Stoichiometric imbalance-promoted step-growth polymerization based on self-accelerating 1,3-dipolar cycloaddition click reactions

Abstract

A library of self-accelerating click reactions was developed based on the 1,3-dipolar cycloaddition of sym-dibenzo-1,5-cyclooctadiene-3,7-diyne (DIBOD) and varied 1,3-dipoles, such as diazo, sydnone, and nitrone groups. A common feature of these reactions was that the reaction of a 1,3-dipole and the first alkyne moiety of DIBOD activated in situ the second alkyne moiety, which consequently reacted with a 1,3-dipole at a much faster rate than did the original DIBOD alkyne group. Because these were polymerization reactions, a novel kind of stoichiometric imbalance-promoted step-growth polymerization method was developed specifically to prepare high molecular weight (>10⁵ g mol⁻¹) polymers containing five-membered heterocycles inside polymer backbones. The self-accelerating property of the DIBOD-based 1,3-dipolar cycloadditions enabled the use of step-growth polymerization to prepare high molecular weight polymers under stoichiometric imbalance conditions using an excess of DIBOD over bis-dipole monomers. The click characteristics of the DIBOD-based 1,3-dipolar cycloadditions assisted the step-growth polymerization so that polymers could be prepared under ambient and catalyst-free conditions. In addition, the varied five-membered heterocycle structures inside the backbones endowed the resultant polymers with distinctly unique properties and functions. The polymers with isoxazoline groups inside the backbones demonstrated self-degradation behavior, where higher molecular weights resulted in greater degradation. The polymers with pyrazole groups inside the backbones had excellent thermal properties: the decomposition temperature at 5% weight loss could reach 576 °C, the glass transition temperature could not be measured up to 400 °C, and the char yield at 800 °C was as high as 71%.

---

Introduction

The 1,3-dipolar cycloaddition of 1,3-dipoles and dipolarophiles (alkene or alkyne) is a classic reaction in organic chemistry for constructing five-membered heterocyclic rings. A 1,3-dipole is defined as a species with zwitterionic resonance structures that undergoes cycloaddition with dipolarophiles in a [3 + 2] process.¹ Since the first 1,3-dipole of diazoacetic ester was discovered in 1883,² dozens of 1,3-dipoles have been reported and their reactivities with dipolarophiles have been systematically investigated.¹ Basically, these 1,3-dipoles can be categorized into three types: allyl anion type, propargyl/allenyl anion type, and mesoionic compound.³ To date, 1,3-dipolar cycloaddition reactions have been widely used in the fields of biochemistry,^4–7 polymer synthesis,^8,9,10 and material science.^11–14

Due to its high reactive efficiency and capability to form five-membered heterocycles, 1,3-dipolar cycloaddition should theoretically play a vital role in step-growth polymerization to prepare high performance polymers containing five-membered heterocycles inside backbones. As a matter of fact, research into 1,3-dipolar cycloaddition in step-growth polymerization was concentrated in the 1960s and 1970s. At that time, cycloaddition reactions involving varied 1,3-dipoles such as sydnone,^15,16 nitrilimine,¹⁷ nitrile oxide,^18,19 nitrone,²⁰ diazo,²¹ azides²¹ and dienophiles like p/m-diethynylbenzene, p-benzoquinone, and bis-maleimides were well explored in step-growth polymerization. A series of polymers with good thermal stabilities was developed because of the introduction of five-membered heterocycles and benzene rings into the polymer backbones. After this period of focus, however, the research into 1,3-dipolar cycloaddition in step-growth polymerization diminished and no high-performance polymers were developed that were widely applied in practice. The main reason lies in the fact that the traditional 1,3-dipolar cycloaddition reactions used at that time shared the distinct disadvantages of relatively low reaction efficiency, unavoidable side reactions, and harsh reaction conditions. This led to resultant polymers with low molecular weights, usually less than 2 × 10⁴ g mol⁻¹. Such low molecular weights seriously deteriorated the physical and mechanical properties of the polymers and limited their applications in practice.

Fortunately, the concept of click chemistry introduced by Sharpless in 2001 rejuvenated the application of 1,3-dipolar cycloaddition reactions in step-growth polymerization.²² By appropriately matching the 1,3-dipoles to dipolarophiles, several 1,3-dipolar cycloaddition reactions could meet the criterion of click chemistry with characteristics such as quantitative reaction yield, good selectivity, and mild reaction conditions.^23–25 Among them, azide-based cycloaddition reactions have received unprecedented attention because the azide dipole is easy to prepare and handle at ambient temperature. For example, Cu^I-catalyzed azide–alkyne cycloaddition (CuAAC) has been used in step-growth polymerization for preparing poly(triazole)s with interesting photoelectronic and biological properties.^26–31 In addition, the metal-free azide–alkyne cycloadditions have been explored to synthesize poly(triazole)s, in which the activated azide dipole such as azidoperfluorobenzene³² or activated alkyne like aroylacetylene³³ has been used as a reactive group and the polymerization has been performed at the high temperature of 100 °C. Although these 1,3-dipolar cycloaddition click reactions could increase polymer molecular weights up to 5 × 10⁴ g mol⁻¹, the preparation of polymers with molecular weights above 10⁵ g mol⁻¹ was still difficult to achieve. This is caused by the inherent shortcoming of traditional homogeneous step-growth polymerization. Based on the concept of equal reactivity of reactive groups, the Carothers–Flory theory predicts that high molecular weight polymers can be prepared from traditional step-growth polymerization only by keeping a strict stoichiometry of reactive groups during polymerization, and simultaneously achieving an extremely high conversion of reactive groups after polymerization.^34,35 The exact stoichiometry of reactive groups, however, is hard to achieve practically due to monomer impurity and side reactions of polymerization. Moreover, even in the presence of stoichiometry, it is difficult to reach an extremely high conversion of reactive groups in a limited polymerization time.

