3.1. Inhibitors and Precautions
Anionic polymerization is routinely inhibited by addition of parts per million of strong organic or mineral acids. These act by proton transfer to terminate anionic or zwitterionic species, before significant chain growth occurs. Anionic inhibitors have included Lewis acids, such as BF
3 complexes [
48], as well as, acetic acid [
50,
51,
52,
53], dichloroacetic acid [
54], trifluoroacetic acid [
55], methanesulfonic acid [
47], and 1,3-propanesultone [
51,
56,
57,
58]. Excessive acidity should, however, be avoided, due to the potential for the hydrolysis of the monomer/polymer ester functionality leading to interference with the polymerization (see above).
Commercial CAs will often contain radical inhibitors, such as hydroquinone, catechol,
p-methoxyphenol, butylated hydroxyanisole, and related phenolic compounds to suppress radical polymerization, which can be triggered with high temperatures or UV-light. In order to study the radical polymerization of CAs, inhibitors should be removed by distillation prior to commencing the reaction [
58]. Any glassware that will come in contact with CAs should be acid washed, acetone rinsed, and oven dried before use. Distilled monomers should ideally be stored in tightly closed, high-density, polyethylene (HDPE) containers, at sub-zero temperatures. For solution polymerizations, attention should be paid to the inhibitor contained within commercial solvents which can often be nucleophilic in nature [
46]. Similarly, any non-solvents used to isolate poly(CAs) by precipitation should be inhibited with strong acid (e.g., MeSO
3H) to prevent anionic polymerization of residual monomers [
58].
3.3. Mechanism and Kinetics of Radical Homopolymerization
Various common azo- and peroxide radical initiators have been employed to initiate CA polymerization. AIBN and benzoyl peroxide (BPO) are the most frequently used, with the former being preferred. When BPO is used as a radical initiator, both benzoyloxy and phenyl radicals participate in initiation as the benzoyloxy radical decomposes to give a phenyl radical through loss of carbon dioxide (
Scheme 9). Methyl 2-cyanoacrylate showed low reactivity towards the benzoyloxy radical, with initiation occurring largely through the addition of the more nucleophilic phenyl radical, as detected using doubly isotopically labelled BPO with carbon-14 and hydrogen-3 (tritium, T) [
56]. For the polymerization at 60 °C, the number of phenyl α-end groups on the polymer was found to be approximately twice that of benzoyloxy end groups.
Chappelow et al investigated the use of tri-
n-butyl borane oxide (TBBO) as the radical initiator for several highly reactive monomers: ethyl 2-isocyanatoacrylate, 2-isocyanatoethyl methacrylate, and ethyl 2-cyanoacrylate [
60]. The initiating radical was generated from the auto-oxidation of the alkylboron compound with air to give the alkyl borane peroxide and butyl radicals [
61]. Homopolymerizations were carried out using 1.8 wt% of TBBO in THF at room temperature. After 21 h, poly(ethyl 2-cyanoacrylate) was isolated by precipitation into hexane at greater than 70% conversion, however, polymer formation via anionic polymerization cannot be ruled out since no anionic inhibitor was employed.
The earliest report on the kinetics for the radical polymerization of CAs was by Canale et al. in 1960, who carried out polymerizations of methyl 2-cyanoacrylate using a boron trifluoride acetic acid complex as an anionic inhibitor and AIBN as the initiator [
48]. The bulk parameter,
kp/
kt0.5, (where,
kp and
kt are the rate coefficients for propagation and termination, respectively) was determined to be 0.021 L·mol
−1·s
−1 at 60 °C. This was significantly greater than the
kp/
kt0.5 = 0.0144 L·mol
−1·s
−1 for MMA and
kp/
kt0.5 = 0.00166 L·mol
−1·s
−1 for styrene (St), measured in the same study, under the same conditions. Poly(methyl 2-cyanoacrylate) was found to be insoluble in common aromatic solvents, such as benzene and toluene. The homopolymer was also reported to be insoluble in alcohols, ketones (acetone, 2-butanone) and chlorinated solvents (chloroform, 1,2-dichloroethane).
Bevington et al. used 1,3-propanesultone (0.3 wt%) as an inhibitor in the radical polymerization of methyl 2-cyanoacrylate in bulk and in solutions of 1,4-dioxane at 60 °C using AIBN and BPO as initiators [
57]. The inhibitor was reported to have negligible effects on the rate of polymerization at 60 °C, even at high concentrations, above 2.6 × 10
−2 M. Although the polymer was found to be insoluble in dioxane, it was deemed an acceptable diluent for the polymerization at 60 °C. In a more recent study, looking at the preparation of highly branched CA containing polymers, the homopolymerization of ethyl 2-cyanoacrylate in two different solvents at 65 °C was carried out [
55]. The polymerization was 50% faster in toluene than in acetonitrile due to an apparent poorer solubility of the polymer in the former, less polar solvent.
