Synthesis and properties of novel polyimide fibers containing phosphorus groups in the main chain

Abstract

A series of polyamic acid copolymers (co-PAAs) containing phosphorous groups in the main chain were synthesized using different ratios of two diamines, i.e., bis(3-aminophenyl)methyl phosphine oxide (DAMPO) and 4,4′-oxydianiline (ODA), with 3,3′,4,4′-biphenyltetracarboxylic dianhydride (s-BPDA) by polycondensation in N,N′-dimethyacetamide (DMAc). The co-PAA solutions were spun into fibers by a dry-jet wet spinning process, and then polyimide copolymer (co-PI) fibers were obtained by thermal imidization. ATR-FTIR spectra and elemental analysis confirmed the chemical structure of PAA and PI fibers. The as-prepared PI fibers have smooth and dense surface as well as uniform diameter. Compared with the blank PI-0, the T_g values of co-PI fibers increased considerably with the increase in DAMPO content. TGA results indicated that the co-PI fibers possessed good thermal stability up to 510 °C and a residual char yield of up to 61% at 850 °C. All co-PI fibers exhibited excellent elongation, and their tensile strength and modulus can reach 0.9 GPa and 14.97 GPa when the molar ratio of DAMPO/ODA was 6/4 and the draw ratio was 3.0. The relationship between microstructure and mechanical property is also discussed.

---

Introduction

Many types of high-performance fibers, ranging from polymeric to carbon, boron, and ceramic fibers, are commercially available. However, each type of fiber has certain additional advantages. For example, in oxygen-free environments, carbon fibers can retain their strength at extremely high temperature, whereas polymeric fibers are much lighter than carbon fibers.

Among high-performance fibers, aromatic polyimide (PI) fibers are known for their excellent resistance to heat, thermal stability, chemical and solvent resistance, and radiation stability. Owing to these outstanding properties, aromatic PI fibers are widely used in manufacturing, microelectronic, engineering, and aerospace industries. Since the earliest work on PI fibers was reported by Irwin,¹ a few decades later, the preparation of high-performance PI fibers was mainly carried out by Japanese^2–10 and Soviet researchers.^11–14 Until recent years, commercial PI fibers include not only P84® fibers (originally introduced in a patent by the Upjohn company, developed by Lenzing AG, and now offered by Inspec Fibers/Degussa), but also Yilun® and ASPI™ fibers produced by Hipolyking and Aoshen in China. However, other kinds of PI fibers are still in the laboratory stage.

Two methods are usually employed for the fabrication of PI fibers. First, PI solution is directly used to spin PI fibers into a coagulation bath. The resulting PI fibers produced by a one-step method usually possess a high tensile strength and initial modulus.^15–17 Cheng¹⁷ et al. reported a kind of PI fiber, which was synthesized by 3,3′,4,4′-biphenyltetracarboxylic dianhydride (s-BPDA) and 2,2′-bis(trifluoromethyl)-4,4′-diaminobiphenyl (PFMB), had a tensile strength of 3.3 GPa and an initial modulus of 130 GPa. However, the solubility of PI solutions and the toxicity of solvents hinder the large-scale production for the one-step method.^18–20 Second, a typical two-step technique is primarily adopted in the preparation of PI fibers.^21–23 A precursor polymer solution, such as polyamic acid (PAA) or poly(amic ester) (PAE), is spun and followed by thermal or chemical imidization. Using the two-step method to fabricate PI fibers can overcome the solubility problem of PI solutions. However, the formation of voids from volatilization of small molecules by thermal imidization reduces the mechanical properties of PI fibers. Park and Farris reported²¹ that pyromellitic dianhydride/oxidianiline (ODA) fibers had tensile strength of 0.4 GPa and initial modulus of 5.2 GPa, respectively. Clair²⁴ reported that the tensile strength and initial modulus of 3,3′,4,4′-benzophenonetetracarboxylic dianhydride/ODA fibers are 0.19 GPa and 3.6 GPa, respectively. Recently, Liu²⁵ and Wu²² reported that controlling the aromatic rigid units (AAQ monomer), hydrogen bonding^26,27 (between AAQ and cyclic imide units), and draw ratios can enhance the tensile strength and modulus to as high as 2.8 GPa and 115 GPa, respectively.

PI fibers are commonly applied in high-temperature resistant filters and fireproof materials. However, in some special applications, such as spaceships, space suits, and satellite ropes, commercial PI fibers cannot meet the needs of flame retardancy and high-temperature resistance. Meanwhile, phosphorus materials exhibit good flame retardancy and thermal stability. Here, phosphorous-containing monomer, i.e., bis(m-aminophenyl)methyl phosphine oxide (DAMPO), was synthesized and a series of co-PI fibers were prepared by adjusting the percentage of DAMPO content through a two-step approach. The high-temperature resistance and mechanical properties of as-prepared PI fibers were discussed.

