Chain-end-functionalized polyphosphazenes via a one-pot phosphine-mediated living polymerization.

A simple polymerization of trichlorophosphoranimine (Cl3 P = N-SiMe3 ) mediated by functionalized triphenylphosphines is presented. In situ initiator formation and the subsequent polymerization progress are investigated by (31) P NMR spectroscopy, demonstrating a living cationic polymerization mechanism. The polymer chain lengths and molecular weights of the resulting substituted poly(organo)phosphazenes are further studied by (1) H NMR spectroscopy and size exclusion chromatography. This strategy facilitates the preparation of polyphosphazenes with controlled molecular weights and specific functional groups at the α-chain end. Such well-defined, mono-end-functionalized polymers have great potential use in bioconjugation, surface modification, and as building blocks for complex macromolecular constructs.

1. Introduction

Poly(organo)phosphazenes are an extremely versatile class of hybrid inorganic–organic polymers whose properties and functionality can be easily tuned by suitable organic side groups attached to the inorganic backbone,[1] and thus can be used to prepare a wide-array of interesting materials, ranging from superhydrophobic polymers[2] and high-performance elastomers,[3] to thermoresponsive polymers[4] and biodegradable polymers[5] for a multitude of biomedical applications,[6] such as tissue engineering[7] and advanced drug delivery systems.[8]

Although the precursor polymer, polydichlorophosphazene, is commonly obtained via a thermal ring-opening polymerization[9] with high molecular weights and broad molecular weight distributions, polydichlorophosphazene with controlled molecular weights can also be prepared via the room temperature living cationic polymerization of trichlorophosphoranimine (Cl3P = N−SiMe3) in solution.[10] Furthermore, tris(organo)phosphoranimines (R3P = N−SiMe3) can be used as end-groups in the presence of PCl5 to induce and/or quench the living polymerization of Cl3P = N−SiMe3, a method which has been used to prepare a number of macromolecular constructs including branched structures,[10a,11] and block copolymers.[12] Other studies, however, have shown that R3P = N−SiMe3 bearing alkoxy or aryloxy groups did not form a stable cationic species to induce the polymerization of Cl3P = N−SiMe3, and were only able to quench the polymerization.[13] The preparation of functional organophosphoranimines also often requires multiple step synthesis, for example, via a bromophosphoranimine precursor,[14] and the repeated vacuum distillations required for purification lead to extensive and time-consuming preparation and extensive handling of toxic phosphoranimines. In an alternative approach, it has been shown that, in the absence of PCl5, dichlorophosphoranes can also initiate the polymerization of Cl3P = N−SiMe3[15] and that this could be exploited to prepare organometallic-inorganic block copolymers.[16]

In an attempt to find a simple alternative to incorporate specific reaction sites at polyphosphazene end-groups, the aim of this study was to employ commercially available triphenylphosphines with (protected) functional groups which, after chlorination, could act as initiating species for the cationic polymerization of Cl3P = N−SiMe3 and thus provide a simple, one-pot method to obtain polyphosphazenes with a (protected) functional group at the α-chain end. Well-defined, mono-end-functionalized polymers have become indispensable building blocks for macromolecular engineering,[17] and thus the methods described here are intended to facilitate the preparation of complex architectures of polyphosphazenes.

2. Experimental Section

2.1. Living Polymerization with PCl5

PCl5 (3.58 mg, 17.2 μmol, 1 equiv.) was dissolved in a mixture of anhydrous CH2Cl2/CDCl3 (4/1) in an NMR tube sealed with a septum. Then, Cl3P = N−SiMe3 (53.80 mg, 0.24 mmol, 14 equiv.) in anhydrous CH2Cl2/CDCl3 was added via a syringe and the reaction was monitored by 31P{1H} NMR spectroscopy until the signal of Cl3P = N−SiMe3 completely disappeared. The theoretical number of chain length ncalc is 28, calculated assuming two initiator molecules are required per chain.[18]