In 1983 Yamada et al. reported the first determination of absolute
kp and
kt for the radical polymerization of ethyl 2-cyanoacrylate using the rotating sector method [
51]. The rotating sector method is so named as it refers to a rotating disc from which a sector-shaped portion is cut out; the disc is placed in between the reaction system and a non-laser light source to cause periodic interruption of light [
62]. The cycling of light and dark periods allowed the parameter τ
s, known as the average lifetime of a growing radical under steady-state conditions, to be calculated. Once τ
s is known, from equation 1,
kp and
kt values can be determined.
The values of
kp and
kt were determined for the bulk polymerization of ethyl 2-cyanoacrylate at 30 °C in the presence of two different anionic inhibitors. Using 7.0 wt% acetic acid,
kp = 1622 L·mol
−1·s
−1 and
kt = 4.11 × 10
8 L·mol
−1·s
−1, and similar values of
kp = 1610 L·mol
−1·s
−1 and
kt = 4.04 × 10
8 L·mol
−1·s
−1 were obtained when using 0.5 wt% 1,3-propanesultone. The close proximity of these values indicated that anionic polymerization was adequately suppressed by both inhibitors, and the rate coefficients were representative of the radical polymerization. The
kp values for ethyl 2-cyanoacrylate were close to those determined at 30 °C using the same method for ethyl 2-chloroacrylate and were attributed to the nitrile group attached to the α-carbon of the double bond having a similar stabilization effect on the propagating radical as the chlorine substituent at the same position [
52].
Table 3 gives a comparison of experimentally measured
kp values using the rotating sector method. The
kp values for ethyl 2-cyanoacrylate are over five times higher than those measured for MMA and
n-butyl methacrylate (
n-BuMA) (
Table 4). Yamada’s determination of
kp for the polymerization of ethyl 2-cyanoacrylate is in agreement with data derived from more recent density functional theory (DFT) computational modelling [
63].
In 2015, the
kp for
n-butyl 2-cyanoacrylate was estimated to be about seven times lower (
kp = 226 ± 32 L·mol
−1·s
−1) than that reported by Yamada et al for ethyl 2-cyanoacrylate [
54]. The
kp was obtained by extrapolation of experimental data for the copolymerization of the CA with MMA using the more accurate pulsed-laser polymerization coupled with size exclusion chromatography (PLP-SEC) technique. PLP uses pulsed laser irradiation to oscillate between the light and dark periods; however, the pulse width is greatly shorter (nanoseconds) compared to the cycle time of the rotating sector method (seconds). The key advantage of PLP is that it allows for
kp to be calculated directly without the need to couple it to
kt, and has led to greater reproducibility of kinetic results; compared to the discrepancies in experimental data observed for the rotating sector method. Consequently, PLP is now the International Union of Pure and Applied Chemistry (IUPAC) preferred method for
kp determination [
64].
Table 4 gives a comparison of experimentally measured
kp for
n-butyl 2-cyanoacrylate and other common monomers using PLP, and the reader is directed to benchmark Arrhenius parameters in authoritative reports [
64,
65]. It is difficult to draw direct comparisons between
kp values obtained from the rotating sector method and those obtained by PLP, since variations in experimentally reported
kp values from different studies can reflect differences in data interpretation and the dependence of kinetic parameters on polymerization conditions. The discrepancy in the
kp measured for the two different CA monomers by two different techniques is, however, significant, but, it is worth highlighting that the value obtained by Rooney et al. is an extrapolation rather than a direct PLP measurement [
54].
As with radical polymerization of methacrylate monomers, termination in CA polymerizations is thought to predominantly occur by disproportionation. The hydrogen-atom terminated (non-functionalized) polymer chains are susceptible to the same degradation by base-catalyzed unzipping that occurs with anionically formed poly(CAs), where the chains with the unsaturated terminus do not unzip and are more stable [
45,
46]. Consequently, radically formed poly(CAs) exhibit greater stability than those formed by anionic polymerization. This improved stability was noted by Robello et al. during degradation studies of poly(CAs) [
46]. Samples of poly(CA) derived from nucleophilic initiation completely degraded when incubated in acetonitrile at 50 °C for 24 h whereas only a small portion of samples derived from radical polymerization degraded under the same conditions. Manipulation of GPC data using a Gaussian numerical algorithm function for peak fitting enabled deconvolution of the GPC curves, which allowed an approximation that 80% of these polymer samples were terminated through disproportionation and 20% by combination.