Experimental

Chemicals and materials

4,4′-Oxydianiline (4,4′-ODA, >99.5%) and 3,3′,4,4′-biphenyltetracarboxylic dianhydride (s-BPDA) were purchased from Shanghai Research Institute of Synthetic Resins, and s-BPDA was dried in vacuum at 260 °C overnight prior to use. Triphenylphosphine, N-methlypyrrolidinone (NMP), N,N-dimethylformamide (DMF), tetrahydrofuran (THF), trifluoroacetic acid (TFA), LiCl, CCl₄, dimethyl sulfoxide (DMSO) and EtOH were purchased from Shanghai Darui Fine Chemical Co., Ltd. without further purification. N,N′-Dimethyacetamide (DMAc, analytical purity of ≥99.5%) was purchased from Tianjin Fine Chemical Co., Ltd. and used as received. All other chemicals were obtained and without further purification.

Synthesis of bis(m-aminophenyl)methyl phosphine oxide (DAMPO)

DAMPO was synthesized according to a modified method.^28–30 The synthetic route is illustrated in Scheme 1 and the details are as follows.

Scheme 1 Synthetic route of bis(m-aminophenyl)methyl phosphine oxide (DAMPO).

Synthesis of methyl triphenylphosphonium iodide (MTI) 1 (ref. 31–33). CH₃I (7.5 mol, 1064 g) was slowly added into a chilled solution of 7.5 mol Ph₃P (1967.18 g) in 6 L of toluene, which was reacted in a 10 L round bottom flask in the ice bath. The mixed solution was stirred for more than 6 h at 0 °C, and then filtered. The residue was washed with fresh toluene and dried in the oven. White powder (Ph₃PCH₃I, MTI) was obtained with more than 98% yield. FT-IR (KBr): 2920, 2869 cm⁻¹ (–CH₃). ¹H NMR (400 MHz, DMSO-d₆) δ (ppm), 7.92 (m, 3H), 7.88–7.73 (m, 12H), 3.24 (d, J = 14.5 Hz, 3H). ³¹P NMR (162 MHz, DMSO-d₆) δ (ppm) 22.30. ¹³C NMR (101 MHz, DMSO-d₆) δ (ppm) 134.71, 133.19, 133.08, 130.05, 129.93, 120.18, 119.31, 7.75, and 7.20. Theoretical calculation for elements of C₁₉H₁₈IP was C: 56.46%, H: 4.49%, P: 7.66%. Found elemental contents were C: 56.89%, H: 4.48%, P: 7.60%.

Synthesis of methyl biphenylphosphine oxide (MDPO) 2 (ref. 34). MTI 1, (4.37 mol, 1767 g) was transferred into a 10 L round bottom flask equipped with a mechanical stirrer, a reflux condenser, and a nitrogen inlet. Sodium hydroxide (64.29 mol, 2571.43 g) and water (6250 mL) were then added into the flask. The mixture was stirred and heated to reflux at about 100 °C for ∼6 h, which produced a cloudy suspension. A clear layer was formed (benzene layer) on the top of the reaction mixture. The reaction was monitored by TLC (dichloromethane : methanol = 9 : 1). The mixture was extracted with dichloromethane, washed for several times with water, dried with magnesium sulfate, and then the solvent was removed. The mixture was finally dried in a vacuum oven for 24 h. The product was obtained as a white solid with 95% yield, mp. 114–116 °C. FT-IR (KBr): 1172 cm⁻¹ (P=O). ¹H NMR (400 MHz, DMSO-d₆) δ (ppm) 7.83–7.73 (m, 4H), 7.58–7.44 (m, 6H), 2.04 (d, J = 13.4 Hz, 3H). ³¹P NMR (162 MHz, DMSO-d₆) δ (ppm) 27.27. ¹³C NMR (101 MHz, DMSO-d₆) δ (ppm) 135.62, 134.64, 131.39, 130.19, 130.10, 128.58, 128.46, 16.25, 15.52. Theoretical calculation for elements of C₁₃H₁₃OP was C: 72.21%, H: 6.06%, P: 14.33%. Found elemental contents were C: 72.35%, H: 5.93%, P: 14.41%.

Synthesis of bis(m-nitrophenyl)methyl phosphine oxide (DNMPO) 3 (ref. 28). Methyl biphenylphosphine oxide (MDPO) 2 (2 mol, 432.44 g) was charged into a 10 L round bottom three-neck flask equipped with a dripping funnel and a mechanical stirrer. The concentrated sulfuric acid (2000 mL) was added into the flask carefully and the system was allowed to cool to 0 °C in the ice bath with stirring until the solid dissolved. The mixture of concentrated sulfuric acid (2000 mL) and fuming nitric acid (4000 mL) was added dropwise with the dripping funnel. The reaction was stirred for 4 h at 0–5 °C and maintained further for 24 h at room temperature. The reaction was monitored by TLC (dichloromethane : methanol = 9 : 1). The reaction mixture was poured into the ice water, and then the precipitated solid was extracted with dichloromethane, washed with aqueous sodium bicarbonate solution until neutral pH was reached, and then the dichloromethane organic layer was dried with magnesium sulfate. The solvent was removed and the obtained product recrystallized with absolute glycol ether to obtain light yellow solid with 93% yield, mp. 245–248 °C. FT-IR (KBr): 1530, 1351 cm⁻¹ (–NO₂) (as shown in Fig. S1†). ¹H NMR (400 MHz, DMSO-d₆) δ (ppm) 8.72–8.52 (m, 2H), 8.40 (dd, J = 8.2, 2.3 Hz, 2H), 8.29 (dd, J = 11.0, 7.6 Hz, 2H), 7.84 (td, J = 7.9, 2.8 Hz, 2H), 2.32 (d, J = 14.1 Hz, 3H) (as shown in Fig. S2 blue†). ³¹P NMR (162 MHz, DMSO-d₆) δ (ppm) 26.62. ¹³C NMR (101 MHz, DMSO-d₆) δ (ppm) 147.92, 147.79, 136.99, 136.67, 136.57, 136.01, 130.75, 130.63, 126.62, 125.01, 124.90, 15.59, and 14.86 (as shown in Fig. S2 black†). Theoretical calculation for elements of C₁₃H₁₃OP was C₁₃H₁₁N₂O₅P was C: 50.99%, H: 3.62%, N: 9.15%, P: 10.12%. Found elemental contents were C: 51.15%, H: 3.71%, N: 9.16%, P: 10.09%. (The spectra of FTIR, H¹ and C¹³ NMR have been supported in ESI†).

