Preparation of a versatile precursor of novel functionalized polymers: the influence of polymerization conditions on the structure of poly [1-(2-hydroxyethyl)aziridine].
Hydroxylated polymers can be regarded as versatile precursors in the preparation of new polymers through chemical modification, which can eventually lead to properly functionalized macromolecules for specific applications. Based on this approach, polymers like poly(vinyl alcohol) (PVA) and poly(glycidol) have been extensively used in the production of hydrogels, membranes, emulsifiers, stimuli-responsive materials, etc. (1-7).
The primary advantage of chemical modification with respect to polymerization lies in its simplicity. Actually, any organic reaction which can be conducted on small molecules can be applied to obtain a modified polymer using this approach (esterificaiion, nucleophilic substitution, oxidation, etc.)
Hydroxylated polyamines can be considered suitable starting materials for the preparation of liquid crystalline polyamines by means of chemical modification, which can combine the existence of an ordered structure with the presence of basic and coordinative positions. Such materials could, in principle, be applicable in the fields of ion separation and/or selective transport For instance, liquid crystalline polyethers have been successfully applied in the field of proton transport membranes (8). Moreover, only a few examples of liquid crystalline polyamines have been reported to date (9-12). Polyamines can be prepared hy ring-opening polymerization of aziridine and aziridine through a cationic mechanism, as the three- and four- membered rings contain enough ring strain to be polymerizable (13). This polymerization reaction gives rise to competition among many undesirable side reactions, such as proton transfer termination, branching and macrocycle formation. Proton transfer termination can be avoided by using tertiary amines as monomers, but the other two undesirable reactions can be only reduced by means of an appropriate choice of the monomer structure and the polymerization conditions. Thus, the polymerization of aziridine and azetidine implies the comprehensive optimization of the reaction conditions (14). For this reason, the chemical modification approach may be especially suitable in the preparation of polyamines. However, to date only a few functionalized polyamines have been described in the literature (15-18).
Aziridines are useful intermediates in organic chemistry due to their highly strained ring systems which allow for a wide range of reactivity. Three-membered heterocyclic rings offer an uncommon combination of reactivity, synthetic flexibility, and atom economy (19). This work deals with the polymerization of a hydroxylaled aziridine, 1-(2-hydroxyethyl)aziridine (EA-OH), which could lead to a suitable starting hydroxylaled polymer for the preparation of more complex polyamines through chemical modification. The polymerization of EA-OH using [BF.sub.3][middot][Et.sub.2]O as an initiator was reported by Rivas et al. (17) in the early 90s. Copolymers of this monomer with acrylamide, aziridine, and 2-oxazoline have also been reported to be useful as polychelalogens and gene delivery systems (20), (21). Despite the interesting properties exhibited by these systems, the use of poly[1-(2-hvdroxyethy1)aziridine] was largely abandoned after its discovery in the 1990's, probably due to the irregularity of its structure, as we will discuss later. Actually, Rivas et al. (17) called for a non-branched structure of polyl[1-(2-hydroxyethyl)aziridine] but, though the reported [[.sup.1]H] and [[.sup.13]C] NMR spectra showed that the expected signals were prevalent, ihey also showed small signals which were nol discussed by the authors and which can arise from end groups, branching or macrocycle formation, as well as from side reactions which involve the participation of the hydroxyl group.
For these reasons, in this study we investigated in more depth the cationic polymerization of EA-OH by using several initiators ([BF.sub.3][middot][Et.sub.2]O, [BF.sub.3][middot][EtNH.sub.2], La[(OTf).sub.3] Methyl triflale, etc.) and reaction conditions in order to obtain a polyamine with a structure as linear as possible and a molecular weight suitable for possible future applications in fields such as membrane technology. The resulting polymers were characterized by inherent viscosity determination, differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), matrix-assisted laser desorption/ionizalion-time of flight (MALDI-TOF), size exclusion chromatography coupled to a multi-angle laser light-scattering detector (SEC-MALLS), and [.sup.1]H- and [.sup.13]C-NMR. We also recorded the [.sup.15]N-NMR spectra, which allowed us to detect quaternary ammonium end groups, The best results were obtained when [BF.sub.3][middot][EtNH.sup.2] was used as an initiator. In this case, the NMR analysis identified the formation of morpholine end groups arising from the attack of neighboring hydroxyl groups, which possibly correspond to the main chain transfer reaction. Nevertheless, kinetic analysis showed that the reaction corresponded to a first-order process for the first 2 h: therefore, this evidence would suggest a pseudo-living behavior in initial polymerization steps.
MATERIALS AND METHODS
All reagents (l-(2-hydroxyeihyl)aziridine (95%), borontrifluoride diethyl etheraie (46.5% [BF.sub.3]basis), boron trifluoride ethylamine complex, lanthanum trifluoromethane-sulfonate (99.9%), scandium trifluoromethanesulfonate (99%), methyl trifluoromethanesulfonate (98%). ethylamine (2.0 M solution in THF) trimethylsulfonium iodide (98%), triphenylcarbenium hexachloroantimonate, triethyloxonium tetrafluoroborate, and benzylamine (99%)) were supplied by Fluka or Aldrich and used as received.
The solvents (acetonitrile, N,N-dimethyiformamide) were purified as described in the literature (22).