3.4. Mechanism and Kinetics of Radical Copolymerization
One key benefit of radical polymerization, which is difficult to achieve through anionic means, is the ease of copolymerization with other monomers, generating polymers with unique and varying properties. CAs form alternating head-to-tail copolymers when radically copolymerized with electron-rich monomers, such as vinyl ethers (
Scheme 10) [
69,
70].
For a vinyl monomer, M
1, being copolymerized with monomer M
2, the reactivity ratios, r
1 and r
2, represent the tendency of the propagating chain-end species to self-propagate with its own monomer over that of the other monomer. Although it has been met with criticism, historically, the
Q and
e scheme has been used as a measure of monomer and radical reactivity on a quantitative basis, in terms of correlating structure with reactivity, and can be used to predict reactivity ratios [
71,
72]. The
Q and
e values are related to the extent of resonance stabilization and polarity of the monomer and are approximated from its copolymerization kinetic data, whereby the parameter
Q describes the resonance factor (and to a certain degree the steric factor) of the monomer and the parameter
e describes the polar factor.
Q and
e values assigned to monomers are correlated against the arbitrarily selected reference values of
Q = 1 and
e = −0.80 for St. The values of
Q and
e increase with increasing monomer reactivity and increasing electron deficiency of the carbon–carbon double bond. Negative values of
e indicate an electron-rich double bond. The values for methyl 2-cyanoacrylate were reported by Otsu and Yamada as
Q = 17 and
e = 2.48 from the bulk copolymerization with MMA in the presence of acetic acid in a sealed tube at 60 °C using AIBN as the initiator [
50]. The authors noted that the monomers showed a high tendency for alternation, with rates of copolymerization decreasing as the CA monomer concentration increased. The modes of derivation for
Q and
e values without the necessity for an arbitrary assignment or without equating the polarities of conjugate monomers and radicals have led to improved methods of assessment for reactivity ratios [
73,
74]. Recent values of
Q and
e for methyl 2-cyanoacrylate have been reported as 4.91 and 0.91, respectively [
75]. The values of
Q and
e for some common monomers are shown in
Table 5.
Kinsinger et al. examined the radical copolymerization of methyl 2-cyanoacrylate with a wide variety of monomers [
76]. The reactivity ratios for methyl 2-cyanoacrylate (M
1) with a number of reference monomers in benzene at 60 °C using AIBN as the initiator were determined, the most important of which was that of MMA (M
2), reported as r
1 = 0.25 and r
2 = 0.04, given that MMA has been one of the more frequently studied co-monomers with CAs. In 2016, the reactivity ratios of ethyl 2-cyanoacrylate and MMA were simulated on PROCOP software by Tang and Tsarevsky, based on their experimental copolymerization data in toluene at 65 °C [
55]. The terminal kinetic model was used, and the theoretical kinetic curves were fitted to the experimental data. The values reported were r
1 = 0.15 and r
2 = 0.02, which are not too dissimilar to those reported by Kinsinger. Over the years, a series of reactivity ratios for CAs with a variety of monomers have been reported, which are summarized in
Table 6.
Han and Kim carried out extensive studies on the stability and degradation of copolymers from the radical copolymerization of ethyl 2-cyanoacrylate and MMA using AIBN at 60 °C with MeSO
3H as an anionic inhibitor [
47]. These copolymers were found to be random in nature with a strong alternating tendency. Due to this alternation in the copolymer system, the inclusion of MMA units in the polymer backbone close to the chain terminus was believed to be the reason for suppressing the unzipping degradation of the polymer. The copolymers were found to have increasing thermal stability with increasing MMA content.
Like CAs, methylene esters of malonic acid are a group of electron deficient monomers which can be polymerized both anionically and radically. Polyakova et al. studied the anionic and radical copolymerization of a series of alkyl and fluoroalkyl methylene malonates with ethyl 2-cyanoacrylate using a molar feed of 5–50% [
77]. Bulk anionic copolymerizations were carried out using a
tert-BuLi as the initiator at 20 °C, whereas bulk radical copolymerizations were carried out using dicyclohexyl peroxydicarbonate (DCHPC) and BPO at 40 °C and 60 °C, respectively, to generate random copolymers (
Scheme 11). The fluoroalkyl methylene malonates were observed to have the highest reactivity toward ethyl 2-cyanoacrylate, as determined by elemental analysis. The mechanical and chemical properties of cured poly(CA) based adhesive formulations containing 10% of alkyl and fluoroalkyl methylene malonates were tested. Formulations containing fluoroalkyl methylene malonate were found to have improved thermal and hydrolytic stability, compared to the control formulation based on ethyl 2-cyanoacrylate alone.