Synthesis of bis(m-aminophenyl) methyl phosphine oxide (DAMPO) 4 (ref. 28). DNMPO (1.385 mol, 424 g) and ethanol (3000 mL) were charged into a 5 L round bottom three-neck flask fitted with a dripping funnel, a mechanical stirrer, a reflux condenser, a nitrogen inlet, and an oil bath. The 10% Pd/C was added before heating the mixture. Hydrazine (8.308 mol, 415.9 g) was dropped into the solution carefully using a dripping funnel when the mixture was heated to 90 °C. The black mixture was then refluxed rapidly at 90 °C. The reaction was monitored by TLC. Finally, the black mixture was filtered using Buchner funnel via diatomite to afford ethanol solution. The solvent was removed, and the obtained product was recrystallized by absolute ethanol with 96% yield, mp. 155–156 °C. FT-IR (KBr): 3367 and 3220 cm⁻¹ (N–H) (as shown in Fig. S1†). ¹H NMR (400 MHz, DMSO-d₆) δ (ppm) 7.12 (td, J = 7.8, 3.7 Hz, 1H), 6.96–6.87 (m, 1H), 6.80 (dd, J = 11.4, 7.4 Hz, 1H), 6.73–6.65 (m, 1H), 5.35 (s, 2H), 1.82 (d, J = 13.1 Hz, 1H) (as shown in Fig. S3 blue†). ³¹P NMR (162 MHz, DMSO-d₆) δ (ppm) 28.17. ¹³C NMR (101 MHz, DMSO-d₆) δ (ppm) 148.79, 148.65, 136.16, 135.18, 129.10, 128.97, 117.10, 117.00, 116.46, 115.19, 115.08, 16.37, and 15.65 (as shown in Fig. S3 black†). Theoretical calculation for elements of C₁₃H₁₅N₂OP was C: 63.41%, H: 6.14%, N: 11.38%, P: 12.58%. Found elemental contents were C: 63.61%, H: 6.18%, N: 11.48%, P: 12.61%. (The spectra of FTIR, H¹ and C¹³ NMR have been supported in ESI†).

Preparation of PAA solutions

A series of PAAs solutions with different phosphorus contents were successfully synthesized by adjusting molar ratios of DAMPO/ODA. For example, in the reaction with co-PAA at a molar ratio of 2/8, 192.23 g (0.96 mol) ODA and 59.10 g (0.24 mol) DAMPO were charged into a flask equipped a stirrer, a nitrogen inlet, and a thermometer. The mixture was stirred in DAMc of 3735 mL. The s-BPDA (353.06 g, 1.20 mol) was quickly added into the mixture when the solid of diamines was completely dissolved (15% solid content, w/w). The isotropic solution was obtained with stirring for another 24 h at room temperature. Homo-PAA (DAMPO/ODA = 0/10) and co-PAAs (DAMPO/ODA = 1/9 to 6/4) were prepared. We found that gel particles were formed when the molar ratio of DAMPO/ODA was higher than 6/4, which was harmful for extrusion of fibers from spinnerets. The proportions of the PAA solutions and the corresponding inherent viscosities were listed in Table 2. The polymerization route of the PAA solutions with phenyl phosphine oxide (PAA-PPO) was shown in Scheme 2.

Table 1 Elemental analysis of PI fibers

No.	Theoretical value (%)			Experimental value (%)
	C	H	N	C	H	N
PI-0	73.35	3.08	6.11	73.25	3.20	6.09
PI-1	72.88	3.11	6.05	72.34	3.14	5.99
PI-2	72.42	3.15	5.99	72.02	3.19	5.94
PI-3	71.97	3.18	5.93	71.39	3.23	5.81
PI-4	71.53	3.21	5.87	70.56	3.26	5.75
PI-5	71.09	3.25	5.82	69.31	3.30	5.59
PI-6	70.67	3.28	5.76	69.37	3.21	5.56

---

Table 2 Inherent viscosities of PAA solutions and mechanical properties of PI fibers