The monomer (2.17 g, 0.02 mol), initiator (mol as indicated in Tables 1 and 7) and 3 ml of solvent (in the case of reaction in solution) were placed in a round-bottom glass flask with an argon inlet tube. The mixture was heated at 45[deg]C (except for polymers P-23 and P-26. see Table 7) under magnetic stirring. After 24 h the polymers were isolated by one of the two following methods: for the reactions performed in solvents, polymers were recovered by pouring the reaction mixture into diethyl ether (which contained 1 vol% methanol) and then decanting and drying: for the reactions carried out without solvent, 2-3 ml of methanol were added to dissolve the polymer, which was then precipitated twice in diethyl ether and dried.
TABLE 1. Cationic polymerization reaction conditions and results. Polymer Initiator (a) Solvent P-l [BF.sub.3][middot][Et.sub.2]O [CH.sub.3]CN P-2 DMF P-3 [BF.sub.3][middot][EtNH.sub.2] [CH.sub.3]CN P-4 DMF P-5 [Sc(OTf).sub.3] [CH.sub.3]CN P-6 DMF P-7 [La(OTf).sub.3] [CH.sub.3]CN P-8 DMF P-9 [CF.sub.3][SO.sub.2][OCH.sub.3] [CH.sub.3]CN P-10 DMF P-l1 [([CH.sub.3]).sub.3]SI [CH.sub.3]CN P-l2 DMF P-13 [([c.sub.6][H.sub.5]).sub.3]CSb[Cl.sub.6] [CH.sub.3]CN P-14 DMF P-l5 [([c.sub.2][H.sub.5]).sub.3]O([BF.sub.4]) [CH.sub.3]CN P-16 DMF P-17 HBr [CH.sub.3]CN P-l8 [BF.sub.3][middot][Et.sub.2]O + [C.sub.7] [CH.sub.3]CN [H.sub.9]N P-l9 [BF.sub.3][middot][EtNH.sub.2] + [C.sub.7] [CH.sub.3]CN [H.sub.9]N P-20 [NH.sub.2][CH.sub.2][CH.sup.3] + [CH.sub.3]CN [Sc(OTf).sup.3] Polymer Conversion (%) Viscosity ([[eta].sub.inh]) (dL/g) P-l 72 0.81 P-2 92 0.66 P-3 81 0.77 P-4 86 0.65 P-5 49 1.82 P-6 86 1.67 P-7 79 1.47 P-8 82 1.31 P-9 49 0.84 P-10 55 0.55 P-l1 52 0.97 P-l2 64 1.20 P-13 85 1.58 P-14 89 1.12 P-l5 43 0.69 P-16 51 0.65 P-17 66 0.81 P-l8 64 0.80 P-l9 57 1.70 P-20 42 1.21 Polymer [M.sub.w] X 10 [M.sub.n] X [10 [M.sub.w] .sup.-3] (b) (g/mol) .sup.-3] (b) (g/mol) /[M.sub.n] P-l 9.1 7.4 1.2 P-2 9.2 7.3 1.3 P-3 6.2 2.8 2.2 P-4 6.2 2.9 2.1 P-5 170 22 7.8 P-6 12 3.5 3.6 P-7 - - - P-8 - - - P-9 - - - P-10 - - - P-l1 10 5.3 1.9 P-l2 6.7 3.6 1.9 P-13 7.4 1.7 4.3 P-14 6.1 1.4 4.3 P-l5 - - - P-16 - - - P-17 - - - P-l8 - - - P-l9 - - - P-20 - - - (a) Initiator/monomer ratio 1 mol%. (b) Determined by SEC-MALLS. Error: [PLUS-MINUS] 10%.
Characterization and Techniques
Thermogravimetric Analysis. Thermal stability studies were carried out in ALU OXIDE crucibles of 70 [mu]l (ME- 24123) with a Mettler TGA/SDTA85le/LF/1100 at temperatures ranging from 30 to 600[deg]C with a heating rale of 10[deg]C/min using aboul 10 mg of sample in nitrogen atmosphere (100 ml/min). The equipment was previously calibrated with indium (156,6[deg]C) and aluminum (660.3[deg]C) pearls.
Differential Scanning Calorimetry. Calorimetrie studies were performed at the heating rate of 10[deg]C/min using about 5 mg of sample with a Mettler DSC822e thermal analyzer, nitrogen as a purge gas (100 ml/min) and liquid nitrogen for syslem cooling. The equipment was previously calibrated using indium (156.6[deg]C) and zinc (419.58[deg]C) pearls and aluminum standard 40 [mu]l crucibles with pin ME-27331.
Inherent Viscosity ([[eta].sub.inh.]). Inherent viscosity was measured on an Ubbelohde viscometer (DIN 510 10/I) connected to an automatic Schott Gerate AVS310 meter equipped with a Schott Gerate CT 050/1 thermostatic bath and a Selecta 6000382 external cryoscopic unit. Measurements were taken with a Schott Gerate type 50110/1 capillary viscometer at 30[deg]C using a 2 g/1 sample diluted in dimethyl sulfoxide (DMSO).
Nuclear Magnetic Resonance. Polymers were characterized by [.sup.1]H-, [.sup.13]C-. and [.sup.15]N-NMR. which were performed using chloroform ([CDCl.sup.3]) and tetrachloroethane ([CI.sup.2][CDCDC1.sup.2]) as deuterated solvents with a Varian Gemini 400 MHz spectrometer ([.sup.1]H - 400 MHz. TMS; [.sup.13]C - 100 MHz, TMS; [.sup.15]N - 40 MHz, [CH.sup.3][N0.sup.2]) at room temperature with pulse delay time of 5 s for the [.sup.1]H-NMR spectrum.