Based on a double-strain-promoted azide–alkyne cycloaddition (DSPAAC) reaction,^36–40 we recently developed a stoichiometric imbalance-promoted step-growth polymerization method specifically to prepare polymers with molecular weights above 10⁵ g mol⁻¹.^41–43 In this approach, sym-dibenzo-1,5-cyclooctadiene-3,7-diyne (DIBOD) and bis-azide dipoles were used as monomer pairs. In DSPAAC, the reaction of an azide dipole and the first alkyne moiety of DIBOD activated the second alkyne moiety, which then reacted with an azide dipole much faster than the original alkyne groups of DIBOD had.^36,37 As a result, the DSPAAC reaction demonstrated an interesting self-accelerating property, which could significantly reduce the problem of excess DIBOD on the reaction of the azide dipole and activated alkyne moiety of DIBOD. In a polymerization reaction, the self-accelerating DSPAAC reaction ingeniously eliminated the fundamental concept of equal reactivity of reactive groups seen in traditional step-growth polymerization. In this situation, the usage of excess DIBOD could not only accelerate the polymerization reaction but also facilitate a quantitative conversion of azide dipoles during polymerization.⁴¹

To date, the azide group is the only reported 1,3-dipole to demonstrate the self-accelerating property when reacting with DIBOD. Considering the similar reaction mechanism of 1,3-dipolar cycloaddition, we believed that the cycloaddition of DIBOD and other 1,3-dipoles should also demonstrate the self-accelerating property. In the first part of this paper, we explore the cycloaddition of DIBOD and three types of 1,3-dipoles, namely allyl anions, propargyl/allenyl anions, and mesoionic compounds, which indeed demonstrated the self-accelerating click reaction property. Based on this, we further applied this library of self-accelerating click reactions in step-growth polymerization and built a series of stoichiometric imbalance-promoted step-growth polymerization methods. They could all efficiently prepare polymers with molecular weights above 10⁵ g mol⁻¹ using a molar excess of DIBOD to 1,3-dipole monomer. In addition, the different five-membered heterocycles inside the backbones endowed the resultant polymers with diverse functional properties and potential applications.

Results and discussion

1. Evaluating the self-accelerating property of 1,3-dipolar cycloaddition reactions between DIBOD and varied 1,3-dipoles

Diazo, sydnone, and nitrone groups were randomly selected as representatives of the three kinds of propargyl/allenyl anionic, mesoionic, and allyl anionic 1,3-dipoles, whose reactions with DIBOD were used to demonstrate the universal self-accelerating property of DIBOD-based 1,3-dipolar cycloaddition. Scheme 1 displays the model reactions performed under ambient conditions with [1,3-dipole]₀/[alkyne]₀ = 0.625. The rate constant ratio (K = k₂/k₁) of the second (k₂) and first (k₁) dipole–alkyne cycloaddition reactions was calculated to evaluate the self-accelerating property of these reactions. The detailed calculation process is illustrated in the Experimental section of the ESI and in Fig. S1 to S6.† Taking the cycloaddition of DIBOD and diazo compound 2 (Scheme 1) as an example, Fig. S1C† displays the ¹H-NMR spectrum of the reaction mixture at 20 min. All proton signals could be accurately ascribed to the substrates of DIBOD (Fig. S1A†) and diazo compound 2 (Fig. S1B†) and the bis-cycloaddition product 8. No extra proton signals were observed for the presumed mono-cycloaddition intermediate 5. In addition, the isolated yield of the final bis-cycloaddition product 8 was around 98%, suggesting the click property of cycloaddition between DIBOD and diazo compound 2. Based on the time-dependent consumption of DIBOD from ¹H-NMR characterization, the K and k₁ values were calculated as 180 and 3.14 × 10⁻² M⁻¹ s⁻¹ for the cycloaddition of DIBOD and diazo compound 2 in DMSO-d₆ (Fig. S2†). Similarly, the self-accelerating click property was demonstrated for the 1,3-dipolar cycloadditions between DIBOD and sydnone compound 3 (Fig. S3†), and between DIBOD and nitrone compound 4 (Fig. S5†). The corresponding K and k₁ values were calculated as 246 and 1.39 × 10⁻¹ M⁻¹ s⁻¹ for the cycloaddition of DIBOD and sydnone compound 3 (Fig. S4†), and 230 and 1.14 × 10⁻¹ M⁻¹ s⁻¹ for the cycloaddition of DIBOD and nitrone compound 4 (Fig. S6†).

Scheme 1 Model 1,3-dipolar cycloaddition reactions between DIBOD and varied 1,3-dipoles.

Besides, the isolated yields of the final bis-cycloaddition products 9 and 10 were all as high as 98%, in line with the characteristic of click chemistry.

2. Imbalance-promoted step-growth polymerization from self-accelerating DIBOD-based 1,3-dipolar cycloaddition reactions

2.1 Imbalance-promoted step-growth polymerization of DIBOD and M1. Typical 1,3-dipolar difunctional monomers were designed with diazo (M1), sydnone (M2), and nitrone (M3) end groups (Scheme 2). The step-growth polymerizations of 1,3-dipolar difunctional monomers and DIBOD were all performed under ambient conditions without catalysts (Scheme 2). The stoichiometric imbalance-promoted property of these DIBOD-based step-growth polymerization reactions was exemplified in detail using M1 and DIBOD as a monomer pair, in which the effect on polymerization behaviour of the monomer ratio (S = [DIBOD]₀/[M1]₀) and the M1 monomer concentration ([M1]₀) were studied thoroughly.

Scheme 2 Imbalance-promoted step-growth polymerization based on self-accelerating DIBOD-based 1,3-dipolar cycloaddition reactions; one isomer is drawn to show the macromolecular structures.