2.2. Polymer 4a

4-(Diphenylphosphino)styrene 1a (7.50 mg, 26.0 μmol, 1 equiv.) was reacted with C2Cl6 (6.77 mg, 28.6 μmol, 1.1 equiv.) in 0.5 mL CD2Cl2 for 24 h at room temperature to yield 2a. The reaction was transferred to an NMR tube and sealed with a septum. First, 1 equiv. of Cl3P = N−SiMe3 (5.84 mg, 26.0 μmol), then 25 equiv. Cl3P = N−SiMe3 (0.15 g, 0.65 mmol) were added via a syringe to yield polymer 4a.

2.3. Polymer 4b

4-Diphenylphosphanyl benzoic acid-2-(trimethylsilyl) ethyl ester 1b (50 μL of 0.5 m in THF, 10.16 mg, 25.0 μmol, 1 equiv.) and C2Cl6 (6.51 mg, 27.5 μmol, 1.1 equiv.) were stirred for 72 h at room temperature. The reaction mixture of 2b was transferred to an NMR tube followed by the addition of Cl3P = N−SiMe3 (0.15 g, 0.65 mmol, 26 equiv.).

Both reactions (4a and 4b) were followed with 31P{1H} NMR spectroscopy until monomer consumption was complete (up to NMR detection limits). The ratio of the monomer concentration, measured at specific time points during polymerization (Mt), to the initial monomer concentration (M0) was estimated from the absolute signal intensities in the 31P{1H} NMR spectra.

2.4. Poly(organo)phosphazene Series 5b and 6b

ncalc was varied from 10, 25, 50, 100, up to 200 by adjusting the ratio of monomer to the chlorinated phosphine 2b followed by macromolecular substitution. Further experimental details and characterization are included in the Supporting Information.

3. Results and Discussion

3.1. Initiation and Polymerization

Aryl phosphines bearing either a styrene (1a) or a trimethylsilyl ester (TMSE) (1b) functional group were investigated as potential starting materials for the polymerization of trichlorophosphoranimine (Cl3P = N−SiMe3) (Figure 1). For this purpose, a small excess of the mild chlorinating agent C2Cl6 (1.1 equiv.) was reacted separately with 1 equiv. of the phosphine 1a and 1b, respectively, with the reductive dechlorination of C2Cl6 yielding the dichlorophosphoranes 2a and 2b. 31P{1H} NMR spectroscopy showed the disappearance of the phosphine signal at −5.9 ppm of 1a and an arising signal at 56.4 ppm, indicating the formation of 2a, with complete conversion of the functional aryl phosphines being observed after 24 h. With 1b, however, significantly longer chlorination times (up to 72 h) were required until the phosphine signal at −5.1 ppm had disappeared and 2b was formed as evidenced by the new signal at 56.4 ppm.

Figure 1

Synthetic route to mono-end-group-functionalized polydichlorophosphazenes in a one-pot reaction using a functional phosphine as the initiating species and macromolecular substitution of 4b with various chain lengths using N-(3′-aminopropyl)-2-pyrrolidone (polymers 5b) and Jeffamine (x = 19, y = 3) (polymers 6b), respectively.

The polymerization progress of Cl3P = N−SiMe3 in the presence of the chlorinated 2a and 2b, as polymerization initiators, was monitored via 31P{1H} NMR spectroscopy (Figure S1, Supporting Information). After the addition of 1 equiv. of Cl3P = N−SiMe3, the signal at 64.1 ppm instantly diminished and upon addition of a further 25 equiv. Cl3P = N−SiMe3, a rapid decrease in the monomer signal at −53.8 ppm was observed, with monomer consumption being complete in 1.5 h. A parallel increase in the signal for the interior [Cl2P = N] groups of the polyphosphazene chain at −16.7 ppm (4a) was also observed. Although these results gave a clear indication of the polymerization progress, suggesting a similar mechanism to reported works,[19] details of the precise mechanism are a matter of future investigation.