Kinetic studies were performed in deuterated acetonitrile at 45[deg]C using [.sup.1]H-NMR, The sample chamber had been previously equilibrated at 45[deg]C by means of a temperature control unit. Then the sample tube containing the freshly prepared reaction mixture and deuterated acetonitrile was inserted into the sample chamber. The spectra were recorded every 5 min for the first 35 min and then every 15 min for the following 90 min.
Quantitative [.sup.13]C-NMR spectra were performed using deuterium oxide as a solvent ([D.sup.2]0) at room temperature with a pulse delay time of 5 s. Delay time was selected on the basis of the relaxation times determined for the monomer 1-(2-hydroxyelhyl)aziridine.
Matrix-Assisted Laser Desorption/Ionization-Time of Flight (MALDI-TOF) MALDI-TOF MS measurements were taken with a Voyager DE-RP mass spectrometer (Applied Biosystems. USA) equipped with a nitrogen laser delivering 3 ns laser pulses at 337 nm. 1,8,9-anlrace- nelriol was used as matrix and potassium or silver triti uoroacetaie as a dopant. Samples were prepared from 3 mg/ml of polymer in chloroform. Matrix was prepared from 40 mg/ml of matrix and 0.1 mg/ml of dopant in THF. The polymer sample was spotted on the MALDI plate, dried, and matrix was placed on top.
Size-Exclusion Chromatography Coupled to a Multi-Angle Laser Light-Scattering Detector (SEC-MALLS). SEC measurements were performed using a PolarGel-M column (Polymer Labs, UK), an Agilent Series 1200 isocratic pump (Agtlent Tech. USA), a K230I differential refractive index (RI) detector (Knauer. Germany) and a MiniDAWN MALLS detector (Wyatt Tech, USA). The eluenl was WA-dimethylacetamide (DMAc) with 3 g/1 LiCl and a flow rate of 1.0 ml/min. The dn/dc value was calculated from the measurements as 0.134 dl/g. The molecular weights were calculated using Astra 4.9 software and Corona software (Wyatt Tech, USA). The fit of the molecular weight/elution volume dependence or sample specific calibration (23) was used to estimate the molecular weight distribution of the lower molecular weight samples. The samples were measured at room temperature. In order to preveni aggregation, the solutions were pretreated at 60[deg]C for 24 h.
RESULTS AND DISCUSSION
As mentioned above, the aim of this work was to examine in more depth the cationic polymerization of 1-(2-hydroxyethyl)aziridine by using several different initiators and reaction conditions in order to obtain a posamine with a structure as linear as possible and a molecular weight suitable for future applications, such as membrane preparation. It is known that aziridines are not reactive towards nucleophilic compounds unless the nitrogen has been quatemized, protonated or has formed a Lewis acid adduct.
As a matter of example, the [BF.sub.3]-catalyzed polymerization mechanism is shown in Scheme 1. The propagation reaction takes place by means of an [S.sub.N]2 attack of the monomer nitrogen atom in the methylene group of the cyclic ammonium active ion species. This attack can also occur with a nitrogen atom of !he polymer molecule and, in the polymerization of tertiary aziridines, the thus formed branched or macrocyclic quaternary ammonium ion is unreactive toward propagation and this reaction is a real termination reaction. As a consequence, the polymerizations of N-substitutcd aziridines generally do not proceed to completion and do not lead to high molecular weight polymers (13).
However, when the starting tertiary aziridines possess a suitable degree of steric hindrance, polymerization can lead to high molecular weight polyamines and, in some cases, even has an almost "living" character: although bulky groups hinder all [S.sub.N]2 attacks compared with nitrogen in the monomer, the nitrogen atoms in the polymer chain are much more sterically hindered (14).
We studied the polymerization of EA-OH initiated with different catalysts. The most commonly used initiators for cationic polymerization are protic acids, Lewis acids, stable organic salts and alkylating or acylating covalent compounds. The chemical structures of the initiators used in this article are reported in Fig. 1. Protic acids can act as initiators of the cationic polymerization of heterocyclic monomers. In particular, strong acids with a non-nucleophilic anion should be used, as in this case the acid-base balance should be shifted to the right-hand side and they should provide weakly nucleophilic anions to prevent the collapse of active ion species. In the polymerization of highly nucleophilic cyclic monomers (oxazolines, cyclic amines), however, monomers may compete successfully with B[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] or 1 counterions. For these monomers, initiation with HBr, for example, may lead to high molecular weight polymers (24).
Among the Lewis acids, [BF.sub.3] is the most frequently used as an initiator for cationic ring-opening polymerization. This is at least partly due to the fact that [BF.sub.3] forms a stable, well-defined complex with ethers (25). Also, [BF.sub.3][middot][Et.sub.2]O is relatively easy to handle and to purify.
Initiators of stable organic salts can be divided into several differenl categories: carbenium salts, oxacarbe- nium salts, carboxonium salts, oxonium salts, etc. (26). Very often these initiators eliminate the drawbacks presented by the more common protic and Lewis acid catalysts in that they are chemically well defined and allow a quantitative initiation reaction to proceed without complicated side reactions or the need for cocatalysts and promoters. Stable organic cations have provided a range of initiator systems for a wide variety of cationic polymerizations of olefmic and heterocyclic monomers (27).
Covalent compounds, which are strong alkylating or acylating agents, may initiate the cationic polymerization of heterocycles. Esters of very strong protic acids (trif1uoromethanesulfonie, fluorosulfonic, etc.) are sufficiently powerful alkylating agents to initiate the polymerization of even weakly nucieophilic monomers. Furthermore, their anhydrides (i.e., triflic anhydride) are efficient initiators (25). The preparation of poly(oxazoline) containing tapered minidendritic side groups with methyl trillale has been reported in other studies in which the polymerization took on a living character, yielding 100% monomer conversion and macromolecules with well-defined molecular weights and narrow molecular weight distributions (28).