The effect of the monomer ratio S on the step-growth polymerization of DIBOD and M1 was studied by setting [M1]₀ at a constant 0.2 M in DMF. Fig. 1A displays the gel permeation chromatography (GPC) curves of Poly 1 prepared with different S values of 0.91, 1.0, 1.1, 1.2, 1.5, and 2.0 (runs Poly 1-1 to Poly 1-6, respectively), whose quantitative analyses are detailed in Table 1. Fig. S7† shows that for all cases with S ≠ 1, the GPC curves of Poly 1 overlapped after 3 h and 6 h reaction times indicating that these polymerizations finished efficiently within 3 h. In the case of S = 1, however, a 15 h reaction time was required to finish the polymerization. Thus, the use of a stoichiometric imbalance could significantly speed up the step-growth polymerization rate. The UV-Vis characterization (Fig. S8†) showed the disappearance of the DIBOD absorption peak at 272 nm in the case with an S value of 0.91 (Fig. S8A†), while FT-IR characterization showed the disappearance of the diazo absorption peak at 2105 cm⁻¹ in the cases with S values of 1.1, 1.2, 1.5, and 2.0 (Fig. S8B†) after 3 h polymerization. These results indicated that the usage of an excess of one monomer could efficiently consume the other deficient polymerization groups under the given polymerization conditions. The GPC curves in Fig. 1A clearly demonstrated that the polymerization behavior of Poly 1 heavily depended on the monomer ratio. For the reaction deficient in the DIBOD monomer with an S value of 0.91, the polymerization produced Poly 1-1 having a broad GPC peak distribution (Fig. 1A, black curve) with a relatively low M_n,GPC of 13 150 (Table 1, run Poly 1-1). This is consistent with the theoretical M_n of polymers predicted by the well-known equation [(1 + r)/(1 + r − 2rp)] × M₀ for traditional step-growth polymerization using AA and BB type monomers. In this equation, r, p, and M₀ represent the monomer feed ratio, the extent of the polymerization reaction, and the average molecular weight of the two monomers, respectively. With r = S = 0.91 and p = 1, the theoretical M_n calculated from the classical equation was 7085 (Table 1, run Poly 1-1), closely related to the measured M_n,GPC of 13 150. The value of p was assumed to be 1 since the deficient DIBOD was completely consumed in the polymerization. As a result, the polymerization behavior of Poly 1 followed the classical step-growth polymerization theory in the presence of deficient DIBOD, and the usage of deficient DIBOD seriously decreased the molecular weight of Poly 1. In the case of S = 1, polymerization produced Poly 1-2 having a GPC curve with a bimodal distribution. As shown in Fig. 1A (red curve), the major peak corresponding to 81.1% by weight of products at the shorter elution time had an apparent M_n,GPC of 218 200 (Table 1, run Poly 1-2). In addition, in the cases using excess DIBOD (S > 1), the polymerization products Poly 1 all had GPC curves with bimodal distributions. For the GPC curve of Poly 1-3 from S = 1.1 (Fig. 1A, blue curve), the major peak (78.0% of products) at the shorter elution time had an apparent M_n,GPC of 207 300 (Table 1, run Poly 1-3). Although slightly lower than that of Poly 1-2 (218 200), it was much higher than the theoretical M_n of 7085 predicted by the classical equation with r = 1/S = 0.91 and p = 1. The p value was assumed to be 1 since the deficient diazo group was totally consumed in the polymerization. When S values were increased to 1.2, 1.5, and 2.0, the major peaks representing about 80% of the products at the shorter elution times had apparent M_n,GPC values of 173 800, 152 300, and 118 100 for the resultant Poly 1-4 (Fig. 1A, magenta curve), Poly 1-5 (Fig. 1A, olive curve), and Poly 1-6 (Fig. 1A, navy curve). These values were all much higher than the respective theoretical M_n values of 3632 (Table 1, run Poly 1-4), 1687 (Table 1, run Poly 1-5), and 1012 (Table 1, run Poly 1-6) predicted by the classical equation with r = 1/S and p = 1. Thus, the presence of excess DIBOD in this stoichiometric imbalance-promoted step-growth polymerization did not reduce the molecular weight of Poly 1 but rather increased it significantly. Furthermore, although the increase of the S value above 1 resulted in a slight decrease of the Poly 1 molecular weight, the polymerization time required to prepare high molecular weight Poly 1 could be significantly shortened. In Fig. S7,† the overlapping GPC curves for the different reaction times suggested that the polymerization finished within 15 h for Poly 1-2 (S = 1), but within only 3 h and 1 h for Poly 1-3 (S = 1.1) and Poly 1-5 (S = 1.5), respectively. In addition to the major peaks for high molecular weight products, the GPC curves (Fig. 1A) of Poly 1 prepared with S > 1 showed minor peaks for lower molecular weight products, which could be assigned to the cyclic oligomers obtained in the initial stage of polymerization.⁴¹ To demonstrate this, the oligomers were isolated from Poly 1-8 by precipitation in methanol; their GPC curve is shown in Fig. S9A† (blue). Fig. S10† shows the corresponding MALDI-TOF MS of the oligomers, in which the peak distributions could be accurately assigned to cyclic Poly 1 ionized with Na⁺ and K⁺.

Fig. 1 (A) GPC curves of Poly 1 prepared from [M1]₀ = 0.2 M and S = 0.91 (black), 1.0 (red), 1.1 (blue), 1.2 (magenta), 1.5 (olive), and 2.0 (navy), in which the polymerization time was 15 h in the case of S = 1.0, and 3 h for the other ratios. (B) GPC curves of Poly 1 prepared from S = 1.1 and [M1]₀ = 0.05 M (black), 0.1 M (red), 0.2 M (blue), 0.3 M (magenta), and 0.5 M (olive), in which the polymerization time was 6 h in the cases of [M1]₀ = 0.05 M and 0.1 M, and 3 h for the other concentrations. (C) GPC curves of raw Poly 1-10 (black) prepared from [M1]₀ = 0.5 M and S = 1.1 after 3 h, and precipitated Poly 1-10 (red). (D) GPC curves of precipitated Poly 1 prepared from [M1]₀ = 0.5 M, S = 1.1, and [mono-diazo compound 2]₀/[M1]₀ = 0% (black), 1% (red), 2.5% (blue), 5% (magenta), and 10% (olive) after 3 h. (E) GPC curves of raw Poly 2-1 (black) prepared from [M2]₀ = 0.5 M and S = 0.91 after 1 h, raw Poly 2-2 (red) prepared from [M2]₀ = 0.5 M and S = 1.1 after 1 h, and precipitated Poly 2-2 (blue). (F) GPC curves of raw Poly 3-1 (black) prepared from [M2]₀ = 0.5 M and S = 0.91 after 0.5 h, raw Poly 3-2 (red) prepared from [M2]₀ = 0.5 M and S = 1.1 after 0.5 h, and freshly precipitated Poly 3-2 (blue). The polymerizations were carried out under ambient conditions; DMF was used as the solvent for Poly 1 and Poly 2, and CHCl₃ was used as the solvent for Poly 3.