Polymer No.	Diamine ratio (m : n)	Draft^a/draw^b ratio	ηinh (dL g^−1)	Mechanical properties
				Strength (GPa)	Modulus (GPa)	Elongation (%)
a The ratio of the take-up speed to the extrusion speed.b The ratio of after-drawing speed to before-drawing speed passing the heat tube.
PI-0	0 : 10	3.8/1.6	1.77	1.02 ± 0.05	16.94 ± 1.11	13.23 ± 0.63
PI-1	1 : 9	3.8/1.6	1.60	0.91 ± 0.04	16.95 ± 0.96	15.06 ± 1.52
PI-2	2 : 8	3.8/1.6	1.59	0.82 ± 0.06	17.38 ± 1.56	16.94 ± 0.83
PI-3	3 : 7	3.8/1.6	1.42	0.63 ± 0.03	15.77 ± 0.85	20.85 ± 1.63
PI-4	4 : 6	3.8/1.6	1.40	0.59 ± 0.03	12.71 ± 0.89	29.12 ± 2.26
PI-5	5 : 5	3.8/1.6	1.43	0.49 ± 0.06	10.82 ± 1.26	51.84 ± 2.61
PI-6	6 : 4	3.8/1.6	1.27	0.43 ± 0.03	8.88 ± 0.81	61.64 ± 3.74
PI-6′	6 : 4	3.8/2.0	1.27	0.57 ± 0.05	9.81 ± 1.47	21.10 ± 1.96
PI-6′′	6 : 4	3.8/3.0	1.27	0.90 ± 0.04	14.97 ± 1.42	18.29 ± 1.21

---

Scheme 2 Synthetic route of PAA-PPO solutions (m/n = 0/10 to 6/4).

Preparation of co-PAA fibers and co-PI fibers

The above co-PAA solutions were spun through a dry-jet wet spinning method to form co-PAA fibers. Then, the as-prepared co-PAA fibers were subjected to thermal imidization to produce co-PI fibers, as shown in Scheme 3.

Scheme 3 Two-step procedure for preparing co-PI fibers.

The PAA-PPO solution was filtered and degassed at room temperature prior to use. The process of preparation of fibers for spinning was shown in Fig. 1. Fibers were spun (3) via the dry-jet wet spinning method with 15% solid content into the coagulation bath (4) which was consisted of water and DMAc in the specified proportion. The solidifying filament entered into a second bath (6), and then into a heating tube (8 and 9), finally collected on the winder (10). The collected fibers were called co-PAA fibers. The co-PAA fibers dried in the oven for about 12 h, and then drew in a heating tube to produce co-PI fibers.

Fig. 1 Fabrication flowchart of PAA fibers.

Characterization

Melting point was determined on a Thomas Hoover capillary melting point apparatus. Glass transition temperature (T_g) was determined by differential scanning calorimetry (DSC) with TA Q100 at a heating rate of 5 °C min⁻¹. The inherent viscosity (η_inh) of PAA solution was measured with an Ubbelohde viscometer at a concentration of 0.5 dL g⁻¹ in DMAc at 30 °C. Fourier transform infrared (FTIR) spectrum was recorded on a VERTEX 70 spectrometer with the scanning wavenumber in the range of 400–4000 cm⁻¹. The measurement of mechanical properties was carried out on XQ-1 instrument with ASTM standard (D3379-75, edition 1987) at a cross-head speed of 20 mm min⁻¹. More than 10 monofilaments for one sample were tested and the average data were used to characterize the mechanical property of the sample. TA-Q50 instrument was employed to analyze the thermal stability of the fibers at a heating rate of 10 °C min⁻¹ under nitrogen atmosphere. Dynamic mechanical analysis (DMA) was performed on the Rheometric scientific DMTA-V in 1 Hz at a heating rate of 10 °C min⁻¹ in the range of 50–400 °C. The surface morphology of PI fibers was determined by XL-30 and scanning electron microscope (SEM). X-ray scattering (SAXS) profiles were obtained by Nanostar-u (Bruker AXS INC). The wavelength was 0.154 nm. A CCD X-ray detector (HI-STAR) was used at a distance of 1062 nm from the sample. A software package FIT2D processed the SAXS 2D images.

Results and discussion

Structure characterization of PAA and PI fibers containing-phosphorus groups

The FTIR spectra of PAA fibers was shown in Fig. 2. The peaks at 3300 and 2948–2848 cm⁻¹ were related to the absorptions of COOH and NH₂, CH₃ groups, respectively. Additionally, peaks at 1712, 1652, and 1533 cm⁻¹ corresponded to the absorptions of C=O(COOH), C=O(CONH) and C–NH, respectively. Peaks at 788–684 cm⁻¹ corresponded to the absorption of P–C in PPO groups. Furthermore, peaks of CH₃ and P–C became stronger with increasing DAMPO content, as shown in Fig. 2.

Fig. 2 FTIR spectra of PAA fibers.

However, in FTIR spectra of PI fibers (Fig. 3), a broad band at 3137–3000 cm⁻¹ corresponded to NH₂⁺ group. The band at 2948–2848 cm⁻¹ was related to the absorption of CH₃ groups. Peaks at 1774 and 1722 cm⁻¹ belonged to the C=O group of asymmetric and symmetric stretching vibrations of the aromatic imide ring, respectively. The peak at 1360 cm⁻¹ was attributed to the C–N stretching vibration of imide ring. The peak at 734 cm⁻¹ implied the appearance of imide ring bending. The peaks at 1712, 1652, and 1533 cm⁻¹ from PAA fibers cannot be found in Fig. 3. The characteristic peaks of PI fibers emerged with 1774, 1722, 1360, and 734 cm⁻¹. The band of CH₃ group at 2948–2848 cm⁻¹ increased in intensity with increasing DAMPO content. The abovementioned results indicated that PAA fibers were successfully converted to PI fibers. The FTIR spectra were similar to those reported in literature.³⁵

Fig. 3 FTIR spectra of PI fibers.