We performed polymerizations in polar solvent [N,N- dimethylformamide or acetonitrile) because they yield higher conversions than less polar solvents such as dichloromethane or toluene, as reported by Rivas et al. (17). The reaction conditions and results obtained by varying the nature of the initiator and the solvent are summarized in Table 1.
When DMF was used, in most cases polymer conversion increased while viscosity decreased. Because inherent viscosity ([[eta].sub.inh]) may decrease as a consequence of either lower molecular weight or branching, we can conclude that polymerization in DMF led to more branched structures and/or lower molecular weights, possibly due to the occurrence of branching and chain transfer side reactions (Table 1). However, these polymers are expected to exhibit certain levels of aggregation due to the presence of alcoholic protons, which favor the formation of hydrogen bonds; for this reason, viscosity is probably also affected by the presence of aggregation phenomena via hydrogen bonding.
For selected samples, absolute molecular weights were determined by SEC-MALLS in N,N-dimethylacetamide and the results are repotted in Table 1, In order to reduce aggregation, SEC was performed both on virgin samples and after 24 h of annealing the sample solution at 60[deg]C in the presence of LiCl. The aggregates were visible only on the light scattering (LS) signal and not on the RI signal, meaning that very few were present. Figure 2 shows the LS and RI signals versus elution volume for sample P-3 as an example. It is clear that after annealing the high molecular weight mass shoulder of the peak, attributable to the presence of aggregates, was greatly reduced. In all cases, we found low aggregate concentrations, as it can be concluded from the RI signal in the higher molar mass region. These aggregates almost completely disappeared after annealing. The molecular weights reported refer to the samples after annealing. The weight average molecular weight determined by SEC-MALLS (Table 1) generally ranged between 6000 and 12,000. with the exception of polymer P-5 (Supporting Information. Fig. S2). The remarkable [M.sub.w] value found for this system can be attributed to both the high acidity of scandium triflate and its oxophilicity (29): the former favors the initiation step, thus making polymerization more effective, while the latter eases the participation of hydroxyl groups in side reactions, which would give rise to a highly branched and irregular structure. This effect is less important in the case of P-6, whose polymerization was performed in DMF, which has greater hydrogen-bond acceptor properly compared to aceionitrile (30), (31).
The structure of polymers was characterized by [.sup.1]H-, [.sup.13]C-, and [.sup.15]N-NMR spectroscopy. Table 2 shows the main structure and the expected head group of most synthesized polymers. In principle, one should expect a secondary amine as a head group for the proton catalyzed mechanism: however, for many other Lewis acid catalysts. ([BF.sub.3], [Sc(OTf).sub.3], etc.) the same head group will be obtained after workup.
TABLE 2. The main structure and head group of polymers. [.sup.1]H-NMR [.sup.13]C-NMR Signal (ppm) Assignment Signal (ppm) Assignment 2.6 1. 2. 3. a, b 46.8 1 36 4, c 51.7 3 4.9 A, d 52.9 a 54.4 2 56.6 b 59.9 c 61.0 4
The signals corresponding to the main chain ol the polymer (Table 2, assignments a, b, c and d) were observed by means of [.sup.1]H-NMR and [.sup.13]C-NMR. while the signals of the head group (Table 2, assignments 1, 2, 3, 4) were detected as small signals only by [.sup.13]C-NMR because their proton signals overlapped.
As previously mentioned, the expected termination reaction occurs by means of the attack of the main chain nitrogen atoms onto cyclic ammonium active ion species (Scheme 1), which leads to the formation of branched or cyclic quaternary ammonium groups. In the case of macrocycle formation (over nine-membered rings), one expects the same NMR signals as in the case of branching. However, pyrazine structures may eventually be formed if the attack is executed by the neighboring nitrogen atom.
Figure 3 shows the [.sup.13]C NMR spectrum of P-3, as an example, which demonstrates the main chain and head group signals already discussed. Moreover, we were able to delect some signals which were assigned as arising from branched or macrocyclic end groups of polymers. These groups did not produce resolved signals in the [.sup.1]H NMR spectrum, while in the [.sup.13]C NMR spectra, only carbon 6 and 7 signals appeared resolved (Table 3). We estimated the expected chemical shifts for carbon atoms arising from the pyrazine end group on the basis of literature data for similar structure as well as from empirical calculations (32). In most cases they were expected to overlap with the signals already discussed, while carbon IV (Figure 3) could be assigned to a small signal at 47.3 ppm. Finally, the small signals at 66.7 and 54.2 ppm were assigned to carbons 1 and II of a morpholine ring end group (32) formed by the attack of neighboring hydroxyl groups: this constitutes a chain transfer reaction (Scheme 1). Moreover, morpholine end groups gave rise to a resolved small signal in the [.sup.1]H-NMR spectrum at around 2.4 ppm (see [.sup.1]H-NMR spectrum on Fig. 4).
TABLE 3. Quaternary ammonium end group observed in polymers Pl-P17. [.sup.13]C-NMR Signals (ppm) Assignment 6.1.0 5 (a) 49.2 6 69.2 7 56.6 8 (b) (a) Overlapped with signal 4. (b) Overlapped with signal b.