Table 1 The synthesis and characterization of Poly 1

Polymer^a	[M]0 ^b (mol L^−1)	S	Ratio^d (%)	M n,theory ^e (g mol^−1)	M n,GPC ^f (g mol^−1)	Đ	Polymer ^f (%)	Oligomer^f (%)
a Polymerization was performed under ambient conditions in DMF. b [Bis-diazo monomer]0. c [DIBOD]0/[bis-diazo monomer]0. d The initial molar ratio of the mono-diazo compound to the bis-diazo compound, [benzyl diazoacetate]0/[M1]0. e Determined from the equation M0 × (1 + r)/(1 + r − 2rp) for Poly 1-1 to Poly 1-10 with p = 1 and r = 1/S; determined from the equation M0 × 2/(2 − pfavg) for Poly 1-11 to Poly 1-14 with p = 1 and favg = 2 × (2 + ratio)/(1 + 1.1 + ratio). f Determined via GPC characterization with a refractive index (RI) detector, where THF was used as the eluent and polystyrene standards were used for calibration.
Poly 1-1	0.2	0.91		7085	13 150	1.44
Poly 1-2	0.2	1.0		∞	218 200	1.84	81.1	18.9
Poly 1-3	0.2	1.1		7085	207 300	1.83	78.0	22.0
Poly 1-4	0.2	1.2		3632	173 800	1.80	79.4	20.6
Poly 1-5	0.2	1.5		1687	152 300	1.80	81.7	18.3
Poly 1-6	0.2	2.0		1012	118 100	1.77	79.9	20.1
Poly 1-7	0.05	1.1		7085	51 710	1.66	39.8	60.2
Poly 1-8	0.1	1.1		7085	104 100	1.64	62.0	38.0
Poly 1-9	0.3	1.1		7085	242 900	1.86	85.2	14.8
Poly 1-10	0.5	1.1		7085	293 800	1.85	90.5	9.5
Poly 1-11	0.5	1.1	1	7119	101 400	1.77	91.7	8.3
Poly 1-12	0.5	1.1	2.5	7170	47 780	1.77	91.9	8.1
Poly 1-13	0.5	1.1	5	7254	25 940	1.66	92.4	7.6
Poly 1-14	0.5	1.1	10	7423	18 800	1.62	92.9	7.1

---

The effect of monomer concentration on the step-growth polymerization of DIBOD and M1 was studied by setting the S value at a constant 1.1. In Fig. S11,† the overlapping GPC curves of Poly 1 from different [M1]₀ at the designated reaction times demonstrated that an increase in [M1]₀ shortened the polymerization time for the formation of Poly 1. The polymerization finished within 9 h for [M1]₀ = 0.05 M but within only 1 h for [M1]₀ = 0.5 M. Fig. 1B displays the GPC curves of Poly 1 prepared from different [M1]₀, and the quantitative analyses are shown in Table 1, runs Poly 1-7 to Poly 1-10. As shown in Fig. 1B, all GPC curves had bimodal distributions. The major peaks at the shorter elution times were ascribed to the high molecular weight Poly 1, while the minor peaks at longer elution times were assigned to the low molecular weight cyclic oligomers formed at the beginning of polymerization.⁴¹ With increasing [M1]₀, the proportion of cyclic oligomers gradually decreased while the molecular weight of Poly 1 continuously increased. It is known that intramolecular cyclization is favored at low polymerization concentrations to form stable cyclic oligomers and this wastes the monomers that are supposed to synthesize high molecular weight polymers. The increase of monomer concentration can efficiently inhibit the cyclization at the beginning of polymerization and thereby increase the molecular weight of the polymers. As shown in Fig. 1B, the highest [M1]₀ of 0.5 M produced Poly 1-10 with the lowest proportion (9.53% for the minor peak) of cyclic oligomers and the highest M_n,GPC of 293 800 for the main polymer peak (Table 1, run Poly 1-10). Although the increase of [M1]₀ could not eliminate the formation of cyclic oligomers during polymerization, they could be conveniently removed by a simple precipitation process. As shown in Fig. 1C, after precipitating the raw Poly 1-10 in methanol three times, the cyclic oligomers were completely removed and a unimodal GPC curve was observed for the precipitated Poly 1-10. The absolute weight-average molecular weight (M_w,absolute) of precipitated Poly 1-10 was measured to be as high as 599 800 by GPC characterization with a multi-angle laser light scattering detector (GPC-MALLS); this is the highest molecular weight reported for polymers prepared from the diazo 1,3-dipolar cycloaddition reaction. The corresponding ¹H-NMR and ¹³C-NMR spectra of the precipitated Poly 1-10 are shown in Fig. 2A and Fig. S12.†

Fig. 2 ¹H-NMR spectra of precipitated Poly 1-1 (A) in DMSO-d₆, and Poly 2-2 (B) and Poly 3-2 (C) in CDCl₃.