The elemental analysis results of PI fibers were presented in Table 1. With increasing DAMPO content, carbon content decreased, whereas both hydrogen and nitrogen contents increased. As shown in Table 1, the C, H and N elemental contents for PI-0 were in good agreement with the theoretical values. With increase of DAMPO content, the C and N elemental contents decreased. For example, for PI-2 and PI-6, their C contents were found to be 72.02% and 69.37%, only slightly lower than the theoretical values (error less than 0.5% and 1.8%, respectively), the N contents were 5.94% and 5.56% (error less than 0.8% and 3.5%, respectively). It was noticed that the PI-6 exhibited bigger errors in C and N contents. Possibly, the difference between the experimental and theoretical values was due to incomplete burning for these containing-phosphor polyimides. Faghihi²⁹ has reported the results of elemental analysis of poly(amide-imide)s containing phosphine oxide. The similar results showed that the experimental values in C and N contents were lower than theoretical values of the containing-phosphor poly(amide-imide)s. Although experimental errors affected the results to some extent, the trend of C, H, and N elemental contents decreased with containing-phosphor monomers like those of theoretical values.

Inherent viscosity of PAA solutions

For the PAA copolymer solutions, the molar ratios of DAMPO/ODA varied from 0/10 to 6/4. Their inherent viscosities, as listed in Table 2, were as following: 1.77, 1.60, 1.59, 1.42, 1.40, 1.43, and 1.27 dL g⁻¹, corresponding to molar ratios of 1/10, 1/9, 2/8, 3/7, 4/6, 5/5, and 6/4, respectively. The value of viscosity decreased with increasing of DAMPO content. The polymerization procedure of PAA is a nucleophilic reaction, in which diamine is the nucleophilic reagent. The reaction rate increases with increase of the charge density of the amino nitrogen.³⁶ DAMPO and ODA have different structures. The oxygen atom in ODA is the charge donor that results in increasing charge density of amino nitrogen. Thus, the reactivity of ODA is higher. In the case of DAMPO, the electron-withdrawing group (P=O) reduces the charge density of the amino nitrogen. Subsequently, the nucleophilic attack of DAMPO is reduced. Thus, increasing DAMPO content will decrease the viscosity of the PAA solutions to some extent.

Surface morphology of PI fibers

The surface morphology of PI fibers (a–g) was observed by SEM, as shown in Fig. 4. All the fibers exhibited smooth surfaces, and the diameters were uniform. The average diameter was about 14 μm. The dense surfaces of all fibers came from the quick dual-diffusion process in the coagulation bath and the uniform diameters with process controlling.

Fig. 4 Surface morphologies of fibers ((a–g) corresponding to PI-0 to PI-6).

Glass transition temperatures of PI fibers

The T_g of PI fibers was measured by both DSC and DMA. The DSC curves were shown in Fig. 5. Their T_gs were 261, 265, 273, 279, 290, 295 and 297 °C, which corresponded to molar ratios of DAMPO/ODA 0/10, 1/9, 2/8, 3/7, 4/6, 5/5, and 6/4, respectively. Definitely, the T_gs of co-PI fibers were higher than the blank one (PI-0) and increasing DAMPO content lead to an obvious increasing of T_gs. It was considered that higher T_g of co-PI fibers was related to the rigid structure of DAMPO (without ether group in main chain) than ODA, thereby requiring more energy (higher temperature) to move the chain segments. The monomer, DAMPO, contains more rigid structure than ODA. Hence, increasing of DAMPO content will result in higher T_gs of copolymer. Furthermore, the T_gs of PI fibers (DAMPO/ODA 6/4) with different draw ratios (1.6, 2.0, and 3.0) were also studied in Fig. 5(b). Results indicated that increasing the draw ratio has no influence on glass transition temperature.

Fig. 5 DSC spectra of PI fibers with DAMPO ratios (a) and draw ratios (b).

T_gs were also obtained using DMA, as presented in Fig. 6(a–d). In the range of 250–310 °C, the α relaxation of PI fibers was observed and corresponded to the glass transition temperature. The T_gs of PI-0, PI-1, PI-2, PI-3, PI-4, PI-5, and PI-6 were 260, 264, 271, 276, 285, 293 and 296 °C, respectively, indicating that the introduction of DAMPO efficiently improved the glass transition temperature. The rigid structure of DAMPO leaded to higher T_g. Meanwhile, the α relaxation of PI fibers nearly did not change with draw ratio (1.6, 2.0, and 3.0) of PI-6 fiber, as shown in Fig. 6(b and d). The abovementioned results were consistent with those of DSC.

Fig. 6 tan δ (a, b) and storage modulus (c, d) as a function of temperature.