In the case of P-3 and P-4, which were obtained by using [BF.sub.3][middot][EtNH.sub.2], the monoethylamine could have also acted as a nucleophile in the termination step, thus producing a secondary amine as a linal group. However, the corresponding signals, expected in the [.sup.13]C-NMR spectrum at about 15 and 44 ppm on the basis of empirical calculations (32), were not detected in the spectra of these polymers. This indicates that this termination reaction does not occur to a great extent, though it cannot be completely ruled out.
In the case of P-3. by comparing the signals at 69.2 ppm (7 + VIII) and at 51.7 ppm (3), we were able to estimate the amount of quaternary ammonium end groups (branched, macrocycle and pyrazine end groups) from the [.sup.13]C-NMR spectrum recorded under quantitative conditions. In this way, we discovered that quaternary ammonium constituted 47% of the end groups; therefore, if we ignore the ethylamine end groups, we can assume that 53% of the polymeric chains contained morpholine end groups. This is roughly consistent with the determined number average molecular weight (2800 g/mol, polymerization degree (DP) approximately equal to 32). As a matter of fact, in the case of P-3, conversion was 81% and the initiator/monomer ratio was 1 mol%, which should have given rise to a DP = 81 in the case of no chain transfer reactions: actually, in our case 47% of the quaternary ammonium end groups were found, which should correspond to a DP of [Tilde]38.
The [.sup.13]C-NMR of P-4 (recorded under standard conditions) yielded results very similar to those for P-3. However, the [.sup.13]C-NMR spectra of most polymers (P-l, P-2. P-5 through P-17) exhibited many additional peaks which could not be univocally attributed (Fig. 5). The resulting signals suggest that side reactions other than those already discussed take place, possibly involving the participation of the hydroxyl groups. As an example, we also performed [.sub.13]C-NMR quantitative spectroscopy for polymer P-1 in order to estimate the amount of ammonium end groups versus head groups, as described earlier for P-3. In this case, a very similar value to P-3 was obtained, i.e., 49%, which suggested a similar contribution of the termination reactions leading to quaternary ammonium end groups. However, while P-l and P-3 exhibited similar inherent viscosities. P-3 exhibited a considerably lower molecular weight (Table 1). This suggested that P-l possessed a more branched structure, which is consistent with the extra signals found in its [.sub.13]C-NMR spectrum (Fig. 5a). Therefore, in the case of P-1 (as well as possibly P-2, and P-5 through P-17), end groups other than quaternary ammonium and morpholine can be envisaged.
All polymers were also characterized by means of MALDI-TOF. In all spectra we were able to identify a set of peaks corresponding to a multiple of the monomer molar mass (M), together with a sodium ion, a silver ion or a proton. This is compatible with polymer chains terminated either with quaternary ammonium or with morpholine groups. As a matter of example. Fig. 6 shows the MALDI-TOF spectrum for P-3. It must be noted that for P-3 and P-4, no peaks corresponding to ethylamine-terminated chains were found, thus confirming that the contribution of ethylamine as an end group was negligible.
In order to verify whether a primary amine might be involved in the termination reaction in some cases, we repeated the P-l and P-3 experiments in the presence of benzylamine in the same molar proportion as the catalyst (P-18 and P-19, Table 1). This amine was chosen because it contains an aromatic moiety that can be easily assigned in both [.sup.1]H- and [.sup.13]C-NMR. In both cases we observed the aromatic signals characteristic of this moiety as an end group (Table 4). though P-l8 contained a higher proportion of benzylamine end groups, as inferred from [.sup.1]H-NMR spectra. Moreover, in the MALDI-TOF spectrum of P-18 we identified a set of peaks corresponding to benzyl- amine terminated chains together with a sodium cation (M + [Na.sup.+] = 652, 739, 826, 913, 1000); this evidence was not found in the MALDI-TOF spectrum of P-l 9.
TABLE 4. End group of polymers P-18, P-19. *Partially overlapped with carbon II from morpholine. [.sup.1]H-NMR [.sup.13]C-NMR Signal (ppm) Assignment Signal (ppm) Assignment 3.7 8 53.9 8* 7.2-7.4 10. 11. 12 127.0-127.5 12 128.3-128.7 10 129.2-129.3 11 139,7 9
The high molecular weight of P-5 together with its considerably high viscosity value encouraged us to repeat this experiment in the presence of ethylamine in a 1:1 ratio with respect to the catalyst (Exp. P-20). We intended to take advantage of the high efficiency exhibited by [Sc(OTf).sub.3, while at the same time reducing the side reactions of the hydroxyl group by adding a better nucleophile, which should have been more competitive in the termination step. According to our hypothesis, a polymer with a more regular structure than P-5 should have been obtained. This, however, was not the case, as demonstrated by the presence of additional signals in the 45-72 ppm region of [.sup.l3]C NMR spectrum (Supporting Information, Fig. S1). Nevertheless, it is important to mention that in this case the [.sup.l3]C-NMR spectrum showed the signals attributable to ethylamine as an end group (Table 5).
TABLE 5. Ethylamine end group of polymer P-20. [.sup.13]-NMR Signals (ppm) Assignmeril 14.7 14 43.9 13
We also recorded the [.sup.l5]N-NMR spectra, which allowed us to detect some end groups by taking advantage of the extremely regular structure of the main chain. [.sup.l5]N-NMR spectroscopy has become increasingly desirable for the characterization of nitrogen-containing compounds, such as amines and ammonium compounds. The [.sup.l5]N-NMR experiment is much less sensitive than [.sup.l]H and [.sup.l3]C experiments. Consequently, the [.sup.l5]N spectrum alone is rarely sensiiive enough and the projection of a heteronuclear correlation (short range and long range) has to be used to improve sensitivity (33).