Besides roughly tuning the molecular weight of Poly 1 by changing the monomer ratio (Fig. 1A) and initial monomer concentration (Fig. 1B), mono-diazo compounds could be employed in the stoichiometric imbalance-promoted step-growth polymerization to finely manipulate the molecular weight over a very wide range. Taking the polymerization with [M1]₀ = 0.5 M and S = 1.1 as an example, benzyl diazoacetate could be utilized to manipulate the molecular weight of Poly 1 when it was copolymerized with M1 and DIBOD. As shown in Fig. S13,† all GPC curves of raw Poly 1 with different amounts of benzyl diazoacetate exhibited bimodal distributions, with similar proportions (about 90%) for the major high molecular weight Poly 1 peak at the shorter elution time (Table 1, run Poly 1-10 to Poly 1-14). This implied that the usage of benzyl diazoacetate did not affect the cyclization at the beginning of polymerization. As shown in Fig. 1D, all GPC curves of precipitated Poly 1 exhibited unimodal distributions but the peak position continuously moved in the lower molecular weight direction as the content of benzyl diazoacetate was increased. From Table 1 (run Poly 1-10 to Poly 1-14), the increase in [benzyl diazoacetate]₀/[M1]₀ from 0% to 10% decreased the M_n,GPC of the resultant Poly 1 from 293 800 to 18 800.

2.2 Imbalance-promoted step-growth polymerization of DIBOD with M2 and M3. Inspired by the successful preparation of high molecular weight Poly 1 based on 1,3-dipolar cycloaddition between the bis-diazo compound and DIBOD, the 1,3-dipolar cycloaddition reactions of DIBOD and sydnone/nitrone groups were explored to prepare polymers by step-growth polymerization. As shown in Scheme 1, difunctional monomers with sydnone (M2) and nitrone (M3) groups were polymerized with DIBOD to prepare the corresponding Poly 2 and Poly 3. The polymerization reactions were all performed with a constant initial dipole monomer concentration of 0.5 M and S values of 0.91 and 1.1 under ambient conditions.

The polymerization behavior of M2 and DIBOD was similar to that of M1 and DIBOD. Fig. S14† shows the GPC characterization of raw Poly 2 with a constant [M2]₀ of 0.5 M, in which the overlapping GPC curves after 1 h and 3 h reaction times indicated that the polymerization of M2 and DIBOD efficiently finished within 1 h regardless of the S value of 0.91 or 1.1. As shown in Fig. 1E, the reaction that was deficient in DIBOD (S = 0.91) produced the low molecular weight Poly 2-1 with a unimodal GPC peak (black) having a measured M_n,GPC of 20 060, which was close to the M_n,theory of 9608 calculated from the classical equation (Table 2, run Poly 2-1). The presence of excess DIBOD (S = 1.1), however, prepared Poly 2-2 with a bimodal GPC peak distribution (red). The major peak at the shorter elution time was from the high molecular weight Poly 2-2 with a measured M_n,GPC of 237 000, which was much larger than the M_n,theory of 9608 calculated from the classical equation (Table 2, run Poly 2-2). The minor peak at the longer elution time was from the low molecular weight cyclic oligomers (corresponding to 4% product), which could be completely removed by precipitation (Fig. 1E, blue curve). The ¹H-NMR and ¹³C-NMR spectra of the precipitated Poly 2-2 are shown in Fig. 2B and Fig. S15,† respectively. The M_w,absolute of the precipitated Poly 2-2 was measured as 370 000 by GPC-MALLS. To further confirm the cyclic topology of the low molecular weight oligomers, they were isolated from Poly 2-3 prepared with [M2]₀ = 0.1 M and S = 1.1 by precipitation in a mixture of methanol and acetone and the GPC curve is shown in Fig. S9B† (blue). Fig. S16† shows the corresponding MALDI-TOF MS, in which the two main peak distributions could be accurately assigned to cyclic Poly 2 ionized with Na⁺ and K⁺.

Table 2 The synthesis and characterization of Poly 2, 3, and 4

Polymer^a	[M]0 ^b (mol L^−1)	S	M n,theory ^d (g mol^−1)	M n,GPC ^e (g mol^−1)	Đ	Polymer^e (%)	Oligomer^e (%)
a Polymerization was performed under ambient conditions, where DMF was used for Poly 2, CHCl3 was used for Poly 3, and NMP was used for Poly 4. b [Bis-dipole monomer]0. c [DIBOD]0/[bis-dipole monomer]0. d Determined from the equation M0 × (1 + r)/(1 + r − 2rp) with p = 1 and r = 1/S. e Determined via GPC characterization with an RI detector, where THF was used as the eluent for Poly 2 and Poly 3, DMF was used as the eluent for Poly 4, and polystyrene standards were used for calibration.
Poly 2-1	0.5	0.91	9608	20 060	1.76
Poly 2-2	0.5	1.1	9608	237 000	2.11	96.0	4.0
Poly 2-3	0.1	1.1	9608	173 900	1.76	81.2	18.8
Poly 3-1	0.5	0.91	9041	15 670	1.53
Poly 3-2	0.5	1.1	9041	142 900	2.00	92.3	7.7
Poly3-3	0.1	1.1	9041	48 270	1.74	79.4	20.6
Poly 4	0.5	1.1	5654	181 300	1.73	95.3	4.7

