One-Pot Synthesis of P(O)-N Containing Compounds Using N-Chlorosuccinimide and Their Influence in Thermal Decomposition of PU Foams

Synthesis of intermediate containing P(O)-Cl bonds is the key to converting P(O)-H bonds to P(O)-N. In this work we have performed chlorination reactions of different H-phosphinates and H-phosphonates using N-chlorosuccinimide as an environmentally-benign chlorinating agent. The chlorination reaction showed high yield and high selectivity for transformation of P(O)-H bonds into P(O)-Cl analogues, resulting in an easily separable succinimide as the by-product. Using a one-pot synthesis methodology, we have synthesized a series of P(O)-N containing derivatives whose synthesis was found to be dependent on the reaction solvents and the starting materials. The synthesized P(O)-N compounds were incorporated in flexible polyurethane foam (FPUF) and screened for their influence in thermal decomposition of FPUFs using thermogravimetric analysis (TGA) and a microscale combustion calorimeter (MCC). All solid P(O)-N compounds influenced the first-stage decomposition of FPUFs, which resulted in an accelerated decomposition or temporary stabilization of this stage. However, the liquid P(O)-N derivatives volatilize at an earlier stage and could be active in the gas phase. In addition, they also work in condensed phase via acid catalyzed decomposition for FPUFs.


Introduction
Over the past century, the use of organophosphorus compounds has been extensively explored due to their versatility [1,2]. This exploration has paved the way to develop new synthetic methodologies and further the design of new organophosphorus compounds [3,4].
Among the latter, the P(O)-N containing organophosphorus compounds, namely phosphoramidates and phosphonamidates, have received a great deal of attention owing to their diverse applications. They have been used for treating and controlling various diseases, making them useful candidates for pharmaceutical products [5,6]. Some of them have shown to be excellent flame retardants on polyurethane [7][8][9][10][11], polycarbonates [12,13], polyamide 66 [14], epoxy resins [15][16][17], and cellulose-based fabrics [18][19][20]. Because of this wide range of applications, production of P(O)-N containing organophosphorus compounds has gained commercial importance. For the commercial exploitation of such compounds, industrially feasible synthetic strategies need to be developed. Since 1945, Atherton-Todd reaction has been widely employed as a synthetic methods, and stored under N 2 [52]. Materials used in the production of flexible PU foams: PO 56 (propylene oxide polyol with a hydroxyl value of 56 mg KOH/g and molecular weight of 3000 g/moL) from Covestro, formerly Bayer MaterialScience AG (Leverkusen, Germany), KOSMOS EF (catalyst of tin ricinoleate) from Evonik Industries AG (Essen, Germany), Tegoamin ZE1 (amine catalyst, blowing catalyst) from Evonik Industries AG. STABI AKTIV (silicone surfactant) from Momentive Performance Materials GmbH (Leverkusen Germany) were supplied by FoamPartner AG (Wolfhausen, Switzerland). T80 (technical grade; a mixture of 2,4-toluene diisocyanate and 2,6-toluene diisocyanate isomers with ratio 80:20, respectively) was purchased from Sigma-Aldrich, Switzerland.

Thermal and Fire Characterization, ICP-OES and NMR
Heat release rates (HRRs) of FPUFs were determined using a microscale combustion calorimeter (MCC) (Fire Testing Technology Instrument UK) following ASTM D7309. Each sample (about 2 mg) was exposed to a heating rate of 1 • C·s −1 from 100 • C to 700 • C in the pyrolysis zone and at a combustion temperature of 900 • C. The thermogravimetric analysis (TGA) of FPUFs was performed on a NETZSCH TG 209 F1 instrument (NETZSCH-Gerätebau GmbH, Selb, Germany). Tests were conducted on 2-5 mg sample under N 2 atmosphere and under air at a heating rate of 10 • C/min from 25 • C to 800 • C. The TGA curves represent the average of two measurements. Differential scanning calorimetry (DSC) analyses were performed on DSC 214 Polyma instrument (NETZSCH-Gerätebau GmbH, Selb, Germany). All experiments were conducted at a controlled heating rate of 10 • C/min under nitrogen flow of 50 mL/min. Phosphorus-contents of FPUFs samples and P(O)-N derivatives were determined using PerkinElmer OPTIMA 3000 ICP-OES (PerkinElmer AG, Rotkreuz, Switzerland). The parameters of the NMR spectroscopy are summarized in the Supporting Information and fully assigned 1 H, 13 C and 31 P NMR spectra are shown in Figures S1-S12.

Synthesis of Phosphonamidates (General Procedure)
To a stirred solution of DOPO (5.0 g, 23 mmoL) in dichloromethane (50 mL), NCS (3.40 g, 25 mmoL) was gradually added under nitrogen atmosphere at 0 • C. The reaction mixture was then allowed to warm to ambient temperature and stirred for additional 2 h. Triethylamine (3.87 mL, 28 mmoL) and the corresponding amine (28 mmoL) were then added drop-wise to the reaction mixture at 0 • C. The reaction mixture was then stirred at room temperature overnight. Volatiles were then completely removed and the solid was stirred in water and then separated by filtration (except 6-(diethylamino)dibenzo[c,e][1,2]oxaphosphinine 6-oxide (DA-DOPO)). All compounds were fully characterized and found identical to the published data [9,29,53].