Thermal stability of PI fibers

The thermal stability of the PI fibers was characterized by thermogravimetric analysis (TGA), and their TGA and derivative curves were shown in Fig. 7. Nearly no weight loss was observed from room temperature to 500 °C. In Fig. 7(b), PI-0 fiber only exhibited the highest decomposition rate at 600 °C. However, the co-PI fibers showed two quicker decomposition rates: the first one at 522 °C was associated with the decomposition of P–C bond, whereas the second one at 600 °C was related to the decomposition of PI backbone. With increasing DAMPO content, the decomposition started earlier because of easier degradation of P–C bond. However, the phosphorous element had a positive effect on the flame retardant. The temperatures of 5% and 10% weight loss (T_5% and T_10%, respectively) were listed in Table 3. The T_5%s decreased from 559 °C to 514 °C, and T_10%s also decreased from 583 °C to 523 °C along with DAMPO content. The residues at 850 °C were 58.9, 64.4, 66.7, 63.0, 62.0, 61.3, and 61.9% corresponding to DAMPO from 0% to 60%, which indicated that introduction of DAMPO facilitated more char yield, i.e., better flame retardancy.

Fig. 7 TGA and their derivative curves of co-PI fibers.

Table 3 TGA data of PI fibers

Sample	Ratio of diamine DAMPO/ODA	T5%^a (°C)	T10%^b (°C)	Residue at 850 °C (wt%)
a Corresponding temperature at 5% weight loss.b Corresponding temperature at 10% weight loss.
PI-0	0/10	559	583	58.9
PI-1	1/9	569	592	64.4
PI-2	2/8	534	568	66.7
PI-3	3/7	515	535	63.0
PI-4	4/6	521	533	62.0
PI-5	5/5	515	527	61.3
PI-6	6/4	514	523	61.9

---

Mechanical properties of PI fibers

For polymeric fibers, the mechanical property is considered to be a significant parameter, particularly in engineering applications. Fig. 8 and 9 showed the variation of mechanical properties with DAMPO content and draw ratio, and the details were listed in Table 2. From PI-0 to PI-6, the tensile strengths were 1.02, 0.91, 0.82, 0.63, 0.59, 0.49, 0.43 GPa and the tensile modulus were 16.94, 16.95, 17.38, 15.77, 12.71, 10.82, 8.88 GPa. Elongations increased obviously from 13.23% to 61.64%. Both strength and modulus decreased with DAMPO content. The abovementioned results indicated that introduction of DAMPO efficiently improved the elongation of co-PI fibers. In particular, the elongation of PI-6 was nearly five times higher than that of PI-0. However, the draw ratio also influenced the mechanical property. In Fig. 9, PI-6′ and PI-6′′ had higher draw ratios of 2.0 and 3.0 comparing with PI-6 (draw ratio = 1.6). Fibers with higher draw ratio exhibited higher strength, modulus, and lower elongation. The abovementioned results showed that 60% DAMPO co-PI fiber can also achieve adequate mechanical property like PI-0 fiber merely by increasing the draw ratio.

Fig. 8 Variation of mechanical properties with DAMPO content for PI fibers.

Fig. 9 Variation of mechanical properties with draw ratio of co-PI fibers.

Molecular chains packing

For aromatic PI fibers, drawing can lead to higher orientation of molecular chains along the axial direction. During the heat drawing process, polymer chains rearranged themselves to form an improved molecular alignment. In Fig. 10(a), a series of broad peaks in XRD patterns in the range of 15–30° can be observed and with increase of DAMPO content, the shape of XRD pattern was similar, thereby implying typical amorphous polymer chains in the equator direction. In Fig. 10(b), the resulting fibers PI-6, PI-6′, and PI-6′′ (with different draw ratios) were also considered amorphous state and increasing the draw ratios doesn't seem to influence the condensed state. Possibly, the molecular chain rearrangement during drawing locally forms small-ordered structure and higher orientation, which ultimately contribute to increased tensile strength and modulus.

Fig. 10 XRD patterns of co-PI fibers with equator direction. (a) Different DAMPO contents; (b) different draw ratios.

Micromorphology

During dry-jet wet-spinning process, the coagulation of fibers is a dual diffusion process, namely, solvent diffuses into the bath, and the coagulant diffuses into the fibers. However, the inner and external forces cannot always reach equilibrium because of different concentration gradients for the solvent and coagulation, leading to defects like micro-voids inside the fibers. Some researchers^37–41 have suggested that the elongated streak in the SAXS patterns was the scattering by long thin voids and parallel to the fiber axis. In this study, the SAXS patterns of PI-6 with different draw ratios are presented in Fig. 11. As a common feature, it can be clearly seen that the scattering patterns exhibit a sharp and elongated streak in intensity because of the needle-shaped micro-voids and aligned parallel to the fiber direction. When the draw ratios increase, the elongated shapes become sharper, thereby illustrating the presence of morphological changes in the fibers.

Fig. 11 2D SAXS patterns of PI-6 fiber with draw ratio of 1.6 (a), 2.0 (b) and 3.0 (c).

The average length and the misorientation of micro-voids from the fiber axis are calculated based on the method of Ruland.⁴² Zhang found that the azimuthal scans of the streak at different scattering vectors can be described by the Lorentzian/Cauchy- type functions.⁴³ The micro-voids length (L) and the misorientation width (B_Φ) can be calculated from the following equation:

where B_obs is the full width at half maximum of the azimuthal profile from the streak fitted with a Lorentzian function. q (q = 4π sin θ/λ) is the scattering vector. 2θ is the scattering angle. λ is the X-ray wavelength.