---

Furthermore, the polymerization behavior of M3 and DIBOD was similar to that of M1 and DIBOD. In Fig. S17,† the overlapping GPC curves of raw Poly 3-1 with different reaction times of 0.5 h and 3 h indicated that the polymerization finished within 0.5 h using [M3]₀ = 0.5 M and S = 0.91. As shown in Fig. 1F, the reaction that was deficient in DIBOD (S = 0.91) produced the low molecular weight Poly 3-1 with a measured M_n,GPC of 15 670, which was close to the M_n,theory of 9041 calculated from the classical equation (Fig. 1F, black curve; Table 2, run Poly 3-1). The usage of excess DIBOD (S = 1.1) prepared high molecular weight Poly 3-2 corresponding to 92.3% product and having an M_n,GPC of 142 900, which was much higher than the M_n,theory of 9041 calculated from the classical equation (Fig. 1F, red curve; Table 2, run Poly 3-2). Moreover, only a small amount of lower molecular weight cyclic oligomers formed for this case (7.7% product), which could be removed completely by precipitation (Fig. 1F, blue curve). The ¹H-NMR and ¹³C-NMR spectra of the precipitated Poly 3-2 are shown in Fig. 2C and Fig. S18,† respectively. The M_w,absolute of precipitated Poly 3-2 was measured as 231 200 by GPC-MALLS. To further confirm the cyclic topology of the low molecular weight oligomers, they were isolated from Poly 3-3 prepared with [M3]₀ = 0.1 M and S = 1.1 by precipitation in methanol and the GPC curve is shown in Fig. S9C† (blue). Fig. S19† shows the corresponding MALDI-TOF MS, in which the main peak could be accurately assigned to cyclic Poly 3 ionized with H⁺.

In summary, a novel type of stoichiometric imbalance-promoted step-growth polymerization was successfully developed based on a library of DIBOD-based 1,3-dipolar cycloaddition reactions with a self-accelerating property. Thus, a series of high molecular weight polymers with varied molecular structures and functions were prepared.

3. Chemical and physical properties

3.1 Thermal properties of Poly 1 and Poly 2. The thermal properties of Poly 1 and Poly 2 were evaluated by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC). From Fig. 3A, the decomposition temperatures (T_d) at 5% weight loss were measured as 283 °C and 373 °C for precipitated Poly 1-10 (black curve) and Poly 2-2 (red curve), respectively, by TGA under a nitrogen atmosphere. From Fig. 3B, the glass transition temperatures (T_g) were measured as 171 °C and 196 °C for precipitated Poly 1-10 (black curve) and Poly 2-2 (red curve), respectively. This data demonstrated that Poly 1 and Poly 2 with the heterocycles and benzene rings inside the polymer backbones indeed had good thermal stability. It is known that the introduction of benzene and heterocyclic rings with high molar contents inside the structure can significantly increase the thermal properties of polymers.³⁵ On the basis of this concept, a bis-sydnone monomer of M4 ⁴⁴ (Fig. 3C) was synthesized containing a diphenylether linker between both sydnone terminals to contrast with M2. After polymerizing with DIBOD, the resultant Poly 4 was expected to have enhanced thermal stability because of the increased molar content of phenyl and pyrazole rings inside the macromolecular structure when compared to that of Poly 2. The step-growth polymerization of DIBOD and M4 was performed in N-methyl-2-pyrrolidone (NMP) under ambient conditions for 1 h with [M4]₀ = 0.5 M and S = 1.1 (Fig. 3C). Fig. S20† shows the bimodal GPC curve (black) of raw Poly 4, in which a low content (4.7%) of cyclic oligomers was observed in the low molecular weight region, which could be completely removed by precipitation. As shown in Fig. S20,† a unimodal GPC peak (red) was observed for precipitated Poly 4 with an M_n,GPC of 181 300 (Table 2, run Poly 4). The corresponding ¹H-NMR and ¹³C-NMR spectra of precipitated Poly 4 are shown in Fig. S21 and S22,† respectively. As shown by the TGA characterization (Fig. 3A, blue curve), Poly 4 had a T_d of 576 °C at 5% weight loss under a nitrogen atmosphere and a high char yield of 71% at 800 °C. According to the DSC characterization (Fig. 3B, blue curve), Poly 4 did not show a glass transition in the measured temperature range of up to 400 °C. In addition, Poly 4 had good ability to form films (Fig. S23†) due to its high molecular weight and good solubility in common polar solvents such as DMF, DMAc and NMP.

Fig. 3 (A) TGA curves of precipitated Poly 1-10 (black), Poly 2-2 (red) and Poly 4 (blue). (B) DSC curves of precipitated Poly 1-10 (black), Poly 2-2 (red), and Poly 4 (blue). (C) The imbalance-promoted step-growth polymerization of M4 and DIBOD; one isomer is drawn to show the macromolecular structures.

3.2 Self-degradation of Poly 3. Fig. 4A shows the time-dependent polymerization behavior of Poly 3-2, in which the molecular weight of Poly 3-2 increased significantly and reached a maximum in a short reaction time of 0.5 h. After that, the molecular weight decreased continuously but the rate of decrease slowed down with increasing polymerization time (Fig. S24†). This indicated that the molecular structure of Poly 3-2 was not stable and that its degradation happened quickly after the formation of the high molecular weight Poly 3-2 during polymerization. Fig. S25A† shows the GPC characterization of the precipitated Poly 3-2 stored in the solid state. The overlapping GPC curves of precipitated Poly 3-2 after storage times of 12 h and 4 days indicated that the molecular structure of Poly 3-2 showed good stability in the solid state for at least 4 days. When the storage time was increased to 8 and 20 days, however, the GPC peak shifted in the lower molecular weight direction, suggesting that the degradation slowly happened even for the polymer in the solid state.

Fig. 4 (A) GPC curves of raw Poly 3-2 prepared from [M3]₀ = 0.5 M and S = 1.1 with different polymerization times. (B) The DSC curve of precipitated Poly 3-2. (C) GPC curves of freshly precipitated Poly 3-2 before (black) and after (red) DSC measurements. (D) GPC curves of raw Poly 3-1 prepared from [M3]₀ = 0.5 M and S = 0.91 with different polymerization times.