Synthesis of Phosphoramidates (General Procedure)
Synthesis of DPP based phosphoramidates: To a stirred solution of diphenyl phosphite (4.09 mL, 21 mmoL) in dichloromethane (50 mL), NCS (3.14 g, 24 mmoL) was gradually added under nitrogen atmosphere at 0 • C. The reaction mixture was then allowed to warm to ambient temperature and stirred for additional 2 h. Triethylamine (3.28 mL, 24 mmoL) and the corresponding amine (24 mmoL) in dichloromethane (10 mL) were then added drop-wise at 0 • C. The reaction mixture was then stirred at ambient temperature overnight. Volatiles were then completely removed under vacuum and the residue was stirred in diethyl ether and filtered. The volatiles were then completely removed from the filtrate under vacuum. After purifications, all compounds were fully characterized and found identical to the published data [8,[54][55][56].
Synthesis of DEP based phosphoramidates: To a stirred solution of diethyl phosphite (4.66 mL, 36 mmoL) in dichloromethane (50 mL), NCS (5.32 g, 40 mmoL) was gradually added under nitrogen atmosphere at 0 • C. The reaction mixture was then allowed to warm to ambient temperature and stirred for additional 2 h. Triethylamine (5.56 mL, 40 mmoL,) and the corresponding amine (40 mmoL) were then drop-wise added at 0 • C. The reaction mixture was stirred at ambient temperature overnight. Volatiles were then completely removed and the oily residue was stirred in diethyl ether (except diethyl benzylphosphoramidate (BA-DEP) and tetraethyl ethane-1,2-diylbis(phosphoramidate) (EDA-DEP)) and filtered. The volatiles were then removed from the filtrate under vacuum. After purification, all the compounds were fully characterized and found identical to the published data [19].

Preparation of FPUF
FPUFs were manufactured by incorporating suitable concentration of synthesized phosphonamidates and phosphoramidates compounds so as to obtain 2 wt % phosphorus in the resulting foams [8]. The required quantities of the polyol (100 parts), phosphorus compounds (parts based on wt % of P-content), water (1.6 parts), Tegoamine ZE1 (0.4 parts), STABI AKTIV (0.5 parts), and KOSMOS EF (0.45 parts) were mixed in a plastic cup with digital homogenizer Ultra Turrax T18 from IKA at speed of 9000 rpm for 90 s. Subsequently, T80 (26.20 parts) was added during continuous stirring of the reaction components. The mixture was then stirred for a further 6 s and immediately transferred into a container to rise freely. The samples were then cured at 80 • C for 1 h and allowed to cool down to ambient temperatures. All prepared FPUFs were conditioned for 48 h at 25 • C and 50% of relative humidity prior to further analysis.

Synthesis of the P(O)-N Compounds
The advantage of the transformation reaction of P(O)-H bonds into its corresponding P(O)-N bonds via intermediate chlorinating step with NCS is the production of succinimide as a benign by-product. The latter can be recovered to regenerate N-chlorosuccinimde, mitigating any potential chemical by-product wastes [57,58]. Thus, the use of NCS can be considered as an alternative sustainable green chlorination agent compared to CCl 4 , SO 2 Cl 2 , and trichlorocyanuric acid (Scheme 1) [29].
Polymers 2018, 10, x FOR PEER REVIEW 4 of 16 mixture was stirred at ambient temperature overnight. Volatiles were then completely removed and the oily residue was stirred in diethyl ether (except diethyl benzylphosphoramidate (BA-DEP) and tetraethyl ethane-1,2-diylbis(phosphoramidate) (EDA-DEP)) and filtered. The volatiles were then removed from the filtrate under vacuum. After purification, all the compounds were fully characterized and found identical to the published data [19].

Preparation of FPUF
FPUFs were manufactured by incorporating suitable concentration of synthesized phosphonamidates and phosphoramidates compounds so as to obtain 2 wt % phosphorus in the resulting foams [8]. The required quantities of the polyol (100 parts), phosphorus compounds (parts based on wt % of P-content), water (1.6 parts), Tegoamine ZE1 (0.4 parts), STABI AKTIV (0.5 parts), and KOSMOS EF (0.45 parts) were mixed in a plastic cup with digital homogenizer Ultra Turrax T18 from IKA at speed of 9000 rpm for 90 s. Subsequently, T80 (26.20 parts) was added during continuous stirring of the reaction components. The mixture was then stirred for a further 6 s and immediately transferred into a container to rise freely. The samples were then cured at 80 °C for 1 h and allowed to cool down to ambient temperatures. All prepared FPUFs were conditioned for 48 h at 25 °C and 50% of relative humidity prior to further analysis.