Kinsinger et al. detailed a lack of success in attempting to copolymerize 1-octene with methyl 2-cyanoacrylate, which largely resulted in CA homopolymer being formed [
76]. Sperlich and Eisenbach overcame this issue by employing CF
3CO
2H or ZnCl
2.OEt
2 as complexing agents in the copolymerization of ethylene with ethyl 2-cyanoacrylate to form alternating copolymers [
78]. Copolymerizations were carried out in a pressure reactor using DCHPC as initiator, with reaction components added together at −80 °C before being heated to 40 °C and maintained at that temperature for 15 h. The concentration of complexing agent was found to have a significant influence on the degree of alternation of the copolymer.
Dikov et al. reported on the radical graft copolymerization of ethyl 2-cyanoacrylate onto poly(butadiene-
co-AN) using bulk polymerizations at 60–80 °C in the presence of up to 2 wt% of the dissolved polymer with BPO as the initiator and TsOH (1 wt%) as an anionic inhibitor [
79]. Graft polymerizations are useful as they allow combination of two polymers which would otherwise be incompatible. Solution copolymerizations were also carried out in toluene at 80–90 °C with up to 75 wt% poly(butadiene-
co-AN) present. IR spectra of the resulting polymers gave superposition of the individual polymer spectra and were considered a good indication of a successful graft copolymerization.
Polymers can act as piezoelectric materials whereby polymers’ functionality can repel or attract each other when an electrical field is applied. Hall Jr et al. synthesized a large number of nitrile containing copolymers and studied their piezoelectric properties in terms of their pyroelectric coefficients [
80]. Copolymers of CA with vinyl acetate (VAc) and isopropenyl acetate were prepared using AIBN in benzene at 60–65 °C with 7 wt% of acetic acid. A copolymer of ethyl 2-cyanoacrylate and VAc gave a value of
p = 6.7 μC·m
−2·K
−1, one of the highest values recorded in the study, in contrast to the value of
p = 1.94 μC·m
−2·K
−1 measured for a copolymer of acrylonitrile (AN) and VAc.
When CAs are added directly to certain electron-rich vinyl monomers, spontaneous copolymerization can occur without the requirement of an external initiator, even at room temperature—this is believed to occur through the formation of diradical intermediates (
Scheme 12) [
81,
82]. After the formation of the intermediate diradical species, the electron-deficient CA radical rapidly adds onto another St monomer to form a new benzylic radical, which, in turn, further reacts with an CA monomer, and an alternating copolymer is generated.
3.6. Nanoparticles
Colloidal nanosystems for drug delivery alter the pharmacokinetics of numerous drugs, which can often improve their therapeutic indexes [
83]. These colloids are usually spherical, submicron in size, stable in bodily fluids, biodegradable, and have no systematic toxicity or immune response. Ideally, these nanoparticles can be loaded with various drugs and targeted to a specific location in the body, allowing drugs to be delivered to certain organs or cells but not to others. Site-specific targeting gives increased drug concentration to infected or abnormal cells and low concentration to normal cells, thus decreasing drug toxicity and undesirable side effects. Many synthetic polymers, such as poly(lactic acid) and poly(caprolactone), have found utility in the preparation of various drug nanocarriers, with poly(CA) becoming an established colloidal delivery system, since its introduction almost 40 years ago [
84]. Poly(CA) is a favoured nanocarrier due to its ease of polymerization, biodegradability, and relatively low toxicity. Broadly speaking, polymeric nanoparticles used for drug delivery are either solid particles or hollow particles, and can be prepared by a wide variety of methods. In the case of solid particles, the drug would typically be distributed throughout the particle or located at/near the interface. In the case of hollow particles (capsules), the drug is normally solubilized in a liquid core of either water or oil, surrounded by a polymer shell (
Figure 2).