Fig. 4B shows the DSC curve of freshly precipitated Poly 3-2 after 0.5 h reaction time, in which a sharp exothermic peak appeared between 100 and 200 °C peaking at 166 °C. The exothermic peak, however, never appeared in the following cooling and reheating scans (Fig. S26†). After DSC measurement, the white solid powder of freshly precipitated Poly 3-2 became a brownish red bulk solid. This indicated that the molecular structure of purified Poly 3-2 had been chemically changed during the first heating process. Fig. 4C shows the GPC curve (red) of Poly 3-2 after DSC measurement, in which the peak position was completely shifted in the lower molecular weight direction compared to that (black curve) of freshly precipitated Poly 3-2. The corresponding M_n,GPC was measured as 5010, which was much lower than the value of 142 900 for the freshly precipitated Poly 3-2 and even lower than 18 420 value of the degraded Poly 3-2 after 4 days in solution (Fig. 4A, violet curve). This clearly indicated that the thermal degradation of Poly 3-2 efficiently happened during the first DSC heating scan and that the thermal degradation rate was much faster than that found in the solution state. The TGA characterization of freshly precipitated Poly 3-2 (Fig. S27†) showed that Poly 3-2 started to lose weight at 220 °C, which was a much higher temperature than the exothermic peak temperature of 166 °C found in the DSC characterization. This indicated that the thermal degradation of Poly 3-2 during the first DSC heating scan released no volatile gas but only cleaved the polymer backbone.

According to the literature,⁴⁵ the N–O bond of the isoxazoline ring formed from the cycloaddition of a nitrone dipole and benzannulated cyclooctyne showed poor stability; it could easily break and rearrange to form the azo-methine ylide. The subsequent hydrolysis of the azo-methine ylide intermediate produced the final α-amino-ketone and benzaldehyde compounds. In addition, the difference between the macromolecular structures of Poly 2 and Poly 3 lay only in the five-membered heterocyclic rings of the polymer backbones. The structural stability of Poly 2 clearly indicated that the degradation of Poly 3 came from the cleavage of the isoxazoline ring inside the Poly 3 backbones. Fig. S28B† shows the ¹H-NMR spectrum of the precipitated Poly 3-2 in CDCl₃ after 4 days. In comparison to that of freshly precipitated Poly 3-2 (Fig. S28A†), a new proton signal appeared at 10.09 ppm that could be ascribed to the resultant aldehyde end group. This strongly supports the above presumed degradation mechanism of Poly 3.

Furthermore, the self-degradation behavior of Poly 3 was closely related to its molecular weight. Fig. 4D shows the time-dependent polymerization behavior of Poly 3-1. The overlapping GPC curves of Poly 3-1 after polymerization times of 0.5 to 24 h demonstrated that Poly 3-1 with its lower molecular weight presented much better stability compared to the higher molecular weight Poly 3-2. This is reasonable since the higher molecular weight Poly 3-2 had larger numbers of labile isoxazoline rings inside single polymer chains compared to the lower molecular weight Poly 3-1. This led to a higher degradation probability of Poly 3-2 and the relatively better stability of Poly 3-1. Fig. S25B† shows the GPC characterization of the precipitated Poly 3-1 stored in the solid state. The overlapping GPC curves of precipitated Poly 3-1 after storage times of 12 h and 20 days suggested that the low molecular weight Poly 3-1 showed good stability in the solid state for at least 20 days.

Conclusions

A novel library of self-accelerating click reactions was developed by reacting DIBOD with varied 1,3-dipoles, such as diazo, sydnone, and nitrone groups. In these DIBOD-based 1,3-dipolar cycloaddition reactions, the reaction of the 1,3-dipole and the first alkyne moiety of DIBOD activated the second alkyne moiety in situ, which then reacted with the 1,3-dipole much faster than the original alkyne groups of DIBOD had. As a result, the mono-cycloaddition reactive intermediate could not be isolated from these reactions and they always produced bis-cycloaddition products with quantitative yields. Based on this library of self-accelerating click reactions, a series of stoichiometric imbalance-promoted step-growth polymerization methods was developed using DIBOD and various bis-dipole compounds as monomer pairs. The distinct advantages of these novel step-growth polymerization methods can be summarized as follows: by virtue of the self-accelerating property of the DIBOD-based 1,3-dipolar cycloaddition polymerization reactions, it was possible to efficiently prepare polymers with heterocycle-rich backbones and high molecular weights above 10⁵ g mol⁻¹ under stoichiometric imbalance conditions with excess DIBOD over bis-dipole monomers. Taking advantage of the click nature of the DIBOD-based 1,3-dipolar cycloaddition polymerization reactions, high molecular weight polymers could conveniently be produced under ambient and catalyst-free conditions in less than 1 h. The wide diversity of the 1,3-dipoles and linker groups of bis-dipole monomers enabled the preparation of high molecular weight polymers with significantly varied macromolecular structures and correspondingly unique properties and functions.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Generous support was primarily provided by the National Science Foundation of China (21622406 and 21871273) and the National Key Research and Development Program of China (No. 2019YFA0210400).