Synthesis of the P(O)-N Compounds
The advantage of the transformation reaction of P(O)-H bonds into its corresponding P(O)-N bonds via intermediate chlorinating step with NCS is the production of succinimide as a benign byproduct. The latter can be recovered to regenerate N-chlorosuccinimde, mitigating any potential chemical by-product wastes [57,58]. Thus, the use of NCS can be considered as an alternative sustainable green chlorination agent compared to CCl4, SO2Cl2, and trichlorocyanuric acid (Scheme 1) [29]. The transformation reaction of P(O)-H bonds into P(O)-Cl (of respective diethyl phosphorochloridate, diphenyl phosphorochloridate, and 6-chlorodibenzo[c,e][1,2]oxaphosphinine 6-oxide reaction) was confirmed by 31 P NMR spectroscopy of the reaction mixture after the first step. All chloro-derivatives were also isolated and their chemical and physical properties were found to be consistent with their reported data [59,60]. The initial reaction screening showed that the transformation of the P(O)-H bond into its corresponding P(O)-Cl is affected by the impurities present in the starting material (Table 1). For instance, in the synthesis of diphenyl phosphorochloridate (Table 1, entry 2) we observed the formation of triphenyl phosphate, which is due to the presence of residual phenol in diphenyl phosphite. The transformation reaction of P(O)-H bonds into P(O)-Cl (of respective diethyl phosphorochloridate, diphenyl phosphorochloridate, and 6-chlorodibenzo[c,e][1,2]oxaphosphinine 6-oxide reaction) was confirmed by 31 P NMR spectroscopy of the reaction mixture after the first step. All chloro-derivatives were also isolated and their chemical and physical properties were found to be consistent with their reported data [59,60]. The initial reaction screening showed that the transformation of the P(O)-H bond into its corresponding P(O)-Cl is affected by the impurities present in the starting material (Table 1). For instance, in the synthesis of diphenyl phosphorochloridate (Table 1, entry 2) we observed the formation of triphenyl phosphate, which is due to the presence of residual phenol in diphenyl phosphite. atmosphere. Toluene (30 mL) was then added and the reaction mixture was cooled to 0 °C. NCS (254 mmoL) was then added in small portions and the reaction was then allowed to warm up to ambient temperature and stirred for 2 h. The precipitate was then removed by filtration using Schlenk-frit and the volatiles were removed under vacuum, affording the desired product; (b) The resulting products were analyzed by GC-MS and NMR spectroscopy; (c) 31 P NMR spectra were recorded in CDCl3; (d) Diphenyl phosphite contained ≤15% phenol and used as received; (e) DOPO was recrystallized from toluene and dried in vacuum at 80 °C prior to use.
The transformation reaction of P(O)-H bonds into P(O)-Cl bonds using NCS was optimized using PA-DOPO as a model derivative. Thus the synthesis of PA-DOPO was optimized using the general synthetic procedure, using different solvent/base pairs. The results are summarized in Table  2. The transformation reaction of P(O)-H bonds into P(O)-Cl (of respective diethyl phosphorochloridate, diphenyl phosphorochloridate, and 6-chlorodibenzo[c,e][1,2]oxaphosphinine 6-oxide reaction) was confirmed by 31 P NMR spectroscopy of the reaction mixture after the first step. All chloro-derivatives were also isolated and their chemical and physical properties were found to be consistent with their reported data [59,60]. The initial reaction screening showed that the transformation of the P(O)-H bond into its corresponding P(O)-Cl is affected by the impurities present in the starting material (Table 1). For instance, in the synthesis of diphenyl phosphorochloridate (Table 1, entry 2) we observed the formation of triphenyl phosphate, which is due to the presence of residual phenol in diphenyl phosphite. atmosphere. Toluene (30 mL) was then added and the reaction mixture was cooled to 0 °C. NCS (254 mmoL) was then added in small portions and the reaction was then allowed to warm up to ambient temperature and stirred for 2 h. The precipitate was then removed by filtration using Schlenk-frit and the volatiles were removed under vacuum, affording the desired product; (b) The resulting products were analyzed by GC-MS and NMR spectroscopy; (c) 31 P NMR spectra were recorded in CDCl3; (d) Diphenyl phosphite contained ≤15% phenol and used as received; (e) DOPO was recrystallized from toluene and dried in vacuum at 80 °C prior to use.
The transformation reaction of P(O)-H bonds into P(O)-Cl bonds using NCS was optimized using PA-DOPO as a model derivative. Thus the synthesis of PA-DOPO was optimized using the general synthetic procedure, using different solvent/base pairs. The results are summarized in Table  2.  The transformation reaction of P(O)-H bonds into P(O)-Cl (of respective diethyl phosphorochloridate, diphenyl phosphorochloridate, and 6-chlorodibenzo[c,e][1,2]oxaphosphinine 6-oxide reaction) was confirmed by 31 P NMR spectroscopy of the reaction mixture after the first step. All chloro-derivatives were also isolated and their chemical and physical properties were found to be consistent with their reported data [59,60]. The initial reaction screening showed that the transformation of the P(O)-H bond into its corresponding P(O)-Cl is affected by the impurities present in the starting material (Table 1). For instance, in the synthesis of diphenyl phosphorochloridate (Table 1, entry 2) we observed the formation of triphenyl phosphate, which is due to the presence of residual phenol in diphenyl phosphite. atmosphere. Toluene (30 mL) was then added and the reaction mixture was cooled to 0 °C. NCS (254 mmoL) was then added in small portions and the reaction was then allowed to warm up to ambient temperature and stirred for 2 h. The precipitate was then removed by filtration using Schlenk-frit and the volatiles were removed under vacuum, affording the desired product; (b) The resulting products were analyzed by GC-MS and NMR spectroscopy; (c) 31 P NMR spectra were recorded in CDCl3; (d) Diphenyl phosphite contained ≤15% phenol and used as received; (e) DOPO was recrystallized from toluene and dried in vacuum at 80 °C prior to use.
The transformation reaction of P(O)-H bonds into P(O)-Cl bonds using NCS was optimized using PA-DOPO as a model derivative. Thus the synthesis of PA-DOPO was optimized using the general synthetic procedure, using different solvent/base pairs. The results are summarized in Table  2. The transformation reaction of P(O)-H bonds into P(O)-Cl (of respective diethyl phosphorochloridate, diphenyl phosphorochloridate, and 6-chlorodibenzo[c,e][1,2]oxaphosphinine 6-oxide reaction) was confirmed by 31 P NMR spectroscopy of the reaction mixture after the first step. All chloro-derivatives were also isolated and their chemical and physical properties were found to be consistent with their reported data [59,60]. The initial reaction screening showed that the transformation of the P(O)-H bond into its corresponding P(O)-Cl is affected by the impurities present in the starting material (Table 1). For instance, in the synthesis of diphenyl phosphorochloridate (Table 1, entry 2) we observed the formation of triphenyl phosphate, which is due to the presence of residual phenol in diphenyl phosphite. atmosphere. Toluene (30 mL) was then added and the reaction mixture was cooled to 0 °C. NCS (254 mmoL) was then added in small portions and the reaction was then allowed to warm up to ambient temperature and stirred for 2 h. The precipitate was then removed by filtration using Schlenk-frit and the volatiles were removed under vacuum, affording the desired product; (b) The resulting products were analyzed by GC-MS and NMR spectroscopy; (c) 31 P NMR spectra were recorded in CDCl3; (d) Diphenyl phosphite contained ≤15% phenol and used as received; (e) DOPO was recrystallized from toluene and dried in vacuum at 80 °C prior to use.
The transformation reaction of P(O)-H bonds into P(O)-Cl bonds using NCS was optimized using PA-DOPO as a model derivative. Thus the synthesis of PA-DOPO was optimized using the general synthetic procedure, using different solvent/base pairs. The results are summarized in Table  2. The transformation reaction of P(O)-H bonds into P(O)-Cl (of respective diethyl phosphorochloridate, diphenyl phosphorochloridate, and 6-chlorodibenzo[c,e][1,2]oxaphosphinine 6-oxide reaction) was confirmed by 31 P NMR spectroscopy of the reaction mixture after the first step. All chloro-derivatives were also isolated and their chemical and physical properties were found to be consistent with their reported data [59,60]. The initial reaction screening showed that the transformation of the P(O)-H bond into its corresponding P(O)-Cl is affected by the impurities present in the starting material (Table 1). For instance, in the synthesis of diphenyl phosphorochloridate (Table 1, entry 2) we observed the formation of triphenyl phosphate, which is due to the presence of residual phenol in diphenyl phosphite. atmosphere. Toluene (30 mL) was then added and the reaction mixture was cooled to 0 °C. NCS (254 mmoL) was then added in small portions and the reaction was then allowed to warm up to ambient temperature and stirred for 2 h. The precipitate was then removed by filtration using Schlenk-frit and the volatiles were removed under vacuum, affording the desired product; (b) The resulting products were analyzed by GC-MS and NMR spectroscopy; (c) 31 P NMR spectra were recorded in CDCl3; (d) Diphenyl phosphite contained ≤15% phenol and used as received; (e) DOPO was recrystallized from toluene and dried in vacuum at 80 °C prior to use.