As an example, the [.sup.l5]N-NMR spectrum of polymer P- 3 is shown in Fig. 4. According to our results, the [.sup.l5]N- NMR chemical shift values are all located in the region between -200 and -350 ppm. The signals corresponding to the chemical shift at -200 ppm (Fig. 4. assignment C) were assigned as quaternary nitrogen atoms (34). These signals suggested the presence of branching or cyclic end groups. Only polymers P-3 and P-4 showed a single signal at -200 ppm, while in all other cases more than one signal of quaternary nitrogen was found. According to these results, the most regular polymer structures were obtained using boron trilluoride ethylamine complex as the initiator. The expected signals from the main chain tertiary amine (Fig. 4. assignment B), secondary amine head group (Fig. 4. assignment A) and morpholine tertiary amine (Fig. 4. assignment D) overlapped at -350 ppm.
Boron trifluoride ethylamine complex (P-3 and P-4) yielded quite regular structures compared with other tested systems. Therefore, we decided to perform a kinetic study at 45[deg]C on selected systems (P-l, P-3, P-5) in order to see whether these structural differences might occur for kinetic reasons,
[.sup.1]H-NMR and [.sup.13]C-NMR spectroscopic techniques are among the most useful methods with which to study the kinetics of reactions. This is due to the high accuracy and precision of these techniques in terms of on-line gain and data recording in relation to reaction progress. Of course, [.sup.1]H-NMR spectroscopy is preferred over [.sup.13]C-NMR because of its greater sensitivity and rate of data collection.
The [.sup.1]H-NMR spectra of the P-3 polymerizing mixture in deuterated acetonitrile at different times are shown in Fig. 7. Chemical shifts of aziridine [-CH.sub.2] protons of 1-(2-hydroxyethyl)aziridine monomer appear at 1.2 and 1.65 ppm. After polymerization, chemical shifts of methylenic protons in the polymer appear at 2.6 and 3.5 ppm. Thus, the chemical shifts of aziridine and polymer protons are sufficiently far apart and their peak area provides a good scale for quantitatively tracking reaction progress with regard to kinetic aspects.
Comparing the monomer and polymer peak areas allowed us to calculate the conversion versus time. If we assume that the initiation step is fast enough to be completed before measurements start and that during measurement time (120 min) both termination and chain transfer reactions are much slower than propagation reaction, a pseudo-first order polymerization mechanism can be postulated:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)
where M is the monomer and 1 is the initiator. According to our hypothesis. [I] coincides with the concentration of polymer-growing chains, which holds constant as long as termination reactions are negligible.
Integration of (1) between t=0 and t gives:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)
where [k.sub.app]=[k.sub.p][middot] is the apparent kinetic constant.
Figure 8 shows the In([[M].sub.0]/[[M].sub.t]) versus time plots for polymerizations P-l [BF.sub.3][middot][Et.sub.2]O, P-3 [BF.sub.3][middot][EtNH.sub.2]. and P-5 [Sc(OTf).sub.3] and demonstrates that polymerization is faster with [Sc(OTf).sub.3], as expected based on the high acidity of scandium tritiate (29). The figure also shows that polymerization with [BF.sub.3][middot][Et.sub.2]O is faster than with [BF.sub.3][middot][EtNH.sub.2]. The trend of ln([[M].sub.0]/[[M].sub.t],) versus time plot, in the case of P-5, is clearly nonlinear, while for P-l it looks linear for the first 50 minutes of reaction, though [r.sup.2]= 0.90 ([k.sub.app] = 1.9 X [10.sup.-5] [s.sup.-1]). Finally, in the case of P-3. the trend is linear over the entire time range studied, and the least squares fitting of the plot gave a correlai ion coefficient of [r.sup.2] = 0.98 and [k.sub.app] = 1.85 X [10.sup.-5] [s.sup.-1]. Given the previously stated assumptions, we can therefore conclude that termination, chain transfer and other side reactions have a non-negligible role in the case of P-5, as weil as in the case of P-l from 50 min onward. Thus, our results suggest that more active initiating systems induce a greater contribution of side reactions in the overall polymerization mechanism.
To sum up, the best results were obtained using boron trifluoride ethylamine complex (P-3 and P-4), which achieved a quite regular structure and good conversions. For these polymers, acceptable molecular weights were obtained, which corresponded to a polymerization degree of around 32 monomeric units. Moreover, kinetic studies performed during the first 120 min showed that polymerization is consistent with a pseudo-first order mechanism.
Therefore, to further improve these results, we tried modifying the reaction conditions while using boron trifluoride ethylamine complex. Our results are summarized in Table 6. together with molecular weights determined on selected samples. As an example, the gel permeation chromatograni of P-27 after annealing is shown in the Supporting Information, Fig. S3.