Monomer conversion over time was monitored via the intensity of the Cl3P = N−SiMe3 signal at −53.8 ppm in the 31P{1H} NMR spectra (Figure 2), with the plots of ln(Mt/M0) over time showing pseudo-first-order kinetics. The well-investigated PCl5-initiated reaction is also shown as a comparison with a similarly linear, living chain growth being observed. In contrast, this phosphine-mediated polymerization results in monodirectional chain growth, as oppose to the bidirectional growth observed with the polymerization of Cl3P = N−SiMe3 and PCl5,[18] which otherwise requires [R3P = N = PCl3][PCl6] end-groups for a mono-directional growth to be achieved.[19]

Figure 2

Kinetic study of polymerization initiated with a) PCl5, b) 2a, and c) 2b monitored by 31P{1H} NMR spectroscopy. Plots of ln(Mt/M0) over time show a linear relationship indicating the living nature of the polymerization.

3.2. NMR and SEC Characterization of Polymer Series

A series of polydichlorophosphazenes, with varying ratios of monomer to phosphine-initiator was then prepared, with ncalc in the range of 10 to 200, assuming one chlorinated phosphine group per chain. To this end, a solution of 2b was reacted separately with different amounts of Cl3P = N−SiMe3 to obtain the respective polydichlorophosphazenes (4b) with various polymer chain lengths. 31P NMR spectra of the precursor polymers whereby the monomer:initiator feed ratio (ncalc) is 10, 25, and 50 showed narrow signals at −18 ppm and complete disappearance of the signal at −54 ppm indicating complete consumption of the monomer after 24 h reaction time (Figure 3c). For ncalc is 100 and 200, however, polymerization was slow and did not proceed to completion within 24 h. The reaction time had to be increased to 48 h to ensure complete disappearance of the monomer signal in the 31P NMR spectra.

Figure 3

a) Representative 1H NMR spectrum of polymer 5b-25 in D2O. Values for ncalc and thus Mn could be obtained via integration of the protons of the TMSE end-group (δ = 0.01 ppm) versus the −CH2 signals of N-(3′-aminopropyl)-2-pyrrolidone side groups (δ = 2.91 ppm); b) Plot of Mn versus ncalc of polymer series 5b, Mn calc was calculated from the M:I feed ratio, Mn exp (experimental value) via 1H NMR spectroscopy; c) 31P{1H} NMR spectra of the functionalized triphenylphosphine 1b, the dichlorophosphorane initiator 2b, dichloropolyphosphazene 4b, and the macrosubstituted poly(organo)phosphazene 6b. (* = Phosphine oxide: Since this peak is not present in sample 4b, but only in the aliquot of 2b and 1b taken for NMR spectroscopy, it is assumed that this is formed by hydrolysis and/or oxidation in the NMR sample tube); d) SEC chromatographs of polymer series 6b with ncalc are 10, 25, 50, and 100, respectively.

Due to the hydrolytic instability of polydichlorophos­phazenes, the chlorine atoms were then substituted in order to give more hydrolytically stable polymers for further characterization (Figure 1). For this purpose, the polydichlorophosphazene series 4b, with increasing ncalc, were substituted with low-molecular-weight N-(3′-aminopropyl)-2-pyrrolidone to give the poly(organo)phosphazene series 5b. 1H NMR spectroscopy confirmed incorporation of the TMSE-phosphine end-group into the polymer chain (Figure 3a) and the experimental chain length was calculated by integration of the TMSE end-group at 0.01 ppm versus the methylene group at 2.91 ppm of the attached side group (Table1). For lower values of n, approximate agreement was observed with the expected number of repeat units ncalc, thus providing good evidence not only of the phosphine end-group incorporation but also the controlled nature of the polymerization and the ability to regulate the polymer chain length via the initiator to monomer molar feed ratio. However, the polymerization appears to be inhibited, when it comes to higher ratios of monomer to chlorinated phosphine. As can be seen in Figure 3b, an increasing deviation of Mn from the calculated values was observed. Since complete consumption of the monomer was observed in all 31P NMR spectra of the polydichlorophosphazene precursors, simultaneous side-reactions resulting in a consumption of monomer not participating in chain growth is assumed. As well as the peaks corresponding to the growing polymer chain, 31P NMR spectroscopy showed signals at around 20 ppm (Figures S1 and S2, Supporting Information). Further analysis of this region of the spectrum with HMBC spectroscopy (Figure S3, Supporting Information) shows one peak, which correlates to the phosphine end-group, but also peaks corresponding to the formation of a new species, including a peak at 21.4 ppm, corresponding to cyclic by-products.[10a,20] This new species increases with increasing reaction time and represents a competing reaction to the chain growth, particularly chain lengths >50. Interestingly, these peaks were also observed in the 31P NMR spectra of the highly concentrated sample of the PCl5-initiated polymerization.