The transformation reaction of P(O)-H bonds into P(O)-Cl bonds using NCS was optimized using PA-DOPO as a model derivative. Thus the synthesis of PA-DOPO was optimized using the general synthetic procedure, using different solvent/base pairs. The results are summarized in Table  2. The transformation reaction of P(O)-H bonds into P(O)-Cl (of respective diethyl phosphorochloridate, diphenyl phosphorochloridate, and 6-chlorodibenzo[c,e][1,2]oxaphosphinine 6-oxide reaction) was confirmed by 31 P NMR spectroscopy of the reaction mixture after the first step. All chloro-derivatives were also isolated and their chemical and physical properties were found to be consistent with their reported data [59,60]. The initial reaction screening showed that the transformation of the P(O)-H bond into its corresponding P(O)-Cl is affected by the impurities present in the starting material (Table 1). For instance, in the synthesis of diphenyl phosphorochloridate (Table 1, entry 2) we observed the formation of triphenyl phosphate, which is due to the presence of residual phenol in diphenyl phosphite. atmosphere. Toluene (30 mL) was then added and the reaction mixture was cooled to 0 °C. NCS (254 mmoL) was then added in small portions and the reaction was then allowed to warm up to ambient temperature and stirred for 2 h. The precipitate was then removed by filtration using Schlenk-frit and the volatiles were removed under vacuum, affording the desired product; (b) The resulting products were analyzed by GC-MS and NMR spectroscopy; (c) 31 P NMR spectra were recorded in CDCl3; (d) Diphenyl phosphite contained ≤15% phenol and used as received; (e) DOPO was recrystallized from toluene and dried in vacuum at 80 °C prior to use.
The transformation reaction of P(O)-H bonds into P(O)-Cl bonds using NCS was optimized using PA-DOPO as a model derivative. Thus the synthesis of PA-DOPO was optimized using the general synthetic procedure, using different solvent/base pairs. The results are summarized in Table  2. Toluene (30 mL) was then added and the reaction mixture was cooled to 0 • C. NCS (254 mmoL) was then added in small portions and the reaction was then allowed to warm up to ambient temperature and stirred for 2 h. The precipitate was then removed by filtration using Schlenk-frit and the volatiles were removed under vacuum, affording the desired product; (b) The resulting products were analyzed by GC-MS and NMR spectroscopy; (c) 31 P NMR spectra were recorded in CDCl 3 ; (d) Diphenyl phosphite contained ≤15% phenol and used as received; (e) DOPO was recrystallized from toluene and dried in vacuum at 80 • C prior to use.
The transformation reaction of P(O)-H bonds into P(O)-Cl bonds using NCS was optimized using PA-DOPO as a model derivative. Thus the synthesis of PA-DOPO was optimized using the general synthetic procedure, using different solvent/base pairs. The results are summarized in Table 2. The transformation reaction of P(O)-H bonds into P(O)-Cl (of respective diethyl phosphorochloridate, diphenyl phosphorochloridate, and 6-chlorodibenzo[c,e][1,2]oxaphosphinine 6-oxide reaction) was confirmed by 31 P NMR spectroscopy of the reaction mixture after the first step. All chloro-derivatives were also isolated and their chemical and physical properties were found to be consistent with their reported data [59,60]. The initial reaction screening showed that the transformation of the P(O)-H bond into its corresponding P(O)-Cl is affected by the impurities present in the starting material (Table 1). For instance, in the synthesis of diphenyl phosphorochloridate (Table 1, entry 2) we observed the formation of triphenyl phosphate, which is due to the presence of residual phenol in diphenyl phosphite. atmosphere. Toluene (30 mL) was then added and the reaction mixture was cooled to 0 °C. NCS (254 mmoL) was then added in small portions and the reaction was then allowed to warm up to ambient temperature and stirred for 2 h. The precipitate was then removed by filtration using Schlenk-frit and the volatiles were removed under vacuum, affording the desired product; (b) The resulting products were analyzed by GC-MS and NMR spectroscopy; (c) 31 P NMR spectra were recorded in CDCl3; (d) Diphenyl phosphite contained ≤15% phenol and used as received; (e) DOPO was recrystallized from toluene and dried in vacuum at 80 °C prior to use.
The transformation reaction of P(O)-H bonds into P(O)-Cl bonds using NCS was optimized using PA-DOPO as a model derivative. Thus the synthesis of PA-DOPO was optimized using the general synthetic procedure, using different solvent/base pairs. The results are summarized in Table  2. It was observed that the reaction mixtures in chloroform, toluene, and acetonitrile ( Table 2, Entries 1-3) turned dark brown upon addition of PA. After removal of the reaction volatiles, dark residue was obtained which required further purification steps to afford a clean PA-DOPO. On the other hand, a white residue was obtained when dichloromethane and triethylamine were used as  Table 2, Entry 4). On washing the white residue with water, a clean PA-DOPO was obtained without any further purification procedure. By changing the organic base to inorganic base (Table 2, Entry 5), a relatively lower yield of PA-DOPO was obtained which can be attributed to the poor solubility of the inorganic base in organic solvents. Subsequently, all phosphonamidates and phosphoramidates ( Figure 1) were synthesized based on the optimized synthetic approach ( residue was obtained which required further purification steps to afford a clean PA-DOPO. On the other hand, a white residue was obtained when dichloromethane and triethylamine were used as a solvent and an organic base, respectively ( Table 2, Entry 4). On washing the white residue with water, a clean PA-DOPO was obtained without any further purification procedure. By changing the organic base to inorganic base ( Table 2, Entry 5), a relatively lower yield of PA-DOPO was obtained which can be attributed to the poor solubility of the inorganic base in organic solvents. Subsequently, all phosphonamidates and phosphoramidates ( Figure 1) were synthesized based on the optimized synthetic approach ( Table 2, Entry 4). The phosphonamidate derivatives of DOPO compound are water insoluble, which simplifies their cleaning procedure. They are purified by washing away the crude products with water to remove the triethylamine hydrochloride salt and the succinimide, affording clean derivatives in relatively moderate to high yields (Table 3, Entry 1). Following the same cleaning methodology, the phosphoramidate derivatives of DPP were cleaned by washing the crude products with water. However, both PA-DPP and DA-DPP derivatives ( Table 3, Entry 2) required a modified cleaning procedure using column chromatography for satisfactory purity. This procedure was found to reduce the overall reaction yield. To overcome the low yield of PA-DPP and DA-DPP, attempts were made to isolate diphenyl phosphorochloridate from succinimide by vacuum distillation to avoid the side reactions. As the boiling point of diphenyl phosphorochloridate is relatively high, the succinimide impurity could separate and become condensed in the distillation apparatus. However, the phosphoramidate derivatives of DEP are water soluble and cannot be purified by washing them with water. Accordingly, these derivatives were isolated from the reaction mixture either by dissolving them in diethyl ether or THF or by vacuum distillation (Table 3, Entry 3). It is noteworthy to mention that due to the low boiling point of diethyl phosphorochloridate at low pressure, it can be quantitatively isolated from the reaction mixture in high purity via distillation. It can be then stored and subsequently used for further chemical reactions. The phosphonamidate derivatives of DOPO compound are water insoluble, which simplifies their cleaning procedure. They are purified by washing away the crude products with water to remove the triethylamine hydrochloride salt and the succinimide, affording clean derivatives in relatively moderate to high yields (Table 3, Entry 1). Following the same cleaning methodology, the phosphoramidate derivatives of DPP were cleaned by washing the crude products with water. However, both PA-DPP and DA-DPP derivatives ( Table 3, Entry 2) required a modified cleaning procedure using column chromatography for satisfactory purity. This procedure was found to reduce the overall reaction yield. To overcome the low yield of PA-DPP and DA-DPP, attempts were made to isolate diphenyl phosphorochloridate from succinimide by vacuum distillation to avoid the side reactions. As the boiling point of diphenyl phosphorochloridate is relatively high, the succinimide impurity could separate and become condensed in the distillation apparatus. However, the phosphoramidate derivatives of DEP are water soluble and cannot be purified by washing them with water. Accordingly, these derivatives were isolated from the reaction mixture either by dissolving them in diethyl ether or THF or by vacuum distillation (Table 3, Entry 3). It is noteworthy to mention that due to the low boiling point of diethyl phosphorochloridate at low pressure, it can be quantitatively isolated from the reaction mixture in high purity via distillation. It can be then stored and subsequently used for further chemical reactions.