TABLE 6. [BF.sub.3][middot][EtNH.sub.2] polymerization reaction conditions and results. Polymer Initiator Solvent T ([degree]C) P-21 [BF.sub.3][middot][EtNH.sub.2] [CH.sub.3]CN 45 P-22 [CH.sub.3]CN 45 P-23 [CH.sub.3]CN 27 P-24 -- 45 P-25 -- 45 P-26 -- 27 P-27 -- 45 Initiator/monomer Conversion Viscosity Polymer Ratio (mol%) (%) [eta].sub.inh](dL/g) P-21 0.2 25 0.62 P-22 2 64 0.61 P-23 1 60 0.69 P-24 0.2 23 0.65 P-25 2 61 0.60 P-26 1 52 0.62 P-27 1 75 0.75 [M.sub.w] X 10.sup.-3 (a) [M.sub.w] X 10.sup.-3 (a) [M.sub.w]/ Polymer (g/mol) (g/mol) [M.sub.n] P-21 -- -- -- P-22 -- -- -- P-23 -- -- -- P-24 8.3 3.7 2.2 P-25 4.8 4.8 2.0 P-26 -- -- -- P-27 7.7 3.8 2.0 (a) Determined by SEC-MALLS. Error: [PLUS-MINUS] 10%
First, we decided to vary the initiator concentration (P- 21 and P-22). We found that the conversion and the viscosity of the polymers decreased in both cases, with a remarkable decrease in conversion with 0.2 mol% catalyst. We also modified the reaction conditions by reducing the reaction temperature (27[deg]C, P-23). In this case, the conversion and viscosity of the polymer decreased as the reaction temperature decreased. The [.sup.13]C-NMR spectra of P-21 and P-23 were no different from those shown in Fig. 3 for P-3, while the spectrum for P-22 exhibited additional peaks in the 50-62 ppm region (Supporting Information. Figs. S4 and S5).
We performed a kinetic study under the same experimental conditions described above for polymerizations P-21 and P-22. Figure 9 reports the ([[M].sub.0]/[[M].sub.t],) versus time plots of the polymerization performed using an initiator/monomer ratio of 0.2, 1, and 2 mol%, respectively. As the figure shows, in these cases the most active initiator system (2 mol%) also traced a nonstriclly linear plot, consistent with the more irregular structure found by the [.sup.13]C-NMR.
The fact that our monomer was liquid encouraged us to try conducting the polymerization without solvent (P-24 through P-27. Table 6) because it is easier to handle and to dry and more environmental friendly. Among the experiments performed, the results obtained in the case of P-27 were especially interesting, as the molecular weight increased remarkably (corresponding to a polymerization degree of 44 monomeric units), while conversion was only slightly reduced compared to P-3. It is important to stress that polymers P-26 and P-27, both obtained with 1 mol% catalyst, exhibited a structure as regular as the one obtained in the case of polymers P-3 and P-4, as indicated by [.sup.13]C-NMR (Supporting Information, Figs. S6 and S7).
The thermal characteristics of all the prepared polymers are reported in Fig. 10. As an example, DSC traces of P-3 and P-5 are reported in the Supporting Information, Fig. S8. They all behaved as amorphous materials with glass transition temperatures ([T.sub.g]s) ranging between -35 and 14[deg]C. The polymers that displayed a more regular structure (P-3, P-4, P-21, P-23. P-26, and P-27) exhibited [T.sub.g]s at around -33[deg]C. Higher [T.sub.g]s values were found in most of the other cases and could be explained by taking into account a more branched structure produced by side reactions, as already mentioned. The presence of these branching points can increase chain entanglements and hence raises [T.sub.g].
The thermal stability of the polymers was studied using TGA. As an example, TGA of P-3 and P-27 are reported in the Supporting Information, Fig. S9 The onset of thermal decomposition (determined as the temperature corresponding to a 5% loss in mass) ranged between 210 and 280[deg]C. Polymer P-27 behaved the best in this respect, exhibiting a mass loss onset temperature of 275[deg]C.
In this article we examined in more depth the cationic polymerization of l-(2-hydroxyethyl (aziridine, which had already been reported by Rivas et al., using [BF.sub.3][middot][Et.sub.2]O. Cationic catalysts of different natures were tested under several different conditions. The most regular structures were obtained using [BF.sub.3][middot][EtNH.sub.2] and selected conditions. In these structures, the expected head group (secondary amine) and the cyclic or branched quaternary ammonium moiety as an end group were detected by NMR; moreover, morpholine end groups were also recognized, which indicate the participation of hydroxyl groups leading to chain transfer reactions. Kinetic studies conducted during the first 120 min of the reactions showed that the more active the calalylic system, the more irregular the polymer structure. The best results were obtained when polymerization was performed with 1 mol% [BF.sub.3][middot][EtNH.sub.2] in the absence of solvent at 45[deg]C. which yielded a polymerization degree of about 44 monomeric units. This functionalized polymer has proven to be an interesting, versatile starling material for chemical modification reactions, which could lead, for instance, to liquid crystalline polyamines.
(1.) H. Frey and R. Haag. Rev. Mol. Biotechnol., 90. 257 (2002).
(2.) J.-W. Rhim. H.B. Park, C.-S. Lee. J.-H. Jun. D.S. Kim, and Y.M. Lee. J. Membr. Sci., 238. 143 (2004).
(3.) D.S. Kim. H.B. Park. J.W. Rhim, and Y. Moo Lee. J. Membr. Sci.. 240, 37 (2004).
(4.) M. Jamroz-Piegza. A. Utrata-Wesolek, B. Trzebicka, and A. Dworak, Eur. Polym. J., 42, 2497 (2006).
(5.) F. Liu, Prog. Polym. Sci., 35, 3 (2010).
(6.) B. Obermeier, F. Wurm, and H. Frey, Macromoleclues. 43, 2244 (2010).
(7.) S. Salehpour and M.A. Dube, J. Macromol. Sci., Part A. 49, 103 (2012),
(8.) B. Tylkowski, N. Castelao, M. Giamberini. R. Garcia-Valls, J.A. Reina, and T. Gurmi, Mater. Sci. Eng. C, 32, 105 (2012).