Table 1

Size characterization of the polymer series 5b and 6b

Polymer	nexpb)	Mn NMRd)(Mn calc)c) [kg mol−1]	Polymer	nexpb)	Mn NMRd)(Mn calc)c) [kg mol−1]	Mn SECe)(Mn calc)c) [kg mol−1]	Mw/Mn
5b-10a)	8	3.6 (4.3)	6b-10a)	9	21.8 (24.9)	8.2 (24.9)	1.2
5b-25a)	21	7.7 (9.2)	6b-25a)	23	50.5 (55.6)	14.6 (55.6)	1.1
5b-50a)	38	13.3 (17.4)	6b-50f)	45	96.5 (106.7)	26.3 (106.7)	1.1
5b-100a)	50	17.4 (33.8)	6b-100a)	–g)	–g)	34.6 (208.9)	1.3
5b-200a)	90	30.5 (66.5)

Number indicates ncalc, expected degree of polymerization calculated from the molar feed ratio of monomer to initiator;

Experimental degree of polymerization, as determined by 1H NMR spectra;

Calculated from ncalc;

Calculated from nexp;

Apparent molecular weights measured by SEC using conventional calibration versus linear polystyrene standards;

Polymer 6b-50 was synthesized via subsequent addition of two aliquots of 25 molar equiv. of monomer;

End-group integral too low for a reliable calculation.

In order to achieve higher molecular weight polymers, a further series of poly(organo)phosphazenes was prepared (6b) with a significantly larger organic component. The SEC traces of the polymers 6b are shown in Figure 3d. With the addition of increasing M:I ratio, a clear shift to lower retention volumes is observed. The molecular weights measured are, however, highly underestimated due to the different hydrodynamic behavior of the polystyrene calibration standards (Table1). In order to confirm that the living chain ends indeed remained active, the precursor of polymer 6b-50 was synthesized with a second, subsequent addition of 25 molar equiv. of Cl3P = N−SiMe3 after complete consumption of a first portion of 25 molar equiv. of the monomer. 31P NMR spectroscopy of the precursor, as well as SEC of the macrosubstituted polymer (Figure 3d) confirmed complete consumption of the second addition of monomer and thus the living nature of the polymerization.

4. Conclusions

The development of a novel approach to synthesize mono-end-functionalized polyphosphazenes via phosphine-based chain initiators has been described. This route yields polydichlorophosphazenes in a one-pot fashion with tunable chain lengths due to the living nature of the polymerization process. Various functional groups can be easily incorporated by careful choice of the phosphine used, thus yielding polyphosphazenes with (protected) functional groups at one chain end that are accessible for subsequent further modifications. Although further investigations into the factors, which could affect the reaction rates of the phosphoranimine polycondensation (e.g., solvent polarity, temperature or concentration) are required, this phosphine-mediated polymerization is shown to be a promising route to prepare controlled, chain-end-functionalized poly(organo)phosphazenes, which could have wide ranging usage, for example, in bioconjugation or as biodegradable building blocks for advanced macromolecular constructs.