Proposed Chlorination Mechanism
The H-phosphinate (DOPO) and H-phosphonates (DEP and DPP) are considered tetra-coordinated P(III) compounds [2]. As these compounds resemble the chemical structure of the tetra-coordinated of P(V) compounds (phosphonates), the P-center is predicted to be an electrophilic center. The electrophilic property can be modified to be a nucleophilic due to their tautomeric equilibrium in a solution between the tetra-coordinated (1a) and tri-coordinated (1b) forms (Scheme 2).

Proposed Chlorination Mechanism
The H-phosphinate (DOPO) and H-phosphonates (DEP and DPP) are considered tetracoordinated P(III) compounds [2]. As these compounds resemble the chemical structure of the tetracoordinated of P(V) compounds (phosphonates), the P-center is predicted to be an electrophilic center. The electrophilic property can be modified to be a nucleophilic due to their tautomeric equilibrium in a solution between the tetra-coordinated (1a) and tri-coordinated (1b) forms (Scheme 2). Scheme 2. Proposed mechanism of the chlorination reaction using NCS.
The Cl atom has a positive charge due to the influence of the electron-withdrawing ability of an oxygen atom on NCS, making it susceptible to attack by a nucleophilic center [61], namely the tricoordinated P-atom; this produces the succinimide and the corresponding chlorophosphate derivative. Based on a previous report with regards to chlorination of secondary amine using NCS, the chlorination reaction occurs in stoichiometric ratio and the exchange of chlorine for hydrogen happens directly (Scheme 2) [62]. The kinetic studies of the chlorination reaction using NCS are reported to take place via first order with respect to NCS [63,64].