Fig. 12 shows the Ruland plot of PI-6 with different draw ratios and related results are presented in Table 4. With increasing draw ratio from 1.6 to 2.0 and 3.0, the length of the micro-void L decreases from 167 nm to 162 nm and 153 nm, respectively. Consequently, the micro-void misorientation (B_Φ) decreases greatly with draw ratios, from 12.9° in draw ratio of 1.6 to 7.7° in draw ratio of 3.0. Fig. 13 illustrates the micro-void in PI fibers. From the abovementioned analysis, increasing the draw ratio increases the orientation of amorphous state, which contributes to increased strength and modulus. Thus, increasing draw ratios is an efficient approach to improve the mechanical properties of PI fibers.

Fig. 12 Ruland plot of PI-6 with different draw ratios, (a) 1.6, (b) 2.0, (c) 3.0.

Table 4 The micro-voids parameters of PI fibers

Draw ratio	L (nm)	BΦ
		(rad)	(°)
1.6	167	0.225	12.9
2.0	162	0.221	12.7
3.0	153	0.134	7.7

---

Fig. 13 Schematic diagram of the micro-void in the orientated fiber.

Solubility of fibers

The solubility of PI fibers in nine solvents (NMP, DMAc, DMF, THF, TFA, DMF/LiCl, CCl₄, DMSO, and EtOH) was investigated through 0.001 g of PI fibers in 20 mL of different solvents. As shown in Table 5, all the fibers showed high resistance to these solvents, including polar solvents such as EtOH, NMP, DMF, non-polar solvents such as CCl₄. Among the fibers, only PI-4 and PI-5 partially dissolved in TFA and only PI-6 sample can fully dissolve in TFA. All above indicate that the polyimide fibers we have made have a high resistance to solvent and the incorporation of DAMPO can improve solubility for polymers to some extent.

Table 5 Solubility of the PI fibers

Polymer No.	Diamine ratio (m : n)	Solvents^a
		NMP	DMAc	DMF	THF	TFA	DMF/LiCl	CCl4	DMSO	EtOH
a ++, wholly soluble; +−, partially soluble; −−, not soluble; NMP, xxx.
PI-0	0 : 10	−−	−−	−−	−−	−−	−−	−−	−−	−−
PI-1	1 : 9	−−	−−	−−	−−	−−	−−	−−	−−	−−
PI-2	2 : 8	−−	−−	−−	−−	−−	−−	−−	−−	−−
PI-3	3 : 7	−−	−−	−−	−−	−−	−−	−−	−−	−−
PI-4	4 : 6	−−	−−	−−	−−	+−	−−	−−	−−	−−
PI-5	5 : 5	−−	−−	−−	−−	+−	−−	−−	−−	−−
PI-6	6 : 4	−−	−−	−−	−−	++	−−	−−	−−	−−

---

Conclusions

A series of PAA copolymer solutions containing phosphorous groups in main chain were synthesized and then spun into PAA fibers via the dry-jet wet spinning processing. The PI fibers were obtained by thermal imidization of PAA fibers. The main results are summarized as follows. (i) The T_gs of PI fibers increase from 268 °C to 295 °C by DSC and from 262 °C to 297 °C by DMA with increasing phosphorous monomer (DAMPO). (ii) TGA results indicate a good thermal stability of up to 510 °C. The residual char amounts at 850 °C are in the range of 61–67% for all co-PI fibers, and these amounts are higher than that of PI fiber without phosphorous moieties. (iii) The co-PI fiber with DAMPO/ODA = 6/4 and draw ratio of 3.0 exhibits a tensile strength and an initial modulus of 0.90 GPa and 14.97 GPa, respectively, which means the co-PI fibers can also reach adequate mechanical property like PI-0 fiber only by increasing the draw ratio. (iv) All fibers are amorphous state, and increased strength and modulus with draw ratio are attributed to the increased orientation of amorphous regions in fibers. (v) All fibers we have made have a high resistance to solvent. In summary, the co-PI fibers exhibit good thermal stability and mechanical properties with the incorporation of phosphorous monomer in the main chain.

Acknowledgements

The authors thank the National Basic Research Program of China (973 program, Key Project: 2014CB643604) and the National Natural Science Foundation of China (Grants 51373164) for the support.