Poly(CA) nanospheres are classically prepared by anionic emulsion or mini-emulsion polymerizations of the monomer in an acidic aqueous medium, typically of pH ≈ 3, containing a surfactant as a colloidal stabilizing agent. This relatively simple process was first reported by Couvreur et al. in 1979 [
84], and numerous studies have investigated the polymerization parameters and kinetics [
85,
86,
87]. Alternatively, poly(CA) nanospheres can be produced by radical polymerization using traditional radical initiators, such as AIBN, though at a very low pH of close to 1 [
88]. These radical mini-emulsion polymerizations have the advantage of achieving high molecular weight poly(CAs), whereas, in the case of anionic polymerization, the molecular weight is strongly dependent on the pH of the medium and typically results in polymers below 8000 g·mol
−1. Typically, nanoparticles have a hydrophilic shell based on poly(ethylene oxide, PEO) and hydrophobic core, based on poly(
n-butyl 2-cyanoacrylate). [
88,
89] PEO can not only impart hydrophilicity on polymers, but modify toxicity (see below).
Chauvierre et al. employed redox radical polymerization using cerium ammonium nitrate, (NH₄)₂Ce(NO₃)₆, (CAN) in HNO
3 to initiate polymerization in the presence of various polysaccharides, such as dextran, heparin or chitosan [
90]. Initiation and propagation of isobutylcyanoacrylate (
i-BuCA) occurred as the monomer was added to the polysaccharide and CAN mixture to form the polysaccharide-poly(
i-BuCA) copolymer (
Scheme 14). The rate of polymerization was found to be high due to the fast initiation step. The high radical initiation rate at a low pH renders any anionic polymerization negligible within the timescale of the reaction and allows radical propagation to predominate. This initiation process has since been applied to emulsion polymerizations in the preparation of poly(CA)-based nanospheres with various polysaccharides on the surface [
91,
92].
Poly(CA) nanocapsules are typically prepared by anionic interfacial polymerization techniques in water-in-oil or oil-in-water emulsion systems [
93,
94]; however, nanocapsules can also be formed by nanoprecipitation. Based on the nanoprecipitation method, preformed polymers (such as those made by radical polymerization) are dissolved in an organic solvent, typically acetone, and then added dropwise to an aqueous solution of surfactant where self-assembly gives nanocapsules [
95]. Using this approach, poly[α-maleic anhydride-ω-methoxypoly(ethylene glycol, PEG)-
co-EtCA] copolymers were prepared by radical solution copolymerization of a PEG macromonomer with ethyl 2-cyanoacrylate at 60 °C using AIBN (
Scheme 15) [
96]. The polymer formed by precipitation was added to acetone containing the drug (ibuprofen), which successfully self-assembled upon addition to water to encapsulate the drug (
Figure 3) [
97]. PEGylated particles (also termed “stealth” nanoparticles) are of great significance as they can escape immuno-recognition to give long-circulating drug delivery vehicles [
98,
99].
Nanoparticles based on CAs are considered promising polymer colloidal drug delivery systems, and have been significantly studied for cancer therapy with certain formulations reaching Phases III in clinical trials [
100].
3.7. Controlled/Living Radical Polymerizations
There has been only one report on controlled/living radical polymerization of CAs, which used reversible addition-fragmentation chain transfer (RAFT) [
58]. RAFT is perhaps the most versatile reversible deactivation (controlled/living) radical polymerization technique for making precision polymers [
101]. Given that CA propagates via a tertiary radical, poly(MMA), a reasonably bulky macroRAFT agent was used to make the first block copolymers containing poly(CAs) (
Scheme 16). ACN (1,1’-azobis(cyclohexanenitrile)) and 1,3-propanesultone were used as the respective radical initiator and anionic stabilizer at 90–95 °C.
Controlled/living character in CA polymerizations were indicated by a shift in the molecular weight distributions (MWDs) to higher MW with increasing conversion, with the MWD remaining mono-modal and relatively narrow throughout. However, the refractive index (RI) GPC MWDs disguised an underlying problem, which was the inherent instability of the CA block copolymers through self-elimination of the RAFT end group to generate dithiobenzoic acid (
Scheme 17). The degradation of the RAFT end group, which was not apparent by RI detection of GPC, could be visualized by UV analysis of polymer samples. UV showed several low MW peaks that were absent in the RI data, which increased in intensity with increasing conversion.
It seems that the acidity of the methylene at the chain terminus is enhanced by the inductively electron-withdrawing nitrile and ester functionalities, leading to the formation of an unsaturated end group. The polymerization is then complicated by the many possible combinations and disproportionation reactions between propagating and/or dithiobenzoic acid-derived radicals, which account for the observed increase in the number of polymer chains obtained through analysis of GPC RI data. For all three CAs investigated (ethyl 2-cyanoacrylate, n-butyl 2-cyanoacrylate and phenylethyl 2-cyanoacrylate), loss of livingness was about 70% by the end of the polymerizations.