Thermal Analysis
The thermogravimetric analysis is viewed as an assisted tool to investigate the possible effect of the incorporated flame retardant additives on the polymer matrix [65]. Therefore, the influence of the synthezised phosphorus compounds on FPUF decomposition was investigated in a non-oxidative environment and relevant data are summarized in Table 4.

Scheme 2.
Proposed mechanism of the chlorination reaction using NCS.
The Cl atom has a positive charge due to the influence of the electron-withdrawing ability of an oxygen atom on NCS, making it susceptible to attack by a nucleophilic center [61], namely the tri-coordinated P-atom; this produces the succinimide and the corresponding chlorophosphate derivative. Based on a previous report with regards to chlorination of secondary amine using NCS, the chlorination reaction occurs in stoichiometric ratio and the exchange of chlorine for hydrogen happens directly (Scheme 2) [62]. The kinetic studies of the chlorination reaction using NCS are reported to take place via first order with respect to NCS [63,64].

Thermal Analysis
The thermogravimetric analysis is viewed as an assisted tool to investigate the possible effect of the incorporated flame retardant additives on the polymer matrix [65]. Therefore, the influence of the synthezised phosphorus compounds on FPUF decomposition was investigated in a non-oxidative environment and relevant data are summarized in Table 4. The thermogravimetric analysis of virgin FPUF ( Figure 2) showed a two-stage decomposition process from 200 • C to 420 • C, which is consistent with the literature data [7,66]. The thermogravimetric analysis of virgin FPUF ( Figure 2) showed a two-stage decomposition process from 200 °C to 420 °C, which is consistent with the literature data [7,66].

Sampless
. The first stage decomposition occurs between 200 °C and 320 °C with around 22% mass loss (Figure 2), which is attributed to decomposition of the urethane linkages to produce isocyanate and its derivatives [67]. The second stage decomposition corresponds to the decomposition of the polyether main chain at about 320 °C till 420 °C, with around 77% mass loss [68]. After which, a gradual mass loss takes place with a negligible char residue (Table 4). On the other hand, the addition of flame retardants to FPUF alters the decomposition profile of the FPUF either via volatilization of The first stage decomposition occurs between 200 • C and 320 • C with around 22% mass loss (Figure 2), which is attributed to decomposition of the urethane linkages to produce isocyanate and its derivatives [67]. The second stage decomposition corresponds to the decomposition of the polyether main chain at about 320 • C till 420 • C, with around 77% mass loss [68]. After which, a gradual mass loss takes place with a negligible char residue (Table 4). On the other hand, the addition of flame retardants to FPUF alters the decomposition profile of the FPUF either via volatilization of the phosphorus compounds or their decomposition products during the first stage decomposition of FPUFs. For example, the thermal decomposition of FPUF containing PA-DOPO, AA-DOPO, and DA-DOPO showed an accelerated decomposition in the first-stage (Figure 3), which may be attributed to the catalytic influence on the decomposition of the urethane bond by the degradation products of phosphorus additives. the phosphorus compounds or their decomposition products during the first stage decomposition of FPUFs. For example, the thermal decomposition of FPUF containing PA-DOPO, AA-DOPO, and DA-DOPO showed an accelerated decomposition in the first-stage (Figure 3), which may be attributed to the catalytic influence on the decomposition of the urethane bond by the degradation products of phosphorus additives. It indicates that these phosphorus additives can act in the gas phase at a relatively low temperature [9]. After the first-stage decomposition, the second-stage decomposition profile of the FPUF contained DA-DOPO is identical to the virgin FPUR, resulting in a complete decomposition and no residue (Table 4). On the contrary, both PA-DOPO and AA-DOPO derivatives influence the second-stage decomposition by decreasing the degradation rate of the polyether main chain, affording a relative increase in the char residue of the treated FPUF. This clearly indicates the possibility of a condensed phase action of the additives. BA-DOPO did now show any influence on the first-stage decomposition of FPUF, but accelerated the rate of decomposition of the second stage, leading to complete decomposition of the FPUF. However, EDA-DOPO containing FPUF showed an increase in temperature for both first-stage and second-stage decomposition, which is attributed to the high-thermal stability of the EDA-DOPO additive and its possible condensed phase action [9]. One can speculate that EDA-DOPO is relatively stable and melts at a temperature (266 °C) which matches well with the first-stage decomposition of FPUF. At this stage, the molten EDA could react with isocyanate released to form allophanate structures and prevents its volatilization resulting in increased residue. The FPUFs containing DPP based phosphoramidates showed similar decomposition profiles (Figure 4).
.  It indicates that these phosphorus additives can act in the gas phase at a relatively low temperature [9]. After the first-stage decomposition, the second-stage decomposition profile of the FPUF contained DA-DOPO is identical to the virgin FPUR, resulting in a complete decomposition and no residue (Table 4). On the contrary, both PA-DOPO and AA-DOPO derivatives influence the second-stage decomposition by decreasing the degradation rate of the polyether main chain, affording a relative increase in the char residue of the treated FPUF. This clearly indicates the possibility of a condensed phase action of the additives. BA-DOPO did now show any influence on the first-stage decomposition of FPUF, but accelerated the rate of decomposition of the second stage, leading to complete decomposition of the FPUF. However, EDA-DOPO containing FPUF showed an increase in temperature for both first-stage and second-stage decomposition, which is attributed to the high-thermal stability of the EDA-DOPO additive and its possible condensed phase action [9]. One can speculate that EDA-DOPO is relatively stable and melts at a temperature (266 • C) which matches well with the first-stage decomposition of FPUF. At this stage, the molten EDA could react with isocyanate released to form allophanate structures and prevents its volatilization resulting in increased residue. The FPUFs containing DPP based phosphoramidates showed similar decomposition profiles (Figure 4). the phosphorus compounds or their decomposition products during the first stage decomposition of FPUFs. For example, the thermal decomposition of FPUF containing PA-DOPO, AA-DOPO, and DA-DOPO showed an accelerated decomposition in the first-stage (Figure 3), which may be attributed to the catalytic influence on the decomposition of the urethane bond by the degradation products of phosphorus additives. It indicates that these phosphorus additives can act in the gas phase at a relatively low temperature [9]. After the first-stage decomposition, the second-stage decomposition profile of the FPUF contained DA-DOPO is identical to the virgin FPUR, resulting in a complete decomposition and no residue (Table 4). On the contrary, both PA-DOPO and AA-DOPO derivatives influence the second-stage decomposition by decreasing the degradation rate of the polyether main chain, affording a relative increase in the char residue of the treated FPUF. This clearly indicates the possibility of a condensed phase action of the additives. BA-DOPO did now show any influence on the first-stage decomposition of FPUF, but accelerated the rate of decomposition of the second stage, leading to complete decomposition of the FPUF. However, EDA-DOPO containing FPUF showed an increase in temperature for both first-stage and second-stage decomposition, which is attributed to the high-thermal stability of the EDA-DOPO additive and its possible condensed phase action [9]. One can speculate that EDA-DOPO is relatively stable and melts at a temperature (266 °C) which matches well with the first-stage decomposition of FPUF. At this stage, the molten EDA could react with isocyanate released to form allophanate structures and prevents its volatilization resulting in increased residue. The FPUFs containing DPP based phosphoramidates showed similar decomposition profiles (Figure 4).
.  The first-stage decomposition of FPUF was accelerated in the presence of PA-DPP, AA-DPP, BA-DPP, and DA-DPP. This influence can be attributed to the acid-catalyzed depolymerization of the urethane bond from the acidic by-products formed from the decomposition of the phosphorus compounds [8]. This observation clearly indicates condensed phase action of these phosphorus compounds. On the contrary, the EDA-DPP containing FPUF exhibited only one stage decomposition process, with a slight increase in the residue. This observation indicates that EDA-DPP can also act in the condensed phase. The thermal decomposition FPUF containing phosphoramidate additives of DEP showed clear three-stage decomposition, except EDA-DEP containing FPUF ( Figure 5).
Polymers 2018, 10, x FOR PEER REVIEW 10 of 16 The first-stage decomposition of FPUF was accelerated in the presence of PA-DPP, AA-DPP, BA-DPP, and DA-DPP. This influence can be attributed to the acid-catalyzed depolymerization of the urethane bond from the acidic by-products formed from the decomposition of the phosphorus compounds [8]. This observation clearly indicates condensed phase action of these phosphorus compounds. On the contrary, the EDA-DPP containing FPUF exhibited only one stage decomposition process, with a slight increase in the residue. This observation indicates that EDA-DPP can also act in the condensed phase. The thermal decomposition FPUF containing phosphoramidate additives of DEP showed clear three-stage decomposition, except EDA-DEP containing FPUF ( Figure 5).  Figure S13a-d). The volatilized additives can thus act in the gas phase. The remaining trapped additives in the FPUF matrix accelerate the first-stage decomposition, possibly due to the formation of non-volatile acidic species formed from their thermal decomposition [7][8][9]69]. PA-DEP, AA-DEP, BA-DEP, and DA-DEP did not influence the second stage thermal decomposition corresponding to polyether main chains. However, EDA-DEP showed an accelerated degradation rate of the second-stage decomposition of the FPUF, but with a slight increase of the residual mass. In light of this, the phosphoramidate derivatives of DEP act mainly in the gas phase with limited action in the condensed phase for both DA-DEP and EDA-DEP.