(9.) S. Ujiie, S. Yano, and M. Sato, Trans. Mater. Res. Soc. Jpn.. 24, 437 (1999).
(10.) S. Ujiie. S. Ojiro. Y. Yano, and A. Mori, Transact. Mater. Res. Soc. Jpn., 30. 373 (2005).
(11.) Y. Kawakami, M. Baba, Y. Murate, H. Kunisada, and Y. Yuki, Polym. J., 28. 1006 (1996).
(12.) Y. Yano, S. Ujiie, and A. Mori. Mol. Cryst. Liq. Cryst., 365. 255 (2001).
(13.) E, J. Goethals. Cationic Ring-opening Polymerization: Amines and N-containing Heterocyclcs, in Comprehensive Polymer Chemistry, G. Allen and J. C. Bevington. Pergamon Press, Oxford, 837 (1989).
(14.) E.J. Goethals and J.M. Geeraert. Macromolecular Design of Polymeric Materials, Plastics Engineering, New York ( 1997).
(15.) J.R. Meyer. H. Keul, and M. Moller. Macromolecules 44, 4082 (2011).
(16.) Y. Rousselin, N. Sok, and F. Boschetti. Eur. J. Org Chem., 9. 1688 (2010).
(17.) B.L. Rivas, K.E. Geckeler, and E. Bayer, Eur, Polym. J., 27. 1165 (1991).
(18.) B.L, Rivas and I. Concha. Polym. Bull., 30, 191 (1993).
(19.) T. Eicher and S. Hauptmann, The Chemistry of Hetero- cycles: Structure, Reactions, Syntheses, and Applications, Wiley-VCH, Germany (2003).
(20.) B.L, Rivas and I. Moreno-Villoslada, J. Appt. Polym. Sci.. 69. 817 (1998).
(21.) D. Fischer, A. von Harpe, K. Kunath, H, Petersen, Y. Li, and T. Kissel. Bioconjugate Chem., 13. 1124 (2002).
(22.) B.S. Furniss, A.J. Hannaford, P.W.G. Smith, and A.R. Tatchelli, Vogel's Textbook of Practical Organic Chemistry, Pearson Education, London (1989).
(23.) A. Lederer. D. Voigt. D. Appelhans, and B. Voit. Polym. Bull., 57, 329 (2006).
(24.) M. Bednarek, K. Brzezinska, J. Stasinski. P. Kubisa, and S. Penczek, Die Makromolekulare Chemie, 190, 929 (1989).
(25.) D.J. Brunelle. Ring-Opening Polymerization: Mechanism. Catalysis. Structure. Utility, Hanser Publishers. Munich (1993),
(26.) K. Matyjaszewski, Cationic Polymerization: Mechanism, Synthesis, and Applications. Marcel Dekker, New York, 437 (1996).
(27.) A. Ledwith and D. Sherrington, Stable Organic Cation Salts: Ion Pair Equilibria and Use in Cationic Polymerisation. Springer, Berlin (1975).
(28.) V. Percec. M.N. Holerca, S. Uchida, D.J.P. Ycardley, and G. Ungar. Biomacromalecules, 2. 729 (2001).
(29.) S. Kobayashi. Eur. J. Org. Chem., 15 (1999).
(30.) S.H. Gellman, G.P. Dado. G.-B. Liang, and B.R. Adams, J., Am. Chem. Soc., 113. 1164 (1991).
(31.) P. Ruelle and U.W. Kesselring, Pharm. Res., 201 (1994).
(32.) E. Pretsch, P. Buhlmann, and C. Affolter, Structure Determination of Organic Compounds: Tables of Spectral Data. Springer 78. 121 (2000).
(33.) G.J. Martin. M.L. Martin, and J.P. Gouesnard, 15N-NMR Spectroscopy. Springer, Verlag (1981).
(34.) G.C. Levy and R.L. Licher, Nitrogen-15 Nuclear Magnetic Resonance Spectroscopy, Wiley, New York, 221 (1979).
Asta Sakalyte, (1) Jose Antonio Reina, (2) Marta Giamberini, (1) Albena Lederer (3)
(1) Departament d'Enginyeria Quimica, Universitat Rovira i Virgili, Av. Paisos Catalans, 26, 43007 Tarragona, Spain
(2) Departament de Quimica Analitica i Quimica Organica, Universitat Rovira i Virgili, Carrer Marcel-li Domingo s/n, 43007 Tarragona, Spain
(3) Leibniz-Institute of Polymer Research Dresden, Hohe Str. 6, 01069 Dresden, Germany, and Technische Universitat Dresden, 01062 Dresden, Germany
Additional Supporting Information may be found in the online version of this article.
Correspondence to: Marta Gianiberini;e-mail: firstname.lastname@example.org
Contract grant sponsor Ministerio de Ciencia e Innovacion; contract grant number: MAT2008-00456/MAT.
Published online in Wiley Online Library (wileyontinelibrary.com).
[c] 2013 Society of Plastics Engineers
|Printer friendly Cite/link Email Feedback|
|Author:||Sakalyte, Asta; Reina, Jose Antonio; Giamberini, Marta; Lederer, Albena|
|Publication:||Polymer Engineering and Science|
|Date:||Mar 1, 2014|
|Previous Article:||Copolymers based on epoxidized soybean oil and diglycidyl ether of bisphenol A: relation between morphology and fracture behavior.|
|Next Article:||Preparation and characterization of chitosan-stearate complexes and in vitro evaluation on the adsorption of deoxycholic acid salt.|