 2 and the corresponding carbon-13 chemical shifts in Table 3.                 For comparison the corresponding DOPO derivatives of bisphenol A and tetrabromobisphenol A were prepared. Structures are shown in Figure 1. For comparison the corresponding DOPO derivatives of bisphenol A and tetrabromobisphenol A were prepared. Structures are shown in Figure 1. For assessment of the effectiveness of these compounds, blends with DGEBA epoxy at levels to provide 1% or 2% phosphorus and in two cases 5% phosphorus were prepared [58]. The flammability of the blends was determined using limiting oxygen index (LOI), microscale combustion calorimetry (MCC), and the underwriter Laboratories vertical burn test (UL94). Results are presented in Table 4.
As may be noted the presence of any of the additives sharply decreases the flammability compared to that for epoxy containing no flame retardant (LOI 19, PHRR 498 W/g, unrated in UL 94). For the derivatives of bis-2,5-(hydroxymethyl)furan, the DOPO adduct is the most effective. At a loading to provide 2% phosphorus in the blend, an LOI of 29.1, a PHRR of 305 W/g and a UL 94 rating of V0 are observed. Derivatives of DOPO (low level of oxygenation at phosphorus) are predominately gas-phase active and function by extruding PO radical to the gas phase to quench combustion For assessment of the effectiveness of these compounds, blends with DGEBA epoxy at levels to provide 1% or 2% phosphorus and in two cases 5% phosphorus were prepared [58]. The flammability of the blends was determined using limiting oxygen index (LOI), microscale combustion calorimetry (MCC), and the underwriter Laboratories vertical burn test (UL94). Results are presented in Table 4.      As may be noted the presence of any of the additives sharply decreases the flammability compared to that for epoxy containing no flame retardant (LOI 19, PHRR 498 W/g, unrated in UL 94). For the derivatives of bis-2,5-(hydroxymethyl)furan, the DOPO adduct is the most effective. At a loading to provide 2% phosphorus in the blend, an LOI of 29.1, a PHRR of 305 W/g and a UL 94 rating of V0 are observed. Derivatives of DOPO (low level of oxygenation at phosphorus) are predominately gas-phase active and function by extruding PO radical to the gas phase to quench combustion propagating reactions [61][62][63][64][65][66]. On the other hand, derivatives of diethyl-and diphenylphosphite are probably not gas-phase active and a higher loading (5% phosphorus) of these additives is required to achieve comparable flame retardancy.
A comparison of the effectiveness of the DOPO derivatives of the diacrylates of bis-2,5-(hydroxymethyl)furan, bisphenol A and tetrabromobisphenol A is presented in Table 5. A comparison of the effectiveness of the DOPO derivatives of the diacrylates of bis-2,5-(hydroxymethyl)furan, bisphenol A and tetrabromobisphenol A is presented in Table 5. At a loading to provide 2 wgt% phosphorus, the impact of the additives is quite similar: LOI of 30, 40% reduction in PHRR and a V0 UL 94 rating. Certainly, the impact of the furan derivative is comparable to that of the corresponding bisphenyl compounds. It is interesting that the brominated bisphenyl compound is little more effective than the analogous compound containing no bromine [67,68]. This may suggest that phosphorus is primarily responsible for the flame retardant effect.

Methods and Instrumentation
Instrumentation and methods for characterization using spectroscopic, chromatographic and thermal techniques have been described previously [58]. Infrared spectra were obtained by attenuated total reflectance (ATR) using a Thermo Scientific (Waltham, MA, USA) Nicolet 380 FT-IR spectrophotomerter. Absorptions were recorded in wavenumbers (cm -1 ), and absorption intensities were classified in the usual fashion as very weak (vw), weak (w), medium (m), strong (s), and very strong (vs) relative to the strongest band in the spectrum. Nuclear magnetic resonance (NMR) spectra were obtained using a 5-15% solution in deuterochloroform or dimethylsulfoxide-d6 and a Varian A comparison of the effectiveness of the DOPO derivatives of the diacrylates of bis-2,5-(hydroxymethyl)furan, bisphenol A and tetrabromobisphenol A is presented in Table 5. At a loading to provide 2 wgt% phosphorus, the impact of the additives is quite similar: LOI of 30, 40% reduction in PHRR and a V0 UL 94 rating. Certainly, the impact of the furan derivative is comparable to that of the corresponding bisphenyl compounds. It is interesting that the brominated bisphenyl compound is little more effective than the analogous compound containing no bromine [67,68]. This may suggest that phosphorus is primarily responsible for the flame retardant effect.