Small Scale Fire Test with Micro Scale Combustion Calorimeter
The objective of this work was to screen series of analogues P(O)-N containing compounds to gain preliminary info on their flame retardant behavior and efficacy. MCC is a useful instrument to evaluate the flame retardant performance in milligram scale. Although the milligram scale analysis may not represent the fire performance in large scale, it can, however, provide valuable information regarding mode of action of flame retardant additives [65]. To investigate the efficacy of synthesized phosphorus compounds, FPUFs were manufactured with suitable concentration of phosphorus compounds with P-content of around 2 wt %. The MCC data of the pure FPUF and the flame retarded FPUFs were summarized in Table 5.  Figure S13a-d). The volatilized additives can thus act in the gas phase. The remaining trapped additives in the FPUF matrix accelerate the first-stage decomposition, possibly due to the formation of non-volatile acidic species formed from their thermal decomposition [7][8][9]69]. PA-DEP, AA-DEP, BA-DEP, and DA-DEP did not influence the second stage thermal decomposition corresponding to polyether main chains. However, EDA-DEP showed an accelerated degradation rate of the second-stage decomposition of the FPUF, but with a slight increase of the residual mass. In light of this, the phosphoramidate derivatives of DEP act mainly in the gas phase with limited action in the condensed phase for both DA-DEP and EDA-DEP.