References

1. R. S. Irwin, Newark and C. E. Smullen, US Pat., 3,415,782, 1968.
2. S. Hara, T. Yamada and T. Yoshida, US Pat., 3,829,399, 1974.
3. M. Minami and M. Tanigchi, US Pat., 3,860,559, 1975.
4. P. T. Patterson and N. J. Cinnaminson, US Pat., 4,064,389, 1977.
5. H. Makino, Y. Kusuki and T. Harada, US Pat., 4,370,290, 1983.
6. K. Ohmura, I. Shibasaki and T. Kimura, US Pat., 4,377,652, 1983.
7. K. Nagaoka and Hirakata, US Pat., 4,448,957, 1984.
8. T. Ohta, Y. Yanaga and S. Hino, US Pat., 4,438,256, 1984.
9. I. Sasaki, H. Itatani and M. Kashima, US Pat., 4,247,443, 1984.
10. H. Itatani, T. Inaike and S. Yamamoto, US Pat., 4,568,715, 1986.
11. N. A. B. M. I. Adrova, L. Laius and A. P. Rudakovetc, Polyimides (A New Class of Thermally Stable Polyers), Technomic Publishing Co. Inc, Stanford, 1970.
12. M. M. Koton, F. S. Florinsky and L. N. Korzhavin, UK Pat., 2,025,311, 1980.
13. M. M. Koton, S. Y. Frenkel and N. A. Adrova, GB Pat., 1,183,306, 1968.
14. K. Mittal, Polyimides: synthesis, characterization, and applications, vol. 1 & 2, 1984.
15. T. Kaneda, T. Katsura, K. Nakagawa, H. Makino and M. Horio, J. Appl. Polym. Sci., 1986, 32, 3151–3176.
16. T. Kaneda, T. Katsura, K. Nakagawa, H. Makino and M. Horio, J. Appl. Polym. Sci., 1986, 32, 3133–3149.
17. M. Eashoo, Z. Q. Wu, A. Q. Zhang, D. X. Shen, C. Tse, F. W. Harris, S. Z. D. Cheng, K. H. Gardner and B. S. Hsiao, Macromol. Chem. Phys., 1994, 195, 2207–2225.
18. S. Z. D. Cheng, Z. Q. Wu, M. Eashoo, S. L. C. Hsu and F. W. Harris, Polymer, 1991, 32, 1803–1810.
19. M. Eashoo, D. X. Shen, Z. Q. Wu, C. J. Lee, F. W. Harris and S. Z. D. Cheng, Polymer, 1993, 34, 3209–3215.
20. H. B. Xiang, Z. Huang, L. Q. Liu, L. Chen, J. Zhu, Z. M. Hu and J. R. Yu, Macromol. Res., 2011, 19, 645–653.
21. S. K. Park and R. J. Farris, Polymer, 2001, 42, 10087–10093.
22. H. Q. Niu, S. L. Qi, E. L. Han, G. F. Tian, X. D. Wang and D. Z. Wu, Mater. Lett., 2012, 89, 63–65.
23. J. J. Chang, W. W. Liu, M. Y. Zhang, L. Cao, Q. Y. Ge, H. Q. Niu, G. Sui and D. Z. Wu, RSC Adv., 2015, 5, 71425–71432.
24. W. E. Dorogy and A. K. S. Clair, J. Appl. Polym. Sci., 1991, 43, 501–519.
25. G. Gao, L. Dong, X. Liu, G. Ye and Y. Gu, Polym. Eng. Sci., 2008, 48, 912–917.
26. H. Q. Niu, M. J. Huang, S. L. Qi, E. L. Han, G. F. Tian, X. D. Wang and D. Z. Wu, Polymer, 2013, 54, 1700–1708.
27. X. Y. Liu, G. Q. Gao, L. Dong, G. D. Ye and Y. Gu, Polym. Adv. Technol., 2009, 20, 362–366.
28. B. Tan, C. N. Tchatchoua, L. Dong and J. E. McGrath, Polym. Adv. Technol., 1998, 9, 84–93.
29. K. Faghihi, J. Appl. Polym. Sci., 2006, 102, 5062–5071.
30. K. Faghihi and M. Hagibeygi, Turk. J. Chem., 2007, 31, 65–73.
31. H. S. Gill, J. Labelled Compd. Radiopharm., 1995, 36, 425–430.
32. K. van Alem, G. Belder, G. Lodder and H. Zuilhof, J. Org. Chem., 2005, 70, 179–190.
33. W. K. Chin, M. D. Shau and W. C. Tsai, J. Polym. Sci., Polym. Chem., 1995, 33, 373–379.
34. L. H. B. Baptistella and A. M. Aleixo, Liebigs Ann. Chem., 1994, 1994, 785–789.
35. Y. Jin, G. Zeng, D. Zhu, Y. Huang and Z. Su, Chin. J. Appl. Chem., 2011, 28, 258–262.
36. V. A. Zubkov, M. M. Koton, V. V. Kudryavtsev and V. M. Svetlichnyi, Zh. Org. Khim., 1981, 17, 1682–1688.
37. J. Crawshaw and R. E. Cameron, Polymer, 2000, 41, 4691–4698.
38. X. M. Chen, C. Burger, D. F. Fang, D. Ruan, L. Zhang, B. S. Hsiao and B. Chu, Polymer, 2006, 47, 2839–2848.
39. X. Chen, C. Burger, F. Wan, J. Zhang, L. Rong, B. S. Hsiao, B. Chu, J. Cai and L. Zhang, Biomacromolecules, 2007, 8, 1918–1926.
40. W. O. Statton, J. Polym. Sci., 1956, 22, 385–397.
41. M. E. Vickers, N. P. Briggs, R. N. Ibbett, J. J. Payne and S. B. Smith, Polymer, 2001, 42, 8241–8248.
42. W. Ruland, J. Polym. Sci., Polym. Symp., 1969, 28, 143–151.
43. C. Q. Yin, J. Dong, D. B. Zhang, J. Y. Lin and Q. H. Zhang, Eur. Polym. J., 2015, 67, 88–98.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra02344d

---