Notes and references

1. R. Huisgen, Angew. Chem., Int. Ed. Engl., 1963, 2, 565–598.
2. T. Curtius, Ber. Dtsch. Chem. Ges., 1883, 16, 2230–2231.
3. K. V. Gothelf and K. A. Jørgensen, Chem. Rev., 1998, 98, 863–909.
4. J. C. Jewett and C. R. Bertozzi, Chem. Soc. Rev., 2010, 39, 1272–1279.
5. O. Boutureira and G. J. Bernardes, Chem. Rev., 2015, 115, 2174–2195.
6. P. An, T. M. Lewandowski, T. G. Erbay, P. Liu and Q. Lin, J. Am. Chem. Soc., 2018, 140, 4860–4868.
7. E. Decuypère, L. Plougastel, D. Audisio and F. Taran, Chem. Commun., 2017, 53, 11515–11527.
8. G. W. Goodall and W. Hayes, Chem. Soc. Rev., 2006, 35, 280–312.
9. K. Kacprzak, I. Skiera, M. Piasecka and Z. Paryzek, Chem. Rev., 2016, 116, 5689–5743.
10. D. Estupinan, T. Gegenhuber, J. P. Blinco, C. Barner-Kowollik and L. Barner, ACS Macro Lett., 2017, 6, 229–234.
11. A. Marrocchi, A. Facchetti, D. Lanari, S. Santoroa and L. Vaccaro, Chem. Sci., 2016, 7, 6298–6308.
12. G. Guo, J. Sun, C. Zhao, Y. Liu and C. M. Liu, Green Chem., 2016, 18, 1278–1286.
13. J. F. Lutz, Angew. Chem., Int. Ed., 2007, 46, 1018–1025.
14. S. M. Hodgson, E. Bakaic, S. A. Stewart, T. Hoare and A. Adronov, Biomacromolecules, 2016, 17, 1093–1100.
15. J. K. Stille and M. A. Bedford, J. Polym. Sci., Part A: Polym. Chem., 1968, 6, 2331–2342.
16. K. K. Sun, Macromolecules, 1987, 20, 726–729.
17. J. K. Stille and F. W. Harris, J. Polym. Sci., Part A: Polym. Chem., 1968, 6, 2317–2330.
18. C. G. Overberger and S. Fujimoto, J. Polym. Sci., Part C: Polym. Symp., 1967, 16, 4161–4168.
19. Y. Iwakura, K. Uno, S. J. Hong and T. Hongu, Polym. J., 1971, 2, 36–42.
20. L. Vretik and H. Ritter, Macromolecules, 2003, 36, 6340–6345.
21. Y. Gilliams and G. Smets, Macromol. Chem. Phys., 1968, 117, 1–11.
22. H. C. Kolb, M. G. Finn and K. B. Sharpless, Angew. Chem., Int. Ed., 2001, 40, 2004–2021.
23. B. C. Sanders, F. Friscourt, P. A. Ledin, N. E. Mbua, S. Arumugam, J. Guo, T. J. Boltje, V. V. Popik and G. J. Boons, J. Am. Chem. Soc., 2011, 133, 949–957.
24. E. Decuypère, L. Plougastel, D. Audisioa and F. Taran, Chem. Commun., 2017, 53, 11515–11527.
25. C. Kesornpun, T. Aree, C. Mahidol, S. Ruchirawat and P. Kittakoop, Angew. Chem., Int. Ed., 2016, 55, 3997–4001.
26. D. J. V. C. van Steenis, O. R. P. David, G. P. F. van Strikdonck, J. H. van Maarseveen and J. N. H. Reek, Chem. Commun., 2005, 4333–4335.
27. S. Bakbak, P. J. Leech, B. E. Carson, S. Saxena, W. P. King and U. H. F. Bunz, Macromolecules, 2006, 39, 6793–6795.
28. N. G. Angelo and P. S. Arora, J. Am. Chem. Soc., 2005, 127, 17134–17135.
29. K. N. Lau, H. F. Chow, M. C. Chan and K. W. Wong, Angew. Chem., Int. Ed., 2008, 47, 6912–6916.
30. G. Lu, S. Lam and K. Burgess, Chem. Commun., 2006, 15, 1652–1654.
31. E. Schwartz, K. Breitenkamp and V. V. Fokin, Macromolecules, 2011, 44, 4735–4741.
32. Q. Wang, H. K. Li, Q. Wei, J. Z. Sun, J. Wang, X. A. Zhang, A. J. Qin and B. Z. Tang, Polym. Chem., 2013, 4, 1396–1401.
33. A. J. Qin, C. K. W. Jim, W. X. Lu, J. W. Y. Lam, M. Häussler, Y. Q. Dong, H. H. Y. Sung, I. D. Williams, G. K. L. Wong and B. Z. Tang, Macromolecules, 2007, 40, 2308–2317.
34. P. J. Flory, Chapter III, in Principles of polymer chemistry, Cornell University Press, Ithaca, New York, 1953.
35. G. Odian, Chapter 2, in Principles of polymerization, John Wiley & Sons, New York, 4th edn, 2004.
36. I. Kii, A. Shiraishi, T. Hiramatsu, T. Matsushita, H. Uekusa, S. Yoshida, M. Yamamoto, A. Kudo, M. Hagiwara and T. Hosoya, Org. Biomol. Chem., 2010, 8, 4051–5055.
37. P. Sun, J. Q. Chen, J. A. Liu and K. Zhang, Macromolecules, 2017, 50, 1463–1472.
38. K. Sharma, A. V. Strizhak, E. Fowler, X. Wang, W. Xu, C. H. Jensen, Y. Wu, H. F. Sore, Y. H. Lau, M. Hyvönen, L. S. Itzhaki and D. R. Spring, Org. Biomol. Chem., 2019, 17, 8014–8018.
39. D. A. Sutton and V. V. Popik, J. Org. Chem., 2016, 81, 8850–8857.
40. Y. H. Lau, Y. Wu, M. Rossmann, B. X. Tan, P. de Andrade, Y. S. Tan, C. Verma, G. J. McKenzie, A. R. Venkitaraman, M. Hyvönen and D. R. Spring, Angew. Chem., Int. Ed., 2015, 54, 15410–15413.
41. J. Q. Chen, L. Xiang, X. F. Liu, X. P. Liu and K. Zhang, Macromolecules, 2017, 50, 5790–5797.
42. L. C. Zhang, X. Ren, Y. X. Zhang and K. Zhang, ACS Macro Lett., 2019, 8, 948–954.
43. L. Xiang, Z. Li, J. A. Liu, J. Q. Chen, M. H. Zhang, Y. Wu and K. Zhang, Polym. Chem., 2018, 9, 4036–4043.
44. N. V. Handa, S. Li, J. A. Gerbec, N. Sumitani, C. J. Hawker and D. Klinger, J. Am. Chem. Soc., 2016, 138, 6400–6403.
45. D. A. MacKenzie, A. R. Sherratt, M. Chigrinova, L. L. Cheung and J. P. Pezacki, Curr. Opin. Chem. Biol., 2014, 21, 81–88.

---

Footnote

† Electronic supplementary information (ESI) available: An Experimental section and Fig. S1–S19. See DOI: 10.1039/c9py01362h

---