Small Scale Fire Test with Micro Scale Combustion Calorimeter
The objective of this work was to screen series of analogues P(O)-N containing compounds to gain preliminary info on their flame retardant behavior and efficacy. MCC is a useful instrument to evaluate the flame retardant performance in milligram scale. Although the milligram scale analysis may not represent the fire performance in large scale, it can, however, provide valuable information regarding mode of action of flame retardant additives [65]. To investigate the efficacy of synthesized phosphorus compounds, FPUFs were manufactured with suitable concentration of phosphorus compounds with P-content of around 2 wt %. The MCC data of the pure FPUF and the flame retarded FPUFs were summarized in Table 5. The heat release rate (HRR) of the pure FPUF as a function of temperature ( Figure 6) showed a two-stage heat release, which is consistent with the TGA data. The first stage of HRR between 300 • C to 320 • C represents the heat release due to thermal decomposition of the urethane bond, affording the volatile isocyanate products. However, the second-stage HRR (320 • C-450 • C) corresponds to the thermal decomposition of the polyether main chain.  The heat release rate (HRR) of the pure FPUF as a function of temperature ( Figure 6) showed a two-stage heat release, which is consistent with the TGA data. The first stage of HRR between 300 °C to 320 °C represents the heat release due to thermal decomposition of the urethane bond, affording the volatile isocyanate products. However, the second-stage HRR (320 °C-450 °C) corresponds to the thermal decomposition of the polyether main chain. It is noteworthy to mention that reduced HRR is evidence of reduced fuel production. In MCC, reduced heat release rate and increased char formation is a clear sign of the condensed phase mode of action. The HRR profile of FPUF/EDA-DOPO ( Figure S14a) compared to the virgin FPUF shows a significant decrease of the HRR in the first-stage decomposition of the FPUF, which can be attributed to its positive influence in retarding the 1st stage decomposition of FPUF. However, the THR of flame retarded FPUF compared to the virgin FPUF (Table 5, Figure S15) did not show any significant decrease, indicating that phosphoramidates and phosphonamidates additives act primarily during the gas phase. On the other hand, BA-DOPO and EDA-DPP showed a decrease in HRC ( Figure S17), revealing that the pyrolysis of these two additives interfere with the second-stage pyrolysis of FPUF, thus reducing the degradation rate of the main chain polyether, allowing these two additives to act somehow in the condensed phase. However, the absence of the char residue clearly shows that these additives cannot create thermally stable residue, and thus may primarily work in the gas phase. It is noteworthy to mention that reduced HRR is evidence of reduced fuel production. In MCC, reduced heat release rate and increased char formation is a clear sign of the condensed phase mode of action. The HRR profile of FPUF/EDA-DOPO ( Figure S14a) compared to the virgin FPUF shows a significant decrease of the HRR in the first-stage decomposition of the FPUF, which can be attributed to its positive influence in retarding the 1st stage decomposition of FPUF. However, the THR of flame retarded FPUF compared to the virgin FPUF (Table 5, Figure S15) did not show any significant decrease, indicating that phosphoramidates and phosphonamidates additives act primarily during the gas phase. On the other hand, BA-DOPO and EDA-DPP showed a decrease in HRC ( Figure S17), revealing that the pyrolysis of these two additives interfere with the second-stage pyrolysis of FPUF, thus reducing the degradation rate of the main chain polyether, allowing these two additives to act somehow in the condensed phase. However, the absence of the char residue clearly shows that these additives cannot create thermally stable residue, and thus may primarily work in the gas phase.

Conclusions
The synthesis of various P(O)-N containing phosphoramidates and phosphonamidates from their corresponding H-Phosphonates and H-Phosphinates, respectively, was investigated using one pot two-step process with N-chlorosuccinimide as a versatile and environmentally friendly chlorinating agent. N-chlorosuccinimide as a chlorinating agent showed high selectivity and high yield, affording succinimide as a benign by-product. Accordingly, this synthetic methodology can be considered as an eco-friendly approach. However, more studies are needed to evaluate the economic and commercial benefits of this method. In this study, we mainly focused on in situ synthesis of P(O)-Cl to investigate the feasibility for production of corresponding P(O)-N containing derivatives and evaluating their influence in the thermal decomposition and possible mode of action in FPUF. In this work we used amines such as propylamine, allylamine, benzylamine, diethylamine, and ethylenediamine and converted them in situ to their corresponding P(O)-N derivatives. The final products were obtained in high yields and high purity even in the presence of succinimide, which can be removed by using a suitable purification procedure. Preliminary investigations regarding their influence in thermal decomposition of FPUFs was evaluated using TGA and MCC and their possible mode of flame retardant action was proposed. Comparing the TGA data of pure FPUF with those containing phosphoramidate or phosphonamidate showed that they have an influence on the decomposition pathways of the FPUF, which is generally limited to the first-stage decomposition of the FPUF, where a catalyzed depolymerization of the urethane linkages takes place. All solid P(O)-N compounds influenced the first stage decomposition of FPUFs which resulted in an accelerated decomposition or temporary stabilization of this stage. However, the liquid P(O)-N derivatives volatilize at an earlier stage and could be active in the gas phase. In addition, they also work in condensed phase via acid-catalyzed decomposition for FPUFs. THR, HRC, and PHRR data obtained in MCC experiments further highlighted the potential gas phase flame inhibition of these P(O)-N substances, with a limited condensed phase mode of action. The aromatic-containing FR additives were noticed to act in the condensed phase, but resulted in negligible char residue at 800 • C. This is attributed to the complete thermal decomposition of FPUF and absence of any dehydration process which prevents formation of the voluminous char. In general, use of N-chlorosuccinimide as a chlorinating agent is a relatively simple and clean synthetic approach to produce P(O)-N containing phosphorus derivatives. Therefore, future work will involve synthesis of phosphates, phosphonates, and phosphinates by using this synthetic approach. In addition detailed fire evaluation of all synthesized phosphorus compounds and a comprehensive mechanistic study will also be carried out and correlated to the data obtained in this research.
Author Contributions: K.A.S. and F.F. synthesized and characterized the P(O)-N containing substances. D.R. performed the NMR measurements and analyzed the data. K.A.S. and S.G. designed the project and prepared the first version of the manuscript and all authors contributed to the final version.

Funding:
The NMR Hardware used for measurements in this work was partially granted by the Swiss Science Foundation (SNF, grant no. 150638).