Methods and Instrumentation
Instrumentation and methods for characterization using spectroscopic, chromatographic and thermal techniques have been described previously [58]. Infrared spectra were obtained by attenuated total reflectance (ATR) using a Thermo Scientific (Waltham, MA, USA) Nicolet 380 FT-IR spectrophotomerter. Absorptions were recorded in wavenumbers (cm -1 ), and absorption intensities were classified in the usual fashion as very weak (vw), weak (w), medium (m), strong (s), and very A comparison of the effectiveness of the DOPO derivatives of the diacrylates of bis-2,5-(hydroxymethyl)furan, bisphenol A and tetrabromobisphenol A is presented in Table 5. At a loading to provide 2 wgt% phosphorus, the impact of the additives is quite similar: LOI of 30, 40% reduction in PHRR and a V0 UL 94 rating. Certainly, the impact of the furan derivative is comparable to that of the corresponding bisphenyl compounds. It is interesting that the brominated bisphenyl compound is little more effective than the analogous compound containing no bromine [67,68]. This may suggest that phosphorus is primarily responsible for the flame retardant effect.

Methods and Instrumentation
Instrumentation and methods for characterization using spectroscopic, chromatographic and thermal techniques have been described previously [58]. Infrared spectra were obtained by attenuated total reflectance (ATR) using a Thermo Scientific (Waltham, MA, USA) Nicolet 380 FT-IR spectrophotomerter. Absorptions were recorded in wavenumbers (cm -1 ), and absorption intensities were classified in the usual fashion as very weak (vw), weak (w), medium (m), strong (s), and very strong (vs) relative to the strongest band in the spectrum. Nuclear magnetic resonance (NMR) spectra were obtained using a 5-15% solution in deuterochloroform or dimethylsulfoxide-d6 and a Varian (Santa Clara, CA, USA) Mercury 300 MHz or an INOVA 500 MHz spectrometer. Proton and carbon 29 A comparison of the effectiveness of the DOPO derivatives of the diacrylates of bis-2,5-(hydroxymethyl)furan, bisphenol A and tetrabromobisphenol A is presented in Table 5. At a loading to provide 2 wgt% phosphorus, the impact of the additives is quite similar: LOI of 30, 40% reduction in PHRR and a V0 UL 94 rating. Certainly, the impact of the furan derivative is comparable to that of the corresponding bisphenyl compounds. It is interesting that the brominated bisphenyl compound is little more effective than the analogous compound containing no bromine [67,68]. This may suggest that phosphorus is primarily responsible for the flame retardant effect.

Methods and Instrumentation
Instrumentation and methods for characterization using spectroscopic, chromatographic and thermal techniques have been described previously [58]. Infrared spectra were obtained by attenuated total reflectance (ATR) using a Thermo Scientific (Waltham, MA, USA) Nicolet 380 FT-IR spectrophotomerter. Absorptions were recorded in wavenumbers (cm -1 ), and absorption intensities were classified in the usual fashion as very weak (vw), weak (w), medium (m), strong (s), and very strong (vs) relative to the strongest band in the spectrum. Nuclear magnetic resonance (NMR) spectra

V0 330
Molecules 2020, 25, x FOR PEER REVIEW 7 of 14 A comparison of the effectiveness of the DOPO derivatives of the diacrylates of bis-2,5-(hydroxymethyl)furan, bisphenol A and tetrabromobisphenol A is presented in Table 5. At a loading to provide 2 wgt% phosphorus, the impact of the additives is quite similar: LOI of 30, 40% reduction in PHRR and a V0 UL 94 rating. Certainly, the impact of the furan derivative is comparable to that of the corresponding bisphenyl compounds. It is interesting that the brominated bisphenyl compound is little more effective than the analogous compound containing no bromine [67,68]. This may suggest that phosphorus is primarily responsible for the flame retardant effect.

Methods and Instrumentation
Instrumentation and methods for characterization using spectroscopic, chromatographic and thermal techniques have been described previously [58]. Infrared spectra were obtained by attenuated total reflectance (ATR) using a Thermo Scientific (Waltham, MA, USA) Nicolet 380 FT-IR spectrophotomerter. Absorptions were recorded in wavenumbers (cm -1 ), and absorption intensities were classified in the usual fashion as very weak (vw), weak (w), medium (m), strong (s), and very strong (vs) relative to the strongest band in the spectrum. Nuclear magnetic resonance (NMR) spectra 30.4 V0 262 At a loading to provide 2 wgt% phosphorus, the impact of the additives is quite similar: LOI of 30, 40% reduction in PHRR and a V0 UL 94 rating. Certainly, the impact of the furan derivative is comparable to that of the corresponding bisphenyl compounds. It is interesting that the brominated bisphenyl compound is little more effective than the analogous compound containing no bromine [67,68]. This may suggest that phosphorus is primarily responsible for the flame retardant effect.

Methods and Instrumentation
Instrumentation and methods for characterization using spectroscopic, chromatographic and thermal techniques have been described previously [58]. Infrared spectra were obtained by attenuated total reflectance (ATR) using a Thermo Scientific (Waltham, MA, USA) Nicolet 380 FT-IR spectrophotomerter. Absorptions were recorded in wavenumbers (cm −1 ), and absorption intensities were classified in the usual fashion as very weak (vw), weak (w), medium (m), strong (s), and very strong (vs) relative to the strongest band in the spectrum. Nuclear magnetic resonance (NMR) spectra were obtained using a 5-15% solution in deuterochloroform or dimethylsulfoxide-d 6 and a Varian (Santa Clara, CA, USA) Mercury 300 MHz or an INOVA 500 MHz spectrometer. Proton and carbon chemical shifts are reported in parts per million (δ) with respect to tetramethylsilane (TMS) as an internal reference (δ = 0.00). Phosphorus chemical shifts are in δ with respect of triphenylphosphate as internal reference (δ = −18.00). Electrospray ionization mass spectrometry (ESI-MS) was carried out using a Waters (Milford, MA, USA) Acquity/LCT Premier XE unit with samples introduced as dilute solutions in acetonitrile/water. Matrix assisted laser desorption ionization (MALDI) time of flight mass spectrometry was performed using a Bruker (Billerica, MA, USA) Daltonics Autoflex unit and 2,6-dihydroxybenzoic acid as matrix. Thermal transitions were determined by differential scanning calorimetry (DSC) using a TA instruments (New Castle, DE, USA) Q2000 instrument. Samples, contained in standard aluminum pans, were analyzed at a heating rate of 5 or 10 • C min −1 . The cell was subject to a constant purge of dry nitrogen at 50 cm 3 min −1 . Thermogravimetry was performed using a TA instruments Q500 instrument. Typically, a heating rate of 5 or 10 • C min −1 was used. Samples (4-10 mg) were contained in a platinum pan. The sample compartment was purged with dry nitrogen at 50 cm 3 min −1 during analysis. Peak heat release rates were determined using a Fire Testing Technology, Ltd. (East Grinstead, UK) (FTT) microscale combustion calorimeter, model FAA-PCFC. Limiting oxygen index values were determined using an FTT Oxygen Index unit. Vertical burn tests were conducted in an FTT UL 94 test chamber.

Test Specimen
Standard plaques for flammability testing were prepared from DGEBA epoxy using 2-ethyl-4-methylimidazole as hardener [58]. Samples were prepared by dissolving sufficient additive to provide a loading of one percent phosphorus in digycidyl ether of bisphenol A (DGEBA) epoxy at 90 • C. Hardener 2-ethyl-4-methylimidazole was added, mixed thoroughly, and the whole was poured into Teflon molds of appropriate dimensions which had been allowed to equilibrate at 95 • C. The blends were cured initially at 95 • C for 1.5 h and then at 130 • C for 1.5 h. The samples were then allowed to cool slowly (0.6 • C/min) to room temperature.
To a stirred solution of 21.60 g (0.17 mol) of bis-2,5-(hydroxymethyl)furan and 50 mL (0.37 mol) of triethylamine, 200 mL of anhydrous THF maintained near 0 • C (external ice bath) was added, dropwise, over a period of 2 h, a solution of 30 mL (0.37 mol) of acryloyl chloride in 150 mL of anhydrous THF. The resulting mixture was allowed to warm to room temperature and stirred for 6 h. Water (200 mL) was added dropwise, followed by 200 mL diethyl ether. The layers were separated and the ether solution was washed, successively, with three 80-mL portions of 5% aqueous hydrochloric acid solution, 80 mL of 5% aqueous sodium hydroxide solution and 80 mL of saturated aqueous sodium chloride solution. The ether solution was dried over anhydrous sodium sulfate and the solvent was removed by rotary evaporation at reduced pressure. The crude product was purified by column chromatography (silica gel; 3:1 ethyl-acetate/hexane as eluent) to provide 29.08 g (72.4% yield) of the diacrylate as a colorless oil: T dec 128 The acrylate esters of bis-2,5-(hydroxymethyl)furan, bisphenol A and tetrabromobisphenol A were converted to phosphorus derivatives by Michael addition of phosphites using a previously described procedure [58]. Other adducts were prepared in an analogous manner.

Conclusions
Bioderived bis-2,5-(hydroxymethyl)furan has been utilized as a base for the generation of a series of phosphorus compounds. The dihydroxy furan was converted to the corresponding bis-acrylate which was, in turn, subjected to Michael addition of phosphite to generate phosphorus derivatives with phosphorus linked through a P-C bond. These compounds act as effective flame retardants in DGEBA epoxy. The DOPO derivative, 2,5-di[(3-dopylpropanoyl)methyl]furan, is the most effective. At a loading to provide 2% phosphorus in a DGEBA blend an LOI of 30, a UL94 rating of V0 and a 40% reduction in peak heat release rate for combustion compared to that for resin containing no additive is observed. This performance is comparable to that observed for the analogous derivatives of bisphenol A and tetrabromobisphenol A diacrylates.